Kantian and Sidgwickian Ethics: The Cosmos of Duty Above and the Moral Law Within [Hardcover ed.] 1138498092, 9781138498099

Citation preview

Kantian and Sidgwickian Ethics

Immanuel Kant and Henry Sidgwick are towering figures in the history of moral philosophy. Kant’s views on ethics continue to be discussed and studied in detail not only in philosophy but also theology, political science, and legal theory. Meanwhile, Sidgwick is emerging as the philosopher within the utilitarian tradition who merits the same meticulous treatment that Kant receives. As champions of deontology and consequentialism respectively, Kant and Sidgwick disagree on many important issues. However, close examination reveals a surprising amount of consensus on various topics including moral psychology, moral epistemology, and moral theology. This book presents points of agreement and disagreement in the writings of these two giants of philosophical ethics. The chapters will stimulate discussions among moral theorists and historians of philosophy by applying cutting-edge scholarship on each philosopher to shed light on some of the more perplexing arguments and views of the other, and by uncovering and examining points of agreement between Sidgwick and Kant as possible grounds for greater convergence in contemporary moral philosophy. This is the first full-length volume to investigate Sidgwick and Kant side by side. It will be of major interest to researchers and advanced students working in moral philosophy and its history. Tyler Paytas is Research Fellow in the Dianoia Institute of Philosophy at Australian Catholic University. His articles have appeared in Journal of the History of Philosophy, Kantian Review, Phronesis, and Utilitas. He is the co-author (with Nicholas Baima) of Plato’s Pragmatism: Rethinking the Relationship between Ethics and Epistemology (Routledge, forthcoming). Tim Henning is Professor and Chair of Practical Philosophy and History of Philosophy at Universität Stuttgart. His articles have appeared in Ethics, Philosophy & Phenomenological Research, and The Philosophical Review. He is the author of From a Rational Point of View: How We Represent Subjective Perspectives in Practical Discourse.

Routledge Studies in Ethics and Moral Theory

Perspectives in Role Ethics Virtues, Reasons, and Obligation Edited by Tim Dare and Christine Swanton Self, Motivation, and Virtue Innovative Interdisciplinary Research Edited by Nancy E. Snow and Darcia Narvaez Morality in a Realistic Spirit Essays for Cora Diamond Edited by Andrew Gleeson and Craig Taylor Comparative Metaethics Neglected Perspectives on the Foundations of Morality Edited by Colin Marshall An Intersectional Feminist Theory of Moral Responsibility Michelle Ciurria The Principle of Double Effect A History and Philosophical Defense David Černý Apologies and Moral Repair Rights, Duties, and Corrective Justice Andrew I. Cohen Kantian and Sidgwickian Ethics The Cosmos of Duty Above and the Moral Law Within Edited by Tyler Paytas and Tim Henning For more information about this series, please visit: www.routledge.com/ Routledge-Studies-in-Ethics-and-Moral-Theory/book-series/SE0423

Kantian and Sidgwickian Ethics The Cosmos of Duty Above and the Moral Law Within Edited by Tyler Paytas and Tim Henning

First published 2020 by Routledge 52 Vanderbilt Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2020 Taylor & Francis The right of Tyler Paytas and Tim Henning to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Names: Paytas, Tyler, editor. | Henning, Tim, 1976– editor. Title: Kantian and Sidgwickian ethics : the cosmos of duty above and the moral law within / edited by Tyler Paytas and Tim Henning. Description: New York and London : Routledge, 2020. | Series: Routledge studies in ethics and moral theory | Includes bibliographical references and index. Identifiers: LCCN 2020015617 (print) | LCCN 2020015618 (ebook) | ISBN 9781138498099 (hardback) | ISBN 9781351016995 (ebook) Subjects: LCSH: Kant, Immanuel, 1724–1804. | Sidgwick, Henry, 1838–1900. | Ethics. Classification: LCC B2799.E8 K365 2020 (print) | LCC B2799. E8 (ebook) | DDC 170—dc23 LC record available at https://lccn.loc.gov/2020015617 LC ebook record available at https://lccn.loc.gov/2020015618 ISBN: 978-1-138-49809-9 (hbk) ISBN: 978-1-351-01699-5 (ebk) Typeset in Sabon by Apex CoVantage, LLC

Contents

Note on Referencesvii Introduction

1

TYLER PAYTAS AND TIM HENNING

PART I

From Theory to Practice11   1 Practical Ethics in Sidgwick and Kant

13

ANTHONY SKELTON

  2 Self-Defeating Moral Theories and Kant’s Doctrine of Right

40

TIM HENNING

PART II

Metaethics59   3 Sidgwick and Kant on Practical Knowledge and Rational Action

61

CARLA BAGNOLI

  4 Sidgwick’s Kantian Account of Moral Motivation

84

DAVID PHILLIPS

PART III

Moral Epistemology105  5 On Seeing What Is Right: Sidgwick, Kant, and Philosophical Intuitionism ROGER CRISP

107

vi  Contents   6 Peerless: Sidgwick, Kant, and Peer Disagreement

118

ROBERT SHAVER

PART IV

Freedom of Will139   7 Freedom and Happiness: Sidgwick’s Critique of Kant

141

PAUL GUYER

  8 Kant and Sidgwick on Freedom of Will, Morality, and Responsibility

163

MARIKO NAKANO-OKUNO

PART V

Ultimate Ends185   9 Aristotelian Eudaimonism and the Dualism of Practical Reason

187

SUKAINA HIRJI

10 Beneficent Governor of the Cosmos: Kant and Sidgwick on the Moral Necessity of God

210

TYLER PAYTAS

Acknowledgments245 List of Contributors246 Index249

Note on References

References to Kant’s writings refer to the standard edition of Kant’s works, Kants gesammelte Schriften (KGS), edited by the Royal Prussian (later German, then Berlin-Brandenburg) Academy of Sciences, 29 vols. (1900 Reimer/de Gruyter). This standard edition of Kant’s works is traditionally referred to as the Academy Edition. References are given according to volume and page number, with the exception of the Critique of Pure Reason, which is cited in terms of the A/B page numbers. Translations and abbreviations used in each chapter are listed in the respective bibliographies. Sidgwick’s Methods of Ethics, 7th ed. (1907) is abbreviated as ME. Earlier editions are abbreviated as ME1, ME2, etc. Sidgwick’s other works are referenced by year of publication.

Introduction Tyler Paytas and Tim Henning

In the preface to On What Matters, Derek Parfit refers to Sidgwick and Kant as his two masters—the historical figures who had the greatest influence on his ethical views (2011, xxxiii). This reverence is hardly insignificant given that Parfit is among the most brilliant philosophical minds of the past 100 years and thus highly influential in his own right. Parfit claims that Kant is the greatest moral philosopher since the ancient Greeks, adding the memorable remark that “in the cascading fireworks of a mere forty pages, Kant gives us more new and fruitful ideas than all the philosophers of several centuries” (2011, xxxiii, 183). Equally unreserved in praise of his other master, Parfit boldly claims that Sidgwick’s Methods of Ethics is “the best book on ethics ever written” (2011, xxxiii). While some may classify these remarks as overstatements, they are certainly not unfounded. Kant’s central importance in the history of ethics is undeniable; his rationalist, deontological approach to moral philosophy has spawned generations of disciples who have forcefully interpreted and championed Kantian doctrine (Rawls 1971; O’Neill 1975; Herman 1993; Baron 1995; Wood 1999). Further, Kant’s writings have inspired a distinctive and influential brand of metaethical constructivism (Korsgaard 1996; Reath 2006; Hill 2012), while also making a significant impact in applied areas such as medical ethics (Cohen 1986), animal ethics (Regan 1983), and law (Shiffrin 2014). Although Sidgwick’s affect on moral philosophy has not been as pervasive or consistent as Kant’s, his work was a strong influence on some of the most important ethical theorists of the twentieth century, including Moore (1903), Ross (1930), Rawls (1971), and Parfit (1984; 2011). Indeed, the Sidgwickian ideas permeating Parfit’s writings are partly responsible for a recent surge of interest in Sidgwick. Since the publication of On What Matters in 2011, several monographs on Sidgwick have been published by major presses (Phillips 2011; Nakano-Okuno 2011; de Lazari-Radek and Singer 2014; Crisp 2015), with others in the pipeline (Kagan ms; Skelton forthcoming). This is in addition to important volumes of broader scope that nonetheless contain extensive coverage of

2  Tyler Paytas and Tim Henning Sidgwick (Hurka 2014; Schultz 2017). The result of these developments has been an expansion of the paradigm for study of classical utilitarianism, from the duo of Bentham and Mill to the triumvirate of Bentham, Mill, and Sidgwick. Still, Parfit’s dual exaltation of Kant and Sidgwick may initially seem surprising given that the two philosophers champion ethical theories commonly seen as antithetical to each other (deontology and consequentialism). But this begins to look less puzzling in light of Parfit’s deeply felt desire to find convergence within moral philosophy. Like Sidgwick, Parfit was highly sensitive to the skeptical threat posed by genuine philosophical disagreement, which seems especially pervasive in ethics. If the most careful and reflective thinkers are in a state of perpetual disagreement over fundamental ethical principles, this casts doubt on our ability to grasp objective ethical facts and even whether such facts exist at all. Hence, philosophers such as Parfit, who believe in objective and knowable ethical truths, incur a philosophical burden of explaining why the apparent disagreement does not warrant skepticism. One strategy for avoiding this conclusion is to carefully study the arguments and insights of figures such as Kant and Sidgwick in order to better understand the respective theories and search for points of contact between them. The present volume is motivated in large part by this idea. At first blush, it might appear that comparative study of Kant and Sidgwick will only serve to deepen worries about irresolvable ethical disagreement. After all, Kant is not ambiguous regarding his disdain for teleological and consequentialist methods: “Woe to him who creeps through the serpent-windings of utilitarianism” (Kant 1996, MM 6:332). Though less vitriolic, Sidgwick is equally doubtful of deontological approaches, describing them as “to some extent doubtful and confused; and sometimes, even when clear, as merely dogmatic, unreasoned, and incoherent” (ME xvii). Moreover, Sidgwick offers sharp criticisms of some of Kant’s central arguments and views. For instance, in considering Kantian moral theology as a possible solution to the apparent rational conflict between prudence and morality, Sidgwick writes: I cannot fall back on the resource of thinking myself under a moral necessity to regard all my duties as if they were commandments of God, although not entitled to hold speculatively that any such Supreme Being really exists. I am so far from feeling bound to believe for purposes of practice what I see no ground for holding as a speculative truth, that I cannot even conceive the state of mind which these words seem to describe, except as a momentary half-willful irrationality, committed in a violent access of philosophic despair. (ME 507 n. 3)

Introduction  3 Other criticisms include the charge that Kant equivocates in his appeals to freedom as the cornerstone of morality (ME 511–516), as well as the allegation of a “paralogism” in Kant’s attempt to derive a principle of benevolence from the idea of humanity as an end in itself (ME 389–90). Not all of Sidgwick’s references to Kant were critical, however. After presenting his ethical axioms in Book III of the Methods, Sidgwick notes that his confidence in them is buoyed by the agreement of those moralists who have been “most in earnest” in their philosophical investigations. Along with Samuel Clarke, Sidgwick cites agreement from Kant, who he describes as “especially noted for his rigour in separating the purely rational element of the moral code” (ME 384–85). While acknowledging some disagreement in the details, Sidgwick views Kant as an important source of corroboration for his principle of Rational Benevolence as well as his axiom of Justice, which he takes to be a corollary of the Universal Law formulation of Kant’s Categorical Imperative (ME 386). Sidgwick also expresses a feeling of kinship with Kant regarding the philosophical desire for system and coherence and the resulting difficulties: “Oh how I sympathise with Kant! With his passionate yearning for synthesis and condemned by his reason to criticism” (Sidgwick and Sidgwick 1906, 177). Although Kant was hostile toward utilitarianism, there is reason to think that he would have appreciated several aspects of Sidgwickian ethics (as compared to the systems put forth by Bentham and Mill). One respect in which Sidgwick differs from his utilitarian predecessors is that he recognizes the need for a rational intuition that can provide a secure foundation for a principle of universal benevolence. Appeals to emotion or the heroic nature of self-sacrifice were not enough for Sidgwick. He memorably expresses this concern in an autobiographical note from the preface to the sixth edition of the Methods: It was no use to say that if I was a moral hero I should have formed a habit of willing actions beneficial to others which would remain in force, even with my own pleasure in the other scale. I knew that at any rate I was not the kind of moral hero who does this without reason; from blind habit. Nor did I even wish to be that kind of hero; for it seemed to me that that kind of hero, however admirable, was certainly not a philosopher. I must somehow see that it was right for me to sacrifice my happiness for the good of the whole of which I am a part. (ME xviii) The preference for rationalism over sentimentalism is one of the factors that seemingly brings Sidgwick closer to Kant than initial appearances suggest. Another point of contact between the two is their recognition of

4  Tyler Paytas and Tim Henning the philosophical significance of conflicts between duty and self-interest. Kant and Sidgwick are generally in agreement that human happiness consists in agreeable states of consciousness, and that there is no reliable connection (at least in the earthly realm) between adhering to duty and partaking in such happiness. That said, the two philosophers draw divergent conclusions from the apparent inevitability of moral-prudential conflicts. Whereas Kant believes that the rational authority of morality could not be undermined by conflicts with self-interest, Sidgwick sees an unresolved “dualism of the practical reason” that significantly threatens the enterprise of philosophical ethics (more on this below). In an essay on Kant’s Critical Philosophy, Sidgwick confesses that “I have often longed to call up the philosopher of Königsberg himself,” to ask about interpretations of his views (1883, 71). Although this was an idle wish, we may perhaps obtain some of the fruits of this imagined encounter by exploring the work of both philosophers together. Such comparative historical studies are useful not merely as a means of seeking philosophical convergence but also for clarifying and strengthening our own thinking on the relevant topics. As C. D. Broad explains in the introduction to his classic work Five Types of Ethical Theory: The best preparation for original work on any philosophic problem is to study the solutions which have been proposed for it by men of genius whose views differ from each other as much as possible. The clash of their opinions may strike a light which will enable us to avoid the mistakes into which they have fallen; and by noticing the strong and weak points of each theory we may discover the direction in which further progress can be made. (Broad 1930, 1–2) In this spirit, the present volume aims to advance both contemporary ethical thought and our understanding of Kant and Sidgwick by using the work of each philosopher to illuminate that of the other. In addition to the question of the extent to which common ground between deontological and consequentialist approaches is possible, the juxtaposition of Kant and Sidgwick will shed light on several important topics including free will, moral epistemology, and moral theology. The volume begins in Part I with discussions of Kant’s and Sidgwick’s views on normative, practical ethics, and of possible connections and divergences between them. Anthony Skelton’s chapter offers an overview of Kant’s and Sidgwick’s views on central issues in practical ethics. Skelton focuses especially on their contrasting positions on the morality of lying and the demandingness of the duty of beneficence. In addition, he investigates how far the two authors can claim to be in general agreement with the morality of common sense. This last point is of major importance because both Kant

Introduction  5 and Sidgwick hold metaethical views that lead one to expect that divergence from common sense should be a rare exception rather than the norm. According to Skelton, it turns out that Kant and Sidgwick do disagree with each other and with common sense on a number of counts. But all in all, he argues that Sidgwick’s approach is more promising when it comes to accommodating common-sense verdicts. Tim Henning’s contribution highlights a point of agreement between Sidgwick and Parfit, on the one hand, and Kant on the other. Among many other things, Sidgwick has become famous for his differentiated views on the question of when a criterion of rightness should serve as the motive for moral action. In particular, Sidgwick argues that the true criterion of rightness may in fact demand that some other consideration should be motivating rather than the criterion itself. Henning follows Parfit’s careful analysis of the ways in which such practical self-defeat can occur, and he suggests that Kant’s Categorical Imperative may be subject to a particular form of indirect self-defeat, much like Sidgwick and Parfit have claimed with regard to utilitarianism. Henning argues that this selfdefeat plays a major role in the basic distinctions and the overall architecture of Kant’s Doctrine of Right (Rechtslehre). Henning ends with the suggestion that this late Kantian work should be read as developing a form of ‘indirect deontology’ on the model of indirect consequentialism. Part II of the volume follows up with discussions of the topics of practical knowledge, rational deliberation, and moral motivation. Carla Bagnoli begins her chapter by emphasizing similarities in Sidgwick’s and Kant’s theoretical projects. As Bagnoli stresses, both authors reject a naturalistic outlook, according to which moral demands can be fully explained in terms of the nature of humans and their circumstances. Likewise, both Kant and Sidgwick reject certain forms of intuitionism. Against naturalistic and (some) intuitionist views, they insist on the practical force of moral judgments, as directly capable of binding the rational subject’s will. At the same time, however, both philosophers reject the idea that moral judgments are simply expressions of antecedently existing motivating states of the subjects, thus making moral judgments depend on contingent psychologies. After Bagnoli carefully lays out how Sidgwick’s version of intuitionism is supposed to live up to these requirements, she suggests that Kant’s view ultimately holds greater promise of success. This, Bagnoli argues, is because Kant conceives of moral knowledge as a kind of practical knowledge, i.e., as a knowledge of principles that arises directly out of the nature of practical decision making. David Phillips’s contribution explores an important line of argument in Sidgwick’s metaethics. Sidgwick endorses the non-naturalist view that terms like ‘ought’ cannot be paraphrased in naturalistic terms (e.g., in terms of desire-fulfillment), and he also holds the Kantian view that reason alone can motivate. Now, these two views are presented by Sidgwick in a way that suggests that there is an argument linking them. Phillips

6  Tyler Paytas and Tim Henning discusses whether there is such a connecting argument and, if so, what this argument is. He begins by arguing, contra J. Schneewind (1977), that there is indeed an argument linking the non-naturalness of ‘ought’ and the Kantian (or anti-Humean) view of moral motivation. Phillips then argues against one reconstruction (proposed by, among others, R. Shaver (2000)) of this argument. According to this reading, Humeans need to deny non-naturalness. In order to explain away apparent conflicts between reason and passion, they need to re-interpret the verdicts of reason as being concerned with prosaic facts about desires after all. Phillips rejects this reconstruction and offers a different interpretation. On Phillips’s reading, it is Kantian accounts of motivation that are committed to non-naturalness. This is because the Kantian view of moral motivation receives its intuitive support from the appearance that ‘ought’-judgments play a distinctive role in motivation. And this appearance could not be vindicated if such judgments could be paraphrased away by statements expressing or reporting ordinary, Humean motivation. After reconstructing the argument, Phillips then shows that Sidgwick’s positive account of moral motivation very closely resembles that of a contemporary nonnaturalist, T. M. Scanlon (1998; 2014). Part III of the volume offers detailed discussions of Sidgwick’s and Kant’s relation to an intuitionist moral epistemology, and of the significance of peer disagreement in moral matters—including apparent disagreements between Kant and Sidgwick. Roger Crisp’s contribution discusses the question of whether there is a fundamental disagreement between Sidgwick’s and Kant’s moral epistemologies. First, Crisp discusses the character of Sidgwick’s philosophical intuitionism, contrasting it with perceptional and dogmatic intuitionism. Famously, Sidgwick himself considered Kant an ally in this epistemological approach—a claim that has been met with much criticism. Crisp, however, replies to these criticisms and argues that considering Kant as a philosophical intuitionist may be tenable after all. One apparent difference between Sidgwick and Kant (which is also discussed in Crisp’s chapter) concerns the significance of dissensus or disagreement. Sidgwick is often cited as an adherent of a conciliatory view regarding peer disagreement. Specifically, he is thought to have held that in cases of peer disagreement, suspension of judgment is called for. Robert Shaver carefully reconstructs Sidgwick’s views, and he proposes an interpretation according to which peerhood is not the default assumption. While Sidgwick includes among one’s epistemic peers all those for whom one has no more reason to suspect of error than oneself, Shaver proposes a view according to which, in a case of disagreement, peerhood has to be established by positive argument. Shaver then goes on to argue that this modified interpretation fits well with Sidgwick’s discussion of intertemporal disagreement in the history of thought. Importantly,

Introduction  7 Shaver concludes that Sidgwick need not have regarded Kant as his epistemic peer in the domain of philosophical ethics. The topic of Part IV is the ethical significance of free will. Whereas Kant famously views freedom as central to morality, Sidgwick argues that the question of free will is ultimately insignificant for ethics. In Sidgwick’s view, when faced with a choice between alternative modes of conduct, the question of whether all of our decisions and actions are ultimately determined by antecedent causal forces is irrelevant. Sidgwick also accuses Kant of equivocating between two distinct notions of freedom— one which is manifest whenever we are guided by reason rather than impulse, and another which is exhibited when we exercise the capacity to choose between right and wrong regardless of which one we choose (ME 511–512). In her chapter, Mariko Nakano-Okuno argues that while Sidgwick’s accusation of equivocation was off base, his skepticism about the practical significance of free will was largely on target. Nakano-Okuno analyzes and defends Sidgwick’s argument for the claim that our capacity to act from rational principles, or in Kantian terms, to set universal laws for ourselves, does not require the assumption of our will’s complete freedom from antecedent biological and environmental causes. Further, while moral emotions such as guilt and remorse may lose their motivational force in the absence of belief in freedom, we can still be motivated to act rightly by the powers of imagination, sympathy, and love of goodness. Regarding the issue of holding others responsible, Nakano-Okuno defends the Sidgwickian idea that prospectively oriented practices of censure and punishment, which do not require the assumption of free will, can be every bit as effective as retributive concepts and practices that do presuppose free will. Nakano-Okuno optimistically concludes that if Kant were to jettison his commitment to the moral significance of freedom, then there would be no insurmountable discrepancy between his ethical system and that of Sidgwick. In his contribution, Paul Guyer agrees that freedom is the fundamental source of divergence between Kant and Sidgwick. Indeed, Guyer suggests that the disagreement between Kant and Sidgwick over the moral significance of freedom is central to the general conflict between utilitarianism and deontological ethics, which he sees as largely concerning the derivation and scope of the duty of benevolence. However, Guyer does not see the downplaying of freedom as a live option for Kantian ethics. Guyer’s central claim is that Sidgwick fails to appreciate the merits of Kant’s having derived the duty of benevolence not from the value of happiness but rather from the value of freedom. This is a crucial point insofar as the freedom-based conception of morality entails constraints on the promotion of happiness that constitute the core tension between deontological and utilitarian approaches.

8  Tyler Paytas and Tim Henning The promotion of happiness is also pertinent to the final part of the volume (Part V), which addresses Kant’s and Sidgwick’s respective views on ultimate ends. As a consequentialist, Sidgwick conceives of ethics as largely a matter of determining the ultimate end of reasonable conduct. And while Kant believes that morality is composed of universal rational principles that apply irrespective of consequences, he acknowledges that human practical reason has a deep need to posit an ultimate rational end (1998, Rel 6:5). Both philosophers recognize a difficulty arising from the idea of an ultimate rational end, which is that such an end would seemingly need to contain two elements that are not reliably c­ onnected—duty (which for Sidgwick is a matter of impartial benevolence) and one’s own happiness. The fact that dutiful conduct seems to at least occasionally require sacrifice of one’s own happiness is perhaps the most fundamental problem for philosophical ethics (see Shaver 1999; White 2002; Crisp 2019). One way of responding to this problem is to claim that the appearance of moral-prudential conflicts is illusory. This is the approach taken by ancient eudaimonist theories, which hold that the most morally virtuous life is also the happiest. However, Kant and Sidgwick are both highly skeptical of this approach. Among their chief objections are that eudaimonist views either posit implausible conceptions of happiness or they are objectionably egoistic. These critiques of eudaimonism are the subject of Sukaina Hirji’s contribution. After outlining and analyzing Kant’s and Sidgwick’s objections, Hirji argues that there are resources available within the eudaimonist tradition to overcome them. Hirji focuses specifically on Aristotelian eudaimonism, which is the version she takes to be best equipped to meet the challenges pressed by Kant and Sidgwick. She advocates an interpretation of Aristotle involving a version of “weak eudaimonism,” in which acting virtuously is a matter of performing right actions with the appropriate reasons and desires, and that acting in this manner is constitutive of human happiness. Hirji argues that this interpretation allows us to see why the Aristotelian agent need not be objectionably egoistic. As to whether the proposed conception of happiness is sufficiently plausible, Hirji provides some considerations in defense of Aristotle while acknowledging that the disagreement over the nature of human happiness is not easily resolvable. One of the puzzling similarities between Kant and Sidgwick is that, despite their mutual advocacy of secular ethics (as opposed to divine command approaches), they both conclude that belief in God is morally necessary. In his chapter, Tyler Paytas argues that this appeal to God is related to their shared views about the idea of an ultimate rational end. As mentioned earlier, both philosophers recognize that one’s own happiness must be included in any plausible conception of an ultimate rational end, but also that apparent moral-prudential conflicts raise serious difficulties on this score. Belief in a divine being who aligns duty and

Introduction  9 self-interest with post-mortem rewards is one way out of this predicament. While Kant appeals to these considerations as practical grounds for faith, Sidgwick expresses skepticism. However, Paytas argues that Sidgwick’s considered position is more similar to Kant’s than initial appearances suggest. Paytas argues further that the predicament that leads Kant and Sidgwick to their need for faith can be avoided by abandoning their hedonistic accounts of well-being. Kant’s Critique of Practical Reason ends on a memorable note of inspiration: “Two things fill the mind with ever new and increasing admiration and reverence, the more often and more steadily one reflects on them: the starry heavens above me and the moral law within me” (1997, CpV 5:162). Sidgwick was less hopeful. At the conclusion of the first edition of the Methods, he somberly describes the implications of the dualism of practical reason: “The Cosmos of Duty is thus really reduced to a Chaos: and the prolonged effort of the human intellect to frame a perfect ideal of rational conduct is seen to have been foredoomed to inevitable failure” (ME1 473). In the preface to his Volume Three, Parfit reaffirms a qualified optimism that falls somewhere in between Kant and Sidgwick. Quoting his own earlier remarks concerning the skeptical challenge from ethical disagreement, Parfit writes: “Non-Religious ethics is at a very early stage. We cannot yet predict whether . . . we will all reach agreement. Since we cannot know how Ethics will develop, it is not irrational to have high hopes” (2017, xii; 1984, 454). Integrated study of Parfit’s two masters may result in progress toward the convergence he hoped to find. But even if it does not, it will undoubtedly improve our understanding of the most important disagreements so that we might continue our ethical investigations with greater clarity and sympathy.

References Baron, M. 1995. Kantian Ethics Almost Without Apology. Ithaca: Cornell University Press. Broad, C. D. 2013 [1930]. Five Types of Ethical Theory. New York: Routledge. Cohen, C. 1986. “The Case for the Use of Animals in Biomedical Research.” The New England Journal of Medicine 314: 865–869. Crisp, R. 2015. The Cosmos of Duty: Henry Sidgwick’s Methods of Ethics. Oxford: Oxford University Press. Crisp, R. 2019. Sacrifice Regained: Morality and Self-Interest in British Moral Philosophy from Hobbes to Bentham. Oxford: Oxford University Press. de Lazari-Radek, K., and P. Singer. 2014. The Point of View of the Universe: Sidgwick and Contemporary Ethics. Oxford: Oxford University Press. Herman, B. 1993. The Practice of Moral Judgment. Cambridge, MA: Harvard University Press. Hill, T. 2012. “Kantian Constructivism in Ethics.” Ethics 99 (4): 752–770. Hurka, T. 2014. British Ethical Theorists from Sidgwick to Ewing. Oxford: Oxford University Press.

10  Tyler Paytas and Tim Henning Kagan, S. Unpublish ms. Thinking Through Sidgwick. Kant, I. 1996. Metaphysics of Morals. Edited by Mary Gregor. Cambridge: Cambridge University Press. Kant, I. 1997. Critique of Practical Reason. Edited by Mary Gregor. Cambridge: Cambridge University Press. Kant, I. 1998. Religion Within the Boundaries of Mere Reason. Edited and translated by A. Wood and G. Di Giovanni. Cambridge: Cambridge University Press. Korsgaard, C. 1996. The Sources of Normativity. Cambridge: Cambridge University Press. Moore, G. E. 2003 [1903]. Principia Ethica. Cambridge: Cambridge University Press. Nakano-Okuno, M. 2011. Sidgwick and Contemporary Utilitarianism. Basingstoke: Palgrave Macmillan. O’Neill, O. 1975. Acting on Principle: An Essay on Kantian Ethics. New York: Columbia University Press. Parfit, D. 1984. Reasons and Persons. Oxford: Clarendon Press. Parfit, D. 2011. On What Matters, 2 vols. Oxford: Oxford University Press. Parfit, D. 2017. On What Matters, vol. 3. Oxford: Oxford University Press. Phillips, D. 2011. Sidgwickian Ethics. Oxford: Oxford University Press. Rawls, J. 1999 [1971]. A Theory of Justice. Cambridge, MA: Belknap Press of Harvard University Press. Reath, A. 2006. Agency and Autonomy in Kant’s Moral Theory. Oxford: Oxford University Press. Regan, T. 1983. The Case for Animal Rights. Berkeley: University of California Press. Ross, W. D. 1930. The Right and the Good. Oxford: Clarendon Press. Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scanlon, T. M. 2014. Being Realistic About Reasons. Oxford: Oxford University Press. Schneewind, J. B. 1977. Sidgwick’s Ethics and Victorian Moral Philosophy. Oxford: Clarendon Press. Schultz, B. 2017. The Happiness Philosophers: The Lives and Works of the Great Utilitarians. Princeton: Princeton University Press. Shaver, R. 1999. Rational Egoism: A Selective and Critical History. Cambridge: Cambridge University Press. Shaver, R. 2000. “Sidgwick’s Minimal Metaethics.” Utilitas 12: 261–277. Shiffrin, S. 2014. Speech Matters: Lying, Morality and the Law. Princeton: Princeton University Press. Sidgwick, A., and E. M. Sidgwick. 1906. Henry Sidgwick: A Memoir. London: Macmillan. Sidgwick, H. 1874. The Methods of Ethics. 1st ed. London: Macmillan. Sidgwick, H. 1883. “A Criticism of the Critical Philosophy.” Mind 8 (29): 69–91. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Skelton, A. forthcoming. Sidgwick’s Ethics. Cambridge: Cambridge University Press. White, N. 2002. Individual and Conflict in Greek Ethics. Oxford: Oxford University Press. Wood, A. 1999. Kant’s Ethical Thought. Cambridge: Cambridge University Press.

Part I

From Theory to Practice

1 Practical Ethics in Sidgwick and Kant Anthony Skelton

Sidgwick engaged both Kant’s theoretical and his practical philosophy. The effect of the engagement with Kant’s theoretical philosophy was not insignificant. Sidgwick opined that post-Kantian German metaphysics was “a monstrous mistake,” and that “we must go back to Kant and begin again from him” (Sidgwick and Sidgwick 1906, 151).1 Sidgwick found the study of Kant’s metaphysics edifying, declaring “I shall always look on him [Kant] as one of my teachers” (Sidgwick and Sidgwick 1906, 151; also 159) and that the study of Kant’s theoretical system provides “a most valuable metaphysical education” (1905, 30). He was, however, unwilling to call himself a Kantian in metaphysics (Sidgwick and Sidgwick 1906, 151). He described the Critique of Pure Reason as putting forward a “false system” (Sidgwick and Sidgwick 1906, 386) and Kant’s work in the book as “deficient in . . . self-criticism” (1883, 72). While Kant is, Sidgwick said, “one of the most original, penetrating, ingenious, and laboriously systematic of modern thinkers . . . he is a profoundly inconsistent thinker, profoundly unaware of his own inconsistency” (1905, 30–31). Because Sidgwick’s work on Kant’s metaphysics focuses mostly on these inconsistencies rather than on building a rival system with Kantian inspiration, it is unsurprising to find it described as “not of the first importance” (Passmore 1975, 195). Sidgwick’s engagement with Kant’s practical philosophy is, while critical, more central to his main research preoccupations and therefore likely of more enduring interest. He claimed Kant as one of his moral philosophical “masters” (ME xx) and he addresses many of Kant’s most important views in ethics. Sidgwick worked out a comprehensive rival ethical system in which he was clear about what he did and did not accept in Kant’s practical philosophy and why. Many of Sidgwick’s criticisms of Kant are of lasting value. Sidgwick’s objection that in his ethics Kant unwittingly relies to his peril on two distinct conceptions of freedom, good freedom and neutral freedom, is among the most noteworthy (ME 512).2

14  Anthony Skelton One is free on the first conception only when one is motivated to do the right thing because it is the right thing and not because of some desire one has for some end (e.g., KpV 5:78; G 4:446–447).3 But, Sidgwick argues, if one is free only when one acts from duty, one cannot be held responsible for wrong actions since such actions are never done from duty; they always involve motivation by inclination, in which case one is determined by sensible motives and therefore not free. One is free on the second conception when one has the power to choose right or wrong (e.g., KpV 5:98). Kant requires this kind of freedom, Sidgwick contends, in order to hold agents responsible for wrongdoing. The difficulty, for Kant, is that neutral freedom implies that one is free even when one acts wrongly, and if Kant admits this, he has to agree “the scoundrel must exhibit and express his characteristic self-hood in his transcendental choice of a bad life, as much as the saint does in his transcendental choice of a good one” (ME 516). In this case the “spiritstirring appeal to the sentiment of Liberty [that Kant gets by connecting freedom and motivation by reason] must be dismissed as idle rhetoric” (ME 516). Sidgwick’s objection is vexing to Kant’s admirers; accordingly, it has received detailed attention.4 Less sustained attention has been paid to Sidgwick’s other interactions with Kant’s ethics. Kant and Sidgwick are noteworthy for having written on both theoretical and practical ethics, making it possible to fruitfully compare and contrast them on a range of ethical issues. In this chapter, I explore the relationship between the two on some issues in practical ethics (broadly understood). In § I, I outline the main element of Kant’s theoretical ethics that Sidgwick endorsed. In §§ II and III, I outline and adjudicate two of their sharpest disagreements in practical ethics, on the permissibility of lying and on the demands of beneficence. In § IV, I argue that compared to Kant, Sidgwick has a better strategy for dealing with disagreement in practical ethics. §V sums things up.

I For Sidgwick, claiming someone as a master did not involve hagiography or signal agreement. Of his masters, Mill is perhaps the most prominent. Through Mill, Sidgwick became an adherent of utilitarianism, the view that the sole criterion of morality is the promotion of aggregate happiness (ME xvii; 1998, 8). Sidgwick nevertheless held a dubious opinion of Mill’s work in ethics. In an obituary he wrote for Mill, he remarks that Utilitarianism is “hastily put together, and the system seems incompletely reasoned and even incoherently expounded” (Sidgwick 1873, 193). Sidgwick was especially dissatisfied with Mill’s proof of utilitarianism. Mill argued that each individual’s happiness is alone good because she

Practical Ethics in Sidgwick and Kant  15 desires only it, from which he famously inferred that the sum or aggregate of each individual’s happiness is the only good and therefore what we ought to promote (Mill 1969, 235–239). Sidgwick rightly noted that there is a gap in this argument. From the fact that each individual’s happiness is good it does not follow that the sum of each individual’s happiness is good. This is because no desire for the sum of happiness is present in the aggregate. The gap in Mill’s argument would be closed, on Mill’s own reasoning, only if the aggregate desired its own happiness. But, Sidgwick complains, an aggregate of actual desires, each directed towards a different part of the general happiness, does not constitute an actual desire for the general happiness, existing in any individual; and Mill would certainly not contend that a desire which does not exist in any individual can possibly exist in an aggregate of individuals. There being therefore no actual desire . . . for the general happiness, the proposition that the general happiness is desirable cannot be in this way established. (ME 388) Sidgwick disagreed with Mill’s view that one can infer claims about what is valuable from claims about what is desired. The former argued that instead the justification of the claim that happiness is the only good “is properly . . . reached . . . by a more indirect mode of reasoning” (ME 389). The manner of reasoning involves, for Sidgwick, appeal to “intuitive judgment after due consideration” and “a comprehensive comparison of the ordinary judgements of mankind” (ME 400). The move from the claim that each individual’s happiness is good to the claim that the happiness of the aggregate is good and therefore what we ought to promote is justified, on Sidgwick’s view, only by appeal to the intuitively known self-evident proposition that one ought to aim at good generally not merely at a particular part of it, giving equal value to equal quantities of happiness (ME 382).5 Sidgwick called this the axiom or principle of rational benevolence.6 Although Sidgwick appealed to intuition to justify utilitarianism, and so rejected Mill’s empiricism, his appeal did not allay his most pressing concern. He was struck by the difficulty of reconciling duty and selfinterest. He fell under the sway of another master, Joseph Butler. Sidgwick rejected the anti-utilitarian elements of Butler’s theory of virtue and vice, thinking that together with a great many other elements of common-sense morality these could be incorporated into utilitarianism (ME xx, 423–459, 496). But he agreed with Butler in holding that there is a self-standing obligation to promote one’s own happiness that competes with other rational demands (ME 7, 119). In some cases, this obligation

16  Anthony Skelton conflicted in practice with utilitarianism and nothing he read in Mill convinced him of the rationality of sacrificing one’s own happiness in order to promote the aggregate’s happiness.7 He was stuck with a dualism of practical reason. Sidgwick turned to Kant, one of his other masters, for inspiration. He agreed with Kant that there are categorical requirements of reason applying to rational agents irrespective of their desires (ME 7, 35; G 4:414–416). He appears to agree with Kant that the recognition that some act is right arouses in rational agents a motivation to do it (ME 34; KpV 5:47ff.). Sidgwick was, he said, especially “impressed with the truth and importance of. . . [Kant’s] fundamental [categorical imperative] principle:—Act from a principle or maxim that you can will to be a universal law” (ME xix; italics in original; also 210; G 4:402, 421, 437; KpV 5:69–70). Sidgwick was not that impressed, however. He dissents from Kant’s principle in two main ways. First, he does not think Kant’s maxim is sufficient to cover the whole of morality. Thinking that “all particular rules of duty can be deduced from. . . [this] one fundamental rule . . . appears to me an error analogous to that of supposing that Formal Logic supplies a complete criterion of truth” (ME 209–210; also xix). The main function of this fundamental rule is to protect ourselves against the danger which besets the conscience, of being warped and perverted by strong desire, so that we too easily think that we ought to do what we very much wish to do. For if we ask ourselves whether we believe that any similar person in similar circumstances ought to perform the contemplated action, the question will often disperse the false appearance of rightness which our strong inclination has given to it. (ME 209; also 319, 380)8 Second, Sidgwick says he is keen on the idea that “whatever is right for me must be right for all persons in similar circumstances” (ME xix; also 209, 318). But he concedes that he does not accept this truth in the “precise form in which he [Kant] stated it” (ME 210n2; also 486). What he accepts is in fact a very slim principle: it cannot be right for A to treat B in a manner in which it would be wrong for B to treat A, merely on the ground that they are different individuals, and without there being any difference between the natures or circumstances of the two which can he stated as a reasonable ground for difference of treatment. (ME 380; also 209, 496) He takes this to be self-evident (ME 380; also 318). The practical upshot is that “an act, if right for any individual, must be right on

Practical Ethics in Sidgwick and Kant  17 general grounds . . . for some class of persons” (ME 486; italics in original). Sidgwick had, of course, turned to Kant for help in trying to deal with the fact that duty and self-interest are in practice often not coincident. As in the case of Mill, and unsurprisingly, the principle Sidgwick agreed to in Kant did not, Sidgwick felt, “meet” the difficulty that perplexed him (ME xx): “it did not settle finally the subordination of Self-Interest to Duty” (ME xix). On Sidgwick’s reckoning the egoist could easily will his own principle as a universal law for all rational beings: He might say, “I quite admit that when the painful necessity comes for another man to choose between his own happiness and the general happiness, he must as a reasonable being prefer his own, i.e. it is right for him to do this on my principle. No doubt, as I probably do not sympathise with him in particular any more than with any other persons, I as a disengaged spectator should like him to sacrifice himself to the general good: but I do not expect him to do it, any more than I should do it myself in his place.” (ME xx)9 Moreover, Sidgwick’s qualifications of Kant’s “truth” led to quite significant differences in many of their most important practical moral conclusions. The next two sections outline two of the sharpest of those disagreements.

II Kant and Sidgwick disagree sharply about the permissibility of lying. Kant holds that lying is always wrong because, on his own understanding of his principle, maxims on which, say, one makes statements one knows to be false in order to get oneself out of a jam, cannot be willed as universal laws (G 4:402–403, 4:422; KpV 5:69). One cannot will that one lie in order to convince someone to lend one money when there is no other way to get it. According to Kant, such a maxim is self-contradictory, for it is impossible to achieve the purpose of getting oneself out of a jam by lying if it is a universal law of nature that people lie in such cases. Sidgwick has a laxer approach to lying (ME 315–316, 448). He maintains that even when you cannot “Act as though the maxim of your action were to become a universal law of nature,” it does not follow that your action is wrong (ME 318). It might be the case that in one’s circumstances one’s maxim is not (or will not be) universally accepted, making it possible for one to produce more benefit by doing what cannot be willed as a universal law of nature. Sidgwick notes that it may appear that lying is ruled out by the “ethical axiom” that “what

18  Anthony Skelton is right for me must be right for ‘all persons under similar conditions’ ” (ME 318). But if the situation is such that one’s maxim is not or will not be universally adopted and one’s lie is all told beneficial, then it is permissible to lie by the criterion. Kant is, of course, unlikely to be moved by this (KpV 5:69–70). Sidgwick’s best tack is to point out that common sense holds that it is permissible, if enough good is at stake, to do something—lie—that cannot be willed as a universal law for all rational beings. He rightly points out that if we may even kill in defence of ourselves and others, it seems strange if we may not lie, if lying will defend us better against a palpable invasion of our rights: and Common Sense does not seem to prohibit this decisively. (ME 315) Common sense seems to concede in other cases that when enough good is at stake lying is permissible (ME 316). In which case, Kant’s view of lying might well be counterintuitive. Sidgwick, of course, has his own problems capturing common sense. He has to argue that in any case in which lying is all told beneficial one is permitted to lie. But Kant is likely right that it is wrong to lie in order to get someone to lend you money even if it is all told slightly more beneficial to lie. Common sense does not admit that if lying yields slightly more benefit than not, it is permissible to lie. Common sense might admit that lying is permissible only to prevent a “palpable invasion of our rights” or cognate invasion. The best Sidgwick can do to save his view is argue that indirect utilitarian considerations capture the commonsense view in its entirety (ME xxiii, 424ff.). But the possibility of this remains, at best, uncertain. Sidgwick might not see the force of Kant’s position on lying because for the most part he ignores the humanity formulation of the categorical imperative, requiring agents to “Act in such a way that you treat humanity, whether in your own person or in any other person, always at the same time as an end, never merely as a means” (G 4:429; also KpV 5:87). Lying involves using another as a mere means to an end because it involves bypassing their will to pursue an end that they do not (or cannot) share and so treating them (merely) like an instrument at your disposal. Kant claims that humanity or rational nature—the capacity for morality (MS 6:434)—is the objective value on which this formulation is based; the value of humanity constrains what it is permissible to do in the service of one’s goals (G 4:428–429; KpV 5:131–132). Kant thinks that rational nature is a “self-sufficient” end (G 4:437). It is possible that Sidgwick fails to contend with the humanity formulation because he misunderstands the nature of a self-sufficient or, as he calls it, a self-subsistent end (ME 380). Sidgwick thinks of an

Practical Ethics in Sidgwick and Kant  19 end as something to be promoted, e.g., pleasure or knowledge (ME 390; 1902, 274–5). He fails to see that for Kant the end of humanity or the capacity for morality is an end that constrains or limits what we may do to each other. Humanity is to be respected or honored rather than promoted. We are not permitted to use humanity as a mere means to an end, however worthy that end may be. This prevents lying since deceiving someone involves treating them as a mere means. Sidgwick might confess to failing to see the nature of a self-subsistent end while denying that this makes a difference to his rejection of Kant’s view on lying. Kant holds that the universal law formulation of the categorical imperative and the humanity formulation are in fact identical or that the one reduces to the other (G 4:436). The view is that when one wills one’s maxims as a universal law for all rational beings one is constraining one’s behavior by reference to what other agents can permissibly act on. Willing only what can function as a law for all rational beings involves treating rational beings as constraints on what it is permissible to will and therefore involves treating them as ends. Sidgwick might argue, then, that in so far as he rejects the universal law formulation he rejects the humanity formulation, since he is in effect arguing that what can only serve as a law for all rational agents is not a constraint on what it is permissible for one to do and therefore he rejects the claim that rational nature is an absolute constraint on lying. However, it is possible that the two formulations are not the same, and that the humanity formulation of the categorical imperative contains resources that the universal law formulation does not.10 Kant would, then, be a kind of pluralist. Pluralism seems contrary to Kant’s general orientation, however. The Groundwork, he says, “aims . . . to seek out and establish the supreme principle of morality” (G 4:392; italics in ­original; also 4:421, 4:440). In the Lectures on Ethics, he claims that “[w]here there are already many principles in ethics, there are certainly none, for there can only be one true principle” (LE 27: 266; also MS 6:224). In any case adopting pluralism would be unlikely to matter to Sidgwick, given his general attitude to lying (not to mention pluralism) and so (it seems) his general attitude to using others as mere means. He is likely to hold that it is more plausible that lying is warranted in some cases than that humanity is an absolute constraint on lying. Kant does not always appeal to the humanity formulation of the categorical imperative to establish the impermissibility of lying. Consider, for example, the famous case of the inquiring murderer discussed in his essay “On a Supposed Right to Lie from Philanthropy.” In this case a murderer arrives at your door inquiring about the whereabouts of your friend who is the murderer’s target. If your friend is in your house, your obligation is to tell the murderer the truth: “[t]o be truthful (honest) in all declarations is . . . a sacred command of reason prescribing unconditionally, one not

20  Anthony Skelton to be restricted by any inconveniences” (OSR 8:427; italics in original; also OSR 8:426, 8:429). Most reply that this conclusion is absurd and even Kant’s disciples try to block it (Korsgaard 1986; Varden 2010). After all, it cannot be wrong in this case to lie to protect a friend. Kant demurred. This is unsurprising. However, what is surprising is that when Kant argues that it is wrong to lie to the inquiring murderer, he does not appeal to the universal law formulation or to the humanity formulation of the categorical imperative. There is good reason for him to refrain from so doing. It would expose these versions of the categorical imperative to serious ridicule. If lying in this case would involve either using the murderer as a mere means or a failure to universalize, it would seem to be these factors not the lying that would be impugned. In the case of the inquiring murderer, Kant argues that while lying to the murderer is wrong, the wrong is not done to the murderer. The murdered has no right to what one sincerely believes. Lying in this case is a wrong in general: I . . . do wrong in the most essential part of my duty in general by such falsification . . . I bring it about, as far as I can, that statements (declarations) in general are not believed, and so too that all rights which are based on contracts come to nothing and lose their force; and this is a wrong inflicted upon humanity generally. (OSR 8:426; italics in original) Sidgwick has two potential replies. First, he could suggest that even if lying to the murderer does affect the extent to which statements are believed, it might be worth the cost. By lying in such cases we might help to protect innocent lives and in addition protect ourselves from being employed by murderers to assist them in their task. Second, he might argue that it is not clear in what way lying to murderers threatens the force of all rights based on contracts. The conditions for permissible lying to murderers would, after all, presumably not hold in the case of (most) contracts. It is not obvious how the fact that we have a permission to lie to inquiring murderers makes it the case that I have less assurance that you will not lie to me in the context of drawing up a contract to paint my house. Contracts are not obviously at issue in the case of the inquiring murderer. Of course, Kant might have a point. Once we permit lying we subordinate truth to “other considerations” and this does damage to our assurance that each will sincerely express what each holds to be true “with respect to all unavoidable necessary statements in general” (OSR 8:429; italics in original). But how moved we are by this will depend on how compelling those other considerations happen to be, and to what extent (if at all) the damage to assurance affects contracts.

Practical Ethics in Sidgwick and Kant  21 In Practical Ethics, Sidgwick takes up the issue of the morality of lying in a number of places. In two essays on clerical veracity, he discusses the obligations that officiating ministers owe to their religious communities in the event that the former come sincerely to reject a core conviction of their religion. Sidgwick argues that while clergy may in preaching various religious doctrines—e.g., that Jesus was born of a virgin—or in reciting creeds take advantage of the full breadth of interpretations and various meanings in understanding their statements, they ought not to “utter . . . hard, flat, unmistakable falsehood” (1998, 81). It would be wrong to profess belief in the virgin birth if a clergy member does not believe it. Sidgwick argues that “no gain in enlightenment and intelligence which the . . . ministry may receive from the presence of such men can compensate for the damage done to moral habits, and the offence given to moral sentiments, by their example” (1998, 77; emphasis added). This statement led Hastings Rashdall (1897, 138) to dub Sidgwick a “Kantian rigorist.” Sidgwick might, then, in his writings in practical ethics come closer to Kant. But, in fact, Sidgwick repeats arguments familiar from the Methods, noting that deception may legitimately be practised for the good of the person deceived. Under a physician’s orders I should not hesitate to speak falsely to save an invalid from a dangerous shock. And I can imagine a high-minded thinker persuading himself that the mass of mankind are normally in a position somewhat analogous to that of such an invalid; that they require for their individual and social well-being to be comforted by hopes, and spurred and cured by terrors, that have no rational foundation. (1998, 73–74; also ME 316) Kant’s worry about the corrosive impact of the permission to lie on assurance might, then, be well warranted. He might be right that there is a significant tension between the permission to lie and assurance with respect to all “unavoidable necessary statements.” At best, Sidgwick can reply that this might just be a warning to be extra careful about articulating the conditions under or contexts in which lying is permissible, Sidgwick’s so-called “similar conditions.” In fact, in another essay in Practical Ethics Sidgwick clarifies his position. In “Public Morality,” he argues that there is no difference in the moral norms binding individuals and states. He argues that when there is divergence in the rules binding the two it is a result of differences in circumstances, not principles. He argues with respect to both states and individuals that “urgent need” and previous non-compliance with veracity on the part of those to whom one intends to lie are “required . . . in order completely to justify a breach of veracity” (1998, 44).

22  Anthony Skelton In his own, more practical treatise, Kant stays with his view expressed in the Groundwork and elsewhere that lying is impermissible. However, he offers a different argument for the claim: “[t]he greatest violation of a human being’s duty to himself regarded merely as a moral being . . . is the contrary to truthfulness, lying” (MS 6:429; italics in original). Kant is clear that the wrongness of a lie has nothing to do with its harmfulness in terms of setbacks to the happiness or well-being of others or of oneself. The wrongness of lying does not rest on the fact that it is a violation of one’s duties to others or that it is an error of prudence. Instead, in lying an individual “violates the dignity of humanity in his own person” (MS 6:429). This seems to follow from the fact that there is a tension between lying or intentionally communicating to others the contrary of what a speaker thinks and “the natural purposiveness of the speaker’s capacity to communicate his thoughts” (MS 6:429). Sidgwick misunderstands this argument, thinking it has something to do with lying being prohibited by a code of honor, and he argues (perhaps rightly) that it is not clear that such codes rule out all lying or that all lying is dishonorable and so a threat to dignity (ME 315). This argument misses the essential feature of Kant’s explanation of the threat to dignity, however, namely, the view of the natural purposiveness of the capacity to communicate one’s thoughts. But Sidgwick’s objection might be recast. It is not clear that or how the natural purposiveness of the capacity to communicate rules out all lying. At the very least more needs to be said about the nature of this purposiveness and (especially) its normative relevance. Sidgwick and Kant are, then, in sharp disagreement over the permissibility of lying. Sidgwick is stubborn in permitting lying even if in only select cases. Kant seems equally persistent in his view that (in principle and in practice) lying is always wrong.11 Both have views that court conflict with common-sense attitudes toward lying.

III Sidgwick’s agreement with Kant’s “fundamental principle” did not prevent disagreements over the requirements of beneficence, either. For Sidgwick, of course, beneficence is the whole of morality (putting egoism aside). He is committed to utilitarianism, which requires that agents “produce the greatest amount of happiness on the whole; that is, taking into account all whose happiness is affected by the conduct” (ME 411; also 1889, 485). What he means is that morality requires the promotion of the greatest surplus of pleasure over pain for the aggregate of all sentient beings. This is a demanding standard, as Sidgwick allows. It requires that individuals make extreme sacrifices for the benefit of others. Some think the

Practical Ethics in Sidgwick and Kant  23 demandingness impugns utilitarianism (see Hills 2010). Sidgwick tries at various stages to argue that the impartiality that utilitarianism requires in theory is muted in practice. On Sidgwick’s reckoning, the principle of rational benevolence—that one ought to aim at good generally not merely at a particular part of it, giving equal value to equal quantities of happiness—is the “rational basis for the Utilitarian system” (ME 387). Right after he declares in favor of this principle, he quickly attempts to disabuse readers of the belief that it demands too much in practice. He remarks that I before observed [ME 252–253] that the duty of Benevolence as recognized by common sense seems to fall somewhat short of this [principle]. But I think it may be fairly urged in explanation of this that practically each man, even with a view to universal Good, ought chiefly to concern himself with promoting the good of a limited number of human beings, and that generally in proportion to the closeness of their connexion with him. (ME 382; italics in original) In his discussion of the relationship between common-sense morality and utilitarianism, he reiterates that there is no “conflict” with the partiality of common-sense morality since the “practical application” of the “theoretical impartiality of Utilitarianism is limited by several important considerations” (ME 431). Sidgwick endorsed an indirect form of utilitarianism on which the general happiness is “more satisfactorily attained if men frequently act from other motives than pure universal philanthropy” (ME 413). The best motives are, for Sidgwick, the main rules (suitably modified) comprising common-sense morality. Accordingly, Sidgwick argues on utilitarian grounds that one is permitted to give one’s own good greater weight than the good of others and to give wide scope to one’s own self-interest because of one’s greater knowledge of one’s own good and the means to securing it and because “it is under the stimulus of self-interest that the active energies of most men are most easily and thoroughly drawn out” (ME 431). In addition, he argues that it is “tolerably obvious” that it is conducive to the general happiness “that each individual should distribute his beneficence in the channels marked out by commonly recognized ties and claims” (ME 433). We are permitted on this view to give greater weight to those with whom we have close relations and who have a special claim on our services, including our own children, our lovers, our friends and those in special need (e.g., the prudent poor or unexpectedly endangered). In this way some worries about demandingness might be deflected and utilitarianism thereby partially reconciled with common-sense morality.

24  Anthony Skelton However, it is not clear how plausible this is as a tack to ward off worries about demandingness. Some of Sidgwick’s remarks suggest that it is not completely effective even given his indirect utilitarianism. He argues that, though partiality is permitted by utilitarianism in general, in cases of unforeseen calamity or emergency I am required to adopt utilitarian impartiality: if I am made aware that, owing to a sudden calamity that could not have been foreseen, another’s resources are manifestly inadequate to protect him from pain or serious discomfort, the case is altered; my theoretical obligation to consider his happiness as much as my own becomes at once practical; and I am bound to make as much effort as to relieve him as will not entail a greater loss of happiness to myself or others. (ME 436) This may be too demanding for common sense. It is not clear that common sense would grant that we are required in this case to be impartial as between our own good and another’s. Common sense might allow one in calamity cases to give a bit more weight to oneself or one’s intimates over strangers. Suppose that two people are hit by a calamity and are short on the resources needed to prevent (equal amounts of net) suffering to themselves, but it is not possible to help both. Suppose that one of the individuals whom you can help is your child and one is a complete stranger. It seems plausible that you are not required to be indifferent between saving your child or a stranger. It is permissible if not obligatory to prevent your child’s suffering. Common sense might permit deviations from this kind of impartiality in a different way. Suppose that I can save a stranger from suffering due to an unforeseen calamity, but that helping costs me slightly more than it benefits the stranger. However, I help nonetheless. Sidgwick calls this “self-sacrifice” and he agrees that common sense praises it (ME 431). He tries to reconcile this to utilitarianism in a number of ways. He might use the same arguments to deflect this worry about his unforeseen calamity principle. He says that we frown on such sacrifices when “the disproportion between the sacrifice and the benefit is obvious and striking” (ME 432). But this leaves open that in cases without such disproportion (like the one under discussion) common sense approves. Sidgwick notes that it is possible that we approve anyway “because such conduct shows a disposition far above the average in its general tendency to promote happiness, and it is perhaps this disposition that we admire rather than the particular act” (ME 432). But it is just as likely that we approve of the act itself apart from the disposition and that we approve of the disposition in advance of our knowledge of its impact on the promotion of general happiness.

Practical Ethics in Sidgwick and Kant  25 Kant might offer a better account of the demands of beneficence. Sidgwick might be accused of offering a standard that is too exacting. Sidgwick says that he and Kant agree on the principle of rational benevolence (ME 386). But this is hard to believe, given what Kant actually says about the duty of beneficence. Kant expresses a variety of standards of beneficence. In the Groundwork, the duty of beneficence is described variously as “a duty to help others [in distress] where one can” (G 4:398), a duty to contribute at least something to the well-being of those who experience “great hardships” or “distress” and can be assisted easily (G 4:423), and a duty to contribute to the happiness of others more generally as a requirement of treating humanity as an end (G 4:430). Treating others as ends involves, as far as one can, furthering “the ends of others. For the ends of any person who is an end in himself must, if this idea is to have its full effect in me, be also, as far as possible, my ends” (G 4:430; italics in original). In the Critique of Practical Reason, Kant mentions in passing that it is wrong to look on the needs of others with “complete indifference” (KpV 5: 69). In general, it is unclear how much these standards demand. The very last one calls only for us not to be completely indifferent. It is possible that avoiding complete indifference is consistent with doing only very little for others. Have I avoided indifference if I promote the needs of others only very rarely when it is very easy for me to do so? To what extent I have avoided it will depend on what not being completely indifferent amounts to. But Kant says little about what complete indifference involves. It is hard to know what the general requirement to further others’ happiness involves. It might be quite demanding. It will depend on how one interprets Kant’s claim that an agent is to further the ends of others “as far as he can” (G: 4:430). It is possible that the first two standards from the Groundwork are quite demanding, though Kant gives no account of how much we are required to help others in distress or those experiencing great hardships. In one case he says that the obligation is to help when doing so is easy, so it seems that he does not intend anything terribly demanding. Kant’s most robust discussion of the demands of beneficence are found in the Metaphysics of Morals, where he states that The happiness of others is . . . an end that is also a duty . . . I ought to sacrifice a part of my welfare to others without hope of return, because this is a duty, and it is impossible to assign determinate limits to the extent of this sacrifice. How far it should extend depends, in large part, on what each person’s true needs are in view of his sensibilities, and it must be left to each to decide for himself. For, a maxim of promoting others’ happiness at the sacrifice of one’s own happiness, one’s true needs, would conflict with itself if it were made a universal law. Hence this duty is only a wide one; the duty has in it a

26  Anthony Skelton latitude for doing more or less, and no specific limits can be assigned to what should be done.—The law holds only for maxims, not for determinate actions. (MS 6:393; italics in original) This suggests a very undemanding standard of beneficence. The strength of the obligation of beneficence rests, seemingly, on what an individual decides based on her view of her true needs. The standard seems to state that you should do something for the well-being or happiness of others, though how much is left largely up to your “sensibilities.” The main idea seems to be that you should at least not be indifferent. Kant also emphasizes that we are permitted to give quite significant priority to our own well-being or happiness over the well-being of strangers in discharging the duty of beneficence. He notes that If one is closer to me than another (in the duty of benevolence) and I am therefore under obligation to a greater benevolence to one than to the other but am admittedly closer to myself (even in accordance with duty) than to any other, then it would seem that I cannot, without contradicting myself, say that I ought to love every human being as myself, since the measure of self-love would allow for no difference in degree.—But it is quite obvious that what is meant here is not merely benevolence in wishes, which is, strictly speaking, only taking delight in the well-being of every other and does not require me to contribute to it (everyone for himself, God for us all); what is meant is, rather, active, practical benevolence (beneficence), making the well-being and happiness of others my end. For in wishing I can be equally benevolent to everyone, whereas in acting I can, without violating the universality of the maxim, vary the degree greatly in accordance with the different objects of my love (one of whom concerns me more closely than another). (MS 6:451–452; italics in original) Kant is not clear on exactly how much greater weight one is permitted to grant to one’s own happiness or well-being. But it seems from what he says that we may give quite a bit more priority to our own happiness. Kant might, then, easily be accused of developing a view of beneficence that is not demanding enough in the face of the needs of others and the capacity that many of us have to assist. Kant ought to have developed a more demanding standard. He is, of course, no stranger to demanding moral standards. He marks out a distinction between perfect or narrow duties and imperfect or wide duties. Duties to refrain from suicide in order to avoid a miserable condition or from lying are perfect or narrow duties, allowing of no latitude in how and when they are discharged, while the duty of beneficence is an

Practical Ethics in Sidgwick and Kant  27 imperfect or wide duty which allows for latitude in how and when it is discharged (G 4:421–424). Respecting the requirements to refrain from suicide to avoid a miserable condition and from lying (as we have seen) can be very demanding. You are required to refrain from lying presumably even to save your own child from an inquiring murderer, and you are required not to kill yourself even if all that your future life promises is unbearable and irremediable suffering. Perhaps fortunately a more stringent standard of beneficence may be extracted from Kant. He notes that in discharging one’s duties of beneficence, sacrifice of part of one’s well-being or happiness is required (MS 6:393, 6:453). He does not say exactly how much sacrifice is required. He seems to suggest that the extent of any sacrifice be left up to the individual’s sensibilities. He can do better than this, lamentably, subjective standard. In his discussion of the vice of avarice, the avoidance of which is a duty to oneself as a moral being, Kant suggests that “acquiring the means to living in excess of one’s true needs” amounts to “greedy avarice” and therefore to a “violation of one’s duty (of beneficence) to others” (MS 6:432; italics in original). The suggestion is that greed involves keeping for oneself more than is required to satisfy one’s true needs with the intention of enjoying it, and that it is wrong to do so because it is a violation of one’s duty of beneficence to other people. By true needs Kant seems to mean “comforts necessary to enjoy life” (MS 6: 433). To deny oneself these, Kant suggests in his discussion of the vice of miserliness, is to “leave one’s own true needs unsatisfied” (MS 6:432) and, he further suggests, to “come to need the beneficence of others” (MS 6:454). It is plausible that when Kant talks about requiring that one sacrifice part of one’s happiness or well-being for others, he means that one sacrifices the part of one’s happiness not related to one’s true needs. Perhaps, then, the duty of beneficence might involve (at least) obliging one to sacrifice one’s resources that are in excess of what is needed to satisfy one’s true needs in order to benefit others, where to benefit them involves providing for their true needs or the comforts necessary to enjoy life. This fits with the focus on “distress” or to what others “need” in his various discussions of beneficence (G 4:398, 4:423; KpV 5:69; MS 6:452, 6:453; LE 27:455) and on his injunction “not to avoid the places where the poor who lack the most basic necessities are to be found but rather to seek them out” (MS 6:457) (since this is important to cultivating compassion, a “means to promoting active and rational benevolence” (MS 6:456)). Of course, Kant holds that the closer one is to another the “greater” the claim or obligation of beneficence one has to that other, and so we have therefore greater duties to our own true needs and to the true needs of those to whom we have close ties (including our spouses, children, and friends). This in some way limits the demands of

28  Anthony Skelton beneficence to strangers and so might generate as a more plausible standard of beneficence than the one Sidgwick puts forward. Still, this might reveal a quite demanding obligation both to those close to one and to strangers. It is plausible that if one thought clearly about what fell into the category of true needs, it would turn out that what is necessary to enjoy life would not include much of what forms the focus of the resources of most affluent people. Any account of true needs is going to leave many such people with quite demanding duties to the true needs of strangers (not to mention those to whom they are closest). This would be demanding even given Kant’s view that perfect or narrow duties take precedence over wide or imperfect duties (such as the duty of beneficence). Famously, Sidgwick rejects the general argument that Kant provides for the duty of beneficence. In a number of places, Kant argues that I am required to be beneficent to others because I cannot will as a universal law that I do not help others in distress or advance others’ happiness, for when I am in need or in distress I will very much will that I be helped. According to Kant, a will that intended not to help others in distress would “be in conflict with itself, since many situations might arise in which the man needs love [that is, to be ‘helped in case of need’ (MS 6: 393)] and sympathy from others, and in which, by such a law of nature generated by his own will, he would rob himself of all hope of the help he wants” (G 4:423; also MS 6:393, 6:453). Sidgwick argues against Kant that Kant cannot derive the duty of benevolence in this way. Sidgwick notes that the argument depends on an empirical claim, namely, that a man will want that others help him when he finds himself in distress. However, Sidgwick argues, it is possible to conceive of a man “in whom the spirit of independence and the distaste for incurring obligations would be so strong that he would choose to endure any privations rather than receive aid from others” (ME 389). And, moreover, Sidgwick says, even if we all as a matter of fact will that we be helped by others when we find ourselves in need, we can imagine that a strong man, after balancing the chances of life, may easily think that he and such as he have more to gain, on the whole, by the general adoption of the egoistic maxim; benevolence being likely to bring them more trouble than profit. (ME 389) Kant could reply to this by reliance on the idea that we have true needs as autonomous beings. He may argue that the independent and the strong man cannot fail to will that others help them at least with respect to their true needs, since fulfilment of these are the preconditions for living as strong and independent beings. Strong and independent men must will

Practical Ethics in Sidgwick and Kant  29 that others help them with at least the necessities for maintaining strength and independence because they do not have complete control over these necessities or means, though they will that they have these means as part of willing their specific form of life. At least with respect to some of the necessities, only other agents have some control over them. And in so far as the strong and independent men will their form of life, they will the necessities or means and so will that others provide them when the strong and independent men find themselves in distress or are unable to obtain these necessities or means on their own. If they cannot fail to will that our true needs be met, Kant can deflect Sidgwick’s objection.12 If one cannot fail to will that one’s true needs be met when one is in need, one cannot refuse to benefit others in terms of their true needs when others are in distress. Sidgwick and Kant disagree on the demands of beneficence. Sidgwick articulates a pretty demanding standard. Kant is not careful in his articulation of the requirement of beneficence. The most plausible account that might be extracted from Kant is quite demanding, though it is not perhaps as demanding as Sidgwick’s account. Both offer views that challenge conventional attitudes respecting beneficence.

IV Kant and Sidgwick disagree about the permissibility of lying and about the demands of beneficence. Their disagreements on lying and beneficence affect their practical recommendations. Their disagreements are not merely theoretical. How might these disagreements be dealt with in light of the need to achieve some consensus over what we ought to do in practice? In reply to the disagreement, Kant might argue that he provides the best representation of common-sense moral thinking. We might think that providing the best representation of this thinking is a sure sign that a philosopher has secured the proper kind and degree of consensus. Kant claims in numerous places that the categorical imperative undergirds common-sense morality. In the Critique of Practical Reason he says The rule of judgment under laws of pure practical reason is this: ask yourself whether, if the action you propose were to take place by a law of nature of which you were yourself a part, you could indeed regard it as possible through your will. Everyone does, in fact, appraise actions as morally good or evil by this rule. (KpV 5:69) In the Groundwork, he remarks, of the categorical imperative, that “Common human reason, when engaged in making practical judgements, also agrees with this completely and has that principle constantly in view” (G 4:402; also 4:404).

30  Anthony Skelton It is hard for Kant to sustain this position. He seemingly contradicts common-sense morality in his official views about lying and possibly about beneficence. It is far from clear that common-sense morality forbids lying in every case and that when it comes to beneficence it counsels only that we avoid complete indifference. He cannot credibly claim to be representing common-sense morality in these cases. He is revising it. That Kant might be revising common-sense morality is actually a problem for him. He is famous for having accepted that each is equally capable of accessing and being motivated by the moral law (see Schneewind 1999). He was convinced of this by Rousseau: I myself am a researcher by inclination. I feel the entire thirst for cognition and the eager restlessness to proceed further in it, as well as the satisfaction at every acquisition. There was a time when I believed this alone could constitute the honour of humankind, and I despised the rabble who knows nothing. Rousseau has set me right. This blinding prejudice vanishes, I learn to honour human beings, and I would feel by far less useful than the common laborer if I did not believe that his consideration could impart a value to all others in order to establish the rights of humanity. (Bem 20:44; italics in original) It is important for Kant to show that his principle underlies our common morality; otherwise, he might have to claim that the moral philosopher has special access to moral principles which s/he communicates to “the rabble.” If Kant cannot show that the categorical imperative is the principle underlying common-sense morality, he either has to give up the principle in favor of one more representative of common-sense morality or give up attempting to represent common-sense morality and the claim that each of us is equally capable of figuring out and being motivated by the claims of morality. If he sticks with his principle, he cannot claim that “even the most ordinary human intelligence can easily be brought to a high degree of correctness and completeness in moral matters” (G 4:391). This might be unfair to Kant. Kant’s claim is that ordinary human intelligence exhibits a high but not perfect degree of correctness and completeness in moral matters. This presumably leaves room for the moral philosopher to play some role in correcting and completing our ordinary moral views. The philosopher might perfect it in some way. The problem is that Kant’s corrections to and completion of ordinary moral views are not entirely agreeable. Sidgwick suggests at points that his favoured principle best represents common-sense morality. He argues that common-sense morality is inchoately and unconsciously utilitarian (ME 424, 427, 453, 454, 462) and that utilitarianism solves the problems besetting common-sense morality

Practical Ethics in Sidgwick and Kant  31 (involving vagueness in and conflict between its various elements) in “general accordance with the vague instincts of Common Sense, and is naturally appealed to for such solution in ordinary moral discussions” (ME 425). Unfortunately, Sidgwick concedes that utilitarianism, not to mention egoism, conflicts with common-sense morality and that in part the philosopher’s function is to “tell men what they ought to think, rather than what they do think” (ME 373). He notes that the limits of the philosopher’s deviation from common sense are “firmly, though indefinitely, fixed” (ME 373). But it is far from obvious that Sidgwick’s deviations fall within what escapes controversy. If neither Kant nor Sidgwick is able to secure consensus using their favoured principles, how can moral disagreement be dealt with for the purpose of practice? In his most practical treatise on ethics, Metaphysics of Morals, Kant says the right view for practical ethical reasoning is the categorical imperative. He argues that The supreme principle of the doctrine of virtue is: act in accordance with a maxim of ends that it can be a universal law for everyone to have.—In accordance with this principle a human being is an end for himself as well as for others, and it is not enough that he is not authorised to use either himself or others merely as a means (since he could then still be indifferent to them); it is in itself his duty to make man as such his end. (MS 6:395) So Kant’s practical ethics is just an extension, and (in the broad sense) an application, of his theoretical ethics. As we have seen, Sidgwick disagrees with Kant’s understanding of his own principle and, of course, Kant rejects Sidgwick’s utilitarianism. In the Methods, Sidgwick remarks that in the face of disagreement like this, either one is reduced to a state of neutrality regarding the matter about which there is disagreement, or one has shown that those with whom one disagrees are more likely than one is to be in error: if I find any of my judgments, intuitive or inferential, in direct conflict with a judgement of some other mind, there must be error somewhere: and if I have no more reason to suspect error in the other mind than in my own, reflective comparison between the two judgements necessarily reduces me temporarily to a state of neutrality. (ME 342) In the Methods, he was not reduced to a state of neutrality with respect to the things about which he and Kant disagreed and so did not yield to disagreement with Kant. This might have been due to the fact that Sidgwick thought that his many objections to Kant’s arguments meant that

32  Anthony Skelton he is less likely to be in error than Kant.13 In this case he could carry on with his utilitarianism in practical ethics.14 That is certainly one way to go. However, there is another, more plausible way to do practical ethics. Sidgwick suggests this way in his book Practical Ethics which includes, among other things, lectures given before practical ethics societies. In one lecture, Sidgwick recounts his experience with the Metaphysical Society, of which he was a member from 1869–1880. He says that the society broke down because the various sides in the controversies of which it treated were unable to come to agreement, and each, philosophically, “remained exactly where we were . . . and no one being in the least convinced by any one else’s arguments” (1998, 4). In part, the reason for the inability to overcome disagreement in the context of the Metaphysical Society was that the aim of such a society was to get to the bottom of the things on which it focused including the meaning of life and the essence of virtue (1998, 5). In practical ethics, Sidgwick argues, it is easier to avoid what brought the Metaphysical Society to an end. This can be done by giving up “altogether the idea of getting to the bottom of things, arriving at agreement on the first principles of duty or the Summum Bonum” (1998, 5). This is possible in part because the focus of practical ethics is “not knowledge but action: and with this practical object it is not equally necessary that we should get to the bottom of things” (1998, 5). We can give up trying to defend or rely on utilitarianism or the categorical imperative. In so doing, he argues, it might be possible to “reach some results of value for practical guidance and life” (1998, 5). Instead of trying to work out the most basic requirements of morality and then applying them to practical problems, we ought to stay as close as we can to the “region of middle axioms” (1998, 7). This region comprises, roughly speaking, the “broad agreement in the details of morality which we actually find among thoughtful persons who profoundly disagree on first principles, and among plain men who do not seriously trouble themselves about first principles” (1998, 6). The broad agreement coalesces roughly around the main rules of common-sense morality, including requirements of veracity and of beneficence, among others. The function of practical ethical thinking is, then, to bring into a more clear and consistent form the broad and general agreement as to the particulars of morality which we find among moral persons, making explicit the general conceptions of the good and evil in human life, of the normal relation of man to his fellows, which this agreement implies. (1998, 7; also 6) One shape this form of reasoning takes is illustrated in what Sidgwick calls casuistry, or reasoning concerning “how far, in the particular

Practical Ethics in Sidgwick and Kant  33 circumstances of certain classes of persons, the common good demands a special interpretation or modification of some generally accepted moral rule” (1998, 12). This involves, among other things, dealing with exceptions to the requirement of veracity and the extent of the demands of beneficence. On Sidgwick’s reckoning, refusal to deal with such problems would leave any approach to practical ethics inadequate for the practical guidance of men engaged in the business of the world; since modifications of morality to meet the special needs of special classes of people are continually claimed, and more or less admitted by serious and well-meaning persons. (1998, 12) This might be an anathema to Kant. But in fact it is not. First, Kant is a great admirer of common-sense morality so sticking with it is not contrary to his view. In addition, although Kant espouses strong views on lying, as recounted earlier, at one point he relaxes his view in line with Sidgwick’s description of casuistry. He notes that it might, after all, be permissible to lie in cases in which “men are malicious”: it is true that we often court danger by punctilious observance of the truth, and hence has arisen the concept of the necessary lie, which is a very critical point for the moral philosopher. For seeing that one may steal, kill or cheat from necessity, the case of emergency subverts the whole of morality, since if that is the plea, it rests upon everyone to judge whether he deems it an emergency or not; and since the ground here is not determined, as to where the emergency arises, the moral rules are not certain. For example, somebody, who knows that I have money, asks me: Do you have money at home? If I keep silent, the other concludes that I do. If I say yes, he takes it away from me; if I say no, I tell a lie; so what am I to do? So far as I am constrained, by force used against me, to make an admission, and a wrongful use is made of my statement, and I am unable to save myself by silence, the lie is a weapon of defence; the declaration extorted, that is then misused, permits me to defend myself, for whether my admission or my money is extracted, is all the same. Hence there is no case in which a necessary lie should occur, save where the declaration is wrung from me, and I am also convinced that the other means to make a wrongful use of it.15 (LE 27:448) Note that Kant is here doing exactly what Sidgwick thinks it is important to do in practical ethics. Kant is sticking close to a conviction core to common-sense morality, that lying is wrong. He is then working out an exception to it for a particular class of person in a particular kind of

34  Anthony Skelton circumstance. Note, too, that he says that this exception is a critical point for a moral philosopher, suggesting that he or she might have some special role to play in working out the correct view. Second, Kant himself leaves us with lots of casuistical questions in the Metaphysics of Morals, which his own principle of practical ethics seems not to answer definitively. For example, he wonders whether one should reply honestly when one’s colleague asks about the quality of his work (MS 6:431). He is wondering about exceptions to common-sense moral rules. To the extent that Kant leaves us with lots of such questions, he ought to join Sidgwick in his insistence on the importance of casuistry under strict conditions, not only because we want answers to such questions, but because, as Sidgwick remarks, it is likely that there is a serious danger that when the need of such relaxations [of moral rules] is once admitted they may be carried too far . . . and no result could be more unfavourable than this to the promotion of ethical interest. (1998, 13) Discussion of such exceptions may well be a solid protection of the rights of humanity. Sidgwick notes that one purpose germane to his conception of practical ethics is to make the rules on which we agree “more explicit and clear than . . . [they are] in ordinary thought” (1998, 6). Kant wonders, for example, about the extent of the demands of beneficence (MS 6:454). Sidgwick wonders the same (1998, 6, 10). Sidgwick’s claims about what is important to practical ethics should resonate with Kant, who himself had trouble working out a full view. Sidgwick goes on to work out a view about the role of the philosopher and the plain man in practical ethics. The philosopher serves as a kind of moral expert in terms of clarifying the rules on which we agree. Her expertise is a function of her training (1998, 19–20). Unsurprisingly, the philosopher’s job is not to work out her own favored position like utilitarianism or the categorical imperative and then to apply it in practice. This would invite the “dangers of sectarianism” and “conflicts” among rival moral factions (1998, 25). It would be disrespectful simply to force a moral view on others in the face of what appears to be perfectly reasonable disagreement (for more on this in Sidgwick, see Skelton 2006). The philosopher is of course to keep her reasoning focused on the “agreement among thoughtful persons,” that is, the main elements of common-sense morality on which both Kant and Sidgwick draw. What is the role for the plain man? Sidgwick’s approach to practical ethics requires a “broad and comprehensive basis” (1998, 20). Accordingly, the plain man’s “special moral experience” is something to which the moral philosopher ought to defer (1998, 23). Indeed, the moral

Practical Ethics in Sidgwick and Kant  35 philosopher “should always study with reverent care and patience . . . the Morality of Common Sense” (1998, 22; also 22, 23). This is for two reasons. First, the philosopher lacks the plain man’s “comprehensive and varied knowledge of the actual opportunities and limitations, the actual needs and temptations, the actually constraining customs and habits, desires and fears, of all the different species of that ‘general man’ ” to whom a view of right will apply (1998, 21). Second, the judgements of practical men as to what ought to be done in particular circumstances are often far sounder than the reasons they give for them; the judgements represent the result of experience unconsciously as well as consciously imbibed; the reasons have to be drawn from that more limited part of experience which has been the subject of conscious observation, information, and memory. (1998, 21–22) Kant might dissent from the claim that the philosopher is some kind of moral expert. But given the training and the time philosophers have to devote to moral questions, it seems not unwarranted to think that they may be at least somewhat better placed than most others to perform certain forms of moral labor (e.g., articulation of moral rules and their exceptions) (Shaver 2012; Skelton 2014). The plain man, as Sidgwick points out, is not well placed to do that. Sidgwick dissents from Butler’s claim that the plain man knows pretty much what to do in most practical contexts (1998, 6, 23) on the grounds that matters in the world are a good deal more complex than they were in Butler’s time. If this was true of Sidgwick’s time (1998, 24), it is certainly true of our own, where commonsense morality is often perplexed by moral issues arising from the use of new technologies. Kant might seem to agree with Butler (G 4:404). But Kant ought to agree that common sense cannot be our sole guide. He raises, as noted, many casuistical questions on which we are, to use Sidgwick’s word, “perplexed.” In addition, he concedes that in some sense the moral philosopher perfects the ordinary person’s thoughts. In general, Sidgwick’s view provides a way to respect both the authority of the researcher/philosopher and the authority of the plain man. Each has their own “special” moral expertise to which each ought to defer. Sidgwick is quite insistent that the main features of common-sense morality aid and control the philosopher’s constructions or clarifications of shared rules (1998, 23). This should appeal to Kant in part because of his view of the “correctness” and “completeness” of the ordinary man’s views. He accords great respect to “common moral judgement” (G 4:412). In addition, the insistence on respect in common-sense morality helps to ward off a worry that Kant has about the judgements of moral philosophers who “may easily be confused by a mass of strange and irrelevant considerations and caused to turn from the right path” (G 4:404).

36  Anthony Skelton But it leaves room for the philosopher to help, in cases where common sense is unclear on “perplexed,” and where the philosopher’s tools might well come in handy. Sidgwick’s general approach to practical ethics is respectful of reasonable disagreement. It offers a way forward in practical ethics that attempts to arrive at results that are both widely agreeable and useful in practice. Sidgwick’s broad approach to practical ethics gives space to the judgement of the common person and to the philosopher. It is an attractive general model for how best to conduct practical moral reasoning in light of theoretical moral disagreements. Kant has good reason to agree. He notes that common-sense morality is important but that it leaves some things unanswered. He notes that the philosopher might have some role to play in clarifying and making up for the deficiencies, even if only slight, in the morality of common sense.

V Sidgwick claimed Kant as one of his moral philosophical “masters.” His engagement with Kant’s ethics is sustained and fruitful. Sidgwick and Kant disagree on a great many moral issues in theory and in practice, including the morality of lying and the demands of beneficence. Neither offers a view that is completely agreeable. Dealing with the conflict between them (and others) in practice is important to making progress in solving practical moral problems. Sidgwick offers a way to handle disagreements in practice that is novel and interesting. Kant fails to deal effectively with disagreement in practical ethics. Sidgwick might well have something to teach his master.16

Notes 1. Sidgwick took a dim view of post-Kantian philosophers in general (Sidgwick and Sidgwick 1906, 233). 2. For the same claim about Sidgwick’s objections to Kant on beneficence, see Hooker (2000, 347). 3. Immanuel Kant’s works are abbreviated herein as follows: Bem = Observations on the Feeling of the Beautiful and the Sublime and other Writings (Kant 2010); G = Groundwork for the Metaphysics of Morals (Kant 2002); KpV = Critique of Practical Reason (Kant 1997a); LE = Lectures on Ethics (Kant 1997b); MS = Metaphysics of Morals (Kant 1996a); OSR = “On a Supposed Right to Lie from Philanthropy” in Immanuel Kant: Practical Philosophy, ed. M. Gregor (Kant 1996b). 4. This objection is discussed in the chapters by Guyer and Nakano-Okuno in this volume. For earlier treatments, see, for example, Korsgaard (1996, 159–187), Timmerman (2007, 164–167), and Wuerth (2014, 247–254). For a penetrating defense of Sidgwick’s thesis, see Fugate (2015). 5. Sidgwick also rejected Mill’s qualitative hedonism (ME 94–95, 121). 6. For detailed accounts of Sidgwick’s argument for utilitarianism, see Skelton (2008), Skelton (2010), Shaver (2014), and Crisp (2015).

Practical Ethics in Sidgwick and Kant  37 7. Sidgwick rejected Mill’s appeals to sanctions to close the gap (ME 499–503). For discussion of Mill and Sidgwick on sanctions, see Harrison (2001). For Mill, see Mill (1969, 228–233). 8. For the same sentiment, see Ross (1954, 9). The “great value” of Kant’s universal law formulation of the categorical imperative, Ross writes, “lies in its putting us in the right attitude, by requiring us to ignore our own particular wishes and to adopt an impersonal point of view, in which we are less likely to be led astray by considerations of personal interest.” 9. Sidgwick did not engage fully with Kant’s attempts to deal with the conflict between duty and self-interest. 10. Kant could perhaps drop the universal law formulation. This might be plausible in light of the problem of false negatives to which Sidgwick refers (ME 487). 11. In MS, Kant wonders whether some lies are permissible in the Casuistical Questions he considers following his discussion of lying (MS 6:431). He considers whether one should reply honestly when one’s colleague asks about the quality of the work she has produced. Kant does not clearly state a view. Some think in this case he favors lying (e.g., Parfit 2011, xliv). For more on this, see section IV. 12. For a similar proposal, see Herman (1984). Herman has a different account of true needs on which they amount to what is necessary to support the capacity to set ends in rational beings (586). The passages to which she points in Kant do not support her reading (6:392, 6:393). For more on Sidgwick’s critique of Kant’s derivation of the duty of benevolence, see Paul Guyer’s chapter in this volume. 13. Sidgwick raises objections to many of Kant’s views beyond those discussed here, including Kant’s view that we have only an indirect duty to promote our own happiness (ME 386), that the good will is the only thing good without limitation (ME 222, 394–95) and that it does not involve an emotional element (ME 223), that we cannot promote others’ perfection (ME 240), and that we have only indirect duties to non-human animals (ME 241, 414, 431). 14. For more on Sidgwick’s views on peer disagreement and how this pertains to his views on Kant, see Robert Shaver’s chapter in this volume. 15. This contradicts Kant’s official view discussed earlier in the case of the inquiring murderer. In that case the murderer means to make wrongful use of my declaration and the declaration of my friend’s whereabouts is wrung from me. 16. I wish to thank Lisa Forsberg and Tyler Paytas for helpful comments on previous versions of this chapter and Corey Dyck and Dennis Klimchuk for helpful conversations about Kant’s ethics.

References Crisp, R. 2015. The Cosmos of Duty: Henry Sidgwick’s Methods of Ethics. Oxford: Oxford University Press. Fugate, C. D. 2015. “On A Supposed Solution to the Reinhold/Sidgwick Problem in Kant’s Metaphysics of Morals.” European Journal of Philosophy 23: 349–373. Harrison, R. 2001. “The Sanctions of Utilitarianism.” In Henry Sidgwick, edited by R. Harrison, 93–116. Oxford: Oxford University Press. Herman, B. 1984. “Mutual Aid and Respect for Persons.” Ethics 94: 577–602. Hills, A. 2010. “Utilitarianism, Contractualism and Demandingness.” Philosophical Quarterly 60: 225–242.

38  Anthony Skelton Hooker, B. 2000. “Sidgwick and Common-sense Morality.” Utilitas 12: 347–360. Kant, I. 1996a. Metaphysics of Morals. Edited by Mary Gregor. Cambridge: Cambridge University Press. Kant, I. 1996b. Immanuel Kant: Practical Philosophy. Edited by Mary Gregor. Cambridge: Cambridge University Press. Kant, I. 1997a. Critique of Practical Reason. Edited by Mary Gregor. Cambridge: Cambridge University Press. Kant, I. 1997b. Lectures on Ethics. Edited by P. Heath and J. B. Schneewind. Cambridge: Cambridge University Press. Kant, I. 2002. Groundwork for the Metaphysics of Morals. Edited by T. Hill, Jr. and A. Zewig. Oxford: Oxford University Press. Kant, I. 2010. Observations on the Feeling of the Beautiful and the Sublime and Other Writings. Edited by P. Frierson and P. Guyer. Cambridge: Cambridge University Press. Korsgaard, C. M. 1986. “The Right to Lie: Kant on Dealing with Evil.” Philosophy and Public Affairs 15: 325–349. Korsgaard, C. M. 1996. “Morality as Freedom.” In Creating the Kingdom of Ends, edited by C. M. Korsgaard, 159–187. Cambridge: Cambridge University Press. Mill, J. S. 1969. Utilitarianism. In Essays on Ethics, Religion and Society, Volume X of the Collected Works of John Stuart Mill, edited by J. Robson. Toronto: University of Toronto Press. Parfit, D. 2011. On What Matters, 2 vols. Oxford: Oxford University Press. Passmore, J. 1975. A Hundred Years of Philosophy. Harmondsworth: Penguin. Rashdall, H. 1897. “Professor Sidgwick on the Ethics of Religious Conformity: A Reply.” International Journal of Ethics 7: 137–168. Ross, W. D. 1954. Kant’s Ethical Theory. Oxford: Oxford University Press. Schneewind, J. B. 1999. The Invention of Autonomy. Cambridge: Cambridge University Press. Shaver, R. 2012. “Ethical Non-naturalism and Experimental Philosophy.” In Ethical Naturalism, edited by S. Nuccetelli and G. Seay, 194–210. Cambridge: Cambridge University Press. Shaver, R. 2014. “Sidgwick’s Axioms and Consequentialism.” Philosophical Review 123: 173–204. Sidgwick, A., and E. M. Sidgwick. 1906. Henry Sidgwick: A Memoir. London: Macmillan. Sidgwick, H. 1873. “John Stuart Mill.” Academy (May 15): 193. Sidgwick, H. 1883. “A Criticism of the Critical Philosophy.” Mind 8: 69–91. Sidgwick, H. 1889. “Some Fundamental Ethical Controversies.” Mind 14: 473–487. Sidgwick, H. 1902. Outlines of the History of Ethics. 5th ed. London: Macmillan. Sidgwick, H. 1905. Lectures on the Philosophy of Kant and Other Philosophical Lectures and Essays. London: Macmillan. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Sidgwick, H. 1998. Practical Ethics: A Collection of Addresses and Essays. Edited by S. Bok. Oxford: Oxford University Press. Skelton, A. 2006. “Henry Sidgwick’s Practical Ethics: A Defense.” Utilitas 18: 199–217.

Practical Ethics in Sidgwick and Kant  39 Skelton, A. 2008. “Sidgwick’s Philosophical Intuitions.” Etica & Politica/Ethics & Politics 10: 185–209. Skelton, A. 2010. “Henry Sidgwick’s Moral Epistemology.” Journal of the History of Philosophy 48: 491–519. Skelton, A. 2014. “On Henry Sidgwick’s ‘My Stations and its Duties’.” Ethics 125: 586–591. Timmerman, J. 2007. Kant’s Groundwork of the Metaphysics of Morals: A Commentary. Cambridge: Cambridge University Press. Varden, H. 2010. “Kant and Lying to the Murdered at the Door . . . One More Time: Kant’s Legal Philosophy and Lies to Murderers and Nazis.” Journal of Social Philosophy 41: 403–421. Wuerth, J. 2014. Kant on Mind, Action, and Ethics. Oxford: Oxford University Press.

2 Self-Defeating Moral Theories and Kant’s Doctrine of Right Tim Henning

Introduction Will the truth about morality, if we find it, serve us as a guide to moral action? This is a natural idea. Aristotle agrees: “Surely, then, knowledge of the good must be very important for our lives? And if, like archers, we have a target, are we not more likely to hit the right mark?” (2000, NE 1094a, 23f). But what if the truth about morality is that we had better ignore the truth about morality? This question does not just raise the possibility that knowledge of the good may turn out to be of little practical value. (This was not alien to Aristotle who, after all, doubted that knowledge of the good in the guise of knowledge of general principles makes us virtuous.) The question is a disconcerting one: Maybe knowledge of the good, if used as a guide, has a tendency to make us miss the mark? If this were true, morality would put us in a weird intellectual predicament indeed, prescribing either that we refrain from finding out the truth about it (a prescription which, of course, is bound to come too late), or that we take great pains to ignore it as soon as we turn to action (see Stocker 1976). On the other hand, the possibility is not exactly farfetched either. Since we mentioned Aristotle: Some authors have argued that virtue ethics seems to invite the disconcerting thought. They suspect that great concern with the virtuous state of one’s own character, and with the flourishing of one’s own life, might actually get in the way of achieving these aims (see Keller 2007). The idea that morality might condemn its own use as a practical guide has received attention in H. Sidgwick’s Methods of Ethics, and it has become the object of even greater scrutiny in Sidgwick’s successor, D. Parfit, in book 1 of Reasons and Persons. Due to these authors, the idea that morality might be practically self-defeating has been most thoroughly discussed in the utilitarian tradition. In this chapter, I want to argue that the disconcerting thought can also be traced to Kant. I will argue that Kant’s Doctrine of Right is best understood as motivated by the insight that the Categorical Imperative

Self-Defeating Moral Theories  41 (as presented in earlier works like the Groundwork and the second Critique) is practically self-defeating in certain interpersonal contexts. Kant’s so-called duties of right, I claim, are best understood on the model of an “indirect deontology.” They are principles that do not follow from the Categorical Imperative, and which are meant to guide us in certain interpersonal contexts precisely because the Categorical Imperative should not do so. If this reading is correct, this means that there is another, hitherto neglected line of convergence between Kant and Sidgwick.

Self-Defeating Moral Theories in Sidgwick and Parfit—and Kant? In Chapter I of Book IV of the Methods of Ethics, Sidgwick describes our target phenomenon: Finally, the doctrine that Universal Happiness is the ultimate standard must not be understood to imply that Universal Benevolence is the only right or always best motive of action. For, as we have before observed, it is not necessary that the end which gives the criterion of rightness should always be the end at which we consciously aim; and if experience shows that the general happiness will be more satisfactorily attained if men frequently act from other motives than pure universal philanthropy, it is obvious that these other motives are necessarily to be preferred on Utilitarian principles. (ME 413) This is an influential, and characteristically clear, statement of the view that a moral theory like utilitarianism may be self-defeating in a practical sense—in the sense that utilitarianism itself may require that we do not make satisfaction of the utilitarian criterion of rightness our conscious aim. Sidgwick gives a range of examples in which the utilitarian criterion of rightness does require motives that differ from impartial benevolence. These include special relations (ME 434), and certain acts or “services” that are more valuable when done out of kindness rather than out of duty (ME 429). Finally, he discusses exceptional situations in which certain types of actions may be conducive to the greatest good which, in more typical cases, are harmful. In these cases, too, Sidgwick argues that utilitarianism may require that most people should believe, guide themselves by, and be taught, simplified rules (an “esoteric morality,” ME 489 f) which ignore these exceptions. Here, too, morality requires that our guiding principles should not be the same as the truth about morally right action. Parfit (1984) distinguishes carefully between different ways for normative theories (moral and non-moral ones) to lead to practical self-defeat.

42  Tim Henning Parfit first assumes that all normative theories can be characterized by “what they tell us to try to achieve,” or our “aims” (1984, 3). The aims prescribed by a theory T are “T-given aims.” The relevant practical forms of self-defeat for a theory T then arise if certain endeavors to achieve T-given aims make it the case that they “will be, on the whole, worse achieved” (op. cit., 5). In Parfit’s further analysis, such self-defeat can be direct or indirect. A theory T is indirectly self-defeating if the attempt to achieve T-given aims makes it the case that these aims will be, on the whole, worse achieved. It is directly self-defeating if successful achievement of a T-given aim will make it the case that our T-given aims will be, on the whole, worse achieved. Next, both direct and indirect self-defeat can be individual or collective. Thus, T is individually self-defeating if the (in the direct case: successful) attempt to achieve T-given aims by one individual will make it the case that this individual’s T-given aims will be, on the whole, worse achieved. T is collectively self-defeating if our (in the direct case: successful) attempts to achieve our T-given aims make it the case that our T-given aims will be, on the whole, worse achieved. According to Parfit, an agent-neutral consequentialist ethics like utilitarianism could not be directly self-defeating. This is because the aim given by that theory is one and the same for all of us. So if an action makes it the case that this aim is, overall, worse achieved, it is by that very same token not itself a successful attempt at achieving the T-given aim. It may, however, still be an attempt to achieve the T-given aim, albeit not a successful one, so indirect self-defeat is very well possible. And in fact, the examples discussed by Sidgwick can be seen to be of the indirect kind. E.g., in the case of “kind services,” the positive effects derive in part from the fact that they are not done from motives of duty. So here, it is the attempt to follow moral principles that must lead to self-defeat. (One is tempted, here, to think of Stocker’s 1976 famous example of visiting a friend in a hospital.) Both Sidgwick and Parfit agree that indirect self-defeat does not count decisively against a theory. Indirect self-defeat may make it the case that a theory is, in Parfit’s terminology, self-effacing. That is, it may recommend, of itself, that it should not be believed by agents. But, as Parfit puts it, it is not the aim of a theory to be believed (see 1984, 11). More generally, as long as a theory does not make it one of its aims that actions should be guided by it, the recommendations of the theory need not pull in different directions, and the theory need not be “failing on its own terms” (ibid.). I now turn to the question whether and how a theory like Kant’s could be self-defeating. I think that direct self-defeat is not possible in this case either. On Kant’s ethics, as on utilitarianism, it could not be the case that an action both causes the given aims to be worse achieved and at the same time successfully achieves these aims itself. The reason is

Self-Defeating Moral Theories  43 interestingly different from the case of utilitarianism, though. The reason is not that all of our actions are subjected to one and the same aim. The reason is, rather, that on Kant’s view, any action serves its own exclusive aim, so to speak. To illustrate: On Kant’s view, I may not tell a lie on a particular occasion even if that would achieve the aim that fewer lies are told overall. This holds both within and across persons. So, first, I may not tell a lie on a particular occasion even if that will keep others from telling more, or greater, lies. (This is the feature of “agent-centered restrictions” that many authors find troubling; see Scheffler 1982.) Second, I may not tell a lie on a particular occasion even if this ensures that I myself tell fewer lies overall. (This feature goes beyond agent-centered restrictions. We might say that on Kant’s view, restrictions are also act-centered. Surely, this is also a feature which many people find puzzling, and which contributes to the so-called “paradox of deontology.” See Heuer 2011.) Accordingly, if a single action of mine conforms to the Categorical Imperative, this is the only aim that is even pertinent to the moral evaluation of that particular act. No other achievements or non-achievements of that theory’s aims by further actions matter to the evaluation of the former. Is there a possibility of indirect self-defeat? This is not clear. Here is one possibility: All moral theories that incorporate so-called agent-­centered restrictions have been accused of being self-centered (see Scheffler, op. cit.). If a moral theory tells me not to tell a lie even if this would lead to fewer lies being told overall, then this theory will lead me to put a special priority on my veracity and to be concerned with keeping my hands clean. However, while such self-centeredness is surely unattractive, it is not clear that it has to conflict with the aims given by the theories in question. And as for Kant’s theory, I have already said that it is act-centered. And the idea that in a moral assessment of an action, that action needs to play a special role does not seem morally problematic. But there may be other ways for Kant’s theory to be indirectly selfdefeating. Authors like Stocker (1976) and Williams (1981) argue that being motivated by respect for the Moral Law gets in the way of good action. To take Stocker’s famous case: If I visit a friend in a hospital and do so out of respect for the Moral Law, this seems incompatible with the personal concern for my friend that would make this a recommendable action. Surely such fetishism for an abstract law would be unattractive. But this objection rests on a misunderstanding of Kant. Especially, it misunderstands Kant’s theory of action. As authors like Reath (2006) and Korsgaard (2008) have explained, we must distinguish between (1) firstorder maxims governing actions and (2) maxims governing the choice of first-order maxims. The Moral Law tells us that we should act only on certain maxims. It is therefore a higher-order maxim, governing the choice of maxims for action. But once it has directed us to a certain

44  Tim Henning maxim for action, that latter maxim gives the reason for which we act. For example: Suppose that the Moral Law directs me to act on the maxim “I visit my friend to see if she is well.” If I am guided by the Moral Law, it will then be the latter maxim which motivates my action. There will be a hierarchy of reasons: My wish to see if my friend is well will be my reason for visiting her in hospital. The Moral Law, by contrast, will be my reason for making this my reason. So far, it is not clear that Kant’s theory is subject to indirect self-defeat. However, I will argue that it is. But I begin with a different way for a moral theory to be practically self-defeating. I will argue that Kant was aware of it, and that it plays an important role in his Doctrine of Right (Rechtslehre). I will then go on to argue that given this threat of selfdefeat, the Categorical Imperative can also be seen to be indirectly selfdefeating in Parfit’s sense. This latter point, I will argue, provides a key to the overall architecture of Kant’s Doctrine of Right, and to his mature moral philosophy in general.

Morality as an Indirect Guide Here is what I have in mind. According to many moral theories, morality can be a guide not just in the direct application to our own actions. It can also be our guide in an indirect fashion by being applied to the actions of others and then determining our reactions. We criticize or blame others because of the moral status of their action; we rebuke our children if they fail to act morally good, etc. And some people try to prevent certain actions of others because they think they are immoral. We can thus distinguish direct and indirect guidance. Of course, indirect guidance and direct guidance typically work in tandem. We show certain reactions to the actions of others on the ground that those target actions have a certain moral status, but we do so because and insofar as we think that these reactions of ours thereby also have a certain moral status (permissible, required). We sometimes treat those who act immorally in ways that we would find immoral if it were not for their prior immorality. The connection between indirect and direct guidance raises interesting issues. The most general question is, If someone else performs an immoral act, what consequence does this have for what I am morally permitted or required to do? If we hold that there is no such consequence, we say that the moral status of others’ actions is never a legitimate business of ours. As I said, most people seem to reject that view. They seem to hold that the fact that someone else acts immorally may make morally permissible at least some form of reaction (criticism, or blame, or avoidance, or advice, or intervention) that would not otherwise be morally permissible. These people agree that morality has an indirectly guiding function, and indeed that it demands to have it.

Self-Defeating Moral Theories  45 I propose that this fact opens up another possibility for moral theories to be practically self-defeating. We can say that a theory T is reactively self-defeating if using it as an indirect guide in our reactions to others makes it the case that the reactions thus guided will fail to achieve T-given aims. (Note that reactive self-defeat thus defined is indirect in Parfit’s sense. I do not say that T actually licenses the reaction, so that it successfully achieves T-given aims and in consequence T-given aims will be overall worse achieved. Rather, it is the attempt to use T as an indirect guide that is self-defeating.) So sometimes, if we react to a person’s action on the basis of our conviction that this person acted immorally, this reaction of ours will thereby be morally objectionable as well. In this case, morality will forbid that we use morality as an indirect guide in the relevant kind of case. Morality will tell us, in these cases, that the immorality of another’s action is none of our concern.

Legal Moralism Are there examples of this kind of self-defeat? This is not uncontroversial, but we can think of examples of moralistic behavior. For example, I think it is immoral to buy and consume meat. But I also think that it would be morally wrong for me to react in certain ways to others on the grounds that they act in this immoral way. I think it would not only be impolite but even morally problematic for me to give strangers lectures, or shame them, if they are ordering steak at the table next to me. Sometimes, it seems, attuning our reactions to others’ actions to the moral quality of these actions can itself be morally problematic. One area in which this concern has played an important role is the philosophy of law. A famous debate between Lord Devlin (1965) and H.L.A. Hart (1963) has centered on a thesis that Hart called legal moralism. As he defines it, legal moralism is the affirmative answer to the question: “Is it morally permissible to enforce morality as such?” (Hart 1963, 4). It is in the nature of the law to enforce certain types of behavior, if necessary through coercive sanctions. The question is whether it is permissible to do so to secure specifically moral conduct. Of course, it is important to note the qualification “morality as such.” It is not disputed that it can be the case that it is permissible to enforce a certain type of behavior which also happens to be required by morality. The question at issue is whether it is morally permissible to coercively enforce or prohibit behavior on the ground that it is required or forbidden by morality. To illustrate the difference, Hart himself cites the case of Mill’s harm principle. Clearly, Mill thought that the harm that results from a certain action can be both a reason to coercively prevent it and a moral wrong-making feature of that action. But the point of Mill’s principle is precisely that the latter fact must not be the reason to use coercion.

46  Tim Henning I claim that Kant was an opponent of legal moralism in this sense. He held this view because the Moral Law, applied as a basis for coercion (legal and otherwise), is reactively self-defeating. I first describe how this view can be seen to be at work in certain crucial distinctions of his late Doctrine of Right. Then I develop a systematic argument for legal antimoralism in the spirit of Kant.

Kant on Duties of Right Kant’s late Metaphysics of Morals (1795) is divided into two parts, the Doctrine of Right and the Doctrine of Virtue.1 In the background of this division is a conceptual distinction between two types of duties, a distinction that Kant has not been explicit about in his earlier work. The first part concerns duties of right (Rechtspflichten, juridische Pflichten), the second, duties of virtue (Tugendpflichten, ethische Pflichten). Both kinds of duty are species of a common genus, which Kant generally denotes as moral duties. In the introduction, Kant tries to explain the new distinction by appeal to other distinctions from his earlier work—e.g., perfect vs. imperfect duties, duties to oneself vs. duties to others, categorical vs. hypothetical imperatives. But I agree with Willaschek (1997) that the relation between the distinction of duties of right and duties of virtue and the other distinctions is opaque. The best interpretive strategy will be to tie the distinction between duties of right and duties of virtue to the two criteria that play significant systematic roles in the Metaphysics of Morals itself. The first criterion is one of indifference to motivating reasons. What duties of right require is only a certain “external behavior,” by which Kant means action in abstraction from the reasons that motivate it. So whether you satisfy a duty of right depends only on what you are doing, not on why you are doing it. By contrast, duties of virtue require the right actions for the right reasons. So whether or not you satisfy a duty of virtue depends not only on what you do but on why you do it—specifically, on whether you are motivated by respect for moral considerations. (In fact, Kant claims that for each duty of right, there is a corresponding duty of virtue which requires the same external action but further demands that conforming to the former duty be the motivating reason.) Here is how Kant describes the distinction in one typical passage of the Metaphysics of Morals: All lawgiving can therefore be distinguished with respect to the incentive (even if it agrees with another kind with respect to the action that it makes a duty, e.g., these actions might in all cases be external). That lawgiving which makes an action a duty and also makes this duty the incentive is ethical. But that lawgiving which does not

Self-Defeating Moral Theories  47 include the incentive of duty in the law and so admits an incentive other than the Idea of duty itself is juridical. (1900ff, VI: 218 f, 45) It should be noted that by duties of law, Kant does not mean positive laws. He insists that he is concerned with rights and duties in a moral sense. He is, Kant says, interested in “the concept of Right, insofar as it is related to an obligation corresponding to it (i.e., the moral concept of Right)” (1900ff, VI: 230, 56). The idea is that morality itself makes certain demands that are indifferent to motives. The second major difference between duties of right and duties of virtue is that the former may legitimately be coercively enforced. As a matter of fact, Kant claims that his doctrine of duties of right just is his doctrine of permissible coercion. He makes this equivalence very explicit, thus: Right should not be conceived as made up of two elements, namely an obligation in accordance with a law and an authorization of him who by his choice puts another under obligation to coerce him to fulfill it. Instead one can locate the concept of Right directly in the possibility of connecting universal reciprocal coercion with the freedom of everyone. . . . Right and authorization to use coercion therefore mean one and the same thing. (1900ff, VI: 232, 57f) So first, duties of right only govern external behavior, in abstraction from motivating reasons, and second, duties of right are precisely those duties that define the limits of permissible coercion. This already shows the relevance of the distinction to our main topic. If duties of right are indifferent to the agent’s motivating considerations, this means that they are indifferent to the maxim of the agent (see, also, 1900ff, VI: 225 f)—i.e., to his or her “subjective principle of action.” But in Kant’s previous moral theory, such maxims are the primary objects of moral assessment. Whether or not we act in conformity with the Moral Law depends precisely on what the maxim of our action is. If we know what an agent does but we do not know his or her subjective principle of action, we cannot decide whether he or she satisfies the Categorical Imperative. But we can, Kant now insists, decide whether he or she satisfies duties of right. This is crucial: The rules that govern whether we may coercively interfere in an agent’s action are indifferent to whether it satisfies the Moral Law. What is the connection between these two criteria? One natural answer, which is explicit at various points in Kant’s text, is that only external behavior can be coercively enforced. You can use force to make an agent do certain things, but force just cannot make her do it for the right reasons.

48  Tim Henning This is certainly part of the rationale. But a glance at the text shows that there is more to it: When one’s aim is not to teach virtue but only to set forth what is right, one may not and should not represent that law of Right as itself the incentive to action. (1900ff, VI: 231, 57) An important detail here is the presence of two modal verbs in the second part, “may not and should not.” Clearly, there is more than one reason not to coercively enforce motives. Similarly: [S]trict Right, namely that which is not mingled with anything ethical, requires only external grounds for determining choice; for only then is it pure and not mixed with any precepts of virtue. . . . [I]f it is to remain pure, this consciousness [of obligation in accordance with law] may not and cannot be appealed to as an incentive to determine his choice in accordance with this law. (1900ff, VI: 232, 57) Again, the final sentence contains two modal verbs. And again, the point is that duties of right may not require certain incentives, not (just) that they cannot. Finally, here is a further relevant passage: But just because ethical lawgiving includes within its law the internal incentive to action (the Idea of duty), and this feature must not be present in external lawgiving, ethical lawgiving cannot be external. (1900ff, VI: 219, 46) This is as explicit as can be: the internal incentive must not be considered in external lawgiving. So Kant’s view is not just that incentives or maxims of an agent cannot be the target of coercion. Rather, they may not or must not be targets of coercion. And the first quotation indicates that what matters is that in appealing to duties of right, “one’s aim is not to teach virtue.” In fact, Kant suggests that the very point of duties of right is that they allow us to avoid such moralizing: Thus, when it is said that a creditor has a right to require his debtor to pay his debt, this does not mean that he can remind the debtor that his reason itself puts him under obligation to perform this; it means instead that coercion which constrains everyone to pay his debts can coexist with the freedom of everyone, including that of debtors, in accordance with a universal external law. (1900ff, VI: 232, 58)

Self-Defeating Moral Theories  49 The last part (“it means instead”) simply states what the appeal to duties of right involves, on Kant’s view. The point I mean to highlight is: If we appeal to such duties, we need not assume the position of the voice of reason. Instead we say that a system of equal freedoms permits us to coerce him. True, this is still coercion. But at least it is not coercion in the guise of rational patronizing. In his interpretation of the Rechtslehre, J. Habermas (1992, 45 ff), too, lays great emphasis on Kant’s idea that the appeal to duties of right is dissociated from moral discourse. But Habermas thinks of this as a pragmatic necessity in modern societies, relieving social interaction from burdens of moral argument. But the quotes suggest that Kant had more in mind. It is not only useful not to have to debate the moral merits of others’ actions. In contexts of coercion, it is not legitimate to do so.

Kant on the Duty to Live in a “Civic State” This claim finds further support if we turn to Kant’s political philosophy. According to Kant, if people live so close that their life paths and plans will inevitably cross, it is their “unconditional and first duty” to enter into a “civic state [bürgerlicher Zustand]” (1900ff, VIII: 289, 73) with one another. Why does Kant think so? To give the answer in advance: Such a state is required because it enforces duties of right, and thus it guarantees that we do not have to use a moral form of coercion. A state with laws can regulate and interfere in behavior without claiming the moral high ground. Before the 1790s, however, there is little evidence of this view in Kant’s writings. On the contrary, in “Idea For a Universal History” (1784), Kant’s account of the state seems to take a Hobbesian form. He appeals to human inclinations, which make it the case that if humans live together in “wild freedom,” the result will be catastrophic (cf. 1900ff, VIII: 22, 46). It bears emphasis, however, that in this text, Kant does not aim for a justification of the duty regarding the state. A little less than a decade later, Kant makes it clear that such observations on human nature, though they may play a role in anthropology and philosophy of history, could not provide the basis for a duty to live in a state. In “On the Common Saying,” the section on Right has the subtitle “Against Hobbes.” Here, Kant stresses that a “civil constitution” is an “absolute and primary duty” (1900ff, VIII: 290; 73), and he insists that this duty is neither instrumental nor grounded in human needs. It is a “requirement of reason.” And as such, of course, it has to be “a priori” (ibid.) in Kant’s view. So there must be an a priori reason why, even abstracting from natural tendencies and inclinations of humans, they need a state with coercive laws. According to Kant, this reason derives from “the concept of freedom.” Here is how he explains it: Man’s freedom as a human being, as a principle for a constitution of a commonwealth, can be expressed in the following formula. No-one

50  Tim Henning can compel me to be happy in accordance with his conception of the welfare of others, for each may seek his happiness in whatever way he sees fit. (1900ff, VIII: 290; 74) The reason we have a duty to live in a state with coercive laws is that not doing so necessarily conflicts with this freedom. Why is this? Suppose that in a state of nature, our plans conflict. What to do? Sometimes, arguments and exchanges of reasons may help. They may lead to an agreement on a judgment concerning what would be best. But sometimes, they may not. How, then, can we resolve conflicts in plans and judgments? Even if our conflicts do not result in the use of force, there is no other way to resolve such conflicts than by a subordination of the judgment of at least one opponent. In irresolvable disagreements in practical judgments, not all of them can prevail. And so someone will have to live by someone else’s judgment concerning what is best, or worthwhile, etc. And this, of course, is what is incompatible with the previous formula. Even if it turns out that the losing party fares extremely well (maybe his opponent was right about what is best for him), he has to live in accordance with the judgment of someone else. This, Kant maintains, is problem enough in itself. If we keep this suggestion in mind, we also get a better understanding of the parallel passage in the Doctrine of Right. At first glance, it may in fact seem that Kant’s justification of the duty to live in a state is Hobbesian in letter and spirit. But we should now see the quotation in a different light: [H]owever well disposed and law-abiding men might be, it still lies a priori in the rational Idea of such a condition (one that is not rightful) that before a public lawful condition is established, individual men, peoples, and states can never be secure against violence from one another, since each has its own right to do what seems right and good to it and not to be dependent upon another’s opinion about this. (1900ff, VI: 312; 124) At first glance, this passage should seem puzzling. Why should we expect violence even if we imagine that people are ideally “well disposed and law abiding”? And even if our nature is such that there are limits to our good dispositions, isn’t this the kind of appeal to empirical, contingent necessities that Kant abjures, as inadequate grounds for what is supposed to be an a priori duty? But if we take the parallel to the passage from “On the Common Saying” seriously, we can see that there is a better way to interpret this passage. We should read this passage, roughly, as follows: The “right to do what seems right and good” is not the source of the violence in the imagined state of nature. Rather, this right is what is violated; it is the target.

Self-Defeating Moral Theories  51 So the danger of violence lies in the fact that in the state of nature, there is no guarantee that one will not depend upon another’s opinion. This makes it clear why duties of right are supposed to be blind to the agent’s motives and convictions. Their very task is to keep people from assessing or evaluating one another’s judgment of what is “right and good.” Their purpose is to keep us from intruding into one another’s minds. This general interpretation of Kant’s views concerning duties of right and their enforcement by the state is not very widespread. But it has been defended by Waldron (1996). Waldron relies on the passage I have just quoted. And he argues that this passage shows that Kant was what he calls a normative positivist about the law. What does he mean? According to Waldron, Kant is a positivist in the sense that for Kant, the content of a legitimate coercive law can be identified independently of moral considerations. This, I think, is right. What makes Kant a normative positivist, by Waldron’s lights, is that Kant insists that there are good normative, indeed moral reasons why the law should be independent in this way. According to Waldron, Kant held that the moral purpose of the law is precisely to avoid taking a stand on the disagreeing parties’ judgments. I will now explain why Kant should have taken this position. I will argue that the reason is precisely the kind of reactive self-defeat I have described. If coercion is based on a judgment of the moral merits of the agent’s plans, then this coercive action will itself violate this Moral Law.

A Systematic Argument: Deliberation-Invasive Coercion Kant’s moral theory provides a systematic reason why it cannot be permissible to coercively interfere in others’ actions on the ground that those target actions are immoral. To see how, note first that such coercion would have to be sensitive to the maxims of the target actions. An action’s maxim is, after all, that which determines the moral status of an act. So if these moral statuses are the decisive criteria for our use of coercion, then this coercion will track agent’s maxims, as follows: Deliberation-Invasive Coercion If we coercively interfere in an action on the basis of its maxim, then we would have interfered in other behavior if it had been done from the same maxim, but we would not necessarily have interfered in that same behavior done from another maxim.

Such coercion does not, in the first instance, constrain how others act, but the reasons from which they can decide to act. It aims to constrain the range of reasons that can successfully lead to action. Coercive intervention that is sensitive to maxims in this way will be defined as deliberation-invasive.

52  Tim Henning If we use moralistic coercion, we act on a maxim that threatens others with deliberation-invasive coercion. What I want to suggest is that no such maxim can be willed as a universal law. Why? Famously, Kant held that when we deliberate and decide on our own maxims, we must do so “under the idea of freedom.” That is to say: “[O]ne cannot possibly think of a reason that would self-­ consciously receive directions from elsewhere with regard to its judgements” (1900ff, IV: 448). This makes it clear that Kant does not claim that we must assume that we will never encounter any external obstacles in our attempt to act on our maxims. But we have to think of the process of reasoning, of deliberating and judging, as free and unconstrained. If we weigh reasons, we must presuppose that we are solely guided by the weight of the reasons that present themselves to us. If we think of the outcome of our thinking as in part determined by other forces (which lack the significance of reasons, defeaters, etc.), we cannot think of our activity as a balancing or weighing of reasons at all. But this would mean that I cannot at the same time think of myself as a reasoning agent and as being subject to deliberation-invasive external force. For I would know that whatever judgment I reach regarding the weight of the relevant reasons, it will lead to action only insofar as it satisfies an independent criterion. In particular, it will lead to action only if it conforms to what someone else thinks is the rational and moral conclusion. And this changes the understanding of the whole process. It is clear that I will act in accordance with someone else’s judgment of what should be done. So why would I even weigh normative reasons? If anything, I could use them as a basis for a prediction about what I will be forced or allowed to do, not as a basis for a judgment about what to do. Thus, I cannot conceive of myself as reasoning and as subject to deliberation-invasive force, simply because I cannot conceive myself as reasoning and as “receiving directions from elsewhere” as regards the constitutive aims and rules of this reasoning. This also means that I cannot conceive of myself as reasoning and as being subject to a universal law of deliberation-invasive intervention. But with this lemma in hand, we have all the ingredients for a general moral argument:

Kantian Argument for Legal Anti-Moralism The moral status of an action is determined by its maxim.   Intervention based on the moral status of the target action is therefore based on the maxim of the target action.   So a maxim of this type of intervention will make someone else subject to deliberation-invasive coercion.   I cannot rationally will a maxim while at the same time thinking that I am subject to deliberation-invasive coercion.

Self-Defeating Moral Theories  53   Thus, I cannot rationally will my maxim of intervention based on the moral status of the target action, and at the same time will that maxim as a universal law.   Therefore, a maxim of intervention based on the moral status of the target action is immoral.

In other words: Coercion that is guided by the moral status of the target action will require an asymmetry in the ability to act on one’s own judgment. Thus, it cannot be willed as a universal law.

Repercussions of the Argument for Kant’s Political Philosophy So there is a reason why duties of right are blind to those features of actions that the Categorical Imperative focuses on. In Kant’s ethics, enforcing morality as such, in Hart’s terms, is itself immoral. So far, this does not amount to self-defeat in any of the senses that Parfit has defined. For all we have said, acts of coercion may perfectly well be directly guided by moral considerations (in the sense described earlier). While the moral status of the target action may be none of our business, it may still be the case that acts of coercion can (or must) themselves be based on moral reasons of their own. In other words, it still seems perfectly legitimate to let the Categorical Imperative guide our acts of coercion, as long as it does not do so via a moral assessment of the target action. I will now argue that even this possibility is ruled out. To put it briefly: In Kant’s moral theory, coercion cannot be directly guided by the Moral Law without being indirectly guided by it. When it comes to coercion, the moral evaluation of our act and of the target act go hand in hand. To begin, however, we must take a look at the relation between duties of right and the Moral Law. This relation is notoriously difficult to understand, and there is great controversy not only on the details but even on the major questions. In a recent monograph on the issue, Horn (2014) observes that there is a fundamental opposition between a Separation Thesis (Trennungsthese) and a Dependence Thesis (Abhängigkeitsthese). The former Thesis denies, while the second thesis affirms, that duties of right are derived from, or based on, the Moral Law (or the Categorical Imperative). Both of these theses count some of the most prominent Kant scholars among their adherents. In Horn’s assessment, the purely textual evidence, all in all, weighs in favor the Dependence Thesis; but when it comes to substantive and systematic considerations, there is much to be said for the Separation Thesis (cf. op. cit., 9). Horn’s conclusion, with which I agree, is that this situation suggests that Kant’s actual views must be more complex than the typical interpretations envisage.

54  Tim Henning One reason in favor of the Separation Thesis is Kant’s claim that there is an absolute duty to obey the positive law of the state without so much as questioning its moral validity. (For evidence and discussion, see Horn 2014, 18 and passim.) But a second and maybe even more important reason is Kant’s own systematic derivation of duties of right. He explicitly does not derive them from the Categorical Imperative, but from another principle called the Universal Principle of Right, which says: “[S]o act externally that the free use of your choice can coexist with the freedom of everyone in accordance with a universal law” (1900ff, VI: 231, 56). This principle is not equivalent to the Categorical Imperative. Just like the more specific duties of right, it governs only external actions. Thus, it does not ask whether my maxim in acting can be universal but whether my external action can be universal. Kant stresses that this Universal Principle itself is not derived from the Categorical Imperative. Instead, he calls it a “a postulate that is incapable of further proof” (ibid.) (and so, a fortiori, incapable of being derived from the Categorical Imperative). Finally, Kant observes that there are substantive differences between the principles. The Universal Principle of Right does not yield any duties of right regarding lying and suicide, which Kant, notoriously, took very seriously. So it certainly seems that Kant is very explicit about a certain kind of independence between duties of right and the Moral Law. At the same time, of course, this is quite puzzling. Why does Kant stress this independence? Why does he bother deriving principles of legitimate coercion from a new “Universal Principle”? After all, he is convinced that he has found the true Moral Law! This has been a source of frustration. In an oft-cited passage, Brown expresses it as follows: [A]ll Kant needs to do in order to complete his program in philosophy of law is to show how the Categorical Imperative may be used to test the moral status of the rules in a body of positive law. If the test is met, the law is what it ought to be. If the test is failed, the law is morally defective and ought to be changed. At this point in the argument, Kant’s task seems almost certain of accomplishment. . . . But in fact, the argument is never advanced beyond this point. Instead of showing how the Categorical Imperative may be applied to test the rules of positive law, Kant introduces a number of different principles which range in degree of generality between the extremes of the Categorical Imperative and the rules of positive law. Many of these principles have no discernible logical relationship to the Categorical Imperative and no clear application to positive law. (Brown 1962, 36)

Self-Defeating Moral Theories  55 So why is Kant’s derivation of duties of right keeping a distance to the Moral Law? What I want to suggest is that my previous argument about reactive self-defeat may help to illuminate this question. Again, I have argued that Kant is an opponent of legal moralism. He holds that it is not morally permissible to enforce morality as such. I will now suggest that this view has important ramifications for Kant’s general moral framework. I begin with the following observation: Acts of intentional coercion, as such, are prima facie problematic in Kant’s ethics. If I decide to prevent you from performing a certain action, I give priority to my verdict as to what you should do over your verdict. I thus assume a certain authority over your actions, and I have to deny you that same authority over your actions. I keep you from deciding whether you φ; but since I evidently count on my ability to decide whether you φ, I cannot will that you, too, keep me from deciding whether you φ. Clearly, this asymmetry is bound to generate problems when it comes to universalizability. Famously, Kant’s view on the matter is this: Since we each see ourselves as having the final authority over our decisions, we must (“as a consequence of just the same rational ground that also holds for me” (1900ff, IV: 429)) grant others the same authority over their decisions. This is why, as I said, coercive interference is prima facie morally problematic. Now, the crucial point is this: If I am to see my coercive actions as morally permissible, I can only do so if there is something special about the target action that makes it permissible in this case. In fact, it seems that my coercive interference does not damage your authority in a unilateral way only if you did not have that authority in the first place. In other words: If the maxim of your action cannot be willed as a universal law to begin with, then my maxim of preventing this action would not, as a universal law, be a problematic restriction. Kant, it seems to me, sees this point. This is why he writes: Now whatever is wrong is a hindrance to freedom in accordance with universal laws. But coercion is a hindrance or resistance to freedom. Therefore, if a certain use of freedom is itself a hindrance to freedom in accordance with universal laws (i.e., wrong), coercion that is opposed to this (as a hindering of a hindrance to freedom) is consistent with freedom in accordance with universal laws, that is, it is right. Hence there is connected with Right by the principle of contradiction an authorization to coerce someone who infringes upon it. (1900ff, VI: 231, 57) I think Kant should be understood as having in mind the stronger point that there is authorization to coerce only if someone infringes upon freedom in accordance with universal laws. In other words, if we are to think of a coercive action of ours as legitimate, we must also take a certain

56  Tim Henning view of the target action, namely as one that is—in itself—in conflict with freedom under universal laws. Now Kant’s own description is not about moral evaluation. It is concerned only with the status of conforming to duties of right. And he claims that if a coercive action conforms to them, that can only be because the target action does not. But I have suggested that when it comes to the Moral Law, something similar holds: Given that we necessarily assume authority over our own decisions, we can will a maxim of coercive interference as a universal law only if we assume that something about the target action makes it the case that its maxim cannot so be willed. Thus, if I think of my coercion as morally permissible, I have to think of the target as morally impermissible. But this latter thought, we have seen, should not be part of our reasons in coercive decisions, on pain of reactive self-defeat. (Again, if I coercively interfere because I think of the target as morally impermissible, this itself is immoral.) Hence, when it comes to coercive interference in others’ actions, there is also indirect self-defeat in the sense described by Parfit. If my coercive action is a conscious attempt to obey the Moral Law, then this makes it the case that it fails to obey this Law. If this is right, then there is a compelling reason for Kant to appeal to a different supreme Principle in his account of legitimate coercion—i.e., of Right. If we justify coercion in terms of that latter principle, this justification is not based on any judgment about our maxim and its relation to the Categorical Imperative and also not on any corresponding judgment about the maxim of the target action. We justify our coercive act solely by pointing out that it can coexist with the external actions of others under universal laws. So if the Universal Principle of Right is different from the Moral Law, we have justified our action while successfully avoiding taking a stand on moral matters.

Practical Postulates and “Indirect Deontology” Where does this idea leave us with regard to the general interpretive issues mentioned earlier? On the one hand, it seems to support the Separation Thesis. After all, I have just argued that there is a good reason why the Principle of Right itself should not be seen as derived from the Moral Law. On the other hand, the reason I have given is itself a moral reason. My diagnosis has been one of indirect self-defeat, which means that the Moral Law itself requires that we must sometimes not take it as a guide. In fact, I think that this opens up an interesting new interpretative possibility. Maybe the correct view of the architecture of Kant’s moral philosophy will take a shape that we know from the utilitarian tradition. Maybe, that is, Kant has offered us an indirect moral theory. On such a theory, there is a criterion of rightness (in Sidgwick’s terms), but there is also a

Self-Defeating Moral Theories  57 further set of principles. These latter principles are not logically related to the criterion of rightness. The relation is a different one. In Parfit’s terms: If we attempt to fulfill the aims given by the further principles, we better succeed in fulfilling the aims given by the criterion of rightness. If the reason for this is that the criterion of rightness is indirectly self-defeating, these other principles have more than an instrumental status. It is morally obligatory to let these principles be our guides and motives. Could anything like this be Kant’s view? It is hard to tell. But in addition to the systematic argument developed earlier, there is at least one interesting piece of textual evidence. Kant, after all, calls his Universal Principle of Right a “postulate.” In the context of his practical philosophy, this term has a rather definitive technical meaning, which is wellknown from the so-called doctrine of postulates in the second Critique. There, Kant claims that the immortality of the soul and the existence of god are postulates. And his point is precisely this: The claims that the soul is immortal and that god exists are not in any way logical consequences of the Moral Law (unlike the claim that we are transcendentally free). Still, any human agent who is properly motivated to conform to the Moral Law will subscribe to these beliefs. I suggest that when Kant calls the Principle of Right, and presumably also its consequences, a postulate, he has something similar in mind. He holds that although these are not logically derived from the Moral Law, any agent who is properly motivated to conform to the Moral Law will accept these principles and use them as a guide in questions of coercion instead of the Moral Law. In this sense, Kant’s view of coercion is an indirect deontology.

Note 1. All references to Kant’s writings in this chapter are to the original German in the Academy Edition (Ausgabe der Akademie der Wissenschaften).

References Aristotle. 2000. Nicomachean Ethics. Translated by R. Crisp. Cambridge: Cambridge University Press. Brown, S. M., Jr. 1962. “Has Kant a Philosophy of Law?” The Philosophical Review 71: 33–48. Devlin, P. 1965. The Enforcement of Morals. Oxford: Oxford University Press. Habermas, J. 1992. Faktizität und Geltung. Beiträge zu einer Diskurstheorie des Rechts und des demokratischen Rechtsstaats. Frankfurt a. M.: Suhrkamp Verlag. Hart, H. L. A. 1963. Law, Liberty and Morality. Oxford: Oxford University Press. Heuer, U. 2011. “The Paradox of Deontology, Revisited.” In Oxford Studies in Normative Ethics, edited by M. Timmons, 236–267. Oxford: Oxford University Press.

58  Tim Henning Horn, C. 2014. Nichtideale Normativität. Berlin: Suhrkamp. Kant, I. 1900ff. Kants Werke. Ausgabe der Akademie der Wissenschaften. Keller, S. 2007. “Virtue Ethics is Self-Effacing.” Australasian Journal of Philosophy 85: 221–231. Korsgaard, C. M. 2008. “Acting for a Reason.” In The Constitution of Agency, edited by C. M. Korsgaard, 207–229. Oxford: Oxford University Press. Parfit, D. 1984. Reasons and Persons. Oxford: Clarendon Press. Reath, A. 2006. Agency and Autonomy in Kant’s Moral Theory. Oxford: Clarendon Press. Scheffler, S. 1982. The Rejection of Consequentialism. Oxford: Clarendon Press. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Stocker, M. 1976. “The Schizophrenia of Modern Ethical Theories.” Journal of Philosophy 73: 453–466. Waldron, J. 1996. “Kant’s Legal Positivism.” Harvard Law Review 109. Willaschek, M. 1997. “Why the Doctrine of Right Does Not Belong in the Metaphysics of Morals.” Jahrbuch für Recht und Ethik 5: 205–227. Williams, B. 1981. “Persons, Character and Morality.” In Moral Luck: Philosophical Papers 1973–1980, edited by B. Williams. Cambridge: Cambridge University Press.

Part II

Metaethics

3 Sidgwick and Kant on Practical Knowledge and Rational Action Carla Bagnoli

While few philosophers in the early analytic tradition read Kant with charity, most readily agreed with John Stuart Mill about the extent of his “grotesque failure” (ME xix). Henry Sidgwick is an exception. His masterpiece The Methods of Ethics shows that his project for the foundation of practical knowledge takes shape in response to distinctively Kantian questions. Sidgwick is struck by a feature of Kant’s account of practical knowledge that has only recently become the focus of debate in analytic ethics and action theory, that is, that there is constitutive relation between norms governing the self and norms governing agency (ME 516). He believes that Kant’s conception of autonomy does not ultimately succeed in vindicating this idea, but he agrees that Kant’s project of grounding morality on rationality is the primary task of philosophical ethics.1 Like Kant, Sidgwick holds that the foundational question must be settled by considering internal standards of moral judgment, that is, standards that concern action as intentionally planned, rather than its outward aspects, such as, e.g., its consequences. In this chapter, I compare and contrast Kant’s and Sidgwick’s arguments in defense of moral cognition as objective practical knowledge. Kant focuses on practical truths in terms of practical laws governing the mind in action, while Sidgwick is concerned with practical truths about action. I argue that this is a crucial difference in the understanding of practical knowledge, which is matched by a different understanding of moral phenomenology and of the significance of subjective experience in accounting for the authority of moral obligations. Key to these differences is a more fundamental divergence regarding the nature of reflection, its dynamics, and its impact on agency (ME 383).

I. Moral Knowledge as Practical Knowledge: A Philosophical Task Sidgwick and Kant share a commitment to prove the objectivity and rational authority of moral knowledge. They both think of moral knowledge as “practical” and relevant for planning actions. Correspondingly,

62  Carla Bagnoli they share the task of vindicating the practical significance of ethical judgments as ordinarily conceived, namely, as conveying practical knowledge. That moral judgments motivate is a claim they share with Hume, but they offer an alternative explanation of moral motivation, which centers on the efficacy of reason. In contrast to Hume, they believe that ethical problems are amenable to rational treatment and invoke a rationalist method for resolving moral disputes.2 Both think that universality is the form of practical knowledge, and although they both take seriously common moral cognitions, neither of them grant common sense ultimate authority. Rather, they both consider a prominent task of philosophy that of elucidating, systematizing, and providing rational justification for what is commonly held as a moral truth. Finally, they aim at the same polemical targets, in that they intend to avoid the pitfalls of dogmatism which mars some versions of ethical intuitionism and rationalism on the one hand, and the risks of subjectivism and skepticism on the other hand. Their argument implicates the falsity of reductive naturalism as a model for understanding the objectivity and authority of moral truths. What is striking and instructive about these philosophers’ agenda is the claim that the objectivity of moral cognition should be vindicated along with its practicality. Contrary to the mainstream meta-ethics of our days, they do not presume any tension between the alleged practicality and objectivity of moral judgment. In fact, they think these features of ordinary moral judgment would go hand in hand and call for a philosophical explanation. While they are both objectivist about practical knowledge and rationalist about the methods of ethics, hence denying the epistemic authority of ordinary morality, Kant and Sidgwick sharply differ in their philosophical explanation of moral judgment. It is against this shared meta-ethical background that the most interesting differences become salient: they ultimately concern the way Sidgwick and Kant understand the practical significance of ethical judgment in what they take to be its paradigmatic form, i.e., the form of ought-judgment containing moral obligation. In what follows, I contrast Sidgwick’s and Kant’s views on moral judgment as conveying practical knowledge (sections 2 and 7), their negative arguments against competitive views of moral judgment (sections 3, 4, 8, and 9), their positive argument for understanding moral knowledge as practical knowledge (sections 5 and 10), and their prospects of succeeding in doing so (sections 6 and 11). My conclusion is that Kant’s theory of moral cognition as conveying objective practical knowledge is more promising than Sidgwick’s because it establishes an internal relation between rational agency and the rational requirements of morality.

II.  Sidgwick’s Definition of Moral Judgment In the Introduction to the 7th edition of The Methods of Ethics, Sidgwick asks why we do not simply endorse some first-order moral principle

Practical Knowledge and Rational Action  63 and state what ought to be done, without considering also how we learn such moral truths and by which route or faculty we can apprehend them. Just as mathematicians and physicists focus on answering their respective questions rather than examining the mental faculties that allow them to do so, perhaps moral philosophers should simply focus on answering ethical questions rather than investigating the faculties we use to arrive at these answers (ME ch. 1). Sidgwick proposes an explanation: One answer is that the moralist has a practical aim: we desire knowledge of right conduct in order to act on it. We cannot help believing what we see to be true, but we can help doing what we see to be right or wise, and in fact often do what we know to be wrong or unwise: thus we are forced to notice the existence in us of irrational springs of action, conflicting with our knowledge and preventing its practical realization: and the very imperfectness of the connexion between our practical judgment and our will impels us to seek for more precise knowledge as to the nature of that connexion. (ME 5) The thought is that human agents need guidance and instruction as to what to do because of their moral and cognitive imperfection, and, perhaps more importantly, they need aids to comply with moral demands, since the human will is weak and the emotions recalcitrant. The primary practical function of moral judgments is to bind the agent’s will. To adequately account for this practical function, ethical theory must explain how moral judgement exerts its authority. To take the practical question into consideration, Sidgwick focuses on a specific category of moral judgments, namely, those containing the terms “ought” and “right,” which entail “an authoritative prescription” (ME 106). Furthermore, Sidgwick distinguishes two uses of ought-judgments. In the narrow sense, ought-judgments are “always thought capable of being brought about by the volition of any individual to whom the judgment applies. I cannot conceive that I ‘ought’ to do anything which at the same time I judge that I cannot do” (ME 33). Instead, in the wide sense, ought-judgements refer to the stance of the wiser judge: “I ought to know what a wiser man would know, or feel as a better man would feel, in my place, though I may know that I could not directly produce myself such knowledge or feeling by any effort of the will” (ME 33). In the former case, the judgment is prescriptive. In the latter case, instead, ought judgments convey a normative assessment from a third-person stance, which highlights the limitations of agency in regard to higher normative standards. Both kinds of ought-judgments convey knowledge,3 and thus command agreement; by judging that something ought to

64  Carla Bagnoli be done, one implies that every rational agent would converge: “what I judge ought to be must, unless I am in error, be similarly judged by all rational beings who judge the matter” (ME 33). In sum, both kinds of ought-judgments claim universal authority and demand coherence in their applications across contexts and across individuals.4 Furthermore, moral judgments demand the same emotional responses from all rational beings, and if the conduct approved is “really right,” “it cannot, without error, be disapproved by any other mind” (ME 27, 33). In the narrow sense, ought-judgment seems directly implicated in action, and exhibits two distinctive features. First, it applies when the agent is capable of bringing about what she ought to bring about (ME 33). In order for moral knowledge conveyed by the judgment to be efficacious, the “ought implies can” principle should be established. Second, the action demanded is a dictate of reason so that “in rational beings as such this cognition gives an impulse or motive to action” (ME 34; see, also, 77). There is a scholarly disagreement as to whether and how Sidgwick is thoroughly committed to internalism; but the passage states that reason is capable of guiding action by producing a motive, at least typically.5

III. Sidgwick’s Argument Against Naturalism: The ‘Reversed Burden of Proof’ It is in this connection that Sidgwick mounts his attack on naturalism. He urges that the notion of ‘ought’ or ‘moral obligation’ as used in our common moral judgments, does not merely import that there exists in the mind of the judging person a specific emotion (whether complicated or not by sympathetic representation of similar emotions in other minds); nor that certain rules of conduct are supported by penalties which will follow on their violation (whether some penalties result from general liking or aversion felt for the conduct prescribed or forbidden or from other source). (ME 32) Contrary to these views, and in a Kantian spirit, Sidgwick sets out to show that ought-judgements are inherently obligating, and their authority cannot be explained away on the basis of some naturalistic reduction. Right and wrong cannot be defined because “the notion which these terms have in common is too elementary to admit of any formal definition” (ME 32). Sidgwick is content with the claim that we “must accept as elementary what introspection carefully performed declares to be so; and using this criterion I find that the notion we have been examining, as it now

Practical Knowledge and Rational Action  65 exists in our thought, cannot be resolved into any more simple notions” (ME 32). Thus, the task of philosophical exploration is not to provide a reductive analysis of the key moral terms in terms of other notions, or to give a formal definition of them, but to elucidate and clarify these concepts “by determining as precisely as possible its relation to other notions with which it can be connected in ordinary thought” (ME 33). The pressing question is how the defense of the practical function of reason, hence of the practical significance of moral judgments, can be sustained. Sidgwick is clear about how it cannot be sustained. His remarks against Jeremy Bentham’s definition of “right” in terms of what is conducive to the general happiness have been regarded as a first sketch of the anti-naturalist move akin to G.E. Moore’s open question argument (ME 26n).6 Against reductive hedonism, Sidgwick says that “good” cannot mean “pleasurable” because “good means pleasurable” is a significant proposition and not a mere tautology (ME 109). Sidgwick points to an unappealing trait of linguistic naturalism, which purports to establish the semantic equivalence of two different concepts, but his general argument is that the naturalistic models of moral judgments leave unaccounted the normative force of the precepts of reason, which cannot be explained in terms of some given natural properties.7 It is, then, the claim about the practical import of reason that rules out naturalist reductions. Moral concepts have cognitive contents and are understood as representations, but they are not reducible to representations of empirical facts. This is what blocks any naturalistic reduction.8 Focusing on the primary prescriptive meaning of ought judgment, Sidgwick reverts the burden of proof in dealing with moral skepticism. Hume holds that the possibility of the rational knave shows that reason does not motivate, and it does not affect the will and impel action. Taking the features of ordinary moral judgment at face value allows Sidgwick to hold that the burden of proof does fall onto the skeptic. Typically, reason commands and prescribes: the normative force of moral judgments is firmly based on “the reflections of individuals on their practical judgements and reasoning” (ME 25; cf. Deigh 1992; Shaver 2000; Phillips 2011). It is for the Humean skeptic about practical reason to prove that this is not so, and that reason needs a concomitant congruent desire in order to fulfill its practical function, despite the appearances.

IV. Sidgwick’s Critique of Dogmatic and Perceptional Intuitionism The critique of naturalism is only part of the negative argument for practical reason. Sidgwick builds upon a critique of Intuitionism. Intuitionism is generally taken to involve three claims: (1) a semantic claim

66  Carla Bagnoli that normative concepts are unanalyzable in terms of non-normative concepts; (2) an ontological claim that normative properties are irreducible to non-normative properties; and (3) the epistemological claim that to know right from wrong, one needs to have a special sense or faculty because normative properties are not natural properties, hence not discoverable or apprehended by empirical methods of investigation. Importantly, the semantic and epistemological claims are grounded on the ontological claim. This sort of Intuitionism has been the target of many criticisms.9 Its main difficulty is that it does not explain the normative dimension of ethical judgments, precisely because it equates practical knowledge to a sort of perception, hence implausibly assuming the existence of nonnatural properties which are inherently motivating.10 It is dubious that Sidgwick subscribes to anything like the intuitionist view that attracts such criticisms, however. To begin with, it is unclear even what the professed ontological commitments of Sidgwick’s Intuitionism are. In the Preface to the 1st edition, for example, he holds the view that for any set of circumstances, there is something “right or reasonable to do, and that this may be known” (ME1 v). It is not at all obvious that this claim counts as an ontological commitment to the existence of “reasonable properties.” This non-committal interpretation seems confirmed by the following passage: “I make no further assumption as to the nature of the object of moral knowledge,” and the general assertion that ethics lacks “some department of actual existence for its subject matter” (ME1 vi, 2). Second, Sidgwick uses the metaphor of vision but nowhere defends any special sense or faculty akin to perception.11 Nor does he seem to think that we apprehend practical truths by engaging in peculiar activities of the mind, e.g., projecting sentiments or applying normative concepts to a valueless reality. He rejects Perceptional Intuitionism, the view that in each situation of choice what is right is simply “seen” or “apprehended” (ME 99–100), on the grounds that immediate individual judgments on a particular case are often unreliable, vary across time and across competent evaluators, and are subject to the distorting influence of emotions.12 Ultimately, such a view displaces the role of rationality in ethics, dismissing “as superfluous all modes of reasoning to moral conclusions” (ME 100). He also rejects Dogmatic Intuitionism, the view that certain general rules implicit in common-sense morality can be discerned with really clear and finally valid intuition, on the ground that it fails to satisfy the requirements of methodological adequacy (ME 101). Such requirements are the following: (1) that intuitions are clear and precise, (2) that they are ascertained by careful reflection, (3) that they are consistent with other propositions held as self-evident, and (4) that they are free from radical disagreement (ME 338–342).13

Practical Knowledge and Rational Action  67

V.  Sidgwick’s Philosophical Intuitionism The methodological critique of Perceptional and Dogmatic Intuitionism does not undercut the claim that some normative propositions are “absolutely and undeniably true and evident” (ME 102).14 There is a scholarly disagreement about the exact number of practical truths elected by Sidgwick, and whether they do satisfy the methodological desiderata that Sidgwick put forward. The following is a representative list, which includes five of them:15 1. “it cannot be right for A to treat B in a manner in which it would be wrong for B to treat A, merely on the ground that they are different individuals” (ME 380); 2. “the mere difference of priority and posteriority in time is not a reasonable ground for having more regard to the consciousness of one moment than to that of another” (ME 381); 3. “the good of any one individual is of no more importance, from the point of view (if I may say so) of the Universe, than the good of any other” (ME 382); 4. “as a rational being I am bound to aim at good generally . . . not merely at a particular part of it” (ME 382); 5. “each one is morally bound to regard the good of any other individual as much as his own, except in so far as he judges it to be less, when impartially viewed, or less certainly knowable or attainable by him” (ME 382). These intuitions are supposed to identify a third variety of Intuitionism, Philosophical Intuitionism, which can deliver a fully rational system of morality and explain why a certain set of practical norms is the correct one (ME 102). To say that such intuitions are self-evident means that it suffices for justified belief in them that they are properly understood. (ME 229; see, also, Crisp 2015, 107–109 and Crisp’s chapter in this volume). As for the route by which we come to know them, it is by “direct reflection” (ME 383). Sidgwick does notice that intuitions are not infallible (ME 211, 400). This feature marks a difference between Philosophical and Dogmatic Intuitionism, which is committed to the existence of propositions implicit in ordinary moral reasoning (ME 101) and holds that ordinary people have “the power of seeing clearly that certain kinds of actions are right and reasonable in themselves, apart from their consequences, from which other consequences admitted to be possibly good or bad are definitely excluded” (ME 200). Dogmatic Intuitionism, we may say, assumes a deontological format, in stark contrast to consequentialism. This contrast does not appear central to Philosophical Intuitionism. In fact, Sidgwick holds that the self-evident philosophical intuitions provide

68  Carla Bagnoli the proof of Utilitarianism: they are supposed to determine the intellect to accept Utilitarianism (ME 387, 406–407).16 Whether Sidgwick succeeded in providing such a proof is a matter of contention. Critics object that the philosophical intuitions that Sidgwick retains as self-evident fall short of his own criteria, and in particular those in whose confidence the acceptance of Utilitarianism rests.17 I am inclined to say that his argument for utilitarianism is ultimately unsuccessful insofar as he is unable to resolve the dualism of practical reason, which makes the choice between egoism and utilitarianism a reasonable one. However, my main concern is whether such philosophical intuitions could ground the sort of objectivity and practical significance that Sidgwick sets out to vindicate for moral knowledge. Such philosophical intuitions are not the sort of first-order principles implicit in ordinary moral reasoning, and hence they cannot guide action in the way in which first-order moral cognitions do. They are theoretical in nature, and their role is to furnish a “permanent basis of the common conviction that the fundamental precepts of morality are essentially reasonable” (ME 383).

VI. Sidgwick’s Failure to Establish That Moral Knowledge Is Practical Intuitionism and Utilitarianism can both be recognized as legitimate methods of ethics once their domain is aptly differentiated: Intuitionism accounts for the irreducibility of fundamental self-evident moral truth, while Utilitarianism accounts for how we reason and solve moral problems.18 The question is how the philosophical intuitions listed earlier stand in relation to such methods of justification. If all moral truths are apprehended by moral reasoning, whether moral reasoning takes the shape of deontological, consequentialist, or virtue-ethicist methods, it means that such moral truths are not directly apprehended: hence, they are not self-evident intuitions at all. The term “self-evident” means that the moral propositions are “immediately known to be true” (ME 98), and there are no truths of this sort because all particular moral truths are derived from the philosophical intuitions. Self-evidence concerns propositions and is characterized in the following ways, which are listed in order of strength: (1) truth is “manifest” (ME 379); (2) a matter of “apprehension” (ME 383); (3) “immediately known to be true” (ME 98); (4) “truth is known immediately, and not as a result of reasoning” (ME 211). If philosophical intuitions are second-order claims, it is hard to see how Sidgwick keeps faith to the claim that practical knowledge is intuitional, in contrast to knowledge mediated by inference and reached by reasoning (selfevident propositions are such whose “truth is known immediately, and not as a result of reasoning” (ME 211).19 If he means that there must always be a foundational, non-inferential intuition even in cases of

Practical Knowledge and Rational Action  69 practical inference, then he would ultimately hold a dogmatic form of Intuitionism. It is also hard to see how Sidgwick can defend the claim that some moral truths are self-evident in themselves, e.g., regardless of reflection on the consequences of action. Furthermore, it is hard to see how this core mark of Intuitionism can be salvaged by referring to the philosophical intuitions that serve as axioms in practical reasoning. It appears that in order to preserve the claim that there are self-evident truths that are not only manifest (ME 379) but also “immediately known to be true” (ME 98), independently of external standards about the effects (ME 211), one must appeal to some form of Dogmatic Intuitionism. This is to say that Philosophical Intuitionism is no Intuitionism at all, unless it confirms and relies on some first-order moral principles that makes it akin to Dogmatic Intuitionism.20 The failure to ground the core claim of Intuitionism is also a failure to displace naturalist reduction. It is a failure to explain that and why moral judgments have practical relevance, and why moral motivation is not reducible to the mechanic workings of psychology. Ultimately, we are provided with a definition of practical knowledge that is too peripheral to rational agency and so, in some important sense, not practical at all. The justification by the way of philosophical intuitions tells us nothing about the struggle of the will to attain an ideal that is normative for it, nor does it tell us how the will is educated, shaped, and guided by rational norms. In short, this theory fails to account for all moral phenomena that it sets out to explain insofar as they are irreducible to psychological events. This is the way in which we are let down in the debate against moral skepticism because—by Sidgwick’s own admission—we have lost the cause in the debate about the possibility of practical reason. This failure to support the objectivity and authority of moral knowledge ultimately depends on the fact that Sidgwick is speculating on the nature (the content) of the propositions that carry practical knowledge, rather than on the dynamics of judging and willing rationally. This is a setback in relation to Kant’s own achievement. Sidgwick’s theoretical understanding of practical knowledge severs the links between reason and agency, and thus exposes the gap between knowing the commands of reason and the reasons to act upon them. In which sense, then, moral knowledge as knowledge of moral truths is practical?

VII. 

Kant’s Conception of Moral Judgment

For Kant, ethical judgments are not propositions about the right standards of action in the sense conveyed by Sidgwick.21 Kant aims to capture our ordinary understanding of moral judgment as conveying practical knowledge of moral requirements. For a moral judgment to be valid, it needs to be universally valid. The implicit universality of moral judgment

70  Carla Bagnoli is to be understood in reference to the subject (i.e., it binds all relevant practical subjects) and to the object of judgment (i.e., it is based on grounds that are valid in all relevantly similar situations). This feature of universality generates agreement among all rational agents about the strict requirements of morality, in the same circumstances of action at all times.22 The act of judgment is an exercise of reason, understood as a as spontaneous, a self-consciously active power. Both in its theoretical and practical understanding, judgment conveys knowledge and thus is cognitive. While theoretical cognition involves also receptive capacities, cognition in itself always bears the mark of autonomy, as it is the result of the self-conscious activity of reason. It is the result of a self-conscious activity, even though in the theoretical case, it also depends on some receptive powers. Cognition implicates an agential effort to cognize and thus counts as an achievement of sort, to be imputed to reason. This is particularly clear in the case of moral judgement. There is a difference between practical and theoretical knowledge. In both cases, there is an exercise of agentive powers. However, the epistemic agent is constrained by the object in its representation. The practical agent, instead, is autonomous in a fuller sense, as she is in charge of deliberating about what to do, albeit within the constraints of rationality. But the constraints of rationality are importantly internal to and constitutive of her agency. Thus, such constraints do bear on rational deliberation, but not in a sense that undercut the agent’s autonomy. On the contrary, Kant’s claim is that the requirements of rationality cohere and deeply resonate with the constitutive features of the agents, insofar as they are rational. In the case of practical knowledge, the feature of self-conscious activity that is distinctive of the exercise of reason becomes efficacy. Possessing practical knowledge is not merely to be informed about what are the right standards of action and determine whether this particular action conforms to the standard. Rather, possessing practical knowledge is efficacious: it determines agents to act in conformity to rational norms. Ordinarily, knowing what one should do ordinarily implicates that being determined to act accordingly. Of course, things can go wrong, in which case the agent would appropriately feel the pangs of conscience. These feelings are a remainder of the authority of moral requirements, which have been violated. The efficacy of practical knowledge is explained by the autonomy of reason, in a way that distinguishes the practical from the theoretical function of reason. The difference is crisply captured by the constructivist formula that the right ends of action are the upshot of practical reasoning rather that its starting points. From the perspective of rational agency, there are no moral requirements (no moral facts) prior to and independently of the autonomous exercise of practical reasoning. This is not to say that the contents of morality are invented or that they are the result of individual, possibly self-serving, confabulation. On the

Practical Knowledge and Rational Action  71 contrary, it means that moral requirements are both objective and efficacious because and insofar as they are the result of an autonomous activity of self-legislation, in which the agent spontaneously engages. In the theoretical case, the actuality of the object of knowledge does not depend on the agent’s self-conscious exercise. Theoretical knowledge presumes the actuality of what one knows, and this also means that the objects of theoretical knowledge stand prior to and are independently of the activity of knowing them. This is why theoretical knowledge relies also on receptive faculties, e.g., the senses, and is grounded on the distinction between the knowing subject and the object of knowledge. By contrast, in the practical case, the actuality of what one knows depends on the actuality of practical knowledge. That is, practical knowledge determines its objects insofar as it implicates the agent’s knowledge of oneself as the efficacious cause of her action. To this latter extent, then, practical knowledge (of the right objects of choice) is always and at the same time knowledge of oneself as an agent, and thus qualifies as a form of self-knowledge. Most importantly, in this account of practical knowledge, the cognitive and the practical aspects of moral judgment cannot be prised apart. The efficacy of moral judgment is explained by its being a cognition of oneself under the representation of agency.23

VIII. Kant’s “Practical Proof” and the General Argument Against Heteronomous Doctrines To defend the objectivity of ethical standards we have to show that such standards can determine humans to act by guiding their mind in action. This is the “practical proof” that Kant endeavors to provide. This sort of proof amounts to a rationalization of ordinary moral cognitions. The purpose of philosophy is to transform common moral cognition into rational cognitions so as to provide them with authority.24 Reason alone can provide moral cognitions with genuine authority, even though there may be different ways to acquire their contents. For instance, one can learn from testimony that breaking a promise is wrong, trust a friend’s advice that it is better to tell the truth, or follow common wisdom that lies have a short life. But in any of such cases one must make up one’s own mind about the matter in order to be guided by testimony, advice, or rules of thumb. Ultimately, one must judge that it is wrong to break a promise and right to tell the truth. Unless one makes a practical judgment about what to do, one’s action lacks any moral value and, in extreme cases, it also lacks autonomy.25 Judgement is the fundamental element of cognition and moral cognition and it is governed by the claim that any attempt to reach beyond the boundaries of possible experience is bound to fail. Importantly, this limitation refocuses the debate on the validity of judgment and the procedures to test it (C1 Bxxi; see Allison 2004; Korsgaard 1996).26 In contrast to a tradition of practical thought that is still present

72  Carla Bagnoli in Sidgwick’s work, Kant’s view is that rationalism should reconceive of the standards of validity of judgments away from the metaphysics of self-evident and absolute truths and clear ideas. Truths cannot be grasped directly through discursive understanding but must be approached in terms of judgments susceptible to proof, that is, such that they can be grounded on reasons that everybody can share.27 Kant argues that the failure to account for the authority of moral obligations is due to a failure to fully understand the practical function of reason. This diagnosis is at the heart of Kant’s critique of dogmatic rationalism (e.g., in its Intuitionist varieties), and also of Humean skepticism. Neither of these ethical theories could vindicate the objectivity and authority of moral obligations because they both misunderstand the autonomous function of reason. This mistake is characteristic of the whole category of heteronomous doctrines, even though there are many ways to commit such a mistake. The fact that Kant focuses on the rational authority of moral obligations is revealing because they constitute a problem only for agents who are neither fully determined by reason nor fully determined by instinct. It is only for such agents that the problem of ethical objectivity arises as a practical question. The practical question is addressed by reconsidering the practical function of reason. This is the question that all ethical theories are said to elude and the positive argument for the autonomy of reason builds upon the critique of the heteronomy of previous doctrines (G 4:41–444; C2 5:35–41, 153, 157). All such doctrines fail to make sense of the subjective authority of moral obligations because they do not adequately account for the scope of practical reason. In contrast to such doctrines, Kant’s theory is designed to prove that the laws of pure practical reason can be a matter of practical knowledge and directly guide rational agents in acting (C2 5:151–153). Finite rational agents are susceptible to the moral law and thus capable of regulating their thinking and their acting in relation to it; but they come to know the moral ends by reasoning rather than by direct acquaintance (MM 6:389, 6:379). The moral law represents the idealized form of a moral community which serves as normative standard for social interactions. This normative standard is a constitutive element of the person’s self-conception. This explains how autonomy is preserved and indeed exercised in practical reasoning. On this view, then, to prove the objectivity of practical reason one needs to account for the subjective conditions of susceptibility to duty, which Kant identifies by the moral feeling of respect as reverence for the law.28 I take this much to be established by the argument of the “fact of reason.”29 While the argument establishes an internal relation between practical reason and morality, it does not rule out immoralism as logically incoherent. Rather, it points out that while conceivable, immoralism cannot be universally sustainable because it defeats human agency, given

Practical Knowledge and Rational Action  73 the peculiar conditions of human sensibility and the circumstances of finitude and inter-dependence. This feature marks the novelty of Kant’s theory of practical reason in contrast to all previous ethical views, which are all rejected on the ground that they have misunderstood the proper scope of reason (G 4:441–444; C2 5:35–41, 153, 157). Humean theory is the primary target of this objection, since Hume considered reason inert and deprived of causal influence on the will: there is no practical reason, and hence there is no rational will.

IX.  The Argument Against Dogmatic Rationalism The same charge is moved against Dogmatic Rationalism, and for reasons that differ from those invoked by Sidgwick against Dogmatic Intuitionism; Humean and Dogmatic Rationalism alike fail to adequately characterize how we apprehend moral truths because they do not recognize the practical or productive function of reason. Kant’s argument moves from a conception of the activity of reason that is quite different from Sidgwick’s and is marked by the claim that reason is autonomous in respect to its objects: the objects that practical reason knows are posterior to its activity. The differences can be articulated around three related claims, concerning the nature of reason, the structure of reflection, and the subjective experience of morality. While for Sidgwick reason is a capacity to grasp truths or to relate them together in an inference, for Kant it is a dynamic and dialogical capacity for judgment. The Kantian view is that practical reflection provides genuine knowledge because it is principled. To be guided by practical principles is a necessary condition for selfgovernance. The primary role of principles is holding agency together and making action not just possible but real, from a distinctively firstpersonal perspective. To be efficacious, practical principles must operate at the level of consciousness. They are not mere rules of thumb but represent the internal structure of rational agency and rational action. Reason is a legislative activity that organizes the self in acting and thinking. To say that the activity of reason is legislative indicates that it applies universally but also that in reasoning, one should regard oneself as the genuine author of one’s judgments and as accountable to others qua colegislators. To this extent, reasoning is marked by mutual respect and recognition (Kant G 4: §2; O’Neill 1989, ch. 2, 2015).30 Second, while for Sidgwick reflection takes the form of introspection and observation, for Kant reflection is an activity of the mind that constitutively implicates others. Reflection is generated by the characteristic reflexivity of reason. Yet this reflexivity should not be thought to be like an inner monologue. Nor is it like a movement of stepping back and forth from one’s individual stance.31 These metaphors are both inadequate to capture the dialogical significance of the lawlike activity of reason. On

74  Carla Bagnoli the contrary, Kant’s forensic metaphors suggest a polyphonic interpretation, where different voices and perspectives are audited and none is granted special authority (C1 Axi-xii).32 The reflexivity of reason makes room not only for the conceptual possibility of others but also for disagreement. For such divergence to be heard, mutual respect and recognition must be in place. This integrated element makes sure that reasoning is constitutively connected to the recognition of other differing voices.33 So, finally, while for Sidgwick moral sensibility plays virtually no role, for Kant the subjective experience of morality is the experience of autonomy and mutual recognition. Kant holds that in the case of finite rational agency the activity of reason occurs and is experienced in a distinctive way, and primarily under the guise of moral feeling. In contrast to Hume, Kant thinks that this moral feeling of reverence for the law is not pathological but originates in the contemplation of the moral law. It is through this moral feeling of respect that reason has an impact on human will. This latter claim may sound a concession to Humean psychology or to the theory of moral sense, but it is neither. The moral feeling of respect does not mediate between reason and action, as desires do; hence, it is not an incentive to conform to moral duty. On the contrary, it is moral duty itself under the guise of an incentive: that it, in its subjective aspect. This claim also separates Kantian rationalism from Dogmatic Rationalist in a way that anticipates Sidgwick’s critique of Dogmatic or Perceptual Intuitionism (G 4:441–443). Dogmatic Rationalism treats reason as inert and incapable of providing the rational agents with moral motives.34 Differently than Sidgwick, Kant holds that the problem with Dogmatism lies in the fact that it underappreciates the capacity of pure practical reason to be a practical guide for humans. Kant’s point is that such authority is not explained on rational grounds; and hence, it is arbitrary. Reasoning finds its starting points from arbitrary first-order principles whose authority is merely asserted rather than proven, so that the authority of reasoning is derivative and hence spurious. Reason’s authority is not derivative, but this is not to say that it is self-evident or evident without proof. On the contrary, the authority of reason is always subjected to rational scrutiny. Reason is self-vindicating: its authority depends on its reflexive exercise, and it is defeated or reinforced only by reasoning. The novelty of Kant’s take on rationalism is that he captures the issue of objectivity as it arises for finite rational agents, as agents that are sensitive to incentives and that need incentives in order to act. The proof of ethical objectivity resides in the autonomy of reason, and this claim must be clarified in relation to finite rational agents.

X. Kant’s Practical Proof: Practical Laws as Laws for Practical Subjects To capture Kant’s conception of moral judgments as “practical” it is useful to recall that moral judgments do not depend on the object but on the

Practical Knowledge and Rational Action  75 principle of rational volition (G 4:399–400). Relatedly, albeit at another level of abstraction, the laws of reason are also laws of subject. The principles of rationality are also principles that govern the constitution of rational selves. To say that moral obligations are rational requirements, then, does not mean only that moral duties are categorically commanded by reason but also that they matter to rational agents because they are constitutive principles of their agency. This much is clear to Sidgwick and it is precisely what he finds attractive in Kant. In the Appendix, while voicing his disappointment at Kant’s failure to reconcile rationality and morality,35 he writes: Nothing in Kant’s ethical writing is more fascinating than the idea that a man realizes the aim of his true self when he obeys the moral law, whereas, when he wrongly allows his action to be determined by empirical or sensible stimuli, he becomes subject to physical causation, to laws of a brute outer world. (ME 516) Whether and why Kant’s attempt to establish morality on a rational foundation was indeed a complete failure is a matter of scholarly disagreement.36 However, Sidgwick seems to have missed the relation that Kant establishes between practical reason and respect, which is crucial to understand his defense of the unity of practical reason. Both Sidgwick and Kant are conversant with Joseph Butler, as their remarks about the conflict between reason and self-love suggest (ME xx–xxii, 7). They agree that there is a natural tendency to self-love, which is the main hindrance to morality. But it is important to construe correctly the obstacles posited by self-love, and Kant sets the matter straight. Both morality and self-love produce incentives and grounds for action, but it is the agent that determines the ranking of such incentives. That is why, for finite rational beings, the moral law is normative, not in the sense that it exhibits regular patterns, but in the sense that it is prescriptive and regulative: It sets the normative standards that rational agents should observe as they are deliberating about what to do (G 4:397, 414). Such agents do not ask how to reason about what to do, since their rational agency is governed by the constitutive norm that requires them to act on principle. But such a norm is also a normative standard of correctness, hence such rational agents can fail to reason correctly about what to do. They have practical knowledge of the standard of reasoning, and they also have theoretical knowledge, that is, they know how to reason, and they know that they can reason. Moreover, they have practical knowledge of their action, that is, they know that they are the causes of their actions. Their awareness plays a key constitutive role, which eludes Sidgwick’s theory.

76  Carla Bagnoli Kant’s critique of the Moral Sense Theory anticipates Sidgwick’s critique of Perceptional Intuitionism.37 Moral sense theorists assume a faculty akin to perception and thus conceive of practical knowledge as akin to perceptual knowledge. Kant does think that moral sensibility is involved in accounting for the objectivity of moral obligations, but not in the sense that it is the source or the ground of moral obligations. Differently than perceptual knowledge of empirical objects, practical knowledge does not require the cooperation of intellect and sensibility in the perception of practical objects (i.e., the ends of rational action). The role of moral sensibility in the guise of reverence of the law is to account for the subjective authority of moral obligations. In defending this view, Kant frames his theory of practical reason in contrast to Moral sense theory, sentimentalism, and Intuitionism. In contrast to the former, he denies that moral sensibility is the ultimate basis of justification of moral obligations. In contrast to the latter, he affirms that moral sensibility plays a crucial role in explaining how it is that finite rational agents, sensitive to natural incentives and inclination, become interested in acting for the sake of duty. His account of reverence as the sole moral incentive shows that the role of moral sensibility is not ancillary and posterior to the acquisition of some sort of propositional knowledge. Knowing practically the rational ends of action is, indeed, knowing how to act rationally by being driven by a specific sort of motivation, i.e., by the very thought that something is right. This thought is normative for finite rational agents because they face the deliberative question about what to do from a distinctive practical starting point, that is, by representing themselves as free and equal co-legislators in a cooperative system of peers. I take this to be a fair rendering of the third formula of the categorical imperative, which is meant to be a compendium of the law formula and the formula of humanity. In a way, to take up the practical standpoint is not open for refusal, since this is the very stance of rational agency. But undertaking this stance requires that the agent claims full responsibility for what she chooses to do. Respecting oneself as an agent requires treating oneself as capable of undertaking action responsibly, so that one can be imputed and answer for what one has done. As a form of practical knowledge, the categorical imperative imposes some features to reasoning from the practical point of view. But such features do not necessitate finite agents to act. Rather, they impose on them the demand that they act in conformity or for the sake of duty. This remark settles some of the issues raised by Sidgwick in his Appendix. The sort of rational autonomy exhibited by finite rational agents does not guarantee that they always act for the sake of duty. This is not because there is an irremediable divide between prudential and moral reasons, as Sidgwick holds. Rather, this is because finite rational agents are not

Practical Knowledge and Rational Action  77 necessitated either by instinct or by reason, hence they act on the basis of specific practical principles, which they build by universalizing their subjective maxims. Understood as the form of practical knowledge, the categorical imperative does not track down an independent realm of moral objects. If this were so, the question of moral motivation would be mysterious. The mystery is dispelled by noticing that the categorical imperative shapes and sustains the self-representation of finite agents as free and equal. To fulfill their cognitive and normative purposes, reasons for action (i.e., the specific practical principles that the agent constructs in specific deliberative scenarios) must fit the form of the law. The formal requirement is that the agent checks whether the proposal of action can be willed as a law. However, the test is not formal in a sense opposite to sensible. In fact, in order for the formal requirement to be guiding, some subjective conditions of moral sensibility must be in place.38 The form of the categorical imperative produces practical knowledge and provides moral constraints insofar as we are endowed with a specific sort of moral sensibility. Thus, to show that practical reason achieves practical knowledge one has to establish not only that the form constraints the content of the maxims but also that a specific sort of moral sensibility is required. To this extent, moral sensibility marked by respect and recognition is not only subjectively necessary for explaining the subjective authority of reason but also objectively necessary for the production of objective practical knowledge.

Conclusion This comparison allows us to set the record straight on a more general issue, concerning the prospects for ethical theory of defending objective standards for action. Bernard Williams criticizes objectivist views for discounting the perspective of agency and holds that this is the unwelcome result of casting objectivity in terms of the viewpoint of the universe. He claims that any ethical theory is bound to enjoin this cosmic viewpoint, and as a consequence, incoherent because it introduces an unbridgeable gap between the objective requirements that are mandatory from the cosmic perspective of the theory and the personal commitments of any one individual agent. Because of its cosmic perspective, ethical theory is necessarily and absolutely alienated from the agent’s own life. It is hard to see how its objective requirements can be authoritative since they cannot have a foothold on the agent’s character. If my comparative arguments are correct, Kant eschews this objection because it defends objectivity as a practical requirement. Objectivity is a mark of practical knowledge as it pertains to agents insofar as they are finite practical subjects, fragile, mutually dependent, and cognitively bounded but also endowed with a moral sensibility which allows them to

78  Carla Bagnoli be effectively guided by the objective moral standards that are designed for them.39

Notes 1. This amounts to “conceptual rationalism,” that is, the view that moral judgments are judgements about what is rational or reasonable to do. On the distinction between “conceptual” and “substantive” rationalism, see Smith (1994). In accordance with Sidgwick’s usage, I do not distinguish between “rational” and “reasonable.” Strictly speaking, the term “irrational” indicates that one is inconsistent whereas “unreasonable” means that one can be subjected to criticisms for one’s views. 2. Sidgwick describes a method of ethics as “any rational procedure by which we determine what individual human beings ‘ought’—or what it is ‘right’ for them—to do, or to seek to realise by voluntary action” (ME 1). There are several methods “implicit in our common moral reasoning”: egoism, utilitarianism, and intuitionism (ME 14); and the task is to recompose the “diversity of methods applied in ordinary practical thought” (ME 6). Such a reconciliation results from an investigation and definition of the mutual relations among the three methods, where they seem to conflict (ME 14). 3. “What ought to be is a possible object of knowledge” (ME 33). 4. “The moral truth apprehended is implicitly conceived to be intrinsically universal” (ME 34). 5. Sidgwick seems to use desires, impulses and motives interchangeably. We should notice that “impulse” is ambiguous when it is equated to “motive,” which can be interpreted as the mediating term, so that the cognition activates a motive which mediates between judging and acting. Brink holds that there is textual evidence supporting both the internalist and the externalist interpretations, but he favors the latter because it fits better with the general account; see Brink (1994). Deigh opts for the internalist reading, focusing on the centrality of the anti-naturalist argument. According to Hurka, it is unclear whether Sidgwick really commits to internalism because he claims that moral motivation is mediated by desires, Hurka (2014, 980). Hurka traces the claim that the precepts of reason motivate only in combination with a desire to do what is right to the Methods (5, 39, 52, 101, 204, 206, 447); but these passages are amenable to different interpretations given that the term desire is used ambiguously. Similar concerns arise about Kant; see, e.g., Korsgaard (1986). 6. Moore acknowledges that Sidgwick anticipated the anti-naturalist argument, see Moore (1903, 17). To some interpreters, this is an exaggeration, see Hurka (2014, 96ff); Rashdall (1885). All of these philosophers seem to ignore that Sidgwick casts the problem in a manner very much like Kant. 7. According to John Deigh, Sidgwick infers from the claim about indefinability that reason directly motivates, without the mediation of desire, Deigh (1992, 251–258). But failure of naturalistic analysis does not itself justify the conclusion that reason can motivate without desire. Deigh has the argument backwards: Sidgwick moves from a conception of reason being practical (productive moral cognitions that guide action) to the rejection of naturalism. 8. It is a mistake to read Sidgwick’s anti-naturalist argument as leaning toward or pressing a non-cognitivist agenda, as in Shaver (2000). 9. On the connection between the unanalyzability of normative concepts and the commitment of a special faculty, see Frankena (1992, 103); Korsgaard (1989, 331); Gibbard (1990, 107, 154, 317). This is Frankena’s main objection to

Practical Knowledge and Rational Action  79 non-naturalism, Frankena (1992, 105); but it is also, ultimately, Mackie’s objection, Mackie (1977, 40ff), see, also, Darwall (1997, 702). To his credit, Brandt differentiates Sidgwick’s Intuitionism from the perceptual model, see Brandt (1959, 198–201); see, also, Shaver (2000). 10. This critique is articulated in Humean terms: it is assumed that the practical significance of moral judgments amounts to their motivating power, and that reason is inert. 11. “I must somehow see that it was right for me to sacrifice my happiness for the good of the whole of which I am part” (ME xviii). On this issue, see Crisp’s chapter in this volume. 12. On inconsistency across competent evaluators, see ME 100. On the distorting influence of emotions, see ME 214. It seems that we are less vulnerable to the epistemically pernicious influence of emotion and desire when we consider abstract ethical principles rather than when relying on our immediate judgments about particular cases. I am grateful to Tyler Paytas for urging me to clarify this point. 13. Notice that (4) is more controversial than might sound. Rawls reads Kant as making room for radical disagreement not about moral obligations—on which we may find a sort of “overlapping consensus”—but about the individual conception of the good life. Such conceptions cannot be “radically different” in the sense of outstepping the bounds of reason, of course. Kant forbids this option by claiming that pure reason binds all rational beings as such. But there can be a plurality of conceptions of the good life which radically different and yet within the bounds of reason. Deep ethical disagreement about the meaning of life must still be articulable by reason: it should be reasonable in order to be really significant. The other side of this view is that also unreasonable disagreement, that is, disagreement can be dealt with by sheer force (i.e., “by coercive respect,” duties of respect are duties of right). But (2) is also problematic for Kant, unless we specify the norms constitutive of reflection as norms constitutive of reason: reflection does not mean reflective detachment plus endorsement, pace some Kantian constructivists. 14. Elsewhere Sidgwick refers to the task of providing “self-evident principles relating to ‘what ought to be” ’ (ME 1 n.1). 15. For Schneewind (1977, 296) there are only three. Skelton (2010, 519) adds a sixth, which does not appear in The Methods of Ethics: “The greater quantum of pleasure is to be preferred to the less, and that ex vi termini the larger sum made up of less intense pleasures is the greater quantum of pleasure” (Sidgwick 1902, 110). 16. Compare Mill (1879, 6–7, 16–17). The other element is that the utilitarian principle is able to systematize and explain the common-sense moral rules (ME 422, 497–497). 17. See e.g. Williams (1995), Skelton (2008, 206 n 43), Crisp (2015, 111), Donagan (1977), Hurka (2014), Phillips (2011), Shaver (2014). For defense of Sidgwick’s argument, see de Lazari-Radek and Singer (2014). 18. There is, instead, a radical opposition between Egoism and Utilitarianism, which cannot be reconciled but on theological grounds, i.e., assuming the moral government of the world (ME xxii). For more on Sidgwick’s dualism, see Paytas’s chapter in this volume. 19. I do not take Sidgwick to hold that ethical judgements about particular conduct (i.e., that you ought to donate your extra income to charity rather than buying some luxury good) are self-evident. Philosophical Intuitionism is the view that abstract and general ethical propositions provide a foundation for an ethical theory that yields reliable verdicts about what we ought to do in particular cases.

80  Carla Bagnoli 20. To clarify, on a current reading, Dogmatic Intuitionism is dogmatic in the sense that it assumes that the rules of common-sense morality are largely correct. Sidgwick’s appeal to the axioms as the foundations for a properly justified and action-guiding ethical theory is not dogmatic in this sense. However, in another sense, Sidgwick’s approach is dogmatic because it presumes that rational agents would converge on axioms. But this convergence does not rest on any proof. 21. Or, better, there is a sense in which Kant does acknowledge a function of ethical judgment that is similar to that advocated by Sidgwick: this is the sense of moral judgment properly speaking, which is judgment about past action; see Kant (MM 6:399–403). But this sense of ethical judgment is different from the paradigmatic case of ought-judgments, understood as prescriptive. 22. Compare Sidgwick’s four criterion of adequacy discussed earlier. 23. For a thorough account of Kant’s conception of practical knowledge, see Engstrom (2009). For an argument that practical cognitivism amounts to constructivism, see Bagnoli (2012, 2013), and compare Engstrom (2012, 2013). 24. This is the purpose of section II of the Groundwork. 25. My claim here is that by following somebody else’s prescriptions about what to do and acting without judging, agents produce acts lacking autonomy, which is a necessary feature of moral agency. I am only hinting at a longstanding debate about the relation between autonomy and moral agency, which is far from settled. 26. Kant’s works are abbreviated herein as follows: C1 = Critique of Pure Reason (Kant 1998); C2 = Critique of Practical Reason, G = The Groundwork of the Metaphysics of Morals, MM = The Metaphysics of Morals (Kant 1996). Page number of the Prussian Academy Edition of Kant’s collected works. 27. Crisp provides a useful discussion of this theme in his contribution to this volume. 28. Respect for the law in its subjective aspect is called moral feeling, Kant (MM 6:465). “Any consciousness of obligation depends upon moral feeling to make us aware of the constraint present in the thought of duty, there can be no duty to have moral feeling or to acquire it,” Kant (MM 6:399). On the subjective conditions of our moral susceptibility, see Kant (MM 6:399–402). 29. On the consciousness of the moral law as “the fact of reason” see Kant (C2 5:46–48, 5:30; cf. 5:155f). Contrary to the mainstream reading, I take the argument of the fact of reason to show that ethical objectivity cannot be understood in the realist fashion because its epistemic function is not evidential. For a constructivist interpretation of the argument, differing from Rawls (2000) and O’Neill (1989), see Bagnoli (2015). 30. On the relation between self-legislation, co-legislation and publicity, see Bagnoli (2017). Kant’s repeated use of metaphors of construction and collaboration in his discussions of reasoning make it natural to speak of his approach and method as constructivist, and of his aim as the construction of reason’s authority; see O’Neill (2015, 4). For a more extensive argument, see Bagnoli (2017). 31. There is a considerable amount of confusion on this point. Kant’s conception of self-consciousness is not to be understood as introspection, nor is the operation of an inner sense. It is also mistaken to treat it as a second-order attitude, achieved by “stepping back” and “reflective distance,” as often said by contemporary commentators such as Korsgaard (1996) and G. Williams (2018). The dynamics of self-consciousness take the dialogical form of colegislation, not the form of a monologue; see Bagnoli (2017). 32. Kant (C1 Axi–xii). The forensic and political metaphors show that Kant opposes the Cartesian and rationalist methods of inquiry and proposes a

Practical Knowledge and Rational Action  81 “political,” or public, conception of reason; see O’Neill (1989, 3–28) and Bagnoli (2017). 33. I developed this interpretation in Bagnoli (2017). Rawls (2000, 171, 209– 214) refers to “public recognition” in explaining the requirements of the categorical imperative in terms of the “law of nature” formulation, and then in the formulation of the kingdom of ends. 34. To take seriously Kant’s claim that moral truths are universal and categorical practical cognitions, Kant’s constructivism should be dissociated from antirealism so as to distinguish it from some contemporary forms of constructivism as well as from R. M. Hare’s universal prescriptivism, both of which claim Kant as their precursor; see Hare (1952, 20) and Bagnoli (2013). 35. According to Sidgwick, Kant’s theory of practical freedom is inadequate to support the claim that one is truly free when one obeys the moral law. If one acts freely only when one obeys the moral law, then one cannot be blamed for failing, since failure depends on lack of freedom. This objection relies on specific conditions for the appropriateness of blame, which are debatable. It is also worth remarking that Kant’s ethics is deliberative and forward-looking, not backward looking. Blame is a characteristic reactive attitude, which makes sense in a backward-looking perspective. A second objection is that if one is equally free, whether one obeys the law or transgresses it (because of transcendental idealism), then sentiment of freedom cannot ground the authority (and incentive/motive) of morally right conduct. But in a Kantian perspective, the moral feeling associated with the experience of freedom does not have any evidential role; on the contrary, it is part of practical knowledge as knowledge of oneself as a cause; see Bagnoli (2015). 36. In the Preface to the 7th edition, Sidgwick confesses to having read Kant’s ethics “somewhat unintellingently, under the influence of Mill’s view as to its ‘grotesque failure’ ”; see ME xix and cf. Mill (1879, 5–6). Sidgwick’s reservation is that one might accept Kant’s formula of the categorical imperative and yet remain an egoist. That is because he treats self-regard or self-love as undeniable as the rationality of self-sacrifice; see ME xx. 37. Sidgwick (ME 366) considers Hutcheson the “opposite pole of Intuitional Ethics,” since he grounds moral knowledge on the faculty of moral sensibility while Intuitionism grounds moral knowledge on the faculty of reason. Kant would agree but would regard both forms of cognitivism inadequate in that they ignore or dismiss the practical and productive function of reason, hence misunderstanding the most relevant meaning of practical knowledge. 38. Rawls explains the purpose of the categorical imperative as a decision procedure device, but he is clear that it demands a specific form of moral sensibility: “Kant takes for granted that the application of this procedure presupposes a certain moral sensibility and a capacity for moral judgment, MdS Introduction 6: 399–402” (Rawls 2000, 165). Cf. Bagnoli (2011, 2013, 2015). This complementary moral psychology is absent or irrelevant for any other sort of deliberative procedure, albeit universalistic, such as Utilitarianism. See Rawls’s discussion of Kantian constructivism in contrast to rational intuitionism (1993, §1, 2, 5). For a discussion of the relevance of Kant to meta-ethics and the so-called “practical turn,” see Bagnoli (2017, 356–359). 39. This chapter was delivered as a paper at the Moral Philosophy Seminar at the University of Oxford, on May 6, 2019. I would like to thank that audience and Roger Crisp, Jeremy Fix, Sabina Lovibond, Tyler Paytas, Caj Strandberg, and Marcel van Ackeren for their helpful comments on earlier drafts.

82  Carla Bagnoli

References Allison, H. 2004. Kant’s Trancendental Idealism: An Interpretation and Defense (revised edition). New Haven and London: Yale University Press. Bagnoli, C. 2011. “Emotions and the Categorical Authority of Moral Reasons.” In Morality and the Emotions, edited by C. Bagnoli, 62–81. Oxford: Oxford University Press. Bagnoli, C. 2012. “Morality as Practical Knowledge.” Analytic Philosophy, 53: 60–69. Bagnoli, C. 2013. “Constructivism About Practical Knowledge.” In Constructivism in Ethics, edited by C. Bagnoli, 153–182. Cambridge: Cambridge University Press. Bagnoli, C. 2015. “Moral Objectivity: A Kantian Illusion?” The Journal of Value Inquiry 49: 31–45. Bagnoli, C. 2017. “Kant in Meta-ethics: The Paradox of Autonomy, Solved by Publicity.” In The Palgrave Kant Handbook, edited by M. C. Altman, 355– 377. New York: Palgrave Macmillan. Brandt, R. 1959. Ethical Theory. Englewood Cliffs, NJ: Prentice-Hall. Brink, D. 1994. “Common Sense and First Principles in Sidgwick’s Methods.” Social Philosophy and Policy 11: 179–201. Crisp, R. 2015. The Cosmos of Duty: Henry Sidgwick’s Methods of Ethics. Oxford: Oxford University Press. Darwall, S. 1997. “Learning from Frankena: A Philosophical Remembrance.” Ethics 107 (4): 685–705. Deigh, J. 1992. “Sidgwick on Ethical Judgment.” In Essays on Henry Sidgwick, edited by B. Schultz, 241–258. Cambridge: Cambridge University Press. de Lazari-Radek, K., and P. Singer. 2014. The Point of View of the Universe: Sidgwick and Contemporary Ethics. Oxford: Oxford University Press. Engstrom, S. 2009. The Form of Practical Knowledge. Cambridge, MA: Harvard University Press. Engstrom, S. 2012. “Bringing Practical Knowledge into View: Response to Bagnoli, Hill, and Reath.” Analytic Philosophy 53 (1): 89–97. Engstrom, S. 2013. “Constructivism and Practical Knowledge.” In Constructivism in Ethics, edited by C. Bagnoli, 133–152. Cambridge: Cambridge University Press. Frankena, W. 1992. “Sidgwick and the History of Ethical Dualism.” In Essays on Henry Sidgwick, edited by B. Schultz, 174–198. Cambridge: Cambridge University Press. Gibbard, A. 1990. Wise Choices, Apt Feelings: A Theory of Normative Judgment. Oxford: Oxford University Press. Hare, R. M. 1952. The Language of Morals. Oxford: Clarendon Press. Hurka, T. 2014. British Ethical Theorists from Sidgwick to Ewing. Oxford: Oxford University Press. Johnson, R., and A. Cureton. 2018. “Kant’s Moral Philosophy.” In The Stanford Encyclopedia of Philosophy (Spring 2018 Edition), edited by Edward N. Zalta. https://plato.stanford.edu/archives/spr2018/entries/kant-moral/ Kant, I. 1996. Practical Philosophy. Edited and translated by M. Gregor. Cambridge: Cambridge University Press. Kant, I. 1998. Critique of Pure Reason. Edited and translated by P. Guyer and A. Wood. Cambridge: Cambridge University Press.

Practical Knowledge and Rational Action  83 Korsgaard, C. M. 1986. “Kant’s Formula of Humanity.” Kant-Studien 77: 183–202. Korsgaard, C. M. 1989. “Kant’s Analysis of Obligation: The Argument of Foundations I.” The Monist 72 (3): 311–339. Korsgaard, C. M. 1996. The Sources of Normativity. Cambridge: Cambridge University Press. Mackie, J. L. 1977. Ethics: Inventing Right and Wrong. Harmondsworth: Penguin. Mill, J. S. 1879. Utilitarianism. 7th ed. London: Longman’s, Green, and Co. Moore, G. E. 1903. Principia Ethica. Cambridge: Cambridge University Press. O’Neill, O. 1989. Constructions of Reason. Cambridge: Cambridge University Press. O’Neill, O. 2015. Constructing Authorities: Reason, Politics and Interpretation in Kant’s Philosophy. Cambridge: Cambridge University Press. Phillips, D. 2011. Sidgwickian Ethics. Oxford: Oxford University Press. Rashdall, H. 1885. “Professor Sidgwick’s Utilitarianism.” Mind 10: 200–226.

Rawls, J. 1993. Political Liberalism. New York: Columbia University Press. Rawls, J. 2000. Lectures on the History of Political Philosophy. Cambridge, MA: Harvard University Press. Schneewind, J. B. 1977. Sidgwick’s Ethics and Victorian Moral Philosophy. Oxford: Clarendon Press. Shaver, R. 2000. “Sidgwick’s Minimal Metaethics.” Utilitas 12 (3): 261–277. Shaver, R. 2014. “Sidgwick’s Axioms and Consequentialism.” Philosophical Review 123: 173–204. Sidgwick, H. 1874. The Methods of Ethics. 1st ed. London: Macmillan. Sidgwick, H. 1902. Lectures on the Ethics of T. H. Green, Mr. Herbert Spencer, and J. Martineau. Edited by E. E. Constance Jones. London: Macmillan. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Skelton, A. 2008. “Sidgwick’s Philosophical Intuitions.” Etica & Politica 10: 185–209. Skelton, A. 2010. “Henry Sidgwick’s Moral Epistemology.” Journal of the History of Philosophy 48: 491–519. Smith, M. 1994. The Moral Problem. Oxford: Blackwell. Williams, B. 1995. “The Point of View of the Universe: Sidgwick and the Ambitions of Ethics.” Reprinted in his Making Sense of Humanity and Other Philosophical Essays, 153–171. Cambridge: Cambridge University Press. Williams, G. 2018. “Kant’s Account of Reason.” In The Stanford Encyclopedia of Philosophy (Summer 2018 Edition), edited by Edward N. Zalta. https:// plato.stanford.edu/archives/sum2018/entries/kant-reason/

4 Sidgwick’s Kantian Account of Moral Motivation David Phillips

In the short intellectual autobiography E.E. Constance Jones included in the Preface to the 6th edition of the Methods, Sidgwick at one point calls Kant and Mill his two “masters” (ME xx). It is tempting to say that Sidgwick combines a Kantian metaethics with a Millian normative theory. Like any such simplification, of Sidgwick in particular, this one is of course too simple. Among the qualifications to be noticed immediately are that Sidgwick explicitly resists Kant’s “resting of Morality on Freedom” (ME xix); and that anyone who accepts, even reluctantly, a dualism of practical reason is no straightforward utilitarian. Nonetheless, the most important metaethical chapter in the Methods, Methods 1.3, does have a strikingly Kantian flavor. It begins with a contrast between an explicitly Humean and an implicitly Kantian view of moral motivation. And the later articulation of the latter view is framed in self-consciously Kantian terms: Further, when I speak of the cognition that ‘X ought to be done’ . . . as a ‘dictate’ or ‘precept’ of reason to the persons to whom it relates, I imply that in rational beings as such this cognition gives an impulse or motive to action . . . In fact, this possible conflict of motives seems to be connoted by the term ‘dictate’ or ‘imperative,’ which describes the relation of Reason to mere inclinations or nonrational impulses by comparing it to the relation between the will of a superior and the wills of his subordinates. This conflict seems also to be implied in the terms ‘ought’, ‘duty’, ‘moral obligation,’ as used in ordinary discourse: and hence these terms cannot be applied to the actions of rational beings to whom we cannot attribute impulses conflicting with reason. We may, however, say of such beings that their actions are ‘reasonable’ or (in an absolute sense) ‘right’. I am aware that some persons will be disposed to answer all the preceding argument by a simple denial that they can find in their

Sidgwick’s Kantian Account  85 consciousness any such unconditional or categorical imperative as I have been trying to exhibit. (ME 34–35) The main part of the chapter is devoted to arguing against naturalistic reductions of the fundamental notion represented by the word “ought” or “right.” In the two formulations of the key claim: The fundamental notion represented by the word “ought” or “right” [is] essentially different from all notions representing facts of physical or psychical experience. (ME 25) And The notion which these terms have in common is too elementary to admit of any formal definition. (ME 32) I have tried to argue elsewhere, against in particular some very interesting and provocative work of Robert Shaver’s (2000, 2006; see Phillips 2011, chs. 2, 3), that in making these claims Sidgwick commits himself to nonnaturalism rather than standard alternative metaethical positions; that this non-naturalism is much more to him than an optional premise in an argument for a claim about moral motivation; and in particular that it is crucial to his argument for epistemic intuitionism, a key element of his own distinctive view, which is sometimes characterized as “Utilitarianism on an Intuitional basis.”1 I don’t want to relitigate these matters here. I want instead to take up what is, for me anyway, unfinished business. My topics here are the argument of the chapter—how the irreducibility claim is connected to the view of moral motivation—and the nature and defensibility of that view of moral motivation. With respect to the first topic, I will defend what I take to be a new interpretation of the argument; with respect to the second, I will claim that the contemporary view of moral motivation Sidgwick’s view most closely resembles is T. M. Scanlon’s, as articulated particularly in Lecture 3 of Being Realistic About Reasons (2014).

The Structure of the Argument of Methods 1.3: The Initial Puzzle The first thing I want to persuade you of is that there is something puzzling or unclear about the apparent argument of 1.3 in the later editions (including the 7th) of the Methods. To try to do so I’ll begin by describing

86  David Phillips the puzzle as I encounter it in reading the chapter, then try briefly to reinforce the idea that there is a puzzle by reflecting on the comparison between Sidgwick’s and Moore’s articulations of non-naturalism. Sidgwick begins the chapter by contrasting two familiar but conflicting views of moral motivation. On one view Wrong conduct is essentially irrational, and . . . appeals to the reason are an essential part of all moral persuasion. (ME 23) On the other As Hume says, “Reason, meaning the judgment of truth and falsehood, can never of itself be any motive to the Will” . . . the motive to action is in all cases some Non-rational Desire. (ME 23) Sidgwick continues by trying to “define the issue raised as clearly as possible.” On the one hand there are experiences familiarly understood as conflicts between non-rational or irrational desires and reason. On the other, again, there is the Humean view. Carefully laid out, it allows that intellectual processes “modify” the resultant forces of non-rational desires and the volitions to which they point. But intellectual processes play only this subsidiary motivational role. The question at issue between the Humean and the anti-Humean is whether there really is such a thing as conflict between reason and desire. Let us label the claim that there is, the thesis upheld on the anti-Humean side, The Conflict Claim: There is such a thing as conflict between reason and desire. It will also be useful to introduce names for the two sides in this dispute. Call the person who accepts the conflict claim, “Sidgwick’s Kantian,” and call the person who rejects it “Sidgwick’s Humean.” Having defined the question at issue, Sidgwick makes the key transition. He rejects the Humean picture he has just finished sketching: I hold that this is not the case; that the ordinary moral or prudential judgments which, in the case of all or most minds, have some—though often an inadequate—influence on volition, cannot legitimately be interpreted as judgments respecting the present or future existence of human feelings or any facts of the sensible world; the fundamental notion represented by the word “ought” or “right,” which such judgments contain expressly or by implication, being

Sidgwick’s Kantian Account  87 essentially different from all notions representing facts of physical or psychical experience. (ME 25) And he rejects the Humean picture (apparently) by appealing to what we can call The Non-Naturalness Claim: The fundamental notion represented by the word “ought” or “right” [is] essentially different from all notions representing facts of physical or psychical experience. As a reader I feel comfortable enough with the conflict claim. And I am comfortable enough in itself with the non-naturalness claim. But I am puzzled by the transition: Why does Sidgwick turn from the conflict claim to the non-naturalness claim? Why and how is the truth of the non-naturalness claim supposed to show that the conflict claim is correct? It will not be profitable to spend much more time trying to reinforce the idea that there is a prima facie puzzle here. But one further point may be helpful. The most famous presentation of a non-naturalist position like Sidgwick’s is Moore’s; Moore (wrongly) suggests Sidgwick is the only ethical writer clearly to anticipate his own position.2 Like Sidgwick, Moore draws epistemic conclusions from his non-naturalism. But (so far as I am aware) nothing in Chapter 1 of Principia draws any connection between non-naturalism and moral motivation. So we might ask why Moore should entirely ignore a connection that is central to the way in which Sidgwick presents non-naturalism, a connection between nonnaturalism and a particular view of moral motivation.

Three Interpretations of the Argument My first concern in this chapter is with this structural problem: the problem of how the non-naturalness claim is supposed to support the conflict claim. I will reject two interpretations and argue for a third. The first interpretation is that the chapter does not contain an overall argument whereby the conflict claim is connected to the non-naturalness claim. The second interpretation is that there is a connection: Sidgwick’s Humean needs to reject the non-naturalness claim in order to explain in properly Humean terms the role of categorical ought-judgments in motivation. The third interpretation is that there is a different connection: Sidgwick’s Kantian needs to accept the non-naturalness claim in order to vindicate the distinctive role of categorical ought-judgments in motivation. I hope the distinction between the latter two alternatives will become clearer as we proceed.

88  David Phillips The Denial That There Is an Overall Argument in 1.3 The first interpretation clearly resolves the structural puzzle. If there is no overall argument in the chapter, if there is not supposed to be any connection between the conflict claim and the non-naturalness claim, then of course we don’t need to worry about how the connection goes. And this interpretation has, in effect, an extremely distinguished advocate: Jerome Schneewind. He says: Sidgwick aims to show three things: first, that the notion [expressed by “ought,” “right,” etc.] is essentially different from notions representing what is the case; second, that it is simple and indefinable; and third, that it makes propositions in which it is an essential term subject to rational considerations. (Schneewind 1977, 216) Schneewind goes on to interpret the chapter as supporting these three claims at three different points. He takes the non-naturalness claim to be established at the beginning of the chapter, before Sidgwick gives any argument for the simplicity claim. Having noted Sidgwick’s initial contrast between the conflict claim and the Humean view, Schneewind writes: It is not disputed, in other words, that we do control our actions in accordance with our ‘deliberate judgments’. But while common sense takes them to be determined by rational considerations, the contrary opinion is that they are, or only work through, desires and feelings . . . Which view is correct? With the issue posed in this manner, Sidgwick is in a position to hold that his first point is simply accepted by all parties. Deliberate judgments about what we ought to do are different from other judgments in that they provide explicit direction of action. (Schneewind 1977, 216–217) Having taken the first couple of pages to establish the non-naturalness claim, Schneewind is committed to treating the main part of the argument as designed only to support the simplicity claim, and not the nonnaturalness claim. As he puts it: Sidgwick’s mode of arguing for his second point, the simplicity of the ‘common notion’, is to try to bring the reader to recognize a set of distinctions among ideas which are easily mistaken for one another. Once the distinctions are made, he thinks, the reader will be able to pick out the idea which is involved in moral terms, and to see that it is simple. (Schneewind 1977, 217)

Sidgwick’s Kantian Account  89 Finally (219ff.), Schneewind picks out part of the argument—the part devoted to rejecting subjectivist or moral sense theories—as “settling the third point, the involvement of the notion with reason.” Schneewind’s interpretation faces a number of specific problems with a common source: That his reading conflicts with a good deal of textual evidence to the effect that the chapter is supposed to contain an overall argument. Each of his suggestions that a part of the chapter contains an independent argument for a distinct claim is problematic. Consider first his suggestion that Sidgwick takes the first couple of pages of the chapter to establish the non-naturalness claim. The problem here is that Sidgwick does not appear to take himself to have established anything at the beginning of the chapter. Rather, he clearly seems to be setting up an argument. It is very hard to read the start of 25 as suggesting that any issue is already settled. Instead, the natural reading is that the Humean denies the simplicity/non-naturalness claim, and that the (at that point forthcoming) argument of the chapter is designed to refute him. The next paragraph, moreover, reinforces this impression. It begins “In considering this question,” again strongly suggesting that nothing has yet been settled. The problems carry over to the second part of Schneewind’s reading. He takes the bulk of the argument of the chapter to be designed to establish the simplicity claim, but not the non-naturalness claim. This flies in the face of the clear evidence in the main part of the chapter that Sidgwick takes himself to be arguing both for the non-naturalness claim and for the simplicity claim. Schneewind’s reading involves treating Sidgwick as much clearer than Moore in sharply distinguishing the non-naturalness claim from the simplicity claim. Pleasing as it would be for the Sidgwick enthusiast to find such clarity, it is very difficult to see it in the text. Sidgwick appears to alternate somewhat shoddily between the non-naturalness claim (found on Methods, pp. 25 and 31–2) and the simplicity claim (most explicitly found on Methods, p. 32). At best, insofar as Sidgwick does distinguish them, on Methods, p. 32, he gives no indication that he takes the non-naturalness claim to have been already articulated and established at the beginning of the chapter, as Schneewind’s reading would require. Finally, consider Schneewind’s claim that the rejection of moral sense theories is supposed by itself to establish the involvement of the fundamental notion with reason. Again, this conflicts with clear textual evidence that Sidgwick treats subjectivist or moral sense theories as one species of the genus of interpretations which define the fundamental notion in factual terms. He appears to take the failure of interpretations of this species to be part of the evidence for the simplicity and non-naturalness claims (which, again, like Moore, he does not appear to be careful in distinguishing), rather than as addressing a separate and independent issue.

90  David Phillips Schneewind’s reading is, thus, problematic in multiple ways. It is better to read the chapter as containing a structured overall argument than to read it, as Schneewind does, as containing no such argument. The Shaver/Deigh Interpretation A second interpretation is that there is a connection between the conflict claim and the non-naturalness claim: Sidgwick’s Humean needs to reject the non-naturalness claim in order to explain in properly Humean terms the role of categorical ought-judgments in motivation. Both Robert Shaver and John Deigh give this interpretation.3 To see how it works, begin with Shaver’s helpful presentation of an example. He says: Sometimes our judgment that we ought not to act imprudently or unjustly stops us from acting imprudently or unjustly. On the face of it, this is an example of reason motivating. Sidgwick notes that there is a Humean tactic for handling this potential counterexample. The Humean understands the moral judgment as describing facts about either consequences of action or means to ends that combine with pre-existing desires. Thus (1) I desire to hit Jones (2) I hold that I ought not to hit Jones Therefore I do not hit Jones becomes (say) (1) I desire to hit Jones (2)’ I believe that not hitting Jones is a necessary means to satisfying my desire to use Jones’s swimming pool this summer. (3) I desire to use Jones’s swimming pool this summer more than I desire to hit Jones. Therefore I do not hit Jones Sidgwick rejects this account of the conflict. . . . The argument thus runs as follows: Humean analyses of (2) in the manner of (2)’ are unsatisfactory; hence the Humean cannot account for some cases in which reason seems to motivate; hence we should reject the claim that reason does not motivate; hence we can keep the claim that moral judgments are judgments about what is reasonable that sometimes motivate. (Shaver 2006, 10–11) Unlike Schneewind’s interpretation, the Shaver/Deigh interpretation gives the chapter an overall argument. It explains how and why the conflict

Sidgwick’s Kantian Account  91 claim is connected to the non-naturalness claim: Sidgwick’s Humean needs to reject the non-naturalness claim in order to give a properly Humean explanation of the role of ought-judgments in motivation, as in Shaver’s swimming pool example. Nonetheless, I think the interpretation is problematic. The first problem is that, on reflection, it does not seem necessary to reinterpret ought-judgments in order to give them a perfectly ordinary Humean role in motivation. To see why, go back to Shaver’s helpful example. How does reinterpreting “I ought not to hit Jones” as “not hitting Jones is a necessary means to satisfying my desire to use Jones’s swimming pool this summer” make it turn out that the example involves only perfectly ordinary Humean motivation? Because we assume (Shaver’s (3)) that I desire to use Jones’s swimming pool this summer. That is, giving the ought-judgment an ordinary kind of factual content explains its motivational role on the assumption that there is a desire with that same content: to use Jones’s swimming pool this summer. But now notice that we can use just the same tactic without reinterpreting the ought-judgment. Suppose I have a desire to do what I ought to do. Then the motivation involved in the example has a perfectly ordinary Humean character. It combines a belief—that I ought not to hit Jones— with a desire with the appropriately corresponding content: to do what I ought to do. No reinterpretation of the ought-judgment is needed to explain away the counterexample.4 Shaver, that is, explains the connection between the conflict claim and the non-naturalness claim by appeal to something Sidgwick’s Humean needs. He suggests that Sidgwick’s Humean needs to reject the non-­ naturalness claim in order to deal with problematic counterexamples. But this is false. Sidgwick’s Humean can deal with the counterexamples without rejecting the non-naturalness claim. It might be urged in response that this is too quick. Surely the Humean cannot allow irreducibly non-natural concepts to play a key role in motivation; absent reduction, the motivation will not have a properly Humean character. I think this tempting response is mistaken. To see why, it may help to consider a different example of a motivating thought that crucially involves an irreducibly non-natural concept.5 Suppose I am motivated to keep a promise because I believe God commanded that promises be kept. The motivation here crucially involves a concept—God—that very plausibly cannot be naturalistically reduced. But it still seems that the motivation can have a perfectly ordinary Humean character. What motivates me in this case is the combination of a desire, the desire to do what God has commanded, with a belief with appropriately corresponding content, the belief that God commanded that promises be kept. That that content includes a concept that cannot be naturalistically reduced does not make the motivation non-Humean.

92  David Phillips A second problem with the Shaver/Deigh interpretation is that it does not directly explain why Sidgwick argues for non-naturalism in general. If Sidgwick were really only concerned with the tactic Shaver nicely illustrates, he ought to focus only on reinterpretations of ought-judgments as means-ends beliefs, like (2). But he doesn’t. Sidgwick rejects all naturalistic accounts of the meaning of ought-judgments. He doesn’t restrict himself to the naturalistic accounts that are employed by Humeans who adopt the tactic Shaver illustrates in the swimming pool example. These two problems show, I think, that the Shaver/Deigh interpretation involves a subtle but important mistake about the dialectic in Methods 1.3. As the Shaver/Deigh interpretation pictures things, the move from consideration of the conflict claim to consideration of the non-naturalness claim is the product of a need of Sidgwick’s Humean. Sidgwick’s Humean requires a particular kind of naturalistic analysis of ought-judgments to be true. By contrast, I shall argue, the move from consideration of the conflict claim to consideration of the non-naturalness claim is the product of a need of Sidgwick’s Kantian. Sidgwick’s Kantian requires all naturalistic analyses of ought-judgments to be false. A Third Interpretation To see the argument of 1.3 aright, I claim, we need to see that it is Sidgwick’s Kantian rather than Sidgwick’s Humean who has to take sides on the non-naturalness claim. The point is that Sidgwick’s Kantian needs to affirm the non-naturalness claim rather than that Sidgwick’s Humean needs to deny it.6 The best textual evidence I know of in support of the interpretation comes from the 2nd edition version of 1.3. You might think this is suspicious; that surely the 7th edition is always the authoritative one. I tend to reject this view both in general and in particular with respect to Sidgwick’s metaethics. I tend to reject it in general because all the editions are carefully written and illuminating, and Sidgwick tells us more than once that he removed passages not because he thought they were false but because they could be misleading or cause unnecessary controversy. And I reject the prejudice in favor of the 7th edition with respect to the metaethics and 1.3 in particular because I think Sidgwick had a fairly stable view on those matters but always struggled to find the best way to present it. And, indeed, it seems to me there is some good reason to find the 2nd edition version of 1.3 (as I do) especially illuminating. That’s because, as Sidgwick tells us in the Preface to the 2nd Edition, he rewrote the chapter in part in response to Leslie Stephen’s review of the 1st edition. Even before the appearance of Mr. Leslie Stephen’s review in Fraser (March 1875), I had seen the desirability of explaining further my

Sidgwick’s Kantian Account  93 general view of the ‘Practical Reason,’ and of the fundamental notion signified by the terms ‘right,’ ‘ought,’ etc. With this object I have entirely rewritten chap. iii of Book I. (ME x) And if we look at Stephen’s review, we find him sharply articulating the kind of Humean challenge to which Sidgwick is, in 1.3, partly responding. Stephen says, inter alia [Mr. Sidgwick] tells us that he assumes that there is something, which, under any circumstances, it is ‘right’ or ‘reasonable’ to do; an assumption that he finds in all ethical treatises. I confess, however, that I should like to see a judicial investigation of several preliminary questions. . . . I have a suspicion that . . . in the difference of the possible answers lies the explanation of some differences between Mr. Sidgwick and myself. I suspect that the popular classification assumed in the word moral is often incoherent and inconsistent; and that a scientific morality would therefore require to be based upon psychological and social data. (Stephen 1875, 307) and Reason, in my view of the case, is not, properly speaking, a faculty which can directly prompt to action. It is the faculty by which we recognise truth. It tells us what are the consequences of our actions, and the conditions by which we are bound. It lays down a map of the country, but does not induce us to follow one route rather than another. A full intelligence without emotion would be absolutely quiescent in an eternal Nirvana. We act simply because we feel. (Stephen 1875, 324) The revision of 1.3 in the 2nd edition is thus (in part) directly prompted by the forceful articulation of a Humean challenge. It would not then be surprising, to me anyway, if that were the version of 1.3 where the character of Sidgwick’s response to that Humean challenge came out most clearly. In any case, in the 2nd edition Sidgwick begins 1.3 by distinguishing two propositions By saying, then, that ‘Reason prescribes or dictates an end’ I mean to imply two things; first that in judging that ‘this action ought to be done’ . . . I am exercising what Hume calls the “judgement of truth or falsehood”; and secondly that this intellectual process is, or is inseparably connected with, a motive to action. It will be convenient to

94  David Phillips begin with the first proposition. This is hardly likely to be met with a direct counter-statement. No one expressly denies that this proposition ‘I (or you) ought to do A’ is in form legitimate, and that some propositions of this form are true and others false. But the common meaning of such propositions is by some writers implicitly rejected; and that chiefly in one or other of two ways. (ME2 25) What follows is a rejection of moral sentiment and penal interpretations of “ought” very like the arguments in favor of the non-naturalness claim in sections 1 and 2 of the 7th edition version of 1.3. At the end of this discussion, intended to support the first proposition, Sidgwick returns explicitly to the connection between it and the second proposition At the outset of this chapter I noticed that it was a disputed question ‘whether Reason acts as a motive’ at all: I do not, however, think that any one who accepts the view of ethical judgments maintained in the preceding section is likely to deny that such a judgment, in rational voluntary agents, is at least inseparably combined with an impulse to action of a peculiar kind, which it is necessary to distinguish from nonrational desires or inclinations. If this be granted, the question whether a bare cognition (or any purely intellectual operation) can or can not influence volition, does not seem to me one of practical importance, however interesting it may be from a psychological point of view. (ME2 31) As I read these passages: ought-judgments seem to play a unique and distinctive role in motivation (they supply “an impulse to action of a peculiar kind”). This appearance would be undermined if we could reject “the common meaning” of ought-judgments and reinterpret them as merely ordinary empirical judgments. But such reinterpretations fail. Hence, the original appearance is vindicated: ought-judgments do play a unique and distinctive role in motivation; they do supply “an impulse to action of a peculiar kind.” There is then (genuine) conflict between reason and desire. The question then is how further to understand and whether to accept these claims.

Sidgwick on Moral Motivation: The Positive Account A standing temptation in doing the history of philosophy is to interpret historical figures in the light of the latest philosophical fashions. No doubt one should be wary when assailed by this temptation. I shall nonetheless succumb to it and argue that Sidgwick’s account of moral

Sidgwick’s Kantian Account  95 motivation closely resembles Scanlon’s account of moral motivation, as articulated in Being Realistic About Reasons (2014). Scanlon distinguishes in Lecture 1 different questions about motivation, and there indicates that the major question for Lecture 3 is “the internal question of motivation”: the “problem coming down to us from Hume” of “how a person can be motivated by the thought that some fact is a reason for action if this thought is a mere belief that something is the case” (Scanlon 2014, 9). In Lecture 3, Scanlon both develops his own positive view and discusses and rejects various expressivist alternatives. I would distinguish three main elements in Scanlon’s positive view, which I will present in a different order than he does. First, Scanlon emphasizes that his view crucially depends on the distinctive content of beliefs about reasons: My account differs. . . [from expressivist accounts] in the nature of the reflective state with which it begins. . . . On my view . . . the reflective state is one of judging something to be true, of having a certain belief. The practical import of this state lies in its distinctive content as a judgment or belief about reasons. (Scanlon 2014, 57) Second, beliefs with this distinctive content play the motivational role they do in rational agents. Scanlon says, Even if the belief that one has a reason to act in a certain way can rationalize that action—that is to say, make sense of it—more needs to be said about how, and in what sense, such a belief can explain action. On my view, this explanation relies on the idea of a rational agent. A rational agent is, first, one that is capable of thinking about the reasons for certain actions and attitudes. . . . Second, a being is a rational agent only if the judgments that it makes about reasons make a difference to the actions and attitudes that it proceeds to have. (Scanlon 2014, 54) Third, the conception of rationality in play is both commonsensical and philosophically controversial. (Scanlon 2014, 56) In all three ways it seems to me Scanlon and Sidgwick are strikingly similar. To be explicit about the parallels: First, as I emphasized earlier, on my interpretation anyway, it is crucial to Sidgwick’s view about moral motivation that the fundamental notion expressed by “ought,” “right,” etc., has the distinctive content it does. Absent this distinctive content, the kind of motivation in question is not available.

96  David Phillips Second, the idea of a rational agent is just as central to Sidgwick’s motivational account as it is to Scanlon’s. Remember again the passage from p. 34 of the 7thth edition When I speak of the cognition or judgment that ‘X ought to be done’ . . . as a ‘dictate’ or ‘precept’ of reason to the persons to whom it relates, I imply that in rational beings as such this cognition gives an impulse or motive to action. (ME 34; italics added) Third, Sidgwick has his own version of Scanlon’s idea that the conception of rational motivation is both commonsensical and controversial: As we saw, when he sets up the motivational issue at the start of the 1.3, he remarks, In the first chapter I spoke of actions that we judge to be right . . . as being “reasonable” or “rational” . . . and I contrasted the motive to action supplied by the recognition of such reasonableness with “non-rational” desires and inclinations. This manner of speaking is employed by writers of different schools and seems in accordance with the common view and language on the subject . . . On the other hand it is widely maintained that, as Hume says, “Reason . . . can never of itself be a motive to the Will.” (ME 22) I should be clear about what I am and am not claiming here. I am claiming that Sidgwick’s view of moral motivation closely resembles Scanlon’s view of moral motivation. I am not claiming that the arguments of Methods 1.3 and Scanlon’s Lecture 3 are the same. They are not. I have presented my line on how the argument of Sidgwick’s chapter works: It crucially involves vindicating the irreducible non-natural concept that plays an essential role in the story about rational motivation, and that nineteenth-century Humeans like Leslie Stephen are disposed to reject. Scanlon, of course, is arguing in a very different dialectical context against much more sophisticated and nuanced Humean alternatives. So one would not expect his arguments to be the same as Sidgwick’s; and they aren’t. It’s the view that’s shared, not the argument strategy. I have claimed that Sidgwick’s view of moral motivation is very similar to Scanlon’s view of moral motivation. In this way I disagree (again) with some of Shaver’s very interesting work. In a couple of papers, Shaver (2000, 2006) draws on Michael Smith to illuminate the argument and view of Methods 1.3. I think Sidgwick has less in common on this issue with Smith than he has with Scanlon. Smith’s treatment of what he calls “the moral problem” in the book of the same title involves a number of distinctive philosophical commitments

Sidgwick’s Kantian Account  97 (Smith 1994). One such commitment is to the truth of the Humean theory of motivation: the idea that every action must be motivated by a combination of beliefs and desires (with appropriate content). Combine this view with the idea that moral judgments appear to be beliefs, that they appear to be intrinsically motivating, and that only desires are intrinsically motivating, and we get the problem. Smith’s solution crucially involves a kind of analysis of beliefs about reasons as beliefs about the motivation of rational agents; roughly, What we have normative reason to do is what we would desire that we do if we were fully rational. (Smith 1994, 150) This “platitude” developed into an analysis explains the distinctive content of moral beliefs and allows them to play a distinctive motivational role without requiring that we reject the Humean theory of motivation. Smith’s view certainly has some resemblance to Sidgwick’s. But in two important ways, I think Scanlon is closer to Sidgwick than Smith is. First and most crucially, Smith proposes to analyze the concept of a normative reason in terms of the concept of a rational agent. I take Sidgwick, like Scanlon, to be committed to rejecting any such analysis. Scanlon (and, on my view, Sidgwick) agree that if I have decisive reason to do X and I am fully informed and rational, then I will be motivated to do X. But that is not because the concept of a rational agent is fundamental and the concept of a reason analyzable in terms of it. Rather, it is the concept of a reason that is fundamental and unanalyzable; and it is in virtue of what a reason is that a rational agent will be moved by the thought that she has a reason. Second, Sidgwick’s attitude to the Humean theory of motivation is more like Scanlon’s than it is like Smith’s. Smith is committed to the Humean theory of motivation as a deep truth that any adequate solution to the moral problem has to respect. Sidgwick has no such commitment. As we saw, at the start of the 2nd Edition version of 1.3, Sidgwick characterizes his view thus: By saying, then, that ‘Reason prescribes or dictates an end,’ I mean to imply . . . that this intellectual process is, or is inseparably combined with, a motive to action. (ME2 25; italics added) Again on p. 31, when the argument has largely been completed, he says: I do not, however, think that anyone who accepts the view of ethical judgments maintained in the preceding section is likely to deny that such a judgment, in rational voluntary agents, is at least inseparably

98  David Phillips combined with an impulse to action of a peculiar kind, which it is necessary to distinguish from non-rational desires or inclinations. (ME2 31) Sidgwick, that is, has no commitment either way on the Humean theory of motivation. If the judgment is an impulse to action, the Humean theory of motivation is false; if the judgment is inseparably combined with an impulse to action, the Humean theory of motivation is true. Sidgwick regards the choice between these alternatives as Not . . . of practical importance, however interesting it may be from a psychological point of view. (ME2 31) One thing these reflections reveal is the importance of distinguishing The View of Sidgwick’s Humean from The Humean Theory of Motivation. Sidgwick’s Humean denies that there is any distinctive kind of rational motivation. Sidgwick’s Humean is committed, inter alia, to the Humean theory of motivation. But there is no converse commitment. Sidgwick’s Humean could be wrong even if the Humean theory of motivation were true. It could be that, as Nagel argues in The Possibility of Altruism (1970, ch. 5), all motivation involves combinations of desires and beliefs, but some motivation is distinctively rational because it involves motivated desires: desires that are themselves generated by normative beliefs. As I see it, Sidgwick is neutral with regard to the Humean theory of motivation. He is neutral, that is, between a view like Nagel’s and a view like Scanlon’s. Either way there is a distinctive kind of rational motivation, and the conflict claim is true. It could be that this distinctive kind of rational motivation is compatible with the Humean theory of motivation, as Nagel suggests; or it could be that it is incompatible with the Humean theory of motivation and involves beliefs which directly motivate, as Scanlon suggests. Sidgwick says that the choice between these two alternatives is psychologically interesting but not of practical importance. If I am right about this, Sidgwick’s attitude to the Humean theory of motivation is very different from Smith’s. But it is also importantly different from Scanlon’s. Scanlon is committed to rejecting the Humean theory of motivation, as is explicit in Being Realistic About Reasons.7 And in a way that is less explicit in Being Realistic About Reasons but very evident in What We Owe to Each Other (1998), that is because Scanlon rejects the concept of desire in terms of which the Humean theory is framed. By contrast, Sidgwick has no problem with the standard Humean concept of desire. Sidgwick’s motivational picture of human beings is a hybrid picture. He thinks of them as capable of being motivated in two distinct ways: in a non-reflective way, by desires and beliefs; and in a distinctively

Sidgwick’s Kantian Account  99 rational way by thoughts about what they ought to do. He thus cannot and would not reject the concept of desire; whether or not the Humean theory of motivation is true, desire plays an important role in Sidgwick’s account of human motivation. Someone might object to my claim that Sidgwick is neutral with regard to the Humean theory of motivation by arguing that an important passage from the 7th edition indicates that, at least by the time he wrote the 6th edition, Sidgwick had changed his mind and was committed to the truth of the Humean theory of motivation.8 Here is the passage: In the preceding chapter I have left undetermined the emotional characteristics of the impulse that prompts us to obey the dictates of reason. I have done so because these seem to be very different in different minds, and even to vary much and rapidly in the same mind, without any corresponding variation in the volitional direction of the impulse. For instance, in the mind of a rational Egoist the ruling impulse is generally what Butler and Hutcheson call a “calm” or “cool” self-love; whereas in the [utilitarian] . . . the desire to do what is reasonable as such is commonly blended in varying degrees with sympathy and philanthropic enthusiasm. Again, if one conceives the dictating Reason . . . as external to oneself, the cognition of rightness is accompanied by a sentiment of reverence for authority . . . at other times, however, the identity of Reason and Self presents itself as an immediate conviction, and then Reverence for Authority passes over into Self-respect. . . . Other phases of emotion might be mentioned, all having with these the common characteristic that they are inseparable from an apparent cognition—implicit or explicit, direct or indirect—of rightness in the conduct to which they prompt. There are, no doubt, important differences in the moral value and efficacy of these different emotions . . . but their primary practical effect does not appear to vary so long as the cognition of rightness remains unchanged. It is then with these cognitions that Ethics, in my view, is primarily concerned. (ME 39–40) This passage, it might be said, suggests that the cognition that one ought to do something can only have a motivational effect in combination with what Sidgwick here labels emotions or sentiments; it is then the emotions or sentiments that really have the motivational effect, and the emotions or sentiments are desires. Thus Sidgwick here endorses the Humean theory of motivation. This objection cannot be right, at least in the suggestion that the passage reflects Sidgwick changing his mind between the 2nd edition (where, as we just saw, he is explicitly agnostic about the Humean theory of motivation) and the 6th edition. For essentially the same passage

100  David Phillips is to be found at the start of the 2nd edition version of I.IV (ME2 34–35). But if Sidgwick is not here endorsing the Humean theory, what is going on? First and most importantly, the emotional elements Sidgwick here refers to are not desires; in discussing the utilitarian, Sidgwick explicitly distinguishes “the desire to do what is reasonable as such” from “sympathy and philanthropic enthusiasm.” The latter is an emotion or sentiment, not the former. Second, I don’t think the passage supports the idea that it’s the emotions or sentiments that provide the motivation. Rather, the passage suggests, the motivation is liable to be steady while the emotions are liable to be varying; so I think the motivational force is not attributable to the emotions or sentiments. So where and how does the emotional fit with the motivational in the picture articulated in this passage? I suggest the following: the “impulse” to do what we ought is or is inseparably connected with the cognition that we ought to do it. As we saw, Sidgwick is agnostic between these alternatives. On the first alternative, the Humean theory of motivation is false; on the second alternative, it is true. Whichever alternative we pick, the emotional is something further and different: whether the impulse is the cognition, or is a distinct thing inseparably connected with the cognition, in a being that is emotional as well as rational, the impulse will tend to acquire (or perhaps will necessarily acquire) one of a number of possible emotional characteristics or flavors. But these flavors are neither the core of the impulse nor essential to its motivational force. These reflections are also germane to the question of the significance of Sidgwick’s Kantian account of moral motivation for Kantian ethicists. I incline to a mixed verdict on that question. On the one hand, Sidgwick’s account (as I understand it) should be attractive to Kantian moral theorists. Insofar as one problem for Kantian moral theorists is to provide an account of motivation that does not depend on prior desire, Sidgwick offers an attractive solution. He argues that there is a distinctive kind of motivation that involves the thought that I ought to do something. If I have the concept expressed by ‘ought” and think that I ought to do x, I will be irrational if I do not do x; and there is some motivational mechanism in human beings, about whose details we can be agnostic, that ensures that this kind of motivation is normally effective. There is, though, one crucial feature of Sidgwick’s view to note: as I read him, the concept involved in this distinctive kind of motivation is not morally loaded. It is instead what Gibbard calls the “flavorless” ought. To the extent that Kantian moral theorists are concerned to vindicate distinctively moral rather than generically rational motivation, Sidgwick’s view may be less attractive to them. Whether Sidgwick’s view will be helpful to Kant interpreters is a different question. On one standard account, the interpretive debate about

Sidgwick’s Kantian Account  101 Kant on moral motivation turns on the role of the feeling of respect for the moral law. As Richard McCarty puts it, Intellectualists hold that respect for the moral law is, or arises from, a purely intellectual recognition of the supreme authority of the moral law, and that this intellectual recognition is sufficient to generate moral action independently of any special motivating feelings or affections. [By contrast] . . . Affectivists . . . maintain that [Kantian moral motivation] also depends on a peculiar moral feeling of respect for the law. (McCarty 1993, 423) As I read him, Sidgwick does not primarily focus on the issue that divides intellectualists from affectivists. Like more contemporary moral philosophers including Nagel and Smith, his primary concern is with the motivational role of desire rather than the motivational role of feeling. The passage, discussed earlier, where Sidgwick does focus on feeling rather than desire is an afterthought from the start of 1.4 rather than a central passage from 1.3. Moreover, even if we set aside this difference of focus, Sidgwick’s attitude to the psychological details would seem hard to square with the texts Kant interpreters must parse. A Kant with Sidgwick’s attitude would be concerned to vindicate the ordinary idea that reason can motivate and be content to leave “respect for the law” as the name for the detailed psychological mechanism involved, whatever it turned out to be. He would think that the nature of respect for the law was not of practical importance, however interesting it might be from a psychological point of view. I take it that the problem for interpreters of Kant on moral motivation arises in part because Kant seems not to have this attitude; he seems to be committed to a specific account of the nature of respect for the law and seems to think that the detailed story about respect for the law is of practical rather than merely psychological importance.

The Defensibility of Sidgwick’s View Sidgwick’s view of moral motivation (as I interpret it) strikes me as very plausible. I will not have space here to offer much by way of defense. But I will try to indicate very briefly the general lines along which I would defend it. The defense would have two elements. First, I would draw on Scanlon’s defense of his view. Second, I would argue that to the extent Sidgwick’s view differs from Scanlon’s, Sidgwick’s view may have important advantages. In Lecture 3, Scanlon argues that his cognitivist account of moral motivation is superior to competing expressivist accounts. As I see it, the argument has two sides. On the one hand, Scanlon emphasizes

102  David Phillips explanatory advantages of treating normative judgments as beliefs about reasons rather than as attitudinal or planning states. Doing so gives a more straightforward unity to first-personal and third-personal normative claims and makes clearer sense of normative discussion and advice. On the other hand, he argues that the supposed advantages of expressivist accounts are largely illusory: Expressivist accounts rely just as much as does his cognitivist account on the idea of a rational agent. And insofar as motivational explanations are causal, in neither the case of motivation by attitudinal or planning states nor the case of motivation by cognitive states is there direct experience of causation. Instead, If such experiences are generally followed by action this is because of some underlying neural mechanism that is equally causal in the two cases and in neither case an object of experience. (Scanlon 2014, 54) Sidgwick’s account could be defended in just the same way. And insofar as his account differs from Scanlon’s, the difference may be to Sidgwick’s advantage. Scanlon is largely neutral on the nature of the underlying neural mechanism but, in effect, rules out the possibility that the neural mechanism involves a desire or desires. Sidgwick, as we saw, is neutral on this matter too. He thinks there is a kind of motivational mechanism that does involve desire—a more primitive motivational mechanism—and leaves it open whether the distinctive rational motivational mechanism also always involves desire.9

Conclusions The most important metaethical chapter of the Methods, 1.3, has a strikingly Kantian flavor. Sidgwick there develops an argument that is difficult to interpret for a distinctive anti-Humean view of moral motivation. The trick with the argument is to understand the connection between the conflict claim and the non-naturalness claim. I proposed a novel interpretation of the argument whereby it is Sidgwick’s Kantian who needs to accept the non-naturalness claim in order to vindicate the existence of a distinctive rational form of motivation. I then argued that the contemporary view of moral motivation closest to Sidgwick’s is Scanlon’s, suggested that Sidgwick’s view is plausible, and indicated briefly the lines along which it could be defended.10

Notes 1. For this way of speaking see, in particular, the short intellectual autobiography E. E Constance Jones included in the Preface to the 6th edition of the Methods (ME xxii), where Sidgwick says “I was then a Utilitarian again, but on an Intuitional basis,” and Constance Jones’s remark on pp. v-vi of the Preface to the Lectures on the Ethics of T.H. Green, Mr. Herbert Spencer, and

Sidgwick’s Kantian Account  103 J. Martineau (1902): “[Professor Sidgwick] most often calls [his own system] Utilitarianism, but it might be described as Utilitarianism on an intuitional basis, by which phrase he has occasionally referred to it.” 2. Moore explicitly acknowledges Sidgwick’s influence on the arguments of the first chapter of Principia Ethica (1903): ““Good,” then, is indefinable; and yet, so far as I know, there is only one ethical writer, Prof. Henry Sidgwick, who has clearly recognized and stated this fact”(Moore 1903, 14). For recent assessments of Sidgwick’s influence on Moore, and comparisons generally to Sidgwick’s advantage, see Hurka (2003), Shaver (2003), and Phillips (2011, ch. 2). 3. Shaver does so in the two papers referred to previously, “Sidgwick’s Minimal Metaethics” and “Sidgwick on Moral Motivation.” Deigh does so in “Sidgwick on Ethical Judgment” (1992). Deigh’s piece is obscure enough on this matter that Shaver wrongly takes Deigh to disagree with him and responds in note 50 on p. 275 of “Sidgwick’s Minimal Metaethics.” 4. Shaver notices this point on p. 12 of “Sidgwick on Moral Motivation.” He considers it as a Humean response to the Kantian rather than as a reason to be suspicious of his own account of the dialectic. 5. The example is a salient one because, of course, Sidgwick himself considers interpretations of “ought” in terms of divine penalties in ME 1.3 particularly on p. 31. It is instructive that Sidgwick seems to take it to be necessary to reject such interpretations. 6. The point can also be put in terms of entailment relations. The Shaver-Deigh interpretation suggests Sidgwick thinks that the falsity of the conflict claim entails the falsity of the non-naturalness claim. So Sidgwick’s Humean, who rejects the conflict claim, has to reject the non-naturalness claim. The third interpretation has Sidgwick see the entailment relations differently. According to it, Sidgwick thinks the conflict claim entails the non-naturalness claim. So Sidgwick’s Kantian, who endorses the conflict claim, has to defend the non-naturalness claim. 7. In particular on p. 65, where Scanlon responds to direction of fit arguments. 8. The edition numbering here may be confusing to the uninitiated. Though the 7th edition (1907) is now the standard edition, the 6th was the last Sidgwick had any role in revising. As E. E. Constance Jones says at the start of the Preface to the 6th edition, “The revision of The Methods of Ethics for this edition was begun by Professor Sidgwick and carried through by him up to p. 276, on which the last of his corrections on the copy were made. The latter portion of his revision was done under the pressure of severe illness, the increase of which prevented him from continuing it beyond the point mentioned; and by the calamity of his death the rest of the book remains without the final touches which it might have received from his hand” (ME xvi). The 7th edition is a reprint of the 6th with very minor changes. 9. This difference will, of course, be an advantage only if Scanlon is wrong in his criticism of the supposed motivational role of desire in Section 8 of Chapter 1 of What We Owe to Each Other. 10. Many thanks to Max Hayward, Tim Henning, Tyler Paytas, Mark van Roojen, Tim Scanlon, Rob Shaver, Mark Timmons, a virtual audience at the 2016 Kant and Sidgwick workshop in Stuttgart, and a live audience at RoME 2018 for very helpful comments on earlier versions of this chapter.

References Deigh, J. 1992. “Sidgwick on Ethical Judgment.” In Essays on Henry Sidgwick, edited by Bart Schultz, 241–258. Cambridge: Cambridge University Press. Hurka, T. 2003. “Moore in the Middle.” Ethics 113: 599–628.

104  David Phillips McCarty, R. 1993. “Kantian Moral Motivation and the Feeling of Respect.” Journal of the History of Philosophy 31: 421–435. Moore, G. E. 1903. Principia Ethica. Cambridge: Cambridge University Press. Nagel, T. 1970. The Possibility of Altruism. Oxford: Clarendon Press. Phillips, D. 2011. Sidgwickian Ethics. New York: Oxford University Press. Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scanlon, T. M. 2014. Being Realistic About Reasons. Oxford: Oxford University Press. Schneewind, J. 1977. Sidgwick’s Ethics and Victorian Moral Philosophy. Oxford: Oxford University Press. Shaver, R. 2000. “Sidgwick’s Minimal Metaethics.” Utilitas 12: 261–277. Shaver, R. 2003. “Principia Then and Now.” Utilitas 15: 261–278. Shaver, R. 2006. “Sidgwick on Moral Motivation.” Philosophers’ Imprint 6 (1): 1–14. Sidgwick, H. 1877. The Methods of Ethics. 2nd ed. London: Macmillan. Sidgwick, H. 1902. Lectures on the Ethics of T.H. Green, Mr. Herbert Spencer, and J. Martineau. London: Macmillan. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Smith, M. 1994. The Moral Problem. Oxford: Blackwell. Stephen, L. 1875. “Sidgwick’s Methods of Ethics.” Fraser’s Magazine II (63): 306–325.

Part III

Moral Epistemology

5 On Seeing What Is Right1 Sidgwick, Kant, and Philosophical Intuitionism Roger Crisp

Nearly three decades ago, Tim Williamson said to me that one good test of an argument is whether it stands up to restatement in different ways. That is in part my excuse for revisiting issues of Sidgwickian and Kantian interpretation which, directly or indirectly, have been much discussed in philosophical ethics since the 1970s—I am thinking primarily of, on the one hand, the revival of interest in ethical intuitionism and in Sidgwick’s intuitionism in particular, and, on the other, the project of developing a form of Kantian constructivism initiated by John Rawls. When it came to moral epistemology, Sidgwick tended to see Kant as an ally. In the important third section of the chapter in the Methods on philosophical intuitionism (3.13), Sidgwick states what he sees as “certain absolute practical principles, the truth of which, when they are explicitly stated, is manifest” (ME 379). These are his principles of justice (or, as we might put it, universalizability), prudence, and rational benevolence. In the following section, given what Sidgwick has said about the importance of avoiding dissensus with epistemic peers (ME 341–342), it should be no surprise to find discussion of Kant. But what is surprising—and not just because Sidgwick continues to focus on consensus, as he had in earlier editions, rather than dissensus—is that Kant is produced, alongside Clarke, as an ally of Sidgwick’s. Though there are differences between them, these are said to be “superficial,” and Kant’s agreement is important for Sidgwick since Kant counts among “those moralists who have been most in earnest in seeking among commonly received moral rules for genuine intuitions of the Practical Reason” (ME 384). He goes on to find common ground with Kant on justice and rational benevolence, failing to discuss prudence altogether. This will all seem very odd to those familiar with, for example, Kant’s discussions of heteronomy in part I of the second Critique and the second chapter of the Groundwork, and with the Kantian constructivist project. My aim in this brief chapter is to look at this apparent point of conflict between Sidgwick and Kant, and to see whether Sidgwick’s use of Kant

108  Roger Crisp as an authority for rational intuitionism, largely as Sidgwick conceived of it, is as unreasonable as it might appear.

Sidgwick’s Intuitionisms First, then, let me briefly remind you of Sidgwick’s understanding of intuitionism. Sidgwick suffered badly from the lack of a good editor, who would never have allowed him to give the name “intuitionism” to three different, albeit related, positions. The first he calls perceptional intuitionism. On this view, each person should follow their immediate judgement about the rightness or wrongness of particular actions, without reference to moral rules or principles. Sidgwick’s main objection to this position is the doubt most of us feel about such immediate intuitions because, for example, of their inconsistency with our own intuitions at other times, or with the intuitions of others. Further, such immediate judgements may often be influenced by one’s own personal interests and require impartial rational assessment in the “cool hour” of reflection, in the light of abstract principles which can distinguish the various morally relevant factors in the situation in question from one another. The next phase of intuitionism—dogmatic intuitionism—seeks to do just that in the light of moral rules, which seem immediately correct in the same sort of way as judgements of conscience on individual actions. Sidgwick’s main problem with this view is broadly the same as that he found in perceptional intuitionism: It is insufficiently reflective. Of course, it is not as unreflective as the perceptional view, since it is seeking to ground immediate judgements about actions in principles which are seen by the agent as plausible in themselves. The rules one will come up with will be those of common-sense morality (this I presume is why Sidgwick suggests that the dogmatic intuitionist sees her job as to elucidate and systematize these common-sense rules). But the dogmatic intuitionist is too ready to accept these principles as foundational, and that means she deprives herself of resources which may be available to make more precise exactly what we are required to do in particular cases. Those satisfied with dogmatic intuitionism—and Sidgwick is almost certainly thinking of William Whewell—lack a “philosophic mind.” Common sense is too messy and unsystematic for a true philosopher to be satisfied with it: “without being disposed to deny that conduct commonly judged to be right is so, we may yet require some deeper explanation why it is so”; what we want is “one or more principles more absolutely and undeniably true and evident, from which the current rules might be deduced, either just as they are commonly received or with slight modifications and rectifications” (ME 102). It has to be admitted that Sidgwick offers little in the way of argument for the claim that dogmatic intuitionism lacks supporting principles. On a charitable reading, we might take him to be pointing out that dogmatic

On Seeing What Is Right  109 intuitionists have not demonstrated that there are no such principles, which at the very least justifies one in looking further into the question. The history of philosophical ethics we should expect to be “a history of attempts to enunciate, in full breadth and clearness, those primary intuitions of Reason, by the scientific application of which the common moral thought of mankind may be at once systematised and corrected” (ME 373–374). These attempts, then, involve the final phase of intuitionism: philosophical intuitionism. The aim, or the hope, is to find normative principles that are genuinely self-evident—that is, principles captured by propositions which are true and evident independently of other propositions. Further, philosophical intuitionism, as Sidgwick understands it, includes a commitment to testing any alleged intuitive principles in the light of four conditions (ME 338–342): (1) Clarity: “The terms of the proposition must be clear and precise”; (2) Reflection: “The selfevidence of the proposition must be ascertained by careful reflection”; (3) Consistency: “The propositions accepted as self-evident must be mutually consistent”; (4) Non-dissensus. This is not stated directly, but indirectly as follows: Since it is implied in the very notion of Truth that it is essentially the same for all minds, the denial by another of a proposition that I have affirmed has a tendency to impair my confidence in its validity. . . . And it will be easily seen that the absence of such disagreement must remain an indispensable negative condition of the certainty of our beliefs. For if I find any of my judgments, intuitive or inferential, in direct conflict with a judgment of some other mind, there must be error somewhere: and if I have no more reason to suspect error in the other mind than in my own, reflective comparison between the two judgments necessarily reduces me temporarily to a state of neutrality. And though the total result in my mind is not exactly suspense of judgment, but an alternation and conflict between positive affirmation by one act of thought and the neutrality that is the result of another, it is obviously something very different from scientific certitude. (ME 341–342) I take it that Kant would have no objection to clarity, reflection, and consistency in themselves, questioning perhaps Sidgwick’s construal of them in rational intuitionist terms. The epistemology of disagreement is more tricky, since I presume that Kant would not have accepted Sidgwick’s fourth condition.2 But in fact it is anyway independent of rational intuitionism itself; indeed, most intuitionists have not explicitly accepted it.

110  Roger Crisp

Kant as Philosophical Intuitionist How plausible is Sidgwick’s claim that Kant is a philosophical intuitionist?3 Consider the Categorical Imperative, in the standard universal law formulation: CI: “Act only on that maxim by which you can at the same time will that it should become a universal law.” (G 4:421)4 According to Sidgwick, this, “duly restricted, is an immediate practical corollary” of Sidgwick’s principle of justice, the canonical form of which is: [I]t cannot be right for A to treat B in a manner in which it would be wrong for B to treat A, merely on the ground that they are two different individuals, and without there being any difference between the natures or circumstances of the two which can be stated as a reasonable ground for difference of treatment. (ME 380) The most immediate “practical corollary” of Sidgwick’s principle would be something like: “Do not treat one person differently from another without reason for doing so.” One might use this idea to elucidate CI. But that would make CI a less immediate practical corollary of Sidgwick’s principle, and also raise the question of why one should prefer using CI to using the more immediate practical corollary, which on the face of it seems more straightforward, avoiding as it does the notions of acting on a maxim and willing maxims to become universal laws. The answer of course is that Kantian ethics is entirely different from Sidgwick’s. In Sidgwick’s view, the right way to decide how to act—the right “method” of ethics—is determined by a principle that governs actions in general, including the actions of deciding on a decision-procedure and of applying that decision-procedure. In Kant’s view, this gets things the wrong way around. The correct account of decision-procedures comes first, and actions are right in so far as they constitute the application of, or issue from, that decision-procedure. This is a major difference between Sidgwick and Kant, and one Sidgwick’s focus on agreement rather than disagreement in the final section of 3.13 leads him to disregard. Sidgwick says that CI must be “duly restricted” if it is to be a practical corollary of his principle of justice. What restrictions does he have in mind? The first arises in his discussion of veracity, in Methods 3.7.3 (see, also, pp. 486–487). There are occasions when the agent’s circumstances make the application of CI more limited than it might at first appear. Consider a case in which I am considering lying, and in which I accept

On Seeing What Is Right  111 that lying in general would be evil. It might be thought that the Kantian must forbid me to lie. But what if this case is unusual: I know that my maxim is not universally accepted, and I have a “reasoned conviction” that my lying would not tend to make it so to any great extent? These unusual circumstances should in effect be included within my maxim, so that what I must be able to will as a universal law is lying—for some good reason—when the agent is sincerely (and, perhaps we should add, reasonably, though Sidgwick does not mention this) convinced that her action will not encourage lying by others to any great extent. As Sidgwick himself later says: “It should . . . be observed that the admission of an exception on general grounds is merely the establishment of a more complex and delicate rule, instead of one that is broader and simpler” (ME 485). It is unclear, then, why Sidgwick should describe this as a restriction of CI, especially when we remember that the role of the CI is not to produce moral rules directly, but to test subjective maxims. Kant would have agreed that the agent should be clear about which maxim she is testing with CI. He may not even have disagreed with the claim that the agent’s beliefs about the likelihood of her lying influencing those of others are relevant, given that the agent’s maxim itself can include reference to the consequences.5 As I have mentioned, Sidgwick also sees Kant as a supporter of his view of rational benevolence: “He regards it is evident a priori that each man as a rational agent is bound to aim at the happiness of other men” (ME 386).6 Now this statement to my ears strongly implies the claim that Kant saw this principle as self-evident, and hence non-­inferential. Indeed, were that not what Sidgwick intended, one would have to wonder why he includes it at this point in his discussion. But he then goes on to suggest that, though he is in broad agreement with Kant’s conclusion, he cannot accept his arguments. He elaborates in a note at the end of 7.13. This note suggests that Kant provided two forms of argument for his principle. The first form involves its deduction from CI. Sidgwick has some good objections to any such deduction, but the very fact that it is a deduction removes it, as far as this form of argument is concerned, from consideration as self-evident. The second form of argument runs as follows. All action of rational beings is done for some end, so there must be some absolute end corresponding to the absolute rule captured in CI. This end can only be reason itself, or the universe of rational beings—that is, “humanity exists as an end in itself.”7 Sidgwick is again unimpressed. It is hard to see how humanity itself is something to be realized; and even if man as a rational being is an absolute end to other rational beings, it is not obvious why man’s nonrational desires should be fulfilled, as required by the principle of beneficence. Again, whatever one might think of Sidgwick’s arguments here, it seems clear that he is not reading Kant as presenting the principle of benevolence as self-evident.

112  Roger Crisp Kant also has a particular argument against rational benevolence as an ultimate moral principle, stated at CpV 5:36: The principle of happiness can indeed furnish maxims, but never such as would be fit for laws of the will, even if universal happiness were made the object. For because cognition of this rests on sheer data of experience, each judgement about it depending very much upon the opinion of each, which is itself very changeable, it can indeed give general rules but never universal rules, that is, it can give rules that on the average are most often correct by not rules that must hold always and necessarily; hence no practical laws can be based on it. This does seem a little unfair to the principle of benevolence. It may be true that that principle cannot support anything other than general rules, and it may be true also—though of course many have doubted it—that CI can be adduced in support of universal and absolute rules, such as a rule prohibiting lying. But it is not obvious why this in itself should count against rational benevolence. What is important is that there be a sound, single, ultimate, absolute principle in one’s ethical theory; whether that principle can support universal or only general rules should be determined at a later point. But whatever one thinks of Kant’s argument, he does appear to be ruling out rational benevolence as a foundational principle. The upshot of all this is that Sidgwick’s case for Kant as rational intuitionist must rest on CI alone. And at this point we may wish to consider some difficulties raised by Terry Irwin (2009, 62, 148) for any attempt to construe Kant as a rational intuitionist. An intuitionist like Clarke claims that moral principles capture certain “fitnesses” of actions, and these principles are intrinsically necessary, in the way that mathematical principles are necessary. Kant does not make these claims. An intuitionist interpreter of Kant, however, may insist that these aspects of Clarke’s thought are not essential to intuitionism in itself. Clarke’s intuitionist position could be stated without reference to the idea of fitness, and the mathematical analogy is entirely avoidable. What matters is the grasp by intuition of self-evident truths. But, Irwin suggests, Kant is in general hostile to the idea of self-evidence, objecting in particular that mere appeal to self-evidence is a poor way to advance a position in as contested an area as morality. Rather, Irwin suggests, “In contrast to intuitionist appeals to self-evidence, Kant argues that categorical imperatives are true and necessary because of their relation to rational agency, not because of some intrinsic necessity” (Irwin 2009, 68). But, as Robert Audi perhaps more than anyone else has demonstrated in recent decades, the claim that some principle is self-evident is equivalent neither to the claim that it is obvious, nor to the claim that it cannot

On Seeing What Is Right  113 be supported by argument (including some argument making reference to rational agency). What is essential to self-evidence, in the Sidgwickian sense, is that the claim in question is non-inferential and can be justified through a grasp of that claim in itself. Sidgwick himself prefers his candidates for self-evident claims to be statable fairly succinctly. But he approaches and buttresses them with a good deal of supporting argument and elucidation, attempting to put his reader in a position in which, given the understanding she now has of the relevant background, she can understand the claims in question and grasp their truth (and it is conceivable that Sidgwick saw Kant as working in a similar way as regards the principle of benevolence). In principle, pace Irwin, there is no reason why a principle fuller than CI, which incorporates the relation between CI and rational agency, should not be self-evident. (There may be more to be said here about a peculiarly Kantian conception of “practical reason,” but this is an independent matter.) In the same way Sidgwick’s statements of his own self-evident principles could be expanded. His actual statements are in fact often highly elliptical: Consider, for example, those expressions of the principle of prudence which refer only to temporal neutrality. Indeed, to return to Clarke, he may have intended one to read his claims about the fitnesses of certain actions as claims about the fitnesses of certain actions that make them peculiarly appropriate for rational beings, thus occupying in effect the same ground as Kant. Now, as Irwin goes on to point out, if Kant is grounding his account of moral principles on an account of rational agency more generally, he will be at most an intuitionist not about moral principles at all but about principles of rational agency. Again, however, that is unlikely to be a problem for Sidgwick, who understands “ethics” to cover the whole domain of practical reason (which of course allows him to count egoism as a method of ethics).

Kant on Heteronomy Toward the end of the second chapter of the Groundwork (4:440–444), Kant argues that all moral theories prior to his own were problematically heteronomous (see, also, Critique of Practical Reason 5:33–41). Since rational intuitionism was around well before Kant, and since he appears to include rational intuitionism within the scope of his arguments, it may seem that Sidgwick was just mistaken to see Kant as any kind of intuitionist, rational, philosophical, or otherwise. The autonomy of the will consists in its legislating to itself, independently of any external object of volition. If we analyze the moral concepts, we will see that the principle of morality must be a categorical imperative, and that this imperative itself commands the autonomy of the will. If one tries to base ethics on inclination (like Hutcheson and other sentimentalists) or on rational ideas (like Wolff and other rationalists),

114  Roger Crisp one will be able to justify only hypothetical imperatives: “You ought to ϕ because you want something else.” And this of course will not justify a genuine moral principle, which will be categorical: “You ought to ϕ, regardless of what you want.” Consider for example an Aristotelian view, according to which a person’s perfect happiness consists in virtue. What is most objectionable about such a view is its “basing morality on sensuous motives which undermine it and totally destroy its sublimity” (G 4:442). When it comes to rationalism, Kant divides such views into two: non-theological versions, such as that of Wolff, and theological versions, which base morality on divine command. The Wolffian version rests on the notion of perfection. As Wolff himself put it: “Do what makes you and your condition more perfect and omit what makes you and your condition less perfect” (Allison 2011, 42). Kant prefers this version of rationalism to the theological variety, objecting to it nevertheless that “it is of course empty, indefinite, and consequently useless . . . it inevitably tends to move in a circle and cannot avoid tacitly presupposing the morality it is meant to explain” (G 4:443). But the main problem with it, of course, is its heteronomy. This may seem unfair to Wolff. He may claim that his perfectionist principle applies to all agents, regardless of what they want. But the problem is the goodness of perfection which lies behind the principle. Even on non-theological versions of rationalism like Wolff’s, “the will . . . never determines itself directly by the thought of an action, but only by the motivation which the anticipated effect of the action exercises on the will” (G 4:444). And the source of that motivation is contingent, in that it depends on the agent’s natural attraction toward the idea of perfection. Some have understood Kant’s argument in these pages as dependent on a narrow understanding of the alternatives to his own position as essentially variations on psychological hedonism. There is a good case to be made against that interpretation. But it does seem that Wolff can insist that at this point Kant needs to sharpen the distinction between normativity or justification on the one hand and motivation on the other. In other words, he may suggest, the perfectionist principle is indeed categorical: It applies to all agents, regardless of their inclinations. This can be the case even if in fact all human agents, given their nature, can act on the perfectionist principle only through a desire for perfection. And if Kant is going to allow us to speak of action arising independently of motivation, Wolff can avail himself of that also. In other words, not only could his theory of morality be autonomous, but so could his theory of moral motivation. But now return to Sidgwick, who is at several points quite careful to distance himself from any version of perfectionism. Kant might try to run the anti-Wolff argument against Sidgwick’s principles of prudence and rational benevolence, the source of their normativity lying in the

On Seeing What Is Right  115 happiness of the agent or others. As we have seen, a distinction between justification and motivation may be enough to avoid that argument. But it will also of course be harder for Kant to employ the argument against Sidgwick’s principle of justice, since this is, as Sidgwick sees it, practically equivalent to CI. In recent years, an alternative reading of Kant’s position on the autonomy of the will has emerged, according to which the key point he is making against rationalism is more deeply metaphysical. This reading resonates with the point made earlier, that Kantian ethics begins with an account of rational decision making—that is, rational agency. Autonomy, Kant believes, is the only way to explain the “bindingness” or authority of morality, and autonomy consists in the will’s legislating for itself. If the will is bound by something external to it, this is heteronomy—subjection to the will of something “other.” What is wrong with dogmatic rationalism or intuitionism, whether in Wolff or Sidgwick, is that on these views moral agents are inactive and merely receptive to objects external to them, as if grasping moral truths were analogous to sense-perception. Whether Kant was indeed a constructivist is a controversial matter, but we need not go into it. In so far as the constructivist view of self-legislation involves a contrast between rational activity in pursuit of some interest or desire, and autonomous rational activity, the distinction between justification and motivation can again be employed. Consider the principle that promises should be kept. According to the rational intuitionist, we have a capacity which enables us to grasp that truth, analogous perhaps in various ways to our capacity to grasp mathematical truth or even to sense-experience.8 Rational activity is indeed in this respect passive or receptive, and the content of the resulting belief is determined externally to the agent, her rationality, and her will, by the way things are, morally speaking. But if the agent then wills to act in accordance with that principle, she is at this point legislating for herself and so making a genuine and autonomous choice, and there is no reason to ascribe to her any motivation for a goal extraneous to that of acting in accordance with the principle, such as a desire to advance her interests or those of others. In other words, as we have already seen, a rational intuitionist can accept Kant’s account of moral motivation if she wishes. What she will not accept is any richer account of self-legislation, in which the agent’s will is entirely unconstrained by anything external to the very exercise of that will itself. At this point, we should note that Kant, despite his view on the ethics of promising, will not allow the principle that one should keep one’s promises to have the same status as CI. That may be in part because there may be rational beings for whom promising is for some reason just not possible, which would make the principle “unfit to be a law” (CpV 5:35). But it is also because he will not allow for a plurality of principles, since this would be to introduce a lack of rationality and system into one’s theory. This may be an issue on which Sidgwick would, in effect,

116  Roger Crisp agree, if Kant’s worry is a concern about conflict. However many principles one believes Sidgwick presents in 3.13, there are certainly more than one. But his hope seems to be that these separate principles will in fact not conflict, and so could in fact be agglomerated into a single principle, such as “Respect justice, benevolence, and prudence.” That hope of course is dashed in the Concluding Chapter of the Methods. It is tempting to think that constructivist interpreters of Kant, especially if they are themselves constructivists, cannot intend to impute to Kant an existentialist account of freedom of the kind I sketched earlier. The CI itself is clear evidence that Kant thought that the will of a rational agent is limited by constraints of practical rationality not of her own making.9 At this point, the constructivist debate may move into more straightforwardly metaphysical territory, where rational intuitionists may be accused of holding a queer, Platonic view of the ontological status of moral properties or truths. But this is really a change of subject, and again there is nothing to prevent a rational intuitionist from refusing to engage with it—as Sidgwick largely does. What matters for the intuitionist is that there are certain normative truths that we can grasp through exercise of a rational faculty. Now even as far as CI alone is concerned, Sidgwick underestimates the differences between it and his own principle of justice. And Kant will also insist that the bare statement of CI on its own is insufficient to grasp it as a statement of reasons grounded in the nature of rational beings. But it does not seem implausible to understand Kant to be claiming that an expanded version of CI is self-evident in Sidgwick’s sense—that is, graspable by reason, on the basis of understanding CI itself rather than on the basis of inference, in such a way that it is a candidate for knowledge. So, to conclude, it is not unreasonable of Sidgwick to see Kant as a rational intuitionist ally, although it remains regrettable that he published the 5th edition of the Methods—in which he moved from a test based on consensus to one based on dissensus— without addressing the serious disagreements between him and Kant that remain and their implications for his own ethical views.10

Notes 1. See ME xviii. 2. I am grateful to Carla Bagnoli for reminding me that Rawls reads Kant as making room for radical disagreement not about moral obligations (on which we may find a sort of “overlapping consensus”) but on the representation of a good life that depends on subjective preferences. 3. In Methods 3.13.4, Sidgwick is not pointing to agreement merely on his principles between himself on the one hand, and Clarke and Kant on the other. I take it that those who have been “most earnest in seeking . . . genuine intuitions” are themselves to be understood as philosophical intuitionists. Having mentioned these earnest seekers, Sidgwick immediately goes on to state that the earlier British intuitionists are more “philosophical” than those who later reacted to Hume (ME 384); and at p. 385, before stating that he and Kant

On Seeing What Is Right  117 converge on similar ethical principles, he says that “Kant is especially noted for his rigour in separating the purely rational element of the moral code.” 4. Kant’s works are abbreviated as follows: CpV = Critique of Practical Reason, G = The Groundwork of the Metaphysics of Morals. All translations are from Practical Philosophy, ed. M. Gregor (Kant 1996). 5. See, e.g., Barbara Herman (1993, ch. 6). 6. As Tyler Paytas pointed out to me, it is important to note that this is not the principle that I am bound to treat the happiness of others as of equal weight to my own. 7. Kant, or a Kantian, might well claim that, though this is an end, it is not an end to be realized, as opposed to being, e.g., respected. 8. The essence of the point being made here is that ethics and mathematics are on all fours. If ethics is to be understood in some constructivist sense, then so is mathematics and vice versa. 9. It may be suggested that, since Kant claims that morality follows analytically from freedom, he has no need for a moral epistemology (see Skorupski, ms, 3.5.3). But that analytic account is itself an epistemology, and the freedom in question is not to be construed as existential. 10. I am grateful to the editors for the opportunity to present an earlier version of this chapter at The Cosmos of Duty Above and the Moral Law Within: A Workshop on Sidgwick and Kant, held at the Institute of Philosophy, University of Stuttgart in May 2016. I wish also to thank the editors, and other participants at the workshop, for very helpful discussion and comments; I am especially indebted for their great help with later drafts to Carla Bagnoli and Tyler Paytas.

References Allison, Henry. 2011. Kant’s Groundwork for the Metaphysics of Morals: A Commentary. Oxford: Oxford University Press. Herman, B. 1993. The Practice of Moral Judgment. Cambridge, MA: Harvard University Press. Irwin, T. 2009. The Development of Ethics, Vol. 3: From Kant to Rawls. Oxford: Oxford University Press. Kant, I. 1996. Practical Philosophy. Edited and translated by M. Gregor. Cambridge: Cambridge University Press. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Skorupski, J. ms. Being and Freedom.

6 Peerless Sidgwick, Kant, and Peer Disagreement Robert Shaver

[T]he denial by another of a proposition that I have affirmed has a tendency to impair my confidence in its validity. . . . [T]he absence of . . . disagreement must remain an indispensable negative condition of the certainty of our beliefs. For if I find any of my judgments . . . in direct conflict with a judgment of some other mind, there must be error somewhere: and if I have no more reason to suspect error in the other mind than in my own, reflective comparison between the two judgments necessarily reduces me temporarily to a state of neutrality. And though the total result in my mind is not exactly suspense of judgment, but an alternation and conflict between positive affirmation by one act of thought and the neutrality that is the result of another, it is obviously something very different from scientific certitude. (ME 341–342) It is obvious that in any such conflict [between intuitions] there must be error on one side or the other, or on both. The natural man will often decide unhesitatingly that the error is on the other side. But it is manifest that a philosophic mind cannot do this, unless it can prove independently that the conflicting intuitor has an inferior faculty of envisaging truth in general or this kind of truth; one who cannot do this must reasonably submit to a loss of confidence in any intuition of his own that thus is found to conflict with another’s. (1905, 464) [I]f we find that an intuitive belief appears clear and certain to ourselves contemplating it, that it is in harmony with other beliefs relating to the same subject, and does not conflict with the beliefs of other persons competent to judge, we have reduced the risk of error with regard to it as low as it is possible to reduce it. . . . [One danger of error] we meet by endeavouring to attain Consensus of Experts. (1905, 465, 466)1

One might worry that Sidgwick’s condition leads to skepticism. Roger Crisp, for example, writes that

Peerless  119 Sidgwick did not fully confront the implications of the non-dissensus condition for his own normative ethics. . . . [H]e fails properly to reflect upon the fact that there will be reflective thinkers who disagree with each of his alleged axioms, however weakly they are interpreted. . . . It remains a mystery why Sidgwick appears not to have recognized the sceptical implications of his non-dissensus condition. (Crisp 2015, 111) One response, as Crisp notes, is to interpret Sidgwick’s axioms very weakly (e.g., Shaver 2014). Here I consider a different reply: perhaps Sidgwick’s condition, rightly understood, does not justify scepticism because it is not so common to encounter those I have no more reason to suspect of error than myself. I think this is a plausible reading of the condition. But if so, Sidgwick should drop his (apparent) claim that Kant is someone he has no more reason to suspect of error than himself. I begin with understanding the condition, then turn to Sidgwick on Kant, then to Crisp.

Peers Sidgwick writes of those I have no more reason to suspect of error than myself, those competent to judge, and experts. I assume that the categories of “those competent to judge” and “expert” are to be understood in terms of reasons to suspect error. Sidgwick says nothing to suggest any other account of competence to judge or expertise. And if, say, someone “competent to judge” is someone I have reason to suspect more likely to err than myself, it is not clear why, if I am faced with the disagreement of someone competent to judge in this sense, I should respond with a loss on confidence.2 By an “epistemic peer,” I intend one whom I have no more reason to suspect of error than myself. This may depart from the usage of those who require that peers share the same evidence and equal ability to interpret the evidence. These similarities seem important because they are grounds for thinking that someone is one I have no more reason to suspect of error than myself. (The similarities are also not necessary for there to be a skeptical threat. You and I might not share evidence, but still be equally reliable, since neither of us has better evidence than the other.) It is similar to the usage of those who require that peers are those one predicts will be equally reliable.3 Sidgwick’s account leaves it open that the number of people who disagree with me can matter. If you and I disagree and I have no evidence about what others think, you might well be someone I have no more reason to suspect of error than myself. But if we disagree and I learn that thousands (independently) agree with me and no one with you, I have more reason to suspect error in you than in me. Similarly, if I learn that thousands (independently) agree with you and no one with me, I have

120  Robert Shaver more reason to suspect error in myself rather than in you. Sidgwick writes that “if I found myself alone contra mundum, I should think it was more probable that I was wrong than that the world was” (1879, 109).4

Conciliation Sidgwick is usually read as holding the “conciliatory” position on peer disagreement. According to conciliationists, in cases of peer disagreement, the correct response is to suspend judgment, or give equal credence to the conflicting beliefs.5 But Sidgwick does not make either claim. In the Criteria passages, he says that the correct response is a “loss of confidence,” or a belief that we have not reduced the risk of error to be as low as possible. In the Methods, confidence is impaired and certainty lost. He does add that he alternates between neutrality and affirmation. But (1) this seems only meant to support the conclusion that certainty is lost, (2) is not the same as claiming that neutrality is the correct response, and (3) does not appear in the Criteria versions. More importantly, in the Methods Sidgwick has no need to claim that suspension of judgment or equal credence is the correct response. He is stating the conditions, the complete fulfilment of which would establish a significant proposition, apparently self-evident, in the highest degree of certainty attainable: and which must be approximately realised by the premises of our reasoning in any inquiry, if that reasoning is to lead us cogently to truthworthy conclusions. (ME 338) He is doing so to argue that his axioms do better than common-sense morality does on these conditions. He has no need to say any more about the correct response to disagreement than that it requires a loss of the highest certainty. Although Sidgwick does not argue for suspending judgment, one might think that he is committed to it (McGrath 2008, 91). I consider this below.

Independence Like conciliationists, Sidgwick requires that one must “prove independently that the conflicting intuitor has an inferior faculty of envisaging truth in general or this kind of truth.” Presumably, he is ruling out the following reasoning, where I believe p and you believe not-p: “the mere fact that you believe not-p is sufficient to show that you are not my epistemic peer.” Anyone who proposes a condition like Sidgwick’s should rule out this reasoning. If one does not, the disagreement condition could

Peerless  121 do little work, since most cases of disagreement would show that you are not my peer.6 Sometimes the worry motivating independence is that if I were allowed to dismiss you as a peer simply in virtue of your disagreement with me, I would beg the question (Vavova 2014, 312; Christensen 2007, 198, 2009, 758, 2011, 2, 18). Say I believe p. You object: “I am as reliable as you. I believe not-p. Therefore, you should suspend judgment about p.” I reply: “Since p is true, you are clearly not as reliable as me. Therefore, I do not need to suspend judgment.” This is question-begging because my defense of p depends on assuming p. Note what these motivations for independence do not rule out. We disagree. I see your reasoning for not-p. You see my reasoning for p. I beg the question if I count your belief that not-p as showing that you are not my peer. But I do not beg the question if I dismiss you as a peer by raising an objection to your reasoning, provided (i) I do not take your reasoning to be faulty simply because it did not result in my belief—say my c­riticism is that you make a factual or logical error in arriving at not-p, and I would find this an error whether or not I believed p (see Christensen 2007, 198); (ii) I have no evidence that you disagree with my objections (and so my objections, unlike p, need not be set aside on the ground that it would be question-begging to defend my objections by assuming that they are correct).7 One might instead object that in deciding whether you are my peer, I must set aside my assessment of whatever reasoning you offer for notp, even if that assessment did not rely on my belief that p and I have no evidence that you disagree with my objection. The motivation might be like the initial motivation given for setting aside my belief that p: if whenever I disagree with you, and see your reasoning, I find some error in your reasoning, there would again be no peer disagreement (in cases where your reasoning is public). But (1) it is false that whenever I disagree with you, and see your reasoning, I find some error in your reasoning. The literature is rife with cases in which I disagree with you but can find no error in your reasoning (e.g., McGrath 2008, 103; Elgin 2010, 58–59; van Inwagen 2010, 24). The disagreement might stem from different non-inferentially justified premises. (2) It is not so costly to admit that there would not be peer disagreement here. There would still be a point to having a peer disagreement condition, since many cases of peer disagreement are cases in which I do not see your reasoning. Independence need not be read as requiring that all features of your belief must be set aside. In some cases, if your belief is by my lights crazy, I can use that to conclude that you are not my peer. Independence requires only that the mere fact that you disagree with me must be set aside when I decide whether you are my peer.8

122  Robert Shaver

Skepticism It is often true that both conditions for not begging the question are met. Regarding (ii), Sidgwick writes that “where the conflicting beliefs are not contemporaneous, it is usually not clear that the earlier thinker would have maintained his conviction if confronted with the arguments of the later” (1905, 464).9 Historians often ask themselves “Would so-and-so have still endorsed that argument had she known about these objections? Would she have disagreed with the reasoning underlying the objections?” Sidgwick notes that “I have often longed to call up the philosopher of Königsberg himself,” to ask his opinion of current reconstructions of his theory (1883, 71). Where I cannot communicate with the author, I often lack evidence about whether the author would agree with my assessment of her reasoning.10 If these conditions are often met, conciliation and independence will not lead to suspension of judgment as often as it seemed they might. If I have an objection to you that meets both conditions, I need not suspend judgment. There is some reason to think Sidgwick agrees. He writes that “I have spoken of the history of thought as revealing discrepancy between the intuitions of one age and those of a subsequent generation.” He then makes the point about non-contemporaneous beliefs. He goes on to note that the “history of thought, however, I need hardly say, affords abundant instances of similar conflict among contemporaries” (1905, 463–464). His view seems to be that disagreement is benign among non-contemporaries but possibly worrisome among contemporaries. But the reason he gives for thinking disagreement is benign among noncontemporaries seems applicable to many cases of disagreement among contemporaries—we are often unclear about whether the person we disagree with would maintain her position when confronted with our arguments. My argument is very close to one given by Nathan King. King thinks of peers as having the same evidence and abilities. Suspension of belief is mandated only when I have a justified belief that you have the same evidence and abilities I have (and we disagree). King argues that it is rare for us to have the same evidence and abilities and rare for me to be justified in believing that we do (King 2012, 254–257, 261–262; also Audi 2008, 489). If giving objections not considered by one’s opponent is thought of as possessing evidence one’s opponent does not, we are not peers.11 One might reject this argument by moving to a view of peerhood that requires only that peers be equally reliable.12 But provided that my considering arguments that you have not (and not vice versa) makes me justified in believing that I am more likely to be right, a King-style argument against suspension of belief still succeeds.

Peerless  123

Onus When Sidgwick writes “unless [I] can prove independently that the conflicting intuitor has an inferior faculty,” he claims that unless I can show that you are inferior, I lose certainty. The same seems true for the Methods: “If I have no more reason to suspect error in the other mind than in my own” suggests that, to avoid a loss of confidence, I must have some positive reason for thinking you are more likely to err than me. The onus is on me; if I do not discharge it, I should think of you as my peer. Distinguish this very demanding view from a less demanding one: I lose certainty only if I have reason to think that you are my peer. Here a case must be made that you are my peer; without such a case, your disagreement does not lower certainty. The less demanding view seems to give a better position: (1) If you disagree with me about everything, I cannot show that you are inferior to me independently of all the things we disagree about, since there is nothing that we do not disagree about; the result of requiring that I show that you are inferior would be skepticism. If instead my belief-revision is required only if I have reason to believe that you are my peer, your disagreement with me about everything is harmless, since here I cannot have reason to believe that you are my peer (see Vavova 2014, 316–318; Christensen 2009, 760, 2011, 14–15). (2) If we disagree, but I know nothing about you or your reasons, it seems wrong for me to lose confidence; I need some reason to think you must be taken seriously.13 I think Sidgwick should hold the less demanding view. But he may think that holding the more demanding view is not worrisome, at least for the disagreements he has in mind. If, as suggested earlier, it is fairly easy to dismiss you as a peer, it is not so costly to hold that the default position is that you are my peer. Sidgwick, then, might avoid skepticism in two ways. He might think that disagreement justifies decreased confidence rather than suspension of judgment. Or he might think that peer disagreement justifies suspension, but that peers are uncommon. There is a limitation to the second way of avoiding skepticism. Sidgwick does not rule out egoism. He must, then, either read his axioms in some weak way such that an egoist will agree to them, or understand egoism in some weak way such that it is compatible with the axioms, or, again, hold that peer disagreement justifies merely decreased certainty rather than suspension.

Is Kant Sidgwick’s Peer? Sidgwick holds these axioms to have the highest degree of certainty:14 justice: “it cannot be right for A to treat B in a manner in which it would be wrong for B to treat A, merely on the ground that they

124  Robert Shaver are two different individuals, and without there being any difference between the natures or circumstances of the two which can be stated as a reasonable ground for difference of treatment” (ME 380). personal irrelevance: “the good of any one individual is of no more importance, from the point of view . . . of the Universe, than the good of any other; unless, that is, there are special grounds for believing that more good is likely to be realised in the one case than in the other” (ME 382). general good: “as a rational being I am bound to aim at good generally . . . not merely at a particular part of it” (ME 382). From personal irrelevance and general good, Sidgwick deduces: benevolence: “each one is morally bound to regard the good of any other individual as much as his own, except in so far as he judges it to be less, when impartially viewed, or less certainly knowable or attainable by him” (ME 382). Some think that Sidgwick adds: prudence: I ought to seek my good on the whole (e.g., Hurka 2014, 159; Crisp 2015, 117–119). After stating his axioms, Sidgwick writes that I should . . . rely less confidently on the conclusions set forth . . . if they did not appear to me to be in substantial agreement—in spite of superficial differences—with the doctrines of those moralists who have been most in earnest in seeking among commonly received moral rules for genuine intuitions of Practical Reason. (ME 384) Of the (better) “earlier intuitional school . . . there is no one who shows more earnestness in the effort to penetrate to really self-evident principles than Clarke” (ME 384). “Among later moralists, Kant is especially noted for his rigour in separating the purely rational element of the moral code” (ME 385). Clarke and Kant are “the two thinkers who in modern times have most earnestly maintained the strictly scientific character of ethical principles” (ME1 357). Sidgwick argues that Clarke and Kant agree with justice and benevolence. In the contents, the Clarke and Kant section bears the description “This [statement of the axioms] is confirmed by a reference to Clarke’s and Kant’s systems” (ME xxxv). It is natural to read Sidgwick’s noting their agreement as showing an absence of disagreement with two peers. But if the account here explains

Peerless  125 how Sidgwick avoids skepticism, he should not think that Kant, at least, is a peer. I consider prudence, benevolence, and justice. Suppose that prudence is an axiom.15 Sidgwick notes that Clarke agrees with prudence (ME 384n4) but that Kant disagrees. Kant holds that “it can only be stated as a duty for me to seek my own happiness so far as I consider it as a part of the happiness of mankind in general,” whereas Sidgwick holds “with Butler that ‘one’s own happiness is a manifest obligation’ independently of one’s relation to other men” (ME 386). Crisp remarks that it “would be somewhat difficult for Sidgwick to undermine Kant’s epistemic capacities regarding Prudence at the same time as appealing to his views on Justice and Benevolence” (Crisp 2015, 126).16 If Kant’s lack of disagreement concerning justice and benevolence is important to note because he is a peer, his disagreement with prudence should scuttle the claim that prudence has the highest certainty. But Sidgwick seems entirely untroubled by Kant’s disagreement with prudence. One explanation, at least for prudence, is that Sidgwick understands the disagreement test as suggested earlier. Sidgwick writes that Kant “explains that to seek one’s own happiness cannot be prescribed as a duty because it is an end to which every man is inevitably impelled by natural inclination” (1896, 275). Kant writes that his own happiness is an end that every man has . . . but this end can never without self-contradiction be regarded as a duty. What everyone already wants unavoidably, of his own accord, does not come under the concept of duty, which is constraint to an end adopted reluctantly. (MM 386; also 387, 451, G 414; R 6n)17 Sidgwick has a related thought: If there is a “psychological law” requiring me always to choose the action that I believe will maximize my happiness, we cannot say that I ought to maximize my happiness, since “this latter must be a rule from which I am conscious that it is possible to deviate” (ME 41; also 34–35). Later Sidgwick makes a slightly different point (about common usage rather than conceptual necessity): we do not commonly apply the term ‘duty’ to right actions . . . when we are so strongly impelled to them by non-moral inclinations that no moral impulse is conceived to be necessary for their performance. Thus we do not say generally that it is a duty to eat and drink enough: though we do often say this to invalids who have lost their appetite. Duties are “Right actions . . . for the adequate accomplishment of which a moral impulse is conceived to be at least occasionally necessary” (ME 217). This is close to the account of duty Sidgwick ascribes to Kant.

126  Robert Shaver Where Sidgwick clearly disagrees is in thinking that (1) it is possible for me to fail to want my happiness, and (2) moral impulses are sometimes needed to get me to pursue my happiness. Sidgwick, like Butler, thinks that I have many (non-moral) desires directed at things other than my happiness or pleasure. It is contingent whether I desire, in addition, the happiness I might get by satisfying these desires. At one point, he directs the point against Kant: It is not (as Kant seems to hold) that the natural determination of the Will is by motives of pleasure and pain, but that when our action is truly rational, a higher law of causation comes into play, apparently— though not really—contradictory to this natural law. Rather (as Butler maintains) in the promptings of Sense no less than in those of Intellect or Reason we find the phenomenon of strictly disinterested impulse: base and trivial external ends may be sought without ulterior aim, as well as the sublime and ideal: and there are many pleasures of the merely animal life which can only be obtained on condition of not being directly sought, no less than the satisfactions of a good conscience. (ME1 39–40; see, in general, ME 1.4) Where I lack a desire for my happiness, moral impulses have a role: We do not all look with simple indifference on a man who declines to take the right means to attain his own happiness, on no other ground than that he does not care about happiness. Most men would regard such a refusal as irrational, with a certain disapprobation; they would thus implicitly assent to Butler’s statement that “interest, one’s own happiness, is a manifest obligation.” (ME 7) Kant himself admits a man in whom “the universal inclination towards happiness has failed to determine his will” and in which there is a place for “the law of furthering his happiness, not from inclination, but from duty” (G 399).18 Thus, Sidgwick is untroubled by Kant’s disagreement with prudence because he thinks an empirical claim in Kant’s argument against prudence is false. Sidgwick does not think that claim is false because it leads to the denial of prudence, and presumably does not find in Kant explicit disagreement with the arguments of Butler or Methods 1.4. Sidgwick does have more reason to suspect error concerning prudence in Kant than in himself, since Sidgwick has non-question-begging reason for rejecting Kant’s argument (and presumably, by his own lights, does not have reason to reject his own argument for prudence).19

Peerless  127 But if so, Sidgwick should not think that Kant’s agreement with benevolence helps to show that benevolence passes the disagreement test. For Sidgwick takes Kant’s argument for benevolence to be no better than his argument against prudence. Sidgwick writes that “Kant’s conclusion appears to agree to a great extent with the view of the duty of Rational Benevolence that I have given:—though I am not altogether able to assent to the arguments by which Kant arrives at his conclusion” (ME 386). The arguments are not “cogent” (ME 389). In the case of the universal law formulation, Kant argues that a “ ‘will that resolved [not to help others in need] would be inconsistent with itself, for many cases may arise in which the individual thus willing needs the benevolence . . . of others’ ” (ME 389; G 423; also MM 393, 453). Sidgwick objects that we can certainly conceive a man in whom the spirit of independence and the distaste for incurring obligations would be so strong that he would choose to endure any privations rather than receive aid from others. But even granting that every one, in the actual moment of distress, must necessarily wish for the assistance of others; still a strong man, after balancing the chances of life, may easily think that he and such as he have more to gain, on the whole, by the general adoption of the egoistic maxim; benevolence being likely to bring them more trouble than profit. (ME 389) Sidgwick makes two points: (1) some might not will to be helped,20 and (2) even if one does will the help of others at the time of need, a “strong man” might adopt a policy of ignoring this willing, on the ground that he gains more by this policy than by a policy of helping.21 In the case of the humanity formulation, Sidgwick objects that the humanity which Kant maintains to be an end in itself is Man. . . in so far as rational. But the subjective ends of other men, which Benevolence directs us to take as our own ends, would seem, according to Kant’s own view, to depend upon and correspond to their nonrational impulses. . . . It is hard to see why, if man as a rational being is an absolute end to other rational beings, they must therefore adopt his subjective aims as determined by his non-rational impulses. (ME 390) Sidgwick does note that Kant argues that “ ‘the ends of any subject which is an end in himself, ought as far as possible to be my ends also, if the conception of him as an end in himself is to have its full effect with me’ ” (1896, 275; ME 390; G 430). Presumably Sidgwick finds this

128  Robert Shaver unsatisfactory because it does not explain why valuing your rational nature requires me to make your ends mine. I could, instead, try only to develop your rational nature (1896, 274) or refrain from interfering with your pursuit of your ends. Perhaps in some cases helping you fails to value your rational nature, since it treats you as someone incapable of working out how to secure your ends, or stops you exercising your rational nature in an attempt to overcome the problem on your own.22 Sidgwick also objects that the “conception of ‘humanity as an end in itself’ is perplexing: because by an End we commonly mean something to be realised, whereas ‘humanity’ is, as Kant says, ‘a self-subsistent end’ ” (ME 390). Sidgwick’s point might not be just about the use of “end” but also about how something “self-subsistent” can underwrite a duty to bring about changes. You have your humanity regardless of what I do, so it is unclear why a concern for your humanity should make me care about making you happier.23 One way, then, of connecting humanity and happiness is ruled out: I do not promote your humanity by increasing your happiness. Sidgwick, then, has more reason to suspect Kant of error regarding benevolence than himself—although they agree on benevolence, Kant is lucky to have arrived at it.24 Justice is trickier. In favor of justice, Sidgwick writes that in the case of ethical objectivity . . . we commonly refuse to admit . . . variations for which we can discover no rational explanation. . . . [I]t will be generally agreed that we cannot admit a[n] unexplained variation. We cannot judge an action to be right for A and wrong for B, unless we can find in the nature or circumstances of the two some difference which we can regard as a reasonable ground for difference in their duties. (ME 208–209) Sidgwick comments that “this appears to me self-evident” (ME 380). He claims that Kant, in giving the universal law test, agrees (ME 209). If so, Kant can be safely cited as a peer. But again, I do not think Kant’s agreement should, for Sidgwick, count for much. First, Sidgwick does not agree with the argument Kant offers for the universal law test. That argument depends on thinking that the good will is the only unconditioned good and that a good will must act out of duty. Sidgwick suggests, against the former, an argument parallel to one Kant offers against happiness: Adding good will to a mistaken conception of one’s duty makes things worse, not better (ME 395; G 393). He also notes that since “some things are likely to be better done, if they are done from other motives than conscious obedience to practical Reason . . . conscious rightness of volition is not the sole ultimate good” (ME 395). Against the claim that a good will must act out of duty, he notes both the foregoing

Peerless  129 point and that we think virtue involves an emotional element (ME 222–223, 239). Second, insofar as Kant arrives at justice by means of the universal law test, Sidgwick disagrees, since he disagrees with the universal law test (as Kant understands it). Sidgwick claims that the “rule of ‘acting on a maxim that one can will to be law universal’ [when] duly restricted, is an immediate practical corollary” of justice (ME 386). When Sidgwick explains the “duly restricted” qualification, he notes that justice allows the maxim “lie, when doing so is useful and I believe that my action will not be widely imitated.” The Kantian principle . . . —at least so far as I have accepted it as self-evident—means no more than that act, if right for any individual, must be right on general grounds, and therefore for some class of persons; it therefore cannot prevent us from defining this class by the above-mentioned characteristic of believing that the act will remain an exceptional one. (ME 486–487; also 318–319) It cannot be right for me to lie, and for you not to lie, just because I am I and you are you; but it can be right for me to lie and for you not to lie, if I believe my lying will not encourage others to lie whereas you believe your lying will encourage others to lie. What matters is just that the difference in whether others will be encouraged to lie is “a reasonable ground for difference of treatment.” Kant’s concern about whether, in a world in which the maxim was universalized, there would be lies, or I could succeed by telling my lie, is irrelevant. Sometimes, Sidgwick views the difference between himself and Kant as a matter of applying the axiom: “I think that Kant, in applying this axiom, does not take account of certain restrictive considerations” (ME 386n1). But Sidgwick’s application shows, I think, that his understanding of the axiom is very different from Kant’s. His “at least so far as I have accepted it as self-evident” hints that he realizes this.25 The point can be made in another way. Sidgwick considers the Kantian thought that if it were once generally understood that lies were justifiable under certain circumstances, it would immediately become quite useless to tell the lies because no one would believe them; and . . . the moralist cannot lay down a rule which, if generally accepted, would be suicidal. One of Sidgwick’s objections is that the ultimate result of general unveracity under the circumstances would be a state of things in which such falsehoods would no longer

130  Robert Shaver be told: but unless this ultimate result is undesirable, the prospect of it does not constitute a reason why the falsehoods should not be told so long as they are useful. (ME 318) My maxim to lie might fail the universal law test, in that, in a world in which my maxim is made a universal law, I cannot succeed by lying. But Sidgwick thinks it could still be permissible to lie. Since Sidgwick (rightly) takes his position on lying to be consistent with justice, but it is not consistent with the universal law test, justice and the universal law test differ. (Sidgwick and Kant may also disagree about the content of benevolence. Kant’s universal law example concerns helping those “with great hardships (and whom he could easily help)” (G 423).26 This suggests Sidgwick’s “lower” duty of benevolence, “to render, when occasion offers, such services as require either no sacrifice on our part, or at least one very much less in importance than the service rendered” (ME 253). Benevolence lacks these restrictions. But Kant’s humanity version of the example notes only the “happiness” of others, and the duty is to “further the ends of others” (G 430). Similarly, in The Metaphysics of Morals, Kant sometimes writes of those “in need,” sometimes of “happiness and well-being” (e.g., MM 452–453). So it is unclear how much Kant and Sidgwick differ.)27

Crisp Crisp agrees that in some cases, I can disregard your disagreement with me—say I have reason to believe that you disagree with me about the identity of a bird because your glasses, unlike mine, are grubby. But Crisp thinks that there is nothing like grubby glasses in a disagreement over utilitarianism or the humanity formulation. “[T]here may be no theory of error available”; nor does the disagreement itself help, since “no ethical theorist can plausibly deny that the second formulation of the Categorical Imperative and utilitarianism are equally reasonable. That is why asserting either one cannot be taken to be evidence of epistemic inferiority” (Crisp 2011, 157). But insofar as Kant and Sidgwick give arguments for the second formulation and for utilitarianism, there is room for something akin to grubby glasses: the grounds for their conclusions, like the grounds of the rival birdwatcher, can be assessed. Crisp does specify that in his example, belief in the humanity formulation (and utilitarianism) is “non-inferential and pre-theoretical: you see it as stating a fundamental principle of ethics” (Crisp 2011, 153). For the humanity formulation, he cites Audi, for whom that is true. But it is not true of Kant. (Nor does Sidgwick present benevolence as non-inferential; it is derived from personal irrelevance and general good. Prudence is (perhaps) derived from the “distinction passage” argument of Methods 498.)

Peerless  131 Kant arrives at the second formulation by means of the complex arguments of Groundwork 428–429. (Paton, for example, finds four different arguments here (Paton 1958, 175–178).) Worse, he takes the value of humanity to be a “postulate” that depends on the sticky arguments of Groundwork III (G 429n). And much of the appeal of Kant, I think, is that he does give arguments here, such as the argument, so prized by Wood and Korsgaard, that to value anything, I must value my rational capacity.28 Nor is Sidgwick’s disagreement with Kant simply a matter of not seeing the truth of some normative claim. In the case of prudence, Sidgwick’s disagreement concerns empirical claims: When Kant argues against prudence, he seems to think, falsely, that my non-moral motivation is always and only directed toward my happiness, and that moral persuasion to get me to pursue my happiness is always unnecessary. In the case of benevolence, Sidgwick makes a point about possibility—not everyone must want help when in need—and a point about the insufficiency of Kant’s premises, whether in the universal law formulation or the humanity formulation, for arriving at benevolence. None of this handles disagreements of the sort that Crisp has in mind. One suggestion is that non-inferential beliefs are justified by seemings (see, for example, Huemer 2005, 99–102). If we disagree about p because our seemings are different, we do not share the same evidence, and so, on one construal, are not peers (King 2012, 256–257; also van Inwagen 2010, 26–27). On a better construal, however, we might still be peers: We might have different evidence but be equally reliable. A better suggestion is that even if the beliefs are non-inferential, there is still a place for objections. One can point to applications of the beliefs that may be unacceptable. For example, Ross writes that “it might seem absurd to suggest that it could be right for any one to do an act which would produce consequences less good than those which would be produced by some other act in his power.” But he thinks the appearance of self-evidence can be dispelled by considering applications—it is not self-evident, or true, he thinks, that I should break a promise to produce 1001 units of good rather than keep it to produce 1000 (Ross 1930, 34–35). Or consider Thomson’s loop trolley case, or Parfit’s many examples in On What Matters, directed against the humanity formulation. Again, the thought is that once one sees some application of the belief—one not foreseen by its advocate— there is reason to think one would give up the belief.29 If so, Sidgwick’s position on peer disagreement need not lead him to skepticism—though it should lead him to discount Kant’s agreement with him.30

Notes 1. Earlier, Sidgwick wrote that a “proposition which presents itself to my mind as self-evident, and is in harmony with all the rest of my intuitions relating

132  Robert Shaver to the same subject, and is also ascertained to be accepted by all other minds that have been led to contemplate it . . . seems to have as high a degree of certainty as I can hope to attain” (1879, 109). 2. For a similar argument, see Elga (2007, 499 n 21) and Dougherty (2013, 222). 3. See, for example, Elga (2007, 484); McGrath (2008, 91); Enoch (2010, 956); Vavova (2014, 307); and Dougherty (2013, 222–223). 4. Lackey (2010) objects to “conciliatory” positions that they cannot give weight to the numbers—on them, the disagreement of just one peer is enough to greatly reduce my confidence, even if many others agree with me. Sidgwick could reply that, where thousands agree with me and none with you, that is reason to think you are not my peer, since that is reason to think you are more likely to have made an error (even if on some other accounts of peerhood, such as equal evidence and training, you are still my peer). 5. Wedgwood (2010, 223–224); Hills (2010, 152); Crisp (2006, 93); Crisp (2007, 33–35); Crisp (2011, 151–152, 159); and Crisp (2015, 111) read Sidgwick as requiring suspension of judgment. Hurka (2014, 114) takes Sidgwick’s point to concern reduction of confidence, though he also takes Sidgwick to share Elga’s equal weight view. McGrath (2008, 91) takes Sidgwick to be making a point about certainty, though thinks his point also justifies holding that knowledge is lost. Vavova (2014, 327n1) lists Sidgwick as a conciliationist. Setiya (2012, 12) says that Sidgwick’s position is “nuanced,” but that “it is in the vicinity of a simpler view, that the correct response to disagreement with an ‘epistemic peer’ is to become agnostic.” 6. The condition could still do work in one case. Say I think you are slightly more reliable than me. I then take your claim that not-p to decrease your reliability to my level. Given that we are now peers, I conclude that I should be less certain that p is true. 7. The case for dismissing you seems even stronger if I can reconstruct how you understandably went wrong. For an example, see Fumerton’s discussion of the Monty Hall problem (Fumerton 2010, 94). Tyler Paytas noted to me that Sidgwick employs this strategy in dismissing the objection that his “maxim of Benevolence” goes beyond the more limited benevolence endorsed by common sense. Sidgwick argues that the more limited benevolence is just what the maxim of benevolence “practically” calls for (ME 382). 8. Crisp agrees: “I take Sidgwick to be claiming not that the proof of inferiority has to be entirely independent of the content of the conflicting intuition itself, but that it has to be independent of the mere fact of disagreement itself” (2011, 225–226). For example, if “your colleague responds to your assertion of the Categorical Imperative with, say, the principle that one should act so as maximally to promote the doctrines of the Unification Church, it would be absurd not to take this as evidence of one’s epistemic superiority to her” (2011, 156–157). Christensen also agrees, adding that this is consistent with independence, since I can invoke dispute-independent considerations to favor myself. For example, in a case where you calculate my share of the bill as greater than the bill as a whole, I can reason that you did not employ the rule “my portion cannot be larger than the bill as a whole” (Christensen 2007, 201, 2011, 9–12). I say “in some cases” because perhaps a charge of craziness that is not backed by some dispute-independent consideration does not take away peerhood. 9. Kelly assumes that if one finds in a “great dead philosopher” “something akin to a clear fallacy, the kind of thing such that if you could only travel back in time and present it to the author, a modicum of intellectual humility

Peerless  133 would compel him to change his mind,” all agree that I need not suspend my judgment that the philosopher’s argument is weak (Kelly 2013, 32). McGrath gives an example of two contemporaneous philosophers: Alice “says that the fetus has a right to life, and that it follows that abortion is impermissible. But I have read Judith Jarvis Thomson’s ‘In Defense of Abortion,’ and so I can supply an argument that even if the fetus does have a right to life it does not follow that abortion is impermissible. I take myself to have rebutted her argument and thus to have more reason to think that she is in error than that I am” (McGrath 2008, 102; see, also, 91–92, 103). One could also see this as showing that they are not peers on other understandings of peerhood— McGrath has evidence that Alice lacks. 10. Hence, the clearest cases of peers who know each other’s reasoning are cases in which they talk to one another. Van Inwagen notes that he and David Lewis “had many conversations and engaged in a rather lengthy correspondence on the matter of compatibilism and incompatibilism, and, on the basis of these exchanges—not to mention his wonderful paper ‘Are We Free to Break the Laws?’—I am convinced beyond all possible doubt that David understood perfectly all the arguments for incompatibilism that I am aware of. . . . It seems difficult, therefore, to contend that, in this matter, he was in epistemic circumstances inferior to mine” (van Inwagen 2010, 23–24). Van Inwagen could not claim that Lewis would agree with him if shown the relevant arguments. 11. When Kelly specifies what he means by “peer,” he notes that “you and I have been exposed to the same evidence and arguments that bear on some proposition: there is no relevant consideration that is available to you but not to me, or vice versa.” We “are thoroughly acquainted with all of the extant arguments, thought experiments, and intuition pumps that the literature has to offer” (Kelly 2010, 111–112; also van Inwagen 2010, 19–26). 12. Vavova (2014, 329n20) makes this objection to King. 13. Paytas objected that if a stranger at a train station disagrees with me about the time of the next train, I should lose confidence, even knowing nothing about the stranger. I think, however, that if I should lose confidence, this is because I make assumptions about the stranger’s ability—I take her to have some ability just given that she is in the station, apparently capable with trains. Perhaps (2) should read “If we disagree, but I have no evidence about you.” 14. I leave aside temporal irrelevance (ME 381). It is not at issue here. 15. My view is that it is not; see Shaver (1999, 74–77), Skelton (2008, 200–202). But even if prudence has a different status than the axioms, Sidgwick certainly believes prudence and certainly is not troubled by Kant’s denial of it. 16. Similarly, Skelton writes that it “would be odd for Sidgwick to appeal to Kant for justification [of his axioms] when he rejects so many of Kant’s other views” (Skelton 2013, para 26). Skelton argues that peer disagreement reduces justification, but agreement does not increase justification. But the oddity Skelton notes raises doubts about why Sidgwick would appeal to Kant’s lack of disagreement. 17. MM = Kant 1996; G = Kant 1964; R = Kant 1960. For one discussion of Kant’s various arguments against prudence, see Hills (2010, 44n15). For an ingenious “Kantian” account very different from that quoted here, and which does not rely on the claim that I unavoidably want my own happiness, see Hill (2002, 186–191). 18. It is unclear how to reconcile this passage with Kant’s argument against prudence. (There is a further textual issue. Wood and Johnson note that Kant’s

134  Robert Shaver view that I “by a natural necessity” (G 415) have my happiness as my end seems inconsistent with Kant’s claim that I am free to set my ends (Wood 1999, 66; Johnson 2002). To avoid this, Wood reads the necessity as rational necessity, but then (as Johnson notes) Kant agrees with prudence, contrary to his official position. For recent textual discussion of Kant’s apparent psychological hedonism, see Reath (2006, ch. 2); Kerstein (2002, 23–32); Herman (2007, ch. 8). 19. Parfit writes that the Methods “contains the largest number of true and important claims. It is not surprising that, though a less great philosopher than . . . Kant, Sidgwick could write a better book. Sidgwick lived later. Unlike later poets or playwrights, who have no advantages over Homer or Shakespeare, later philosophers do have advantages, since philosophy makes progress” (Parfit 2011, I: xxxiii). 20. Kant may have denied this on the basis, again, of the claim that I cannot fail to want my happiness (see Hurka 2014, 277). 21. Sidgwick’s objection has prompted some surprising interpretations of Kant. Rawls takes the objection to require that when I test a maxim, I must imagine myself ignorant of my particular features and of what place I occupy in the world in which my maxim is universalized (Rawls 2000, 175–176). When Hill tries to justify benevolence, he suggests an application of justice that starts, not with a desire to be helped but with the thought that it “makes sense for others to help us when we are in dire need” (hence it would be wrong for me to refuse to help them) (Hill 2002, 122–124). 22. Contemporary Kantians often move very quickly here. Wood claims without further argument that the “happiness of human beings is an end of morality because setting this end shows respect for rational nature” (Wood 2008, 102; also Wood 1999, 149–150). Rawls suggests that I value your rational nature by doing what the universal law formulation shows to be required—which, again, does not explain the connection and inherits whatever difficulties the universal law argument faces (Rawls 2000, 194). Hill, when initially listing the practical consequences of his view of valuing humanity, does not list benevolence (Hill 1980, 93–94). Later he writes that “one should help others to set their own ends and rationally pursue them rather than try to make their lives pleasant independently of their own goals,” but this does not explain why (or even claim that) I should in general help you achieve your ends—it does not explain why, say, I should give you the stamp you need to complete your collection, when doing so has no effect on your ability to set or rationally pursue ends. It also conflicts with Kant’s claim (noted by Sidgwick and Hill) that I have no duty to promote the perfection of others (1896, 274–275; Hill 1980, 95–96, MM 386). Hill comments that “it is hard to see how a duty [of general beneficence] would follow from the principle to treat humanity as an end” (97). O’Neill takes the duty to be a duty to secure the “agency” of others (O’Neill 1989, 140). 23. For this worry directed at Audi, see Hurka (2007, 69–70); for it directed at Kant, see Ross (1952, 52) (noted by Hurka). 24. He is lucky insofar as he really arrives at it by means of his arguments rather than by seeing its truth directly. Paytas noted that perhaps Sidgwick took Kant to have seen its truth directly and then messed up by adding (bad) arguments. This might let Sidgwick count Kant’s agreement as helpful, even if he thinks Kant’s arguments are not. 25. O’Neill (2001, 86–89) notes the large gap between justice and Kant’s universal law test. 26. Thus some think Kant’s version of benevolence is limited to the satisfaction of “true needs,” or even those needs “that cannot be left unmet if he is to

Peerless  135 continue in his activity as a rational agent” (Herman 1993, 67, 69–72; see, also, Rawls 2000, 173–175). 27. Kant also claims that the duty is imperfect, whereas Sidgwick seems to take the upshot of benevolence to be that I must help, provided helping maximizes happiness (for discussion, see Crisp 2015, 126; Hurka 2014, 163; Irwin 2009, 495, 499, 501, 504–505, 510, 511, 512–513, 518). Whether this is a significant difference depends on the circumstances in which Kant thinks I must help, and on whether Sidgwick intends benevolence to be a prima facie duty. For the latter, see Shaver (2014, 185–200). 28. I put aside (persuasive) doubts about whether Kant really gives this argument (see Stern 2012). 29. McGrath sees non-inferential moral beliefs and beliefs about unique green (the shade of green that is neither bluish nor yellowish) to be analogous (McGrath 2008, 99–103). One disanalogy is that most moral beliefs have applications that can be questioned, whereas beliefs about unique green do not seem to imply many other beliefs. 30. Thanks to Joyce Jenkins and Tyler Paytas, for comments on earlier drafts, and to the audience at the Stuttgart Kant-Sidgwick conference.

References Audi, R. 2008. “Intuition, Inference, and Rational Disagreement in Ethics.” Ethical Theory and Moral Practice 11: 475–492. Christensen, D. 2007. “Epistemology of Disagreement: The Good News.” Philosophical Review 116: 187–217. Christensen, D. 2009. “Disagreement as Evidence: The Epistemology of Controversy.” Philosophy Compass 4: 756–767. Christensen, D. 2011. “Disagreement, Question-Begging, and Epistemic SelfCriticism.” Philosophers’ Imprint 11 (6). Crisp, R. 2006. Reasons and the Good. Oxford: Oxford University Press. Crisp, R. 2007. “Intuitionism and Disagreement.” In Rationality and the Good, edited by J. Greco, A. Mele, and M. Timmons. New York: Oxford University Press. Crisp, R. 2011. “Reasonable Disagreement: Sidgwick’s Principle and Audi’s Intuitionism.” In The New Intuitionism, edited by J. Hernandez. London: Continuum. Crisp, R. 2015. The Cosmos of Duty: Henry Sidgwick’s Methods of Ethics. Oxford: Oxford University Press. Dougherty, T. 2013. “Dealing with Disagreement from the First-Person Perspective: A Probabilist Proposal.” In Disagreement and Skepticism, edited by D. Machuca. New York: Routledge. Elga, A. 2007. “Reflection and Disagreement.” Nous 41: 478–502. Elgin, C. 2010. “Persistent Disagreement.” In Disagreement, edited by R. Feldman and T. Warfield. Oxford: Oxford University Press. Enoch, D. 2010. “Not Just a Truthometer: Taking Oneself Seriously (but not Too Seriously) in Cases of Peer Disagreement.” Mind 119: 953–997. Fumerton, R. 2010. “You Can’t Trust a Philosopher.” In Disagreement, edited by R. Feldman and T. Warfield. Oxford: Oxford University Press. Herman, B. 1993. The Practice of Moral Judgment. Cambridge, MA: Harvard University Press.

136  Robert Shaver Herman, B. 2007. Moral Literacy. Cambridge, MA: Harvard University Press. Hill, T. 1980. “Humanity as an End in Itself.” Ethics 91: 84–99. Hill, T. 2002. Human Welfare and Moral Worth: Kantian Perspectives. Oxford: Oxford University Press. Hills, A. 2010. The Beloved Self. Oxford: Oxford University Press. Huemer, M. 2005. Ethical Intuitionism. New York: Palgrave Macmillan. Hurka, T. 2007. “Audi’s Marriage of Ross and Kant.” In Rationality and the Good, edited by J. Greco, A. Mele, and M. Timmons. New York: Oxford University Press. Hurka, T. 2014. British Ethical Theorists from Sidgwick to Ewing. Oxford: Oxford University Press. Irwin, T. 2009. The Development of Ethics, vol. 3. Oxford: Oxford University Press. Johnson, R. 2002. “Happiness as a Natural End.” In Kant’s Metaphysics of Morals: Interpretive Essays, edited by M. Timmons. Oxford: Oxford University Press. Kant, I. 1960. Religion with the Bounds of Reason Alone. Translated by T. Greene and H. Hudson. New York: Harper. Kant, I. 1964. Groundwork of the Metaphysic of Morals. Translated by H. Paton. New York: Harper. Kant, I. 1996. The Metaphysics of Morals. Translated by M. Gregor. New York: Cambridge University Press. Kelly, T. 2010. “Peer Disagreement and Higher-Order Evidence.” In Disagreement, edited by R. Feldman and T. Warfield. Oxford: Oxford University Press. Kelly, T. 2013. “Disagreement and the Burdens of Judgment.” In The Epistemology of Disagreement, edited by J. Lackey and D. Christensen. Oxford: Oxford University Press. Kerstein, S. 2002. Kant’s Search for the Supreme Principle of Morality. Cambridge: Cambridge University Press. King, N. 2012. “Disagreement: What’s the Problem? Or a Good Peer Is Hard to Find.” Philosophy and Phenomenological Research 85: 249–272. Lackey, J. 2010. “What Should We Do When We Disagree?” Oxford Studies in Epistemology 3: 274–293. McGrath, Sarah. 2008. “Moral Disagreement and Moral Expertise.” Oxford Studies in Metaethics 3: 87–108. O’Neill, O. 1989. Constructions of Reason. Cambridge: Cambridge University Press. O’Neill, O. 2001. “Sidgwick on Practical Reason.” In Henry Sidgwick, edited by R. Harrison. Oxford: The British Academy. Parfit, D. 2011. On What Matters, vol. 1. Oxford: Oxford University Press. Paton, H. 1958. The Categorical Imperative. London: Hutchinson. Rawls, J. 2000. Lectures on the History of Moral Philosophy. Cambridge, MA: Harvard University Press. Reath, A. 2006. Agency and Autonomy in Kant’s Moral Theory. Oxford: Oxford University Press. Ross, W. 1930. The Right and the Good. Oxford: Clarendon Press. Ross, W. 1952. Kant’s Ethical Theory. Oxford: Clarendon Press. Setiya, K. 2012. Knowing Right from Wrong. Oxford: Oxford University Press.

Peerless  137 Shaver, R. 1999. Rational Egoism: A Selective and Critical History. Cambridge: Cambridge University Press. Shaver, R. 2014. “Sidgwick’s Axioms and Consequentialism.” Philosophical Review 123: 173–204. Sidgwick, H. 1874. The Methods of Ethics. 1st ed. London: Macmillan. Sidgwick, H. 1879. “The Establishment of Ethical First Principles.” Mind 4: 106–111. Sidgwick, H. 1883. “A Criticism of the Critical Philosophy.” Mind 8: 69–91. Sidgwick, H. 1896. Outlines of the History of Ethics. 4th ed. London: Macmillan. Sidgwick, H. 1905. “Appendix to ‘Criteria of Truth and Error’.” In Lectures on the Philosophy of Kant and Other Philosophical Lectures and Essays. London: Macmillan. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Skelton, A. 2008. “Sidgwick’s Philosophical Intuitions.” Etica & Politica 10: 185–209. Skelton, A. 2013. “Sidgwick’s Argument for Utilitarianism and his Moral Epistemology: A Reply to David Phillips.” Revue d’etudes benthamiennes 12. Stern, R. 2012. “Constructivism and the Argument from Autonomy.” In Constructivism in Practical Philosophy, edited by J. Lenman and Y. Shemmer. Oxford: Oxford University Press. van Inwagen, P. 2010. “We’re Right. They’re Wrong.” In Disagreement, edited by R. Feldman and T. Warfield. Oxford: Oxford University Press. Vavova, K. 2014. “Moral Disagreement and Moral Skepticism.” Philosophical Perspectives 28: 302–333. Wedgwood, R. 2010. “The Moral Evil Demons.” In Disagreement, edited by R. Feldman and T. Warfield. Oxford: Oxford University Press. Wood, A. 1999. Kant’s Ethical Thought. New York: Cambridge University Press. Wood, A. 2008. Kantian Ethics. New York: Cambridge University Press.

Part IV

Freedom of Will

7 Freedom and Happiness Sidgwick’s Critique of Kant Paul Guyer

Among Kantians, Henry Sidgwick is best remembered for his critique of Kant’s treatment of freedom of the will. This was first published in an article in Mind in 1888, 14 years after the publication of the first edition of his magnum opus The Methods of Ethics, and then included in the 1901 sixth edition of Methods, largely completed before Sidgwick’s death in 1900, and retained in the posthumous seventh edition, now standard (ME 511–516). As Sidgwick summed up his critique of Kant in some lecture notes included by his editor E. E. Constance Jones in the Preface to the sixth edition, Kant’s resting of morality on Freedom . . . involves the fundamental confusion of using ‘freedom’ in two distinct senses—‘freedom’ that is realised only when we do right, when reason triumphs over inclination, and ‘freedom’ that is realised equally when we choose to do wrong, and which is apparently implied in the notion of ill-desert. (ME xvii) This criticism and the prominence that it is given are puzzling. The criticism itself is puzzling because Kant seems to have made this distinction, at least by the time of the 1793 Religion within the Boundaries of Mere Reason,1 to which Sidgwick does not refer, although of course not in Sidgwick’s words, but in the form of his insistence upon our radical freedom to choose whether or not to subordinate self-love to the moral law. And the prominence given to this criticism is also puzzling because the issue of free will does not itself seem central to the argument of The Methods of Ethics, being the explicit subject of only one of 35 chapters (although in this chapter Sidgwick does give an exceptionally lucid statement of the position that determinism is the presupposition of all explanations of human action, yet that it is irrelevant to the practical task of decision making and also irrelevant to our practices of holding people responsible and punishing them for their violations as long as we do not accept a retributive theory of punishment; see ME 70–72).2

142  Paul Guyer However, a general discomfort with another aspect of Kant’s position on freedom is central to the main argument of Methods, namely discomfort with Kant’s attempt to derive our determinate duties from freedom of choice—the freedom of each human being to set and pursue her own ends—as the foundational value from which our more determinate duties can be derived, including what Kant calls the imperfect duty to promote the happiness of others and what Sidgwick calls the duty of “Rational Benevolence.” The central debate of The Methods of Ethics, between “dogmatic intuitionism” and utilitarianism, is most centrally a debate over the derivation and determinacy of the duty of benevolence, and Sidgwick’s position is that previous treatments of this duty, including Kant’s, fail in ways that can only be remedied by appeal to the utilitarian principle that “the conduct which, under any given circumstances, is objectively right, is that which will produce the greatest amount of happiness on the whole; that is, taking into account all whose happiness is affected by this conduct” (ME 411). Of course, this is a direct rejection of what Kant himself intended as the alternative to the form of utilitarianism with which he was familiar, that represented, for example, by Francis Hutcheson, for Sidgwick’s claim is that only utilitarianism can give determinate duties, while Kant’s criticism was precisely that the concept of happiness is too indeterminate to give rise to a determinate specification of our duties. In Kant’s words, Happiness is such an indeterminate concept that, although every human being wishes to obtain this, he can still never say determinately and consistently with himself what he really wishes and wills. . . . One cannot therefore act on determinate principles for the sake of being happy, but only on empirical counsels . . . which experience teaches are most conducive to well-being on the average. (G 4:418) In this chapter I will not try to establish whether Sidgwick’s utilitarian principle yields the determinacy of duties that he thought was lacking in previous accounts of benevolence including Kant’s; I would hardly be original if I were to try to argue that Sidgwick fails to establish that meaningful quantifications of happiness and therefore comparisons of quantities of happiness, whether diachronic intrapersonal, synchronic interpersonal, or diachronic interpersonal, are really possible. I will only argue that Kant has a more plausible derivation of the duty of benevolence than Sidgwick gives him credit for; the problem is that this derivation is not a direct derivation from the value of happiness but an indirect derivation that goes through the value of freedom, the foundation of Kant’s approach to moral philosophy to which Sidgwick seems to have been blind, perhaps blinded by his focus on the undoubted confusion of

Freedom and Happiness  143 some of Kant’s treatments of the freedom of the will in the works prior to the Religion. Examining the freedom-based grounds of Kantian benevolence will also help to explain why, unlike Sidgwick, Kant could not have accepted that the duty of benevolence is the sole or supreme principle of morality. The issue between Sidgwick and Kant is to some extent obscured by the facts that Kant’s account of the duty of benevolence is only part of his larger system of duties, as Sidgwick himself recognized, and that Sidgwick’s utilitarian principle itself functions within a larger account of duty ethics that also shares elements with Kant’s own wider scheme. Sidgwick does not reject Kant’s system of duties as a whole, although he does criticize some aspects of the system besides the duty of benevolence, nor does he in fact attempt to replace it entirely and solely with the utilitarian principle. Rather, he does leave in place central features of the commonsense morality that he took Kant to have attempted to formalize. So we have to consider Sidgwick’s critique of Kant’s treatment of the positive duty of benevolence, which reveals his failure to appreciate that Kant was attempting to derive even this duty to promote happiness not from the value of happiness as such but from the value of freedom, in the larger context of Sidgwick’s argument.

Sidgwick’s Strategy Sidgwick considers three “methods of ethics,” namely “egoistic hedonism,” “intuitionism,” and “universalistic hedonism,” or for short “egoism,” “intuitionism,” and “utilitarianism.” By a “method of ethics” Sidgwick means “any rational procedure by which we determine what individual human beings ‘ought’—or what it is ‘right’ for them—to do, or to seek to realize by voluntary action” (ME 1), or a normative principle or set of principles about what we ought to do rather than an epistemological or meta-ethical principle about how we know what we ought to do. But while egoistic and universalistic hedonism do sound like normative positions, commanding either the pursuit of one’s own happiness or the pursuit of the happiness of all as the basis of all duty, intuitionism sounds like a meta-ethical rather than normative method of ethics, that is, it would seem to concern the basis on which duty is known rather than the content of duty, namely by intuition or perception rather than by induction, inference, or some other method. But in examining a third general method of ethics alongside egoism and utilitarianism, Sidgwick focuses on a specific type of intuitionism that he labels “dogmatic intuitionism.” This method involves appeal to the non-consequentialist duties that are found within common-sense morality and purportedly grasped via intuition. What complicates matters is that Sidgwick also uses “intuitionism” more broadly to describe various means of accessing ethical truths non-inferentially and without appeal to evidence,

144  Paul Guyer including immediate judgements of conscience concerning particular circumstances, as well as grasping abstract foundational principles such as those of universalistic and egoistic hedonism. This non-singular usage of “intuitionism,” as well as the somewhat idiosyncratic use of the term “methods” are among the features of Sidgwick’s work that make the general line of argument more difficult to follow than it might have been otherwise (Crisp 2015, 21, 102; see, also, de Lazari-Radek and Singer 2014, 68). Be all that as it may, the real debate of the book is between the nonconsequentialist duty-based morality and utilitarianism; and since Kant is the greatest champion of the former, the real debate of the book can be seen as between Kantianism and utilitarianism. This is precisely as the real debate in moral philosophy was seen by Kant himself, and as it has remained, Sidgwick himself having dispatched what some have recently regarded as a third main approach, namely “virtue ethics,” before it ever got off the ground with the statement that the “so-called Virtues which can be thought to be essentially and always such . . . as Wisdom, Universal Benevolence, and (in a sense) Justice . . . manifestly involve [a] notion of Good, supposed already determinate” (ME 392–393). But given that Sidgwick actually shares ground with common sense and Kant on what he will call the duties of justice and prudence, the real debate between Kant and Sidgwick concerns the duty of benevolence, including its derivation and the question of whether it is the supreme duty (i.e., whether there are also deontological constraints on promotion of the good). Sidgwick sets the terms of this debate by proposing criteria for a successful method of ethics that any rational person would be hard-pressed to deny; they could be taken as criterial of rationality itself, and were they quoted without citation one might easily imagine that they came from Descartes’s Rules for the Direction of the Mind, Spinoza’s Emendation of the Intellect, or Locke’s Conduct of the Understanding. Sidgwick’s four conditions on any ethical proposition that would “lead us cogently to trustworthy conclusions” in our reasoning are (1) “The terms of the proposition must be clear and precise” (here Sidgwick mentions Descartes as a predecessor); (2) “The self-evidence of the proposition must be ascertained by careful reflection,” not asserted on the basis of “mere impressions or impulses” nor “mere opinions, to which the familiarity that comes from frequent hearing and repetition often gives a false appearance of self-evidence”; (3) “The propositions accepted as self-evident must be mutually consistent”; and (4) “Since it is implied in the very notion of Truth that it is essentially the same for all minds, the denial by another of a proposition that I have affirmed has a tendency to impair our confidence in its validity,” thus “the absence of . . . disagreement must remain an indispensable negative condition of the certainty of our judgments” (ME 338–342).3 From the last two requirements in particular, Sidgwick infers the requirement of determinacy: Any

Freedom and Happiness  145 successful moral principle must exclude contradiction among particular judgments about duty within any one agent’s body of moral judgments and between the moral judgments of different agents. If competing claims to duty do present themselves in either of these cases, there must be a way of resolving the apparent conflict. Sidgwick’s general claim is that “we have no means of reducing” such a conflict “to a common standard, except by the application of the Utilitarian—or some similar—method” (ME 342)—except that in his view there is no other, “similar” method that can do this job. Kant would not disagree with Sidgwick’s conditions—he too insists that any genuine practical law must be truly universal (and therefore necessary), and that there can be no unresolved conflicts of duty.4 The difference between them is rather that Sidgwick denies that Kant’s fundamental principle of morality can satisfy this requirement, while Kant denies that any conception of happiness, thus Sidgwick’s method of utilitarianism, could satisfy the requirement. But there is also an area of agreement between Sidgwick and Kant (see, also, Schneewind 1977, 286). This is as follows. Following his critique of common-sense intuitionism, Sidgwick claims that philosophical reflection reveals three fundamental moral obligations, more basic than a variety of other duties (or virtues) that common sense recognizes, such as truthfulness, courage, and humility, namely the obligations to adhere to principles of justice, prudence, and benevolence. Sidgwick takes Kant’s system to replicate the principle of justice with the universal law formulation of the Categorical Imperative (ME 386). Although Sidgwick objects to Kant’s claim that there is no direct duty to promote one’s own happiness, he does not believe the disagreement over prudence is all that deep (1896, 276). The deep disagreement concerns the ground for the principle of promoting the happiness of others. Sidgwick argues above all that Kant does not have a cogent method of deriving the duty of benevolence, a fortiori a method of deriving that duty that precludes the possibility of conflicts. He claims that his own method of utilitarianism does avoid such conflicts, and also that it provides a deep unifying explanation for the various duties found within common-sense morality. Thus, Sidgwick accepts as self-evident the following two principles. The first is that it cannot be right for A to treat B in a manner in which it would be wrong for B to treat A, merely on the ground that they are two different individuals, and without there being any difference between the natures or circumstances of the two which can be stated as a reasonable ground for differences of treatment. He regards this as equivalent to Kant’s requirement of the universalizability of maxims. The second self-evident principle is that “the mere difference of priority and posteriority in time is not a reasonable ground for

146  Paul Guyer having more regard to the consciousness of one moment [than] to that of another” or, in the form in which this principle “practically presents itself to most men,” “a smaller present good is not to be preferred to a greater future good (allowing for difference of certainty),” that is, allowing a discount for the uncertainty of one’s continued existence or capacity to enjoy the good at issue at any particular time in the future (ME 380–381). This is the principle of prudence and at least part of one’s own practical rationality that ought to be perfected by Kant’s lights as well. But as Sidgwick says, explicitly about the first of these principles but in effect about both, such principles “manifestly” do “not give complete guidance.” The first principle merely throws “a definite onus probandi on the man who applies to another a treatment of which he would complain if applied to himself” but does not establish that this burden of proof cannot be satisfied (ME 380); the second principle is not specific enough about what sorts of present and future goods can be compared to each other. But above all, these two principles do not establish that one has any positive duties to increase the good of others in any form. The first is a negative principle, telling you not to treat one person differently from another, including not treating others differently from yourself, but does not tell you in positive terms what you owe to anyone, yourself or others; the second principle patently applies only to oneself. However, Sidgwick does believe that careful reflection leads to a reliable account of positive duties to others.5 His argument begins from the further self-evident principles that “the good of any one individual is of no more importance, from the point of view (if I may say so) of the Universe, than the good of any other,” and further that “as a rational being I am bound to aim at good generally—so far as it is attainable by my efforts,—not merely at a particular part of it.” These two principles together imply that each is to aim at the good of all, that is, “From these two rational intuitions we may deduce, as a necessary inference, the maxim of Benevolence in an abstract form: viz. that each one is morally bound to regard the good of any other individual as much as his own.” From this it follows in turn that one morally ought to aim at one’s own good only insofar as it is considered a part of the aggregate that is the general good, and as no more an important part than any other. Of course, this principle is to be applied in light of any “special grounds for believing that more good is likely to be realised in the one case than in the other” (ME 382), that is, some reason to believe that one can make a greater contribution to the aggregate general good by contributing to one’s own good or that of those nearer to oneself rather than to that of those who are more remote, or absence of reason to believe that one can effectively contribute to the good of those who are more remote. Sidgwick claims that this principle of Rational Benevolence is “required as a rational basis for the Utilitarian system” (ME 387). The principle is not itself sufficient for utilitarianism in part because it does not purport to

Freedom and Happiness  147 be the sole or supreme principle of ethics; it is logically compatible with deontological constraints on promotion of the good. While Sidgwick interprets Kant as accepting Rational Benevolence, he takes issue with Kant’s account of the grounds of this principle, as well as with Kant’s view that the duty to promote happiness is constrained by other duties. As we shall see presently, these disagreements arise ultimately from Sidgwick’s and Kant’s divergent views on the value of freedom and its relation to the basic principles of morality.

Sidgwick’s Critique of Kant Sidgwick’s criticisms of Kant are not confined to his treatment of benevolence, or as Kant puts it, the imperfect duty to promote the happiness of others (G 4:430 MM 6:385–388). In particular, Sidgwick also objects to Kant’s version of the first of the self-evident ethical intuitions, essentially, reason’s demand that like cases be treated alike. But Sidgwick’s most challenging objection to Kant is his objection to Kant’s treatment of the duty of benevolence, and that this is where Sidgwick’s view that Kant does not have a clear conception of freedom is going to be most costly, either to Kant or to Sidgwick himself. So let us look first at Sidgwick’s other objections to Kant and then at this objection. (1) Sidgwick refers to Kant in his first discussion of the “intuitionist” principle, which however he also accepts, that We cannot judge an action to be right for A and wrong for B, unless we can find in the nature or circumstances of the two some difference which we can regard as a reasonable ground for difference in their duties. (ME 209) He says that applying this test to one’s proposed actions “will often disperse the false appearance of rightness which our strong inclination has given to it,” and to that extent it is valuable—a good way of stepping back from impetuousness, one might suggest. But he thinks that Kant went too far when he “held that all particular rules of duty can be deduced” from his version of this principle, namely “the one fundamental rule ‘Act as if the maxim of thy action were to become by thy will a universal law of nature.’ ” Sidgwick actually has two different criticisms to make here. First, he claims that “this appears to me an error analogous to that of supposing that Formal Logic supplies a complete criterion of truth” (ME 209–210). This could be taken to mean that Kant’s formula of universal law (or universal law of nature)6 supplies only a necessary but not a sufficient condition of duty: that any proposed course of action that fails this test is morally wrong, but the principle is not by itself sufficient to determine everything that is morally obligatory.

148  Paul Guyer This would parallel a point that Kant himself insisted upon in the case of theoretical judgment, namely that satisfaction of the principle of noncontradiction is a necessary but not a sufficient condition of truth.7 Kant does seem to suggest that all classes of our duty can be derived from the formula of universal law alone when he illustrates it in the Groundwork (4:422–423). Some commentators have suggested that Kant did not intend the formula of universal law to ground all human duties but rather intended it to function only as part of a complex categorical imperative, a system of laws, also including the formula of humanity, within which system only the latter formula yields our positive duties to self and others (see Wood 2006, 342–380). Sidgwick will object to Kant’s attempt to derive such duties from the concept of humanity, or as he puts, rational being, and we can take this objection, what I am regarding as his main objection, to moot the question of whether Kant really thought he could derive all our duties from the formula of universal law alone. However, Sidgwick makes what I consider a second objection to Kant’s formula of universal law, namely, that it does not exclude any possible maxim, thus does not even act as a necessary condition or constraint on proposed maxims of actions: I conceive that all (or almost all persons) who act conscientiously could sincerely will the maxims on which they act to be universally adopted: while at the same time we continually find such persons in thoroughly conscientious disagreement as to what each ought to do in a given set of circumstances. (ME 210)8 In other words, the sincere Nazi is perfectly happy to have his maxim, to kill as many Jews as possible, universalized, so that everyone should aim to kill as many Jews as possible.9 Of course the conclusion that this maxim is morally acceptable, let alone obligatory, is blatantly contradictory to every pre-theoretical moral judgment of any reasonable person, and the conflict between the Nazi who believes this and the Jew who believes the opposite is precisely the sort of conflict that a successful method of ethics is supposed to avoid. The response to Sidgwick’s criticism is as well-known as the criticism itself: It gets the Nazi’s maxim wrong. The response depends upon the assumption that we can identify the principle an agent is really prepared to act upon, or his “sincere” maxim. In the case of the Nazi this would be the more general principle to kill as many of his enemies as possible, whoever they turn out to be, for that is the principle on which he would act if his enemies happened to be someone other than Jews (for example, just Slavs, or Roma, or homosexuals, who were “enemies” of the Nazis and would have been even if Jews had not been); and the claim is then that the Nazi cannot coherently universalize his maxim if that is what it

Freedom and Happiness  149 really is, for willing that others who might have him as their enemy kill as many of their enemies as possible, thus him if they can, would conflict with his maxim of killing as many of his enemies as possible. The Nazi’s attempt to act on his maxim while also universalizing it is simply a formula for all-out war, the paradigm of immorality (as it was).10 I would now suggest that Sidgwick’s failure to anticipate this obvious response is part of his more general failure to appreciate that Kant does have a coherent conception of freedom and its potential as a normative principle for moral philosophy. For we could say that the agent who attempts to act on a non-universalizable maxim is one who either simply ignores the freedom of others, that is, is willing to privilege his own freedom over that of anyone whom he sees as conflicting with it, or else effectively wills to undermine his own freedom, by willing that others should freely adopt a maxim that could destroy his own freedom, as the maxim that everyone should try to kill his enemies would do. Kant’s supposition that there can be no conflict of duties, understood to apply both intrapersonally and interpersonally, is that in every situation there is some course of action or restraint from action available that will preserve the freedom of all involved, thus either one’s present self and one’s future self, or one’s own self and other selves. The willingness to universalize any old maxim violates the maximal preservation of freedom and should be rejected for that reason. (2) Sidgwick’s next objection to Kant is to his view that “a good will is the only absolute and unconditional Good” (ME 222), which Sidgwick interprets as the view that a morally good agent must not be moved by “the emotion of affection, but merely [by] the resolution to benefit” (or, presumably, to fulfill some other class of duty, such as not to harm oneself or others), “which alone has ‘true moral worth.’ ” Sidgwick does not think that he actually has to appeal to any special utilitarian principles to reject this view, for he thinks that “Common Sense” itself will answer the question: whether an act is virtuous in proportion as it was done from regard for duty or virtue . . . in the negative: for the degree in which an act deserves praise as courageous, loyal or patriotic does not seem to be reduced by its being shown that the predominant motive to the act was natural affection and not love of virtue as such. (ME 223) Rather, Sidgwick supposes, a good part of virtue will consist in the cultivation of natural affections to courage, loyalty, patriotism, and so on, and more generally virtue will consist in the settled disposition to such affections and action upon them and not in regard to duty as “the ultimate spring of action” (ME 223–224). Here Sidgwick could argue that Kant’s position is in conflict with both common sense and

150  Paul Guyer with utilitarianism, for insisting that people focus on duty as such will not be as efficacious a source of happiness as cultivating the emotions of affection. The obvious response to Sidgwick’s objection is that he has misconstrued the strategy of the first section of the Groundwork. Kant is not there providing his full account of the psychology or phenomenology of morally worthy action but is conducting a thought-experiment aimed at elucidating a key feature of the fundamental principle of morality, one that Kant even thinks will suffice to identify it. That is, Kant opposes two kinds of stick-figures: one is the agent who is motivated to perform morally correct actions solely by inclination, even of an ordinarily beneficial kind, but who for that reason cannot be counted upon to act correctly should that inclination disappear, as inclinations often do; the other is an agent who is motivated to act solely by regard to duty, and who can therefore do what is right even in the complete absence of inclination (G 4:397–398). Kant infers from the mere conceivability of the latter figure that the principle of duty cannot be derived from inclination or from the objects of inclination—thus from any conception of happiness as the gratification of some sum of inclinations (G 4:418)—but can only concern the form of our maxims, namely their universalizability (G 4:402)— form being the only alternative to matter in Kant’s view, and universality being the only alternative to particularity in maxims that would be patently immoral. But once the moral law has been thus derived from the thought-experiment, the work of that experiment has been done, and Kant can eventually turn to a fuller and more realistic phenomenology of moral motivation, which is that a human being expresses or exercises her commitment to fulfilling her duty by cultivating a variety of “aesthetic preconditions of the mind’s susceptibility to concepts of duty,” or natural affections that can prompt one to morally correct action and can be acted upon in cases where conscience determines that the action so prompted is in fact morally correct (MM 6:399–403).11 Thus, Kant does not merely “relax” the “rigidity” of his conception of virtue or moral worth in order to allow for the cultivation of and action upon such feelings as “Gratitude” as Sidgwick thinks (ME 223) but rather supposes that cultivation of and action upon moral feelings of several kinds—love of others, sympathy, gratitude, self-esteem—are the “means” that nature has afforded us “to promoting active and rational benevolence” or other duties (MM 6:456), that is, the means through which we naturally and normally express and exercise our underlying regard to duty. Now we could suggest that Sidgwick has exaggerated the difference between Kant on the one hand and both common sense and himself on the other just because of his overemphasis on the Groundwork and corresponding neglect of a late section in the second half of the late work, the Metaphysics of Morals, a work to which Sidgwick does in fact refer several times (e.g., ME 223, 386, 389). But I think his failure to

Freedom and Happiness  151 appreciate Kant’s real position on the relation between moral feeling, regard for duty, and moral worth lies in his general view that Kant’s position on freedom is a morass of confusion—his easy rejection of Kant’s treatment of freedom of the will as a tissue of confusion hurts him here. For to appreciate Kant’s actual position, one has to be ready to accept Kant’s position that a morally worthy regard for duty can be expressed through the cultivation of and action upon moral feelings, which Kant himself models as a relation between the noumenal and the phenomenal. The underlying regard to duty, or as Kant puts it in Religion within the Boundaries of Mere Reason, the choice of the fundamental maxim to subordinate self-love to morality (RBMR 6:35–36), is conceived of as a noumenally free choice, while the cultivation of moral feelings and actions prompted by them is conceived of as something taking place at the phenomenal level, grounded in some way in what has transpired at the noumenal level but not literally caused by a choice at that level for the simple reason that causation is a relation between two temporally successive states of affairs, and choices at the noumenal level are supposedly not temporal and therefore do not have determinate positions in time. Of course, Sidgwick is not friendly to the metaphysics of transcendental idealism, as is clear from his Lectures on the Philosophy of Kant (1905, II and III); indeed, no one is friendly to Kant’s transcendental idealism unless they water it down into a distinction between two “viewpoints” or ways of conceiving of objects.12 But one does not have to take all of Kant’s transcendental idealist model of free will on board in order to appreciate this conception of the relation between underlying commitment to principle and cultivation and action on moral feelings. However one understands responsibility for choice of fundamental principles, one can still hold that one’s fundamental moral commitments are not expressed by an immediate causation of action by principles but through the cultivation of feelings that can then be allowed and counted upon to prompt actions in appropriate circumstances. (3) Sidgwick’s discomfort with Kant’s conception of freedom also underlies his objection to Kant’s doctrine that the perfection of others, specifically the perfection of their virtue, cannot be a duty for me, leaving only their happiness as an end that is also a duty regarding others. This is half of the basis for Kant’s argument that the two ends that are also duties are my own perfection and the happiness of others, the other half of the argument being that my own happiness cannot be a duty for me because I naturally desire it and never have to constrain myself to pursue it by regard for duty (MM Introduction, sections IV and V.B)—a lame argument, to be sure, because of course my immediate desires can get in the way of my interest in my own long-term satisfaction (ME 36). But Sidgwick’s concern is with the other half of the argument, based on Kant’s assumption that one person cannot promote another’s self-perfection. What Kant has in mind is moral self-perfection, and his

152  Paul Guyer assumption is that one’s moral self-perfection can only ever be the product of one’s own free choice to make the moral law one’s fundamental maxim. Sidgwick recognizes that Kant’s argument is that “my neighbor’s Virtue or Perfection cannot be an end to me, because it depends upon the free exercise of his own volition, which I cannot help or hinder.” Yet he takes it to be “undeniable that we can cultivate virtue in others: and indeed, such cultivation is clearly the object not only of education, but of a large part of social action, especially of our expression of praise and blame” (ME 240). Kant certainly does suppose that moral education is a central part of education. But a careful reading of his treatment of moral education in both his critical works and his lectures on pedagogy shows that in his view it should consist above all in demonstration by example to the pupils that they really do have the freedom to live up to the moral law, with such examples making their freedom more graphic to them than the philosopher’s pronouncement that “ought implies can” ever can. Moral education consists in putting before children examples of honest people, to whom they can compare themselves, who choose to do what is right in spite of promises of great rewards for not doing so or threats of punishment for doing so. The example of such people, such as the unnamed man who refused to calumniate Anne Boleyn in spite of Henry VIII’s threats (Kant is thought to refer to the courtier Henry Norris, who was accused of adultery with her but insisted upon her innocence and paid for this with his life), brings home to the learners the “power over the human heart” that a pure morality can have; that is, the fact of their own freedom (PracR 5:156).13 The freedom, the fact of which is thus brought home, however, is the freedom of each to make one’s own choices, including the freedom to make the moral law one’s fundamental maxim; in Kant’s view, moral education reinforces the individual’s recognition of her own freedom rather than in any way diminishing it. More generally, Kant holds that natural processes can develop abilities of the phenomenal self that can be used for good purposes if the individual so chooses but cannot by themselves substitute for the free choice of right over wrong by the noumenal self; that is why “discipline” can be an “ultimate end” of nature in human development but the freedom to use that discipline in behalf of the good remains “a supersensible faculty” (CPJ 5:432). So for Kant there remains a dividing line between everything that anyone can do to help anyone else use their freedom correctly and the choice of the other to so use their freedom. Once again, Sidgwick rejects Kant’s noumenal account of freedom of the will, and there is no such barrier for him; virtue must be a phenomenal characteristic of phenomenal selves achieved by natural causes because there are no others, and there can be no rigid separation between what one person can do to contribute to another’s virtue and what the other must do. In the end,

Freedom and Happiness  153 it must be conceded that if we do drop Kant’s transcendental idealism, which guarantees that everyone always has free will but also that free will exists at a noumenal level impervious to phenomenal factors, then whether and how much one person can contribute to the development of virtue in others must become empirical questions. And for Sidgwick, the empirical answer to such questions is obvious: of course people can help each other cultivate their own virtue, and should do so, not merely provide each other with examples of virtue. (4) The previous two issues turn on Sidgwick’s rejection of Kant’s approach to freedom of the will. But the last and most important of his objections to Kant that I will consider turns instead on his rejection of Kant’s view of freedom of choice as the foundational value on which all of morality can be built. More precisely, the interpretation that Sidgwick gives to Kant’s foundational conception of humanity as the sole unconditional end in itself shows that he does not recognize that this is an expression of the idea that freedom itself is the foundational value of morality, and the objections that Sidgwick makes to Kant’s use of this idea turn on the interpretation that he places upon it.14 The crucial encounter with Kant comes in the chapter on “Philosophical Intuitionism,” the penultimate chapter of Book III of The Methods of Ethics. Here Sidgwick offers a brief review of the history of “philosophical intuitionism” including Socrates, Plato and Aristotle, the Stoics, the English rationalist Samuel Clarke, and finally Kant. His discussion of Kant begins with the remark that Among later moralists, Kant is especially noted for his rigour in separating the purely rational element of the moral code: and his ethical view also appears to me to coincide to a considerable extent, if not completely, with that set forth in the preceding section. (ME 385) In particular, Sidgwick finds agreement from Kant concerning the foundational principles of Justice and Rational Benevolence. Regarding the later, Sidgwick attributes to Kant the view that “the only real ultimate” and positive end that we have with regard to others “is the object of Rational Benevolence as commonly conceived—the happiness of other men.” He claims that “Kant’s conclusion appears to agree to a great extent with the view of the duty of Rational Benevolence that I have given:—although I am not altogether able to assent to the arguments by which Kant arrives at his conclusion” (ME 386). He considers two arguments that Kant offers for this duty and finds them both wanting. The first argument is that which Kant states in his illustration of the formula of universal law in the Groundwork and repeats in the Doctrine of Virtue of the Metaphysics of Morals, namely, as Sidgwick puts it, “We . . . necessarily constitute ourselves an end for others, and claim

154  Paul Guyer that they shall contribute to our happiness; and so, according to Kant’s fundamental principle,” namely, the requirement that moral maxims be universalizable, “we must recognize the duty of making their happiness our end” (ME 389). In Kant’s terms, the will of someone who decided to universalize the maxim of not helping others in need when he could, would conflict with itself, since many cases could occur in which one would need the love and sympathy of others and in which, by such a law of nature arisen from his own will, he would rob himself of all hope of the assistance he wishes for himself. (G 4:423) Sidgwick’s objection to this is that the assumption “that every man in need wishes for the aid of others is an empirical proposition which Kant cannot know a priori,” for “We can certainly conceive of a man in whom the spirit of independence and the distaste for incurring obligations would be so strong that he would choose to endure any privations rather than receive aid from others.” Here we could counter that, as Kant makes clear in the Metaphysics of Morals, while the fundamental principle of morality, the derivation of which Kant calls “metaphysics of morals” in the Groundwork (see G 4:390), must be universal and necessary, valid for all rational beings, and thus knowable entirely a priori, the derivation of particular duties for human beings, what he calls “metaphysics of morals” in the work of that title, is not and cannot be entirely independent of empirically known facts about human nature— a metaphysics of morals cannot dispense with principles of application, and we shall often have to take as our object the particular nature of human beings, which is cognized only by experience, in order to show in it what can be inferred from universal moral principles. (MM 6:217) And this means that the mere fact that we can conceive of a person who is always willing to forego the assistance of others in the achievement of his own ends and the attainment of the happiness that this constitutes is of no real significance in a “metaphysics of morals” as a complete moral theory; the empirical fact that no one can reasonably count on always being able to achieve all his ends by his own means and on remaining content with that restriction is properly used in deriving the duties of human beings from the universal moral principles valid for all rational beings, even if some overly self-confident individual fails to realize this. The problem for this confident individual might not be exactly the same as that for the Nazi, who by universalizing his own maxim to kill his

Freedom and Happiness  155 enemies whenever he can wills his own destruction but is rather that by universalizing a maxim of non-assistance he misunderstands the real conditions of human willing—that no one can ever count on achieving their goals entirely by their own efforts—and thereby contradicts the possibility of his own rational willing under all humanly possible circumstances. And were it to be countered that for some independence from assistance by others is simply more important than any other particular goal, Kant would rightly reply that this is an unrealistic picture of human nature, which is properly considered in the derivation of duties. We might also say that no one who takes proper account of the point of view of others could reasonably impute to all others such an unconcern or distaste for assistance even if that trait were powerful in himself. He would have no reason to think that all others resembled him in this regard, and further could coherently adopt a maxim of helping others who want help when he can even if he reasonably thought there were some others who would never want help—and this would be a perfectly reasonable restriction on his maxim of beneficence, for since no one is ever going to be able to help everyone else who might need help, he is going to have to find some principles by which to limit his beneficence. But this brings us to the question of what is involved in properly treating others as ends in themselves and never merely as means to one’s own ends, and this is the issue in Sidgwick’s objection to Kant’s second, “apparently different line of argument” for “Rational Benevolence.” Here Sidgwick has in mind Kant’s argument for and application of the principle that humanity is always to be treated as an end in itself, never merely as a means. As Sidgwick puts it, Kant lays down that, as all action of rational beings is done for some end, there must be some absolute end, corresponding to the absolute rule before given, that imposes on our maxims the form of universal law. This absolute end, prescribed by Reason necessarily and a priori for all rational beings as such, can be nothing but Reason itself, or the Universe of Rationals; for what the rule inculcates is, in fact, that we should at as rational units in a universe of rational beings. (ME 389–390) ‍ hat is, treat others as “rational units” just as much as I treat myself T as such. “Now, says Kant,” Sidgwick continues, “as long as I confine myself to mere non-interference with others, I do not positively make humanity my end” (see G 4:430); rather, “Virtue is exhibited and consists in the effort to realise the end of Reason in opposition to mere selfish impulses.” Sidgwick’s objection to Kant then turns on his interpretation of “Reason itself” as the absolute end of all moral actions: it is not clear what it would be to make this an end of action, but it is clear that such an end would have nothing to do with ordinary desires and therefore with

156  Paul Guyer happiness, the fulfillment of desires, the end of the duty of benevolence as ordinarily conceived, and of course as conceived in utilitarianism. In Sidgwick’s words, The conception of “humanity as an end in itself” is perplexing: because by an End we commonly mean something to be realised, whereas “humanity” is, as Kant says, “a self-subsistent end”: moreover, there seems to be a sort of paralogism in the deduction of the principle of Benevolence by means of this conception. For the humanity which Kant maintains to be an end in itself is Man (or the aggregate of men) in so far as rational. But the subjective ends of other men, which Benevolence directs us to take as our own ends, would seem, according to Kant’s own view, to depend upon and correspond to their non-rational impulses—their empirical desires and aversions. It is hard to see why, if man as a rational being is an absolute end to other rational beings, they must therefore adopt his subjective aims as determined by his non-rational impulses. (ME 390) ‍ e can set aside Sidgwick’s worry that as a “self-subsistent end” reason W is not even the sort of thing that can be promoted by actions; if, as on his own account, virtue is the sort of thing that one person can cultivate in others, then it would seem as if reason or rationality would be too, precisely the sort of thing that can be promoted or “realised” through education.15 Let us focus instead on the “paralogism” that Sidgwick insists undercuts Kant’s argument from “Reason itself” as the absolute end of morality to “Benevolence.” This objection assumes that reason has nothing to do with the satisfaction of “empirical desires and aversions” but that the happiness that is the object of benevolence consists in the satisfaction of such desires and aversions, that is, that benevolence consist in helping others satisfy their desires and aversions to the extent that one can. This diagnosis depends on interpreting “humanity as an end in itself” as pure rationality and on taking such rationality to have nothing to do with empirical desires and aversions. However, Kant himself suggests a very different interpretation of the idea of humanity as an end in itself. Although Kant sometimes suggests that it is only “humanity insofar as it is capable of morality” that gives a human being “dignity” and moral standing (G 4:435), a principle to make humanity in this narrow sense the end of all our actions would certainly not give rise to the whole range of moral duties that Kant recognizes; so in the Doctrine of Virtue of the Metaphysics of Morals, he defines humanity more generally as that “by which . . . alone [one] is capable of setting himself ends” (MM 6:387), or as “the capacity to set oneself an end—any end whatsoever” (MM 6:392). This does not suggest that humanity has nothing to do with particular ends but rather that

Freedom and Happiness  157 it is precisely the capacity to set oneself particular ends. If ends can be set on the basis of empirical desires or aversions, or not set without some such desires or aversions—they must also, of course, be set in compliance with the constraints of morality—then it is connected rather than disconnected to such desires and aversions. Insofar as happiness consists of the satisfaction of desires and aversions, then humanity is connected to that too. To be sure, setting an end is not just having an empirical desire or aversion, but results from using reason to make the satisfaction of a desire or aversion into the object or goal of an action or maxim of actions by judging that satisfaction of such an empirical inclination is consistent with the fundamental principle of morality and the general system of duties that follows from it under the circumstances of human existence or is even required by this. This is why Kant describes the “realm of ends,” the effort to realize which is prescribed by his final formulation of the categorical imperative in the Groundwork, as “a whole of all ends in systematic connection (a whole both of rational beings as ends in themselves and of the ends of his own that each may set for himself)” (G 4:433). What it is to treat a rational being as an end in itself is to treat the particular ends that it sets for itself as also ends for oneself, as long as that is consistent with treating everyone as an end in him- or herself, and these particular ends that each sets for him- or herself originate with natural desires and aversions, although those do not become ends until they pass through the screens of both prudential and pure practical reason. Several things follow from the assumption that humanity really is the capacity to transform empirical desires and aversions into ends through the exercise of reason. One will be that reason’s own requirement of universalizability—that like cases be treated alike—be applied, and thus that in setting one’s own ends one must take account of others’ ends as well. Perhaps this will be enough to get the duty of “Rational Benevolence” off the ground. But if it is not enough by itself, perhaps it will be when supplemented with a further canon of practical rationality, the complement of what Kant makes explicit as the principle of hypothetical imperatives: If one wills the end, one must also will some adequate means to it (G 4:417); the complement of this is that one cannot rationally set an end for oneself if one does not reasonably believe oneself to have some adequate means to realize it. This then suggests that if one is to make the humanity of all and not just of oneself an end, one must be prepared to make means available to others, when and insofar as one can, for the realization of their ends, that is, to make it possible for them reasonably to set those ends for themselves. Making means available to others (and for that matter oneself) expands the range of ends that they can rationally set for themselves, and thereby expands their freedom, or the possibility for the rational use of their freedom. At the same time, it makes available means to their happiness, although not because happiness is itself the foundational value of morality, or in more Kantian language the direct

158  Paul Guyer object of morality. All human beings getting to rationally set their own ends to the greatest extent possible compatibly with everyone else also getting to set their own ends to the greatest extent possible is the immediate object of morality, but happiness is the natural outcome of being able to set one’s ends with adequate means to realize them—the availability of such means not being required by a separate imperative that rational beings should be able to realize their ends, but by their being able to rationally set ends in the first place—and the greatest happiness possible in the world would thus be the natural outcome of the greatest freedom possible in the world. Another way of putting this point is to say that humanity consists in the freedom of human beings to set their own ends, but that this freedom is restricted when people are deprived of means to ends they could otherwise rationally set for themselves, and conversely that this freedom is promoted and extended when means are afforded to people that allow them to set ends for themselves that they otherwise could not. This is the way in which taking freedom, in the sense of the ability to set and pursue ends rationally, can function as a foundational value for the derivation of the duties, in particular the duty of benevolence. The connection to happiness would not be explicit in the concept of humanity but would follow from the assumption that happiness consists simply in the realization of ends or follows from it. The response to Sidgwick would thus be that a proper understanding of Kant’s conception of humanity as not just “Reason itself” but as freedom as the capacity to set ends for oneself in a rational way does, when subjected to the requirement of universalizability, give rise to the duty of “Rational Benevolence.” Of course it would take more than I have said here to show that this derivation of the duty of benevolence gives rise to no more conflicts than does Sidgwick’s utilitarian goal of the greatest aggregate happiness, or conversely that Sidgwick’s principle of happiness gives rise to no fewer conflicts than Kant’s principle that humanity as the capacity of each to set and pursue his own ends must always treated as an end and never merely as a means in one’s own person and that of every other. Kant himself insists that there can be no genuine “collisions” between duties, only between grounds of duties, because in any such case there is always in fact one right thing to do—but he hardly spells out an argument for this assertion (see MM 6:224). All I will suggest here on this issue is that one way in which Kant reduces conflicts among duties is by suggesting that perfect duties not to destroy others or their freedom always take priority over promoting or enhancing the scope of their agency, that is, satisfaction of perfect duties not to harm takes priority over imperfect duties to benefit; and the same sort of principle is surely built into Sidgwick’s method of ethics, which recognizes the principle of justice before it recognizes the principle of rational benevolence, and presumably prioritizes the former over the latter. To be sure, that leaves open

Freedom and Happiness  159 that even when the satisfaction of perfect duties such as justice is presupposed, there may still be competing demands for anyone’s benevolence that cannot be mechanically resolved, for example, between benevolence toward one’s parents on the one hand or one’s children on the other, or between benevolence toward the small number of one’s own children and the larger number of the children of some community beyond one’s own family. It is not clear that Kant’s derivation of the duty of benevolence provides a mechanical method for the resolution of such conflicts. But I think it is safe to say that Sidgwick would have had to have gone a lot further in the exposition of his utilitarianism to guarantee that his method could resolve such conflicts any better than Kant’s. To conclude, I have argued that the fundamental difference between Sidgwick and Kant is not over their conceptions or applications of the principle of “rational benevolence” but over how they derive it: For Sidgwick, this principle is an intuition on which, he concedes, his utilitarianism ultimately depends, while for Kant it is a consequence of his conception of humanity as the freedom to set one’s own ends as the ultimate basis of morality, the “ground of a possible categorical imperative” as he says in the Groundwork (4:428). But Kant’s account of freedom as the normative basis of morality also entails the deontological constraint that the happiness of all cannot be maximized at the cost of freedom for anyone, or that justice is always prior to benevolence. For Kant, the greatest happiness possible in a moral world is the greatest happiness that is consistent with justice. Whether this is a better argument than Sidgwick has for the constraint of benevolence by justice would be the matter for another essay.

Notes 1. Kant’s works are abbreviated herein as follows: PracR = Critique of Practical Reason, G = The Groundwork of the Metaphysics of Morals, MM = The Metaphysics of Morals (Kant 1996a); CJ = Critique of the Power of Judgement (Kant 2000); RBMR = Religion within the Boundaries of Mere Reason (Kant 1996b); one passage from Kant’s Critique of Pure Reason will be located by the pagination of its 1st (“A”) and 2nd (“B”) editions, included in all modern translations of the work, such as that by Paul Guyer and Allen W. Wood (Kant 1998). 2. On the relevance of Sidgwick’s discussion of freedom of the will to the larger argument of the Methods, see Schneewind (1977, 207–212). 3. de Lazari-Radek and Singer (2014, 90–92) discuss these four criteria under the rubric “Sidgwick on Justifying Ethical Principles.” I would think that the role of these criteria is rather to determine whether a proposed method of ethics is even a candidate for success, leaving the further question of the justification of any method of ethics that might pass this test open. This might better fit the structure of Sidgwick’s overall argument as it evolves through the several editions of his work. In the earlier editions, the argument is that only utilitarianism can pass the test provided by the criteria, so no further question of justification need be broached; in the later editions, Sidgwick

160  Paul Guyer acknowledges that utilitarianism itself must rest on some fundamental justificatory intuition; see (ME xx-xxi). See, also, Schneewind (1977, 286–289). 4. On the universality of any genuine practical law, see PracR 5:19 and MM 6:389. On the impossibility of a conflict of duties, see MM 6:224. 5. See de Lazari-Radek and Singer, Point of View, pp. 119–120. 6. The “formula of universal law” is a common designation for the categorical imperative as Kant formulates it at G 4:402 and 421 and PracR 5:30. The “formula of the universal law” of nature is the common name for the second version of the categorical imperative that Kant provides at G 4:421, where he states that the test for whether one could will that a maxim on which one proposes to act also be universalized is whether one could will it to be a law of nature, i.e., will that a world in which it is a law exist. For the classical discussion of Kant’s formulations of the categorical imperative and the relations among them, see Paton (1947, ch. 13). For my position, see Guyer (1995, 353–385). 7. See Critique of Pure Reason, A 151/B 190. 8. This is a criticism that has been made many times since. See, for example, Singer (1961). For early and still relevant discussion, see O’Neill (1975, chs. 2, 5, with further references at p. 60). 9. For an illuminating discussion of the uniqueness of the Nazi attack upon the Jews within their larger pattern of genocide and within instances of genocide more generally, see Evans (2015, 365–389). 10. This interpretation of the violation of the formula of universal law as a contradiction between the attempt to act upon a proposed maxim and the “universal generalization” of the maxim was diagnosed by O’Neill (1975, 69–71) and adopted by John Rawls (2000), Christine Korsgaard (1996, 77–105), and many others. O’Neill (2013, 13–16) distinguishes between an agent’s conscious intentions and the maxims that may properly be ascribed to them in the Introduction to the 2nd edition of Acting on Principle. On the question of what maxim to properly assign an agent, see, also, Herman (1993, 132–158). 11. For my extended argument for this interpretation, see Guyer (2010, 130–151). 12. Notably Bird (1962, 2006) and Allison (2004). 13. For fuller discussion, see Guyer (2011, 124–138). 14. In a brief but suggestive paper, Onora O’Neill argued, “The lingering sadness of so many passages in The Methods of Ethics reflects Sidgwick’s dispassionate and stoical refusal to assert claims that cannot be supported by reasons and his view that practical reason supplies no more than universal generalisations and instrumental rationality,” thus that “Although his discussion of practical reason is often conducted in Kant’s terminology, it is profoundly unlike Kant’s” (O’Neill 2001, 88). I do not want to paint the difference between Sidgwick and Kant so starkly but do want to argue that Sidgwick does not recognize that Kant’s application of canons of practical reason presupposes an antecedent conception of the foundational value of freedom. My suggestion in what follows is that it is the element of the freedom to choose one’s own ends for action subject to the rest of the constraints of practical reason, which is to say precisely “universal generalisation and instrumental rationality.” For an argument that Kant does not actually begin with an unexplained premise of the value of freedom but from an application of the most general canons of reason as such, beginning with the law of noncontradiction, to the fact of human freedom, see Guyer (2019). 15. This is a quick argument that reason or rationality is the kind of thing that can be adopted as an end to be promoted, not merely acted against; to be

Freedom and Happiness  161 sure, Kant does not suggest that anyone has a duty to produce rationals or rational beings in the sense of procreating. He does not even argue that marriage is morally undertaken only with the goal of procreation, although he does argue that sexual intercourse is moral only within marriage.

References Allison, H. E. 2004. Kant’s Transcendental Idealism: An Interpretation and Defense. Revised ed. New Haven: Yale University Press. Bird, G. 1962. Kant’s Theory of Knowledge. London: Routledge and Kegan Paul. Bird, G. 2006. The Revolutionary Kant: A Commentary on the Critique of Pure Reason. Chicago and LaSalle: Open Court. Crisp, R. 2015. The Cosmos of Duty: Henry Sidgwick’s Methods of Ethics. Oxford: Oxford University Press. de Lazari-Radek, K., and P. Singer. 2014. The Point of View of the Universe: Sidgwick and Contemporary Ethics. Oxford: Oxford University Press. Evans, R. J. 2015. The Third Reich in History and Memory. New York: Oxford University Press. Guyer, P. 1995. “The Possibility of the Categorical Imperative.” Philosophical Review 104: 353–385. Guyer, P. 2010. “Moral Feelings in the Metaphysics of Morals.” In Kant’s Metaphysics of Morals: A Critical Guide, edited by L. Denis, 130–151. Cambridge: Cambridge University Press. Guyer, P. 2011. “Examples of Moral Possibility.” In Kant and Education: Interpretations and Commentary, edited by K. Roth and C. W. Surprenant, 124– 138. London: Routledge. Guyer, P. 2019. Kant on the Rationality of Morality. Cambridge, Cambridge University Press. Herman, B. 1993. The Practice of Moral Judgment. Cambridge, MA: Harvard University Press. Kant, I. 1996a. Practical Philosophy. Edited and translated by M. Gregor. Cambridge: Cambridge University Press. Kant, I. 1996b. Religion and Rational Theology. Edited and translated by A. W. Wood and G. di Giovanni. Cambridge: Cambridge University Press. Kant, I. 1998. Critique of Pure Reason. Edited and translated by P. Guyer and A. W. Wood. Cambridge: Cambridge University Press. Kant, I. 2000. Critique of the Power of Judgment. Edited by P. Guyer, translated by P. Guyer and E. Mathews. Cambridge: Cambridge University Press. Korsgaard, C. 1996. Creating the Kingdom of Ends. Cambridge: Cambridge University Press. O’Neill, O. 1975. Acting on Principle: An Essay on Kantian Ethics. 1st ed. New York: Columbia University Press. O’Neill, O. 2001. “Sidgwick on Practical Reason.” In Henry Sidgwick, edited by R. Harrison. Oxford: Oxford University Press. O’Neill, O. 2013. Acting on Principle: An Essay on Kantian Ethics. 2nd ed. New York: Cambridge University Press. Paton, H. J. 1947. The Categorical Imperative: A Study in Kant’s Moral Philosophy. London: Hutchinson.

162  Paul Guyer Rawls, J. 2000. Lectures on the History of Moral Philosophy. Cambridge, MA: Harvard University Press. Schneewind, J. 1977. Sidgwick’s Ethics and Victorian Moral Philosophy. Oxford: Clarendon Press. Sidgwick, H. 1896. Outlines of the History of Ethics. 4th ed. London: Macmillan. Sidgwick, H. 1905. Lectures on the Philosophy of Kant and Other Philosophical Lectures and Essays. London: Macmillan. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Singer, M. 1961. Generalization in Ethics: An Essay in the Logic of Ethics with the Rudiments of a System of Moral Philosophy. New York: Alfred A. Knopf. Wood, A. W. 2006. “The Supreme Principle of Morality.” In The Cambridge Companion to Kant and Modern Philosophy, edited by P. Guyer. Cambridge: Cambridge University Press.

8 Kant and Sidgwick on Freedom of Will, Morality, and Responsibility Mariko Nakano-Okuno

Henry Sidgwick is known as a pioneer of contemporary analytic moral philosophy and supporter of utilitarian ethics.1 One of the factors that distinguishes Sidgwick from other classical utilitarians is his level of engagement with Kant’s central ideas and his incorporation of some of them into his theory of moral philosophy. Although Sidgwick’s happiness-centered, consequentialist ethics and Kant’s respect-based deontological ethics are often considered to be at odds with each other, I have argued elsewhere (Nakano-Okuno 2007) that their ethical theories are based on essentially the same fundamental principles. Among these principles are (1) the universalizability of moral judgment, which is presented as the axiom of justice in Sidgwick and the universal-law formula of Categorical Imperative in Kant, and (2) the equal treatment of each individual’s good, which is presented as the axiom of rational benevolence in Sidgwick and the formula of humanity or respect for persons in Kant.2 Even their claims about the pursuit of happiness are not incompatible, as Sidgwick contends that the utilitarian obligation to pursue people’s happiness derives not from our natural inclination to do so, but from our sense of duty to follow the fundamental moral axioms that our rationality dictates through our philosophical intuitions. Despite these similarities, there remains one critical difference between Sidgwick’s and Kant’s ethics—their respective views on freedom of will and its practical importance. While Kant believes that morality does not make sense without the notion of free will, Sidgwick questions if it is epistemically necessary and practically meaningful for us to assume this notion in order to act morally. This point of divergence between Kant and Sidgwick is the focus of this chapter. The central thesis of this chapter is that, whereas part of Sidgwick’s criticism can be ascribed to his misconstruing of Kant’s argument, Sidgwick is probably right in demystifying the notion of free will. If Kant drops his claim on the practical importance of freedom, then there will be no insurmountable theoretical discrepancies between Sidgwick and Kant’s moral philosophy.

164  Mariko Nakano-Okuno The discussion of this chapter will proceed as follows.3 In the first section, after contrasting and examining the mostly overlapping yet critically different notions of “will” in Kant and Sidgwick, we will confirm what Kant means by freedom of will in the context of morality and moral agency. We will also attest how this notion of freedom relates to the concept of moral responsibility in Kant, and how freedom is fundamental to his moral philosophy. Then, in the second section, we will focus on Sidgwick’s critique of Kant’s views on freedom. I will argue that, although Sidgwick’s major claim that Kant confused two senses of freedom is mostly unwarranted, Sidgwick is acutely on target when he criticizes Kant’s overemphasis on the necessity and practical importance of free will, and that Kant has no unproblematic replies available. One crucial point made throughout this chapter is that, whereas Kant seems to assume that the will of rational beings is, by his definition, always correct—more precisely, that our reason can correctly grasp a universal moral law that every other rational being would recognize as valid (even though we humans may not act upon it due to our human limitations), Sidgwick observes that our will or reason itself is fallible; a person may fanatically believe that he is autonomously setting a universal moral law and acting out of respect for that law, but the law he has set may turn out to be wrong in the eyes of other rational beings. According to Sidgwick, this wrongness of his moral judgment must be judged not by his will’s independence from natural impulses or natural causality, or by his will’s ability to set universal laws, but by the harmful consequences his willful act brings to other people, whether or not he and these other people are rational and free in Kant’s sense. This critical difference in Kant and Sidgwick led the former to rely on the freedom of will as an indispensable basis of our morality, while the latter denounced the theoretical and practical importance of free will in moral philosophy. This chapter will argue that Sidgwick’s stance is more plausible both in theory and in practice. A few remarks need to be made before we delve into the main discussion. This chapter focuses on the issue of freedom insofar as it relates to ethics and morality and examines whether and in what sense the notion of freedom is important for ethics. “Ethics” here means a field of study that deals with the questions of what ought to be done, or what is the right thing to be done, by an individual who is considered to have his or her own volition, or will.4 As such, ethics inevitably involves an individual’s will, moral judgment, or belief about what ought to be done. It is also commonly believed that ethics in this sense concerns not only an individual’s private beliefs or judgments about the rightness of acts but also the “universal” or “objective” rightness of acts in that the same course of action should be deemed appropriate by other people as well. As I understand it, both Sidgwick and Kant discussed the role of an individual’s will in ethics in this sense.

Kant and Sidgwick on Freedom of Will  165

The Idea of Freedom Freedom is a notion widely used to denote the absence of coercion. However, the precise meaning of “freedom” varies. When one discusses freedom of speech or freedom of expression, or when J. S. Mill advocates freedom in his On Liberty, “freedom” means the absence of explicit or implicit coercion or restraints on one’s actions (such as speeches and expressions) by other individuals, authorities, or organizations. This notion of freedom is different from what Kant considers a central notion of ethics and moral philosophy. What Kant has in mind is freedom of will in the sense that it is not constrained by any natural or sensible causes (including one’s own inclinations) external to one’s faculty of reason. This Kantian notion of freedom is called transcendental freedom, which is explained as “independence from everything empirical, and, consequently, from nature generally, whether it is an object of the internal sense considered in time only, or of the external in time and space” (KpV 5:96–97).5 Impulse, Choice, and Will The freedom that Kant considers when he discusses morality is the freedom of will (Wille) or choice (Willkür). To understand the meaning of freedom central to Kant’s ethics, we need to first understand the relationship between will, choice, and the faculty of reason. One point to note is that both Kant and Sidgwick distinguish between will and choice, or between will and motive, the former of which is tied to a moral judgment, while the latter can vacillate between various stimuli including sensible impulses as well as moral judgments. According to Kant, will and choice are sub-categories of object-seeking impulses.6 The appetitive faculty (Begehrungsvermögen),7 or what I would call the faculty of object-seeking impulse, refers to one’s faculty to make itself, by means of one’s mental representation, the cause of the realization of the object that is being represented in mind (MS 6:211). Even when one does not have an impulse toward a specific object, one may have a general lust (das Gelüsten) after an unspecified something as a kind of pro-attitude. This lust tends to awake an object-seeking impulse when being triggered by the representation of a specific object (see MS 6:213). Thus, when one (1) represents a certain object in mind and (2) seeks to realize it, one has an appetitive impulse, i.e., an objectseeking impulse. When its determining ground toward action originates from the agent herself—in other words, when the very agent (and not any other being or object) spontaneously opts whether to act or not to attain that object, this object-seeking impulse constitutes a power to opt to act or not at its discretion, which can be termed a spontaneous actopting power. When the agent is aware that she can bring about the

166  Mariko Nakano-Okuno object by performing that act—in other words, when the agent’s spontaneous act-opting power is recognized as actionable —, this spontaneous and actionable act-opting power is called choice (Willkür). If the subject does not think or feel that her act can bring about the object, her spontaneous yet non-actionable act-opting power is merely called a wish (see MS 6:211–212). Choice, or the spontaneous actionable act-opting power, can be driven by one’s sensible impulses, determined by one’s faculty of reason, or influenced by the combination of both. When one anticipates a certain pleasure (Lust) and forms an impulse to attain a certain object because of the pleasure that the targeted object brings to him or her, one has a desire (Begierde), which can be understood as a pleasure-based object-seeking impulse. Pleasure is not clearly defined in Kant’s writings—except for his highly abstract definition of pleasure as “the idea of the agreement of the object or the action with the subjective conditions of life” (KpV 5:9). Kant says that pleasure and displeasure are too simple notions to explain further because these are merely subjective feelings, only having relation to the mind of the person feeling it and having no fixed relation to a specific object. To put it simply, we cannot define what pleasure is because different people get pleasure from different objects. But we can at least say that pleasure or displeasure is a purely subjective positive or negative feeling that is, in one’s mind, attached to a certain object. A habitual desire is called inclination (Neigung). On the other hand, when one’s object-seeking impulse is grounded in and determined by one’s reason, one is said to have a will (Wille), which can be understood as a reasonbased object-seeking impulse. Unlike desire and inclination, the will is an appetitive faculty whose internal determining ground lies in reason, not in pleasure (see MS 6:212–213). The description shown here, however, does not yet explain the relationship of one’s reason to one’s will, or how one’s reason “determines” one’s object-seeking impulse. In Kant, the will is also conceived as “a power of determining oneself to action in accordance with the idea of certain laws” (Gr 4:427). To “determine” a will is to determine what sort of object-seeking impulse one ought to have in a way in which the subject can clearly distinguish his commitment to such an act-to-be from all other impulses within his mind. The subject’s capacity for choice (spontaneous and actionable act-opting power) generates a maxim, while his will (a reason-based object-seeking impulse) accompanies the representation of a universal law (MS 6:226). Maxims (Maximen) refer to those practical principles whose condition to determine one’s will is regarded by the subject as valid only for his own will. On the other hand, practical laws (praktische Gesetze) refer to those practical principles whose condition is recognized as valid for the will of every rational being (KpV 5:19). A maxim is a principle on which the subject acts, while a practical law is a principle on which the subject ought to act, and this latter is perceived

Kant and Sidgwick on Freedom of Will  167 as an imperative for those beings who have impulses and inclinations that go against it (Gr 4:421). Kant believes that it is not one’s mere desire or impulse but one’s faculty of reason that determines one’s will under a universal law. The impulsive power of will does not directly motivate one to perform a particular action, but it urges one to internally adopt a universal moral law that dictates one to act in a certain way. (MS 6:213 and 6:226) Thus, a will is a reason-based object-seeking impulse which urges one not directly to perform a particular act but to internally adopt a moral judgment that takes the form of universal law on what ought to be done. To be noted at this point is that will (Wille) and choice (Willkür) are not the same things in Kant. A person may spontaneously opt to perform a wrong act even if her reason has formed a will to adopt a moral law that dictates what ought to be done. She can be fully aware of, and to some extent prompted to act out of, a universal moral law about what ought to be done in the situation, and yet still behave differently. As indicated earlier, choice, or spontaneous actionable act-opting power, can be influenced both by sensible impulses and the faculty of reason. Human choice often oscillates between its incentives—between sensible and moral-judgment-based impulses—as to what to do. At times, one may end up choosing to follow one’s sensible impulse. At other times, one may choose to carry out what one honestly judges to be the right thing. Choice is not determined purely by reason, while will is always determined by reason (MS 6:213–214). Interestingly, when we read Sidgwick’s explanations of impulses, desire, and will, we can find notions quite similar to Kant’s notions of appetitive impulses and will.8 One key difference is that will in Sidgwick is not always considered entirely rational or right. First, in the Methods, we can find descriptions of an object-seeking impulse, or what Sidgwick calls “desire,” which is defined as a felt impulse or stimulus to an action that tends to realize what is desired (ME 43 fn.2; Nakano-Okuno 2011, 114). We can rephrase it as a conscious impulse that aims at a certain object or at an act to obtain it. As we saw in Kant’s explanation of the concept of lust, Sidgwick suggests that we also have vaguer impulses such as instinctive or subconscious impulses when we are not conscious of a targeted object, or of our act to realize the object (ME 24). However, when we represent the targeted object in our mind and consciously seek it, we have a desire in Sidgwick’s term and an appetitive impulse in Kant’s term, both of which mean an object-seeking impulse. Second, when this object is expected to be obtained as a consequence of an action, this object-seeking impulse for the consequence is called a “motive” (ME 202, 362). This notion seems to correspond to Kant’s notion of choice or spontaneous and actionable act-opting power. Moreover, Sidgwick points out that the targeted object of such impulses may or may not be pleasure. Pleasure is understood as a purely subjective

168  Mariko Nakano-Okuno positive feeling that the individual considers desirable or agreeable at the time of feeling it. Sidgwick denies that we always seek pleasure or absence of pain (ME 40, 52–54, among others). This discussion is also in line with Kant’s. Finally, and most importantly, Sidgwick observes that we do have higher-order mental activities that are distinct from our lower-order object-seeking impulses, the former of which are commonly called “volitions (as differentiated from desires in a narrow sense)” or “will (of a superior)” (ME 25, 34, 36, p. 363 fn.1; Nakano-Okuno 2011, 10–11). Sidgwick adds that this higher-order mental faculty operates by making an ought-judgment, which should be universally applicable to “any other action belonging to a certain definable class” and which “in rational beings as such” gives a certain kind of motive to action judged to be right (ME 33–34, 36). Because of these two features, universalizability and a special kind of motivating force, ought-judgments are commonly regarded as being “rational” or based on “Reason” (ME 23, 33). This notion of higher-order motive corresponds to Kant’s concept of will. Not surprisingly, Sidgwick agrees with Kant that an act chosen by will is not always performed because a person may be moved by other, lowerorder object-seeking impulses while being conscious of what action he has consciously chosen as an act-to-be. Sidgwick states, “in human beings, of course, this [impulse or motive generated from one’s cognition of an ought-judgment] is only one motive among others which are liable to conflict with it and is not always—perhaps not usually—a predominant motive” (ME 34). As with Kant, one’s will, which is based on a moral judgment, and one’s motive, which can include both moral and non-moral impulses, are two distinct concepts for Sidgwick. However, there is one additional point about Sidgwick’s view of will and moral judgment. It is one of Sidgwick’s firm convictions that our moral judgments are fallible, as far as we humans are concerned. What we judge to be a “rational” moral law about what ought to be done may often turn out to be “false.” We may fail to universalize our maxim properly or to take into account morally relevant features of the situation. Some people may not even recognize that they are supposed to will their ought-judgments into universal laws and to respect other persons.9 Thus an ordinary person’s ought-judgment-based object-seeking impulse, or “will,” may not be apprehending truly rational moral laws about what ought to be done. In fact, Sidgwick says, this is precisely the reason why we turn to ethics (ME 5–6). If our will can always recognize what truly and rationally ought to be done, it would have been futile to ask, “Why should I do what I see to be right?” Such a question would have been an empty question which means “Why should I rationally do what I see to be what should be rationally done?” Still, we do “widely and continually” ask such a question and do not think it is an empty question. Importantly, this is not simply the matter of our weakness of will in which

Kant and Sidgwick on Freedom of Will  169 humans tend to surrender their rational will to sensible impulses. Humans may not form a rational will in the first place. Sidgwick still believes that, when we seriously ask ourselves the question of why we should do it, we are usually already committed to pursuing whatever is shown upon reflection to be really right, though we are not certain about whether what we think to be rational and right is truly so. The “unphilosophic man” (ME 6) often has a confused idea of what principle or law to use when determining what one ought to do. So he wishes to get rid of such confusions and turns to ethics. We will return to this point later. The Basic Meaning of “Freedom” in Kant As previously stated, Kant’s concept of freedom concerns the property of will or choice in the sense defined earlier.10 Kant distinguishes two aspects of such freedom. The negative concept of freedom means that one’s choice (spontaneous and actionable act-opting power) can work independently of determination by sensible impulses and of any empirical determinants. The positive concept of freedom means that the pure practical reason can give a universal law to oneself— “the property which a will has of being a law to itself” or “the ability of pure reason to be of itself practical” (Gr 4:446–447; KpV 5:33; MS 6:213–214). In either sense, the freedom of will is an idea (Idee), or a pure rational concept for which we can find no corresponding instance in our possible experience. Nevertheless, we may be able to recognize its existence through the representation of moral laws. A moral law that is posited by the free will is “of a special kind” that is distinct from the laws of nature (Gr 4:446). Various interpretations are made as to what precisely Kant meant by this concept of freedom. From the most lenient perspective, we could interpret Kant as using the term freedom as having all the following connotations: 1. It is not coerced by other individuals or authorities (Non-coercion), 2. One’s appetitive impulse originates from one’s mind and from nothing else that is external to one’s mind (Self-emanation), 3. One’s appetitive impulse is independent of natural causality or necessity, laws of nature, and all empirical or sensible grounds (Transcendental freedom), 4. One’s reason can posit a universal law by itself (Self-legislation). Furthermore, 3–4 implies that 5. In performing a certain act, it is always possible that the agent could or should have acted otherwise (Alternative possibilities). Therefore, 6. It can be the ground for attributing responsibility to the agent who formed a will based on the universal law one set for oneself (Moral responsibility).

170  Mariko Nakano-Okuno Of these possible connotations, the unique and vital to Kant’s moral philosophy is transcendental freedom (see 3). What Kant intensively discussed in the third antinomy of his Critique of Pure Reason is freedom as a sort of causality which is different from that of the laws of nature, and it is believed that it is freedom in this sense that constitutes the ground for attributing responsibility. Whether in the negative or the positive sense, the freedom of will or choice in Kant obviously means that it is not constrained by natural necessities. Kant maintains that such freedom is a sort of causality in that it causes a will to posit a law for oneself, as described in 4. This emphasis on self-legislation is another key feature of Kant’s notion of freedom. Additionally, the freedom of being not subject to sensible and empirical grounds is the property of our own faculty, which is called will or choice. It should be noted here that non-coercion and self-emanation (connotations 1 and 2) are features not limited to a rational will that posits a universal law to oneself. When one spontaneously forms a non-rational choice or motive to attain one’s desired object, and when it is not coerced or hindered by others, her choice or motive will have features 1 and 2, and we may want to call such a choice “free” in an ordinary sense. Also, when one is physically constrained and unable to attain her desired object, we may want to state that her freedom of non-rational choice has been ignored or undermined. However, this ordinary sense of freedom of non-rational choice should not be confused with the freedom of reasonbased choice in the sense of 3 and 4. What makes Kant’s concept of freedom unique is the implications of 3 and 4, so we should be mindful of these features.11 It is also important to point out here that transcendental freedom and self-legislation may not be necessarily undermined by physical constraint or coercion, as long as the subject can form a universal law in her mind and choose the act-to-be. It remains unclear how such unique Kantian senses of freedom (3 and 4) can be tied to the ordinary senses of freedom of choice. It seems reasonable to assume that Kant may have either confused all of these connotations under the single term “freedom of choice” or distinguished between those different connotations and used his concept of “freedom” to essentially mean 3 and 4 (with implications 5 and 6 following from them). I shall assume in what follows that Kant takes the latter approach. The Importance of Freedom in Kant’s Moral Philosophy Based on the explanations of choice, will, and freedom described earlier, Kant goes on to argue for the critical importance of freedom of will, or choice, in the whole system of moral philosophy. For Kant, our moral practice does not make sense if we cannot assume that our choice is free. According to him, freedom is an idea that cannot be found directly in our experiences, but for the moral law to exist, our will must be free in

Kant and Sidgwick on Freedom of Will  171 that it can present a law to oneself without being influenced by empirical conditions. By recognizing the moral laws in ourselves, we can realize that we are free in that sense. Thus, Kant states, “freedom is the ratio essendi [existential basis] of the moral principle, and the moral law is the ratio cognoscendi [recognition basis] of freedom” (KpV 5:4 [explanatory terms added]). Kant also maintains that every rational being, as far as he or she possesses a will, must necessarily have the idea of freedom and he or she can act only under this idea of freedom (Gr 4:448). The idea of freedom is thus inseparably linked to the notion of our acting under moral laws. Kant claims further that the idea of freedom of will also relates to the notion of moral responsibility. According to him, imputation in the moral sense, or the judgment that someone is morally responsible, is the judgment by which someone is deemed as the author (free cause) of an act under laws (MS 6:227) This judgment of imputation or moral responsibility comes with the notion of merit. The greater physical and psychological obstacles one had to overcome in performing a moral act, and the clearer one’s moral duty was, the more merit is to be accounted for the act that was done (see MS 6:228). Without transcendental freedom, however, no moral laws are possible, and hence no moral imputation is possible (KpV 5:97). So, Kant’s idea of freedom of will consists of one’s ability to set laws for oneself, being free of sensible impulses, natural inclinations, or other influences bound by the laws of nature. Without this notion, Kant believes, morality and moral responsibility do not make sense. When we act morally, we should be able to assume that our will is free in this sense. Because one acts out of free will in this sense, one is regarded as the agent who freely initiates the act he or she chooses with the representation of a universal law, and thus the act done is morally imputed to him or her. Sidgwick, however, rejects Kant’s account—“Kant’s resting of morality on Freedom did not indeed commend itself to me” (ME xix). Sidgwick’s criticism is twofold. First, Sidgwick points out certain ambiguities and confusions in Kant’s argument on freedom. Second, Sidgwick doubts the importance of assuming the freedom of our will in ethics.12 In my opinion, it is possible for Kant to fend off part of the first criticism, but Sidgwick’s argument on our moral fallibility will leave another part of Kant’s problem unresolved. Furthermore, it is difficult for Kant to play solid defense against Sidgwick’s second criticism. Let us look into each of the two objections in turn.

Sidgwick’s Criticisms of Kant’s Idea of Freedom Two Meanings Confused? Sidgwick’s first criticism concerns what he regards as Kant’s confused use of the two different senses of freedom (ME 511 ff.). According to

172  Mariko Nakano-Okuno Sidgwick, Kant sometimes uses the term “freedom” as equal to rationality, in which case we can say that the more one acts in accordance with reason, that is, the more a person acts according to a universal moral law about what ought to be done, the freer that person is. Sidgwick conveniently labels this notion “rational freedom.” Meanwhile, when Kant discusses moral imputation and considers the situation in which one acted wrongly but could have acted otherwise, he seems to use the term “freedom” in the sense in which we can say that a person’s will is free regardless of whether that person intentionally chooses a moral or an evil act. Sidgwick calls this second notion “neutral freedom.” It is inconsistent, however, to claim on the one hand that a person chooses an evil option just as freely as when he chooses a good option, and on the other hand that the same person is freer if he chooses the morally right option in accordance with his faculty of reason. Sidgwick’s point seems simple and clear. However, it may be possible to understand Kant as not having such confusions. Although Kant certainly associates freedom with reason and moral laws, Kant may not be claiming that freedom means the state in which one acts in accordance with the law set forth by reason. As Sidgwick points out, Kant indeed states that “a free will and a will under moral laws are one and the same” (Gr 4:447). However, Kant clearly and deliberately makes a distinction between humans “under moral laws (unter moralischen Gesetzen)” and humans who “act in accordance with moral laws (verhält sich gemäß den moralischen Gesetzen)” (KU 5:448 fn). When a person consciously makes a rational moral judgment under a certain universal law, her will is under the moral law whether or not she acts in accordance with it. She may succumb to her other impulses or motives and fail to act in accordance with the universal law she has set, but her will is still under the moral law as long as she is aware of what she should have done. Given this distinction, the statement that “a free will and a will under moral laws are one and the same” does not mean that “a free will and a will to act in accordance with moral laws are one and the same.” Sidgwick also refers to Kant’s claim that “[freedom] must be a causality according to immutable laws, but of a peculiar kind” (ME 53).13 From this statement, Sidgwick presumably infers that, if freedom is a causality according to immutable laws, and if the will is a sort of causative faculty that derives acts from laws, then freedom must mean the state in which one’s will brings about an act according to moral laws. However, what these laws “of a special kind” connect in an immutable manner to one’s will should not be the actual performance of an act but an act that ought to be performed. Kant further clarifies that what ought to happen does not always happen (Gr 4:387–388; 4:421). Thus Kant’s statement that “freedom is a special kind of causality according to immutable laws” merely means that freedom is the state in which one’s will directs one to

Kant and Sidgwick on Freedom of Will  173 an act that ought to happen according to such laws. It does not imply the state in which one’s will actualizes such an act. Will is considered a product of our faculty of reason, but the actual wills of humans are simultaneously exposed to various subjective conditions such as inclinations and non-rational impulses. Thus the will of humans is one that, “although it is determined by principles of reason, does not necessarily follow these principles in virtue of its own nature” (Gr 4:413). As such, a human may at times commit wrongdoing while being under moral laws. In his mind, his will is necessarily and immutably associated with “the acts to be.” At the same time, this will is exposed to other subjective conditions such as inclinations and impulsions, and he may end up behaving differently under the influences of both reasonbased object-seeking impulses and non-rational impulses. Nevertheless, this person’s will can be called free in a positive sense, insofar as his will posits moral laws to himself and hence is “under moral laws.” According to Kant, the attribution of moral responsibility to someone is the judgment by which that person is regarded as an author of an act that stands under laws (MS 6:227). A villain who is aware that he has done wrong is still under moral laws in that he knows what ought to have been done. He is therefore responsible for his indulgence in allowing his inclinations and non-reason-based impulses to affect his maxims to the detriment of the rational laws governing his will (Gr 4:457–458). Understood this way, Kant can fend off Sidgwick’s criticism of the alleged confusion of the aforementioned two forms of freedom. A person is free when she forms a choice, or spontaneous and actionable act-opting power, toward what ought to happen under moral laws, which her reason posits to herself. She may end up acting otherwise after succumbing to her sensible stimuli, but she is still free in the sense that her will toward what ought to happen was determined independently of sensible and empirical grounds. Kant did not claim—we can even say that he carefully avoided claiming— that the agent is more free when she does the right thing than when she performs an evil act while being aware of what she should have done. Kant certainly states that such an agent will deserve more merit when she does the right thing (MS 6:228), but it does not mean she was more free. By being aware of what she ought to do, the subject is already free whether or not she acts upon it. If Kant claimed that persons, as rational beings, exhibit even more freedom when they act in accordance with their rational will, then this part of Kant’s argument would rightly be criticized by Sidgwick for containing confusions. However, there is no textual evidence that Kant himself had such confusion between rational and neutral freedom.14 However, this line of counterargument, which Kant could have made against Sidgwick, only applies to situations in which a person has made a rational moral judgment, is clearly aware of it, yet surrenders his rational will to his other sensible temptations and ends up performing wrongdoing.

174  Mariko Nakano-Okuno What if his moral judgment was not rational in the first place? Is he free and morally responsible for what he did out of his respect for what he believed to be moral laws but what turned out to be not rational (by “not rational” I mean what is not genuinely universalizable and lacks respect for persons)? Let us recall Sidgwick’s view about human moral judgment. Humans tend not only to surrender their moral judgment to their inclinations and sensible impulses, but they also tend to make mistakes in making moral judgments. Suppose a terrorist seriously believes in, and acts upon, what appears to him to be a universal moral law which dictates that anyone who tarnished a certain holy script ought to be severely sanctioned and, if they are deemed incorrigible, exterminated. We may also suppose that a cult leader honestly thinks his reason has grasped a universal moral law that humankind ought to proliferate through unrestricted sex including extramarital intercourse and sexual relationships with teenagers. Kant may claim that what they think to be practical laws are not really reason-based moral laws but merely their subjective maxims which are not fully universalizable and not based on respect for humanity in persons. How can we call such a terrorist or a cult leader free and therefore morally responsible for what they do? Aren’t they unfree and not morally responsible because they have not grasped genuinely rational moral laws at the time of acting, probably being hampered by their nonrational, sensible, empirical impulses, by a lack of knowledge, or by other external factors, such as education and local trends, hindering them from growing into rational individuals capable of making truly universalizable and humanity-respecting moral judgments? Supporters of Kant may claim that anyone who has the capacity to formulate genuinely rational moral laws is free and morally responsible for his acts. However, the question arises—how do we determine who has such capacity and who does not? When a terrorist has grown up in an environment that makes one believe strongly that it is the most sacred imperative that all blasphemers be exterminated, how can we meaningfully say that this person is capable of forming a genuinely universalizable, humanity-respecting moral judgment? It sounds more plausible to say that his upbringing and his living environment (which are external empirical factors) have made him incapacitated to make genuine moral judgments, and therefore he was not free in the Kantian sense (that is, in terms of transcendental freedom and self-legislation) at the time of making his ought-judgment. However, if these terrorists and cult leaders can be deemed unfree because of their upbringing and other external factors limiting their capacity, how about us ordinary people? Our capacity to posit genuinely rational moral laws is also limited due to our upbringings, the education we received, the lack of knowledge, the current trend in our time, and many other external factors. The problem is, we are all more or less imperfectly capable of being rational. As Kant admits, freedom is only an idea, which cannot be found directly in any instance

Kant and Sidgwick on Freedom of Will  175 of our experiences. Does this mean, then, that none of us are truly free and morally responsible for our actions? Kant still needs to explain whether and why a person’s imperfectly rational ought-judgment-based impulse should be treated as if it is an ideally rational being’s free will, and whether and why a person’s imperfectly rational ought-judgment could make that person held accountable for the act he ended up doing. Sidgwick’s suggestion of this disparity between ideal rationality and our limited rationality appears to remain valid. Sidgwick thinks that, if the concept of freedom is to make sense, we should construe it in the sense in which we can ascribe moral responsibility to a wrongdoer, and not in the sense in which one becomes freer as he or she acts in accordance with reason (ME 515). Even when understood in this way, however, freedom may not be an indispensable concept in morality.15 We will look into Sidgwick’s second criticism of Kant in the remainder of this chapter. The Unimportance of the Assumption of Freedom in Ethics Kant’s emphasis on the importance of freedom in ethics has two aspects: (1) subjective necessity of the assumption of freedom, and (2) the practical importance of freedom. According to Kant, when having a moral will, we cannot but assume (in our mind) that our will is free. This internal sense of necessity is called the subjective necessity of the assumption of freedom. Furthermore, Kant contends that our moral claims and moral responsibility do not make sense unless we assume the freedom of our will. This contention is called the practical importance of freedom. Subjective Necessity Refuted As we saw earlier in this chapter, Kant maintains that all rational beings who possess a will must also possess the idea of freedom, and only under this idea of freedom can they act. According to Kant, as long as we have the will that presents moral laws to ourselves, we must inevitably presuppose that we are free in the sense that our own faculty of reason can selflegislate what ought to happen independently of our non-reason-based impulses. Kant repeatedly asserts that this presupposition of freedom is only subjectively inevitable, namely, that this assumption is what a subject possessing a will inevitably makes about him- or herself. He states, “its objective validity can in no way be exhibited by reference to laws of nature and consequently cannot be exhibited in any possible experience” (see, for example, Gr 4:459). According to Kant, it is sufficient for our purpose if we can assert the subjective necessity of the assumption of freedom (or the presupposition of freedom merely as an Idea) in all rational beings (Gr 4:448; KpV 5:93–94). Why is it sufficient? One possible reason is

176  Mariko Nakano-Okuno that if each of us inevitably presupposes that he or she is free, we can propose to all rational beings to treat each other as a free being who is ready to undertake the responsibility for what he or she chooses to do. Another reason is that if we can accept this positive concept of freedom as the common feature of us rational beings, then according to Kant, we can also address other issues of philosophical interests, such as the immortality of the soul and the existence of God, among others. Setting this last claim aside, Sidgwick is not against Kant’s claim that, in making moral judgments, we do believe that we are capable of setting universal laws for ourselves. Like Kant, Sidgwick admits that, in our minds, moral decision making, or object-seeking impulses based on one’s moral judgment, is clearly distinguished from mere impulses and selfish desires (ME Bk.1 Ch.3 Sec.1; 36–37). We believe that our moral object-seeking impulses are formed under certain universal laws, while selfish desires or mere impulses are random, capricious, or narrowminded. Such a rational will is considered distinct from other primitive and instinctive impulses, yet it is still a sort of internal impulse we form in our mind. We make a moral judgment based on the fundamental moral principles we perceive as valid and form a will to act according to it (ME 33–34, 36–37). In doing so, we cannot but believe that we are capable of setting such laws by and for ourselves and determining our will with the awareness of such laws by ourselves. Thus far, Sidgwick has no objection. Sidgwick’s most acute criticism is not about Kant’s claim that we do have the ability to set laws for ourselves. Sidgwick’s criticism is about Kant’s claim of the mysterious origin of such an ability, i.e., the claim that our ability to set laws for ourselves, or what Kant calls “practical reason,” must exist apart from any empirical causes. We may inevitably believe that we are capable of setting fundamental moral principles or universal moral laws by ourselves, but it is unclear whether we can also claim that such universal principles or laws are grasped by the faculty that follows a different sort of causality from that of natural laws, as Kant always means by the faculty of reason. It is also unclear whether this faculty itself is completely free of influences by past experiences and other empirical factors.16 Sidgwick, who lived in the postDarwinian era, maintains that Kant’s doctrine of freedom “may be left to conduct an unequal struggle with the modern notions of heredity and evolution” (ME 515). To our mind, our will under moral laws appears to be a special kind of voluntary impulse that follows a certain universal law that is distinct from the laws of nature. However, our faculty to set such seemingly non-empirical laws may not be free from the laws of nature when we see it through the lens of evolutionary theories. We may have acquired this law-forming faculty as the result of natural selection, which gave more chance of survival to those who happened to have more of such a faculty. When we contemplate and

Kant and Sidgwick on Freedom of Will  177 adopt moral laws to guide our actions, we feel as if we do so by ourselves, and we may even believe that we set these laws independently of any external forces, sensible impulses, or other empirical factors, but it is also possible that the universal laws we feel as if we have set by ourselves may not, in reality, be set by our free self. We do not need the subjective assumption of our complete freedom—our will’s complete freedom from all biological and environmental determinants—for us to be able to set universal laws for ourselves. We need no supernatural origin of our faculty of reason. Practical Importance Refuted At this point, it may be objected that, if our will turns out not free but bound by external influences, the notions of moral agency and imputation no longer make sense. Kant would claim that this will make the whole system of morality collapse. However, the most intriguing aspect of Sidgwick’s take on ethics is his denial of the practical importance of freedom. His refutation of the practical importance of freedom appears in Methods Bk.1 Ch.5, in which he deals with two questions. One is the question of whether the subjective necessity of the assumption of free will is critically important in our practice of morality—in other words, whether the failure to conceive of ourselves as free of any empirical determinants damages our moral practice. The second question is whether the concept of freedom is necessary to enable imputation. To both questions, Sidgwick answers in the negative. Simply put, he contends that our practical claims will remain virtually the same whether we assume or deny the freedom of our will. As for the first question of whether it is practically important to think of ourselves as free at the time of determining an action, Sidgwick argues as follows. Even if it turns out that our will is ultimately governed by the laws of nature, the reasons (grounds) based on which we adopt the right act (or the right maxim) remain unchanged. To admit that our will is ultimately determined by natural and empirical causes is not to believe that we can accurately predict what action we will choose by inferring from our past and present experiences. We still have to consider, deliberate, and reason out what we ought to do. Through such a deliberative process, it suffices to assume that, whether we set universal laws by our faculty of reason alone or under the influences of the laws of nature, we are to accept the three fundamental ethical axioms and ethical hedonism presented in Methods as self-evident and significant components of our morality. To borrow Kant’s terminology, we are to conceive of the formulae of Categorical Imperatives as the foundation of our morality and to adopt general happiness as the ultimate end of our actions-to-be. Even if it turns out that our will is ultimately governed by the laws of nature, we will not, and need not, stop making ought-judgments to guide

178  Mariko Nakano-Okuno our actions and test those judgments against these fundamental axioms which remain the basis of our morality. It is possible that, when we assume that our actions and decisions are all ultimately determined by natural and empirical forces, the motivating power that prompts us to the right action may weaken to a greater or lesser degree. Sidgwick himself admits that, if a person regards his actions as “necessary results of causes anterior to his personal existence,” the sentiment of remorse (understood as self-blame) for those actions will tend to vanish from his or her mind (ME 71). Nevertheless, we can still be—equally or even more strongly—motivated to perform moral actions by the power of imagination, sympathy, and the love of goodness. We should not assume that the imagination, sympathy, and love of goodness of those individuals whose will is determined by the laws of nature are less fertile or weaker than those of free beings. What about the second question of whether freedom is necessary to make imputation possible? Kant believes that if there is no a priori practical freedom, then no moral laws and no moral imputation are possible. However, as already stated, Sidgwick does not see the necessary connection between freedom and moral laws, nor does he see the necessary connection between moral responsibility and freedom. (ME Bk.1 Ch.5 Sec.4–5.) Admittedly, our common notions of desert, merit, and responsibility appear to presuppose free will. However, we can use concepts such as sanction, punishment, reward, and responsibility in a perfectly meaningful way without presupposing free will. As utilitarians do, we can give a meaningful sense to the statement “this person is responsible” if we construe it as contending that “it is right to sanction this person for the harmful act done by him.” To prevent similar harmful acts from being done by him and others in the future, we sanction a person who has the manifest tendency to commit the same types of wrongdoing, and this tendency is typically recognized through a person’s record of harmful acts performed in the past.17 Without presupposing the freedom of will, we can regard a person as accountable for the act done by him to the extent to which a series of consequences are expected to result from that action. When we take this view without presupposing the freedom of will, our sentiments of moral praise or condemnation might be less associated with our retrospective judgment of what a person previously did, and more associated with prospective impulses to encourage benevolent acts and to prevent harmful acts in the future. There is no need to believe that the effects of these prospective impulses are going to be weaker than the effects of retrospective evaluation of criminal acts already done. Our discussions in the previous two sections can be summarized as follows. Like Kant, Sidgwick acknowledges the self-emanative feature of our moral will, that is, the feature that our moral judgments and their associated motives originate from our own minds. According to Sidgwick, Kant is respectably unique and important in that he found the

Kant and Sidgwick on Freedom of Will  179 basis of morality in our own faculty. However, Sidgwick believes, we do not have to assume that this faculty (reason or philosophical intuition) is miraculously endowed to us apart from any natural and empirical grounds. Furthermore, whereas Kant considers freedom as the cornerstone of our reality and morality and thinks of it as inseparably linked to the concept of moral responsibility, Sidgwick feels no necessity to presuppose the freedom of our will or to argue for its practical importance. This is because our practical claims will remain virtually the same regardless of what stance we take on the debate over the freedom of will. The dictates of our will may appear to be distinct from what our sensible impulses drive us to do, but that is because our moral judgments have certain features, such as universalizability, time indifference, and personal indifference, which cannot be found in other sensible impulses. We do not have to assume that our faculty of making such moral judgments is free of any empirical and natural grounds. We do not have to presuppose the connection between freedom and moral responsibility either. After all, Kant and Sidgwick significantly disagree on the connections between the notions of reason, moral judgment, empirical and natural causality, and responsibility as well as their practical importance in ethics.

Conclusion Despite remarkable similarities between their foundational claims in ethics, Kant and Sidgwick’s views diverge in their handling of freedom in morality. It should be noted that Sidgwick never concluded that there is no such thing as free will. His point is that, whether or not we presuppose freedom of will, we will, after all, adopt almost the same moral principles and make moral judgments based on them. Therefore, according to Sidgwick, there is no need to decide the issue of freedom of will before we develop a credible ethical theory that can be widely accepted. Still, this reasoning indicates that Sidgwick does not believe the freedom of will is what really matters in ethics. If the followers of Kant still strongly believe in the reality of the freedom of will (in Kant’s sense) and its practical importance, they will have to respond to Sidgwick’s arguments and demonstrate the following two points. First, they will need to plausibly show that our rational ability to set a law is essentially unrelated to sensible impulses and other empirical determinants governed by the laws of nature. Second, they will need to persuasively show that the concept of the freedom of will is vital to our entire system of moral beliefs and that any other explanations, such as those of utilitarians, cannot adequately supply the theoretical grounds for our practice of morality. It will not be easy to complete these two missions. They will not succeed in attaining the proof of transcendental freedom as long as we the imperfectly rational can perceive and understand

180  Mariko Nakano-Okuno only things and events in this phenomenal world. This metaphysical attempt may be irresistible for some philosophers, but it is unlikely that their ever-continuing debate over transcendental freedom can guide us to solve the practical ethical issues we encounter. In that sense, this pursuit of freedom would not make differences in our ethical decision making. Then it would appear worth listening to Sidgwick’s assertion that we can develop a credible ethical theory by adopting the fundamental moral principles whose content is virtually equivalent to what Kant presented as the formulas of Categorical Imperatives, even when we cannot prove the ultimate reality of transcendental freedom. When we have alternative courses of action before us, we stop and think which act we ought to perform, then choose what we consider to be the right action, whether or not our faculty of reason is ultimately constrained by the laws of nature. Whether we are free or unfree, we can appreciate our being in this world and respect each other’s wishes and wills. This is, in my view, all the explanation we need in order to discuss our moral practice in this phenomenal world.

Notes 1. Some may point out that Sidgwick supported not only utilitarianism but also rational egoism, as that is what he clearly indicated in The Methods of Ethics. That is a fair explanation of Sidgwick’s stance in that particular volume. However, I contended elsewhere (Nakano-Okuno 2011, 16–19) that Sidgwick was basically a utilitarian. His rigorous scrutiny in The Methods of Ethics led him to admit that all three major methods of ethics (utilitarianism, egoism, and dogmatic intuitionism) have certain rational bases. Still, we can say that Sidgwick remained a supporter of utilitarianism by claiming that it is the most cogent moral theory insofar as one takes a broader point of view, which Sidgwick called the point of view of the universe (ME 420–421; Nakano-Okuno 2011, 151–154). 2. There seem to be common misunderstandings of Sidgwick’s axiom of rational benevolence. Some equate this axiom to the utilitarian principle of pursuing others’ happiness. This is a sheer misunderstanding because Sidgwick clearly states that this axiom is yet to say anything about happiness. Others seem to identify this axiom with a maximization principle to pursue the greatest sum total of people’s goods (I found this type of interpretation in the early draft of de Lazari-Radek and Singer (2014) and Schultz (2017)). This understanding is also incorrect, as Sidgwick indicates nothing about maximization or aggregation in this axiom. One of the two philosophical intuitions, the combination of which entails the axiom of rational benevolence, does mention the rationality of pursuing peoples’ good on the whole when seen from a universal viewpoint, but pursuing people’s good “on the whole” does not necessarily mean maximizing the sum total of people’s good. For example, Kant has in mind all rational beings’ good on the whole when he discusses the kingdom of ends; John Rawls attempted to pursue people’s good on the whole by equally securing basic liberties to all members of society while giving priority to the least advantaged in distributing societal and economic benefits. Sidgwick’s axioms of prudence and benevolence should be understood simply as the principles that require time and personal indifference

Kant and Sidgwick on Freedom of Will  181 in weighing people’s goods at different points in time. (Nakano-Okuno 2011, 96–103, 112–113). 3. Although it is sometimes said that Kant’s thoughts show some transitions over time in his definitions of concepts and discussions of certain topics, I assume that Kant’s basic stance remained largely consistent throughout the critical period. When apparent inconsistencies among his claims are found, I generally follow Kant’s arguments in his later writings, assuming that those writings reflect his more considered stance on the topics discussed. 4. I intentionally left the meaning of “will” open here. The definitions of “will” in Kant and Sidgwick are explained in the subsequent section. 5. Kant’s works are abbreviated herein as follows: KpV = Critique of Practical Reason (Kant 1996a); Gr = The Groundwork of the Metaphysics of Morals, (Kant 2000a); MS = The Metaphysics of Morals (Kant 1996b); KU = Critique of the Power of Judgment (Kant 2000b). 6. The distinction between choice and will is most clearly stated in Kant’s The Metaphysics of Morals (MS 6:211–220). Some researchers (see, for example, Wuerth 2014) claim that Sidgwick failed to recognize this Wille/ Willkür distinction in Kant, but I would suggest that a similar distinction is found in Sidgwick’s own explanations of moral judgment, desire, and motive (see the first section of this chapter). 7. In Gregor’s English version of MS (Kant 1996b), the German terms “Begehrungsvermögen,” “Begehren,” and “Begierde” are translated as “the faculty of desire,” “desire,” and “desire.” However, I choose to follow J. W. Semple’s 1886 translation and adopt the terms “the appetitive faculty” and “appetitive impulse” for the former two German terms and use “desire” only as a translated term for “Begierde.” Only “Begierde” refers to an impulse toward pleasure, and “Begehren” refers to a much broader impulse targeted at a certain specific object, whether this impulse originates from the expectation for pleasure or the faculty of reason. As we will see later, however, Sidgwick called all these object-seeking impulses “desires.” 8. Sidgwick himself mentions the widespread influence of Kantian theory on the ethical thought in his time (ME 58 fn.1), so it may have been natural for him to adopt similar terminologies whenever doing so seemed appropriate. 9. According to Sidgwick, the three fundamental axioms—the axiom of justice that requires our moral judgment to be universalizable, and the axioms of rational prudence and benevolence that requires time and personal indifference in weighing goods—are considered what our philosophical intuition would find to be the rational basis of morality. Thus, Sidgwick would agree that one’s rational will have to satisfy all these requirements. The problem is, however, that the will of ordinary persons is rarely fully rational. 10. In MS, Kant maintains that “the will, which is directed to nothing beyond the law itself, cannot be called either free or unfree,” and that “only choice can be . . . called free” (MS 6:226). Strictly, our discussion of the freedom of will in this chapter refers to the freedom of choice in Kant. 11. It seems to me that some discussants of the freedom of will often confuse freedom in this ordinary sense with the freedom in the sense that is essential to Kant’s moral philosophy, that is, transcendental freedom and selflegislation. For example, when Paul Guyer discusses that a universalizable moral judgment must be the one that would not invade others’ freedom (see his chapter in this volume), I am not sure if he is talking about freedom in the sense of 3 and 4 or freedom in the ordinary sense that is described in this paragraph. Suppose the Nazis confine Jews in the concentration camp, fanatically believing that their maxim, “we ought to confine the eugenically

182  Mariko Nakano-Okuno inferior race in one place so that they won’t attempt interracial marriage to tarnish the superior race’s gene pool”, is perfectly universalizable. (Guyer’s own example contains the killing of the Jews, which literally deprives the Jews of all kinds of choices, rational or non-rational; thus I modified the story so that we can clarify what kind of confusion his approach can contain.) Guyer may perhaps claim that putting people in the concentration camp will jeopardize those captivated people’s freedom and therefore not universalizable. Confining them in a concentration camp will certainly limit the Jew’s “freedom” in the ordinary sense, as people in the concentration camp are unable to choose many things they desire to have. However, even in a concentration camp, unless they are killed or severely disabled, the Jews must have been able to form a will to act under a universalizable moral judgment (such as “I must not hit the man next to me in order to go ahead and get today’s soup before the supply runs out” or “I should not steal a dead man’s belongings just because I want to get his shoes”—these maxims are what people in the concentration camp actually formed but often failed to put into practice due to their strong natural inclinations, as seen in Frankl (2006)). Therefore, the Jews in the concentration camp were, as long as they were alive and not impaired, not deprived of transcendental freedom by the Nazis. If one claims that their freedom is deprived, one should articulate what that “freedom” means in that context, and how this freedom relates to Kant’s unique notions of transcendental freedom and self-legislation. 12. This point is addressed in Sidgwick’s discussion on the free will controversy in general, not specifically targeted at Kant. In Methods Book 1, Chapter 5, where he discusses free will, Sidgwick indicates that he has intentionally excluded Kant’s concept of free will from the discussion of that chapter (ME 58 fn.1). In my view, however, Kant’s belief in the practical importance of freedom is shared widely by the defenders of free will, and Sidgwick’s discussion in that chapter contains a valid criticism against such a belief. 13. I simply cited this passage from Sidgwick’s book. Its original appears in Gr 4:446 and is translated as “[freedom of will is] a causality conforming to immutable laws, though of a special kind” in Kant (2000a, 107). 14. Paul Guyer provides a further supporting argument on this. See his chapter in the present volume. 15. In his chapter in the present volume, Guyer claims that Sidgwick’s critique of Kant’s reasoning leading to the principle of rational benevolence (ME 389–390) is based on Sidgwick’s more general failure to appreciate how the conception of freedom is foundational to Kant’s moral philosophy. In the following sections, I will attempt to explain why Sidgwick denounced the concept of freedom as a foundational value to moral philosophy. As for my own critique of Guyer’s freedom-based argument, see endnote 11 in this chapter. 16. Here we should recall our point that freedom is the property of the faculty of will or choice. Like Kant’s Categorical Imperatives, Sidgwick’s fundamental moral axioms are not what we can derive via induction from empirical facts, but what we apprehend by our philosophical intuitions (that is, our ability to grasp apparent truths) as “self-evident and significant propositions.” Such propositions are called a priori synthetic propositions in Kant. This indicates that the content of Sidgwick’s three axioms is also not determined by natural causality or sensible impulses. However, it does not follow that our faculty to apprehend such axioms (i.e., philosophical intuition) is entirely independent of natural causality or sensible impulses.

Kant and Sidgwick on Freedom of Will  183 17. Utilitarianism is the view that the morally right act, or what one morally ought to do, is the one, among all feasible courses of action, that is expected to bring about the greatest overall happiness or preference-satisfaction for all parties involved. Happiness is interpreted as equal to pleasure, or as what is constituted by pleasures (ME 92). Pleasure here refers not only to sensual or physical pleasure but also to any kind of agreeable, pleasurable, and satisfied feeling which the person feeling it would find desirable. Sidgwick’s (and Mill’s) Utilitarian ethics should not be confused with what we commonly perceive as pragmatism or materialism. Utilitarianism should not be confused with egoism either. While egoism mainly concentrates on one’s own pleasure or happiness alone, utilitarianism considers the happiness or preference-satisfaction of all individuals who are expected to be affected by one’s action. Kant is often thought to have acutely opposed the idea of the moral will being determined by empirical factors such as happiness, but in NakanoOkuno (2007), I maintained that Sidgwick’s hedonism does not contradict Kant's argument.

References de Lazari-Radek, K., and P. Singer. 2014. The Point of View of the Universe: Sidgwick and Contemporary Ethics. Oxford: Oxford University Press. Frankl, V. E. 2006. Man’s Search for Meaning. Boston: Beacon Press. Kant, I. 1886. The Metaphysics of Ethics, trans. J.W. Semple, ed. with introduction by Rev. Henry Calderwood (Edinburgh: T. & T. Clark, 1886) (3rd edition). Kant, I. 1996a. Critique of Practical Reason. Translated by T. K. Abbott. New York: Prometheus Books. Kant, I. 1996b. “The Metaphysics of Morals.” In Practical Philosophy, edited and translated by M. Gregor. Cambridge: Cambridge University Press. Kant, I. 2000a. Groundwork of the Metaphysics of Morals. Translated by H. J. Paton. London: Routledge. Kant, I. 2000b. Critique of the Power of Judgment. Translated by P. Guyer and E. Matthews. Cambridge: Cambridge University Press. Mill, J. S. 1978. On Liberty. Indianapolis: Hackett. Nakano-Okuno, M. 2011. Sidgwick and Contemporary Utilitarianism. Basingstoke: Palgrave Macmillan. Nakano-Okuno, M. 2007. “Sidgwick and Kant: On the So-called ‘Discrepancies’ Between Utilitarian and Kantian Ethics.” In Henry Sidgwick: Happiness and Religion, edited by P. Bucolo, R. Crisp, and B. Schultz. Catania: University of Catania Press. Schultz, B. 2017. The Happiness Philosophers: The Lives and Works of the Great Utilitarians. Princeton: Princeton University Press. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Wuerth, J. 2014. Kant on Mind, Action, and Ethics. Oxford: Oxford University Press.

Part V

Ultimate Ends

9 Aristotelian Eudaimonism and the Dualism of Practical Reason Sukaina Hirji

Both Kant and Sidgwick recognize a deep tension between moral and prudential reasons. On the basis of what we can observe, both agree that fulfilling one’s moral obligation is not always prudentially optimal. And, both agree, if acting morally must sometimes come at the cost of one’s own happiness, then we should be skeptical about the possibility of having a completely rationalized—that is, a completely systematic and coherent— account of practical reason. Both suggest that, the empirical world being thus, the only way that practical reason can be rendered coherent is if we assume the existence of God, as the being who will ensure that acting morally is always conducive to one’s happiness, through a system of divine rewards and punishments. Ancient eudaimonism might seem to offer a solution to the dualism of practical reason, and without any theological assumptions. Eudaimonism holds that there is indeed a single ultimate rational aim that can accommodate all of our ethical concerns, namely, the pursuit of an agent’s own happiness. Kant and Sidgwick, however, are pessimistic about the possibility of eudaimonist theories successfully reconciling both ethical and moral concerns under a single system of practical reason. In this chapter, I review the challenges that Kant and Sidgwick pose to eudaimonism and argue that there are resources within ancient theories— and with Aristotle in particular—to potentially resolve at least some of their concerns. Much turns on how exactly we understand ancient eudaimonist theories, and I argue here that this question is far from settled.

I Though Kant and Sidgwick ultimately defend very different ethical pictures, they share in common a rejection of the eudaimonism of ancient ethical theories. In particular, both criticize ancient theories for failing to distinguish between the reasons given by morality, and the reasons given by self-interest. In what follows, I will consider two of the main criticisms of eudaimonism from both Kant and Sidgwick.

188  Sukaina Hirji Kant’s first worry, directed primarily at an Epicurean conception of eudaimonia as consisting in pleasure and the absence of pain, is that acting for the sake of one’s happiness is not a moral motive. Kant distinguishes the worth of one’s state or condition (Zustand) from the worth of one’s person. Our state includes pleasure and pain but also everything external to us that causes them, as well as our temperament and natural constitution. By contrast, our person is the object of moral evaluation. Our person includes our actions and character, which express our good or evil will. So also, Kant distinguishes the “natural” good and the “moral” good and criticizes ancient theories for failing to distinguish these two goods (see Wood 2018, 246). Kant insists that any moral theory that takes some conception of the natural good to be fundamental will make moral agency heteronomous insofar as it treats moral principles as hypothetical imperatives. According to such a theory, Kant argues, we only have sufficient reason to act on a moral principle if we have a sufficiently strong inclination toward the end that would be promoted by acting on that principle. But, Kant insists, when we consider the value that we attach to reason, we think its value extends beyond the way in which it can be used to secure the ends pursued by inclination. Indeed, for Kant, essential to rational agency is conformity with the categorical imperative, a non-instrumental principle, and so also conformity to moral requirements themselves. The categorical imperative, the fundamental principle of morality, is the law of an autonomous will—a human will in which the Moral Law is decisive is motivated by the thought of duty, not natural inclinations. Although Kant is especially concerned to reject eudaimonist theories on which happiness is associated with the satisfaction of desires or bodily pleasures, the deeper worry for any eudaimonist theory including Aristotle’s is that, if an agent is motivated by her happiness to perform an action that aligns with her duty, it is only by accident that she has done what duty demands.1 Again, for Kant, it is only from the motive of duty that the agent expresses her determination to act dutifully out of respect for the moral law itself; for her action to have moral worth on Kant’s picture, an agent should treat considerations of duty as sufficient reasons for her conforming to those requirements. Kant insists that when the principle of self-love or “one’s own happiness is made the determining ground of the will, the result is the direct opposite of the principle of morality” (KpV 5:35).2,3 It is only when happiness or Glückseligkeit is brought under the principle of reason that it is an end and good for the agent: “The agreeable . . . must first be brought under the principles of reason by the concept of an end in order to call it, as an object of the will, good” (KU 208; see Engstrom 1998, 104–105). Until then it is “not an ideal of reason but of the imagination, resting merely on empirical grounds” (G 418). It is not enough for this ideal to be viewed as agreeable for it to

Aristotelian Eudaimonism and the Dualism  189 be regarded as an end of action or good, nor indeed “the highest of all practical goods.” A second further worry that Kant presses, directed principally at Stoic theories on which eudaimonia consists in virtuous activity, is that eudaimonism gives us a defective conception of happiness. Kant believes that all human beings have a desire for happiness, where this is understood as “an idea in which all inclinations are summed up” (G 4:399; cf. KU 5:208). He describes happiness as a “system of inclinations brought into tolerable harmony” (KpV 5:73–76). Crucially, this conception of happiness is, according to Kant, entirely different from the Stoic conception. Kant denies that the kind of self-contentment one experiences when one is conscious of having done one’s duty is any part of happiness understood as the sum of the satisfaction of all our inclinations. This is clear from the fact that many inclinations conflict with duty (KpV 5:110–119). So, although Kant thinks that humans should pursue happiness, we should do so only so long as this pursuit does not conflict with duty: This difference of the principle of happiness from that of morality is not therefore an opposition between them, and pure practical reason does not require that we should renounce the claims on happiness; it requires only that we should take no account of them whenever duty is in question. (KpV 5:93) As with Kant, Sidgwick offers two main criticisms of ancient eudaimonia, presented as a kind of dilemma: Either ancient eudaimonism tells us that happiness lies in acting in accordance with virtue without filling in the content of the reasons we have, in which case the view is hopelessly uninformative, or eudaimonism offers a substantive account of happiness as acting in accordance with our common-sense conception of the virtues, in which case it is committed to a vicious circularity. Sidgwick locates two important differences between ancient ethical theories and the central ethical concepts of the modern period. First, he takes ancient ethical theories to be teleological, in the sense that the ultimate good is conceived independently of the right, and the right is conceived as what promotes the ultimate good in some way.4 This is in contrast to intuitional or deontological theories according to which actions are right when they conform to principles of duty that are unconditionally binding; on these theories the right is determined independently of the good (ME 105–106). Second, Sidgwick takes ancient ethics to be fundamentally egocentric in a way that modern ethics is not. Sidgwick understands ancient ethics to be principally concerned with the question of how we should live: ancient ethics tells us that we should make

190  Sukaina Hirji the pursuit of our own supreme good our ultimate end. Ancient theories disagree only in what this supreme good consists in: Indeed it may be said that Egoism in this sense was assumed in the whole ethical controversy of ancient Greece; that is, it was assumed on all sides that a rational individual would make the pursuit of his own good his own supreme aim: the controverted question was whether this Good was rightly conceived as Pleasure or Virtue, or any tertium quid. (ME 91–92; see Santas 1996; Brink 2018) Though Sidgwick is sympathetic to the idea that promoting one’s own happiness is a central demand of practical reason, he is unpersuaded by the sort of eudaimonism that he takes Aristotle to defend. As he understands it, Aristotle eschews a hedonistic account of eudaimonia, and instead locates eudaimonia in virtuous activity. However, Aristotle also insists that the best sorts of pleasures will accompany virtuous activity such that even the hedonist should prefer a life of virtuous activity. Sidgwick suggests that a view like Aristotle’s faces a kind of dilemma depending on how it understands the virtuous activity constitutive of eudaimonia. On the one hand, the view might tell us that happiness lies in acting in accordance with practical reason without filling in the content of the reasons we have. In this case, the results are trivial; the view tells us nothing substantive about how we ought to act if we are to achieve the sort of happiness we desire: If, however, we ask how we are to ascertain the kind of conduct which is properly to be called Virtuous, it does not seem that Plato can tell us more of each virtue in turn than that it consists in (1) the knowledge of what is Good in certain circumstances and relations, and (2) such a harmony of the different elements of man’s appetitive nature, that their resultant impulse may be always in accordance with this knowledge. But it is just this knowledge (or at least its principles and method) that we are expecting him to give us: and to explain to us instead the different exigencies under which we need it, in no way satisfies our expectation. (ME 376) On the other hand, eudaimonism might be understood as offering a substantive conception of the virtuous activity that is constitutive of eudaimonia, where virtue is understood in terms of something like the common-sense morality of the time. Here, Sidgwick argues, the ethical theory looks viciously circular. This is because, as Sidgwick argues, the rules of common-sense morality are subordinate to considerations of expediency. As he argues, when we are deciding how to apply rules or

Aristotelian Eudaimonism and the Dualism  191 specify their limits and relations, the only rational option—and the one that common-sense appeals to—is utility. As such, any eudaimonism that holds that the human good is conforming with common-sense morality is problematic because, as he takes himself to have shown, common-sense morality is itself subordinate to an independent notion of the good. When, however, we examine the controversies to which this question originally led, we see that the investigation which has brought us round to it has tended definitely to exclude one of the answers which early moral reflection was disposed to give it. For to say that ‘General Good’ consists solely in general Virtue,—if we mean by Virtue conformity to such prescriptions and prohibitions as make up the main part of the morality of Common Sense—would obviously involve us in a logical circle; since we have seen that the exact determination of these prescriptions and prohibitions must depend on the definition of this General Good. (ME 392) Even beyond the circularity, on this horn, Aristotle owes us an explanation of why this sort of happiness is one we have independent reason to care about. Why should we think that living a life of justice, temperance, and courage will guarantee us the greatest pleasure or satisfaction of our desires? Sidgwick treats as implausible the idea that the whole of the human good could consist in obedience to moral rules; happiness, he suggests, must also consist at least partly in agreeable consciousness (ME 391). Ultimately, Sidgwick worries that Aristotle’s substantive account of eudaimonia obscures two different ends of practical reason, that can, in principle, conflict: on the one hand, the end of our own happiness, and on the other hand, the end of maximizing utility or communal happiness. Prima facie, what benefits an agent on the one hand, and what makes her a virtuous or excellent agent on the other hand, are distinct, and can come into conflict. But, Sidgwick suggests, Aristotle’s conception of an agent’s ultimate end conflates these two different goods (ME 92). Sidgwick concludes that Aristotle’s theory offers little help in reconciling the fundamental duality in practical reason. To sum up: both Kant and Sidgwick are pessimistic about the possibility that ancient eudaimonist theories can resolve the dualism between prudential and moral reasons. In different ways, they each worry that happiness understood as pleasure, or the satisfaction of our desires and inclinations, cannot adequately ground our ethical concerns. And, they worry, more substantive accounts of happiness on which eudaimonia is constituted in part by ethical concerns seem to rely on an implausible, or deeply confused, conception of what human happiness is. Kant ultimately holds that belief in God is morally necessary, in large part due to the

192  Sukaina Hirji dualism between the rational ends of happiness and duty. Because morality cannot function properly without the assumption that God exists, humans are justified in making this assumption on practical grounds (KpV 5:124–132; CJ 5:448–450; Rel 6:3–9). Sidgwick, like Kant, accepts that God’s existence can render practical reason coherent. Moreover, he acknowledges that this might provide sufficient warrant for theism. He was not, however, ultimately himself persuaded and continued to treat the dualism of practical reason as unresolved (ME 503–509).5 Before I proceed, it is worth forestalling a potential worry. One might worry that there is something hopelessly anachronistic about asking how prudential and moral reasons are related on ancient ethical pictures. That is, one might worry that, insofar as ancient ethical theorists do not explicitly draw such a distinction, the question of how self-interest and morality are related is misplaced. In fact, I think this worry is misguided. Although no ancient ethical theorists explicitly draw the distinction between prudential and moral reasons, it seems clear that thinkers like Plato and Aristotle recognized a conceptual distinction between an agent’s self-interest and ethical demands and thought seriously about how the two might be reconciled. Consider Socrates’s conversation with Cleinias in Plato’s Euthydemus. In this dialogue, Socrates questions a young man, Cleinias, on his conception of happiness. Socrates takes for granted, as an assumption shared by all parties in the conversation, that everyone wants to be happy, and that being happy requires having many good things, and things that benefit us (Euthydemus), 278d-282; see Annas 1998, 43–46). Cleinias espouses conventional views about happiness, believing it to consist in or require goods like health, wealth, and political power. Socrates grants Cleinias the idea that happiness consists in the possession and use of good things but argues that no conventional goods are good for an agent unless she possesses virtue, which equips her to use these goods in the correct ways. Socrates, in this conversation, is not assuming some moralized conception of eudaimonia; after all, Cleinias has no prior interest in such a conception of eudaimonia. Rather, Socrates implicitly recognizes the distinction between Cleinias’s conception of his own self-interest, and Cleinias’s common-sense conception of morality as constituted by the traditional Greek virtues like bravery and justice. Socrates takes on the challenge of showing how Cleinias can best promote his own self-interest—best secure the sort of life he cares about—by living a life of ethical virtue. Likewise, in NE, Aristotle begins by laying out the common-sense opinions or endoxa about eudaimonia, and then offers an account of eudaimonia that he believes can best accommodate these endoxa. He is not, from the beginning, assuming some “moralized” conception of eudaimonia with virtue already built in. Rather, that eudaimonia centrally involves virtue is a substantive claim that Aristotle argues for (see White 2002, chs. 1–2, 5–6).

Aristotelian Eudaimonism and the Dualism  193 This is all to say that we cannot simply dismiss the criticisms of Kant and Sidgwick out of hand. Rather, in order to evaluate whether and in what way their criticisms are apt, we need to get clear on how exactly to understand Aristotle’s eudaimonism. Are our reasons to be just or generous explained in terms of the way doing so promotes our own happiness? Or, do we have reasons independent of our own happiness to be just or generous? If so, how do we reconcile these reasons with eudaimonism, the view that we ought to do all things for the sake of our own happiness?

II To determine whether ancient eudaimonist theories can respond to the challenges posed by Kant and Sidgwick, it is necessary to get clear on exactly what these theories were committed to. Eudaimonism is typically understood as the view that a human being ought, in some sense, to live her life so as to promote her own happiness or flourishing. Different forms of eudaimonism differ in how they conceive of eudaimonia as well as in the degree to which they make central the agent’s pursuit of her own eudaimonia, as opposed to the happiness of others. As Beaton and Whiting note, classical forms of eudaimonism tend to be understood as formally egocentric if not substantively egocentric: they tend to assume that each agent should live so as to realize her own eudaimonia, which however is often construed broadly, to include the interest of others (such as family and friends) to whom she stands in certain special relations. (Beaton and Whiting 2013, 1759) It is often thought that ancient ethical theories—to the extent they capture other-regarding demands—do so by arguing that fulfilling these demands is part of living a eudaimon life. I will focus in what follows on attempting to elucidate Aristotle’s version of eudaimonism, both because it is arguably the most fleshed out of ancient theories, and because I think it offers resources to meet at least some of the challenges pressed by Kant and Sidgwick. It is near consensus that Aristotle, like other ancient Greek philosophers, was committed to some form of eudaimonism. But there is significant disagreement about just what form this eudaimonism takes, and how Aristotle intends to explain our reasons for performing just, generous, or courageous actions that benefit others. It is clear from the beginning of the Nicomachean Ethics that eudaimonia is meant to play some central role in Aristotle’s ethical theory. He identifies eudaimonia with the final good, the one end that we choose only for its own sake, and for the sake of which we choose other things. And, he insists, it is important to discover what this good consists in since this knowledge will affect

194  Sukaina Hirji how we live our lives: “like archers having a target we are more likely to hit what is right” (see Irwin 1991, 384). Consider some of the key passages where Aristotle appears to commit to some form of eudaimonism. Aristotle opens the Nicomachean Ethics by observing that every craft, line of inquiry, action, and decision aim at some good, and that the good appears to be what everything aims at.6 In NE 1.2, he argues that, if there is some one good that we desire entirely for its own sake, and for the sake of which we desire other goods, this will be the best good, happiness. Suppose, then, that the things achievable by action have some end that we wish for because of itself, and because of which we wish for the other things, and that we do not choose everything because of something else—for if we do, it will go on without limit, so that desire will prove empty and futile. Clearly this end will be the good, that is to say, the best good. (NE 1.2, 1094a18–22)7 Aristotle resumes the discussion of our final end in NE 1.7. He begins 1.7 by observing that different actions and arts have different ends, and that if we are trying to locate the end of any specific art, we should consider that for the sake of which “the other things” are done. One way to determine the end of any specific craft is to look at whether the actions and decisions of the practitioner of a craft reveal some one goal for the sake of which they are done. So, in the case of medicine, health is that for the sake of which everything in medicine is done, and it is the end and the good of medicine. Likewise, victory is the end and good of generalship, and a house is the end and good of architecture. Aristotle goes on in 1.7 to distinguish between two kinds of ends. Some ends, such as wealth, flutes and instruments in general, are chosen for the sake of something further while some ends are not chosen for the sake of something further but rather are teleion: “complete,” “end-like,” or “perfect.” And, he insists, whatever the chief good is, it must be something that is teleion or, if there is more than one end that is teleion, the most teleion of these: “therefore, if there is only one teleion end, this will be what we are seeking, and if there are more than one, the most teleion of these will be what we are seeking.” He goes on to explain the standard by which goods count as more or less teleion: We say that an end pursued in its own right is more complete teleion than an end pursued because of something else, and that an end that is never choice-worthy because of something else is more complete than ends that are choice-worthy both in their own right and because of this end. Hence an end that is always choice-worthy

Aristotelian Eudaimonism and the Dualism  195 in its own right, never because of something else, is complete without qualification. (NE 1.7, 1097a30–34) Certain sorts of goods, like wealth and flutes, serve as ends of our action, but they are only ever chosen as intermediary ends, on the way to some further end. Other ends, by contrast, are to varying degrees “end-like” or “complete”: they are chosen at least in part for their own sake and not simply for the sake of some further end. And, whatever the best good is, it will be the end that is most “end-like” or “complete.” So, Aristotle argues, if there is only one complete end, this will be the best good, and if there are several, it will be the most complete among them (1097a28– 30). Aristotle goes on to mention several complete ends—goods that are chosen at least in part for their own sake—including wisdom, certain pleasures, and honors. The most complete end is one that is always choice-worthy for its own sake, and never for the sake of something else; it is complete without qualification. Aristotle concludes that this best good is eudaimonia, since no one ever chooses it for the sake of other goods, but other goods like wisdom, honors, and pleasures are chosen not just for themselves but also for the sake of happiness: Now happiness, more than anything else, seems complete without qualification. For we always choose it because of itself, never because of something else. Honor, pleasure, understanding, and every virtue we certainly choose because of themselves, since we would choose each of them even if it had no further result; but we also choose them for the sake of happiness, supposing that through them we shall be happy. Happiness by contrast no one chooses for their sake, or for the sake of anything else at all. (NE 1.7, 1097a34-b6) This passage, taken together with NE 1.2, seems to provide clear evidence for a eudaimonist reading of the NE: Aristotle tells us that eudaimonia is always choice-worthy for its own sake and never for the sake of anything else, whereas other goods that we value are choice-worthy at least in part for the sake of eudaimonia. That is, the value of everything else we pursue is explained, at least in part, by the way in which it promotes our own eudaimonia. Further evidence is found in NE 1.12, in the context of Aristotle considering whether happiness is something praiseworthy, or something honorable. He argues that happiness is something honorable, citing the following as a piece of evidence: It seems to be so also from the fact that [eudaimonia] is a first principle; for it is for the sake of this that we all do everything else, and

196  Sukaina Hirji the first principle and cause of goods is, we claim, something prized and divine. NE 1.12 1102a2–4 Despite the apparently strong evidence for a eudaimonist reading of Aristotle, there are difficulties in interpreting the details in a way consistent with Aristotle’s other commitments. The first challenge, as pressed by both Kant and Sidgwick, is to explain exactly how an agent’s concern for her own happiness has her reliably performing morally good actions from morally good motives. After all, there seem to be clear cases where moral and prudential concerns are in tension with each other. In his discussion of courage, Aristotle makes clear that sometimes the courageous agent will sacrifice his life for the sake of a noble cause. How, exactly, are we to understand this choice as consistent with the agent pursuing his own happiness above all else? The second challenge is to accommodate what Aristotle says about other goods in the NE 1.7 passage quoted earlier. Aristotle seems to claim that goods like honor, pleasure and virtue are choice-worthy for the sake of eudaimonia, but, he claims, we would also choose them if they had no further result. That is, these goods seem to be valuable, at least in part, in and of themselves, independently of the way they promote eudaimonia. If that is right, the value of these goods is not exhausted by the way they are related to eudaimonia. But again, on the standard view of eudaimonism, ancient ethical theories tell us that we ought to choose all things for the sake of eudaimonia. This version of eudaimonism seems to be in tension with Aristotle’s claim in NE 1.7 that certain goods like pleasure and honor are choice-worthy at least in part for their own sake, independent of the way they might promote eudaimonia. We will return to this difficulty in more detail in the next section.

III Consider first a cluster of interpretations I will call “strong eudaimonism,” according to which an agent does, or ought to, make all her decisions and actions with a view to promoting her own eudaimonia. Some take Aristotle to be committed to a kind of psychological or descriptive thesis about how human beings act: on this interpretation, Aristotle is a eudaimonist insofar as he thinks that all human beings do in fact pursue eudaimonia as their final end, though they disagree about what this consists in. Typically, commentators defend a qualified version of this interpretation to accommodate cases of akrasia: when an agent is acting rationally, she acts so as to promote her own eudaimonia. This way of thinking about ancient eudaimonism in general was famously defended by Vlastos, who posits what he calls “the Eudaemonist Axiom” which he thinks was initially defended by Socrates and “becomes foundational for virtually all subsequent moralists of classical antiquity.” As Vlastos argues, “the Eudaemonist Axiom” should be understood as the claim

Aristotelian Eudaimonism and the Dualism  197 that “happiness is desired by all human beings as the ultimate end (telos) of all their rational acts” (Vlastos 1991, 293). Irwin takes up this style of interpretation of Aristotle, arguing that what distinguishes humans from non-rational animals is the ability to form rational desires (or boulêsis), and that forming rational desire necessarily involves some conception of a final or ultimate end: “the desire for a final good is part of the human essence and pattern of activity” (Irwin 1980, 45). Irwin insists that even if an agent herself does not consciously recognize some final good that she is aiming at, Aristotle seems to hold that we cannot make intelligible an agent’s choices unless there is, implicit in her patterns of decision and action, some such final end. So, Irwin argues, on Aristotle’s view “the explanatory role of psychic states requires us to ascribe a desire for some overall good, some goal including the agent’s other goals,” even if the agent herself does not avow some such good (Irwin 1980, 47). McDowell defends a similar line, insisting that once we properly understand Aristotle’s concept of an action or praxis, the psychological eudaimonism he defends no longer seems wildly implausible. McDowell argues that the concept of action is narrower than just any voluntary or purposive doing. Specifically, he suggests that for Aristotle, an action is what is undertaken as the result of a decision or prohairesis, where this is understood as “a deliberative desire to do something with a view to doing well”: an action is a praxis when it is undertaken as a constituent means to eudaimonia, rather than merely as instrumental to eudaimonia (McDowell 1980, 361). A consequence of McDowell’s reading is that there can never be any genuine conflict between what an agent should do to promote her own eudaimonia and any other considerations: To embrace a specific conception of eudaimonia is to see the relevant reasons for acting, on occasions when they coexist with considerations that on their own would be reasons for acting otherwise, as, not overriding, but silencing those other considerations—as bringing it about that, in the circumstances, they are not reasons at all. (McDowell 1980, 370) Not everyone thinks Aristotle is committed to this sort of psychological or descriptive thesis, but perhaps the dominant view is that Aristotle is at least committed to a kind of normative eudaimonism: an agent ought to structure her actions and decisions in such a way as to promote her own eudaimonia. Lear, for example, argues that Aristotle thinks we should conceive of eudaimonia as the goal or end (telos) of practical deliberation and action (Lear 2005). On the strongest version of this interpretation, an agent only has reasons to perform an action if it promotes her own eudaimonia. The challenge for these interpretations is to explain how, if an agent ought to perform those actions that promote her own

198  Sukaina Hirji eudaimonia, Aristotle’s ethical theory doesn’t turn out to be an unbridled and objectionable form of egoism. How can a fundamentally egocentric theory capture the other-regarding demands that Aristotle himself seems to recognize? Here, commentators employ different strategies. Irwin, for example, suggests that Aristotle’s discussion of close friendships provides a model for understanding the virtuous agent’s altruistic concerns more broadly. He argues that (1) eudaimonia consists in the full realization of an agent’s rational capacities, (2) an agent is only able to fully realize her rational capacities in the context of close friendships and a good political community, and (3) close friendships and a good political community are only possible when an agent has genuinely altruistic motives in acting for the sake of her friends and community. Although Irwin seeks to justify ethical behavior in light of an agent’s own self-interest, he is careful to avoid attributing to Aristotle an unsophisticated ethical egoism. The virtuous agent cares to develop and preserve close friendships because of the way they are good for her, but a friendship itself involves the virtuous agent seeking to promote the good of her friend for the friend’s own sake because of the qualities that make her good. Irwin suggests that we can extend this framework to understand how the virtuous agent can seek to promote the good of the political community as a whole. The virtuous agent is benefitted by her friendships and her political community, but she only realizes the full benefits in the form of rational activity when she chooses the good of her friends and political community for their own sake. One worry for Irwin’s picture is the limited textual evidence for thinking that Aristotle means to model relationships in the political community along the lines of a close friendship. A second worry is whether such an interpretation can give a coherent account of the virtuous agent’s motivations. Irwin argues that friendship provides us a way to see how self-love and altruism are not necessarily in tension: when we love our friends for their own sake, we make possible relationships that provide us security and assistance, as well as beneficiaries and collaborators in our virtuous actions. Extending this to the political community as a whole, Irwin’s picture seems to be one where we choose to cultivate altruistic concerns for our fellow citizens because of the way the community will ultimately benefit us. The worry here is how to spell out the virtuous agent’s motivations in a way that isn’t either objectionably egoistic—treating her friendships and political community as ultimately instrumental to her own well-being—or as implausibly self-effacing—being motivated by altruistic concerns only by ignoring the self-interested reasons she in fact has. Hursthouse offers a different explanation of the relationship between ethical action and an agent’s own eudaimonia. Specifically, she argues that what makes a character trait virtuous is that it reliably promotes an agent’s own flourishing, and what makes an action virtuous is that

Aristotelian Eudaimonism and the Dualism  199 it is the sort of action that reliably issues from a virtuous character. By defining virtue in terms of eudaimonia, Hursthouse can show how virtuous actions are the sorts of actions that reliably promote an agent’s own eudaimonia. As I argue in more detail elsewhere, the sort of explanation that Hursthouse offers for the goodness of virtuous actions does not square well with what Aristotle says in the NE (Hirji 2018). Specifically, I argue, there is good evidence that Aristotle defines what it is to be a virtuous action in terms of the ends of the action and not in terms of the concept of virtue. As Jennifer Whiting also argues, we see in Aristotle’s detailed account of the virtues in Books 3–5 that “each of the canonical virtues is associated with a certain sort of external result at which stereotypically virtuous actions of the relevant sort typically aim” (Whiting 2002, 280). So, for example, the generous agent that Aristotle describes in NE 4.1 is one who gives in the right amounts to the right people at the right times and does not give to the wrong people, or for the wrong reasons, or at the wrong times. And what sets the standard of correctness for paradigmatically generous actions—what makes the generous agent’s use of wealth appropriate in a particular circumstance—seems to be the way in which the generous agent’s spending benefits others (NE 1120a23, 1120b1–3). Something similar holds in the case of other paradigmatic virtues. The courageous agent is one who experiences the appropriate amount of fear with respect to the right things, with the right aim, in the right way, and at the right time (see Irwin 1985, 131–138; Kraut 1989, ch. 2). And the greatest demonstrations of courage are in battle, where the courageous agent is willing to face the prospect of death in order to ensure the safety and security of the polis. Likewise, the temperate agent is one who experiences the appropriate pleasures and pains with respect to bodily appetites. Specifically, he has a moderate desire for the things that are conducive to health and good condition, and for other pleasant things that are not hindrances to these ends (NE 1119a15–18). What sets the standard of correctness for the temperate agent’s pleasures and pains is in part what is conducive to some external end, in this case the health and good condition of the body. So also in the case of justice, Aristotle claims that paradigmatically just actions are those that aim at a proportionate distribution of benefits and harms necessary to ensure the stability of the polis (NE 1132b31–33a2). As Whiting argues, justice aims at a particular kind of external result, namely, “the state of affairs in which each person has his or her fair share of beneficial and harmful things” (Whiting 2002, 278–279). In the case of each of these virtues, Aristotle’s characterization of the paradigmatically virtuous actions is, in part, in terms of the kinds of external ends at which they aim. We find an even stronger statement of this idea in Book 1 of the Rhetoric, albeit in a passage where Aristotle is reporting the endoxa or common opinions. Here, he claims that virtue is typically understood as being a capacity to “provide and preserve good

200  Sukaina Hirji things,” and “to confer many great benefits, and benefits of all kinds on all occasions” (Rhetoric 1.9, 1366a36–1366b4). From the perspective of both Kant and Sidgwick, these sorts of solutions will not be satisfying. Both of these interpretations raise worries both for how we understand an agent’s motivations when she performs virtuous actions, and for how the goodness or rightness of these actions is explained. At the level of an agent’s motivating reasons, we might worry that the virtuous agent, even if she does the right action, does it for the wrong reasons: even when she is performing just or generous actions, her goal in doing so is ultimately to promote her own eudaimonia.8 Similarly, at the level of an agent’s normative reasons, the explanation for an agent’s reasons for performing good actions, so the objection goes, seems to be counter-intuitively self-regarding. If what makes a character trait count as a virtue is that it contributes to an agent’s own flourishing, and if what makes an action count as good is that it is such as to be performed by someone with a virtuous character, it looks as though, ultimately, what explains the goodness of an action is the way it contributes to an agent’s own flourishing. And, according to this objection, this sort of egoistic explanation is at odds with our ordinary conception of the virtues (see Hurka 2013, 15–16). There is also a worry for these interpretations that the connection between virtue and happiness looks somewhat tenuous. Even if we are convinced that character traits like generosity and justice are typically the most reliable path to flourishing, we can imagine situations, or societies, in which this wouldn’t be the case. What prevents us from imagining individuals without the virtues who still live flourishing lives? And, even if being virtuous benefits us most of the time, why think we are better off being fully virtuous, or virtuous all of the time, rather than selectively so? (see ME 170–175).

IV Partly in the face of these kinds of worries, some commentators have defended interpretations of Aristotle according to which it is not the case that everything an agent does or ought to do is for the sake of promoting her own eudaimonia. These views differ significantly in how they understand what eudaimonia consists in, but they share in common the idea that certain goods or actions can be valuable independent of the way they contribute to an agent’s eudaimonia. Although these views can more easily accommodate non-egocentric reasons for action, they face the challenge of explaining away the passages where Aristotle seems committed to a strong form of eudaimonism and, more generally, the challenge of accommodating potential conflicts between non-egocentric reasons for action and actions given by an agent’s own eudaimonia. Call this cluster of views “weak eudaimonism.”9

Aristotelian Eudaimonism and the Dualism  201 Take the case of ethically virtuous actions. We saw earlier that strong eudaimonism seems to involve an egocentric reading of Aristotle’s ethical theory: what makes a virtuous action choice-worthy for an agent is that it promotes her own eudaimonia. To avoid this sort of egocentrism, a weak eudaimonist reading allows that there are reasons for a virtuous agent to perform a virtuous action independent of the way that it promotes her own eudaimonia. But the weak eudaimonist faces a different challenge: what should the virtuous agent do when her reasons to perform a virtuous action conflict with what promotes her own eudaimonia? The weak eudaimonist needs to explain how eudaimonia plays the kind of structuring and guiding role in practical reason that Aristotle seems to describe. Kraut offers one prominent weak eudaimonist reading. He argues that in fact Aristotle does recognize reasons for choice and action that are entirely unconnected with an agent’s own happiness. He defends a narrow, intellectualist reading of perfect happiness as consisting in contemplation alone. However, he argues that there are other goods that are intrinsically good even if they do not contribute to an agent’s happiness understood as contemplation; these goods include at least some virtuous actions. Moreover, he suggests, there are also goods that are choice-worthy for their own sake for an agent even though they do not benefit the agent herself. These goods include benefitting other people for their own sake even when it comes at some personal cost (Kraut 1989). Kraut attempts to capture the central role of eudaimonia by arguing that, even though there are goods that are choice-worthy independent of the way they causally promote eudaimonia, it is also the case that these goods causally promote eudaimonia. In order to live a life devoted to theoretical contemplation over a long period of time, a philosopher must also have friends and engage in practical action. Contemplation is not an alternative to the other goods but rather the ultimate end that hierarchically orders all subordinate ends, including virtue, friendship, and pleasure. Kraut suggests that, for Aristotle, we ought to pursue noneudaimonic goods, but ideally only to the extent that they also causally promote eudaimonia: “the best amount of lower goods to have, from one’s own point of view, is the amount that fully contributes to one’s ultimate end” (Kraut 1989, 12). The upshot is that, even though we have reasons to perform certain actions that are unconnected with our own happiness, we should only perform these actions so long as they instrumentally promote our happiness, understood as theoretical contemplation.10 For Kraut, the way in which our pursuit of eudaimonia constrains our pursuit of other, independently valuable, goods is meant to capture the guiding and structuring role of eudaimonia in our practical reason. But, as Irwin argues, it is not obvious why we should attribute this “eudaimonist constraint” to Aristotle. If, for Kraut, there are reasons unconnected with our happiness for pursuing ethically virtuous action, why think we

202  Sukaina Hirji should only act on these reasons so long as the actions in question also causally promote our own happiness? After all, although it might be true that engaging in some virtuous actions is helpful for our securing a life devoted primarily to theoretical contemplation, this may not be true of all or most virtuous actions: it is easy enough to imagine just, generous, or courageous actions that detract from a life of contemplation. For Kraut, every good in human life is located somewhere within a hierarchy with a single end—contemplation, in the best life—at its pinnacle. But if some of the goods lower on this hierarchy, including virtuous actions, are valuable in part for reasons independent of the way in which they promote contemplation, why think that contemplation determines the extent to which we should pursue these goods? Like Kraut, Lear defends a “weak eudaimonist” interpretation of Aristotle, according to which certain goods are valuable independent of the way in which they promote eudaimonia.11 However, while Kraut argues that these goods are also, at the same time, “for the sake of” eudaimonia because we pursue them only to the extent they causally promote eudaimonia, Lear argues that these goods are “for the sake of” eudaimonia because they “teleologically approximate” eudaimonia, understood as theoretical contemplation. So, for example, Lear argues that courage and temperance, the virtues of war and peace, are fine and choice-worthy for their own sakes because they reflect a virtuous agent’s commitment to the value of uses of leisure that display one’s rational and political nature. On Lear’s view, what makes courageous or temperate actions fine is that they are oriented toward the most excellent and leisurely use of reason. When the phronimos acts for the sake of the fine or kalon in her action, she shows her appreciation that practical truthfulness is determined by this best use of leisure, and so approximates theoretical truthfulness. When an agent performs a courageous or temperate action, she reveals that the rational use of leisure in contemplation is what makes life worth living. Ethically virtuous activity, for Lear, is “for the sake of” contemplation by being an approximation of it; contemplation sets the standard for excellent, practically rational action. The overall picture suggested by Lear is one on which the value of all goods in the ethical domain, including ethically virtuous actions and other goods that are choice-worthy for their own sakes, is ultimately explained by reference to the supreme value of contemplation. Although goods like acting virtuously are choice-worthy for their own sake, the way in which they are choice-worthy for their own sake is by being an approximation of contemplation, and so choice-worthy for the sake of contemplation. Again, a worry for Kraut’s view is that it risks seeming like an unexplained coincidence that goods choice-worthy for their own sake are also at the same time choice-worthy for the sake of eudaimonia. Kraut allows for some possibility of conflict between what promotes contemplation and what is demanded by ethical virtue, but suggests that,

Aristotelian Eudaimonism and the Dualism  203 in general, the two will align in the virtuous agent’s life. What we want, however, is an explanation for why the demands of ethical virtue, in general, align with what promotes one’s own happiness; otherwise, it looks as though Aristotle is trying to have his cake and eat it too, insisting that we do what promotes our own happiness while assuming that this will also fulfill what ethical virtue demands. Rather than being some unexplained coincidence, on Lear’s view, it is precisely because these goods are related to contemplation in a particular way that they are also, at the same time, choice-worthy in and of themselves. This is in contrast to Kraut’s view where these goods seem to have two entirely distinct and unrelated sources of value. Lear’s interpretation ingeniously resolves some of the worries facing Kraut’s account. However, her view is not without challenges. One worry for her picture is whether there really is strong evidence for the kind of approximation relationship she suggests holds between contemplation and other goods. Lear relies heavily on passages outside of Aristotle’s ethical works, and one would have hoped that, if the approximation relation is indeed an important part of Aristotle’s ethical theory, he would have said as much explicitly. A second worry is whether her account can convincingly explain the value of intrinsic goods ranging from pleasure to honor to friendship in terms of some relationship they bear to contemplation. Again, her view seeks to explain how these goods are, by their nature, hierarchically ordered under contemplation, while also at the same time being choice-worthy for their own sakes. It is clear enough how this is meant to work in the case of ethically virtuous activity, since it approximates theoretical contemplation by involving a grasp of practical truth. It is less clear how other goods can be explained in this way. A third worry is that, on her account, the virtuous agent may not realize that her courageous, just, or temperate actions are, ultimately, for the sake of contemplation. To act virtuously, she must act for the sake of the fine, but she may not fully grasp what makes her action good, namely that it approximates theoretical contemplation. This consequence—that most virtuous agents most of the time lack a full understanding of what makes their actions valuable—has struck some commentators as implausible or unpalatable (Annas 2005). Although the interpretations defended by Kraut and Lear face challenges, notice that they are not vulnerable to the objection, pressed by Sidgwick and Kant, that Aristotle fails to distinguish between two different ends of practical reason. On a weak eudaimonist reading, Aristotle does not think that goods are only choice-worthy for an agent if they causally promote her own eudaimonia. Rather, on the weak eudaimonist picture, there are a variety of goods, including the ends of virtuous actions, that are choice-worthy independent of the way they promote eudaimonia. Moreover, on both Kraut’s and Lear’s interpretations, there is an explanation for how these goods are still “for the sake of”

204  Sukaina Hirji eudaimonia, and how, in general, an agent’s pursuit of these goods is not in tension with her pursuit of eudaimonia.

V Let’s take stock. Aristotle is often taken to be committed to a strong form of eudaimonism according to which an agent ought to do all things for the sake of promoting her own eudaimonia. Both Kant and Sidgwick worry that such a view reduces moral reasons to merely prudential ones. Although some have insisted that these criticisms are hopelessly anachronistic, I argue that they are not so easily dismissed. Both Plato and Aristotle take seriously the challenge of explaining how acting in accordance with the common-sense conception of virtue is in an agent’s own interests. To show that Aristotle is not vulnerable to the objections of Kant and Sidgwick, we need an explanation for how eudaimonia plays a central role in structuring and guiding practical reason without it being the case that it is the sole source of value for other goods. Much of what I have hoped to do in this chapter is simply to highlight the difficulties in interpreting Aristotle’s apparent commitment to eudaimonism. On what I call the “strong eudaimonist” readings of Aristotle, eudaimonia is the source of value for all other goods in the ethical domain. These views run afoul of the egoist worries pressed by Kant and Sidgwick. But they also face interpretive challenges. In both the NE and the Rhetoric, Aristotle seems to explain the goodness of virtuous actions in terms of the good ends at which they aim in the political community, and not in terms of the way in which they promote an agent’s eudaimonia. More generally, Aristotle seems to recognize a number of goods ranging from honor to pleasure to friendship that are valuable independent of the way they promote an agent’s eudaimonia. Strong eudaimonist readings are able to explain why performing virtuous actions promotes an agent’s eudaimonia—on these readings, what makes a virtuous action good is that it promotes an agent’s eudaimonia—but they face serious philosophical and interpretive challenges. On “weak eudaimonist” readings, eudaimonia is not the sole source of value for goods in the ethical domain. The challenge for these views is to make sense of passages where Aristotle seems to describe eudaimonia as that “for the sake of which” all other goods are choice-worthy, and passages where he seems to assign it a central role in structuring the ends of practical reason. Moreover, these views need to explain how performing virtuous actions that are good independent of the way they promote an agent’s eudaimonia still somehow promotes an agent’s eudaimonia; these views need to explain how prudential and moral reasons come to be aligned for Aristotle. Where does all this leave us? My main goal here has been to highlight how difficult it is to make sense of precisely what Aristotle is committed

Aristotelian Eudaimonism and the Dualism  205 to in conceiving of eudaimonia as the ultimate end for human beings. There is at least some reason to doubt that the traditional “strong eudaimonist” reading of the NE accurately reflects his ethical views. In fact, I am inclined to think that “weak eudaimonist” readings along the lines of that defended by Lear are more promising. That is, I am inclined to think that the way in which other goods are choice-worthy for the sake of eudaimonia need not be because they causally promote eudaimonia but rather because they are related in the right sort of metaphysical way to eudaimonia. On this sort of view, eudaimonia is the ultimate end for human beings because it consists in the activities that are the full realization of our rational nature: It is the ultimate end for us in a teleological sense, in the same way that the mature functioning of an animal body is itself an end in the teleological order that Aristotle takes to be a basic, mind-independent feature of the world. Much more would need to be said to show that the “weak eudaimonist” interpretation of Aristotle is preferable to the “strong eudaimonist” reading. One challenge is to show how other intrinsically valuable goods in the ethical domain are “for the sake of” eudaimonia by being teleologically related to eudaimonia rather than by causally promoting it. Lear offers one such way of doing this, appealing to the approximation relation. I have some doubts about whether this is the most fruitful approach, but I want to set this challenge aside. A further challenge for the “weak eudaimonist” reading is to make sense of the relationship between what we might now think of as prudential and moral reasons. If not all our reasons are given by what promotes our own eudaimonia—if we have reason to perform virtuous actions because they aim at ends that are good independent of our eudaimonia—how is it also the case that it is in our own interests to be virtuous? After all, Aristotle clearly thinks that acquiring and practicing virtue is a reliable way to promote our own happiness. Elsewhere, I defend one solution to this challenge (Hirji 2018). I argue that “acting virtuously” is the activity an agent engages in when she performs a virtuous action—an action that is good because of the good ends it aims at independent of an agent’s own happiness—with a full appreciation of the goodness of the action, and with the corresponding motivations. The activity of “acting virtuously” is the way in which human beings express their practically rational nature: It is one way of engaging in the excellent rational activity that is constitutive of eudaimonia. On this solution, it is not virtuous actions themselves that contribute to an agent’s flourishing; again, just, generous, or courageous actions will often be good independent of the way they promote an agent’s eudaimonia. Rather, it is performing these actions with the right reasons and desires—acting justly, courageously, or generously—that is constitutive of an agent’s happiness. The upshot is that this view is not egoistic either in its description of an agent’s motivations or in its explanation of the

206  Sukaina Hirji goodness of virtuous actions. The view does not reduce the goodness of a virtuous agent’s actions to the way they ultimately promote her own eudaimonia. Likewise, the virtuous agent’s motives are not self-effacing, nor does she treat her virtuous action as merely instrumental to achieving her own happiness. If all this is right, Aristotle gives us a way to see how prudential and moral reasons might come to be aligned through a certain conception of human agency: what it is to fully realize our distinctively human capacity for practical rationality is to perform good actions with the right reasons and desires. In a Sidgwickian spirit, Aristotle explains the goodness of actions in terms of the way they promote happiness in a community. And, in a Kantian spirit, he explains how prudential and moral reasons are aligned by appealing to the supreme value of rationality in human life. Of course, there is a final worry pressed by Kant and Sidgwick that I have not attempted to directly address, namely, that Aristotle offers an implausible conception of human happiness. Even if we don’t think Aristotle identifies eudaimonia with virtuous activity, we might worry that his conception of happiness as excellent rational activity is far removed from our ordinary use of the term, or the sort of happiness understood by Kant and Sidgwick as agreeable states of consciousness. An initial response is to notice that Aristotle is better understood as providing an account of well-being—an account of what it means for a human to do well or flourish—rather than an account of happiness as it is commonly understood. Specifically, he is defending a perfectionist account on which well-being consists in the development and exercise of the capacities that are central to our nature. But Aristotle also takes it upon himself to argue that eudaimonia as he understands it is indeed the most pleasant life for a human being; it is the most pleasant life because, in the first place, it involves the most valuable activities and these are, by nature, most pleasant. Moreover, it is the life that allows us to best satisfy our desires, insofar as developing a virtuous character helps to harmonize and tame our otherwise conflicting and unruly desires. The disagreement then between ancient eudaimonists and Kant and Sidgwick is not only in how to understand the relationship between an individual’s moral and prudential reasons but, more fundamentally, in the nature of that prudential good itself.

Notes 1. In opposing heteronomous conceptions of morality, Kant is especially concerned to reject the Epicurean idea that the motive for living virtuously lies in the instrumental value of such living for securing pleasure. 2. Kant’s works are abbreviated herein as follows: KpV = Critique of Practical Reason, G = The Groundwork of the Metaphysics of Morals (Kant 1996); CJ = Critique of Judgment (Kant 1987); Rel = Religion within the Boundaries of Mere Reason (Kant 1998).

Aristotelian Eudaimonism and the Dualism  207 3. Moreover, Kant suggests that Aristotle's explanation of an agent's motivating reasons is ultimately incoherent: As he understands Aristotelian eudaimonism, the virtuous agent is supposed to enjoy a particular kind of pleasure or satisfaction from acting from duty. But, he worries, if the virtuous agent is acting so as to secure this pleasure, then she is failing to act from duty. 4. “It is, however, possible to take a view of virtuous action in which, though the validity of moral intuitions is not disputed, this notion of rule or dictate is at any rate only latent or implicit, the moral ideal being presented as attractive rather than imperative. Such a view seems to be taken when the action to which we are morally prompted, or the quality of character manifested in it, is judged to be ‘good’ in itself (and not merely as a means to some ulterior Good)” (ME 105–106). 5. For more on Kant’s and Sidgwick’s respective views on the moral necessity of God, see Tyler Paytas’s chapter in this volume. 6. This passage is sometimes read as an instance of fallacious reasoning: If all choice and action aims at some end, there must be some one end at which all actions and choice aim. See, for example, Anscombe (1963, 34), Geach (1972, 2), Ackrill (1974, 350). On this reading, Aristotle asks us to suppose there is some one end that we choose for itself, and for the sake of which other things are chosen, and then argues there must be some such good or else we would choose everything for the sake of something else, and desire would turn out to be pointless. But, as a number of commentators have argued, this reading of the passage is not only implausible but also not well supported by the grammar of the sentence. 7. Translations from the NE taken from Irwin, Hackett, 2nd ed. 8. Some commentators, in the face of this kind of worry, have argued that Aristotle endorses a kind of self-effacing theory. See, for example, Annas (2008, 205–221, esp. 212). Even if the ultimate explanation for why an agent ought to perform a just or generous action is that it will promote her own eudaimonia, this explanation need not figure in her conscious deliberation. It is unclear however, how this strategy can be reconciled with the way Aristotle describes the agent who possesses practical reason as possessing a grasp of the orthos logos or correct account of actions. Aristotle seems to describe the virtuous agent as someone who correctly grasps the reasons she has. 9. I set aside for the purposes of this discussion the debate about whether Aristotle's conception of eudaimonia is best understood as “inclusivist”— consisting in a set of intrinsically valuable goods—or “intellectualist”—consisting solely in rational activity. However, there is some overlap in these debates. In particular, “weak eudaimonist” interpretations naturally align with “intellectualist” views. It is easier to see, after all, how various intrinsically valuable goods can be choice-worthy for the sake of eudaimonia on an inclusivist view where these goods are "for the sake of" eudaimonia by being parts of a whole 10. See, also, Broadie (1991, 31–32): “Thus the central good functions sometimes as a constraint rather than a goal in the ordinary sense of a positively aimed for objective. I would stop doing what might adversely affect it, even if I was not doing that thing in order to obtain it . . . the position does not imply that we can never, for instance, admire, delight in, love, take an interested in, something else just for what it is . . . Aristotle only means that when we take practical steps towards any object, we should do so having regard to the central good.” 11. Lear herself insists that her theory, unlike Kraut's, is “eudaimonistic” in the sense that every action is, in some way, justified by reference to an agent's

208  Sukaina Hirji own happiness. This is because, on her view, the intrinsic value of goods other than happiness is not independent of the value these goods have by being for the sake of eudaimonia. Still, I count her as a weak eudaimonist here insofar as she thinks that various goods like ethical virtue are valuable even if they do not causally promote eudaimonia.

References Ackrill, J. L. 1974. “Aristotle on Eudaimonia.” In Essays on Aristotle’s Ethics, edited by A. O. Rorty, 15–34. Berkeley: University of California Press. Annas, J. 1998. “Virtue and Eudaimonism.” In Virtue and Vice, edited by E. F. Paul, F. D. Miller, and J. Paul, 37–55. Cambridge: Cambridge University Press. Annas, J. 2005. “Comments on John Doris’s Lack of Character.” Philosophy and Phenomenological Research LXXI (3): 636–642. Annas, J. 2008. “Virtue Ethics and the Charge of Egoism.” In Morality and SelfInterest, edited by Paul Bloomfield, 205–221. Oxford: Oxford University Press. Anscombe, G. E. M. 1963. Intention. 2nd ed. Oxford: Blackwell. Aristotle. 1999. Nicomachean Ethics. Translated by Terence Irwin. Indianapolis: Hackett. Beaton, R. S., and J. Whiting. 2013. “Eudaimonism.” In International Encyclopedia of Ethics, edited by H. Lafollette. doi:10.1002/9781444367072.wbiee292. Brink, D. 2018. “Eudaimonism and Cosmopolitan Concern.” In Virtue, Happiness, Knowledge Themes from the Work of Gail Fine and Terence Irwin, edited by D. O. Brink, S. S. Meyer, and C. Shields, 270–292. Oxford: Oxford University Press. Broadie, S. 1991. Ethics with Aristotle. New York: Oxford University Press. Engstrom, S. 1998. “Happiness and the Highest Good in Aristotle and Kant.” In Aristotle, Kant, and the Stoics, edited by S. Engstrom and J. Whiting, 102–138. Cambridge: Cambridge University Press. Geach, P. T. 1972. Logic Matters. Berkeley: University of California Press. Hirji, S. 2018. “Acting Virtuously as an End in Aristotle’s Nicomachean Ethics.” British Journal for the History of Philosophy 26 (6): 1006–1026. Hurka, T. 2013. “Aristotle on Virtue: Wrong, Wrong, and Wrong.” In Aristotelian Ethics in Contemporary Perspective, edited by J. Peters, 9–26. New York: Routledge. Irwin, T. H. 1980. “The Metaphysical and Psychological Basis of Aristotle’s Ethics.” In Essays on Aristotle’s Ethics, edited by A. O. Rorty, 35–53. Berkeley: University of California Press. Irwin, T. H. 1985. “Aristotle’s Conception of Morality.” Proceedings of the Boston Area Colloquium in Ancient Philosophy 1 (1): 115–143. Irwin, T. H. 1991. “The Structure of Aristotelian Happiness: Aristotle on the Human Good. Review of R. Kraut.” Ethics 101 (2): 382–391. Kant, I. 1987. Critique of Judgment. Translated by W. S. Pluhar. Indianapolis: Hackett Publishing Company. Kant, I. 1996. Practical Philosophy. Edited and translated by M. Gregor. Cambridge: Cambridge University Press. Kant, I. 1998. Religion Within the Boundaries of Mere Reason. Edited and translated by A. Wood and G. Di Giovanni. Cambridge: Cambridge University Press.

Aristotelian Eudaimonism and the Dualism  209 Kraut, R. 1989. Aristotle and the Human Good. Princeton: Princeton University Press. Lear, G. R. 2005. Happy Lives and the Highest Good: An Essay on Aristotle’s “Nicomachean Ethics.” Princeton: Princeton University Press. McDowell, J. 1980. The Role of Eudaimonia in Aristotle’s Ethics’. In Essays on Aristotle’s Ethics, edited by A. O. Rorty, 359–376. Berkeley: University of California Press. Santas, G. 1996. “Goodness: Aristotle and the Moderns. A Sketch.” Philosophical Inquiry 18 (1–2): 43–60. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Vlastos, G. 1991. Socrates: Ironist and Moral Philosopher. Cambridge: Cambridge University Press. White, N. 2002. Individual and Conflict in Greek Ethics. Oxford: Oxford University Press. Whiting, J. 2002. “Eudaimonia, External Results, and Choosing Virtuous Actions for Themselves.” Philosophy and Phenomenological Research 65 (2): 270–290. Wood, A. W. 2018. “Virtue: Aristotle and Kant.” In Virtue, Happiness, Knowledge Themes from the Work of Gail Fine and Terence Irwin, edited by D. O. Brink, S. S. Meyer, and C. Shields, 234–252. Oxford: Oxford University Press.

10 Beneficent Governor of the Cosmos Kant and Sidgwick on the Moral Necessity of God Tyler Paytas The wages of sin is death: if the wages of Virtue be dust, Would she have heart to endure for the life of the worm and the fly? She desires no isles of the blest, no quiet seats of the just, To rest in a golden grove, or to bask in a summer sky; Give her the wages of going on, and not to die. Alfred Tennyson, ‘Wages’ And do not let it matter to you for how long you will be alive in this work: even three hours spent thus are sufficient. Marcus Aurelius, Meditations 6.23

Kant and Sidgwick conduct their respective ethical inquiries in light of two shared beliefs about the relationship between God and morality. The first is that God’s commands cannot be the source of genuine ethical principles. Kant emphasizes the point that true moral principles cannot be grounded in the arbitrary will of an external authority because this would be mere heteronomy, as the normative force of ethical commands would be contingent upon each individual’s fears and desires (Gr 4:443; CpV 5:127–129; ET 8:339).1 Sidgwick rejects divine command ethics partly on the grounds that the very notion of divine justice presupposes an independent moral standard by which God would issue just punishments and rewards (ME 31, 505; 1891, 92; see, also, Gr 4:408–409; CJ 5:444). These points are related to a second shared assumption, which is that we can obtain ethical knowledge through the use of reason without relying on theistic conjectures or religious dogma (Gr 4:389, 4:411; ME 5–6, 79–80; 1891, 92). Yet, despite sharing these views about morality’s independence from theology, both Kant and Sidgwick ultimately conclude that the assumption of God’s existence is necessary for the complete viability of practical reason (including principles of morality) within human beings (CpV 5:124–132; CJ 5:448–450; Rel 6:3–9; ME 503–509; Sidgwick and

Beneficent Governor of the Cosmos  211 Sidgwick 1906, 605). Although the details of their respective positions differ, the basic thought is that only divine justice could secure an ostensibly crucial connection between dutiful conduct and the happiness that such conduct merits. Without belief in such a connection, a commitment to morality will likely seem futile and possibly even irrational. While Kant believes that these considerations provide a type of warrant for theism, Sidgwick remains skeptical, lamenting that without independent evidence of God, the enterprise of philosophical ethics seems to have been “foredoomed to inevitable failure” (ME1 473). The parallel reintroduction of God into the ethics of Kant and Sidgwick is striking not only because of their secular presuppositions (with respect to the grounds and authority of morality) but also because of their championing of divergent ethical frameworks (consequentialism and deontology). The central claim of this chapter is that, despite their philosophical differences, Kant’s and Sidgwick’s respective appeals to the moral necessity of God have a common cause. Both philosophers perceive a deep human need for belief in cosmic justice (i.e., universal harmony between duty and happiness), and their shared hedonic account of well-being colors their conception of what such justice must involve. After explaining how these ideas give rise to the ostensible moral need for faith, I provide a brief defense of an alternative view based on the Stoic belief that the universe is sufficiently just and good, despite the lack of a reliable connection between moral virtue and hedonic well-being.

Kant on the Summum Bonum Although Kant insists that moral principles cannot be sourced in God, he does believe that morality inevitably leads to theism. The first extensive treatment of this idea occurs in the Critique of Practical Reason, where Kant elaborates on his notion of the “highest good,” which was originally introduced in the Canon of Pure Reason in the first Critique.2 Near the start of the Dialectic of Pure Practical Reason, Kant makes note of an analogy between speculative reason and practical reason. Just as pure reason in its speculative use seeks a totality of conditions for appearances (i.e., the unconditioned cause that brings an end to the apparent regress of cause and effect), pure reason in its practical use “likewise seeks the unconditioned for the practically conditioned (which rests on inclinations and natural needs)” (CpV 5:108). In other words, pure practical reason seeks the unconditioned totality of its object. An object of practical reason is an effect possible through freedom (CpV 5:57). We act freely when we act on universalizable maxims, and so an object of practical reason is an effect that we intend to bring about through our maxims. Hence, the key claim at the start of the Dialectic amounts to the idea that pure practical reason seeks an unconditioned totality of the intended effects of adopted maxims. That is to say, human practical reason inevitably seeks

212  Tyler Paytas an ultimate end to which all of our actions (viewed as a whole) are to be directed. Kant adds that this unconditioned totality is sought under the name of the highest good (CpV 5:108). In chapter 2 of the Dialectic, Kant claims that while moral virtue is the supreme good, it is not the complete good because for this happiness is also required. Hence, the highest good of the individual consists of virtue combined with happiness, and the highest good of a possible world (i.e., the whole or complete good) consists of happiness distributed in exact proportion to morality (CpV 5:110).3 Kant does not fully clarify why happiness is required as a necessary part of the highest good. He notes that for a virtuous person to need happiness but not receive it “cannot be consistent with the perfect volition of a rational being that would at the same time have all power” (CpV 5:110). However, there is no explicit explanation for why the perfect volition of an omnipotent rational being is relevant for determining the content of the highest good (we will return to these points presently). After articulating the notion of the highest good and its contents, Kant notes an apparent difficulty. The highest good combines two distinct elements (happiness and morality) that are represented as standing in a necessary relationship. Such a necessary connection must either be analytic (i.e., logical) or synthetic (i.e., causal). However, there is clearly no logical connection between morality and happiness; and experience makes it clear that neither is there a (necessary) causal connection. This is brought out most clearly by the common occurrence of the moral law requiring conduct that is contrary to one’s own happiness (e.g., refusing to give false testimony even under threat of severe punishment (CpV 5:30)). And even if an agent were somehow fortunate enough to avoid frequent conflicts between duty and self-interest, everyone is vulnerable to the natural ills of pain, disease, aging, and death. Hence, even the morally best among us will never receive all of the happiness that they deserve.4 The apparent lack of a reliable connection between duty and happiness is problematic because promotion of the highest good is necessitated by morality. If happiness is not the effect of our right conduct, then morality’s command to promote the highest good would be defective, which means that the authority of morality itself would be called into question. In Kant’s words, “If . . . the highest good is impossible in accordance with practical rules, then the moral law, which commands us to promote it, must be fantastic and directed to empty imaginary ends and must therefore in itself be false” (CpV 5:114). The fact that the moral law commands promotion of the highest good, combined with the absence of grounds for believing that virtue will ever result in proportionate happiness constitutes an “antinomy of practical reason.” Kant claims that in order to resolve this antinomy we must assume our own immortality and the existence of God as postulates of pure practical reason. The assumption of immortality is required because complete

Beneficent Governor of the Cosmos  213 conformity of dispositions with the moral law is the supreme condition of the highest good, and such conformity is impossible for human beings during their earthly existence (because finite creatures are always vulnerable to inclinations). Hence, the moral component of the highest good must be realized in an endless progress that requires the continuation of one’s personality beyond physical death (CpV 5:122). And the assumption of God is required to ensure that virtue always leads to happiness (through divine rewards), something that would otherwise be problematic given the contingency of the relationship between the two in the sensible world (CpV 5:124). These striking claims immediately give rise to several questions: (1) How could promotion of an ultimate end be commanded by a moral law that supposedly takes no account of consequences? (2) If we do in fact need to posit an ultimate end, why can’t the attainment of virtue itself (i.e., without corresponding happiness) play this role? (3) If happiness is a necessary element of the ultimate end, why could the goal not be to bring about as much deserved happiness as humanly possible, in which case we would not need to postulate entities outside of our experience? (4) What exactly is the doxastic state (e.g., knowledge, belief, hope, faith) with respect to the question of God that arises through the idea of the highest good? Although Kant devotes several sections of the Dialectic to the highest good, answers to these questions are difficult to extract. Aware of the need to provide further clarification, Kant offers new discussions of these matters in a handful of later works including Critique of Judgment (1790), “Theory and Practice” (1793), and Religion within the Boundaries of Mere Reason (1794). Let us now turn to these texts in order to get a clearer picture of Kant’s position.5

The Highest Good in Later Works The Need for an Ultimate End Our first interpretive question is why the promotion of an ultimate end is commanded by the moral law in the first place. This issue is especially perplexing given Kant’s repeated emphasis of the point that genuine moral principles are formal and unconditional, and hence do not make any reference to ends or desired effects. As we have seen, in the second Critique, Kant notes that practical reason seeks a final object (i.e., ultimate end) that will bring its principles into systematic unity. There is a clear parallel between this need of pure practical reason and the need of pure theoretical reason to find an unconditioned condition. What’s puzzling is that, in the first Critique, Kant offers a negative take on reason’s desire for ultimate unity and synthesis, explaining how this leads us to confusion and error. Yet, in the Dialectic of the second Critique, Kant

214  Tyler Paytas adopts a positive stance toward practical reason’s quest for the unconditioned and argues that this need of reason provides independent warrant for postulating ideas such as those of God and immortality. Although Kant stresses that these ideas do not expand theoretical cognition, they are given “objective reality” through the needs of pure practical reason (CpV 5:132). But if the lesson of the first Critique was the importance of avoiding the illusions arising from speculative reason overstepping its boundaries through its inherent desire for synthesis and ultimate explanation, should we not proceed with the same caution with respect to the inherent desires of practical reason?6 A potential key to resolving this puzzle is found in the preface to the 1st edition of Religion, where Kant explicitly aims to clarify the connection between morality and belief in God. After reaffirming that we can recognize and fulfill our moral obligations without aiming at an end that precedes determination of the will, Kant concedes that, nonetheless, morality does necessitate representation of an end as the necessary consequence of morally obligatory maxims. This is not because the moral law itself is inherently concerned with consequences but rather because determination of the will is not possible without an intended effect. Although this intended effect is not the determining ground (i.e., incentive or motivating reason) of the action, representation of an effect is necessary for genuine action (Rel 6:4). Since every action is followed by an effect, we must be able to represent and endorse (at least on some level) the likely effect in order to genuinely act. Kant writes: [W]ithout this end, a power of choice which does not [thus] add to a contemplated action the thought of either an objectively or subjectively determined object (which it does or should have), instructed indeed as to how to operate but not as to whither, can itself obtain no satisfaction. (Rel 6:4; see, also, TP 8:279) The basic idea is that, just as we cannot rationally will an end without also willing the necessary means, neither can we rationally will to act without directing our action toward a specified end that reason can justify (Pasternack 2014, 69). And since the laws of morality command action, human adherence to the moral law requires the representation of an end. This is a reasonable explanation for why we must represent some particular end for each of our right actions, despite the fact that the moral law itself makes no reference to consequences. But the question remains as to why we are morally bound to represent a single ultimate end for the whole of our conduct. Why is it not enough for each right action to have its own particular end (e.g., to relieve the suffering of others through aid, to protect someone’s privacy by refusing to divulge confidential

Beneficent Governor of the Cosmos  215 information, etc.)? Kant’s answer in Religion is simply that the representation of an ultimate end fills a “natural need” in human beings (Rel 6:5). This hearkens back to the claims in the second Critique about pure practical reason’s quest for the unconditioned condition (i.e., final end) for all ends. The highest good fills the role of final end by providing a special point of reference for the unification of all ends . . . for only in this way can an objective practical reality be given to the combination, which we simply cannot do without, of the purposiveness [deriving from freedom] and the purposiveness of nature. (Rel 6:5, emphasis added) The suggestion is that the very nature of human practical reason is such that it will inevitably seek an ultimate end that will provide rational order to all of its ends. Since we cannot act without representing an end, and since we have several different ends (both moral and non-moral), we feel compelled to bring order to our practical lives by representing a final end toward which all of our conduct (insofar as we act rationally) is ultimately directed.7 Supposing it is true that human beings by nature feel some sort of impetus to set a single ultimate goal to work toward, we might still wonder in what sense this constitutes a need of practical reason. If the idea of the highest good is to provide a type of warrant for theism (which theoretical reason ostensibly cannot provide), then it must be grounded in a genuine need of high importance—something deeper than a mere inclination or impulse. One way of gaining insight into the nature of this need is to ask what would happen if we were unable to meet it. Supposing it were impossible to represent an ultimate end justified by reason as the effect of our conduct, what exactly would be the problem? Some of Kant’s remarks suggest an answer to this question that centers on the importance of maintaining one’s commitment to the moral law. Kant claims that were it not for our ability to believe that our moral conduct will bring about the highest good, our natural need to represent an ultimate end justified by reason would be a “hindrance to moral resolve” (Rel 6:5). He elaborates on this point in a key footnote shortly thereafter: Now in this end human beings seek something that they can love, even though it is being proposed to them through reason alone. Hence the law that only inspires respect in them, though it does not recognize this sought-after something as [its own] need, nonetheless extends itself on its behalf to include the moral ultimate end of reason among its determining grounds. (Rel 6:7) Given Kant’s frequent insistence that the genuine moral incentive is divorced from sensible feeling, this appeal to love is rather surprising.

216  Tyler Paytas If we are fully capable of acting from the motive of duty, why must we represent an object that we can love? Kant sheds some light on the source of this need in an essay published contemporaneously with the 2nd edition of Religion titled “The End of All Things”: Respect is without doubt what is primary, because without it no true love can occur, even though one can harbor great respect for a person without love. But if it is a matter not merely of the representation of duty but also of following duty . . . then love, as a free assumption of the will of another into one’s maxims, is an indispensable complement to the imperfection of human nature (of having to be necessitated to that which reason prescribes through the law). For what one does not do with liking he does in such a niggardly fashion—also probably with sophistical evasions from the command of duty—that the latter as incentive, without the contribution of the former, is not very much to be counted on. (ET 8:338) Two parts of this passage are especially important for present purposes. The first is the idea that what is of primary concern is not the mere representation of duty but rather actually following it. The second is the concluding remark that without the supplement of love, duty as the incentive for human beings is not very much to be counted on. The suggestion seems to be that, since what matters ultimately is not just the mere knowledge of our duties but actually fulfilling them, we ought to take whatever precautions are necessary to avoid succumbing to our natural tendency to shirk from our obligations. Given the various flaws of human nature, it is crucial that we cultivate a disposition to fulfill our duties gladly (CpV 5:83). These points are relevant for explaining the nature of our need for an ultimate end that we can love because representing such an end is one way of safeguarding ourselves against the weakening of our moral resolve. There are (at least) two respects in which our moral resolve is vulnerable. First, as finite beings with an animal nature we are always susceptible to the influence of inclinations that tempt us to violate the moral law. Second, there are many instances in which fulfilling one’s duty appears to have overall bad effects. For instance, dutifully refusing to sacrifice an innocent person’s life during some crisis may result in the deaths of numerous other innocent people. Moreover, even if all human beings were to become perfectly virtuous, we would still live in a world full of immense suffering and loss due to the pernicious forces of nature (CJ 5:452). Given all this, it is quite easy to see how an otherwise well-disposed and strong-willed person might eventually be overcome by despair and a weakening of moral resolve. Hence, the need to represent

Beneficent Governor of the Cosmos  217 an ultimate end justified by reason as an object to love is a moral need. Since anyone with a genuine moral disposition will be concerned about protecting that disposition from the looming threats of hopelessness and anguish, we have a moral interest in representing a final end which could justify the existence of the world and help us to avoid viewing our moral efforts as ultimately futile. Here one might object that these ideas impute human beings with a type of moral weakness that is damning to the Kantian system of ethics, which emphasizes autonomy and action purely from respect for the moral law (without regard for consequences).8 Indeed, even in the key passages from Religion, Kant acknowledges that duty should be a sufficient incentive for human beings regardless of its external effects: All human beings could sufficiently partake of this incentive [i.e., duty] too if they just adhered (as they should) to the rule of pure reason in the law. What need have they to know of the outcome of their doings and nondoings that the world’s course will bring about? It suffices for them that they do their duty, even if everything were to end with life in this world, and in this life too happiness and desert perhaps never converge. (Rel 6:7) In principle, all rational agents are capable of obeying the law without any regard for consequences. Yet, human beings are hindered by their animal nature that forces them to view the world through the lens of cause and effect. Kant continues: Yet it is one of the inescapable limitations of human beings and their practical faculty of reason (perhaps of that faculty in all other worldly beings as well) to be concerned in every action with its result, seeking something in it that might serve them as an end and even prove the purity of their intention . . . (Rel 6:7) The claim is not that we must represent the highest good as our ultimate aim because it is technically impossible for us always to act strictly from the motive of duty. Such a claim would indeed undermine the account of moral motivation that is central to Kant’s ethics. The claim is, rather, that since we know we are vulnerable to existential despair and diminished moral motivation in the face of cosmic injustice and interminable natural ills, we must take whatever steps that are necessary to prevent this from happening. This holds even for those who might possess an especially strong moral disposition. It is true that any particular individual might be so heroic as to not need a lovable final end in order to reliably carry

218  Tyler Paytas out her duties throughout the duration of her life. But even such a hero wouldn’t be able to know with certainty that she has such remarkable resolve. And so even the strongest of human beings ought to represent a final end justified by reason as the goal of their moral conduct. This will not guarantee that the agent will always prioritize duty over inclination, but it does decrease the chances of being overcome by despair and subsequently losing moral resolve. Kant refers to the need to make the highest good the object of one’s will so as to promote it with all of one’s powers as a duty (CpV 5:142). This cannot be a direct duty because the moral law itself makes no reference to an ultimate end. Hence, the duty to represent the highest good as one’s final end must be an indirect duty; it is an essential step that we must take in order to facilitate the fulfillment of our direct duties. The reasoning is similar in the case of other examples of indirect duties such as promoting one’s own happiness to prevent temptations arising from unhappiness (Gr 4:399), as well as the duty to cultivate and utilize sympathy in order to more effectively fulfill duties of beneficence (MM 6:457).9 In each case, we have an indirect duty to take whatever (permissible) steps are necessary to make us more adept and reliable in fulfilling the duties arising directly from the law of pure practical reason. The Necessity of Happiness We have seen that the answer to our first interpretive question, that of why we need an ultimate end in the first place, is that representing such an end is a necessary precautionary measure for protecting our moral resolve. Our next question asks why the attainment of virtue itself (or at least as much moral perfection as humanly possible) could not serve as our ultimate goal. Part of the answer to this question is that our animal nature forces us to aim at happiness. Given the sort of creatures we are, we simply cannot jettison happiness from our set of ends (CJ 5:451). Of course, we must be capable of self-sacrifice in order for morality to even be possible (CpV 5:30). But even when our wills are good and we prioritize the law over our own happiness, the latter never completely disappears from our conception of what is good and worthy of promoting.10 As Kant stresses repeatedly, this view of the importance of happiness as an end does not arise from mere self-interest but rather from a purely impartial perspective. Recall Kant’s claim that an omnipotent, perfectly rational being could never will a state of affairs in which a person has made herself deserving of happiness and yet does not partake of it (CpV 5:110). He makes a similar claim in Religion, noting that any rational individual who considers the possibility of the highest good would will for it to be actualized even knowing that they might thereby forfeit their own happiness because they are not adequate to this moral ideal (Rel 6:5–6). The point is that, as long as it is deserved, happiness is objectively

Beneficent Governor of the Cosmos  219 good. This is important because the ultimate end must be something that is objectively (and completely) good. For Kant, whatever we call “good” in this sense must be “an object of the faculty of desire in the judgment of every reasonable human being” (CpV 5:61).11 A world in which human beings make themselves deserving of happiness and yet still end up afflicted by the evils of poverty, illness, and death cannot be an object of desire in the judgement of a reasonable person. Of course, such a world would contain all the value inhering in the good wills of those people. But this only serves to make the numerous ills that will inevitably befall them all the more lamentable. Hence, if we are to mitigate against the existential despair that threatens to hinder moral resolve by representing an ultimate rational end, this end must include not only moral excellence but also corresponding happiness. These considerations are relevant to our third interpretive question, which asks why the ultimate goal cannot simply be to attain as much deserved happiness as humanly possible, thus eliminating the need to postulate a divine being.12 This question is answered by the simple fact that a world in which the virtuous are deprived of the happiness that they need and deserve is not objectively good on the whole (even though virtue itself is the supreme good). Hence, even though we would have strong reasons to aim for deserved happiness to the best of our abilities in a godless universe, without assuming a reliable connection between morality and happiness, we will continue to be at risk of diminished moral resolve.13 Moral Faith This leads to our final interpretive question, which concerns the doxastic state with respect to the question of god that arises from representing the highest good as our ultimate end. Kant is careful to note that his moral argument for God does not extend theoretical cognition, meaning that we do not obtain knowledge of God’s existence through these ethical considerations. But if the moral law commands us to aim for the highest good, and the highest good is only possible through God, it would seem that we should be at least as confident in God’s existence as we are in the validity of the moral law. And as Kant stresses, the validity of the moral law is not open to reasonable doubt—it is apodictically certain (CpV 5:47). So why does the doctrine of the highest good not lead to genuine knowledge of God’s existence? And if the resulting doxastic state is not knowledge, what exactly is it? In the second Critique, Kant claims that the ideal of the highest good leads to Glaube (“belief” or “faith”) with respect to God.14 Of course, it is not possible to have a duty to believe a given proposition because belief is not sufficiently under our control. But what is under our control are the goals that we choose to set for ourselves. Since the goal of the highest

220  Tyler Paytas good is not incoherent, we are free to set its realization as our ultimate end. And as we have seen, we have an indirect duty to do so based on a moral need (the need to maintain an optimistic stance toward the world and the effects of our moral efforts). In fulfilling this indirect duty, we will resolve to work toward the highest good through our right conduct. We need not think about the highest good every time we act, but in the moments when we stop to reflect on our ultimate purpose, which should not be infrequent, we will envision the highest good. We might also reflect on this ultimate end in those moments in which it looks as though doing the right thing will lead to an overall bad state of affairs (not merely for oneself, but impartially). While duty must still be the incentive in these cases, we will be aided by the working assumption that the universe is ultimately a just place, and that by adhering to the law we are doing our part in ushering in a state of affairs in which those who might be harmed as a side-effect of our right conduct will eventually receive the happiness that they deserve. All of this could be possible only through assistance from God, and so adoption of this ultimate end will bring with it the assumption of God’s existence. The result is not knowledge because the validity of the law itself in no way depends on God. The duty to promote the highest good is indirect, as it arises from mere contingent facts about human sensibilities and the human faculty of practical reason. The moral law would be no less binding were we to somehow discover that we live in a godless universe (CJ 5:451). Yet, while the relevant doxastic state is belief rather than knowledge, it is not a belief in the ordinary sense, such as the beliefs we have about historical or scientific propositions. The state we are led to through the highest good is a type of moral faith that is made up of two components. The first component is hope, which is essentially an optimistic wish. Consider the notion of not losing one’s hope. When a person claims not to have lost hope, she does not mean merely that she has not lost her wish or desire for the preferred outcome. One can maintain a wish without maintaining hope, as when we wish things were different from how they in fact are. To not lose hope is to maintain some degree of confidence that one’s wish or desire will be fulfilled even when the evidence for this is low. The second component of moral faith is the actions one takes and the assumptions under which they are carried out. In fulfilling my indirect duty to represent the highest good as my ultimate end, I will perform obligatory acts under the assumption that God exists, and I will maintain hope that he does. In the second Critique, Kant refers to this practical faith as “assent in a moral respect” (CpV 5:146). This notion of moral faith might be rendered clearer through an analogy. Imagine that two soldiers have been given an assignment that involves each of them entering different parts of extremely dangerous enemy territory. And imagine that the mission can be successfully

Beneficent Governor of the Cosmos  221 completed only if both soldiers make the lengthy journey (roughly two weeks) and complete their respective tasks without being captured. Suppose further that they will not be able to communicate with each other or any other members of their side. The soldiers are informed that although they are both very likely to be captured before completing their tasks, attempting this mission is the only available means of avoiding a military defeat that would lead to catastrophic consequences for millions of innocent civilians in multiple countries. Now consider the doxastic attitude of each soldier with respect to the other once they are a week into the mission. Given what they were told about the probability of capture, combined with their own experiences of narrowly avoiding capture themselves, each soldier has objective epistemic grounds for believing that the other has been captured. But this does not mean that they must give up hope. Since neither of them is certain that the other has been captured, it is rationally possible for them to maintain hope. And since they have a duty to carry out the mission, it is incumbent upon them to try to uphold a positive view of the success of their counterpart. Since the end they are working toward can be realized only on the condition that they both do their respective parts, their pursuit of this end can be rationally carried out only under the assumption that they will both succeed. Hence, in optimistically pursuing the objective, the soldiers manifest their assent to a proposition for which they lack objective epistemic warrant. Crucially, this assent is justified by the practical need arising from their circumstances. Thus, they are not in a state of knowledge or even ordinary belief. Rather, the soldiers manifest a type of practical faith that is similar to the moral faith in God that ostensibly arises when we adopt the highest good as our ultimate end.15 After outlining the moral grounds for faith in Critique of Judgment, Kant notes that his “proof” is clear enough so that “one could easily adapt [it] to the form of logical precision” (CJ 5:450). The fact that he never attempts to do so is evidence that this claim is an instance of Kant’s characteristic hyperbole. Indeed, the task of reconstructing Kant’s argument in a premise-conclusion form that renders it plausible while staying loyal to the texts is quite challenging. What follows is my best attempt, based on the preceding discussion. 1. We have an indirect duty to minimize threats to the diminishment of our moral resolve. 2. Representing an ultimate moral end that reason can justify is essential for minimizing the risk of diminished moral resolve. 3. The only ultimate moral end that reason can justify is a world full of virtuous agents who receive the happiness that they deserve (i.e., the highest good). 4. Representing the highest good as our ultimate moral end is essential for minimizing the risk of diminished moral resolve. (2–3)

222  Tyler Paytas 5. We have an indirect duty to represent the highest good as our ultimate moral end. (1, 4) 6. Representing the highest good as our ultimate moral end involves moral faith (hope combined with action under the assumption of God’s existence). 7. Adhering to our indirect duties (as we should) will lead us to moral faith. (5–6) Premises 1 and 6 are relatively uncontroversial. If duty is real, then it only makes sense that we would have indirect duties to make ourselves more reliable in fulfilling our direct duties. And it is hard to see how we could represent the highest good as the intended effect of our right conduct in the absence of moral faith. More controversial are premises 2 and 3. Is representing an ultimate moral end that reason can justify really essential for protecting our moral resolve? And if so, is a world of universal virtue and corresponding happiness the only suitable ultimate moral end? Before attempting to answer these questions, it will be useful to consider Sidgwick’s case for the moral necessity of God. While Sidgwick’s argument is puzzling in certain respects, it is more straightforward than Kant’s. And although Sidgwick diverges from Kant on certain points, his view is motivated by some of the key assumptions that underlie Kant’s argument. Hence, by investigating Sidgwick’s argument with Kant’s position in view, we might be better able to assess the merits of both, and ultimately decide for ourselves whether the general idea of the moral necessity of faith is plausible. Let us turn then to Sidgwick’s reintroduction of God into philosophical ethics, which arises from his notorious “dualism of practical reason.”

The Profoundest Problem of Ethics According to Sidgwick, the aim of philosophical ethics is to systematize and free from error the apparent cognitions that most men have of the rightness or reasonableness of conduct, whether the conduct be considered as right in itself, or as the means to some end commonly conceived as ultimately reasonable. (ME 77) Sidgwick approaches this endeavor by focusing on the most influential “methods of ethics,” which he defines as “any rational procedure by which we determine what individual human beings ‘ought’—or what it is ‘right’ for them—to do, or to seek to realise by voluntary action” (ME 1). These methods are logically connected to what are taken to be ultimate (i.e., non-derivative) reasons for action. Sidgwick’s stated aim in The Methods of Ethics is not to argue for or against any particular

Beneficent Governor of the Cosmos  223 method and its associated account of ultimate reasons but rather to clarify the various methods and elucidate their implications and mutual relations. Although many people turn to the ostensible commands of God as the source of true ethical principles, Sidgwick does not view this method as meriting serious attention. As noted earlier, in conceiving of God as acting in good and just ways, we commit ourselves to viewing God as a moral agent. This implies that our notions of “justice,” “right,” and “good” are absolute and independent of God’s will (ME 31, 505; see Crisp 2015, 8–9).16 With God-based morality set aside, Sidgwick homes in on three methods of ethics: egoism, utilitarianism, and a form of common-sense deontology that he calls “intuitional morality.” This choice of focus arises from the fact that there appear to be three candidate sources of ultimate reasons that have widespread appeal: happiness, unconditional duty, and perfection (or excellence).17 Happiness can function as a source of ultimate reasons in either of two ways, depending on whether the object is one’s own happiness (egoism) or universal happiness (utilitarianism).18 The view that unconditional duties (i.e., moral obligations whose justifications are independent of consequences) are the true source of ultimate reasons gives rise to the method of intuitional morality. Although perfection can be understood as an independent end and source of ultimate reasons, Sidgwick subsumes perfectionism under the method of intuitional morality. This is because moral virtue is widely regarded as the chief element of human perfection, and moral virtue is typically understood as involving dispositions associated with the duties of common-sense morality (ME 10–11; 83–83).19 Although the methods of egoism, utilitarianism, and intuitional morality each play a role in shaping common moral judgments and practices, they appear to conflict with each other at least some of the time. Since methods that deliver incompatible verdicts cannot all be valid, we must either find grounds for privileging one method over the others, or else hope that the disagreements are merely superficial (ME 6, 14). Hence, one of Sidgwick’s aims in the Methods is to search for a means of reconciling egoism, utilitarianism, and intuitional morality into a single coherent method that is systematic and comprehensive. His hopes of achieving this aim are buoyed by an apparent resolution of the conflict between utilitarianism and intuitional morality. Among Sidgwick’s chief criticisms of common-sense intuitionism is that the various ostensible duties often conflict, such as when the duty to tell the truth conflicts with the duty to avoid causing harm. In many cases, the only principled resolution involves appeal to considerations of utility. Indeed, this is how most of these conflicts are resolved in practice, as when a person decides not to tell the truth on the grounds that the harms resulting from veracity would be greater than any harms caused by lying

224  Tyler Paytas (ME 316, 348). Not only does appealing to utility allow for a principled resolution to such conflicts, it also provides a deep unifying explanation for the duties themselves—they are the principles which, if generally followed, tend to promote the greatest net sum of happiness in society. Hence, intuitional morality is ultimately subordinate to utilitarianism, and the apparent conflict between these methods is resolved (ME 275, 354, 384, 421).20 Unfortunately, Sidgwick is unable to find a means of reconciling the apparent conflict between utilitarianism and egoism. Although there are many contexts in which the ends of self-interest and universal benevolence cohere, there are also clear cases in which promoting the general good requires a sacrifice of one’s own good. Since two conflicting rules of conduct cannot both be reasonable, the unresolved conflict between egoism and utilitarianism implies that “the apparently intuitive operation of the Practical Reason, manifested in these contradictory judgments, is after all illusory” (ME 508). Since egoism and utilitarianism are ostensibly grounded in well-vetted ethical intuitions, the fact that they contradict each other casts doubt on the general enterprise of philosophical ethics. The problem is expressed most somberly in the first edition of the Methods: “The Cosmos of Duty is thus really reduced to a Chaos: and the prolonged effort of the human intellect to frame a perfect ideal of rational conduct is seen to have been foredoomed to inevitable failure” (ME1 473).

A Divine Solution? Interpretive Puzzles Although Sidgwick is deeply troubled by the dualism, he does not rule out the possibility of a resolution. He suggests that the existence of God would resolve the problem because God would presumably reward those who aim to promote the general good, thus aligning beneficence with prudence. This is based on the idea that, since God is a rational being, and rational beings are “bound to aim at good generally” (ME 382), God’s ultimate end is universal happiness (Sidgwick has by this point concluded that happiness is the sole intrinsic good). Since God is perfectly just, he will presumably reward human beings who make universal happiness their ultimate end and punish those who prioritize self-interest over the greater good (ME 506).21 Unfortunately, Sidgwick does not see any intuitive warrant for theism (ME 507). And despite devoting a substantial amount of time to paranormal investigations, he never found any empirical evidence in support of God’s existence (Schultz 2004). The lack of intuitive or empirical warrant for theistic belief seemingly leaves us with an unresolved dualism. However, the fact that only God could resolve an apparent

Beneficent Governor of the Cosmos  225 contradiction in practical reason might itself provide a type of justification for theism. Sidgwick concludes the Methods by suggesting that our need for moral-theoretical coherence might allow for a solution to the “profoundest problem of ethics.” The idea is that since at least some scientific hypotheses are apparently deemed credible simply because they are required to avoid a contradiction, the fact that the God hypothesis prevents a contradiction in practical reason might similarly be an adequate reason for its acceptance. While a scientifically minded person might balk at this suggestion, Sidgwick warns that such a refusal may have damning implications for our broader system of beliefs: If . . . we find that in our supposed knowledge of the world of nature propositions are commonly taken to be universally true, which yet seem to rest on no other grounds than that we have a strong disposition to accept them, and that they are indispensable to the systematic coherence of our beliefs,—it will be more difficult to reject a similarly supported assumption in ethics, without opening the door to universal scepticism. (ME 509) Despite the fact that the proposed coherentist move represents a potential solution to the problem that most deeply troubles him, Sidgwick never offers an extended treatment of it (Crisp 2015, 234; Schneewind 1977, 377–379). While Sidgwick’s failure to pursue this possible solution is puzzling, a more basic puzzle concerns the suggestion that God’s existence could resolve the dualism in the first place. As C. D. Broad first noted, even if divine punishments and rewards would eliminate practical conflicts between duty and self-interest, there would still be a deep theoretical conflict between egoism and utilitarianism (Broad 1930, 158–159, 244–245, 253; see Crisp 2015, 227–234; Irwin 2009, 528–529). These methods do not merely tell us what we ought to do, they also make claims about the fundamental right-making features of our actions. Whereas egoism says that an action is right solely in virtue of the fact that it promotes one’s own good, utilitarianism says that an action is right solely in virtue of the fact that it promotes universal good. These two theories make “totalizing claims” and thus cannot be rendered (theoretically) coherent by any contingent matter such as God’s existence (Skorupski 2007, 432). Postulating God might provide some practical benefit insofar as it would allow an agent who is convinced that one of either egoism or utilitarianism must be true to be confident that she is acting rightly by promoting universal happiness, since divine sanctions would align this end with selfinterest (Broad 1930, 255). However, as Broad notes, the conscientious agent will not only want to act rightly, she will also want to act from the right principle. And even with knowledge of God’s existence, the agent

226  Tyler Paytas torn between egoism and utilitarianism would not know which principle is true (ibid.). In light of such concerns, Jerome Schneewind suggests that we should aim for an interpretation of the dualism that preserves a contradiction of principles while also explaining how divine sanctions could solve the problem. Since Broad’s criticism is unavoidable if we assume that the two parts of the dualism make totalizing claims in which there can be only one right-making feature, a charitable interpretation must present nontotalizing versions of the egoist and utilitarian principles. Schneewind also argues that our interpretation should bring out the importance of the distinction between own-good and universal good, which Sidgwick often emphasizes. Hence, Schneewind proposes the following formulations of the principles comprising the dualism: Own Good (OG): Maximizing the agent’s own good is an ultimate right-making characteristic. Universal Good (UG): Maximizing the universal good is an ultimate right-making characteristic. (Schneewind 1977, 373; labels altered from original) For clarity, we should note that OG implies that any act that does not maximize the agent’s own good is thereby wrong, and UG implies that any act that does not maximize universal good is thereby wrong. Since neither formulation stipulates an exclusive right-making feature, there is no inherent contradiction between the two. As long as the principles never issue conflicting verdicts about particular acts, they can both be true. A world in which both principles are true would be one in which any act that is prudentially optimific would also be impartially optimific, and vice versa. In such a world, any particular right act will be right in virtue of the fact that it maximizes the agent’s own good and in virtue of the fact that it maximizes universal good (i.e., universal happiness). And any act that fails to maximize either one’s own good or universal good would thereby be wrong.22 What generates the difficulty, according to Schneewind’s interpretation, is the possibility that some acts that maximize one’s own good fail to maximize universal good. Under such conditions, the two principles in question issue conflicting verdicts. And it appears that such conditions obtain in the actual world. Since the dictates of these principles appear not to coincide in the actual world, they cannot both determine actual rightness (Schneewind 1977, 374). However, if God exists, then the appearance of a genuine conflict is illusory because it would always be in the individual’s interest to choose the act most conducive to universal happiness. The No-Self-Sacrifice Interpretation Schneewind’s interpretation has the merit of avoiding the implication that Sidgwick made the significant blunder of suggesting that a

Beneficent Governor of the Cosmos  227 contingent matter such as the existence of God could resolve a theoretical contradiction. The main problem with Schneewind’s interpretation is that it makes no reference to an ultimate end—a notion to which Sidgwick repeatedly appeals (ME 3, 77, 91, 239, 387, 403, 415, 498).23 Sidgwick does not assume from the outset that there must be a single ultimate end; he takes seriously the possibility that ultimate reasons are grounded in unconditional duties (rather than an ultimate end). However, he is drawn to the idea of an ultimate end because it allows for the type of system and synthesis that he deems necessary for the viability of practical reason (1905, 467).24 The idea of an ultimate end makes it easy to systematize various principles and reasons into a coherent whole. This in turn explains why Sidgwick does not seriously consider a view that we might call “partial dualism,” according to which one’s own happiness and the happiness of others ground ultimate reasons of differing strength that must be weighed against each other via one’s judgement (Broad 1930, 244–245; Crisp 2006, 132–143; Parfit 2011, I:134–141; Phillips 2011, 145–151). Of course, an interpretation on which the dualism consists of selfinterest and universal happiness as two distinct ultimate ends, which Schneewind (1977, 373) considers, would lead us to the difficulties mentioned earlier. Even if God’s existence would ensure that we could always pursue both ends simultaneously, there would still be a deep theoretical conflict insofar as two different ends appear to us as ultimate. And such a theoretical conflict would still raise doubts about our ability to reliably grasp basic ethical truths. However, there is e­ vidence that Sidgwick’s considered view of the dualism did not include a principle of egoism according to which self-interest is a plausible ultimate end. Consider the following passage from the start of Book III: The offence which Egoism in the abstract gives to our sympathetic and social nature adds force to the recoil from it caused by the perception of the occasional practical conflict with common notions of duty . . . A dubious guide to an ignoble end appears to be all that the calculus of Egoistic Hedonism has to offer. (ME 199–200; see, also, 500) The perceived implausibility of self-interest as an ultimate end is expressed even more forcefully in a chapter on Ultimate Good: And certainly one’s individual happiness is, in many respects, an unsatisfactory mark for one’s supreme aim, apart from any direct collision into which the exclusive pursuit of it may bring us with rational or sympathetic Benevolence. It does not possess the characteristics which, as Aristotle says, we “divine” to belong

228  Tyler Paytas to Ultimate Good: being (so far, at least, as it can be empirically foreseen) so narrow and limited, of such necessarily brief duration, and so shifting and insecure while it lasts. But Universal Happiness . . . seems an End that satisfies our imagination by its vastness, and sustains our resolution by its comparative security. (ME 403–404) These passages suggest that, insofar as Sidgwick is tempted by a principle of self-interest, it is not a principle of “pure egoism” (Broad 1930, 241), according to which the ultimate reasonable end (and sole right-making feature) is one’s own happiness. Not only does such a principle seem implausible in itself (by denying the existence of basic other-regarding reasons), it stands in theoretical conflict with the principle of Rational Benevolence—“each one is morally bound to regard the good of any other as much as his own”—in which Sidgwick expresses high confidence (ME 382).25 Note further that Sidgwick’s main argument in defense of the egoistic principle comprising half of the dualism offers very little support for pure egoism. He makes an appeal to the “real and fundamental” distinction between persons and the fact that individuals have special concern for their own existence (ME 498; 1889, 484). The metaphysical separateness of persons can perhaps be grounds for believing that one ought to show some degree of basic partiality toward oneself. But it is hard to see how it could support the extreme view that one could be justified in having complete disregard for the welfare of other beings, as pure egoism would allow. That being said, pure egoism is a view that Sidgwick takes seriously, even devoting all of Book II of the Methods to it. And he does mention this view in his most lengthy discussion of the dualism, which occurs in the final chapter of the Methods. Note, however, that in that discussion Sidgwick does not present the egoist position as one that he himself subscribes to or even finds plausible. What he concedes regarding pure egoism is that, unless the egoist acknowledges the idea of universal good, there appears to be no non-question-begging argument that could successfully move the egoist to admit universal happiness as the true ultimate reasonable end. This is certainly troubling, but it is not the true source of Sidgwick’s concern because he does not find the denial of universal good all that plausible (note again the implausible implication that there are no basic other-regarding reasons). Further, the tendency of some reflective persons to accept pure egoism can be explained away along the lines of the debunking explanation that Sidgwick offers regarding those who reject the axiom of Rational Benevolence (ME 382, 339; see, also, Parfit 1984, 130; de Lazari-Radek and Singer 2014, ch. 7).

Beneficent Governor of the Cosmos  229 The principle of self-interest that Sidgwick finds irresistible, and that constitutes half of the dualism, is made explicit in the following key passage from the concluding chapter: And further, even if a man admits the self-evidence of the principle of Rational Benevolence, he may still hold that his own happiness is an end which it is irrational for him to sacrifice to any other; and that therefore a harmony between the maxim of Prudence and the maxim of Rational Benevolence must be somehow demonstrated, if morality is to be made completely rational. This latter view, indeed (as I have before said), appears to me, on the whole, the view of Common Sense: and it is that which I myself hold. (ME 498) Here Sidgwick explicitly describes his own view as admitting the selfevidence of Rational Benevolence while simultaneously holding that self-sacrifice is irrational. While he does refer to the maxim of Prudence, this should not be read as a reference to pure egoism because the existence of God could not render pure egoism theoretically coherent with Rational Benevolence (as Schneewind notes). Moreover, an individual who finds Rational Benevolence self-evident could not simultaneously endorse a principle according to which there are no basic other-regarding reasons. Hence, we might consider an interpretation that emphasizes the explicit appeal to the irrationality of self-sacrifice in the previous passage. On such a reading, the relevant principle of self-interest is not pure egoism but rather the claim that it is irrational to sacrifice one’s own good.26 Further textual support for this interpretation is provided by additional passages in which the self-interested component of the dualism is put in terms of the irrationality of self-sacrifice. For instance, in “Some Fundamental Ethical Controversies,” Sidgwick characterizes the relevant intuition as the idea that “it would be irrational to sacrifice any portion of my own happiness unless the sacrifice is to be somehow at some time compensated by an equivalent addition to my own happiness” (1889, 483; see, also, ME xviii, 109 fn.1, 418, 498).27 The rational constraint against self-sacrifice is theoretically compatible with a principle stating that the ultimate reasonable end (i.e., the final goal toward which all of our conduct should be directed) is universal happiness. The ostensible irrationality of self-sacrifice does not imply that one’s own good is the ultimate end; individuals who ensure their own perfect happiness still have work to do insofar as they still have reasons grounded in the happiness of others. While the principles of universal benevolence and non-self-sacrifice are logically consistent with each other, if it is ever the case that the available

230  Tyler Paytas act that is most conducive to universal happiness requires a sacrifice of one’s own happiness, the principles would issue contradictory verdicts. The fact that maximizing universal happiness often does seem to require self-sacrifice is the source of the “profoundest problem of ethics.” With these considerations in mind, I propose the following interpretation of the dualism and Sidgwick’s suggested divine solution: 1. Practical reason dictates that the ultimate end of reasonable conduct is universal happiness. 2. Practical reason dictates that sacrifice of one’s own happiness is always unreasonable. 3. If God does not exist, then some actions that are most conducive to universal happiness involve a sacrifice of one’s own happiness. 4. If God does not exist, then there are cases in which practical reason dictates that an action is both reasonable and unreasonable. (1–3) 5. If there are cases in which practical reason dictates that an action is both reasonable and unreasonable, then we ought to be skeptical about practical reason in general. 6. Unless we believe in God, we ought to be skeptical about practical reason in general. (4–5) Note that on this interpretation the impartialist element of the dualism is not full-blooded utilitarianism but rather a principle stating that universal happiness is the ultimate reasonable end. This is important because utilitarianism (understood as a totalizing view) is logically incompatible with the principle prohibiting self-sacrifice, and so God would not be able to resolve the dualism so understood. What makes the dualism troubling is not that it renders practical reason incomplete by failing to provide action-guidance in cases of conflict between self-interest and universal good. The problem is, rather, that the presence of contradictory verdicts in such cases means that at least one of the principles must be false. If we know that at least one principle which we judge to be self-evident (after careful reflection) is false nonetheless, this should undermine our confidence in the general enterprise of philosophical ethics. This is the motivation for premise 5, which states that such conflicts would provide warrant for general skepticism about practical reason. As Sidgwick remarks in the second edition of the Methods, a resolution of the dualism is “a matter of life and death to the Practical Reason” (ME2 468). One might object to my interpretation on the grounds that there still seems to be a deep theoretical conflict in play. Specifically, one might worry that the principle prohibiting self-sacrifice implies the falsity of Rational Benevolence. Recall that Rational Benevolence states that “each one is morally bound to regard the good of any other individual as much as his own” (ME 382). The claim that self-sacrifice in particular is always

Beneficent Governor of the Cosmos  231 unreasonable (premise 2) suggests that, contra RB, one ought to have greater regard for one’s own good than the good of others. In responding to this objection, the first thing to note is that Sidgwick’s use of the term “regard” is somewhat misleading. What matters in this context are not patterns of concern but rather ultimate principles (Crisp 2015, 228). One of the two axioms from which RB is deduced states that “as a rational being I am bound to aim at good generally . . . not merely at a particular part of it” (ME 382). The other axiom essentially states that everyone’s welfare constitutes an equal portion of the good (ibid.). What these axioms entail is that one’s ultimate goal should not be promotion of one’s own happiness but rather universal happiness. The truth of this claim is logically consistent with a rational prohibition against self-sacrifice. And so, while there is a sense in which I ought to have special regard for my own good insofar as I must never sacrifice it (whereas sacrifice of others may in principle be justified), the upshot of RB is that the final end to be aimed at is universal good. If God exists, then my efforts to maximize universal happiness will never require me to sacrifice my own happiness, and so there would be no contradiction. On the other hand, if we live in a godless universe, then it would appear that obeying the rational dictate to aim at good generally will require me to violate the rational dictate to never sacrifice my own good. A related objection also concerns the possibility that, even on my proposed interpretation, the existence of God would not resolve the dualism. For even if God does exist, we could still imagine a world (i.e., a godless world) in which some acts that would lead to the most total happiness would require self-sacrifice. The fact that such a practical conflict is conceivable seems just as worrisome for practical reason as it would be if the conflict were actual. Since basic ethical principles are assumed to be necessary truths, if two such principles could conceivably issue contradictory verdicts, they cannot both be true. This worry is expressed forcefully by William Frankena: It seems to me that two ethical principles cannot both be regarded as self-evident if it is in principle possible for them to come into conflict, and that even a postulate of coincidence in practice cannot save them both. For the coincidence might obtain only because of a fortunate accident about the constitution of our world, and not be true of other possible worlds. But this is a hard question and Sidgwick does not consider it. (Frankena 1974, 457–458, cited in Crisp 2015, 228)28 I have two replies to this objection. First, the fact that Sidgwick does not seem to have considered the problem means that the objection is not damning for my proposal as an interpretation of Sidgwick; we may just have to admit that he was guilty of an oversight here. My

232  Tyler Paytas interpretation would still have an advantage over totalizing interpretations insofar as the oversight in question is much more subtle than the mistake of thinking that God could make logically contradictory principles compatible. Second, Sidgwick may not have been guilty of an oversight after all. Perhaps he was operating under the assumption that, if God exists, then the practical coincidence between the two principles would not be merely a “fortunate accident” because the divine nature would render the circumstances in which the principles would conflict metaphysically impossible. Hence, if God exists, then questions about how ethical principles would operate in a godless world may become irrelevant. It would still be the case that ethical principles could not be grounded in God’s will. But there may be no constraint on true ethical principles that they could never issue conflicting verdicts in an imagined godless world. The only relevant constraint may be that it is metaphysically (rather than logically) impossible for the principles to issue conflicting verdicts, and it is reasonable to suppose that God’s existence would render the conflicts metaphysically impossible.29 This second reply is highly speculative and involves complex philosophical issues that could only be sorted out through a lengthy investigation. Rather than engaging in this inquiry, I would like to return to the broad questions that motivate this chapter. Was there a common factor that led both Kant and Sidgwick from secular ethical foundations to positing the moral necessity of God? And, more importantly, how plausible are the key assumptions that motivate their appeals to theism?

Cosmic Justice and Moral Resolve The Need for Cosmic Justice Recall that, for Kant, the moral necessity of God is grounded in an indirect duty to maintain our moral resolve. While the laws of pure practical reason do not depend on God for their validity, human beings are constituted such that we are inevitably concerned with the effects of our conduct, and we are vulnerable to moral despondency in the face of the widespread suffering and injustice that seem to be an inherent feature of existence. In order to protect our moral resolve, we must represent an ultimate end that reason can justify as the consequence of our moral efforts. This ultimate end is the highest good—a world of universal happiness resulting from corresponding virtue. Since we are duty-bound to maintain our moral resolve, we are duty-bound to represent the highest good as our ultimate end. And since this end is realizable only if God exists, a commitment to the moral law will lead us to act under the assumption that God exists. The resulting outlook is not knowledge or even theoretically justified belief but rather a type of moral faith.

Beneficent Governor of the Cosmos  233 One of the key differences between Kant and Sidgwick on this issue is that Sidgwick takes seriously the bleak possibility that the enterprise of practical reason is “foredoomed to inevitable failure” (ME1 473). Sidgwick believes that, if we live in a godless universe, it follows that the ideal of grasping objective ethical truths through practical reason is a mere chimera because two of the vetted ethical intuitions would issue conflicting verdicts. And this is a very real possibility for Sidgwick, as he expresses skepticism about the possibility of a strictly practical warrant for belief in God: Nor can I fall back on the Kantian resource of thinking myself under a moral necessity to regard all my duties as if they were commandments of God, although not entitled to hold speculatively that any such Supreme Being exists “as Real.” I am so far from feeling bound to believe for purposes of practice what I see no ground for holding as a speculative truth, that I cannot even conceive the state of mind which these words seem to describe, except as a momentary half-willful irrationality, committed in a violent access of philosophic despair. (ME1 471; ME 507, n. 3; see, also, Sidgwick and Sidgwick 1906, 615) Yet, despite his stated rejection of practically grounded theistic belief, Sidgwick’s personal views on this matter were quite similar to Kant’s. A letter to a former schoolmate composed in 1880 (six years after publication of the first edition of the Methods) is especially telling: For if I am asked whether I believe in God, I should really have to say that I do not know—that is, I do not know whether I believe or merely hope that there is a moral order in this universe. . . . All I can say is that . . . I cannot accept life on any other terms, or construct a rational system of conduct except on the basis of this faith. . . . Duty is to me as real a thing as the physical world, though it is not apprehended in the same way; but all my apparent knowledge of duty falls into chaos if my belief in the moral government of the world is conceived to be withdrawn. Well, I cannot resign myself to disbelief in duty; in fact, if I did, I should feel that the last barrier between me and complete philosophical scepticism, or disbelief in truth altogether, was broken down. Therefore, I sometimes say to myself “I believe in God”; while sometimes I say no more than “I hope this belief is true, and I must and will act as if it was.” (Sidgwick and Sidgwick 1906, 347–348; see, also, 605) The similarities to Kant’s position are striking. The first similarity is revealed in the claim that duty is “as real a thing as the physical world.”

234  Tyler Paytas Here Sidgwick expresses a strength of conviction in the truths of ethics and the validity of practical reason that is nearer to Kant than what we find in some of the more skeptical passages of the Methods. A further point of contact is the recognition of the human need to believe in cosmic justice, which essentially involves a reliable connection between duty and happiness. Kant recognizes that, given human nature, viewing the universe as inherently unjust (as appearances suggest) is likely to hinder moral resolve (though this would not undermine the validity of the moral law). Sidgwick differs insofar as he thinks a lack of harmony between duty and one’s own happiness would actually cast doubt on practical reason itself because it would mean that the two key ethical intuitions issue conflicting judgements. But perhaps more deeply, Sidgwick’s attitude manifests Kant’s concerns about existential despair. Sidgwick finds the thought of living in a cold and unjust universe intolerable, and not merely from a self-interested perspective but from an objective point of view. Alluding to a poem by Tennyson, Sidgwick writes: When a man passionately refuses to believe that the “Wages of Virtue” can “be dust,” it is often less from any private reckoning about his own wages than from a disinterested aversion to a universe so fundamentally irrational that ‘Good for the Individual’ is not ultimately identified with ‘Universal Good.’ (ME2 468; see, also, ME1 507) This leads to another key similarity, which is the doxastic attitude that both philosophers adopt with respect to the question of God. Because of his deep commitment to duty, Sidgwick avoided nihilistic despair by adopting moral faith. Despite the skeptical tone of the concluding chapter of the Methods, Sidgwick maintained an optimistic hope that God exists and a resolution to act under the assumption that he does. In this way, Sidgwick personified the spirit of Kantian moral faith.30 Do the views leading to Kantian and Sidgwickian moral faith constitute a strong case for the moral necessity of God? The idea that maintaining moral resolve will be difficult in a universe believed to be inherently unjust (or even irrational, as Sidgwick feared) is not implausible. However, it is not clear that God is necessary for rendering the universe just and rational to a degree sufficient for our moral purposes. What ultimately leads Kant and Sidgwick to posit a need for moral faith is not just the importance placed on cosmic justice but also their shared hedonistic conception of well-being (CpV 5:60; ME 395–407). It is because they take the human good to consist of pleasure and the satisfaction of inclinations that Kant and Sidgwick believe that a universe without God (who will unite duty with happiness) could not be adequate to our moral ideals. However, alternative conceptions of human well-being are available

Beneficent Governor of the Cosmos  235 that could allow us to view the world as sufficiently just even in the absence of a deity. A Stoic Alternative The conception of well-being that may be best suited for this purpose can be traced back to the Stoics. In their view, the human good does not consist in pleasure or brute desire satisfaction but rather in practical wisdom (i.e., virtue) and psychological harmony. While positive hedonic states may be choiceworthy when all else is equal, they should never be chosen at the cost of one’s character. Crucially, a person who manifests good character and a stable, coherent psychology cannot be harmed by external circumstances because virtue is completely sufficient for wellbeing. On this view, it is a mistake to look for the “Wages of Virtue” in some place other than in the exercise of virtue itself.31 This unorthodox value theory is based on the premise that faring well is a matter of acting in alignment with one’s nature. This claim applies equally to all living things, whether human, animal, or even plant, with the defining characteristics of a given type of being determining its nature. As was common among the ancient schools, the Stoics held that the defining characteristic of human beings is our capacity for reason. Hence, our lives go well to the extent that our conduct is guided by the verdicts of our reason rather than affective responses or our instinctual desire for pleasure and status. And our lives go poorly not when we experience pain, poverty, or imprisonment but rather when we act in ways that manifest a lack of justice and integrity.32 As a matter of history, these Stoic doctrines originate within a theistic, teleological framework in which the universe is assumed to be divinely ordered and purposive. However, this type of theism is importantly different from the Christian tradition to which Kant and Sidgwick belong. Perhaps the most significant difference is that, unlike traditional Christianity, the emphasis of the Stoic theological outlook is entirely terrestrial. The Stoics make no appeal to an afterlife in which the virtuous will be rewarded with enjoyment and the wicked will be made to suffer.33 Nor is there any notion of a “final judgement” or “end of days,” in which all of humanity will be judged by the divine lawgiver (Long 2002, 146). According to Epictetus, for instance, the punishment for those who fail to live virtuously is “to be just as they are” (Discourses 1.12.21). This earthly focus is important because it means that the Stoic value theory and corresponding ethical code can in principle be divorced from any theological assumptions. The Stoic account does not depend on the agency of supernatural beings in order to be rendered coherent. One need not believe in the existence of Zeus in order to judge that the human good lies not in external circumstances (including bodily states)

236  Tyler Paytas but rather in the excellences of character that we exhibit in properly facing whatever circumstances happen to arise.34 By embracing the Stoic account of well-being rather than the hedonistic conception endorsed by Kant and Sidgwick, the task of maintaining moral resolve in the world as we find it is rendered more feasible. If we accept that the human good consists in having good character and trusting reason to determine what is best, then Sidgwick’s worry about duty requiring self-sacrifice is avoided because doing the right thing will always promote one’s well-being. Further, the Stoic outlook prevents us from becoming discouraged by the ubiquity of suffering and malevolence. Indeed, we can embrace our encounters with challenging circumstances as opportunities to manifest our capacity for true excellence. As Epictetus counsels: Welcome present circumstances and accept the things whose time has arrived. Be happy when you find that doctrines you have learned and analyzed are being tested by real events. If you’ve succeeded in removing or reducing the tendency to be mean and critical, or thoughtless, or foul-mouthed, or careless, or nonchalant . . . then every day can be a feast day—today because you acquitted yourself well in one set of circumstances, tomorrow because of another. (Discourses 4.4.45–46) As noted earlier, both Kant and Sidgwick are skeptical about eudaimonistic views, including Stoicism.35 The central Kantian and Sidgwickian objection to the Stoic view is that it is flagrantly at odds with a sober understanding of human nature. This critique is directed at the claim that pleasure and happiness (understood in the conventional sense) are not necessary for human well-being. Kant suggests that on this particular point the Stoics “could have been sufficiently refuted by the voice of their own nature” (CpV 5:127). Likewise, Sidgwick observes that “the inhuman severity of the paradox that ‘pleasure and pain are indifferent to the wise man’, never failed to have a repellent effect” (1877b, 30). According to Sidgwick, such a paradoxical view is partly explained by a deep-rooted faith that there cannot really be a genuine conflict between duty and well-being (ME 405). While he recognizes the impulse toward such faith in himself, he finds the Stoic formulation of it, with its denial of the necessity of pleasure for human welfare, too paradoxical to take seriously. Kant takes this worry a step further. He argues that by renouncing the need for (conventional) happiness and representing complete virtue as an attainable ideal, the Stoics not only contradicted common knowledge but also ushered in a type of moral enthusiasm that is less stable and reliable than the sober and pure moral discipline inhering in Christian morals. By failing to acknowledge the basic facts of human nature, the Stoic view

Beneficent Governor of the Cosmos  237 fosters an attitude of conceit while increasing the likelihood of moral laxity when confronted with the reality of human ill-being and the futility of our efforts to live up to the ideal of attaining perfect virtue in this life (CpV 5:86; 5:127).36 For Kant and Sidgwick, the preferred alternative is a type of moral theology in which the rationality of the universe and the non-futility of our moral efforts are ensured by an all-loving deity and the promise of an afterlife in which virtue and happiness are brought into alignment. However, a strong case can be made for the superiority of the Stoic position. First, the Stoic value theory isn’t as untenable as Kant and Sidgwick suggest. Kant remarks that we may always laugh at the Stoic who in the most intense pains of gout cried out: Pain, however you torment me I will still never admit that you are something evil (xaxov, malum)! . . . He felt that his pain was an ill, and his cry betrayed that. (CpV 5:60) However, the fact that most everyone, including Stoics, will eventually cry out when pain becomes severe enough does not tell against the claim that no amount of pain can harm the virtuous. This would only follow if we assume that finding one’s state of consciousness agreeable is essential for well-being, which is exactly what is at issue. An alternative assumption with no less prima facie credibility is that well-being is entirely a matter of exhibiting excellence. If we start with the assumption that excellence is what matters, then pain and pleasure are irrelevant because a person’s excellence does not fluctuate based on the valence of their hedonic states. Obviously, this value-theoretic debate is difficult to adjudicate—we eventually reach intuitional bedrock. But let us grant for the sake of argument that Sidgwick is right about the Stoic position being so counterintuitive that it must be at least partly rooted in a type of moral faith.37 If so, the key question is whether this version of moral faith has more or less going for it than the moral faith championed by Kant and Sidgwick. One point in favor of Stoic faith is that it doesn’t require belief in miracles, supernatural beings, or any other metaphysical oddities; it is simply a matter of deviating from common opinion in choosing one’s value theory.38 A more important advantage of the Stoic outlook is that it is more inspiring and emboldening than the traditional Christian alternative. Kant worries that by confining moral progress to our earthly existence, the Stoic view represents virtue as an impossible ideal and is therefore a recipe for moral laxity due to perceived futility. But just the opposite seems true. If we accept Kant’s claim that our moral development can continue beyond our bodily death, then it is easy to see how we might become even more inclined to yield to temptation. If, instead, we hold

238  Tyler Paytas that our sole opportunity for virtue is here and now, and that everything hinges on the choices we make today (Epictetus Enchiridion 51), we may be better able to hold strong in the face of temptation and adversity. A comprehensive adjudication of these issues would require a booklength treatment. What I hope to have made clear is that, even if we follow Kant and Sidgwick in recognizing a moral need to believe in cosmic justice, we are not forced to conclude that belief in God is morally necessary. We are only led to this conclusion if we assume that the welfare of the virtuous depends on their experiencing agreeable states of consciousness. If instead we hold with the Stoics that virtue is its own reward, we can carry on with the belief that the universe is fundamentally rational and just even in the absence of a deity and an afterlife.

Conclusion I have argued that the mutual reintroduction of God in the ethics of Kant and Sidgwick has a common cause related to an ostensible need for cosmic justice (and rational order) in the form of a unification of happiness and duty. While Kant held that this need will lead the dutiful agent to morally grounded faith, Sidgwick’s stated position was more skeptical. However, in his private beliefs and conduct regarding these issues, Sidgwick perfectly embodied the spirit of Kantian moral theology. I have suggested that one can accept the Kantian and Sidgwickian position regarding the need for cosmic justice while denying the moral necessity of God. If the human good does not consist in enjoyable states of consciousness but rather in the cultivation and realization of excellence, then there is no need to despair over the lack of a reliable connection between dutiful conduct and hedonic well-being. The Stoic view of human welfare allows us to maintain an optimistic outlook without resorting to supernatural beliefs. Sidgwick recognized that some people are able to maintain their moral commitments without any religious beliefs, and though he could not adopt such a position himself, he acknowledged its appeal: “The difficulty that I find is in convincing myself that this untheological morality is really abnormal, and does not rather represent the beginnings of a more advanced stage in the development of the moral consciousness” (Sidgwick and Sidgwick 1906, 615).39

Notes 1. Kant’s works are abbreviated herein as follows: CpV = Critique of Practical Reason, Gr = The Groundwork of the Metaphysics of Morals, MM = The Metaphysics of Morals, TP = “On the Common Saying: That May Be Correct in Theory, but It Is of No Use in Practice” (Kant 1996); CJ = Critique of

Beneficent Governor of the Cosmos  239 Judgment (Kant 1987); ET = “The End of All Things”, Rel = Religion within the Boundaries of Mere Reason (Kant 1998); LE = Lectures on Ethics (Kant 1997). 2. The initial presentation of the highest good in the first Critique is distinct from subsequent discussions in that it makes a problematic appeal to the desire for happiness as the source of moral motivation. This does not occur in subsequent discussions of the highest good which were developed after the rationalist account of moral motivation was first presented in the Groundwork. 3. There is controversy over whether Kant’s highest good requires mere proportionality between happiness and virtue or rather universal virtue combined with universal happiness. The main argument for the proportionality interpretation is the fact that at least some people never seem to commit to the pursuit of virtue, thus making the aim of universal virtue futile. However, strong textual evidence for the universalist interpretation is found in TP (8:279) and LE (27:717). Further, two philosophical considerations tell strongly against the proportionality interpretation. First, if some people never attain virtue and thereby do not receive happiness, it is hard to understand why a just God would have created them in the first place (ET 8:329). Second, as Parfit (2011, I:245) notes, “it would always be better if there was more virtue, and more deserved happiness, even if the proportionality condition would be less well met.” For defense of the proportionality interpretation, see O’Connell (2012). 4. Kant considers and rejects ancient Greek views according to which happiness is sufficient for virtue (i.e., Epicureanism) and virtue is sufficient for happiness (i.e., Stoicism). Whereas the Epicureans had an incorrect supreme practical principle (i.e., happiness), the Stoics’ claim that satisfaction of inclinations is not a necessary component of happiness runs afoul of common experience and basic facts of human nature (CpV 5:126). Sidgwick echoes Kant’s criticisms of these views (ME 162–175; 1877b, 29–30). I shall return to the Stoic conception of the highest good below. For a discussion of Kant’s and Sidgwick’s respective critiques of eudaimonism, see Sukaina Hirji’s chapter in this volume. 5. Scholars disagree about whether these later discussions represent a substantial shift in Kant’s thinking or are rather attempts to elucidate the same basic account given in the second Critique. Due to concerns that the later works introduce extraneous teleological considerations, some commentators have attempted to reconstruct a comprehensive account of the highest good (i.e., one that provides answers to the questions raised in the previous paragraph) working solely from the second Critique (e.g., Bader 2015). For a defense of the view that there is a basic underlying unity in the different arguments from the second Critique onward, see Pasternack 2014. 6. See Pasternack (2014, 48). Though I diverge from Pasternack on certain interpretive points, his excellent commentary influenced my analysis in this section. 7. Kant’s idea that reason “seeks maximal unity under principles” is shared by Sidgwick: “But undoubtedly we all recognize that this systemization is demanded of us as reasonable beings: indeed it is with a view to this that we set out on our inquiry for an ultimate end of conduct” (1876, 547, see, also, 1905, 467). Note that this idea has roots in a classical tradition going at least as far back as Aristotle (NE 1:1 1094a). For discussion, see Sukaina Hirji’s chapter in this volume and Wood (2000). 8. See Sussman 2010.

240  Tyler Paytas 9. I defend this “patient-centered” interpretation of the duty to cultivate sympathy in Paytas 2015. 10. Sidgwick is in agreement with Kant on this point. He describes the view that good-willing/right-seeking is the sole ultimate good as a “palpable and violent paradox” (ME 394). 11. Note the similarity to Sidgwick’s definition of objective good as “what as a rational being I should desire and seek to realise, assuming myself to have an equal concern for all existence” (ME 112). 12. See Sussman (2010). Some read Kant as positing a secular version of the highest good (e.g., Reath 1988). For a critique of this interpretation, see Pasternack (2014). 13. Here one might object that we need not believe that the best outcome is possible in order to maintain our moral resolve—it may be enough to believe that at least a fairly good outcome is possible. I suspect Kant would reply by insisting that a world in which we never have a chance to complete our moral development, and the good willing we do manage to exhibit is inevitably rewarded with pain, illness, and eternal darkness, is not even a fairly good world. I thank Rob Shaver for raising this concern. 14. Glaube is distinct from belief in the ordinary sense in that Glaube is a type of assent that is rendered sufficiently justified by non-epistemic considerations. The clearest example of non-epistemically justified assent is the morally grounded faith in God and immortality that I will consider in the rest of this section. For a detailed discussion of the varieties of Glaube, see Chignell (2007). 15. For more on Kantian moral belief see Wood (1970), Hare (1996), and Chignell (2007). 16. Two further considerations tell against viewing divine command ethics as a plausible method. First, since it is hard to see how human conduct could possibly frustrate or facilitate the ends of a deity, the idea that God’s will could be a source of ultimate reasons for human beings is dubious (ME 79; 1891, 62). Second, even supposing that God’s will were a source of ultimate reasons, it is not clear how we could determine what God wills (ME 80, 504). 17. Sidgwick defends a hedonistic account of well-being and ultimate value, and so the two methods directed at happiness as ultimate end both employ a hedonistic conception. 18. A third view holds that ultimate reasons are grounded exclusively in the happiness of some limited class of individuals, such as one’s family, nation, race, or species. Sidgwick rejects this possibility on the grounds that any such boundary would be arbitrary (ME 10). 19. Sidgwick’s restriction of his focus to the three methods of ethics has been the subject of criticism (Schneewind 1977; Irwin 2009). For discussion of this issue, see Crisp (2015) and de Lazari-Radek and Singer (2014). 20. David Phillips (2011) and Thomas Hurka (2014) argue that Sidgwick applied a double standard in his argument for the superiority of utilitarianism over deontology. For discussion of this issue, see Donagan (1977), Crisp (2015), and Shaver (2014). 21. This seems to imply that God’s ultimate end is not universal happiness, per se, but rather universal deserved happiness. This raises interesting questions about Sidgwick’s views on desert and how they fit with his brand of utilitarianism. I do not have adequate space to address those questions here. 22. For a similar interpretation, see Mackie (1976). 23. A second potential problem arises from the fact that, as Schneewind acknowledges, Sidgwick sometimes expresses the principle of self-interest as referring to the sole right-making property (ME 95, 174, 403, 497). However, we

Beneficent Governor of the Cosmos  241 need not assume that whenever Sidgwick is describing a version of egoism he is talking about the same view that comprises half of his dualism. As I shall explain presently, there are good reasons to think that when Sidgwick is describing full-blooded egoism, he is not attempting to articulate a principle that he himself finds highly plausible. 24. This is one of several points on which Sidgwick finds strong agreement from Kant: “Oh how I sympathise with Kant! With his passionate yearning for synthesis and condemned by his reason to criticism” (Sidgwick and Sidgwick 1906, 177). 25. Schneewind’s interpretation avoids this second problem. However, I shall now propose an interpretation that avoids the problem equally well while better accommodating two prominent features of Sidgwick’s discussions of the dualism: (1) the frequent appeal to the idea of an ultimate end, and (2) the frequent appeal to the irrationality of self-sacrifice. 26. This is similar to the principle that Robert Shaver calls “veto egoism” (1999, 2). Veto egoism says that contributing to the agent’s well-being is a necessary (but not sufficient) condition for rational action. I’ve chosen to formulate the principle as a rational prohibition on self-sacrifice because Sidgwick frequently puts it in these terms. For other veto-egoism-like views, see McLeod (2000, 284); Phillips (2011, 134, 136, 140); Crisp (2015, 228–230). 27. There are two passages in which Sidgwick seems to indicate that he finds it plausible that one’s own good is the ultimate reasonable end (ME 404 fn.; 1889, 486). But given his frequent appeals to the irrationality of self-sacrifice, and given that this formulation of the principle avoids the philosophical difficulties concerning pure egoism, the most charitable reading of the allusions to self-interest as a plausible final end is that they are insufficiently careful glosses of the more precise view that I have proposed in this section (i.e., that it is irrational to sacrifice one’s own good). 28. Note that this objection is equally threatening to Schneewind’s interpretation because that also relies on the idea that the dualism would be avoided by the (ostensibly) contingent fact of God’s existence. 29. Considerations along these lines seem to underlie the defense of Sidgwick’s position presented in Coady (1994, 115). See, also, Schultz (2004, 213). For a relevant argument concerning God and metaphysical necessity, see Plantinga (1974). 30. For a thorough treatment of Sidgwick’s spirituality and the role it played in both his private and public life, see Shultz (2004). 31. Elements of Stoic doctrines overlap with the broader category of ancient eudaimonism. Opponents of Stoicism such as the Academics and Peripatetics agreed that virtue is preferable to happiness in the modern conventional sense. Their disagreement with the Stoics concerned the claim that virtue is entirely sufficient for well-being (Sidgwick 1902, 94). Here I focus on the Stoics precisely because their denial of the importance of pleasure and external circumstances makes their position especially well-suited to overcome the concerns that ostensibly necessitate belief in God according to Kant and Sidgwick. For a discussion of Kant and Sidgwick on eudaimonism, see Sukaina Hirji’s chapter in this volume as well as White (2002, ch. 1). 32. This point is forcefully presented throughout Epictetus’s Discourses (e.g., 4.1.119–127). 33. This does not hold for some of the early Stoics such as Cleanthes and Chrysippus. 34. For modern examples of secularized Stoicism, see Becker (1998) and Irvine (2009) 35. Among Sidgwick’s chief criticisms is that such views are circular and thereby fail to issue substantive ethical guidance. The vicious circle is composed of

242  Tyler Paytas the claim that wisdom is the sole good and proper aim of rational beings, along with the claim that wisdom consists of knowledge of the good. One answer to this challenge on behalf of the Stoics invokes their doctrine of “preferred indifferents.” The basic idea is that while wisdom (i.e., virtue) is the sole intrinsic good, things such as health, friendship, and even pleasure can be choiceworthy under the right circumstances. The fact that Stoicism makes room for non-virtue-based goods for people to aim at seemingly undermines the force of Sidgwick’s circularity charge (see Crisp 2015, 114–115). However, whether the doctrine of preferred indifferents is entirely coherent and plausible is a matter of controversy (see Shields 2012, 197–207). 36. These points are related to a further critique offered in the Religion, which is that the Stoics mistakenly located the enemy of morality in the inclinations themselves rather than in the innate human propensity to privilege the inclinations over duty (Rel 6:57–59). However, one can plausibly interpret the Stoics as basically in agreement with Kant on this point. 37. Sidgwick also claims that the Stoic view arises partly from “a confusion of thought between what it is reasonable for an individual to desire, when he considers his own existence alone, and what he must recognize as reasonably to be desired, when he takes the point of view of the larger whole” (ME 405). But it is doubtful that the Stoics were guilty of such a confusion. When Epictetus claims that it would be better to suffer physical pain and loss of social status than to sacrifice one’s integrity, it is unlikely that his underlying thought is that maintaining one’s integrity would be better for the universe as a whole. See, e.g., Discourses 1.2. 38. Note that Epictetus was open to irrational belief-forming processes if they were deemed necessary for psychological harmony: “If I had to be deceived into believing that externals, which lie outside our power, are not man’s proper concern, personally I would consent to such a deception, provided it really could enable me to live an untroubled life, in peace of mind. Which condition you prefer you can determine for yourself” (Discourses 1.4.27; see, also, Fragments 10a). 39. Earlier versions of this chapter were presented at the 2016 Kant-Sidgwick Workshop at Universität Stuttgart, Oklahoma State University, the 2018 ACU Conference on Moral Epistemology, and Murdoch University. I thank audiences at those presentations for their questions and comments. I am especially grateful to Robert Audi, Nicholas Baima, Nevin Climenhaga, and Robert Shaver for helpful comments on earlier drafts.

References Aristotle. 2000. Nicomachean Ethics. Edited and translated by R. Crisp. Cambridge: Cambridge University Press. Aurelius, M. 2006. Meditations. Edited and translated by M. Hammond. New York: Penguin Books. Bader, R. M. 2015. “Kant’s Theory of the Highest Good.” In The Highest Good in Aristotle and Kant, edited by J. Aufgerheide and R. M. Bader, 183–213. Oxford: Oxford University Press. Becker, L. C. 1998. A New Stoicism. Princeton: Princeton University Press. Broad, C. D. 1930. Five Types of Ethical Theory. London: Routledge and Kegan Paul. Chignell, A. 2007. “Belief in Kant.” Philosophical Review 116 (3): 323–360.

Beneficent Governor of the Cosmos  243 Coady, A. J. 1994. “Henry Sidgwick.” In The Nineteenth Century, Routledge History of Philosophy, vol. 7, edited by C. L. Ten, 101–121. London: Routledge. Crisp, R. 2006. Reasons and the Good. Oxford: Clarendon Press. Crisp, R. 2015. The Cosmos of Duty: Henry Sidgwick’s Methods of Ethics. Oxford: Oxford University Press. de Lazari-Radek, K., and P. Singer. 2014. The Point of View of the Universe: Sidgwick and Contemporary Ethics. Oxford: Oxford University Press. Donagan, A. 1977. “Sidgwick and Whewellian Intuitionism: Some Enigmas.” Canadian Journal of Philosophy 7: 447–465. Epictetus. 2008. Discourses and Selected Writings. Edited and translated by R. Dobbin. New York: Penguin Books. Frankena, W. 1974. “Sidgwick and the Dualism of Practical Reason.” Monist 58: 449–467. Hare, J. 1996. The Moral Gap. Oxford: Clarendon Press. Hurka, T. 2014. “Sidgwick on Consequentialism and Deontology: A Critique.” Utilitas 26: 129–152. Irvine, W. B. 2009. A Guide to the Good Life: The Ancient Art of Stoic Joy. Oxford: Oxford University Press. Irwin, T. 2009. The Development of Ethics, vol. 3: From Kant to Rawls. Oxford: Oxford University Press. Kant, I. 1987. Critique of Judgment. Translated by W. S. Pluhar. Indianapolis: Hackett Publishing Company. Kant, I. 1996. Practical Philosophy. Edited and translated by M. Gregor. Cambridge: Cambridge University Press. Kant, I. 1997. Lectures on Ethics. Edited by P. Heath and J. B. Schneewind, translated by P. Heath. Cambridge: Cambridge University Press. Kant, I. 1998. Religion Within the Boundaries of Mere Reason. Edited and translated by A. Wood and G. Di Giovanni. Cambridge: Cambridge University Press. Long, A. A. 2002. Epictetus: A Stoic and Socratic Guide to Life. Oxford: Oxford University Press. Mackie, J. L. 1976. “Sidgwick’s Pessimism.” Reprinted in Essays on Henry Sidgwick, edited by B. Schultz, 163–174. Cambridge: Cambridge University Press. McLeod, O. 2000. “What Is Sidgwick’s Dualism of the Practical Reason?” Pacific Philosophical Quarterly 81: 273–290. O’Connell, E. 2012. “Happiness Proportioned to Virtue: Kant and the Highest Good.” Kantian Review 17 (2): 257–279. Parfit, D. 1984. Reasons and Persons. Oxford: Clarendon Press. Parfit, D. 2011. On What Matters, 2 vols. Oxford: Oxford University Press. Pasternack, L. R. 2014. Kant’s Religion Within the Boundaries of Mere Reason: An Interpretation and Defense (Routledge Philosophy Guidebook Series). London: Routledge. Paytas, T. 2015. “Rational Beings with Emotional Needs: The Patient-Centered Grounds of Kant’s Duty of Humanity.” History of Philosophy Quarterly 32 (4): 353–376. Phillips, D. 2011. Sidgwickian Ethics. New York: Oxford University Press. Plantinga, A. 1974. The Nature of Necessity. Oxford: Oxford University Press.

244  Tyler Paytas Reath, A. 1988. “Two Conceptions of the Highest Good in Kant.” Journal of the History of Philosophy 26 (4): 593–619. Schneewind, J. B. 1977. Sidgwick’s Ethics and Victorian Moral Philosophy. Oxford: Clarendon Press. Schultz, B. 2004. Henry Sidgwick: Eye of the Universe. Cambridge: Cambridge University Press. Shaver, R. 1999. Rational Egoism: A Selective and Critical History. New York: Cambridge University Press. Shaver, R. 2014. “Sidgwick’s Axioms and Consequentialism.” Philosophical Review 123: 173–204. Shields, C. 2012. Ancient Philosophy: A Contemporary Introduction. 2nd ed. New York: Routledge. Sidgwick, A., and E. M. Sidgwick. 1906. Henry Sidgwick: A Memoir. London: Macmillan. Sidgwick, H. 1874. The Methods of Ethics. 1st ed. London: Macmillan. Sidgwick, H. 1876. “A Critical Notice of F. H. Bradley’s Ethical Studies.” Mind 1: 545–549; repr. in Sidgwick 2000 as “Badley’s Ethical Studies”: 185–189. Sidgwick, H. 1877a. The Methods of Ethics. 2nd ed. London: Macmillan. Sidgwick, H. 1877b. “Hedonism and Ultimate Good,” Mind 2: 27–38; repr. in Sidgwick 2000: 89–98. Sidgwick, H. 1889. “Some Fundamental Ethical Controversies.” Mind 14: 473– 487; repr. in Sidgwick 2000: 35–46. Sidgwick, H. 1891. “Is the Distinction between ‘Is’ and ‘Ought’ Ultimate and Irreducible?.” Proceedings of the Aristotelian Society 2: 88–92; repr. as “The Distinction Between ‘Is’ and ‘Ought’ ” in Sidgwick 2000: 59–62. Sidgwick, H. 1902. Outlines of the History of Ethics. 5th ed. London: Macmillan. Sidgwick, H. 1905. Lectures on the Philosophy of Kant and Other Philosophical Lectures and Essays. London: Macmillan. Sidgwick, H. 1981 [1907]. The Methods of Ethics. 7th ed. Indianapolis: Hackett. Sidgwick, H. 2000. Essays on Ethics and Methods. Edited by M. G. Singer. Oxford: Clarendon Press. Skorupski, J. 2007. “Sidgwick and the Many-sidedness of Ethics.” In Henry Sidgwick: Happiness and Religion, edited by P. Bucolo, R. Crisp, and B. Schultz, 410–442. Catania: University of Catania Press. Sussman, D. 2010. “Something to Love: Kant and the Faith of Reason.” In Kant’s Moral Metaphysics, edited by B. J. Bruxvoort Lipscomb and J. Krueger, 133– 148. Berlin: de Gruyter. White, N. 2002. Individual and Conflict in Greek Ethics. Oxford: Oxford University Press. Wood, A. W. 1970. Kant’s Moral Religion. Ithaca: Cornell University Press. Wood, A. W. 2000. “Kant vs. Eudaimonism.” In Kant’s Legacy: Essays Dedicated to Lewis White Beck, edited by P. Cicovacki, 261–282. Rochester: University of Rochester Press.

Acknowledgments

We begin by thanking all of the contributors for their hard work and patience as we were assembling this collection. We are especially grateful to those who took the time to provide feedback on the drafts of other contributors. The editors at Routledge, especially Allie Simmons and Andrew Weckenmann, have been extremely patient and generous in their support of this project. We also thank Robert Audi, Mark Timmons, and three anonymous reviewers for helpful advice. Some of the chapters in this volume were originally prepared for a workshop on Kant and Sidgwick held at the Institut für Philosophie at Universität Stuttgart in 2016. We thank all of those, in addition to the authors, who contributed to the success of that event. The workshop was funded by the VolkswagenStiftung, and we are grateful for their support. We also thank Universität Stuttgart and Washington University in St. Louis for sponsoring the VolkswagenStiftung Postdoctoral Fellowship that brought Tyler to Germany to pursue his research on Kant and Sidgwick. William F. Tate, Dean of the Graduate School of Arts & Sciences at Wash U, and Wilhelm Krull, former Secretary General of the VolkswagenStiftung, deserve special mention for facilitating this opportunity. We also thank Julia Driver for supporting Tyler’s fellowship application (not to mention six years of generous mentorship and supervision). Finally, we are grateful to Australian Catholic University and Universität Stuttgart for providing wonderful environments in which to work as we were collaborating on this project.

Contributors

Carla Bagnoli is Professor of Philosophy at the University of Modena. In addition to articles on moral epistemology and action theory, Bagnoli has published four monographs on moral dilemmas, practical reasoning, moral authority, and responsibility. She is the editor of Morality and the Emotions (Oxford University Press, 2011), and Constructivism in Ethics (Cambridge University Press, 2013). Her current research focuses on rational authority and ascriptive responsibility. Roger Crisp is Professor of Moral Philosophy at the University of Oxford and Uehiro Fellow and Tutor in Philosophy at St Anne’s College, Oxford. His research focuses on normative ethics, metaethics, and the history of ethics. He is the author of Mill on Utilitarianism (Routledge, 1997), Reasons and the Good (Clarendon Press, 2006), The Cosmos of Duty: Henry Sidgwick’s Methods of Ethics (Clarendon Press, 2015), and Sacrifice Regained: Morality and Self-Interest in British Moral Philosophy from Hobbes to Bentham (Clarendon Press, 2019). Paul Guyer is the Jonathan Nelson Professor of Humanities and Philosophy at Brown University and the Florence R. C. Murray Professor in the Humanities emeritus at the University of Pennsylvania. He is the author, editor, and/or translator of more than two dozen books and was the General Co-Editor of the Cambridge Edition of the Works of Immanuel Kant, published in 16 volumes from 1992 to 2016. Among his recent publications are A History of Modern Aesthetics in three volumes (Cambridge University Press, 2014), Virtues of Freedom: Selected Essays on Kant (Oxford University Press, 2016), and Kant on the Rationality of Morality (Cambridge University Press, 2019). Tim Henning is Professor and Chair of Practical Philosophy and History of Philosophy at the University of Stuttgart. He specializes in ethics, metanormativity, Kant, and philosophy of language. His articles have appeared in Ethics, Philosophy & Phenomenological Research, and The Philosophical Review. His book on autonomy titled Person sein

Contributors  247 und Geschichten erzählen (de Gruyter, 2009) was selected for the prestigious Wolfgang-Stegmüller Award. He is the author of a German introduction to Kant’s ethics (Reclam, 2016) and From a Rational Point of View: How We Represent Subjective Perspectives in Practical Discourse (Oxford University Press, 2018). Sukaina Hirji is Assistant Professor of Philosophy at the University of Pennsylvania. She works both in ancient philosophy and contemporary ethics. Her work in these areas has been published in journals such as Philosophy and Phenomenological Research, Pacific Philosophical Quarterly, and The British Journal for the History of Philosophy. Mariko Nakano-Okuno is Assistant Professor of Bioethics at the University of Alabama-Birmingham School of Medicine. She specializes in normative ethics, metaethics, and bioethics. She has published extensively on utilitarian moral philosophy, especially on Sidgwick, including essays in Etica & Politica/Ethics & Politics and Henry Sidgwick: Happiness and Religion (edited by Bucolo, Crisp, and Schultz), as well as multiple articles in Japanese. Her book Sidgwick and Contemporary Utilitarianism (Palgrave Macmillan, 2011) systematically analyzes Sidgwick’s impact on contemporary ethical theories. Tyler Paytas is Research Fellow in the Dianoia Institute of Philosophy at Australian Catholic University. He specializes in ethics and the history of philosophy. His articles have appeared in Journal of the History of Philosophy, Kantian Review, Phronesis, and Utilitas. He is the co-author (with Nicholas Baima) of Plato’s Pragmatism: Rethinking the Relationship between Ethics and Epistemology (Routledge, forthcoming). David Phillips is Professor and Chair of the Department of Philosophy at the University of Houston. His research interests are in ethics and the history of ethics with a particular focus on British moral philosophers. His work in these areas has appeared in journals such as Journal of Philosophy, Philosophy and Phenomenological Research, and Ethical Theory and Moral Practice. He is the author of Sidgwickian Ethics (Oxford University Press, 2011) and Rossian Ethics: W. D. Ross and Contemporary Moral Theory (Oxford University Press, 2019). Robert Shaver is Professor of Philosophy at the University of Manitoba. He specializes in ethical theory and the history of ethics. He has published widely on Hobbes, Hume, Sidgwick, and Ross, with articles appearing in journals such as Philosophical Review, Ethics, and Journal of the History of Philosophy. In Rational Egoism: A Selective and Critical History (Cambridge University Press, 1999), he examines the most influential arguments for rational egoism in modern philosophy and argues that none of them are cogent.

248  Contributors Anthony Skelton is Associate Professor in the Department of Philosophy at the University of Western Ontario. He conducts research in the areas of normative ethics and the history of ethics. His work has appeared in the journals Ethics, Canadian Journal of Philosophy, Journal of the History of Philosophy, and Utilitas. He is the author of Sidgwick’s Ethics (Cambridge University Press, forthcoming). He is an associate editor of the Journal for the History of Analytical Philosophy.

Index

anti-moralism see legal anti-moralism Aristotelianism see eudaimonism Aristotle: Nicomachean Ethics (NE) 204; Rhetoric 204; see also eudaimonism Audi, Robert 112 – 113, 130 Aurelius, Marcus 210 authority 55 – 56, 61 – 64, 69 – 77, 81n35, 99 – 101, 211 – 212 autonomy 61, 70 – 76, 80n25, 113 – 115, 217 avarice 27 Bagnoli, Carla 5, 246 Beaton, R.S. 193 beneficence 22 – 29 benevolence 7, 25 – 28, 37n12, 107, 111 – 116, 124 – 131, 132n7, 134n21 – 22, 134n26, 135n27, 142 – 147, 227 – 230; see also rational benevolence Bentham, Jeremy 2 – 3, 65 Broad, C. D.: Five Types of Ethical Theory 4 Brown, S. M., Jr. 54 burden of proof 64 – 65, 146 Butler, Joseph 15 – 16, 35, 75, 99, 125 – 126 capacity 115, 156 – 158, 174 Categorical Imperative (CI) 3, 5, 85, 132n8, 163, 188; freedom and happiness 145, 148, 157, 159, 160n6; philosophical intuitionism 110 – 113, 115 – 116; practical ethics 16, 37n8; practical knowledge and rational action 76 – 77, 81n33, 81n36,

81n38; self-defeating moral theories 40 – 41, 43 – 44, 47, 53 – 54, 56 choice (Willkür) 165 – 169 Christianity 235 – 237 “civic state”: the duty to live in 49 – 51 clarity 109 Clarke, Samuel 3, 107, 112 – 113, 124 – 125, 153 coercion see deliberation-invasive coercion common sense 4 – 5, 62, 88, 108, 132n7; Aristotelian eudaimonism 190 – 191; freedom and happiness 144 – 145, 149 – 150; the moral necessity of God 229; practical ethics 18, 23 – 24, 31, 35 – 36 conciliation 120 conflict claim 86 – 88, 91 – 92, 98, 102, 103n6 consistency 109 cosmic justice 232 – 238; the need for 232 – 235 Crisp, Roger 6, 119, 124 – 125, 130 – 131, 246 Critique of Judgment (Kant) 192, 210, 213, 216, 218, 220 – 221 Critique of Pure Reason (Kant) 13, 71, 74, 80, 170 Critique of the Power of Judgment (Kant) 172, 188 – 189 Critique of Practical Reason (Kant) 9; Aristotelian eudaimonism 188 – 189, 192; Dialectic of Pure Practical Reason 211 – 213; freedom of will, morality, and responsibility 165 – 166, 169, 171, 175; practical ethics 14, 16 – 18, 25, 27, 29; practical knowledge and rational action 72 – 73

250 Index Deigh, John: interpretation of Methods 1.3 90 – 92 deliberation-invasive coercion 51 – 53 deontology see “indirect deontology” Dependence Thesis (Abhängigkeitsthese) 53 Descartes, René: Rules for the Direction of the Mind 144 desire 2 – 6, 14 – 16, 78n5, 78n7, 126; Aristotelian eudaimonism 188 – 191, 194, 197, 205 – 206; freedom and happiness 155 – 157; freedom of will, morality, and responsibility 166 – 170, 181n7; moral motivation 90 – 91, 96 – 102; the moral necessity of God 213 – 214, 219 – 220, 242n37; motivated 98; non-rational 86, 94, 98, 111; philosophical intuitionism 114 – 115 determinacy 142, 144 – 145 Devlin, Lord 45 dignity 22, 156 discipline 152, 236 dissensus 6 Doctrine of Right (Kant) 5, 40 – 41, 44 – 46, 49 – 50; see also Metaphysics of Morals guidance 32 – 33, 44, 230, 241n35 dogmatic intuitionism 6, 66 – 67, 73, 80n20, 108, 142 – 144 dogmatic rationalism 73 – 74 dualism 187, 191 – 192 duties of right 46 – 49 duty: to live in a “civic state” 49 – 51 egoism 123, 143 – 144, 183n17; Aristotelian eudaimonism 198 – 200, 204 – 206; the moral necessity of God 223 – 229, 241n23, 241n26 – 27 egoistic hedonism 143 – 144, 227; see also egoism empiricism 15 “End of All Things, The” (Kant) 210, 216 endoxa (common opinions) 192, 199 Epictetus 235 – 236, 238, 242n37 – 38 ethical objectivity 72 – 74, 80n29, 128 ethics: the assumption of freedom in 175; the profoundest problem of 222 – 224; see also practical ethics

eudaimonism 187, 193 – 196, 204 – 206; differing interpretations of 200 – 204; rejection of 187 – 193; strong eudaimonism 196 – 200 faith see moral faith feeling 101, 166 – 168, 183n17, 215; moral feeling 72, 74, 80n28, 81n35, 101, 150 – 151 Formal Logic 147 Frankena, W. 78 – 79n9, 231 – 232 freedom 49 – 50, 141 – 143, 160n14, 165 – 171; the assumption of freedom in ethics 175; basic meaning in Kant 169 – 170; importance in Kant’s moral philosophy 170 – 171; Sidgwick’s critique of Kant 147, 149, 151 – 153, 157 – 159, 171 – 179 general good 17, 124, 130, 146, 191, 224 Gibbard, A. 100 Glaube 219, 240n14 God 210 – 211, 238; interpretive puzzles 224 – 226; Kant on the Summum Bonum 211 – 213; moral faith 219 – 222; the necessity of happiness 218 – 219; the need for cosmic justice 232 – 235; the need for an ultimate end 213 – 218; the no-self-sacrifice interpretation 226 – 232; the profoundest problem of ethics 222 – 224; a stoic alternative 235 – 238 Groundwork of the Metaphysics of Morals (Kant): Aristotelian eudaimonism 188 – 189; freedom and happiness 142, 147, 150, 153 – 157; freedom of will, morality, and responsibility 166 – 167, 169, 171 – 173, 175; the moral necessity of God 210, 218, 239n2; peer disagreement 125 – 131; philosophical intuitionism 107, 110, 113 – 114; practical ethics 16 – 19, 22, 25, 27 – 30; practical knowledge and rational action 73 – 75; self-defeating moral theories 41 Guyer, Paul 7, 181 – 182, 246

Index  251 Habermas, J. 49 happiness 142 – 143, 145, 147 – 148, 150 – 151, 153 – 154, 156 – 159; the necessity of 218 – 219 Hart, H.L.A. 45, 53 Henning, Tim 5, 246 – 247 heteronomous doctrines 71 – 73 heteronomy: Kant on 113 – 116 highest good 213 – 222; Summum Bonum 32; Kant on 211 – 213; supreme good 190, 212, 219 Hirji, Sukaina 8, 247 Hobbesian view 49 – 50 Horn, C. 53 – 54 humanity 18 – 20, 127, 130 – 131 Hume, David 62, 65, 73 – 74, 86, 93 – 96, 247; see also Humean view Humean view 6, 72 – 73, 74, 79n10; moral motivation 84, 86 – 93, 96 – 100, 102, 103n4, 103n6 Hurka, Thomas 78n5, 132n5, 240n20 Hursthouse, Rosalind 198 – 199 Hutcheson, Francis 81n37, 99, 113, 142 hypothetical imperatives 46, 114, 157, 188 impulse 165 – 169 inclination 166 independence 120 – 121 “indirect deontology” 5, 41, 56 – 57 indirect moral theory 56 – 57 intuitional morality 223 – 224 intuitionism see dogmatic intuitionism; perceptional intuitionism; philosophical intuitionism Irwin, Terry 112 – 113, 197 – 199, 201 Jones, E. E. Constance 84, 102 – 103n1, 103n8, 141 judgment 71; see also moral judgment justice 107, 115 – 116, 123 – 125, 128 – 130, 134n21; Aristotelian eudaimonism 191 – 192, 199 – 200; cosmic 211, 232 – 235, 238; freedom and happiness 144 – 145, 158 – 159; the moral necessity of God 210 – 211

kalon 202 Kant, Immanuel 163 – 164; argument for legal anti-moralism 52 – 53; basic meaning of “freedom” in 169 – 170; Critique of Judgment 213; Critique of Practical Reason 9; on duties of right 46 – 49; on the duty to live in a “civic state” 49 – 51; Groundwork 19, 25; on heteronomy 113 – 116; the highest good in later works 213 – 222; “Idea For a Universal History” 49; importance in Kant’s moral philosophy 170 – 171; Lectures on Ethics 19; moral knowledge 69 – 71; on the moral necessity of God 210 – 211, 232 – 236; “On the Common Saying” 49 – 51; and peer disagreement 119, 123 – 131; philosophical intuitionism 110 – 113; practical ethics in 13 – 14, 16 – 22, 25 – 36; on practical knowledge and rational action 61 – 62, 64, 69 – 78; practical proof 71 – 73, 74 – 77; Religion within the Boundaries of Mere Reason 141, 143, 213; repercussions of the argument for the political philosophy of 53 – 56; Sidgwick’s criticisms of Kant’s idea of freedom 171 – 179; Sidgwick’s critique of 141 – 159; Sidgwick’s Kantian account of moral motivation 84, 86, 92, 100 – 102; as Sidgwick’s peer 123 – 130; on the Summum Bonum 211 – 213; “Theory of Practice” 213; see also specific works King, Nathan 122 knowledge see moral knowledge; practical knowledge Kraut, R. 201 – 203 laws see practical laws Lear, G.R. 197 – 198, 202 – 203, 205, 207 – 208n11 Lectures on Ethics (Kant) 19, 27, 33 Lectures on the Philosophy of Kant and Other Philosophical Lectures and Essays (Sidgwick) 151 legal anti-moralism 52 – 53 legal moralism 45 – 46 Locke, John: Conduct of the Understanding 144 lying 17 – 22

252 Index maxims 16 – 18, 25 – 26; freedom and happiness 145 – 152, 154 – 155, 160n10; freedom of will, morality, and responsibility 166, 173 – 174; the moral necessity of God 211, 229; peer disagreement 129 – 130, 132n7 – 8, 134n21; philosophical intuitionism 110 – 112; practical knowledge and rational action 77; self-defeating moral theories 43 – 44, 47 – 48, 51 – 56 McCarty, Richard 101 McDowell, J. 197 Metaphysics of Morals (Kant) 2; freedom and happiness 147, 150 – 151, 153 – 158; freedom of will, morality, and responsibility 165 – 173, 181n6; the moral necessity of God 218; peer disagreement 125, 130; practical ethics 18 – 19, 22, 25 – 28, 31, 34, 37n11; practical knowledge and rational action 72, 80n21, 80n28; self-defeating moral theories 46 Methods of Ethics (Sidgwick) 1, 3; denial that there is an overall argument in 1.3 88 – 90; freedom and happiness 141 – 142, 153, 160n14; freedom of will, morality, and responsibility 167, 177, 180n1, 182n12; interpretations of the argument of 1.3 87 – 94; moral motivation 84, 96, 102; the moral necessity of God 222 – 225, 228, 230, 233 – 234; peer disagreement 120, 126, 130, 134n19; philosophical intuitionism 110, 116; practical ethics 21, 31; practical knowledge and rational action 61; self-defeating moral theories 40 – 41; Shaver/Deigh interpretation of 1.3 90 – 92; structure of the argument of 1.3 85 – 87 Mill, J. S. 14 – 17; On Liberty 165; Utilitarianism 14 – 15 Moore, G.E. 65, 78n6, 86 – 87, 89, 103n2 moral faith 219 – 222 moralism see legal anti-moralism; legal moralism moralistic behavior 45 morality as such 45, 53, 55

moral judgment: Sidgwick’s definition of 62 – 64 moral knowledge: Kant’s conception of 69 – 71; as practical knowledge 61 – 62, 68 – 69 moral law 101, 164, 167 – 174; the moral necessity of God 212 – 220; freedom and happiness 150 – 152; practical knowledge and rational action 72 – 75, 81n35; self-defeating moral theories 43 – 44, 46 – 47, 53 – 57 moral motivation 84 – 85, 94 – 102 moral need 211, 217, 220, 238 moral obligation 56 – 57, 147, 214 moral resolve 232 – 238 moral responsibility 164, 169, 171, 173, 175, 178 – 179 Moral Sense School 76 moral theories see self-defeating moral theories motivation see moral motivation Nagel, T. 98, 101 Nakano-Okuno, Mariko 7, 247 naturalism: Sidgwick’s argument against 64 – 65 natural need 215 Nazis 148 – 149, 154 – 155, 181 – 182n11 neutral freedom 13 – 14, 172 non-coercion 169 – 170 non-dissensus 109 non-naturalness claim 6, 87 – 92, 94, 102, 103n6 normative positivist 51 no-self-sacrifice interpretation 226 – 232 object-seeking impulse 165 Observations on the Feeling of the Beautiful and the Sublime and other Writings (Kant) 30 “On the Common Saying: That May Be Correct in Theory, but It Is of No Use in Practice” (Kant) 214, 239n3 “On the Suppose Right to Lie from Philanthropy” (Kant) 19 – 20 onus 123 ought-judgment 91 own good 23 – 24, 146, 190, 224 – 226, 229, 231, 241n27

Index  253 paralogism 3, 156 Parfit, Derek: On What Matters 1; Reasons and Persons 40; selfdefeating moral theories in 41 – 44; Volume Three 9 partial dualism 227 Paytas, Tyler 8 – 9, 117n6, 132n7, 133n13, 134n24, 247 peer disagreement 6, 118 – 119; conciliation 120; and Crisp 130 – 131; independence 120 – 121; Kant as Sidgwick’s peer 123 – 130; onus 123; peers 119 – 120; skepticism 122 peers 119 – 120; Kant as Sidgwick’s peer 123 – 130 perceptional intuitionism 65 – 66, 108 perfection 151 – 152, 223 personal irrelevance 124, 130 Phillips, David 5 – 6, 240n20, 247 philosophical intuitionism 6, 67 – 68, 79n19, 107 – 116 philosophic mind 108, 118 phronimos 202 Plato 116, 153, 190, 192, 204; Euthydemus 192 political philosophy: repercussions of the argument for Kant’s political philosophy 53 – 56 postulates 54, 56 – 57, 131, 212 – 213, 219, 231 practical ethics 13 – 14, 36; and beneficence 22 – 29; disagreements of Kant and Sidgwick 29 – 36; and lying 17 – 22; and Mill 14 – 17 practical importance 177 – 179 practical knowledge 61, 77 – 78; and the argument against dogmatic rationalism 73 – 74; and Kant’s conception of moral judgment 69 – 71; and Kant’s practical proof 74 – 77; moral knowledge as 61 – 62; practical proof and heteronomous doctrines 71 – 73; and Sidgwick’s argument against naturalism 64 – 65; and Sidgwick’s critique of dogmatic and perceptional intuitionism 65 – 66; and Sidgwick’s definition of moral judgment 62 – 64; and Sidgwick’s failure to establish that moral knowledge is practical 68 – 69; and Sidgwick’s philosophical intuitionism 67 – 68

practical laws 74 – 77 practical postulates 56 – 57; see also postulates practical proof 71 – 73, 74 – 77 practical reason 187, 189 – 192, 201, 203 – 204 practical subjects: laws for 74 – 77 praxis 197 Principia Ethica (Moore) 87, 103n2 prohairesis 197 proof see burden of proof; practical proof propositions 68 – 69 prudence 113 – 116, 124 – 127, 130 – 131, 133n15, 144 – 146 Rashdall, Hastings 21 rational action 61, 77 – 78; the argument against dogmatic rationalism 73 – 74; Kant’s conception of moral judgment 69 – 71; Kant’s practical proof 74 – 77; moral knowledge as practical knowledge 61 – 62; practical proof and heteronomous doctrines 71 – 73; Sidgwick’s argument against naturalism 64 – 65; Sidgwick’s critique of dogmatic and perceptional intuitionism 65 – 66; Sidgwick’s definition of moral judgment 62 – 64; Sidgwick’s failure to establish that moral knowledge is practical 68 – 69; Sidgwick’s philosophical intuitionism 67 – 68 rational benevolence 3, 23, 25, 27; freedom and happiness 142, 146 – 147, 150, 153, 155 – 156, 158 – 159; freedom of will, morality, and responsibility 163, 180n2, 182n15; the moral necessity of God 228 – 230; peer disagreement 127; philosophical intuitionism 107, 111 – 112, 114 rational freedom 172 rationalism see dogmatic rationalism; rational action Rawls, John 79n13, 81n33, 81n38, 134n21 – 22, 160n10, 180n2; philosophical intuitionism 107, 116n2 reasoning see process of reasoning Rechtslehre see Doctrine of Right reflection 109

254 Index Religion within the Boundaries of Mere Reason (Kant) 8, 151, 192, 210, 214 – 215, 217 – 218, 239n36 rightness 5, 41, 56 – 57, 164 Ross, W.D. 37n8, 131 Rousseau, Jean-Jacques 30 Scanlon, T. M.: Being Realistic About Reasons 85, 95 – 98; Kant as peer 123 – 130; Lecture 3 96; What We Owe to Each Other 98 Schneewind, J. 6, 88 – 90, 226 – 227, 241n25, 241n28 self-defeating moral theories 40 – 41; collectively 42; deliberation-invasive coercion 51 – 53; directly 42; indirectly 42; individually 42; Kant on duties of right 46 – 49; Kant on the duty to live in a “civic state” 49 – 51; legal moralism 45 – 46; morality as an indirect guide 44 – 45; practical postulates and “indirect deontology” 56 – 57; reactively 45; repercussions of the argument for Kant’s political philosophy 53 – 56; in Sidgwick and Parfit 41 – 44 self-effacing 42, 198, 20 – 6, 207n8 self-emanation 169 – 170 self-legislation 71, 115, 169 – 170, 174 self-love 26, 75, 81n36, 99; Aristotelian eudaimonism 188, 198; freedom and happiness 141, 151 Separation Thesis (Trennungsthese) 53 – 54 Shaver, Robert: interpretation of Methods 1.3 90 – 92, 241n26, 247 Sidgwick, Henry 163 – 164; argument against dogmatic rationalism 73 – 74; argument against naturalism 64 – 65; beneficence 22 – 25, 28 – 29; conciliation 120; Crisp 130 – 131; criticisms of Kant’s idea of freedom 171 – 179; critique of dogmatic and perceptional intuitionism 65 – 66; critique of Kant 141 – 143, 147 – 159; defensibility of 101 – 102; definition of moral judgment 62 – 64; disagreement with Kant 29 – 36; independence 120 – 121; interpretive puzzles 224 – 226; intuitionisms of 108 – 109; Kantian account of moral motivation

84 – 85; Kant as peer 123 – 130; Kant’s practical proof 74 – 77; Kant on the Summum Bonum 211 – 213; Lectures on the Philosophy of Kant 151; lying 17 – 22; “methods of ethics” 143 – 147; Mill 14 – 17; moral faith 219 – 224; moral knowledge 61 – 62, 68 – 69; on moral motivation 94 – 101; on the moral necessity of God 210 – 211, 238; the necessity of happiness 218 – 219; the need for cosmic justice 232 – 235; the need for an ultimate end 213 – 218; no-self-sacrifice interpretation 226 – 232; onus 123; peer disagreement 118 – 119; peers 119 – 120; philosophical intuitionism 67 – 68; practical ethics in 13 – 14, 36; Practical Ethics 21; on practical knowledge and rational action 61, 77 – 78; self-defeating moral theories in 41 – 44; skepticism 122; the Stoic alternative 235 – 238; see also Methods of Ethics Skelton, Anthony 4 – 5, 79n15, 133n16, 248 skepticism 122 Smith, Michael 96 – 98, 101 Spinoza, Baruch: Emendation of the Intellect 144 Stephen, Leslie 92 – 93, 96 Stocker, Michael 42 – 43 Stoics 153, 189, 211, 235 – 238, 239n4, 241n31, 242n35 – 37 Socrates 153, 192, 196 subjective necessity 175 – 177 subordination 50 teleion 194 – 195 telos see ultimate end Tennyson, Alfred 210, 234 transcendental freedom 165, 169 – 171, 174, 180, 181 – 182n11 ultimate end 8, 152, 177; Aristotelian eudaimonism 190 – 191, 197, 201, 205; the moral necessity of God 212 – 220, 224, 227, 229 – 230, 232, 240n21; the need for 213 – 218 universal good 23, 225 – 226, 228, 230 – 231, 234

Index  255 universalistic hedonism see utilitarianism universal law 52 Universal Principle of Right 54 – 57 violence 50 – 51 Vlastos, G. 196 – 197

Waldron, J. 51 Whewell, William 108 Whiting, Jennifer 193, 199 will (Wille) 165 – 169 Willaschek, M. 46 Williams, Bernard 43, 77 Williamson, Tim 107 Wolff, Christian 113 – 115