248 53 955KB
English Pages [188] Year 2021
Kant’s Justification of Ethics
Kant’s Justification of Ethics OW E N WA R E
1
1 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Owen Ware 2021 The moral rights of the author have been asserted First Edition published in 2021 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2020945986 ISBN 978–0–19–884993–3 DOI: 10.1093/oso/9780198849933.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.
Preface I once heard it said that because of the sheer amount of time it takes, the person who begins writing a book is not the same person who finishes it. Nothing could be more true about my experience working on this project. The research began in the fall of 2006 when I enrolled in ‘Kant’s Ethics’, a PhD seminar co-taught by Sergio Tenenbaum and Arthur Ripstein at the University of Toronto. My final paper ended up becoming my first publication on Kant, ‘The Duty of SelfKnowledge’ (2009). After my area examination, I decided to pursue my dissertation project on Kant’s moral philosophy under the supervision of Paul Franks, which I defended in the summer of 2010. In the years that followed, I abandoned some of the views I upheld in my dissertation, went on to pursue different aspects of Kant’s ethics, and even ventured into the world of post-Kantian philosophy, writing a book on Fichte (2020) along the way. How the present book came into being is a bit of a mystery, even to me, but I can identify one set of encounters that made its existence possible. During the 2015–2016 academic year I lived in Frankfurt as a Humboldt Research Fellow at the Goethe-Universität. It was during that year that I had the pleasure of meeting Gabriele Gava, who showed me a hospitality that it seems only Italians are capable of. It was during that year, involving many conversations over espresso, that Gabriele helped me see Kant’s methodology (specifically his distinction of analytic and synthetic procedures) as a key for unlocking the crit ical system. What I learned from Gabriele helped bring about a kind of revolution in my way of understanding Kant, and it was largely thanks to him that I was able to piece together a new way of applying the analytic-synthetic distinction to Kant’s project of moral justification in the Groundwork and the second Critique. In Frankfurt I also benefitted from conversations with Marcus Willaschek and members of his Kant-Arbeitskreis. Among the members of the Arbeitskreis, I was fortunate to become friends with Thomas Höwing, whose warmth and friendliness made my stay in Germany one I shall never forget. Many other individuals helped me prior to, and after, my stay in Frankfurt. In addition to my dissertation committee members and current colleagues, Sergio Tenenbaum and Arthur Ripstein, I must first thank two individuals from my student days at the University of Toronto who I am now proud to call friends: Anthony Bruno and Ariel Zylberman. The amount of care they have both shown in helping me refine my work is staggering. I am also grateful to Karl Ameriks, Barbara Herman, and Robert Stern for providing me with feedback and
vi Preface encouragement during the early years of my post-dissertation life, first as a postdoctoral fellow and then as an assistant professor. One will see in this book just how much the spirit of my reading of Kant is shaped by their scholarship. I have benefitted from regular conversations with colleagues at the institutions where I worked, including Kristen Gjesdal, David Wolfsdorf, Sam Black, Evan Tiffany, Dai Heide, and many others. One could not ask for a better colleague than Dai; and he helped me see—more than anyone else—the importance of Kant’s metaphysics for understanding his practical philosophy. Other people who have helped shape the views I present in this book, either through written or spoken feedback, include Don Ainslie, Stefano Bacin, Jochen Bojanowski, Claudia Blöser, John Callanan, Robert Clewis, Ben Crowe, Janelle DeWitt, Stephen Engstrom, Michael Forster, Markus Gabriel, Sebastian Gardner, Jeanine Grenberg, Hannah Ginsborg, Paul Guyer, Andree Hahmann, Bob Hanna, Chris Herrera, Karolina Hübner, Thomas Khurana, Karin Nisenbaum, Sven Nyholm, Lara Ostaric, Markus Kohl, Bernd Ludwig, Dean Moyar, Heiko Puls, Andrews Reath, Francey Russell, Lisa Shapiro, Feroz Shah, Irina Schumski, Nick Stang, Martin Sticker, Oliver Sensen, Ulrich Schlösser, Joe Saunders, Krista Thomason, Allen Wood, Benjamin Yost, my two Oxford University Press reviewers, and my supportive editor, Peter Momtchiloff. I do not know how I can repay all these debts of gratitude, and all I can say when I think of these individuals is thank you. I am grateful to have the permission to reuse portions of my previously published articles, many of which have been revised or rewritten here. Chapter 1 draws from ‘Skepticism in Kant’s Groundwork’, European Journal of Philosophy (2016). Chapter 2 draws from ‘Rethinking Kant’s Fact of Reason’, Philosophers’ Imprint (2014). Chapter 3 draws from ‘Kant’s Deductions of Morality and Freedom’, Canadian Journal of Philosophy (2017). Chapter 4 draws from ‘Kant on Moral Sensibility and Moral Motivation’, Journal of the History of Philosophy (2014) and ‘Accessing the Moral Law through Feeling’, Kantian Review (2015). And Chapter 5 draws from ‘The Duty of Self-Knowledge’, Philosophy and Phenomenological Research (2009). While writing this book, I have been fortunate to receive the love, support, and companionship of Leah Ware. She makes it all worth it. Last, but not least, I have had the privilege of teaching many talented students at Temple University, Simon Fraser University, and at my current institution, the University of Toronto. When I wrote this book, I had my students at the forefront of my mind. They have helped me grow as a writer, scholar, and aspiring philosopher more than they possibly could know. This book is dedicated to them.
Preface vii Italo Calvino once wrote that ‘every reading of a classic is in fact a rereading’, from which it follows, he added, that ‘a classic is a book that has never finished saying what it has to say.’1 I find these definitions fitting when it comes to the Groundwork and the second Critique, for both of these texts contain—despite their relatively small size—a conceptual richness and systematicity that few other works in the history of ethics have attained. It is perhaps this seemingly inexhaustible quality that underlies the variety of interpretations and controversies that continue to surround Kant’s moral philosophy today. My method for reading Kant in this book is to take the principle of charity to heart and try, as best as I can, to present his arguments in a coherent, consistent, and unified manner. This is not because I think Kant should have the last word, but because I find the principle of charity the most powerful tool for unlocking a philosopher’s position. I do not pretend to have solved all the mysteries of Kant’s project of justifying e thics in the Groundwork and the second Critique; nor do I regard my interpretations here as final or beyond fixing. But this much is clear: Whether or not the reader is sympathetic to Kant, the fact stands that Kant is a ‘classic author’ as Calvino defines it: ‘the one you cannot feel indifferent to, who helps you to define yourself in relation to him, even in dispute with him.’2
1 Calvino (1982, 128).
2 Calvino (1982, 130).
Abbreviations In the case of the Critique of Pure Reason, I follow the standard practice of referring to the 1781 (A) and 1787 (B) editions. For all other texts, citations appear in the order of abbreviation, volume number, and page number from the Akademie Ausgabe, Kants Gesammelte Schriften, edited by Königlich Preussische Akademie der Wissenschaften (29 vols. Berlin: de Gruyter, 1900ff). Translation decisions are my own, though I have consulted (and sometimes followed) The Cambridge Edition of the Works of Immanuel Kant, edited by Paul Guyer and Allen Wood (Cambridge: Cambridge University Press, 1992ff).
Kant’s Works A/B Anth Br FM
G GSE ID
KpV KU MpVT MS ND NVE
OP Päd R
Kritik der reinen Vernunft (Critique of Pure Reason), 1781/87. Die Anthropologie in pragmatischer Hinsicht (Anthropology from a Pragmatic Point of View), 1798. Briefe (Letters), various dates. Welches sind die wirklichen Fortschritte, die die Metaphysik seit Leibnitzens und Wolf ’s Zeiten in Deutschland gemacht hat? (What Real Progress has Metaphysics Made in Germany since Wolff?), 1793/1804. Grundlegung zur Metaphysik der Sitten (Groundwork for the Metaphysics of Morals), 1785. Beobachtungen über das Gefühl des Schönen und Erhabenen (Observations on the Beautiful and the Sublime), 1764. De Mundi Sensibilis atque Intelligibilis Forma et Principiis (Inaugural Dissertation on the Form and Principles of the Sensible and the Intelligible World), 1770. Kritik der praktischen Vernunft (Critique of Practical Reason), 1788. Kritik der Urteilskraft (Critique of the Power of Judgment), 1790. ‘Über das Mißlingen aller philosophischen Versuche in der Theodicee’ (‘On the Miscarriage of All Philosophical Trials in Philosophy’), 1791. Die Metaphysik der Sitten (The Metaphysics of Morals), 1797. Nova dilucidatio (A New Elucidation of the First Principles of Metaphysical Cognition), 1755. ‘Nachricht von der Einrichtung seiner Vorlesungen in dem Winterhalbenjahre von 1765–1766’ (‘Announcement of the Programme of his Lectures for the Winter Semester of 1765–1766’), 1765. Opus Postumum. Pädagogik (Education), 1803. Die Religion innerhalb der Grenzen der bloßen Vernunft (Religion witihin the Boundaries of Mere Reason), 1793.
xii Abbreviations Refl RezSchulz TP
ÜGTP V-Met-L2/Pölitz V-Met/Mron V-Met/Heinze V-Mo/Collins V-Phil-Th/Pölitz V-PP/Herder WDO
Reflexionen, various dates. Recension von Schulz’s Versuch einer Anleitung zur Sittenlehre für alle Menschen (Review of Schulz), 1784. Über den Gemeinspruch: Das mag in der Theorie richtig sein, taugt aber nicht für die Praxis (On the common saying: That may be correct in theory, but it is of no use in practice), 1793. Über den Gebrauch teleologischer Principien in der Philosophie (On the Use of Teleological Principles in Philosophy), 1788. Kant Metaphysik L 2 Pölitz (Lectures on Metaphysics Pölitz), 1790/1791. Vorlesungen Wintersemester 1782/1783 Metaphysik Mrongovius (Lectures on Metaphysics Mrongovius), 1782/1783. Kant Metaphysik L1 Heinze (Lectures on Metaphysics Heinz), 1770–1775. Vorlesungen Wintersemester 1784/1785. Moralphilosophie Collins (Lectures on Moral Philosophy Collins), 1784/1785. Vorlesungen Wintersemester 1783/1784 Philosophische Religionslehre nach Pölitz (Lectures on the Doctrine of Religion Pölitz), 1783/1784. Praktische Philosophie Herder (Practical Philosophy Herder), 1762/1763. Was heißt sich im Denken orientiren? (What Does it Mean to Orient Oneself in Thinking?), 1786.
Plato was right in raising this question and asking, as he used to do, Are we on the way from or to the first principles? There is a difference as there is in a racecourse between the course from the judges to the turning-point and the way back. Aristotle, Nicomachean Ethics (1.4) The schools are instructed to pretend to no higher or more comprehensive insight on any point touching the universal human concerns than the insight that is accessible to the great multitude (who are always most worthy of our respect), and to limit themselves to the cultivation of those grounds of proof alone that can be grasped universally and are sufficient from a moral standpoint. Kant, Critique of Pure Reason (B xxxiii)
Introduction The Quiet Avoidance of Justification
When we survey the history of ethics in the West from the ancient Greeks to the present, we see that approaches to moral justification divide roughly into two camps. The first camp starts with a minimal set of premises, say, a thin conception of what it means to be a rational agent, and it then proceeds to derive a substantive account of moral requirements and their normativity. This would be an ambitious strategy of justification, since it tries to get a lot (moral normativity) from very little (a mere capacity to respond to reasons). While clearly a strong version of the ambitious approach, we can easily identify more moderate examples that still share the same form, such as social contract theories that only appeal to self-interest and the instrumental principle. The second camp goes about the task of justification the other way around. Instead of starting with very little, advocates of this strategy start with more—perhaps a whole lot more—such as a conception of thick ethical virtues, and then they work to explain their function, say, for securing an agent’s flourishing. This is what I would call a modest strategy of justification, for instead of trying to get us into the world of morality, it already presupposes our home within that world. The second approach merely attempts to make the standpoint of morality intelligible for those already committed to it. Ever since the revival of interest among English-speaking scholars in Kant’s philosophy, dating from the 1960s, one can detect a boundary line dividing those who approach his ethics through the framework of the ambitious strategy and those who approach it through the framework of the modest strategy. Among the former we can include the work of Thomas Nagel, David Velleman, and Christine Korsgaard; among the latter, the work of John Rawls, Barbara Herman, and Thomas Hill.1 A parallel boundary line also divides scholars working on Kant’s theoretical philosophy, especially with regard to his transcendental deduction of the categories. Those who give the deduction an ambitious interpretation, according to which we get a lot (the validity of the categories) from very little (a mere capacity for representation) include Peter Strawson, Jonathan Bennett, and Robert Wolff.2 Among those who give the deduction a more modest interpretation, we can 1 See Nagel (1970), Velleman (1989, 2000, 2009), Korsgaard (1996a, 1996b, 2008, 2009), Rawls (1989), Herman (2007), and Hill (2012). 2 See Strawson (1966), Bennett (1966), and Wolff (1963). Kant’s Justification of Ethics. Owen Ware, Oxford University Press (2021). © Owen Ware. DOI: 10.1093/oso/9780198849933.003.0001
2 introduction identify Karl Ameriks, Stephen Engstrom, and Robert Stern.3 What motivates these two boundary lines is difficult to tell, but I believe it is partly the result of the meta-philosophical assumptions these writers uphold. In particular, what seems to divide their positions is an attitude toward skepticism, with the first group arguing that Kant’s aim in his theoretical and moral philosophy is to refute the skeptic, and the second group arguing instead that Kant was either uninterested in skepticism or, at most, only concerned to diagnose it.4 My reasons for drawing attention to the reception of Kant’s philosophy is to show just how surprising it is that after sixty years of scholarship, the literature devoted to Kant’s justification of ethics is still rather thin. Some time ago Barbara Herman described the state of affairs as follows: In the resurgence of work on Kant’s ethics, one notices the quiet avoidance of the issue of justification. This is to some extent the harmless by-product of a new enthusiasm generated by success with the substantive ethical theory. But the other thing at work, I believe, is the suspicion that the project of justification in Kantian ethics is intractable.5
Since Herman wrote these words in 1989, a handful of monographs have appeared devoted to what were once ‘scarcely charted regions’ in Section III of the Groundwork for the Metaphysics of Morals and the first chapter of the Critique of Practical Reason.6 For the most part, however, the suspicion Herman mentions is very much alive, since scholars continue to disagree over nearly every aspect of Kant’s arguments for freedom and morality. While it is not the goal of my book to settle these controversies once and for all, I hope to show that the parameters of these disputes have been unduly narrow, and that Kant’s project of justification links up, in deep and interesting ways, with his theories of moral motivation, moral feeling, and moral conscience. Like advocates of the modest interpretation, I do not think that Kant was seeking to refute the skeptic on the skeptic’s own terms. Nor do I think that Kant ever sought to derive the moral law or its normativity from a thin conception of what it means to be a rational agent, as I will argue in Chapter 3. However, matters are complicated by the fact that Kant’s so-called modest starting point is not a conception of ethical life in the Aristotelian sense, involving thick virtues, or in the Hegelian sense, involving concrete institutions, but in a conception of ‘common 3 See Ameriks (2003), Engstrom (1994), and Stern (1999). 4 There is a precedent in Aristotle, who acknowledges the limited force of arguments in ethics. ‘Now if arguments were in themselves enough to make men good, they would justly, as Theognis says, have won very great rewards, and such rewards should have been provided; but as things are, while they seem to have power to encourage and stimulate the generous-minded among our youth, and to make a character which is gently born, and a true lover of what is noble, ready to be possessed by virtue, they are not able to encourage the many to nobility and goodness’ (1984, X.9). 5 Herman (1989, 131). 6 See Schönecker (1999), Allison (2011), Grenberg (2013), and Puls (2016).
Introduction 3 human reason’ (gemeinen Menschenvernunft). And while the form of Kant’s strategy is modest insofar as it works within this conception, the content is arguably not modest at all. If anything, it is wildly ambitious insofar as Kant frames the ends and interests of common human reason in terms of a moral teleology (regarding our essential ends), a moral theology (regarding ideas of the soul and God), and a seemingly robust metaphysics (regarding a distinction between phenomena and noumena). I want to emphasize this from the outset as a reminder that Kant’s attempt to reorient philosophy from a common standpoint is not an innocent philosophical move. Nor is it simply a matter of describing moral experience from a first-person perspective, free from presuppositions. For Kant, as we shall see, the appeal of a common standpoint is that it reveals a use of reason already bound up with moral, religious, and even metaphysical concepts.7 This aspect of Kant’s position will likely strike readers today as antiquated. But in my view it is something we should acknowledge and, if not come to terms with, then at least accept on the grounds of textual accuracy. For many years it was not unusual to hear philosophers denounce Kant’s approach to ethics on the grounds of its ‘extravagant metaphysical luggage’,8 and many shared Strawson’s view that if there is lasting value to Kant’s philosophy, it lies on the merely conceptual (which is to say non-metaphysical) side of his project.9 This sentiment has enjoyed a long lifespan in the field of Kantian ethics, which to this day appears unaffected by a growing wave of metaphysics-friendly approaches to the first Critique.10 Not that long ago, for instance, one prominent ethicist believed it would attract his readership to declare, quite boldly, that ‘we can be naturalists while preserving the moral and psychological richness of Kant’.11 My own view is that separating Kant’s ethics and his metaphysics cannot be won so easily. Nor am I of the opinion that such a separation would be desirable. At any rate, a conviction that will guide my discussion here is that it is worthwhile to approach Kant on his own terms, extravagant luggage and all, and resist the temptation to make his work conform to contemporary trends. Kant himself was of the view that we ‘always return to metaphysics’ (A850/B878),12 and that such a return is necessary if we want to ground ethics as a science. Accordingly, he tells us, ‘those same people who oppose metaphysics still have an indispensable duty to go back to its principles even in the doctrine of virtue and, before they teach, to become pupils in the classroom of metaphysics’ (MS 6:376–7).13
7 See Ameriks (2000) for further discussion of these presuppositions. 8 Williams (1985, 65). 9 Strawson (1966). 10 See, for example, Adams (1997), Watkins (2005), Ameriks (2003), Hogan (2009), Marshall (2010), Insole (2013), Dyck (2014), Stang (2016), Indregard (2018), and Heide (2020). 11 Velleman (2006, 15). 12 A return, he adds, ‘as to a beloved from whom we have been estranged’ (A850/B878). 13 In this context Kant argues that feeling cannot ground a moral principle. Any such principle of feeling, he adds, ‘is really an obscurely thought metaphysics that is inherent in every human being because of his rational predisposition, as a teacher will readily grant if he experiments in questioning his pupil socratically about the imperative of duty and its application to moral appraisal of his actions’ (MS 6:376–7).
4 Philosophy as Justification Taking this council to heart, my aim here is to understand the basic structure of Kant’s project of moral justification, with a focus on his foundational arguments for the reality of human freedom and the normativity of the moral law. For this reason I shall be limiting my investigation almost exclusively to the two texts where these arguments appear—Groundwork III and the second Critique— though I will often contextualize these arguments within Kant’s corpus at large, including his writings on theoretical philosophy. My chief aim in this book is to give a fresh interpretation of Kant’s justification of ethics that, while true to the spirit of the modest strategy sketched above, reveals the far-reaching significance of his effort to reorient ethics around a shared, pre-theoretical, and hence common standpoint. To this end I will be building upon the excellent work of recent commentators who have shown the extent to which Kant’s commitment to common human reason plays a central role in his philosophy, although I wish to push this reading further by uncovering its deeper systematic function. A distinctive feature of the interpretation I will be developing over the coming chapters is that Kant’s project of justifying the reality of human freedom and the normativity of the moral law turns on a complex set of argumentative strategies that have gone overlooked by most commentators. At their basis, I shall argue, we find a revolutionary view of the relationship between philosophy and what Kant calls our ‘higher vocation’ (höhere Bestimmung).
Philosophy as Justification This revolutionary view is central to the project of critical philosophy Kant sets out in the Critique of Pure Reason, making it a helpful place to begin our investigation. Indeed, the opening lines of the first Critique draw attention to our unavoidable tendency to ask questions that transcend sense experience, such as whether we possess an immortal soul, whether we are free in our actions, or whether God exists. At the same time, Kant is clear that what hangs in the balance of these questions is nothing that will occasion despair for the ordinary person. As he explains, the space where such questions lead constitutes the ‘battlefield . . . of metaphysics’ (A viii). To deprive its combatants of their knowledge-claims is thereby a loss that only effects ‘the monopoly of the schools and in no way the interest of human beings’, since ordinary persons are not entangled in such controversies (B xxxii). After all, Kant asks, have the ‘fine-spun arguments’ of the schools—concerning immortality, freedom, or God—‘ever been able to reach the public or have the least influence over its convictions?’ (B xxxiii). He does not think so. Nor does he think philosophers should pretend to occupy such a position of influence. Instead they should seek ‘no higher or more comprehensive insight on any point touching the universal human concerns than the insight that is accessible to the great multitude (who are always most worthy of our respect),
Introduction 5 and’, Kant adds, ‘to limit themselves to the cultivation of those grounds of proof alone that can be grasped universally and are sufficient from a moral standpoint’ (B xxxiii).14 Upon hearing such a plea for humility, one might think that Kant is rejecting metaphysics altogether. Yet I believe this impression would be mistaken. In add ition to denouncing dogmatism and skepticism in philosophy, Kant goes out of his way to criticize ‘indifferentism’ for its anti-metaphysical character, writing that it signals ‘the mother of chaos and night in the sciences’ (A x).15 As he emphasizes, ‘it is pointless to affect indifference with respect to such inquiries to whose object human nature cannot be indifferent’ (A x), referring once again to those questions that drive human reason to leave its sphere in the sensible world.16 Rather than retreat from the battlefield of metaphysics, then, it is the philoso pher’s duty to establish peace by way of a critical examination of reason itself—its sources, contents, and limits—the process of which Kant says will usher a new era in the development of reason, the era of ‘enlightenment’ (A xi). But contrary to what one might expect, this stage does not privilege the theoretical interests of reason, not even in the domains of well-grounded science such as physics, mathematics, and logic. The aim of depriving speculative reason access to objects beyond the sensible world is to vindicate those objects as items of belief for reason in its practical use. The aim, as Kant famously puts it, is to ‘deny knowledge in order to make room for faith’ (B xxx; see also A744/B772). Behind these claims we find a new model of philosophy at the heart of Kant’s system, what we might call philosophy as justification.17 The inspiration for this model appears to have been occasioned by Kant’s reading of Rousseau, and in a surviving fragment Kant describes the effects of this reading in language reminiscent of a religious conversion: 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 honor
14 Kant nonetheless ends on a more positive note: ‘Yet care is taken for a more equitable claim on the part of the speculative philosopher. He remains the exclusive trustee of a science that is useful to the public even without their knowledge, namely the critique of reason; for the latter can never become popular, but also has no need of being so; for just as little as the people want to fill their heads with fine-spun arguments for useful truths, so just as little do the equally subtle objections against these truths ever enter their minds’ (B xxxiv). 15 See Kelsey (2014) for an excellent account of the threat indifferentism poses in Kant’s theoretical philosophy. 16 ‘Why has providence set many objects, although they are intimately connected with our highest interest, so high that it is barely granted to us to encounter them in an indistinct perception, doubted even by ourselves, through which our searching glance is more enticed than satisfied?’ (A744/B772). 17 This is similar to what Rawls (1989) calls ‘philosophy as defense’ and what Ameriks (2000) calls ‘philosophy as apologetics’, although I mean to capture a broader role in speaking of philosophy as justification.
6 Philosophy as Justification of humankind, and I despised the rabble who knows nothing. Rousseau has set me right. This blinding prejudice vanishes, I learn to honor human beings, and I would feel by far less useful than the common laborer if I did not believe that this consideration could impart value to all others in order to establish the rights of humanity. (Refl 20:44)18
This fragment was composed around 1765, a decade and a half before the appearance of the first Critique, and its influence on Kant’s masterpiece is unmistakable. A recurring theme in the first Critique is that philosophy faces the project of protecting the rights of humanity by making the ground for the ‘majestic moral edifices’ of pure reason ‘level and firm enough to be built upon’ (B376). Kant even defines the true concept of philosophy as the ‘science of the relation of all cognition to the essential ends of human reason’, adding that the highest of these ends constitutes ‘the entire vocation of human beings’ (A840/B868).19 The practical orientation of Kant’s system therefore explains, on the one hand, why he advocates a close engagement with traditional metaphysics, condemning an attitude of indifference to such crucial questions, and yet, on the other, why he ridicules the fine-spun arguments of the schools, claiming that they never have nor ever will influence the hearts and minds of ordinary persons.20 In saying this, however, is Kant recommending that we settle all disputes in philosophy by appealing to common sense? In wanting to give philosophy a foothold in common human reason, one of Kant’s aims is to supply us with a touchstone ‘for passing judgment on the correctness’ of reason in its ‘speculative use’ (LJ 9:57). The touchstone is what he identifies with the entire moral vocation of human beings, the object of which permits an extension of reason beyond the sensible world, but only for the sake of rational belief. In order to have a guiding thread through the ‘immeasurable space of the supersensible’, Kant says the phil osopher must take hold of an interest of reason in its practical use, found already in the most ordinary understanding (WDO 8:137). In this way the philosopher must rid herself of what Kant calls a ‘prejudice against healthy human reason’ (V-Lo/Blomberg 24:193). But this does not obviate the need for insight and science altogether; nor does it render methods of argument and justification unnecessary. 18 For the Rousseau–Kant connection, see Velkley (2002), Ameriks (2012), and Callanan (2019). 19 It is in the sphere of practical reason, Kant argues, that the ‘unquenchable desire to find a firm footing beyond all bounds of experience’ must be directed, once we realize, of course, that such desire will never find satisfaction on speculative grounds alone (A796/B824). ‘Pure reason has a presentiment of objects of great interest to it. It takes the path of mere speculation in order to come closer to these; but they flee before it. Presumably it may hope for better luck on the only path that still remains to it, namely that of its practical use’ (A796/B824). See Ferrarin (2015) and Deligiorgi (2017) for further discussion. 20 Kant makes a similar claim in the 1781 edition: ‘In what concerns all human beings without exception nature is not to be blamed for any partiality in the distribution of its gifts, and in regard to the essential ends of human nature even the highest philosophy cannot advance further than the guidance that nature has also conferred on the most common understanding’ (A831/B859).
Introduction 7 In fact, Kant is careful to warn against what he calls a ‘prejudice for healthy human reason’ that would render such methods redundant—adding, with a bite of sarcasm, that ‘to appeal to ordinary common sense when insight and science run short, and not before, is one of the subtle discoveries of recent times, whereby the dullest windbag can confidently take on the most profound thinker and hold his own with him’ (Prol 4:259).
Methods of Justification in Ethics This last warning is crucial for understanding the topic of this book: namely, Kant’s justification of ethics in Groundwork III and the second Critique. Although Kant thinks moral inquiry must also get a foothold in common human reason, he does not think we can prove or defend the fundamental concepts of ethics by appealing to common sense alone. On the contrary, methods of argument and justification play a key role in Kant’s writings, four of which will occupy our attention in the coming chapters: • First, there is the skeptical method, or the method of doubting a claim (and ‘bringing it to the highest degree of uncertainty’) in order to motivate inquiry into its sources and origins. This procedure is effective for suspending judgment in matters of speculation with the aim of ‘getting on the trail of truth’ (LJ 9:84; see also A423/B451). As we shall see in Chapter 1, the skeptical method plays a central role in the Groundwork. After claiming that only a good will can be considered good ‘without limitation’, Kant considers the suspicion that this idea has its basis in ‘mere high-flown fantasy’ and that we have ‘misunderstood the purpose of nature in assigning reason as the ruler of the will’ (G 4:394–395). But Kant raises this suspicion, not because he harbors any doubt in the value of a good will, but because the suspicion serves to motivate inquiry into the sources and origins of the idea itself. • Second, there is the experimental method, or the method of illustrating a claim by way of a thought experiment. This procedure applies in contexts where the claim in question can be made vivid and intuitively compelling, even though it does not admit of a strict proof. As we shall see in Chapter 2, the experimental method plays a central role in the second Critique. After claiming that consciousness of the moral law reveals our freedom to us, Kant sets up a thought experiment to illustrate how common human reason separates morality from considerations of one’s own happiness (KpV 5:30). The experiment has the reader consider the case of a man facing a conflict between duty and death, with the aim of showing that this man would judge it possible to perform his duty even under threats of execution. What then
8 Proofs in Moral Philosophy becomes vivid is the way we judge that the moral law holds more authority than the sum-total of our sensible inclinations. • Third, there is the polemical method, or the method of defending a claim by countering ‘dogmatic denials’ of it (A739–40/B767–8). This procedure applies in contexts where the claim in question lies beyond the reach of human reason, since we can then show that one is not entitled to reject what one cannot know. As we shall see in Chapter 3, the polemical method plays a central role in Groundwork III. After tracing the moral ‘ought’ to our own intelligible ‘will’, Kant argues that moral motivation requires a ‘causality of reason to determine sensibility in conformity with its principles’ (G 4:460). Yet Kant’s point is that we cannot explain the connection between a nou menal cause and a phenomenal effect in feeling. Thus, ‘where determination by laws of nature ceases, there all explanation ceases as well, and nothing is left but defense, that is, to repel the objections of those who pretend to have seen deeper into the essence of things’ (G 4:459). • Lastly, there is the phenomenological method, or the method of reflecting on a claim as it appears in consciousness. This procedure applies in contexts where the claim in question displays unique features that come to light, not through conceptual analysis, but only through reflective attention. As we shall see in Chapter 4, the phenomenological method plays a central role in the third chapter of the second Critique. After repeating his earlier claim that we cannot comprehend how reason can determine sensibility, Kant adds that we still have room to consider what effects moral consciousness must have on our capacity to feel (KpV 5:72). Aside from eliciting a painful feeling of self-reproach when we see that we have treated happiness as a lawgiving principle, our consciousness of the moral law also elicits a pleasurable feeling of self-esteem, i.e., when we see that as rational beings we are ‘elevated’ above our sensible nature and hence capable of autonomy. As we shall see, what makes the phenomenological approach important as a method of justification is that it warrants our possession of a faculty of sensibility attuned to the demands of pure practical reason, without presuming insight into the causal connection between the two. Similar to the experimental method mentioned above, this procedure is effective for establishing moral concepts that do not admit of a strict proof.
Proofs in Moral Philosophy This brings us to an important question for the present study: What kinds of proof are available within Kant’s moral philosophy? While he never stipulates the exact meaning of a ‘proof ’ (Beweis) in his writings on ethics, Kant does speak to
Introduction 9 this issue in the first Critique and in his lectures on logic. An important distinction we find in these texts concerns what Kant variously calls indirect, negative, or apagogic proofs, on the one hand, and direct, positive, or ostensive proofs, on the other (A7891/B817). Apagogic proofs work to establish the truth of a proposition by demonstrating the falsity of claims opposed to it—hence, the strategy is indirect or negative, since the truth emerges by way of elimination. Ostensive proofs in contrast work to establish the truth of a proposition by revealing its grounds— hence, the strategy is direct or positive, since the truthemerges by way of insight into its source. The advantage of an ostensive proof, Kant explains, is that it discloses, not only the truth of a proposition, ‘but also at the same time its genesis, its generative source’ (V-Lo/Blomberg 24:233). As he puts it, ‘The direct or ostensive proof is, in all kinds of cognition, that which is combined with the conviction of truth and simultaneously with insight into its sources’ (A7891/B817).21 Applying this distinction to Kant’s moral philosophy is helpful. In the Groundwork, for example, Kant explains that it is one thing to clarify the structure of the moral law and another to justify its application to us as a binding imperative. The problem is that merely clarifying a concept does not give us insight into its genesis. The moral law expresses an unconditional demand, and we experience that demand in the form of an ‘ought’, but we cannot tell from where this demand purports to bind us. For all we know, Kant adds, the moral ‘ought’ might arise from the faculty of the imagination, not from the faculty of reason, in which case it would be an illusion. This question leads Kant in Groundwork III to offer a genetic proof or ‘deduction’ of the moral law’s bindingness. As we shall see, the genetic proof involves a critical examination of our faculty of practical reason— separating its empirical and pure uses—in order to show that our will is not just sensibly affected. Kant argues that the ground of our own sensibly affected will contains the idea of a pure will capable of determining itself on the basis of reason alone. The source of moral obligation therefore lies within us: what we ‘ought’ to do as sensibly affected beings is what we ‘would’ do as beings with a pure will. But what about the moral law itself? Does it not also fall within the purview of Kant’s project of justification? On the reading I shall defend in this book, the necessity of the moral law itself is never the object of a genetic proof or deduction
21 The impression we receive upon hearing these remarks is that the model of an ostensive or genetic proof is clearly superior, in Kant’s eyes, to the model of an apogogic or non-genetic proof. And to an extent that is no doubt the case. However, we should not be misled into thinking that apogogic or non-genetic proofs are to be avoided at all costs, as if their presence would threaten to undermine a program of justification. To be sure, although Kant refers to them as ‘more of an emergency aid than a procedure which satisfies all the aims of reason’, he adds that ‘they have an advantage in self-evidence over the direct proofs in this: that a contradiction always carries with it more clarity of representation than the best connection, and thereby more closely approaches the intuitiveness of a demonstration’ (A790/B818). Although I disagree with the details of his interpretation, Guyer’s (2017) proposal that Kant’s procedure in Sections I and II of the Groundwork largely accords with the apogogic model is helpful.
10 Proofs in Moral Philosophy in Kant’s moral philosophy, not even in the semi-critical form it takes in the first Critique.22 A key piece of evidence in support of this reading comes from Kant’s closing statements in Groundwork III, which have not, I am afraid to say, received the amount of attention that they deserve among commentators.23 What Kant makes clear in these statements is that human reason reaches a limit when it comes to the absolute necessity of the moral law, because human reason can alight upon no further condition under which to subsume (and thereby comprehend) the necessity in question (G 4:463). Were we to find some further condition under which to subsume the necessity of the moral law, then by that very token it would not be a species of absolute necessity after all. For Kant, we can gain insight into the moral law’s bindingness, or its ‘necessitation’ (Nötigung), insofar as we can trace this bindingness to our own intelligible will. But when it comes to the moral law itself, or its ‘necessity’ (Notwendigkeit), Kant argues that all we can do is comprehend its ‘incomprehensibility’ (G 4:463). When we take these final remarks in Groundwork III seriously, a novel way of interpreting Kant’s justification of ethics opens up. In the first place, we need not view Kant’s assertion in the second Critique that consciousness of the moral law is an underivable ‘fact’ as marking a ‘great reversal’ from his earlier argument in Groundwork III.24 On the contrary, I shall argue that with respect to the moral law’s necessity as a principle valid for all rational beings, Kant’s position shows no signs of wavering between Groundwork III and the second Critique: in each text he upholds that we cannot comprehend the necessity of the moral law and that it therefore admits of no deduction or genetic proof. What Kant makes salient in his later work is that an experimental method can illustrate our consciousness of this necessity as a fact of common human reason, ‘prior to all speculation about its possibility’ (KpV 5:90). But as we shall see, this is a variation of a theme already present in Groundwork III, since even there Kant argues that the ‘practical use of common reason confirms the correctness’ of his deduction, referring us to the case of a scoundrel who, when one sets before him examples of virtuous conduct
22 See §3.2 and §3.6 for my defense of this claim. 23 Rauscher (2009) and Puls (2016) are two important exceptions to this neglect. 24 This is Ameriks’s (1982, 226) turn of phrase. The list of scholars who subscribe to some version of the reversal reading is long, including Ross (1954), Beck (1960), Henrich (1960), Williams (1968), Korsgaard (1989), O’Neill (1989), Allison (1990), Łuków (1993), Neiman (1994), Hill (1998), Allison (2011), Rawls (2000), Engstrom (2002), Darwall (2006), Sussman (2008), Reath (2012), and Grenberg (2013). Other scholars find a significant change in Kant’s project of justification, not because he gave up a proof of the moral law, but because he gave up a non-moral argument for freedom. For this view, see Schönecker (1999, 2006, 2013, 2014), Guyer (2009), Timmermann (2010), Ludwig (2010, 2012, 2015, 2018), Hahmann (2012), and Bojanowski (2017). Noteworthy exceptions to this trend in the literature include McCarthy (1982), Wolff (2009), Wood (2011), and Tenenbaum (2012). However, with the exception of Puls’s German monograph (2016), there have been no systematic efforts to find continuity in Kant’s project of justification from 1785 to 1788.
Introduction 11 (even ones involving ‘great sacrifices’ of self-interest) wishes ‘that he might also be so disposed’ (G 4:454).25 Further support for my reading comes from what Kant says about his methodology in the Groundwork and the second Critique. In the prefaces of each work Kant tells the reader that his order of exposition will unfold along two paths, one ‘analytic’, and the other ‘synthetic’. The analytic path begins with what is given in common moral experience and works to clarify the highest principle that makes this experience possible. We follow this path, Kant explains, when we separate what is empirical in our faculty of practical reason (as it is conditioned by sens ibility) from what is pure (as it is unconditioned by sensibility) in order to ‘ascend’ to the supreme law of this faculty as a whole—the moral law as a principle of autonomy. Only after discovering this law can we take the synthetic path, which turns back and descends to our original starting point. We follow this path, Kant explains, when we recombine what was previously separated, the empirical and pure parts of practical reason, in order to reveal their necessary connection. In this respect, while both the analytic and synthetic paths constitute a single method, the synthetic path marks the path of justification proper, since it yields a special kind of insight, either insight into the possibility of moral obligation (in the Groundwork) or insight into the possibility of moral motivation (in the second Critique).26 We shall return to this distinction of analytic and synthetic paths more than once in the coming chapters.27 But I should say that my aim here is not to defend a continuity reading for its own sake. The value of this reading is that it clears room for us to explore other possible differences shaping Kant’s project in the second Critique. As I have just hinted at, Kant’s later project is much broader in scope than commentators have traditionally assumed.28 It includes, not just his 25 For a different reading, see Bittner (1989). Sticker (2014, 2015) is one of the few scholars to recognize that Kant’s example of the scoundrel from subsection 4 of Groundwork III signals a return ‘back’ to common cognition. However, like the majority of commentators, Sticker does not context ualize this example with reference to Kant’s methodology, as I plan to do. 26 In an important essay devoted to Kant’s methodology, Gabriele Gava (2015) makes a compelling case for assigning two distinct senses to Kant’s analytic–synthetic distinction in his theoretical phil osophy. A broader sense characterizes them as strategies of exposition and a narrower sense characterizes them as modes of cognition. In the latter case, analysis refers to conceptual clarification and synthesis refers to the special a priori insights afforded by Kant’s transcendental deductions. In the former case, an analytic strategy of exposition begins by separating a faculty of cognition into its basic elements, as it is given in experience, and a synthetic strategy recombines what was previously separated in order to show their necessary unity. Interestingly, these two senses come together in the Groundwork and the second Critique. Kant explicitly organizes each text in terms of analytic and synthetic paths, according to which we ascend from the basic parts of practical reason to its supreme law (the moral law), and then descend from this law back to the parts. Moreover, in each text Kant provides analytic knowledge (based on conceptual clarification) along the analytic path and synthetic cognition (based on necessary a priori connections) along the synthetic path. 27 See §1.1, §1.13, §3.1, §3.3, §3.12, §4.1, §4.11, and §4.12. 28 Beck (1960) and Allison (1990) are standard representatives of this view. For an important exception, see Franks (2005).
12 To Whom? From Where? Against What? doctrine of the fact of reason in the first chapter, but also his theory of moral sensibility in the third chapter. On the reading I shall defend, Kant’s synthetic path in the second Critique extends to the question of how the moral law can be a real motivating ‘incentive’ (Triebfeder), i.e., how it can be a subjective ground of action and not merely an objective (and potentially ineffective) principle. Given our inability to comprehend a causality of reason, Kant appeals instead to a phenomenological method in the third chapter and examines the effects that consciousness of the moral law must have on our capacity to feel.29 The result is confirmation of the moral law’s applicability to beings who, like us, are not only rational but also sensibly affected, since the theory of moral sensibility shows that our ‘hearts’ and not just our ‘heads’ are responsive to the demands of duty. This is consistent with Kant’s project of locating the source of morality within us, the aim of Groundwork III, but it goes further, on my account, by describing a positive interaction between reason and feeling, ultimately showing how our consciousness of the moral law influences our faculty of sensibility through moral feelings of self-reproach and self-esteem.
To Whom? From Where? Against What? If one now wanted a label to capture the spirit of Kant’s metaethics in the Groundwork and the second Critique, as I have presented these texts so far, then I think anti-error theory would be an apt turn of phrase. The error theorist, at least in her traditional guise, tells us that ordinary moral judgments are systematically false. This implies by extension that common human reason is completely mistaken in its conception of ethics. Kant’s metaethics is an anti-error theory to the extent that it gives primacy to the ordinary standpoint of life over the speculative standpoint of philosophy. Yet there is a further point to note in light of Kant’s conversion around 1765, namely, that he seeks to reorient philosophy itself (and not just moral philosophy) to the ends and interests of human beings, making his anti-error theory as much a meta-philosophical view as it is a meta-ethical one. I mention this to help dispel a sense of perplexity readers are likely to have with respect to Kant’s apparent disregard for any form of radical moral skepticism. On this issue I find it instructive to consider Kant’s rejoinder to a critic of the Groundwork ‘who wanted to say something censuring this work’, but in fact, he adds, ‘hit the mark better than he himself may have intended when he said: that no new principle of morality is set forth in it but only a new formula’ (KpV 5:8n). 29 I am sympathetic to Heidegger’s impression of the third chapter—that it is ‘the most brilliant phenomenological analysis of the phenomenon of morality that we have from him’ (1927/1988, 133). But I think there are limits to the phenomenological interpretation of Kant defended by Heidegger (1927/1988) and, more recently, by Schönecker (2013) and Grenberg (2013), which I shall discuss at greater length in §4.10 and §4.11.
Introduction 13 In reply Kant asks rhetorically: ‘But who would even want to introduce a new principle of all morality and, as it were, first invent it? Just as if, before him, the world had been ignorant of what duty is or in complete error about it’ (KpV 5:8n; emphasis added). A response some readers might have is that this hypothetical scenario—that the world has been in complete error about duty—is precisely the kind of skeptical threat the moral philosopher is responsible for addressing. But it is important to see from the outset of this book that Kant views this response as mistaken in its idea of what moral philosophy can accomplish, for the same reasons Rousseau gives, speaking through the character of a Savoyard Vicar, when he asks ‘Were not all books written by men? Why, then, would man need them to know his duties, and what means had he of knowing them before these books were written?’30 This is not to say that Rousseau and Kant are apologists for common sense, since they agree that our everyday judgments require development, education, and cultivation, of which the right kind of philosophy (in Kant’s case, a ‘metaphysics of morals’) plays an active role.31 But neither of these writers take it as a condition of success for their theorizing to somehow refute a skeptic who requests a reason to care about duty or ethical life in general, or who regards common reason as totally mistaken in all moral matters. All of this leads me to conclude that a distinctive feature of Kant’s anti-error theory is a self-directed point of focus, insofar as it aims to vindicate the claims of ethics within a pre-theoretical standpoint. This gives us a preliminary answer to a set of questions Bernard Williams once raised in the context of asking ‘what a justification of ethical life should try to do’.32 In Williams’s view, we should ask any attempted justification the following three questions: ‘To whom is it addressed? From where? Against what?’33 On my reading, while Kant invokes the skeptic’s doubts to provoke inquiry into the sources of morality, he is not addressing the moral skeptic per se. He is not taking up the ‘adversarial stance’ and attempting to deploy an argument the skeptic about morality must accept on pain of contradiction.34 At the same time, we should not lose sight of the fact that Kant was writing for students and specialists of philosophy—not just to professional scholars, but also to individuals who have turned their reflections to moral issues and who want, if possible, rationally satisfying answers. Neither the Groundwork nor the second Critique were intended to serve as popular treatises; here as elsewhere Kant is explicit about giving scholastic rigor priority over widespread appeal when it comes to laying the foundations of a science (G 4:391–392). Because Kant thinks moral philosophy is corruptible, and in some cases corrupting, it is necessary for his project to reveal the fatal flaws of rival ethical theories, especially those based on empirical principles. So to answer Williams’s second question, ‘From where?’, it is clear that Kant develops 30 Rousseau (1762/1979, 303). 31 On this point I agree with Sticker (2015). 32 Williams (1985, 23). 33 Williams (1985, 23). 34 Wright (1991).
14 To Whom? From Where? Against What? his justification of ethics from the standpoint of philosophy, but one suitably equipped to examine the sources of common human reason. This brings us to Williams’s third question, ‘Against what?’ That is, what is Kant’s project of justification working to avoid? If my answers to the first two questions are correct, then it is safe to say that Kant is not working to avoid the threat of immoralism. Kant is optimistic that even a scoundrel recognizes the dignity of moral action, however much he fails (perhaps due to frailty) to conform his will to the requirements of duty (G 4:460). But then what is the problem to which Kant is seeking a solution in his writings on ethics? The answer changes depending on what text we have before us, the Groundwork or the second Critique, but what they share in common, on my reading, is an effort to rescue common human reason from the conflict it experiences between the claims of morality, on the one hand, and the claims of happiness, on the other. This is not a mere speculative problem of ‘doubt’ (Zweifel), such as the kind pure reason generates when it oversteps the field of experience, but a practical conflict at the heart of common moral consciousness—in a word, a problem of ‘despair’ (Verzweiflung).35 Yet the despair in question is not so much a crisis of competing forces as it is a crisis of competing self-conceptions, since what the tension between morality and happiness threatens, at bottom, is a disharmony in our higher vocation. It is a threat, as Kant puts it, of becoming obscure to ourselves (G 4:405). A fourth question we might pose, adding to Williams’s list, concerns what a project of moral justification aims to accomplish. To what end is it directed? If it is correct to say that Kant is working against a threat of self-obscurity, then we can understand his aim in terms of restoring harmony to the idea of our higher vocation. This is the mind-set in which Kant is working to overcome, not doubt, but despair. By making moral experience intelligible to ourselves—either by defending the belief that we are free, or by revealing the source of obligation within us, or by describing our capacity to feel respect for the law—Kant’s goal is to vindicate a lofty yet fragile idea of humanity: namely, the idea of humanity as having a citizenship in a world beyond the sensible one and a destiny beyond the pursuit of happiness. In other words, by making moral experience intelligible to ourselves, the aim of Kant’s justification of ethics is to restore trust in the idea of ourselves as the kind of beings for whom morality applies—that is, finite beings with moral reason and moral sensibility. This answers what we might call a general question about the legitimacy of our moral vocation, which will make up the largest portion of this book (Chapters 1–4). As we shall see, however, it does not address a specific question of how we can take steps toward our moral vocation, since this raises an issue of how we can 35 See Breazeale (2012) and Franks (2008) for two illuminating accounts of the link between ‘doubt’ (Zweifel) and ‘despair’ (Verzweiflung) in post-Kantian skepticism. See also Grenberg (2013) for a detailed treatment of the problem of a practical conflict between morality and happiness.
Introduction 15 know whether our moral progress is genuine. For this reason I have devoted the final chapter of this book (Chapter 5) to the problem of moral self-knowledge, keeping an eye to the obstacles Kant thinks stand in the way of our moral improvement. On my reading, these obstacles constitute two sides of what I call Kant’s opacity thesis: his claim that we can never get to the bottom of our own intentions for acting. One side of the opacity thesis concerns our persistent tendency to deceive ourselves, to construe our intentions in a flattering or praise worthy light. The other side concerns the limit we encounter in trying to understand ourselves without ever gaining access to our underlying characters. On my account, the presence of self-opacity threatens to undermine the intelligibility of moral progress by making it uncertain whether our commitment to our higher vocation is sincere or merely feigned. This is not a question of whether we are warranted in ascribing such a vocation to ourselves. Rather, it is a question of whether we can trust our own moral aspirations, the solution of which, I shall argue, leads Kant to develop a theory of conscience in his later works, Religion within the Boundaries of Mere Reason and the Metaphysics of Morals. Admittedly, the broader question I plan to investigate here—whether there is reversal or continuity in Kant’s justification of ethics—may appear to be of interest only to historians of philosophy. Yet in truth it speaks to a larger issue in contemporary ethics over the normativity of moral requirements, specifically the issue of whether those requirements are derivable from a more foundational conception of action or even from a more basic conception of theoretical rationality (a strategy employed, with varying aims, by Thomas Nagel, Alan Gewirth, Peter Railton, David Velleman, Connie Rosati, and Christine Korsgaard, among others).36 Proponents of what we might call foundationalism in ethics are ready to acknowledge that Kant’s second Critique has a different starting point: our common moral consciousness.37 Nevertheless, many believe that Kant had ventured an argument from a more basic conception of rational agency in his earlier work. It would then be significant if my version of the continuity reading were true, beyond throwing new light on Kant’s intellectual development. Not only may foundationalism lose its historical affiliation with Groundwork III, but Kant’s reasons for resisting a deduction of the moral law from non-moral premises may also be good reasons for resisting foundationalism today. I will return to this set of reflections in the Conclusion where I shall say why Kant’s justification of ethics, although not free of problems, is still of lasting importance. 36 See Nagel (1970), Gewirth (1978), Railton (1997), Velleman (1989, 2000, 2009), Rosati (2003), and Korsgaard (1996a, 2008, 2009). 37 As I am using this label, foundationalism in ethics is a broad category for any strategy of deriving the normativity of moral requirements from a more basic conception of action, agency, freedom, or rationality (including theoretical rationality). There are similarities here to what Ameriks labels ‘strong foundationalism’ in the philosophical programs of the early post-Kantians, who similarly wanted to derive robust knowledge claims (e.g., about the external world) from a more basic conception of representation, consciousness, or self-consciousness. See Ameriks (2000).
1
Moral Skepticism The objections that are to be feared lie in ourselves. We must search them out like old but unexpired claims. – Kant (A777/B805)
1.1 Introduction The aim of this chapter is to set the stage for a closer examination of Kant’s arguments for the reality of human freedom and the normativity of the moral law in Groundwork III. Its central task is to clarify the organization of the Groundwork as a whole, which I will show unfolds according to a skeptical method. My proposal is that Kant’s strategy of making a claim doubtful in order to arrive at its source underlies the various ‘transitions’ of the work, namely, from common to philosophical moral cognition in Section I, from popular moral philosophy to a metaphysics of morals in Section II, and lastly, from a metaphysics of morals to a critique of pure practical reason in Section III. By working through these transitions, my first task is to provide the reader with a sketch of the main arguments constituting the Groundwork, which will prepare the way for a careful assessment of Kant’s project of justification in the third section. One important result of interpreting the Groundwork through this framework is that it highlights a significant but easy to overlook feature of Kant’s methodology. As he tells us in the Preface, the first two sections of the book follow an ‘analytic’ path and the third follows a ‘synthetic’ path (G 4:392). While commentators have debated over the exact import of this distinction, I will propose that the analytic path is equivalent to a procedure of ‘ascending’ to the highest principles of practical reason and the synthetic path is equivalent to the procedure of ‘descending’ from these principles back to the standpoint of common human reason in which we find them employed. In all this it remains puzzling that Kant does not invoke or address the skeptic’s traditional question in Groundwork III, ‘Why be moral?’, as one might have expected. Quite a few commentators have taken Kant’s silence on this issue as evidence that he was uninterested in the sort of concerns that preoccupy ethical and meta-ethical theorizing today. John Rawls, for example, was convinced that Kant did not wish to confront a radical moral skeptic. That was not a problem for
Kant’s Justification of Ethics. Owen Ware, Oxford University Press (2021). © Owen Ware. DOI: 10.1093/oso/9780198849933.003.0002
Moral Skepticism 17 him, ‘however much it may trouble us’.1 Similarly, Thomas Hill has suggested that Kant’s aim in Groundwork III is ‘easily obscured’ by the fact that his audience was not the sort of skeptic or amoralist we are familiar with. He did not take himself to be responding to those ‘who are indifferent to morality and demand that phil osophy supply them with a motive to be moral’.2 Related comments appear in the work of Allen Wood, Jens Timmermann, and Henry Allison.3 There are, however, a few exceptions to this trend. Michael Forster and Paul Guyer have both argued that Kant is actively engaged with a variety of skeptics from the tradition—of Humean, Cartesian, and Pyhronnian origin.4 In Guyer’s view, Rawls and Hill are right to suggest that the project of the Groundwork is not to justify the universal and binding demands of morality to someone who alleges no presumption in their favor, but it would be wrong to conclude from this that Kant’s argument is not intended as an answer to moral skepticism as Kant understood it.5
According to Guyer, Kant is responding to a number of skeptical opponents in his writings, including the ancient Epicureans, as well as modern authors such as Wolff, Baumgarten, Hobbes, Hume, and others.6 In what follows I wish to clear room for a new position within this debate. While I agree that Kant’s goal in the Groundwork is anti-skeptical, contrary to Rawls and Hill, I do not want to characterize it in adversarial terms. In my view, Kant is not advancing an argument the skeptic about morality must accept on pain of contradiction; he is not trying to defeat ‘a real philosophical opponent, the Sceptic, in rational debate’, to borrow Crispin Wright’s description.7 My own view is that characterizing Kant’s ethics in adversarial terms obscures an important feature of his approach: namely, that the skeptic most worth addressing lies within ourselves. I will therefore argue, contrary to Forster and Guyer, that Kant is not speaking to the philosophical tradition in any direct or straightforward manner.8 1 Rawls (2000, 149). 2 Hill (1998, 250). 3 Wood: ‘Kant’s deduction of the supreme principle of morality admittedly does not address (or even appear to take seriously) some of the more extreme forms of skepticism about value which have dominated twentieth-century meta-ethics’ (1999, 381, note 30). Timmermann: ‘Kant does not take the traditional amoralist’s question of why we should be moral at all seriously . . . Moreover, it is doubtful whether anyone would be in a position to persuade a radical amoralist’ (2007, 130, note 21). Allison: ‘[Kant’s] interlocutor is not the familiar skeptical amoralist in search of reasons for obeying the dictates of morality when they clash with self-interest’ (2011, 309–10). 4 Forster (2008) and Guyer (2008). 5 Guyer (2008, 26). 6 Guyer (2008, 7). 7 As Wright makes clear: ‘There are no such real opponents. That generations of philosophers have felt impelled to grapple with skeptical arguments is not attributable to a courtesy due to an historically distinguished sponsorship but to the fact that these arguments are paradoxes: seemingly valid deriv ations from seemingly well supported premisses of utterly unacceptable consequences’ (1991, 89). 8 This makes my position closer to Guyer’s earlier view (2000). There are also resonances between my account and Conant’s (2004), although exploring those resonances is beyond the scope of this book.
18 Background: Kant ’ s Methodology On the reading I will defend, Kant himself employs a skeptical method in the Groundwork as a way of exposing certain obstacles in our ordinary and philosophical thinking about morality. The central obstacle he wants to uncover is practical in character, arising from a natural tendency we have to rationalize against the moral law. In attempting to resolve this tendency, the goal of the Groundwork is to vindicate our common idea of duty and the higher vocation it entails.
1.2 Background: Kant’s Methodology When we open the Groundwork to its first page, we find Kant introducing the reader to a map of human knowledge. At the upper end of the map we find a distinction between two kinds of rational cognition: ‘formal’ and ‘material’ (G 4:387). Rational cognition is formal (as with logic) when it abstracts from any particular object and merely considers rules necessary for thinking. Rational cognition is material when it concerns laws necessary for objects themselves—either ‘laws of nature’ (as with physics) or ‘laws of freedom’ (as with ethics). According to Kant, each branch of material rational cognition admits of a further sub div ision. Physics has an ‘empirical’ part and a ‘pure’ part. The empirical part concerns the application of the most fundamental laws of nature to the sensible world; the pure part concerns those laws themselves, abstracting from their application to the sensible world. Similarly, the empirical part of ethics concerns the application of the most fundamental laws of freedom to human beings; the pure part concerns those laws themselves, abstracting from their application to human beings. To one side of the map, then, we have an empirical physics of nature and what Kant calls ‘practical anthropology’; to the other side, we have a metaphysics of nature and what Kant calls a ‘metaphysics of morals’ (G 4:388). Carefully separating these branches of study is important, Kant explains, because they point to a clear division of intellectual labor. ‘All professions, crafts, and arts have gained’, he writes, ‘by the distribution of labor, namely when one person does not do everything, but each limits himself to a certain task that differs noticeably from others in the way it is carried out, so as to be able to accomplish it most perfectly and with greater ease’ (G 4:388). Kant argues that if we try to do everything at once, we will not be able to make progress in any given field: ‘Where labor is not differentiated and distributed like that, professions still remain in a most barbarous state’ (G 4:388). One difficulty is that mixing the empirical and the pure together is very much in line with public taste, especially in the field of ethics. So Kant thinks that if we are to make real progress in the study of moral philosophy we must abandon, if only for the time being, any desire for widespread appeal. We must attain what he calls ‘determinate insight’ first,
Moral Skepticism 19 and only afterwards speak to the interests of the public (G 4:410). To this end Kant thinks we must divide our labor properly and begin with the pure part of ethics before proceeding to its empirical part. We must begin with a metaphysics of morals. The first question Kant takes up in the Preface is whether a metaphysics of morals is even possible. We have no basis to raise this concern in the case of logic or physics; for these are well-established branches on the map of human know ledge. But the same is not true for the pure part of ethics, the theory of which does not yet exist. Kant’s reply is that we catch a glimpse of the possibility of a metaphysics of morals when we turn to the ‘common idea of duty’ (gemeinen Idee der Pflicht) (G 4:389). For this idea highlights an important feature of our everyday moral discourse: the fact that when we speak of moral obligations, we assume that such obligations carry ‘absolute necessity’ (G 4:389): Everyone must admit that a law, if it is to hold morally, i.e., as the ground of an obligation, must carry with it absolute necessity; that the command: Thou shalt not lie, does not just hold for human beings only, as if other rational beings did not have to heed it; and so with all remaining actual moral laws. Hence the ground of the obligation here must not be sought in the nature of the human being, or in the circumstances of the world in which he is placed, but a priori solely in concepts of pure reason. (G 4:389)
Here Kant is saying that everyday moral experience (i.e., our consciousness that some actions are absolutely required of us) indicates that a metaphysics of morals is possible, for these demands carry a necessity that could not have come from the empirical constitution of our will. That is sufficient, in Kant’s view, to show that a study of the pure part of ethics is not in vain. But now we must ask a related question: Why is a metaphysics of morals necessary? Right away Kant is clear that venturing into this new area of study is not simply a matter of satisfying our intellectual curiosity. A more serious issue is at stake. This is because in everyday life we require a power of judgment ‘sharpened by experience’, not only to apply moral laws in concrete cases, but also to acquire ‘momentum’ for their performance (G 4:389). While it is relatively easy for us to understand the nature of our moral obligations, it is difficult to make those obligations ‘effective’ in our lives (G 4:389). In saying this, Kant is calling attention to the fact that we are not purely rational beings: we are prone to self-deception, especially when the demands of morality conflict with our self-love. A metaphysics of morals is thus ‘indispensably necessary’, he writes, ‘not merely on the grounds of speculation . . . but because morals themselves remain subject to all sorts of corruption as long as we lack that guideline and supreme norm by which to judge them correctly’ (G 4:389–90).
20 Background: Kant ’ s Methodology What is striking, then, is that Kant is saying we need a metaphysics of morals to protect common human reason from itself—in fact, this is the practical aim of his entire work, as we shall see. Kant concludes the Preface by saying that his aim in the Groundwork is twofold: to clarify and justify ‘the supreme principle of morality’ (G 4:392). As he writes, this order is best if one wants to proceed along the analytic path from common cognition to the determination of its highest principle and then turn back [wiederum zurück] along the synthetic path from an examination of this principle and its sources to the common cognition in which we find it used. (G 4:392)
By ascending along the analytic path first, Kant is saying, we will be able to grasp the principle adequate to explain our common idea of duty, the principle of autonomy. By then descending along the synthetic path, we will return to the common idea of duty with which we began, but now with a new insight we did not have before: namely, insight into the source of morality within us. This is the sense in which the Groundwork promises to come full circle back to common cognition. As we will see, after completing the argument of Groundwork III Kant says that the ‘practical use of common human reason confirms the correctness of this deduction’, citing as an example of a ‘malicious scoundrel’ who identifies his better self with the demands of morality (G 4:454).9 In a text composed just prior to the Groundwork Kant is clear that the analytic method ‘signifies only that one proceeds from that which is sought as if it were given, and ascends [aufsteigt] to the conditions under which alone it is possible’, to which he adds, ‘it might better be called the regressive method to distinguish it from the synthetic or progressive method’ (Prol 4:276n). In the practical sphere, what is given is our common cognition of duty, and the first question we must address is what principle articulates this form of cognition. To do this, the analytic procedure has us separate what is pure in our cognition from what is empir ical; that is what allows us to climb up, as it were, to the highest principle governing the cognition in question. Afterwards, when we have grasped this highest principle, we can descend along the synthetic path and recombine what had previously been separated, thereby revealing the necessary connection between what is pure in our cognition and what is empirical. I take it this is why Kant identifies the synthetic path with the path of justification in Groundwork III. For it is only after separating our faculty of practical reason into its pure and
9 Although he does not make reference to Kant’s statement of method in the Preface, Engstrom (2009) offers an evocative description of this stage in terms of a moment of return: ‘the turning back in thought to the original position of practical knowledge in an acknowledgment of what is already implicitly present in all maxims and practical judgments’ (2009, 245).
Moral Skepticism 21 empirical uses that he is able to reveal their necessary connection, thereby locating the ground of the moral ‘ought’ in our own intelligible ‘will’. Of course, Kant did not invent this distinction between an analytic ascent and a synthetic return, and his characterization of these two paths as movements of a single argument has much older roots. Aristotle is reported to have said that Plato ‘was right in raising this question and asking, as he used to do, Are we on the way from or to the first principles? There is a difference as there is in a race-course between the course from the judges to the turning-point and the way back.’10 The idea of a second stage of argument that reverses the order of the first appears vividly in the Republic when Socrates describes the intellect’s discovery of the first principle as the result of an upward ascent, after which ‘it reverses itself and, keeping hold of what follows from it, descends to a conclusion without making use of anything visible at all’.11 In the Platonic and Aristotelian tradition, the way upwards marks the path of discovery, which leads to a first principle (in the present context, the supreme principle of morality). But it is only the ‘way back’—the descent from this first principle—that marks the path of justification (in the present context, the principle’s bindingness upon a human will). Still, from what we have seen in the Preface one might think that returning to a common standpoint requires that we find access to public taste first and foremost—a view many late eighteenth-century German authors upheld, including Herder, Garve, and Mendelssohn. It is then unclear why Kant lashes out against ‘those who, conforming to the taste of their public, are in the habit of peddling the empirical mixed with the rational in all sorts of proportions unknown to themselves’, authors he unflatteringly describes as mere ‘bunglers’ (G 4:388). Despite the harsh tone of this remark, it is important to see that Kant shares the fundamental goal of ‘popular philosophy’ (Popularphilosophie)—that of giving the standpoint of life primacy over the standpoint of philosophy—but he rejects the method that authors like Herder, Garve, and Mendelssohn adopt to attain this goal, the method of mixing the pure part of ethics with the empirical part. Once again Kant’s point is that the possibility of making ethics a science depends on separating the two, for only by doing this can we gain insight into the principle behind our common idea of duty. Kant is clear that making ethics open to the public is ‘very commendable’, but only ‘if the elevation [die Erhebung] to the prin ciples of pure reason has already happened’ (G 4:409). He is not rejecting the aim of popular philosophy, then, as much he is rethinking the path to attain it. On the whole Kant says little else about his plan of organization in the Groundwork beyond the procedural distinction between Groundwork I and II
10 Aristotle (1984, 1.4; emphasis added). 11 Plato (1997, 511b). See Menn (2002) for an in-depth treatment of the Platonic and Aristotelian tradition.
22 Background: Kant ’ s Methodology (the ‘analytic path’) and Groundwork III (the ‘synthetic path’). But he leaves the reader with a further clue in the form an outline at the end of the Preface: 1. First section: Transition from common to philosophical moral rational cognition. 2. Second section: Transition from popular moral philosophy to the metaphysics of morals. 3. Third section: Final step from the metaphysics of morals to the critique of pure practical reason. When we inspect this outline more carefully, an important feature of Kant’s methodology comes into focus. We can see that Kant raises a problem within each section of the book in order to make a ‘transition’ or ‘shift’ of frameworks (an Übergang) necessary.12 In Groundwork I, he raises a problem in the framework of common human reason (with its pre-reflective understanding of morality) that only a transition to philosophical cognition can overcome. In Groundwork II, he raises a problem in the framework of popular philosophy (with its empirical, a posteriori method) that only a transition to a metaphysics of morals can overcome. Finally, in Groundwork III, Kant shows that a metaphysics of morals is also limited (with its conceptual, a priori method) in a way that demands a critical examination (or a ‘critique’) of our faculty of reason. What becomes clear, then, is that each stage of Kant’s investigation reaches a limit, one internal to the stage itself, and that discovering this limit motivates a shift of frameworks, a shift to a new, higher stage of insight. As we will see, one advantage of this reading is that it explains what motivates each transition in the Groundwork: by invoking possible doubts at the level of our ordinary and philosophical thinking about morality, Kant is able to advance his argument to its final, critical step. While it may seem speculative to attribute this skeptical method to Kant, there is textual evidence to support it, which I touched upon in the Introduction. In the Doctrine of Method from the first Critique Kant says that skeptical claims pose a threat to our philosophical commitments even if they have not been voiced by a real person. Their ‘seed’, as he puts it, lies in human reason itself (A778/B806). 12 There is little discussion of the Groundwork’s organization in the literature. One exception is Guyer (2007) who offers a concise overview of the transition-structure in his commentary on the Groundwork. As Guyer points out, one might be tempted to read ‘transition’ (Übergang) in the sense of ‘lead to’, which suggests a continuous unfolding from one stage to another (2007, 34–5). This gives the impression that the stage of common reason in Groundwork I will gradually lead to a metaphysics of morals in Groundwork II, and so on. On the contrary, to make a transition for Kant means something more like ‘to overcome’, indicating the presence of a limit or obstacle. Other commentators who reflect on the meaning of a transition include Schönecker (1997), who focuses on the transition in Groundwork I, and O’Neill (1989), who focuses on the transition in Groundwork III. However, neither explain what unifies the transition-structure of the Groundwork as a whole.
Moral Skepticism 23 Because of this, Kant says it is our responsibility to seek out those claims ourselves, to discover them by using what he calls a skeptical method. ‘Thus, think up for yourself the objections which have not yet occurred to any opponent’, Kant writes, ‘and even lend him the weapons or concede him the most favorable position that he could desire. There is nothing in this to fear, though much to hope, namely that you will come into a possession that can never be attacked in the future’ (A778/B806). Similarly, Kant is reported to have said during lecture that as harmful as skepticism is, ‘the skeptical method is just as useful and purposeful, provided one understands nothing more by this than the way of treating something as uncertain and of bringing it to the highest uncertainty, in the hope of getting on the trail of truth in this way’ (JL 9:84).13
1.3 Skepticism in Groundwork I Kant begins Groundwork I with the well-known claim that a good will is the only thing we can call good ‘without limitation’ (G 4:393). Riches and power are only good when they are possessed by a person with a good will. Happiness—which may seem to be good for everyone who receives it—is only objectively good on the condition that one is worthy of it. Even characteristics like intelligence and a calm disposition lack intrinsic worth, for it is easy to find examples of intelligent people bent on evil designs or courageous people leading criminal lives (G 4:393). Without going into the details of this claim, I merely want to point out that Kant takes up the standpoint of common reason in the fourth paragraph, and he does so in response to a skeptical worry: There is, however, something so strange in this idea of the absolute worth of a mere will, in the estimation of which no allowance is made for any usefulness, that, despite all the agreement even of common understanding with this idea, a suspicion must yet arise that its covert basis is perhaps mere high-flown fantasy and that we may have misunderstood the purpose of nature in assigning reason to our will as its governor. (G 4:394–5)
What is strange is not so much the idea that we have a vocation to cultivate a will that is good, but that pursuing it might require us to suspend or sacrifice our happiness. This is what Kant anticipates will elicit our suspicion. How could it be
13 As Kant puts it in the first Critique: ‘External quiet is only illusory. The seed of the attacks [from skepticism], which lies in the nature of human reason, must be extirpated; but how can we extirpate it if we do not give it freedom, indeed even nourishment, to send out shoots, so that we can discover it and afterwards eradicate it at its root?’ (A778/B806).
24 Skepticism in Groundwork I rational to give up everything we hold dear for the sake of a higher vocation? There is something puzzling about this, so much so that Kant says we are likely to suspect the idea of a good will rests on an illusion. The structure of Groundwork I takes shape as a response. ‘We shall’, he asserts, ‘put this idea to the test’ (G 4:395). This test draws on resources Kant assumes are available to common reason. One is an idea he thinks everyone has about organized life: that each component of a living being is best adapted to its final end (G 4:395). This is a claim about nature’s purposiveness. Kant uses it to show, first, that a rational creature with a hedonic vocation could not be responsible for planning its own happiness. For such a hypothetical creature, nature would have blundered in giving it a rational faculty for seeking the ends of inclination, even, Kant adds, for seeking their best means of fulfilment. Yet a blunder in nature contradicts its purposiveness. On this scenario, then, nature would have been sure to appoint instinct the role of governing the creature’s will, and it would have been careful to assign reason a passive function—that is, to self-consciously admire the efficiency of instinct, not to ‘break forth’ into the sphere of action (G 4:395). This raises the question: Why is reason so ill-suited for the pursuit of happiness? Kant asks us to look at people who have tried to devote themselves to a hedonic vocation. In his view, the ‘more a cultivated reason purposely occupies itself with the enjoyment of life and with happiness, so much the further does a human being get away from true satisfaction’ (G 4:395). This shows that putting our rational faculties into the service of our sensible nature is counterproductive: it can likely have the effect of multiplying the number of our inclinations, driving us further away from our goal. Kant believes this is true even for people who have devoted themselves to so-called higher pleasures: they often end up bringing ‘more trouble upon themselves’ instead of gaining in happiness. Soon they ‘envy rather than despise the more common run of people, who are closer to the guidance of mere natural instinct and do not allow their reason much influence on their behavior’ (G 4:396). After making these observations, Kant shifts subjects: ‘Reason is nevertheless given to us as a practical faculty, that is, as one that is to influence the will’ (G 4:396). The shift is easy to miss, however. At G 4:396 Kant is no longer speaking about a hypothetical creature; he is now speaking about an actual human being. For us, reason does break forth into the sphere of action, and that is enough to show (from the standpoint of common reason) that our vocation cannot merely be to pursue happiness. It would be inconsistent with the wisdom of nature to assign an active rational capacity to a being destined for happiness. So according to nature’s purposiveness, our vocation must be different. As Kant says, ‘where nature has everywhere else gone to work purposively in distributing its capacities, the true vocation [Bestimmung] of reason must be to produce a will that is good, not perhaps as a means to other purposes, but good in itself, for which reason was absolutely necessary’ (G 4:396).
Moral Skepticism 25
1.4 The Need for Philosophical Cognition Aside from any question of how compelling this ‘test’ is, we might wonder what else remains to be shown in Groundwork I. It seems that common reason has all the resources it needs to justify our higher moral vocation. The claim of nature’s purposiveness, combined with the supposed fact that reason is practical for us, shows that it is consistent to assign unlimited value to a mere will. Why does Kant not think this is enough? Recall that the skeptical worry from G 4:394 is that our vocation may rest on an illusion, that its underlying idea—a will that is good without limitation—may be a product of the imagination. While the test shows the consistency of these claims from the standpoint of common reason, it does not clarify the principle of a good will itself. And this highlights an obstacle in our ordinary moral thinking: we are unable to render the principle of our higher vocation explicit. A ‘transition’ to philosophical cognition is needed for this. Kant’s claim is more modest than it sounds, however. The transition in Groundwork I is only meant to clarify what we already know. He says that we have ‘to explicate the concept of a will that is to be esteemed in itself and that is good apart from any further purpose, as it already dwells in natural sound understanding’ (G 4:397). We need philosophical cognition to do this because it contains the right conceptual tools. Thus Kant’s strategy is to analyze how someone with a good will must act when faced with desires coming from his or her sensible nature. The result is that we set before ourselves the concept of ‘duty’ (Pflicht), ‘which contains that of a good will though under certain subjective limitations and hindrances, which, however, far from concealing it and making it unrecognizable, rather bring it out by contrast and make it shine forth all the more brightly’ (G 4:397). By proceeding in this way, Kant is able to show that the principle of a good will can be expressed as a formula of universal law, and that its motive for action can be expressed as a feeling of respect for the law as such. Aware of how abstract this sounds, Kant reminds us that we do not necessarily have a conscious grasp of the moral law in our everyday lives. It is, rather, the implicit normative standard of our judgments, a standard Kant thinks we employ whenever we appraise an action’s inner worth. This is why philosophical cognition does not teach us anything we did not already know. Without ‘in the least teaching it anything new’, Kant writes, we have only ‘like Socrates’ made common reason ‘attentive [aufmerksam] to its own principle’ (G 4:404).14 ‘With this compass in hand’, he 14 Kant alludes to Socratic maieutics in the first Critique in the context of discussing Plato’s theory of archetypes. For Plato, he writes, archetypes ‘flowed from the highest reason, through which human reason partakes in them; our reason, however, now no longer finds itself in its original state, but must call back with toil the old, now very obscure ideas through a recollection [Erinnerung] (which is called philosophy)’ (A313/B370).
26 A Natural Dialectic goes on to say, common reason ‘knows very well how to distinguish in every case that comes up what is good and what is evil, what is in conformity with duty or contrary to duty’ (G 4:404). The transition to philosophical cognition in Groundwork I is therefore a transition within common reason.15
1.5 A Natural Dialectic But there is a problem here. In our everyday frame of mind we are prone to selfdeception. We can easily delude ourselves into thinking that our happiness should have priority over moral claims. As a result, common reason falls into a ‘natural dialectic’: it starts to ‘rationalize against those strict laws of duty, and to cast doubt upon their validity, or at least their purity and strictness’ (G 4:405).16 This is a different problem than before. By subverting the moral law from a position of authority, we are susceptible to a genuine error of deliberation—that of reversing the priority of moral claims, so that considerations of our happiness retain super iority from a practical point of view. We are susceptible, that is, to an excess of self-love (what Kant will call ‘self-conceit’ in the second Critique). Kant is not clear how this excess arises. At first it appears that the source of our propensity to rationalize against the moral law is borne from the desires we have as sensible beings; but on closer inspection this does not hold up. To see why, consider what Kant says: The human being feels within himself a powerful counterweight to all the commands of duty, which reason represents to him as so deserving of the highest respect—the counterweight of his needs and inclinations, the entire satisfaction of which he sums up under the name happiness [Glückseligkeit]. (G 4:405)
This last sentence ends with an important qualification. Kant is not speaking of our needs and inclinations taken individually, but of the conception we have of their sum-total satisfaction. Later, in Groundwork II, Kant will define happiness as an ‘ideal’ of one day having our wants and wishes fulfilled (G 4:418). The ideal is 15 I agree with Allison’s (2011) observation that the transition of Section I is unique. While the transitions of Sections II and III involve the ‘replacement of one philosophical standpoint by another (popular moral philosophy by metaphysics of morals) and one level of philosophical discourse by another (a metaphysics of morals by a critique of pure practical reason)’, Section I resembles what Allison calls ‘a movement in place’. That is to say, ‘its argument consists in a clarification or making explicit of what is supposedly implicit in a shared, pre-philosophical, understanding of morality’ (2011, 71). 16 ‘Dialectic’ has a pejorative meaning in Kant’s writings, referring to a tempting but ultimately erroneous way of thinking. Presumably the dialectic mentioned in the Groundwork is ‘natural’ because it refers to an error all of us are prone to make in matters of deliberation. It is a dialectic in the sense of a back-and-forth exchange—as we will see, a constant ‘quibbling’ with morality—that without a critique will never find rest.
Moral Skepticism 27 empty, however, because it does not specify in advance what will make us satisfied; we still need experience to figure out how to bring our desires into harmony with each other. This is a small yet significant point of clarification. It shows that the dialectic of common reason does not lie between the moral law and the inclinations. Our inclinations, taken individually, do not present us with an end. The dialectic arises between the two ends we have as beings who are both rational and sensible— namely, between morality and our own happiness. They are set in opposition, in Kant’s view, because the demands of morality do not promise anything to our desires. Viewed from the perspective of common reason, then, the strict laws of duty seem to insult us, to show ‘disregard and contempt’ for the interest we have in our well-being (G 4:405). It is only natural that we will try to defend this interest by striking back, i.e., by stripping moral laws of their very authority.17
1.6 Skepticism in Groundwork II In Groundwork II Kant calls upon philosophy for a second time. At first we needed tools of philosophical cognition to clarify the principle of a good will. This was more or less a theoretical need. But having identified a natural dialectic in our ordinary thinking, Kant now argues we need philosophy on practical grounds. We need help in resolving our tendency to rationalize against the moral law. As we will see, the question framing Groundwork II is what kind of philosophical method is fit for this task, and here Kant criticizes popular philosophy because it plays into the hands of skepticism.18 But what kind of skepticism is at stake? And how does Kant use this problem to motivate his own (less popular) alternative? To address the first question, let us consider the two skeptics Kant introduces in Groundwork II.19 The first does not deny the truth or correctness of the moral 17 As Engstrom puts it, the recognition the moral law inflicts upon the ego ‘may prompt it to strike back in anger’, and it may do this by staging a reversal of representation, throwing back the discovery of its own pretense (2010, 115). 18 In Kant’s view popular philosophy suffers from a mixed method: ‘now the special determination of human nature (but occasionally the idea of a rational nature as such along with it), now perfection, now happiness, here moral feeling, there fear of God, a bit of this and also a bit of that in a marvelous mixture’ (G 4:410). Quite a few late eighteenth-century philosophers fall under this description. Johan van der Zande lists Sulzer, Iselin, Mendelssohn, and Herder, among others (1995, 423). Interestingly, one figure van der Zande does not list is the young Kant himself, who in his early work attempted to draw moral conclusions by observing ways in which sentiments of beauty and sublimity arise in human life (see §4.2). 19 The first skeptic I will be speaking of represents the group of philosophers Kant says have always ‘denied the reality of this disposition in human actions [i.e., to act from duty] and ascribed everything to more or less refined self-love’ (G 4:406). The second skeptic represents ‘the wishes of those who ridicule all morality as the mere phantom of a human imagination overstepping itself through selfconceit’ (G 4:407). What this shows, I believe, is that the two skeptics from Groundwork II do not
28 Skepticism in Groundwork II law. Kant tells us his doubts concern human nature: in his view, we are unable to live up to the moral law’s commands, and so he is skeptical about our motiv ational capacities. Kant describes this philosopher as speaking with ‘deep regret’ that human nature is ‘noble enough to take as its precept an idea so worthy of respect but at the same time is too weak to follow it’ (G 4:406). It is not that he thinks the moral law is impure (for example, that duty is mixed up with selfinterest), but that we are imperfect and so incapable of regarding duty as our incentive. The thought that troubles him is not that moral principles lack validity, but that we lack the capacity to be successfully moved by them. Kant’s second skeptic raises a different concern. Rather than express ‘deep regret’ over the moral weakness of our will, he harbors a malicious wish to ‘ridicule all morality as a mere figment of the imagination [Hirngespinst] overstepping itself through selfconceit’ (G 4:407).20 In his eyes, we frequently meet individuals who use the appearance of morality to hide their selfish motives, from which he draws the conclusion that there is nothing behind the appearance, no objective ground from which the claims of duty can bind us. The everyday experience of moral obligation, the second skeptic argues, is not a product of reason: it is an illusion of the imagination.21 Despite their differences, Kant’s point is that each skeptic commits the same error. Both rest their doubts on experience. For the first skeptic, it is the evidence of our weakness of will—the fact that we seem incapable of regarding duty as an incentive—that questions our ability to live up to the moral law’s commands. For the second skeptic, it is the evidence of our selfishness—the fact that we often disguise our motives under the mask of virtue—that questions moral objectivity itself. In Groundwork II Kant uses a strategy of exaggeration to undermine these concerns. First, he admits that we cannot know for certain whether our actions have a moral motive or a motive drawn from self-love. It may be the case that our actions are but effects of the ‘dear self ’, ‘which is always turning up’ (G 4:407). But Kant’s point is that it would be wrong to draw any conclusions from this. The fact
voice problems specific to any traditional school of thought. Rather, they are general problems that arise naturally in the course of speculating about the nature of morality. As we will see, Kant introduces them in order to expose the shortcomings of popular moral philosophy, the rival to his own metaphysics of morals. 20 In German Hirngespinst means ‘phantasm’, ‘fantasy’, or ‘figment of the imagination’. In Kant’s technical language, a Hirngespinst usually refers to a concept empty of content. For example: ‘An imaginary being is a mere phantom of the brain [Hirngespinst], but of which the thought is still possible. What does not contradict itself is logically possible; that is, the concept is indeed possible, but there is no reality there. One thus says of the concept: it has no objective reality’ (V-Met-L2/Pölitz 28:544). 21 In the second Critique, Kant says that some people (not necessarily philosophers) have a tendency to search out a hidden self-interested motive for every moral action, with the idea that ‘human virtue might in the end be held a mere phantom of the brain, and so all striving toward it would be deprecated as vain affectation and delusive self-conceit’ (KpV 5:154).
Moral Skepticism 29 that we cannot find examples of our moral strength of will does not support the inference that we lack the capacity for such strength. Nor does the fact that people more often than not use morality for their selfish purposes support the inference that morality is entirely subjective. By exaggerating their worries, Kant’s aim is to uncover the fallacy these philosophers share in common: the fallacy of thinking that what we cannot perceive must not exist. Kant thinks we can learn a general lesson from this. The lesson is that we cannot settle questions about morality on the basis of experience. Observations of people’s behavior cannot determine what we are motivationally capable of, nor do they give us purchase on normative claims about how we should act (for example, that we should only attend to the ‘interests of the inclinations’, G 4:406). Admittedly, when we look out into the world we see that people—including ourselves—often act selfishly. We need not be cynical in making such reports.22 Yet the important point is this. If we cannot know whether our actions proceed from a moral motive or a motive of self-love—and if empirical evidence cannot distinguish between the two—then we cannot know from experience whether we have the capacity to act from duty. So if we try to counter the skeptic by offering examples of allegedly good conduct, we will in turn commit the same mistake the skeptic has made. We will develop our stance, which may purport to be anti-skeptical, on the assumption that we can settle moral questions from an empirical point of view. And that would draw us back into the very problem we wished to escape. What is at stake in the beginning of Groundwork II is a question of proper philosophical procedure. As we read further, we see that Kant’s criticism is aimed not so much at skepticism but at a moral theory that subscribes to an empirical method. This is what Kant calls popular moral philosophy, a theory based on a ‘disgusting hodgepodge of patchwork observations’, some drawn from reason, others from experience (G 4:409). Since popular philosophy assumes we can gain insight into virtue by observing others and finding examples of good conduct, it is left defenseless against those who wish to cast doubt upon our higher vocation. Once again Kant’s point is that experience cannot be our testing ground for morality and virtue. As he puts it elsewhere, Whoever would draw the concepts of virtue from experience, whoever would make what can at best serve as an example for imperfect illustration into a model from which to derive knowledge (as many have actually done), would make of 22 ‘One need not be an enemy of virtue’, Kant writes, ‘but only a cool observer, who does not take the liveliest wish for the good straightaway as its reality, to become doubtful at certain moments (especially with increasing years, when experience has made one’s judgment partly more shrewd and partly more acute in observation) whether any true virtue is to be found in the world’ (G 4:407). The idea that sharpening one’s judgment makes one prone to skepticism is also a theme found in the first Critique. There Kant speaks of the ‘childhood’ of pure reason as a kind of self-certain dogmatism that comes to be questioned, presumably in reason’s adolescence, but that still requires criticism to reach ‘mature and adult power’ (A761/B789).
30 The Need for a Metaphysics of Morals virtue an ambiguous non-entity, changeable with time and circumstances, useless for any sort of rule. (A315/B371)
No wonder, then, that popular philosophy fails to secure a foundation for ethics. Once we see this, Kant thinks, it is clear why we need an alternative method, even one that risks unpopularity.
1.7 The Need for a Metaphysics of Morals Kant’s solution is straightforward: If an empirical method fails to secure a foundation for morality, then we must see if we can do better by taking up a different method—a method of rational reflection. In taking up this alternative, we must give up the idea that morality can be derived from some special feature of human nature. As Kant explains, ‘duty is to be practical unconditional necessity of action and it must therefore hold for all rational beings (to which alone an imperative can apply at all) and only because of this be also a law for all human wills’ (G 4:425). Kant announces this new method under the title of a metaphysics of morals, which he also calls ‘pure practical philosophy’ (G 4:410). As a method, we can see that a metaphysics of morals ascends higher up along the analytic path of Groundwork I. As Kant puts it, ‘we leave it undecided whether what is called duty is not as such an empty concept’ (G 4:421)—yet what emerges from this path is unexpected. First, when we think of a practical imperative that abstracts from any interest or inclination, we are led to the formula of universal law: ‘Act only in accordance with that maxim through which you can at the same time will that it become a universal law’ (G 4:421). Second, when we ask what end is objectively valid for all rational beings, we see that it can only be humanity in the form of rational nature, since only rational nature has a dignity that sets it apart from things of contingent value. According to Kant, this yields a second formula, the formula of humanity: ‘So act that you use humanity, whether in your own person or in the person of any other, always at the same time as an end, never merely as a means’ (G 4:429). Kant now argues that the first and second formulas combine to yield a third. Quickly put, his claim is that when we think of a being adopting humanity as its end, we have the idea of a being legislating itself by reason. That is, we have the idea of a being acting autonomously, free of any interest or inclination. So by reflecting on the concept of a rational will in general, a metaphysics of morals leads us to the formula of autonomy: ‘[So act] that the maxims of your choice are also included as universal law in the same volition’ (G 4:440; cf., 4:431).23 Without
23 I have removed the italics from the original versions of each formula.
Moral Skepticism 31 discussing the details of these connections, we can at least appreciate what Kant is trying to achieve in Groundwork II. The analytic path (that of ascending to the highest principle of practical reason) shows that what the moral law demands of us is ‘neither more nor less than just this autonomy’: that is the highest concept for thinking of the moral law as a categorical imperative (G 4:440). Had we kept to an empirical method of investigation, the approach of popular philosophy, we would never have been able to grasp this. Despite Kant’s frequent reminders that his procedure in the first two sections of the Groundwork is merely analytic, his closing statement in Groundwork II still comes as a surprise. He tells us it is possible at this stage of the argument that the concept of duty may be empty, that morality may be a ‘figment of the imagination’ (Hirngespinst) (G 4:445). But why is this still a possibility? And why is Kant telling us this now? It may help to remember that Kant thinks we leave ourselves exposed to skepticism if we develop the concept of duty from experience (G 4:407). The problem with popular philosophy is not only that it leaves us with a patchwork of half-rationalized principles, but that its method consists of observing others and finding examples of virtuous conduct. As a result, popular philosophy is left defenseless against those who wish to denounce morality as an illusion, since experience teaches us that people often act from selfish motives. By adopting a rational method of investigation, we have seen that Kant rejects the assumption the two skeptics from Groundwork II share in common, i.e., that what we cannot perceive must not exist. And this suggests Kant’s closing remarks are addressed to a new audience. As he says, ‘whoever holds morality to be something and not a chimerical idea without any truth must also admit the principle of morality brought forward’ (G 4:445; emphasis added).24 My conjecture is that the new audience consists of ‘whoever’ has, with Kant, made the transition to a metaphysics of morals. If this is right, his comment about the possible illusoriness of duty serves only as a reminder that a metaphysics of morals is limited. Yet again we can see that Kant is using a skeptical method to show why a transition or shift of frameworks is necessary. As a procedure, a metaphysics of morals avoids the pitfalls of popular philosophy—it ascends from our common idea of duty to the principle of autonomy—but within this framework we cannot decide whether this principle is actually binding upon us. By the end of Groundwork II it remains to be seen whether the principle of a rational will in general, expressed by the moral law as autonomy, is also a necessitating imperative for beings like us. For Kant, justifying the moral law’s bindingness requires that we take the synthetic path and descend from the principle of autonomy back
24 Even though the first skeptic does not deny the truth of morality—he only questions our motiv ational capacities—I do not think Kant is speaking to him at the end of Groundwork II. This is because he thinks we can settle moral issues from an empirical point of view, and Kant has rejected this assumption in taking up a metaphysics of morals.
32 The Deduction of the Categories to the ‘common cognition in which we find it used’ (G 4:392). This finally brings us to the deduction of Groundwork III.
1.8 The Deduction of the Categories Right away we are likely to interpret the term ‘deduction’ through the context of modern logic, whereby we speak of deducing a conclusion from a valid set of premises. But Kant’s use of the term comes from the older discourse of Romano-canon law. One striking instance of this discourse occurs at the beginning of Kant’s transcendental deduction of the categories in the first Critique. Here he compares this strategy of argument to a juridical procedure: Jurists, when they speak of entitlements and claims, distinguish in a legal manner between the questions about what is lawful (quid juris) and that which concerns the fact (quid facti), and since they demand proof of both, they call the first, that which is to establish the entitlement or the legal claim a deduction. (A84/B116)25
Kant spent years laboring on a deduction of the categories, which he says cost him the ‘most effort’ in the first Critique (A xiv). Unfortunately, he left us few clues for understanding the path he took to arrive at his deduction in Groundwork III. Given the complexities in each text, it would be unfeasible to compare the two in any detail. Yet I believe an outline of Kant’s deduction of the categories will shed light on aspects of Groundwork III we might otherwise miss. In the older discourse of Romano-canon law, the task of a deduction is to vindicate a disputed claim by providing an account of its rightful acquisition. For Kant, pure concepts of the understanding or categories provoke suspicion of their legitimacy because they make claims of epistemic necessity. For example, the category of cause-and-effect says that one event must, under relevant conditions, follow from another. Accordingly it says: ‘Under relevant conditions water mustfreeze when the temperature drops to zero degrees Celsius’, and this goes far beyond a claim of prediction, that the water will probably freeze at this temperature. What is at issue is not whether we cognize necessity in the categories—for in fact we do—but whether our cognition has an objective ground. A transcendental deduction in Kant’s sense must show this, making it akin to what he calls an ostensive or genetic proof, that is, a proof that provides insight into the sources of a concept (A7891/B817). 25 For example, my claim to an estate would be justified by a document (such as a will) identifying me as the estate’s legal inheritor. In a ground-breaking study Dieter Henrich (1989) has argued that the juridical model gives us insight into the structure of Kant’s transcendental deduction.
Moral Skepticism 33 The difficulty is that the categories are not acquired through sense experience. The mere fact that we use pure concepts like cause-and-effect does not weigh in favor of their legitimacy. Who is to say they are objectively valid? Kant observes that terms like fortune and fate ‘circulate with almost universal indulgence’, yet nobody can provide ‘clear legal ground for an entitlement to their use either from experience or from reason’ (A85/B117). So without a deduction or genetic proof we could not tell whether the categories have entered our mind surreptitiously, say, through custom or habit. The worry is urgent because if the basis of the cat egories turns out to be ‘merely subjective’, we will be forced to abandon the idea that they serve as conditions of possible experience. Thus Kant wants to impress upon the reader why undertaking a justification of the categories is necessary. ‘The reader’, he explains, ‘must be convinced of the unavoidable necessity of such a transcendental deduction’ (A88/B121). Lacking conviction on this point, the reader ‘would otherwise proceed blindly, and after much wandering around would still have to return to the ignorance from which he had begun’ (A88/B121). In order to prevent this, Kant appeals to a skeptical thought in order to push his argument in the right direction.26 Tellingly, he labels this move a ‘transition’ (Übergang) in the title of section 14: ‘Transition to the Transcendental Deduction of the Categories’. The skeptical thought Kant appeals to concerns the coordination of our faculties. He says the unity of appearances may not be a product of categories like cause-and-effect, so that what is given to us in sensibility may not relate to the functions of the understanding. The skeptical worry, in other words, is of a potential disharmony of our faculties: Appearances could after all be so constituted that the understanding would not find them in accord with the conditions of its unity, and everything would then lie in such confusion, e.g., in the succession of appearances nothing would offer itself that would furnish a rule of synthesis and thus correspond to the concept of cause-and-effect, so that this concept would therefore be entirely empty, nugatory, and without significance. Appearances would nonetheless offer objects to our intuition, for intuition by no means requires functions of thinking. (A90/B123)
Kant says the reader may wish to escape the difficulties of a deduction and try, instead, to establish the validity of the categories by appealing to examples drawn from experience. Taking the empirical path he might think we can derive causal necessity from the regularity of appearances, e.g., from phenomena like the constant rising of the sun. But Kant points out that experience can only support
26 See Gardner (1999) for a lucid overview of these issues.
34 Skepticism in Groundwork III inductive generalizations (that B will probably follow A), not claims of necessity (that B must follow A). At this crossroads the alternatives are clear: Pure concepts ‘must either be grounded in the understanding completely a priori or else be entirely surrendered as a mere figment of the imagination’ (A91/B123). According to this threat, we might not be entitled to say that an effect, B, necessarily follows a cause, A, only that we must represent ‘A-B’ together, ‘which is precisely what the skeptic wishes most, for then all of our insight through the supposed objective validity of our judgments is nothing but sheer illusion’ (B168). The problem is that the categories might only be rules for thinking objects in general, as thinking does not require sensory data, but they would not be applicable to whatever we perceive, i.e., to objects given to us in space and time. In addition to a first step that would show the validity of the categories for a manifold of intuition in general, Kant’s argument in the first Critique must take a second step, one that would show the application of the categories for a human manifold in particular. Indeed, a second step is necessary because the domain of human intuition is unique—it is characterized by space and time—and we need to know if the categories extend this far. In this light, the success Kant’s deduction depends on showing the connection between the pure concepts of the understanding and the pure forms of sensibility, which had previously been separated for the sake of analysis. The deduction must now re-combine these two modes of cognition and reveal their necessary synthesis.
1.9 Skepticism in Groundwork III It may not be obvious how any of this relates to Groundwork III,beyond the fact that Kant speaks of a ‘deduction’ in each text.27 But we are now in a better pos ition to see how the argument of Groundwork III unfolds. Reviewing our steps, we have seen that in Groundwork II Kant is troubled by an approach to morality that leaves itself exposed to skepticism: the approach of popular philosophy. This is a significant turning point in his argument, for Kant had just shown in Groundwork I that common human reason is susceptible to a natural dialectic which requires it to seek help in philosophy. If common reason turns to popular philosophy for a clarification of its moral commitments, it would likely become more entangled in self-obscurity. Does Kant think bad philosophy can enter the dialectic of common reason? The naturalness of the dialectic suggests that in our everyday frame of mind we do not need the help of empirical theories to justify placing our interest in happiness above the moral law. At the same time, we could readily do this by 27 See Choi (2019) for an illuminating discussion of the parallels between Groundwork III and Kant’s version of the transcendental deduction in the 1787 edition of the first Critique.
Moral Skepticism 35 turning to the reductive explanations such theories endorse, i.e., explanations that downgrade the necessity of the moral law along with our experience of duty.28 If we could convince ourselves that categorical imperatives are really hypothetical, we could justify suspending them in situations where we would like to act out of self-love. So even if empirical theories of morality have an origin outside our natural dialectic, they could still be re-appropriated under the guise of respectable philosophical accounts—accounts that would serve to back up the rationalizations Kant thinks we are prone to make when faced with the strict requirements of duty.29 This idea is worth repeating, because it is easy to forget that one of Kant’s central tasks in the Groundwork is to rescue common reason from itself.30 As we have seen, the practical need for philosophy only becomes explicit at the end of Groundwork I when we discover we have a tendency to rationalize against the moral law (see §1.5). That is why Kant says we are compelled to seek help in phil osophy, not only to satisfy a ‘need of speculation’, he adds, ‘but on practical grounds themselves’, so that we may escape the conflict that inevitably arises between morality and our own happiness (G 4:405). What I now want to show is that the problem of a natural dialectic also plays a central role in the final section of the Groundwork. For in the Preface Kant tells the reader that the dialectic we suffer from in our ordinary thinking will only find rest in a ‘complete critique of our reason’ (G 4:405). This is part of the final step Kant thinks we need to take in Groundwork III, namely, to a ‘critique of pure practical reason’.31
1.10 The Source of Moral Obligation What does this step involve? Early in the Groundwork Kant leaves us with a hint. At the end of Groundwork I he says the need of common reason to seek help in
28 Rawls offers a slightly different point when he speaks of ‘the depth of Kant’s conviction that those without a conception of the moral law and lacking in moral sensibility could not know that they were free. They would appear to themselves as purely natural creatures endowed with rationality, without the essentials of humanity . . . The empiricist “delusion,” as Kant calls it, must not be allowed to take from us the glorious disclosure of our autonomy made known to us through the moral law as an idea of pure reason’ (1988, 133). In §2.11 we will find Kant making a similar point in his reply to Garve. 29 Here I believe we can build upon Guyer’s observation that ‘the risk to our moral self-understanding does not arise from without, from wily artifices of corrupt philosophers who appear out of nowhere to darken our paths’. Rather, ‘the source of sophistry and corruption lies within us, in evil possibilities inherent in our own nature, which can in turn co-opt our own faculty of reason to produce a form of philosophy that would appear to justify our lapses from duty’ (2000, 209). 30 In addition to Guyer (2000), the idea that the Groundwork’s task is educative is voiced by Philonenko (2008, 13–14), but only in passing. 31 Kant takes up the problem of a ‘dialectic’ once more in the second Critique. Yet we should be attentive to the fact that the dialectic of the second Critique is of pure practical reason (cf., KpV 5:107). That would set it apart, as a philosophical problem, from the dialectic of the Groundwork, which arises for sensibly conditioned practical reason. See Klemme (2010) for further discussion.
36 The Source of Moral Obligation philosophy is similar to the need we have in matters of speculation. Here I take Kant to be alluding to the Transcendental Dialectic from the first Critique where he shows that we suffer from a tendency to ‘overstep’ the proper bounds of human cognition in seeking objects of knowledge. We fall into a ‘chaos of uncertainty’ in laying claim to objects outside any relation to sense experience (to God, for example). In the Groundwork, however, Kant implies that common reason suffers from the opposite problem: We cannot consider without admiration how great an advantage the practical faculty of appraising has over the theoretical in common human understanding. In the latter, if common reason ventures to depart from laws of experience and perceptions of the senses it falls into sheer incomprehensibilities and self-contradictions, at least into a chaos of uncertainty, obscurity, and instability. But in practical matters, it is just when common understanding excludes all sensible incentives from practical laws that its faculty of appraising first begins to show itself to advantage. (G 4:404)
In matters of knowledge, our speculative use of reason becomes dialectical when we overstep the sensible conditions that make human cognition possible. By contrast, in matters of deliberation our practical use of reason displays its authority when we ignore sensuous incentives—for then we can judge an action’s worth in terms of its lawful form, stripped from any consideration of self-interest. This means, to bring out the reverse point, that we fall into a ‘chaos of uncertainty, obscurity, and instability’ when we let sensible incentives encroach upon our practical reason. That is when our self-love becomes excessive and when we begin to ‘quibble’ with the demands of duty. All of this goes to show why the Groundwork needs a final shift of frameworks. If we can show that our faculty of reason has a pure use, separate from its empirical use, then we can see that the only principle legislative for our will is autonomy—that of being a law to ourselves—which is equivalent to morality. A critical examination of this sort would show that our tendency to rationalize against the moral law is without basis: it occurs when we allow what is only a part of ourselves—namely, our sensibility—to act as if it constituted ourselves as a whole. If successful, then, Kant’s critique would vindicate the idea of our higher vocation first expressed in Groundwork I, ultimately showing that we have not ‘misunderstood the purpose of nature in assigning reason to our will as its governor’ (G 4:394–5). It goes without saying that the analytic path opened up by a metaphysics of morals is unable to achieve this end. Although Kant leaves us in the dark here, we can begin to understand his strategy better by looking at the title of subsection four in Groundwork III: ‘How is a Categorical Imperative Possible?’ An even more revealing piece of evidence
Moral Skepticism 37 comes a few paragraphs later where Kant draws a parallel to the transcendental deduction of the categories. After summarizing the argument, he adds: ‘this is roughly like the way in which concepts of the understanding, which by themselves signify nothing but lawful form in general, are added to intuitions of the world of sense and thereby make possible synthetic propositions a priori on which all cognition of a nature rests’ (G 4:454). If we take Kant at his word, then the parallel of a transcendental deduction in Groundwork III is to be found in his answer to the question of how a moral ‘ought’ is possible. And this makes sense, considering that our cognition of an ‘ought’ expressed in the moral law provokes the same sort of suspicion raised by the epistemic ‘must’ in the categories. In both cases we are faced with claims marked by necessity, and the question is whether these claims have an objective ground. We want insight into the source of moral obligation, in order to vindicate our possession of it, just as we want insight into the source of the categories.32 The first step of Kant’s deduction in Groundwork III introduces different but compatible ways we can view ourselves as agents. On the one hand, when I view myself from the standpoint of an observer, in the third person, I see that my actions are part of the same ‘world of sense’ that govern objects and events around me. My actions are the product of forces beyond my control. On the other hand, when I view myself as an agent, in the first person, I presuppose a different space: a ‘world of understanding’ or ‘intelligible world’ unaffected by natural influences, including the influences of my sensibility (G 4:452).33 I recognize that I am, not only acted upon, but active myself. Kant uses this distinction to show that when I view myself exclusively from the standpoint of agency, participating in the intelli gible world, I cannot derive a principle of action from my inclinations as a sensible being. I am now considering myself ‘outside’ of the space of causes to which my inclinations belong. So the only principle I can derive is that of being a law to myself, and that is the principle of autonomy. As a participant in the intelligible world, then, I see that a principle of being a law to myself is valid for me without
32 I am following Schönecker (2006), Timmermann (2007), and Stern (2011) in framing the deduction of Groundwork III as an argument, not for the validity of the moral law, but for its bindingness as a categorical imperative. Guyer has recently expressed resistance to this interpretation (2007, 2008). In his view, the bindingness problem only concerns the ‘imperatival character of the fundamental prin ciple of morality for us’, yet this, he concludes, is ‘a statement of a consequence of the deduction rather than the attempted deduction itself ’ (2009, 181, note 7). There is evidence that Kant’s deduction of the categorical imperative in subsection four presupposes a previous deduction of freedom in subsection three. This evidence would make my view of the Groundwork’s proof-structure compatible with Guyer’s. The remaining question—which I will touch upon in §3.8—is whether Guyer is right to interpret the initial deduction of freedom in strongly metaphysical terms. 33 As Kant writes: ‘One resource, however, still remains to us, namely to inquire whether we do not take a different standpoint when by means of freedom we think ourselves as causes efficient a priori than when we represent ourselves in terms of our actions as effects that we see before our eyes’ (G 4:450; emphasis added).
38 The Source of Moral Obligation the mediation of my inclinations: in this way, it is the only principle of my will qua will, the only principle I give to myself immediately. But as a step in a genetic proof, this gives us a limited result. The two-world distinction says that when we regard ourselves as agents we presuppose a space undetermined by sensibility, and from the standpoint of this space our practical reason is ‘more’ than a sensibly conditioned faculty. A critic might concede, however, that the moral law is valid for rational beings, and so for ourselves con sidered only as participants in the intelligible world. Yet this leaves open the question of whether the moral law applies to beings like us, who are affected by sensibility (G 4:450; cf., 4:453). Just as step one from the transcendental deduction leaves open the question of whether the categories are valid in their application to our particular manifold of intuition—as it is characterized by space and time—the first step of Groundwork III leaves open the question of whether the moral law is binding for human agents, as they are characterized by sensible desires.34 What this shows, as before, is that a second step is necessary to complete the deduction. It is important to bear in mind that Kant treats both worlds in Groundwork III as standpoints for thinking of our agency. We can consider ourselves from the standpoint of the sensible world, on the one hand, and from the standpoint of the intelligible world, on the other (G 4:452). We can agree, then, that if the intelli gible world affords us a standpoint for thinking of our ‘pure’ will, the world of sense affords us a standpoint for thinking of our ‘empirical’ will. In this sense Kant’s two worlds stand in a unique relationship to each other because one gives us a concept of our will as self-sufficient, whereas the other gives a concept of our will as dependent. To be sure, each world affords us insight into the nature of our will, I am not denying that, but my point is that only the intelligible world affords us insight into the ‘idea’ of our will as a self-legislating faculty—where an ‘Idea’ signifies, in Kant’s epistemology, the maximum perfection of a concept.35 The world of sense affords us insight into the nature of our will, too, but only as a faculty dependent on impulses and inclinations. 34 The question of parallelism in Kant’s deductions is intriguing yet difficult to settle. All I am committed to here is parallelism between (a) a general step (the validity of the categories for a manifold of intuition in general/the validity of the moral law for a rational being in general) and (b) a specific step (the validity of the categories for a human manifold of intuition/the validity of the categorical imperative for a human will). Guyer has attempted to draw a more elaborate set of connections ranging from Kant’s derivation of the categorical imperative in Section II of the Groundwork to his system of duties in the Metaphysics of Morals. For details, see Guyer (2007). For a critical reply, see Allison (2007). 35 In the first Critique Kant draws inspiration from Plato in this definition of an ‘idea’ (Idee). With respect to the idea of a political constitution ‘providing for the greatest human freedom according to laws that permit the freedom of each to exist together with that of others’, Kant writes: ‘Even though this may never come to pass, the idea of this maximum is nevertheless wholly correct when it is set forth as an archetype, in order to bring the legislative constitution of human beings ever nearer to a possible greatest perfection’ (A317/B373). I have capitalized the English to make this technical feature of Kant’s term salient. This point is developed by Tenenbaum (2012) in an excellent essay, to which I am indebted.
Moral Skepticism 39 In saying this, Kant’s claim is that the Idea I have of my will from the standpoint of the intelligible world is the ‘ground’ of my will as I conceive it from the standpoint of the sensible world. The concept of my will as it is sensibly affected is a limited concept of my will; it only reveals what is a part of my faculty of practical reason. The Idea of my will as it is unaffected by sensibility, i.e., the concept of the maximum perfection of my will, reveals my faculty of practical reason as a selfsufficient whole. Kant indicates as much in the following passage: This categorical ought represents a synthetic proposition a priori, since to my will affected by sensible desires there is added the idea of the same will but belonging to the world of the understanding—a will pure and practical of itself, which contains the supreme condition, in accordance with reason, of the former will. (G 4:453)
This passage is easy to overlook, but in my view it is essential for understanding the second step of Kant’s deduction in Groundwork III. Moral laws carry the force of imperatives for us because our empirical will is not a separate faculty at all. Whenever we exercise our will in the world of sense, we presuppose the idea of our same will yet unaffected by needs, impulses, and inclinations. For this reason moral laws do not concern the activity of a separate faculty, a ‘pure will’ belonging to some other realm. The pure will is nothing other than the Idea of our own, empirical will conceived of in its complete perfection. That is why the moral ‘ought’ expresses my own will as a member of the intelligible world: I experience it as an ‘ought’ only because I must view myself ‘at the same time’ (zugleich) as part of the world of sense (G 4:455). What Kant’s genetic proof shows, then, is that my experience of necessitation in the categorical imperative has an objective basis after all. What the moral ‘ought’ expresses is, not a figment of my imagination, but my own pure ‘will’ (G 4:454).
1.11 Kant’s Scoundrel One advantage of the interpretation just sketched is that it shows why Kant’s deduction in Groundwork III works to resolve the natural dialectic of common human reason. As we have seen, Kant describes this dialectic in Groundwork I by speaking of a ‘counterweight’ we feel ‘to all the commands of duty’ (G 4:405). The problem, he explains, is that reason ‘issues its precepts unremittingly, without thereby promising anything to the inclinations, and so, as it were, with disregard and contempt for those claims’ (G 4:405). To protect the interest we have in our own happiness, we in turn strip moral laws of their status, either by casting doubt upon their validity or by questioning their strictness. By way of rationalization we try to make our compliance with the moral law conditional upon our well-being. And that, for Kant, is how the natural dialectic of common reason arises. What
40 Kant ’ s Scoundrel starts off as a desire to protect our interest in happiness ends up as a choice to subvert the moral law, to dislodge it from its proper place of authority. But what is the origin of this choice? Much of what Kant says is shrouded in mystery, but he leaves us with some suggestive remarks. The natural dialectic arises, he writes, as soon as common reason ‘cultivates itself ’ (G 4:405). We can think of it like this: Our first vocation as sensible beings comes from the idea of our happiness, the imagined end of our sum-total satisfaction. Over time we become aware of other reasons for action, ‘laws’ that do not speak to the fulfillment of our needs and inclinations. A conflict inevitably arises, in turn, because our initial grasp of these laws is distorted: all we see is their unremitting character, without recognizing their true source. At this stage we are capable of moral thought in some basic sense—we can consider duties of action—but we do not yet know where these duties come from. All we see are external claims that pose a threat to our first vocation, happiness. In this way our distorted grasp of duty is a matter of perspective, the perspective we have as beings who are unaware that moral claims come from within us. The example Kant uses to ‘confirm the correctness’ of his deduction in Groundwork III reveals the importance of the problem at hand: There is no one—not even the most malicious scoundrel, if only he is otherwise accustomed to use reason—who, when one sets before him examples of honesty of purpose, of steadfastness in following good maxims, of sympathy and general benevolence (even combined with great sacrifices of advantage and comfort), does not wish [wünsche] that he might also be so disposed. (G 4:454)
Anyone confronted with these examples must see something remarkable. The agent they characterize becomes more sublime in our eyes the more he suffers a loss of personal well-being. The wish we feel in response—to be disposed of such strength of will—indicates that our higher vocation cannot be the pursuit of happiness. If it were, we would not want to be like someone whose moral resolve leads him, in the examples at hand, to great sacrifices of advantage and comfort. Kant makes this point earlier in the Groundwork when he claims, appealing once more to ‘the most ordinary observation’, that an action of integrity ‘done with steadfast soul, apart from every view to advantage of any kind in this world or another and even under the greatest temptations of need or allurement’ has a more powerful effect on us than a similar action motivated by self-interest or love of honor. Only the former, he adds, ‘elevates the soul’ of the spectator and ‘awakens a wish to be able to act in like manner oneself ’ (G 4:411n).36 36 This example from Groundwork III foreshadows Kant’s use of an experimental method in the second Critique, which we will discuss again in §2.5 and §2.7.
Moral Skepticism 41 The moment we express this wish, Kant claims, we must shift standpoints, stepping outside the world of sense to which our needs and inclinations belong (G 4:454). In wanting to emulate the virtuous agent, we must place ourselves in the intelligible world, and only there do we recognize our capacity for pure practical reason. Kant’s point, as we have just seen, is that even a malicious scoundrel must see this—even he must recognize his autonomous self reflected back to him in the examples of moral conduct. Once he shifts standpoints and regards himself in this new light, ‘he is conscious of a good will that, by his own acknowledgments, constitutes the law for his evil will as a member of the world of sense—a law of whose authority he is cognizant even while he transgresses it’ (G 4:455). Having already articulated the supreme principle of morality, Kant has thus come full circle and returned to the common idea of duty first presented in Groundwork I. The scoundrel now before us recognizes the moral ‘ought’ as his same ‘will’ in the intelligible world, and it is ‘thought by him as an “ought” only insofar as he regards himself at the same time as a member of the world of sense’ (G 4:455).
1.12 The Delusion of Skepticism Looking back, we can see why nothing Kant says at the end of the Groundwork would satisfy a person looking for a motive to be moral. The argument of Groundwork III takes up a different question altogether. As Hill points out, Kant is not speaking to those ‘who are indifferent to morality and demand that phil osophy supply them with a motive to be moral’.37 And to this extent Hill and others are right: Kant is not trying to refute a radical skeptic or amoralist. His aim is to resolve the dialectic we are all prone to succumb to by showing that morality is the principle of our will as a whole,considered in itself, whereas happiness only applies to a part of our will, considered in relation to sensibility. Yet it is only when we descend along the synthetic path of Groundwork III that we attain this insight and see the necessary connection between the two. The pure will revealed to us from the standpoint of the intelligible world is nothing other than the ‘Idea’ of our empirical will revealed to us from the standpoint of the sensible world. For this reason we should not turn to Groundwork III for an answer to the skeptic’s standard question, ‘Why be moral?’ If the skeptic is asking how morality will further one of his given desires, then he is begging the whole issue at stake. He is treating morality as a system of hypothetical imperatives, placing his own interests and inclinations in first rank. That is not to say Kant’s attitude to skepticism is entirely dismissive in the Groundwork. The results of Groundwork III 37 Hill (1998, 250).
42 Closing Remarks must speak, at least indirectly, to those who wish to ridicule morality as a figment of the imagination. Like any ordinary person who engages in rationalization, the skeptics Kant considers in Groundwork II would display a genuine practical error if they asserted their happiness above the moral law. By doing so they would be lost in the same self-obscurity Kant says common reason finds itself. They would be guilty of giving priority to the satisfaction of their inclinations, aside from whatever false theory of ethics they uphold. So in the end I believe we can turn the skeptics’ words against themselves. If anyone has succumbed to a delusion, as they claim, it is not the moralist, but the skeptic himself, insofar as he has presumed that sensibility constitutes our whole self and happiness our highest end. This is the sense in which our need for phil osophy is, as Kant says, practical in nature. Skeptical claims are troubling, not because they attack our deepest moral intuitions from the outside, but because they resonate with unspoken doubts we feel within ourselves. I have proposed that we read Kant’s Groundwork as a method for invoking such doubts in order to expose, and overcome, the limits in our ordinary and philosophical thinking about morality. It seems that an actual skeptic stands in need of this help, no less than anyone else. After all, part of the problem Kant is addressing in the Groundwork is how our rational development leads us to come into conflict with the demands of morality. We experience those demands as external attacks to what we initially regard as our sole vocation, the pursuit of happiness. That is why we need insight—and for Kant, this is what the philosopher can offer—into the true source of those demands.
1.13 Closing Remarks In this chapter I have argued that Kant’s project in Groundwork III is best clarified in light of his own distinction between an analytic path and a synthetic path. We follow the analytic path when we ascend, in the first two sections, to the moral law as the highest principle of practical reason. We follow the synthetic path when we descend, in the third section, from the moral law back to the common use of reason in which we find it used. In this way the path of justification in Groundwork III culminates in Kant’s claim that what we recognize we would do as members of the intelligible world is what we should do as beings who belong at the same time to the sensible world. Indeed, after separating the faculty of practical reason from any empirical element, Kant’s task is to show that the highest law of this faculty, the moral law, is binding upon the will of sensibly affected beings like us. It is thus only by taking the synthetic path, where we move from the first principle back to our original starting point, that we can see a real synthetic connection between the moral ‘ought’ and our ‘will’, constituting a genetic proof or
Moral Skepticism 43 deduction of the former. Despite its complex apparatus, then, the project of Groundwork III works to resolve a practical problem, that of our tendency to rationalize against the claims of morality. In this way Kant is not trying to convince us to enter into the world of morality, but rather to make that world intelligible to those of us already committed to it.
2
The Fact of Reason We have at hand examples of reason judging morally. We can analyze them into their elementary concepts and, in default of mathematics, adopt a procedure similar to that of chemistry—the separation, by repeated experiments on common human understanding, of the empirical from the rational that may be found in them. – Kant (KpV 5:163)
2.1 Introduction The third section of the Groundwork constitutes Kant’s most sustained engagement with the project of moral justification. But Groundwork III would not be Kant’s last effort to address questions pertaining to the reality of human freedom and the normativity of the moral law. When we turn to his next major writing on ethics, the second Critique, we find Kant returning to many of the themes canvassed in the Groundwork, including the themes we have examined in the previous chapter. What has struck many commentators as puzzling is that Kant’s strategy in the second Critique appears to have undergone a ‘great reversal’, to use Karl Ameriks’s phrase.1 At the root of this puzzle lies a seemingly new starting point for Kant’s argument: the doctrine of the ‘fact of reason’ (Factum der Vernunft), according to which our consciousness of the moral law admits of no proof—no deduction, genetic or otherwise—but is nonetheless ‘firmly established of itself ’ (KpV 5:46). Insofar as we read Kant’s project in Groundwork III as seeking precisely such a deduction, it is natural to suppose that he reversed, abandoned, or at least seriously modified his method of justification in the second Critique. On one familiar reading, Kant recognized that his deduction of the moral law in Groundwork III failed, and he tried to remedy the situation in the second Critique by putting forth the moral law as an underivable ‘fact’. Yet in the eyes of many scholars, this last move failed, too, since it exposed Kant to the charge of dogmatism. Before turning to compare Kant’s arguments in the Groundwork and the second Critique, my first item of business is to clarify what Kant’s doctrine of the fact of
1 Ameriks (1982, 226). Kant’s Justification of Ethics. Owen Ware, Oxford University Press (2021). © Owen Ware. DOI: 10.1093/oso/9780198849933.003.0003
The Fact of Reason 45 reason amounts to. This is no easy task, as we shall see. I have therefore devoted the present chapter to a close investigation of Kant’s Factum, postponing all comparative questions for the next chapter. My interpretive strategy will be indirect, since a bit of detour is necessary if we want to understand Kant’s doctrine of the fact of reason, bringing us to the traditions of Romano-canon and English common law; to the works of the British experimentalists in the seventeenth century; and finally, to the German translation of Bishop Butler’s Analogy of Religion. While my aim in this chapter is to defend a new systematic reading of Kant’s Factum, I believe that tracing the history of ‘fact’ through these phases is necessary for putting his argument in its proper context. As we shall see, Kant’s Factum shares the meaning of a ‘matter of fact’ (Tatsache),2 referring to the reality of our moral consciousness; and like a Tatsache, it is something we can attest to with the aid of thought experiments. Later on I will address a few possible objections to this reading, one of which I anticipate coming from Dieter Henrich and Ian Proops,3 who have argued that Kant’s Factum is best understood under a legal analogy. Aside from offering a novel perspective on Kant’s intellectual development, the reading I wish to defend in this chapter will have two significant results for my study as a whole. First, it will show why Kant’s mysterious allusions to chemistry in the second Critique are of central importance for understanding his experimental method (KpV 5:92; cf., 5:163). When Kant says a philosopher has an advantage ‘like a chemist’, he means we can set up thought experiments that capture the way we separate duty from happiness and acknowledge the authority of the former. This shows why Kant’s appeal to moral consciousness in the second Critique is not dogmatic, contrary to a prevailing view in the literature.4 On my account, while Kant’s Factum does not admit of a deduction or genetic proof, it does admit of illustration by way of the experimental method. Another payoff to my reading, which I will discuss at the end of this chapter, is that it shows why Kant’s project of moral justification shifts focus from the standpoint of philosophical speculation to the standpoint of ordinary life. It is only when we philosophize from a common standpoint, which is both practical and first-personal, that we can understand what human reason already knows ‘in its heart’. If we can appreciate
2 During the 1790s one finds the expressions ‘Tatsache of reason’ and ‘Factum of reason’ used interchangeably. Among recent commentators, Kleingeld (2010) offers a helpful analysis of these linguistic points. To my knowledge no previous commentator has investigated the Factum/Tatsache connection through the history of fact. 3 See Henrich (1989) and Proops (2003). 4 See Guyer (2007, 462) and Wood (2007, 135) for the charge of dogmatism. Similarly, in Ameriks’s view only ‘some technical peculiarities’ prevent us from labelling Kant’s position in the second Critique ‘fundamentally intuitionistic . . . and to this extent he can be said to have encouraged the return, at least in Germany, to a kind of dogmatic metaphysics’ (2003, 184). Schopenhauer drew this verdict long ago, and in less flattering terms, when he wrote that Kant’s doctrine of the fact of reason opened a ‘gateway’ for all ‘philosophasters and phantasists’ to bring forth their spurious theories (citing the work of Jacobi, Reinhold, Fichte, and Schelling) (1841/1988, 45).
46 A History of ‘ Fact ’ this point, I believe we shall be closer to seeing why Kant’s doctrine of the fact of reason is of lasting philosophical value.5
2.2 A History of ‘Fact’ In the Romano-canon tradition, matters of fact pertained to a sphere distinct from matters of law, as expressed by the maxim: ‘You give me the facts, I give you the law’ (‘Da mihi facto dabo tibi ius’), from Quintilian’s Institutes of the Orator.6 The first entry in the Oxford English Dictionary also defines matters of fact as ‘that portion of a subject of judicial inquiry which is concerned with the truth or falsehood of alleged facts’, citing its first record in English from 1583 (‘He speaketh of a matter of fact’, by Nowell and Day). During the Restoration era, facts came to acquire a more fixed meaning in criminal courts, often referring to the acts and deeds of the accused (coming from the Latin factum, ‘something done’). Facts in this sense required proof. Quite distinct from the meaning they would later acquire, facts were not objects of reasonable belief, but items of evidence to be determined by a jury. Thus, in English common law, ‘matters of fact’ referred to what the accused had done (for example, when he was last seen on the night of the murder), yet their domain was still one of alleged truth. While it is difficult to say when the concept acquired a privileged epistemic status, early modern historians were key in shifting the meaning of fact. For his torians of the time, the relevant contrast was not between matters of fact and matters of law, but between matters of fact and matters of opinion. John Selden, for example, described his Historie of Tithes as a collection of ‘such things of fact’, and others wrote that a faithful historian must report ‘nothing but fact’.7 However, the concept did not yet enjoy an elevated status. As Francis Bacon viewed the matter, ‘a belief of history (as the lawyers speak, matter of fact)’ and a ‘matter of art and opinion’ belong to the same category: that of ‘things weakly authorized’.8 In this respect historical facts required further evidence to warrant assent from others. The job of the historian, like the lawyer, was to convince his audience of matters
5 From this introductory sketch, my position may appear to be at odds with commentators who read the fact of reason as a kind of ‘act’ or ‘deed’, drawing from Kant’s remark that pure reason ‘proves its reality and that of its concepts by what it does [durch die Tat]’ (KpV 5:3). See, for example, Willaschek (1991), Engstrom (2002), and Franks (2005). Nevertheless, I believe we can reconcile these readings if we maintain that moral consciousness only arises through an original act of reason’s selfdetermination. This would still make moral consciousness the primary referent of Kant’s Factum, but it would preserve the important insight by Willaschek, Engstrom, and Franks that the moral law is one we actively give to ourselves. Moreover, the scientific sense of fact. I shall uncover from the work of Boyle and others is consistent with the ‘deed of reason’ interpretation, for the experimental method is meant to reproduce a form of autonomous self-activity in separating morality from happiness. 6 Cited in Shapiro (2000, 9). 7 Cited in Shapiro (2000, 40). 8 Bacon (1605 [1857], 288).
The Fact of Reason 47 of fact; and in both cases personal observation and reliable testimony carried the burden of proof. During the seventeenth century, the concept of ‘fact’ found a new home in the work of Robert Boyle, Robert Hooke, Joseph Glanvill, and other members of the Royal Society of London. One point scholars agree upon is that the practice of giving testimony in courts of law was consciously imitated by the British experimentalists in their efforts to redefine the boundaries of scientific methodology.9 Within this new context, appeals to fact were considered appropriate because the kind of certainty found in logic or mathematics was no longer a plausible goal.10 In many ways advocates of the experimental method wanted to rethink proof and certainty according to an increasingly modest framework of explanation. Boyle, for example, did not claim to understand the causal mechanism of air through his experiments. Rather, he only claimed to witness its effects through the technology of the air-pump. This shift was essential for members of the Royal Society who wanted to secure a foundation of probable knowledge while avoiding contentious issues about causality.11 By the time Locke wrote An Essay Concerning Human Understanding at the close of the seventeenth century, this new way of thinking was common currency. Yet there is no question that Locke contributed to this shift by raising the concept of fact to a philosophical category. In the chapter of the Essay devoted to degrees of assent, Locke identified a matter of fact with ‘some particular Existence’ cap able of observation and testimony, allowing us to ‘reason and act thereupon with as little doubt, as if it were perfect demonstration’.12 As he explained: ‘Thus, if all English-men, who have occasion to mention it, should affirm, that it froze in England the last Winter, or that there were Swallows seen there in the Summer, I think a Man could almost as little doubt of it, as that Seven and Four are Eleven.’13 As a category of knowledge, matters of fact did not admit of strict proof, the kind we find in logic or mathematics, yet they came close to commanding the same degree of assent. ‘These Probabilities rise so near to Certainty’, Locke concluded, ‘that they govern our Thoughts as absolutely, and influence all our Actions as fully, as the most evident Demonstration: and in what concerns us, we make little or no difference between them and certain Knowledge’.14 9 See Shapiro (2002, 250–1). 10 See Serjeantson (1999, 159–60). 11 Although Bacon was the first to transfer juridical methods into natural philosophy, it was Boyle (1661/1772) who developed a strategy of proof specific to English common law, the strategy of multiplying witness testimony. See Shapin (1984), Shapin and Schaffer (1985), Serjeantson (1999), and Shapiro (1994, 2000, 2002). Very often the technology of the experiment was so rare, as with Boyle’s air-pump, that only a few individuals could ever experience the effects first-hand. As Shapin has argued, Boyle attempted to supplement the absence of first-hand witnesses with literary technologies—his own written reports and illustrations of the case—that would reproduce at a public level the results discovered by only a few individuals. Every reader of Boyle’s reports could in theory retrace the steps of the experiment and judge its success or failure, as if he or she were present (1984, 493). 12 Locke (1689/1975, IV.xvi, 6). 13 Locke (1689/1975, IV.xvi, 6). 14 Locke (1689/1975, IV.xvi, 6).
48 From England to Germany This historical overview shows just how much the meaning of fact changed over the course of the early modern period. In the Romano-canon tradition, matters of fact were distinguished from matters of law, referring to human actions relevant to judicial inquiry. In English common law the term became more fixed, referring to the alleged acts and deeds of the accused. While the expression spread rapidly to other contexts over the course of the seventeenth century, the British experimentalists were vital in shifting the reference of facts from human deeds to natural phenomena. In their hands, facts were effects (often reproduced through experiment) that warranted the scientist’s full conviction. Over a period of two hundred years, then, the legal sense of fact gave way to what may be called a scientific sense, coming to signify natural phenomena one could attest to with certainty, even without claiming to know their underlying causes.15 By the time of Locke’s Essay, facts were no longer alleged, doubtful, and in need of proof. They were actual, immediately certain, and firmly established.
2.3 From England to Germany The term ‘matters of fact’ (Tatsachen) entered the German language rather late in the eighteenth century. The expression was coined by the German theologian Johann Joachim Spalding in his 1756 translation of Butler’s Analogy of Religion, first published in 1736. In this work Butler vigorously defended the compatibility of natural and revealed religion, arguing on the basis of fact to establish a conclusive proof of the doctrine of a future state of rewards and punishments.16 Still working within Locke’s framework—where facts rise ‘so near to Certainty’— Butler gave the concept a new twist. If natural religion is based on matters of fact—for example, that we actually possess a moral faculty of judgment—and if such facts are consistent with the doctrines of revealed religion, then we have sufficient reason to believe those doctrines are true. In the absence of demonstrative proof, we have what Butler called practical proof: ‘fully sufficient, in reason, to influence the actions of men, who act upon thought and reflection’.17 Of course, the notion of practical proof assumes that we are free in our actions; and Butler was aware of the objection that we are dictated by causes beyond our control, the ‘objection from Necessity’. Yet in his view the objection has no significance; it only arises from the standpoint of speculation. From the standpoint of deliberation, Butler argued, we are already conscious of a ‘rule of action’ within ourselves—a rule, moreover, ‘of a very peculiar kind: for it
15 For a related claim, see Johnston (2004). I have also benefited from Austin’s (1961) interpretation of the OED entry on ‘fact’. 16 Butler (1736/1896, 369). For further discussion, see the outstanding article by Russell (2004). 17 Butler (1736/1896, 156).
The Fact of Reason 49 carries in it authority and a right of direction; authority in such a sense, as that we cannot depart from it without being self-condemned’.18 This is a matter of fact that no amount of speculation can call into question. Despite the affinities in their thinking, there is no evidence that Kant ever read Butler. Nevertheless, there are good reasons to believe that the English bishop had an indirect influence on him. The mediating link was Spalding himself, whose best-seller, The Vocation of Humankind (Die Bestimmung des Menschen), first published in 1748, bears the unmistakable imprint of Butler’s theistic claims. In this work Spalding argued, like Butler, that we can infer God’s plan for us by reflecting on features of human life in the present. The sense of injustice everyone feels, for example, when vice goes unpunished or virtue unrewarded must guide a reflective person to the conclusion that such oversights will be made up for in a future state, leading to the belief: ‘I am therefore created for another life.’19 Spalding’s book also sparked a major controversy among German theologians during the 1770s and 1780s on the question of humankind’s ‘vocation’ (Bestimmung), and both sides of the controversy appealed to Tatsachen for support.20 On Spalding’s side, for example, Moses Mendelssohn argued that if we want to know what designs Providence has in store for us, we should not ‘work up hypotheses’ but only look to what ‘actually happens’, that is, to matters of fact (Tatsachen).21 By the 1790s, the concept developed further to include what philosophers working in wake of Kant’s philosophy called ‘facts of consciousness’ (Tatsachen des Bewusstseins). Karl Reinhold used this phrase, for instance, to designate
18 Butler (1736/1896, 150). 19 Spalding (1748/1997, 192). There are many noteworthy parallels between the work of Spalding and Kant. Early on in the Vocation, for example, the first-person narrator of Spalding’s text laments: ‘It is an obscure feeling of longing together with a mysterious emptiness within me that I cannot endure, that consumes me. I, the unhappy one! What do I want? And what will help me’ (1748, 12–13). By the end of the text the narrator reaches a religious insight into the higher vocation of humanity, saying: ‘I have searched for the truth; I have strove for the light, since confusion, doubt, and deception surrounded me everywhere; I have investigated the purpose of my life, the ground of my happiness, and the final aim of my wishes’ (1748, 65–6). For helpful discussions of Spalding’s influence on German theology and philosophy in the late eighteenth century, see Zammito (2002), di Giovanni (2005), Brandt (2007), Munzel (2012), Tippmann (2011), and Printy (2013). See also di Giovanni (2011). 20 The same is true for the so-called ‘fragments controversy’ (Fragmentenstreit) that erupted after Lessing published parts of Hermann Samuel Reimarus’s manuscript under the title ‘Fragments from an Unnamed Author’. In this manuscript Reimarus wanted to use the Gospels’ diverging accounts of the Resurrection to challenge doctrines of revealed religion. Among the defenders of Lutheran Orthodoxy, Johann Heinrich Ress made an important reply with his book Geschichte Jesu Christi.At one point, for example, Ress affirmed: ‘My principle is: there is no objection to experience and matters of fact . . . It is experience that the magnet attracts iron; and all metaphysical and physical reasons do not dispute it. It is a matter of fact that Christ rose again, because we have the gospel—a sure success of this great incident—which it brings almost before our eyes’ (Es ist Tatsache, da Christus auserstanden, denn wir haben Evangelium, ein so sicherer Erfolg dieses groen Vorfalls, da er ihn uns beinahe vor die Augen bringt) (1777, 172). Ress was not the only author of the time to characterize the Resurrection as a Tatsache. See also Lavater (1772/73, esp. 151), and Starck (1779, esp. 864). 21 Mendelssohn (1783, II.6).
50 Facts in Kant ’ s Theoretical Philosophy ‘a kind of evidence that was neither deductive nor inductive but nevertheless valid, and that could perhaps play a role in responding to scepticism’.22 In this respect, too, the original influence of Spalding’s translation is easy to detect. At key moments in the Analogy Butler linked ‘facts’ to the felt quality of our cognitive activities. As he explained, ‘that the three angles of a triangle are equal to two right ones, is an abstract truth: but that they appear so to our mind, is only a matter of fact’.23 Butler also insisted that this is beyond doubt. The way a geometrical proposition appears to the mind ‘must have been admitted’, he wrote, ‘if any thing was, by those ancient sceptics, who would not have admitted the former [i.e., the abstract truth]: but pretended to doubt, Whether there were any such thing as truth, or Whether we could certainly depend upon our faculties of understanding for the knowledge of it in any case.’24
2.4. Facts in Kant’s Theoretical Philosophy It will soon become clear that Kant gives a similar characterization of the fact of reason in the second Critique. But first I would like to offer textual evidence of his familiarity with Tatsache as a technical term. To start with, there is a Reflexion from the late 1770s where Kant distinguishes ‘matters of opinion’ (Sachen der Meynung), ‘matters of belief ’ (Glaubenssachen), and ‘matters of fact’ (Tatsachen) (Refl 2765).25 Clearly this division captures a progression from what is less certain (opinion) to what is more certain (fact), very much in line with the sense of ‘fact’ we have seen the British experimentalists employ. Other texts show that by a Tatsache Kant understands what is actual, given, and incontestable. In an essay from the late 1780s, for example, he writes: ‘how much less can a merely arbitrary surmise be placed against facts!’ (ÜGTP 8:176), and elsewhere: ‘The basic prin ciple, that all knowledge begins solely from experience, involves a quaestio facti, and is thus not at issue here, since the fact [Tatsache] is unreservedly granted’ (FM 20:276). There are also places where Kant uses Factum in the same way.26 In the B edition of the first Critique Kant twice claims that our possession of synthetic a priori cognition is a fact, using Tatsache in the first passage and Factum in the second:
22 Franks (2000, 124). See also Franks (2008, esp. 56, note 11). 23 Butler (1736/1896, 368; emphasis added). 24 Butler (1736/1896, 368). 25 The specific date of this Reflexion is unknown. 26 In another Reflexion, this time from the late 1790s, Kant writes that the idea of participating in a cosmopolitan world must be the wish (Wünsch) of all rational human beings, describing this as a Factum whose reality we can call all persons to witness (‘ein Factum, über dessen Wirklichkeit man alle Menschen zu Zeugen rufen kann’) (Refl 8077).
The Fact of Reason 51 Now it is easy to show that in human cognition there actually are such necessary and in the strictest sense universal, thus pure a priori judgments. If one wants an example from the sciences, one need only look at all the propositions of math ematics; if one would have one from the commonest use of the understanding, the proposition that every alteration must have a cause will do . . . [So] we can content ourselves with having displayed [dargelegt] the pure use of our cognitive faculty as a fact [Tatsache]. (B5) The famous Locke, from neglect of this consideration, and because he encountered pure concepts of the understanding in experience, also derived them from this experience. The empirical derivation, however, [to which Hume resorted as well] cannot be reconciled with the reality of the scientific cognition a priori that we possess, that namely of pure mathematics and general natural science, and is therefore refuted by the fact [Factum]. (B127–8)27
Details aside, Kant’s point in these passages is clear. Synthetic a priori cognition is not something we can doubt, since we actually possess it in pure mathematics and general natural science. By itself, this should shake any confidence we might have in the naturalistic programs of Locke or Hume, who would like to reduce pure concepts to empirical sources. Yet Kant thinks there is a positive claim we can draw from this. If we actually possess synthetic a priori cognition, we can infer that it must have a pure source, because experience does not teach us necessity. After all, no empirical faculty could have generated this cognition within us. Even so, Kant’s positive claim invites us to ask, What proof do we have that we possess synthetic a priori cognition? In the B edition of the first Critique, Kant answers this in a rather striking way. Instead of demonstrating the reality of synthetic a priori cognition by means of a deduction or genetic proof, we find him illustrating it using a method of experiment. Speaking now to the reader, he writes: Gradually remove from your experiential concept of a body everything that is empirical in it—the color, the hardness or softness, the weight, even the impene trability—there still remains the space that was occupied by the body (which has now entirely disappeared), and you cannot leave that out. Likewise, if you remove from your empirical concept of every object, whether corporeal or incorporeal, all those properties of which experience teaches you, you could still not take from it that by means of which you think of it as a substance . . . Thus,
27 I have added boldface to ‘fact’ in each passage and removed Kant’s own boldface from B127–8.
52 The Fact of Reason in the Second Critique convinced by the necessity with which this concept presses itself on you, you must concede that it has its seat in your faculty of cognition a priori. (B6)28
By having us employ a procedure of abstraction, Kant wants us to see that certain cognitions are necessary for experience. We cannot conceive the absence of space, for example, so the representation of space must be necessary for the way we apprehend outer appearances. Granting that such epistemic necessity is not something that could arise from an empirical faculty, we must concede that it came from a pure faculty. In this way, the aim of Kant’s thought experiment is to elicit our actual consciousness of epistemic necessity, so that when we ask, ‘How is this consciousness possible?’ we are led directly to its source, whether in a pure faculty of intuition (as with ‘space’) or in a pure faculty of understanding (as with ‘substance’). In the next section we shall see that Kant also employs an experimental method in the second Critique, one that serves to illustrate our actual consciousness of the moral law. The point of the experiment is to show that we possess a pure or ‘higher’ faculty of desire—a faculty where reason and not inclination determines the will.29
2.5 The Fact of Reason in the Second Critique Let me draw two general observations from the texts gathered so far in this chapter. First, it is clear from both his published and unpublished writings that Kant uses ‘Factum’ and ‘Tatsache’ in the newer, scientific sense (i.e., of something actually the case, and so demanding no proof), in contrast to the older, legal sense (i.e., of something allegedly done, and so capable of proof). Kant foreshadows this in his 1783 Prolegomena to Any Future Metaphysics when he claims to be starting with something ‘already known to be dependable’, namely, the propositions of mathematics and natural science, some of which are ‘apodictically certain’ through reason alone (Prol 4:275). As he says: ‘We have some at least uncontested synthetic cognition a priori, and we do not need to ask whether it is possible (for it is actual)’ (Prol 4:275). 28 Initially we might think Kant is advancing a psychological claim in this passage, saying that we cannot conceive the absence of space, for example, due to our peculiar cognitive makeup. Yet on closer inspection it is clear he has a different sense of ‘necessity’ in mind, one that is properly epi stemic. I am borrowing this terminology from Allison (2004, 104–5). In the first Critique Kant further says that the necessity of space grounds the ‘apodictic certainty’ of geometrical principles. For if we drew our representation of space from experience (i.e., ‘from general outer experience’), geometrical principles would have the same contingency we find in perception. We could not say, for example, that a straight line is necessarily the shortest distance between two points, only that experience always teaches us that (A24/B39). 29 As Kant puts it, only ‘insofar as reason of itself (not in the service of the inclinations) determines the will, is reason a true higher faculty of desire’ (KpV 5:24–5).
The Fact of Reason 53 Second, in view of the historical observations sketched earlier, it is noteworthy that Kant identifies a mode of cognition as a fact. With members of the Royal Society, as we have seen, matters of fact shifted in reference from human deeds to natural phenomena. In Germany, nearly a century later, the concept was shifting yet again, coming to signify mental states accessible to any reflective human being. Interestingly, the passages I have cited above still show Kant’s debt to the experimentalist tradition: he is only willing to identify mental states as Tatsachen if they are actual, immediately certain, and prior to speculation.30 So while the German phrase was increasingly associated with facts of consciousness—accessible only from a first-person perspective—Tatsache retained the elevated status Boyle and others originally conferred on the term.31 Having this framework in view, I think we can bring a new perspective to Kant’s project of moral justification in the second Critique. First, there is a noteworthy continuity between his doctrine of the fact of reason and the argumentstructure we have seen in the B edition of the first Critique and the Prolegomena. Kant draws attention to this continuity when he asks, ‘But how is consciousness of that moral law possible?’ (KpV 5:30). In reply, he writes: We can become aware of pure practical laws just as we are aware of pure theoretical principles, by paying attention to the necessity [Notwendigkeit] with which reason prescribes them to us and to the setting aside of all empirical conditions
30 Kant does not make his debt to this tradition a secret. In the first Critique he claims that it was ‘the ingenious Francis Bacon’ who was partly responsible for putting natural philosophy onto the path of science (B xix). In Kant’s view, the examples of mathematics and natural science ‘were remarkable enough that we might reflect on the essential element in the change in the ways of thinking that has been so advantageous to them, and, at least as an experiment, imitate it insofar as their analogy with metaphysics, as rational cognition, might permit’ (B xv–xvi). Further references to the experimental method run throughout the first Critique (see also B xviii; B xxxviii; A356; A804/B832; A821/B849; A826/B854; A838/B866). Perhaps the most striking occurs in a footnote at B xxi. After characterizing how we should emulate the revolution of the sciences in metaphysics, Kant says: ‘This method, imitated from the method of those who study nature, thus consists in this: to seek the elements of pure reason in that which admits of being confirmed or refuted through an experiment.’ For further discussion, see Vanzo (2012). 31 Here it is worth mentioning that Schopenhauer criticized Kant’s immediate successors (Reinhold specifically) for characterizing the moral law as an ‘immediately certain Factum’ (unmittelbar gewisses Factum) and an ‘original Tatsache of moral consciousness’ (ursprüngliche Tatsache des moralischen Bewussteins) (1841/1988, 45). Schopenhauer believed this contradicted Kant’s explicit warning from the Groundwork that we cannot derive the moral law ‘from some particular property of human nature’ (G 4:425). If we render the moral law a fact of consciousness, Schopenhauer wrote, ‘then it would be grounded anthropologically through experience, though inner experience, and hence empirically, which goes directly against Kant’s view and is repeatedly rejected by him’ (1841/1988, 45). However, I believe Schopenhauer was mistaken to read the concept of a fact in a narrow sense, i.e., as a psychological fact of experience. If we understand the fact of reason in terms of our consciousness of moral necessity (which would be common to all rational beings), then Schopenhauer’s objection loses its force.
54 The Fact of Reason in the Second Critique to which reason directs us. The concept of a pure will springs from the first, as consciousness of a pure understanding springs from the latter. (KpV 5:30)32
In the example discussed earlier (from B6), Kant argued that space is not something we can abstract from a body, making space necessary for our representation of a body. Moreover, because this kind of epistemic necessity is not something we can learn from experience (and so does not spring from an empirical faculty), we are justified to infer it must spring from a pure faculty of intuition. In the passage cited above, Kant wants to extend this line of reasoning to the practical sphere. He wants to show that we discover the concept of a higher faculty of desire in the same way: by ‘paying attention’ to the necessity of moral laws, the way they exclude sensible incentives from entering into our maxims.33 Although Kant does not speak in this way, we can say that moral laws express a unique kind of deliberative necessity, as they bear directly upon our will. Unfortunately, what Kant says after he introduces the fact of reason at KpV 5:31 is quite obscure. ‘Our consciousness of the moral law may be called a fact of reason’, he writes, ‘because one cannot reason it out from antecedent data of reason, for example, from consciousness of freedom (since this is not antecedently given to us) and because it instead forces itself upon us of itself as a synthetic a priori proposition that is not based on any intuition, either pure or empirical’ 32 A qualification is in order. In the theoretical sphere, space and time are pure intuitions ‘that contain a priori the conditions of the possibility of objects as appearances’, and for this reason there is no need to establish their objective validity (A89/B121). By contrast, categories like substance are pure concepts that do not require corresponding sensory data, and that is why Kant is responsible for demonstrating their objective application (i.e., by way of a transcendental deduction). Given this asymmetry, the doctrine of the fact of reason is closer to Kant’s metaphysical expositions of space and time than it is to his transcendental deduction of the categories. Moral laws do not require corresponding empirical motives to bind the will of a rational being, and for this reason they are ‘immediately lawgiving’. Showing that they spring from a higher faculty of desire obviates the need to establish their objective authority. Kant draws attention to this parallelism in a Reflexion from the 1780s: ‘The critique of practical reason has as its basis the differentiation of empirically conditioned practical reason from the pure and yet practical reason and asks whether there is such a thing as the latter. The critique cannot have insight into this possibility a priori because it concerns the relation of a real ground to a consequence, thus something must be given which can arise from it alone; and from reality possibility can be inferred. The moral laws are of this sort, and this must be proven in the same way we proved the representations of space and time as a priori representations, only with the difference that the latter concern intuitions but the former mere concepts of reason’ (Refl 7201). For helpful discussion, see Allison (1990, 234–5). 33 I agree with Guyer (2009) that the fact of reason refers not to our consciousness of moral ‘necessitation’ (Nötigung)—the constraint of the law upon our sensibly affected nature—but to our consciousness of its ‘necessity’ (Notwendigkeit)—the authority of the law as a law of pure practical reason. This is an important point of clarification, for it shows that Kant is abiding by the requirement he lays down in the Groundwork: namely, that the authority of the moral law does not depend on any capacity unique to human nature (G 4:447–8). As Guyer notes, our consciousness of moral necessity is ‘something that would be self-evident for any rational being, not just a human being’, and on these grounds we are warranted to infer that the moral law ‘must have a pure source within us’ (2009, 192). On this point of interpretation, I disagree with Rauscher (2002), who equates the fact of reason with moral necessitation, and with Grenberg (2013) and Schönecker (2013), who limit our access to the moral law to the feeling of respect. I shall return to this issue in §5.2
The Fact of Reason 55 (KpV 5:31). As I hinted at earlier, this remark has led many commentators to conclude that Kant is slipping into a kind of dogmatism in the second Critique, resorting to ‘moralistic bluster’ and ‘foot-stamping’, and it certainly sounds as if he is asserting our moral consciousness as a brute given.34 For what else could it mean in saying that a fact ‘forces itself upon us’, if not that rational argumentation— the kind necessary to justify a contested claim—has eluded us?35 Despite these worries, however, what Kant says before this passage points the way to a more charitable reading of the text, to which I now turn.36 Kant asks us to compare moral laws to the rules we find in pure geometry (KpV 5:31). The latter, he says, ‘contain nothing further than the presupposition that one could do something if it were required that one should do it’. By contrast, moral laws say that ‘one ought absolutely to proceed in a certain way’ (KpV 5:31). (This is the deliberative necessity mentioned above.) As we read further, it is clear that Kant also wants to highlight the status of our moral consciousness itself. We are, he explains, actually conscious that moral laws provide us with immediate determining grounds of choice, and for that reason we do not need to ask whether our moral consciousness is possible. Not surprisingly, right after Kant says it is ‘not impossible’ to think of practical laws that direct the will immediately, he employs the language of fact: ‘Consciousness of this fundamental law may be called a Factum of reason’ (KpV 5:31). So within the logic of this paragraph, we can see that Kant’s Factum carries the meaning of a Tatsache in the scientific sense, i.e., as something real or undisputed or actually the case. The Factum refers to our actual consciousness of the moral law’s deliberative necessity.37 34 The first expression is from Guyer (2007, 462); the second from Wood (2007, 135). 35 To make matters worse, Kant seems to allow for a degree of arbitrariness here in his allusion to the tyrannical wife from Juvenal’s Satire 6, writing that the moral law ‘announces itself as originally lawgiving (sic volo, sic jubeo)’ (KpV 5:31). In the context of Satire 6, the wife orders the execution of a slave, and the perplexed husband asks her why, only to receive the reply: ‘sic volo, sic jubeo, sit pro ratione voluntas’ (What I will, I command; my will is reason enough) (1992, 22; translation modified). For further discussion of this passage, see Proops (2003), Ware (2010), and Kain (2010). 36 Readers have also been troubled by Kant’s remark that the moral law ‘cannot be proved by any deduction’ although it is ‘firmly established of itself ’ (KpV 5:46; cf., MS 6:225 where Kant says moral laws are ‘incapable of being proved and yet apodictic’). In Hegel’s memorable phrase, the moral law seems to be the ‘final undigested lump left within the stomach, the revelation given to reason’ (quoted in Allison 1990, 281). However, it is important to keep in mind what Kant means by a deduction. As he explains in the subsequent paragraph, concepts of pure understanding or categories admit of a deduction (i.e., a justification of their objective and universal validity), because they refer to objects of possible experience. We can in turn show that they are cognizable only by virtue of these very categor ies. No such deduction of the moral law is forthcoming, since no exact observation of it can be found in experience (KpV 5:46–7). 37 On this point my reading is close to Kleingeld’s, who also reads ‘Factum’ as a kind of ‘Tatsache’, although there are differences worth mentioning. In Kleingeld’s view, ‘the entire argument [of the fact of reason] can be cast in (presumably ‘non-moral’) terms of a theory of action and be regarded as the articulation of the self-understanding of agents who take themselves to be reasoning about which maxims to adopt and why’ (2010, 70). She first supports this reading with Kant’s remark that we need only analyze ‘the judgments people pass on the lawfulness of their actions’ to confirm the fact of reason (KpV 5:32). As I see things, however, it is debatable whether Kant meant ‘lawfulness’ here in nonmoral terms. Kleingeld then maintains that the fact of reason refers to our consciousness of ‘the
56 The Fact of Reason in the Second Critique Now, however, we must ask the same question as before: What proof do we have that our consciousness of the moral law is actual? As we saw in the theoretical sphere, Kant does not think we need to argue for the reality of synthetic a priori cognition; we need only illustrate it with examples from mathematics and natural science. And Kant wants us to use this same method in the practical sphere. We need only illustrate the reality of moral consciousness, and we can do this, he thinks, by turning to examples of common moral judgment. Thus, after introducing the fact of reason at KpV 5:30, Kant writes: ‘The Factum mentioned above is undeniable’, adding: ‘One need only analyze the judgments people [Menschen] pass on the lawfulness of their actions in order to find that, whatever inclination may say to the contrary, their reason, incorruptible and self-constrained, always holds the maxim of the will in an action up to the pure will, that is, to itself inasmuch as it regards itself as a priori practical’ (KpV 5:32). Later in the second Critique Kant explains why the judgments of common reason have a justificatory role to play in this context (KpV 5:91). He observes that in the theoretical sphere it was ‘easily and evidently proved’ that we possess a pure faculty of cognition, since we have ‘examples from the sciences’ ready at hand. Such sciences ‘put their principles to the test in so many ways by methodic use’, so we need not fear what Kant calls ‘a secret mixture of empirical grounds’ underlying them (KpV 5:91). Turning now to the practical sphere, he writes: But that pure reason, without the admixture of any empirical determining ground, is practical of itself alone: this one had to be able to show from the most common practical use of reason, by confirming the supreme practical principle as one that every natural human reason cognizes—a law completely a priori and independent of any sensible data—as the supreme law of its will. It was necessary first to establish and justify the purity of its origin even in the judgment of this common reason before science would take it in hand in order to make use of it, so to speak, as a fact [Factum] that precedes all subtle reasoning about its possibility and all the consequences that may be drawn from it. (KpV 5:91)38
fundamental law of pure practical reason’, a law that is ‘subsequently called the moral law’ (2010, 66; emphasis added). However, in section 6—before Kant introduces the ‘fundamental law’—he writes: ‘It is therefore the moral law . . . that first offers itself to us and . . . leads directly to the concept of freedom’ (KpV 5:29). Given these discrepancies, I share Ameriks’s doubts (2012) regarding the prospect of reconstructing a ‘non-moral route’ in the second Critique. 38 Citing this passage, Rawls also highlights the importance of common moral judgment for interpreting Kant’s doctrine of the fact of reason. On this point I am indebted to his approach. However, I am not committed to Rawls’s further claim that by the time of the second Critique Kant had developed ‘a constructivist conception of practical reason’ (2000, 268). As Kain points out (2006), there are difficulties facing constructivist readings of Kant’s ethics. Setting these difficulties aside, I am also not sure how Rawls thinks the fact of reason justifies (or ‘authenticates’, as he prefers to say) the moral law. As I understand the structure of the first chapter, Kant begins with a conceptual argument for why there can only be one fundamental law of pure practical reason, the ‘moral law’ (KpV 5:30). Then, after this
The Fact of Reason 57 If we now read this passage in light of KpV 5:30, a clear picture of what motivates Kant’s doctrine of the fact of reason takes shape. A law that commands ‘completely a priori and independent of any sensible data’ is logically coherent. Yet that does not give us the conclusion Kant is seeking: that it originates from a pure (as opposed to an empirical) faculty of desire. We are first required to show that we actually are conscious of this law as a ground of choice before we can infer the condition of its possibility. And that is why the judgments of common reason are indispensable for Kant’s project of moral justification. Just as examples from mathematics and science confirm the reality of synthetic a priori cognition, judgments from ordinary people confirm the reality of moral consciousness. Both provide us with grounds to infer the existence of a pure faculty within us, whether of cognition or of desire.
2.6 The Experimental Method At this point it may be tempting to read Kant’s claim from KpV 5:32—that ‘one need only analyze the judgments people pass on the lawfulness of their maxims’— as an appeal to empirical evidence.39 Now it is true that (most) human beings engage in moral behavior, if only by judging their actions or the actions of others ‘good’ or ‘bad’. But I believe it would be a mistake to read Kant’s strategy in this way. Kant does not want us to observe the standpoint of common reason from a theoretical, third-personal perspective. Rather, he wants us to take up this standpoint ourselves, so that we can illustrate our consciousness of the moral law from a practical, first-personal perspective. Kant calls attention to this point of orientation later in the second Critique (KpV 5:92). In a gesture at once odd and intriguing, he claims that a philosopher has an advantage ‘like a chemist’ in that he can set up experiments with every ordinary person. As he puts it, the philosopher can ‘distinguish the moral (pure) determining ground from the empirical, namely, by adding the moral law (as a determining ground) to the empirically affected will (e.g., that of someone who would gladly lie because he can gain something by it)’ (KpV 5:92).40 To make the conceptual argument, Kant claims that our consciousness of this law is something actual—a fact– which in turn warrants the assumption that our faculty of desire is also pure (and not merely affected by sensible incentives) (KpV 5:31). 39 Moyar (2008, 333) and Grenberg (2013,151) have (correctly, I believe) found Proops guilty of this reading. In Proops’s view, ‘what would reveal the non-empirical origin of the idea of duty would be its constancy across persons whose quality and level of moral education differed widely’ (2003, 226). 40 Returning now to the obscure passage from KpV 5:31, we can see that, far from resorting to a dogmatic position, Kant is merely stating that our consciousness of the moral law is not accessible outside a normatively thick practical perspective. For obvious reasons, it is not accessible from a speculative point of view, because speculative reason only teaches us the law of causal mechanism, the antithesis of freedom. For less obvious reasons, Kant’s point is that our consciousness of the moral law
58 The Experimental Method metaphor vivid, Kant compares the philosopher’s experiment to a process of adding alkali to hydrochloric acid: When an analyst adds alkali to a solution of calcareous earth in hydrochloric acid, the acid at once releases the lime and unites with the alkali, and the lime is precipitated. In just the same way, if a man who is otherwise honest (or who just this once puts himself only in thought in the place of an honest man) is confronted with the moral law in which he cognizes the worthlessness of a liar, his practical reason (in its judgment of what he ought to do) at once abandons the advantage, unites with what maintains in him respect for his own person (truthfulness), and the advantage, after it has been separated and washed from every particle of reason (which is altogether on the side of duty), is weighed by everyone, so that it can enter into combination with reason in other cases. (KpV 5:92–3)41
While they are difficult to see at first, thought experiments of this sort run throughout the second Critique.42 One that deserves special attention occurs at KpV 5:30, only paragraphs before Kant first speaks of the fact of reason. In this section of the text Kant has argued that we only become aware of our freedom through the moral law, and to prove this he introduces a character who reports to have a lustful inclination he cannot control (KpV 5:30). In what follows Kant arranges for us a narrative with two scenes. In the first scene, the lustful man is asked what he would do if, upon satisfying his inclination, he were immediately strung up on a gallows: Suppose someone asserts of his lustful inclination that, when the desired object and the opportunity are present, it is quite irresistible to him; ask him whether, if is not accessible from a thin practical perspective. For when we examine the capacities anyone must possess to function as a rational agent (e.g., a capacity to step back and reflect on one’s desires), there is nothing internal to these capacities that would bind one to a law other than one’s own happiness. That is why, as Kant says elsewhere, the ‘most rational being of the world’ might be oblivious to the moral law and so ‘might still need some incentives, coming to him from the objects of the inclinations, in order to determine his power of choice’ (R 6:27n). In light of these considerations, I find it difficult to associate Kant’s ethics with recent foundationalist arguments for moral normativity, a point I will return to in the Conclusion. 41 See also Rohden (2012, esp. 103), Keller (2010, esp. 124), and Timmermann (2010, esp. 88, note 29). 42 For example, at KpV 5:25, Kant writes: ‘Suppose that an acquaintance whom you otherwise liked tried to justify to you his having given false testimony by first pleading what he asserts to be the sacred duty of his own happiness . . . Or suppose that someone recommends to you as steward a man to whom you could blindly trust all your affairs and, in order to inspire you with confidence, extols him as a prudent human being with masterly understanding of his own advantage.’ In the first case, Kant says, ‘you would either laugh in his face or shrink back from him with disgust’ (KpV 5:26). In the second case, ‘you would believe either that the recommender was making a fool of you or that he had lost his mind’ (KpV 5:26). Thus, Kant concludes: ‘So distinctly and sharply drawn are the boundaries of morality and self-love that even the most common eye cannot fail to distinguish whether something belongs to the one or the other’ (KpV 5:26). For similar examples, see KpV 5:37 and 5:88–9.
The Fact of Reason 59 a gallows were erected in front of the house where he finds this opportunity and he would be hanged on it immediately after gratifying his lust, he would not then control his inclination. One need not conjecture very long what he would reply. (KpV 5:30)43
In the second scene the man is asked what he would do if, threatened by the same execution, a prince demanded him to bear false witness against an innocent person: But ask him whether, if his prince demanded, on pain of the same immediate execution, that he give false testimony against an honorable man whom the prince would like to destroy under a plausible pretext, he would consider it possible to overcome his love of life, however great it may be. He would perhaps not venture to assert whether he would do it or not, but he must admit without hesitation that it would be possible for him. (KpV 5:30)
In the second scene we are considering what the man would do if he took up a moral point of view. We are—to speak in terms of the chemistry metaphor—adding a pure determining ground (‘the alkali’) to the man’s empirically affected will (‘a solution of calcareous earth in hydrochloric acid’) in order to see what must result. Prior to this determining ground it is likely the man would lie, since he would have something very important to gain from it: his own life. But now, if he reflects on his duty, we see that he must reach an entirely different verdict: namely, that he would be worthless as a liar, and that refusing the prince is the only choice that would maintain his self-respect. He would thus abandon whatever advantage he sees in the lie (as the acid ‘at once releases the lime’) and acknowledge the authority of the moral law (as the acid ‘unites with the alkali’).44 In working through the second scene, we stand to witness the man’s transfigur ation, whereby he comes to deliberate without giving priority to his own happiness, and so without letting empirical motives enter into his maxims. In judging what he ought to do, the man is conscious of a law ‘completely a priori and independent of any sensible data’, and he is conscious of it, moreover, as an immediate 43 We must still assume the lustful man is sensitive to considerations of prudence. It is not difficult for us to imagine characters who, in the grips of passion, are ‘in no mood to listen to reason’—as de Sade describes the villains who tell Justine that ‘even if there had been a gallows on the spot, [she] shall still be their prey’ (1791/1999, 27). In Kant’s example, we need not conjecture very long what the lustful man will say, assuming that he is (unlike de Sade’s villains) listening to reason. 44 The experiment would still work even if the man decided he would give in to the prince’s demand, for he would not thereby claim his maxim should become a universal law. As Kant makes clear in the Groundwork, in any transgression of duty ‘we find that we do not really will that our maxim should become a universal law, since that is impossible for us, but that the opposite of our maxim should instead remain a universal law [e.g., refusing the prince], only we take the liberty of making an exception to it for ourselves (or just for this once) to the advantage of our inclination’ (G 4:424).
60 Spectator and Experiment ground of choice. In my understanding, the aim of Kant’s thought experiment is to elicit this fact from the reader, so that he or she may see how people separate duty from self-interest as if the two were unmixable chemical compounds. In this respect, when we work through the steps of the experiment, we take up a delib erative perspective available to all rational beings.45 And that is why Kant thinks we have grounds to treat moral consciousness ‘as a fact that precedes all subtle reasoning about its possibility’ (KpV 5:91), deciding that it must spring from a pure faculty within us. Beyond this, we do not need a long, complex strategy of proof. The Tatsache of moral consciousness shows that reason can be practical of itself, and ‘all subtle reasoning against the possibility of its being practical is futile’ (KpV 5:3).
2.7 Spectator and Experiment Before moving forward, there is an apparent inconsistency I should address. Why does Kant treat the reader as a spectator in the experiment from KpV 5:30? Does this mean he is not trying to elicit our consciousness of the moral law from a practical, first-personal perspective? We can see this perspective in the first Critique when Kant asked us to remove from a body ‘everything that is empirical in it—the color, the hardness or softness, the weight, even the impenetrability’, as a way of illustrating the necessity of space, which experience cannot teach us. My own view is that Kant is using this same strategy in the second Critique; the only difference is that he wants us to apply it to another person’s perspective. When we imagine a man faced with threats of execution, for example, we cannot detect any empirical motive that would secretly press upon him to refuse the prince. So the moment he judges that truthfulness is the only act that ‘maintains in him respect for his own person’, we must conclude that he is deferring authority to a principle other than his own happiness. In this way the experiment has a participatory element, despite the fact that it does not bring us (the readers) into the narrative scene. This participatory element is central to Part II of the second Critique, titled the Doctrine of the Method of Pure Practical Reason.46 Here Kant argues that the moral law must have a more powerful effect on us the more it is isolated from incentives that would link it (if only covertly) to self-interest. To establish this 45 For this reason my account avoids an objection Moyar has brought against first-personal readings of the second Critique. In Moyar’s view, these readings establish the reality of freedom only ‘for the reader who successfully takes up and is moved by the examples considered in the text’. Yet as he points out: ‘Kant surely meant for the deduction to establish the reality of freedom from the practical point of view, not just from my practical point of view’ (2008, 334). On my interpretation, Kant’s thought experiments serve to illustrate a fact for all rational agents, not just for human agents, and certainly not just for readers of the second Critique. 46 For a detailed discussion, see Bacin (2010).
The Fact of Reason 61 controversial thesis, he asks us to imagine telling the following story to a youth around ten years of age. The story centers on an honest man ‘whom someone wants to induce to join the calumniators of an innocent but otherwise powerless person (say, Anne Boleyn, accused by Henry VIII of England)’ (KpV 5:156). This man, Kant explains, is offered gain, that is, great gifts or high rank; he rejects them. This will produce mere approval and applause in the listener’s soul, because it is gain. Now threats of loss begin. Among these calumniators are his best friends, who now refuse him their friendship; close relatives, who threaten to disinherit him (he is not wealthy); powerful people, who can pursue and hurt him in all places and circumstances; a prince who threatens him with loss of freedom and even of life itself. But, so that the measure of suffering may be full and he may also feel the pain that only a morally good heart can feel very deeply, represent his family . . . as imploring him to yield and himself, though upright, yet with a heart not hard or insensible either to compassion or to his own distress; represent him at a moment when he wishes that he had never lived to see the day that exposed him to such unutterable pain and yet remains firm in his resolution to be truthful, without wavering or even doubting. (KpV 5:156)
At this point, Kant concludes, my young listener will be raised step by step from mere approval to admiration, from that to amazement, and finally to the greatest veneration and a lively wish [Wunsche] that he himself could be such a man (though certainly not in such circumstances); and yet virtue is here worth so much only because it costs so much, not because it brings any profit. (KpV 5:156; cf., G 4:411n and 4:454)
From what Kant portrays here, it is clear that we can engage with a story like this on both cognitive and affective levels. As the story progresses, the young listener is led to cultivate his judgment, eventually distinguishing in the honest man’s resolution nothing but pure duty, separated from any motive that would implore him, like his family, to yield to the evil calumniators. At the same time, the boy is led to cultivate his sensibility, first by admiring the honest man’s courage, and finally by expressing the deepest respect for his actions.47 As these cognitive and affective responses reach their peak, the dialectic between spectator and experiment turns inward. For in the final stage we have the boy expressing ‘a lively wish that he himself could be such a man’ (emphasis added), moved by the awareness of his own capacity to act from pure motives. All of this suggests that Kant’s
47 See also KpV 5:158–9 for the example of Phalaris’s bull.
62 Matters of Fact in the Third Critique method for eliciting our actual consciousness of the moral law is not only of philosophical value, i.e., for showing that pure reason is practical. When it enters the mind of a pupil by way of examples, the philosopher’s experiment is also a method of moral education.48 So far I have argued that our consciousness of the moral law is a fact because it is something actual, and it is a fact of reason because—upon reflection—we see that it originates from a higher faculty of desire. On the interpretation I have put forward, Kant is not treating our moral consciousness dogmatically, i.e., as a brute given. Rather, he is treating it as something actual, and what is more, he is setting up thought experiments that generate this fact within us. These points become clear as soon as we read Kant’s doctrine through the notion of a Tatsache that had acquired a meaning so flexible in Germany—thanks to Spalding—that it was adopted primarily by theologians during the 1770s and 1780s. There are, however, a few objections that one could raise against my account in this chapter. One could argue that my interpretation is inconsistent with Kant’s usage of ‘matters of fact’ from the Critique of the Power of Judgment (1790), and that it lies in tension with a claim, defended by Henrich and Proops, that Kant’s Factum is best understood under a legal analogy. More seriously, one could argue that my interpret ation goes no further than the subjective experience of moral obligation, and that it fails to vindicate Kant’s position against someone who denies that consciousness of the moral law is a genuine fact.
2.8 Matters of Fact in the Third Critique First, the only place where Kant speaks of a Tatsache of reason—in section 91 of the third Critique—he does not refer to our consciousness of moral necessity. He defines matters of fact as objects of concepts the ‘reality of which can be proved’, adding parenthetically: ‘whether through pure reason or through experience, and whether in the first case through theoretical or practical data of reason’ (KU 5:648).
48 Similarly, Kant writes in the Metaphysics of Morals that when the attention of a pupil ‘is drawn to the fact that none of the pains, hardships, and sufferings of life—not even the threat of death—which may befall him because he faithfully attends to his duty can rob him of consciousness of being their master and superior to them all, then the question is very close to him: what is it in you that can be trusted to enter into combat with all the forces of nature within you and around you and to conquer them if they come into conflict with your moral principles? Although the solution to this question is completely beyond the capability of speculative reason, the question arises of itself; and if he takes it to heart, the very incomprehensibility in this cognition of himself must produce an exaltation in his soul which only inspires it the more to hold its duty sacred, the more it is assailed’ (MS 6:483). Most commentators overlook the connections between Kant’s doctrine of the fact of reason and his theory of moral education. For exceptions to this trend, see Budick (2010, esp. Ch. 5), Guyer (2012, esp. 134, 137), and Grenberg (2013, esp. Ch. 9). See also Louden (2000) for a masterful study of Kant’s theory of moral education.
The Fact of Reason 63 What is surprising, given our previous discussion, is that Kant goes on to list ‘freedom’ as the only idea we can list among the facts: But what is quite remarkable, there is even one idea of reason (which is in itself incapable of any presentation in intuition, thus incapable of theoretical proof of its possibility) among matters of fact, and that is the idea of freedom. The reality [of this idea] as a particular kind of causality—the concept of which would be excessive from a theoretical point of view—can be established through practical laws of pure reason, and, in accordance with these, in actual deeds, and thus in experience.—It is the only one among all the ideas of pure reason whose object is a matter of fact [Tatsache]. (KU 5:648)
If we take this passage to represent Kant’s official position, it appears we only have textual grounds to speak of a ‘fact of freedom’, not a ‘fact of moral consciousness’, as I have proposed. There is, however, a larger interpretive difficulty at stake here, for even within the second Critique we find Kant offering conflicting character izations of the fact of reason. While it is usually clear that Kant’s Factum refers to our consciousness of the moral law, at one point he speaks of it in terms of our consciousness of freedom, writing: ‘this fact is inseparably connected with, and indeed identical with, consciousness of freedom of the will’ (KpV 5:42; emphasis added). In working toward a solution, all we must do is coordinate two of Kant’s wellknown claims in the second Critique: the reciprocity thesis and the disclosure thesis.49 According to the former, freedom and morality are analytically linked, because when we examine one concept we are led directly to the other, and vice versa (KpV 5:29). According to the latter, it is only through our consciousness of the moral law—as a law that commands without empirical motives—that we acquire a positive idea of our freedom. As Kant puts it, ‘had not the moral law already been distinctly thought in our reason, we should never consider ourselves justified in assuming such a thing as freedom (even though it is not self-contradictory)’ (KpV 5:5n), and elsewhere: ‘this law is the only law that makes us conscious of the independence of our power of choice from determination by all other incentives (of our freedom)’ (R 6:26n). With these distinctions in place, we can return to section 91 with a new perspective. On a second look it is clear that Kant has the disclosure thesis in mind when he says freedom ‘can be established through practical laws of pure reason’ (KU 5:648). We can then qualify Kant’s statement from the third Critique in the following way: Strictly speaking, morality and freedom are one and the same from the viewpoint of the reciprocity thesis, but different from the viewpoint of
49 ‘Reciprocity thesis’ is Allison’s phrase (1990); ‘disclosure thesis’ is my own.
64 Henrich and Proops on the Legal Analogy the disclosure thesis. That is to say, the Factum/Tatsache of moral consciousness has epistemic priority over the Factum/Tatsache of freedom, because it is only through the moral law’s necessity (as a law that commands ‘completely a priori’) that we first become aware of the independence of our power of choice. Qualified in this way, it is consistent with Kant’s overall position to locate moral consciousness among the facts, as I have done.50
2.9 Henrich and Proops on the Legal Analogy In a highly influential essay, Dieter Henrich has argued that the ‘entire first Critique, and the way in which Kant presents its theory as a whole, was thoroughly affected by the decision to adopt juridical procedures as a methodological paradigm’.51 More recently, Ian Proops has extended Henrich’s claim to argue that the legal analogy, more so than anything else, penetrates the obscurity surrounding Kant’s Factum. On the basis of published and unpublished texts, Proops builds what appears to be a strong case for reading Kant’s Factum in terms of the quid facti/quid juris distinction criminal lawyers used during the eighteenth century. What emerges from this distinction, he argues, is the meaning of a factum as something that remains ‘relevant to, but falls short of, a final ruling’.52 To support his reading, Proops cites various places in the second Critique where Kant appeals to the judgments of common human reason, adding: ‘One doubts that these “proofs” of the purity of origin of the moral law can be conclusive, but it is plain that Kant believes they ought to have some probative force—in the way, perhaps, that legal evidence has defeasible, non-demonstrative force.’53 It is no doubt true that Kant was deeply influenced by the juridical procedures of his day, and I think it is useful to emphasize this as Henrich and Proops have. But at the level of details, Proops’s attempt to read such procedures into the second Critique runs up against major difficulties. For instance, if we interpret Kant’s Factum solely in terms of an analogy to eighteenth-century criminal law, we are forced to demote the status of moral consciousness to something allegedly the case, so as to make Kant’s Factum fit the idea of legal evidence bearing ‘defeasible, non-demonstrative force’. To do so, however, is completely at odds with the text and spirit of the second Critique, where Kant repeatedly affirms that the moral 50 In a footnote Kant also explains that he is going beyond the usual meaning of a Tatsache. As he writes: ‘Here I extend the concept of a matter of fact, as seems to me right, beyond the usual meaning of this word. For when the issue is the relation of things to our cognitive capacities it is not necessary, indeed not even feasible, to restrict this expression merely to actual experience’ (KU 5:468n). 51 Henrich (1960/1994, 38). 52 Proops (2003, 215). 53 Proops (2003, 227). He continues: ‘What really matters, for present purposes, is the very fact that Kant envisages such proofs at all. That he does so supports a view of the fact of reason as the factum of the Deduction of Freedom; for as we have seen, a factum is a fact that needs to be proved in the course of a deduction’ (2003, 227; emphasis added).
The Fact of Reason 65 law is ‘apodictically certain’ and ‘firmly established of itself ’ (KpV 5:47). By contrast, the notion of a scientific fact that emerged with the British experimentalists puts this epistemic ultimacy in its proper light. For writers like Boyle, matters of fact were promoted from the category of things allegedly done, and so capable of proof, to that of things actually the case, and so demanding no proof. On the whole, this makes the scientific analogy I have uncovered a better tool than the legal analogy for interpreting Kant’s Factum. The scientific analogy also illuminates Kant’s strategy of justification within his critical philosophy at large.54 When Kant claims that our consciousness of the moral law requires no proof, we need only recall that a year earlier he had used this language in his theoretical philosophy, stating that ‘we can content ourselves with having displayed the pure use of our cognitive faculty as a fact’ (B5). As I have shown, there is a continuity in the argument-structure of these texts. By attending to the epistemic necessity in the concept of space, for example, we can see that it must spring from a pure (rather than an empirically conditioned) faculty of intuition. Likewise, by attending to the deliberative necessity in the moral law, we can see that it must spring from a pure (rather than an empirically conditioned) faculty of desire. Finally, in the same way that examples from mathematics and natural science illustrate the reality of synthetic a priori cognition, judgments from common human reason illustrate the reality of moral consciousness— elevating both to facts that neither admit of nor require further proof.
2.10 Beck’s Dilemma Even if the textual support I have offered in this chapter is compelling, worries about the general plausibility of Kant’s doctrine may still surface. In particular, one might worry that Kant’s doctrine offers us nothing more than a subjective analysis, i.e., an account of what moral obligation is like from the perspective of the deliberating agent. The objection is that there is only so much an analysis of this sort can reveal. Granted it can reveal what we experience, that the law we face appears overriding and inescapable, but it cannot reveal the object of our experience, namely, that it is Kant’s version of the moral law, the principle of autonomy. However detailed or insightful, an account of what our experience of moral obligation is like does not rule out the possibility that we are subject to a non-Kantian law (e.g., a principle of perfection, happiness, or God’s will). A deeper worry one 54 For instance, much of what I am saying about Kant’s strategy of justification in the second Critique resonates with Ameriks’s reading of the transcendental deduction in the first Critique, according to which Kant argues regressively from experience (what is given) to the categories (its conditions of possibility). See his (2000). However, as sympathetic as Ameriks is to Kant’s regressive strategy in the first Critique, he has been a vocal critic of Kant’s doctrine of the fact of reason in the second Critique, despite the fact that this doctrine is a clear illustration of a modest approach to justification.
66 Beck ’ s Dilemma might have is that Kant is making an illicit slide from the fact of reason, as the subjective side of moral experience, to the principle of autonomy, as the objective side of moral experience. This worry is only compounded by the variety of ways Kant speaks of the Factum in the second Critique. On examination the term appears to have at least six different characterizations in the text:
(a) our consciousness of the moral law (KpV 5:31); (b) the moral law itself (KpV 5:31, 5:47, 5:91); (c) consciousness of freedom of the will (KpV 5:42); (d) autonomy in the principle of morality (KpV 45:2); (e) an inevitable determination of the will by the moral law itself (KpV 5:55); and (f) actual cases of actions presupposing unconditional causality (KpV 5:105).55 In the previous section I offered a reading of (c) that makes it consistent with Kant’s claim that freedom is the essence of the moral law, its ‘ratio essendi’, so we need not discuss it further here. This leaves us with five characterizations to consider more closely. Following the work of Lewis White Beck, most scholars acknowledge that these five characterizations fall into two general classes: a subjective class identifies the Factum with our consciousness of the moral law, and an objective class identifies it with the moral law itself.56 A problem Beck drew attention to years ago is that, as it stands, these two classes generate a dilemma. On the one hand, if we align the fact of reason with the subjective class, then it may be true that we have an experience of moral obligation (even a skeptic will grant us this), but that does not establish the reality of Kant’s moral law, the principle of autonomy. On the other hand, if we align the fact of reason with the objective class, then the reality of Kant’s moral law becomes the key point of dispute and it would be nothing short of dogmatic to assert its reality without further argument. As Allison puts it: ‘If the fact is construed subjectively as a mode of consciousness, its existence is readily granted but no inference to the validity of the law is warranted thereby. Conversely, if it is construed objectively and equated with the law itself, then the existence of this fact becomes the very point at issue and can hardly be appealed to in order to ground the reality of moral obligation.’57 To make matters worse, Kant appears oblivious to these contrasting senses of the term, and at one point he speaks confidently of ‘the sole fact of pure reason’ (KpV 5:32; emphasis added), as if he only had one referent in mind. 55 And elsewhere Kant identifies the fact of reason with the categorical imperative (MS 6: 252 and OP 21: 21). 56 See Beck (1960), as well as Allison (1990), Rawls (2000), and Proops (2003). 57 Allison (1990, 232).
The Fact of Reason 67 One might worry that since I have defined Kant’s Factum as a ‘fact of c onsciousness’, my interpretation is at risk of getting caught in the first horn of Beck’s dilemma. In response to this worry, I should say first that when Kant speaks of our consciousness of the moral law he is not referring loosely to any experience of constraint, obligation, or necessity. We experience constraint whenever we apprehend the force of a practical imperative, whether categorical or hypothetical, since the imperative conveys what we ‘ought’ to do. Our consciousness of necessity in such cases belongs to normative experience in general, but not to moral experience in particular. As we know, what makes the fact of reason unique is that it speaks to our consciousness of ‘absolute’ necessity—and this qualification is important, in my view, because it suggests that only Kant’s formula of the moral law is adequate to explain this peculiar necessity. This may be why Kant appears to shift in the second Critique between subjective and objective characterizations of the Factum. Technically speaking, there is only one Factum: our consciousness of absolute moral necessity, and that is the subjective side of moral experience. Yet this consciousness can only refer to the principle of autonomy, as the objective side of moral experience, since no other version of the moral law (in terms of perfection, happiness, or God’s will) can render this necessity intelligible. As Kant argues in a number of places, all previous formulations of the moral law exhibit heteronomy, meaning that they connect the law to our will through an intermediary interest. Moreover, whatever binds our will through an intermediary interest motivates us on the basis of our sensible constitution. As a result, if we were subject to a non-Kantian principle our consciousness of it would not be accompanied by absolute necessity. What is distinctive about moral experience, for Kant, is that we are aware of a command that excludes all, and not just some, of our sensible interests. And that is why Kant thinks there is only one supreme moral principle, the principle of autonomy. To be clear, Kant is not saying our awareness of this abstract formula is a fact of consciousness. The formula merely explicates from a philosophical perspective a law we take to be overriding and inescapable from an ordinary, pre-philosophical perspective. If this is correct, then we no longer have grounds to worry that our subjective moral experience leaves open the possibility that we are subject to a non-Kantian law. On the contrary, when we pay due attention to the nature of the Factum, we can see that Kant’s inference to the principle of autonomy is well-founded, and that shows why the interpretation I have offered does not get caught in Beck’s dilemma.
2.11 The Omnipotence of Theoretical Reason The last objection I would like to consider in this chapter is related to the previous one, but it raises a new challenge. As I noted earlier, many commentators in Kant’s day and our own have found the second Critique disappointing because its
68 The Omnipotence of Theoretical Reason argument begins where they think it should end: with the moral law (or our consciousness thereof). If this alleged fact provides the only basis for inferring that pure reason is practical, it appears the scope of Kant’s position is severely limited. It would not persuade a skeptic, for example, who is already convinced that reason is conditioned by empirical motives all the way down. Here it seems all the skeptic has to do in order to refute the argument of the second Critique is deny what Kant treats as a fact, i.e., our consciousness of absolute moral necessity. A skeptic could reply that, when reflecting upon the contents of his consciousness, he is only aware of hypothetical or prudential imperatives—nothing that would commit him to affirm the reality of a higher faculty of desire. In light of our previous discussion, one could agree that if we define the Factum as our consciousness of the moral law’s necessity, Kant’s inference from the fact of reason to the prin ciple of autonomy is well-grounded. But one could simply deny possessing anything like consciousness of the moral law’s necessity to begin with, and thereby deny Kant’s way of illustrating that pure reason is practical for us.58 Now Kant is ready to admit that our consciousness of the moral law is ‘strange’ (befremdlich), for there is nothing like it in our entire field of practical consciousness. When we attend to the necessity of the moral law, we are led to see that it wholly excludes empirical motives from entering into our maxims. No other principle—hypothetical or prudential—has this unconditional character. For Kant, however, skepticism about the reality of moral consciousness can only arise from a certain standpoint: the standpoint of abstract speculation. Indeed, Kant makes this clear in the second Critique when he says that only philosophers can make the question of pure morality ‘doubtful’ (zweifelhaft) (KpV 5:155). In another work he tells us why: Only those ‘who are accustomed merely to explan ations by natural sciences will not get into their heads the categorical imperative from which these laws proceed dictatorially, even though they feel themselves compelled irresistibly by it’ (MS 6:378). In this way skepticism about moral consciousness can only arise for those who seek to defend the ‘omnipotence of theor etical reason’ (MS 6:378). Although our cognition of the moral law is ‘distinct’ and ‘irrepressible’, such philosophers will nevertheless try to ‘shut their ears’ to it for the sake of protecting their speculative interests (KpV 5:35). Kant distills the essence of this point in his 1793 essay ‘On the Common Saying: That May Be Correct in Theory, but it is of No Use in Practice’. Here we find him responding to Christian Garve’s confession—from his 1792 Essays—that he readily grasps the separation between duty and happiness in his head, but loses 58 J. B. Schneewind raises this objection in the context of justifying the title of his book, The Invention of Autonomy. He writes: ‘Autonomy, as Kant saw it, requires contracausal freedom; and he believed that in the unique experience of the moral ought we are “given” a “fact of reason” that unquestionably shows us that we possess such freedom as members of a noumenal realm. Readers who hold, as I do, that our experience of the moral ought shows us no such thing will think his version of autonomy as an invention rather than an explanation’ (1998, 3).
The Fact of Reason 69 it entirely in the strivings of his heart. In reply Kant writes with a note of humor that he feels no hesitation in contradicting Garve and ‘championing his heart against his head’ (TP 8:285). As a man of integrity, Kant adds, Garve has ‘actually found this separation in his heart every time (in his determination of will), only it would not be reconciled in his head’ (TP 8:285). In other words, Garve is unable to reconcile the ‘possibility of categorical imperatives (such as those of duty are)’ with what Kant calls the principles of psychological explanation, ‘all of which have the mechanism of natural necessity as their basis’ (TP 8:285). It is only from the standpoint of theory, then, that Garve can find obscure what he knows clearly and distinctly from the standpoint of practice. This is the critical part of the reply I imagine Kant would have us give to the skeptic. We must show that the skeptic’s demand to verify the fact of reason with further proof is unfounded. It rests on the mistaken conviction that all facts, including facts of consciousness, must conform to theoretical standards of explanation (‘all of which have the mechanism of natural necessity as their basis’). From what Kant says, however, there is also a constructive reply we can bring forth, and this may be the deeper lesson of his doctrine of the fact of reason. If we are to cease defending the omnipotence of theoretical reason, we must learn to philosophize in a different way, and I take it this is what Kant would have us do by adopting a method ‘similar to that of chemistry’ (KpV 5:163). The way people immediately separate duty from happiness, and acknowledge the former’s authority, is evident by experiment, even if not known by a genetic proof. However strange our moral consciousness may be from the standpoint of speculation, it is a fact we can grasp from a practical point of view. For Kant, then, it is only when we guide ourselves by this fact, and do so as philosophers, that we can finally reconcile in our heads what we already know in our hearts.
2.12 Closing Remarks In this chapter I have argued that Kant’s doctrine of fact of reason does not reduce to a form of dogmatism, as many commentators have maintained, but instead employs an experimental method of justification. I worked to establish this claim by taking the reader through a history of the concept of ‘fact’ in Romano-canon and English common law; to the British experimentalists of the seventeenth century; and lastly, to the German translation of Butler’s Analogy. What this historical detour brought to light was an important shift in meaning, whereby facts did not designate things allegedly the case but things actually the case (and hence requiring no further proof). Kant’s own use of the terms Factum and Tatsache captures this latter sense. Much in the spirit of the seventeenth century experimentalists, however, Kant does not regard our consciousness of the moral law’s necessity as a brute given that the reader must simply accept, as if it were an item
70 Closing Remarks of faith. To the contrary, his strategy involves reproducing this consciousness by way of compelling and intuitively vivid thought experiments (such as a man facing a choice between duty and death), which allow us to attest to our consciousness of the moral law. While not amounting to a deduction or genetic proof of the latter, this method still plays an important role in Kant’s argument by securing the actuality of moral consciousness. For it is on the basis of this actuality that we are warranted to infer its source—namely, in our own faculty of practical reason.
3
Freedom and Obligation We do not indeed comprehend the practical unconditional necessity of the moral imperative, but we nevertheless comprehend its incomprehensibility; and this is all that can fairly be required of a philosophy that strives in its principles to the very boundary of human reason. – Kant (G 4:463)
3.1 Introduction In light of the previous two chapters, the question we must now come to terms with is whether Kant’s doctrine of the fact of reason marks a departure from the argument of Groundwork III. From what we have discussed so far, at least one contrast must have struck the reader as unusual: in Groundwork III, Kant pro ceeds from the standpoint of common human reason in order to commence a search for the supreme principle of morality, and it is precisely the worry that the bindingness of this principle might be a ‘figment of the imagination’ that motiv ates Kant’s transition to the third section. In the second Critique, rather than raise this skeptical worry to motivate a deduction of the moral law, Kant regards our consciousness of the moral law as a ‘fact’ that, while admitting no further proof, nonetheless serves to warrant our possession of a higher faculty of desire (or what amounts to the same, a free will). Even if the interpretation I put forth in the pre vious chapter is compelling, it does not address this larger question of compari son. But how else can we make sense of this apparent change of strategies? Does Kant’s new starting point in the second Critique not give us evidence to think that he abandoned, rightly or wrongly, his earlier argument? My thesis here is that Kant’s project of moral justification is continuous between the Groundwork and the second Critique, and defending this thesis is the task of the present chapter. After surveying the early reception of Kant’s project of moral justification, I shall concentrate on the problem of a ‘hidden circle’ that Kant raises in Groundwork III. A common way of understanding the problem suggests that Kant is seeking a deduction of the moral law that proceeds on the basis of nonmoral premises, although commentators diverge over their interpretation of these premises. On my reading, Kant uses the problem of a hidden circle to motivate a shift to a critique of pure practical reason, which marks the synthetic path of
Kant’s Justification of Ethics. Owen Ware, Oxford University Press (2021). © Owen Ware. DOI: 10.1093/oso/9780198849933.003.0004
72 The Early Reception of Kant ’ s Project descending from the principle of autonomy to the will of human beings. Working through these issues will allow me to return to the details of Groundwork III with a focus on Kant’s two-world distinction. Building upon Chapter 1, I will argue that this distinction provides us with two frameworks for cognizing our faculty of will, with the idea of an intelligible world serving as a model for the determin ation of our pure will. A surprising piece of evidence comes to light when—at G 4:457 and 4:462—Kant appeals to our consciousness of the laws governing an intelligible world, i.e., moral laws, as a basis to warrant our possession of a faculty of pure practical reason. After addressing a number of potential objections to my reading, I shall devote the final section of this chapter to explaining what I take to be the real differences between the Groundwork and the second Critique.
3.2 The Early Reception of Kant’s Project Those familiar with the standard reading will be surprised to learn that Kant’s early commentators detected no rift between Groundwork III and the second Critique. In the years following the publication of the second Critique, commen tators embraced the doctrine of the fact of reason as a paradigm for philosophical inquiry, not only in the domain of morality, but also, more controversially, in the domain of knowledge. As I noted in the previous chapter, Karl Reinhold intro duced the concept of ‘facts of consciousness’ (Tatsachen des Bewusstseins) in the late 1780s,1 a phrase that was quickly picked up by J. G. Fichte in the early 1790s. Neither of these authors spoke of Kant’s Factum as a novel development within his system of ethics. Nor was this unusual for the time.2 In his book-length com mentary on the second Critique (1796–7), for example, Christian Friedrich Michaelis discussed why our consciousness of the moral law is a fact (in the sense of a Tatsache) that admits of no further proof, a point he contrasted more than once to Kant’s transcendental deduction of the categories in the first Critique.3 In this two-volume work, amounting to over 600 pages in length, Michaelis only mentioned Kant’s Groundwork five times, and each time he did so without reference to the fact of reason.4 This is not to say Kant’s arguments from Groundwork III and the second Critique won the immediate approval of his contemporaries. Many interpreters of these books were of a critical bent.5 But what is noteworthy is that none of these 1 See his (1786–87, esp. 17, 78, 80, 83, 87, 110, 138). For a superb study of Reinhold’s role in the development of post-Kantian philosophy, see Ameriks (2000). 2 Although Fichte would soon put Kant’s doctrine of the fact of reason to novel use in developing his ‘doctrine of science’ (Wissenschaftslehre), as I argue in Ware (2020). 3 See his (1796–97, esp. vol. I. 139). For another example, see Brastberger (1792, esp. 57, 71). 4 See his (1796–97, esp. vol. I. 207, 247n, 255, 282, and 352). 5 See Tittel (1786), Pistorius (1786/1975, 1786/2007), and Stattler (1788). In the Conclusion I shall also discuss how Reinhold, Creuzer, and Maimon were, with varying degrees of sympathy, attempting
Freedom and Obligation 73 early writers even suggested that Kant’s later doctrine of the fact of reason marked a change, let alone a reversal, in his project of moral justification.6 In looking at their work now, one gets the impression that commentators were either uninter ested in the question of Kant’s intellectual development from 1785 to 1788, or they were willing to take Kant at his word that the second Critique ‘presupposes’ the Groundwork (KpV 5:8). This is how Georg Albert Mellin organized the rela tionship between these texts in his monumental Encyclopedic Dictionary of Critical Philosophy at the turn of the nineteenth century. After summarizing the three sections of the Groundwork, Mellin paraphrased Kant’s remark that the second Critique ‘presupposes’ the Groundwork insofar as it ‘constitutes preliminary acquaintance with the principle of duty and provides and justifies a determinate formula of it’ (KpV 5:8).7 Like Reinhold, Fichte, and Michaelis before him, Mellin also spoke of the fact of reason as a kind of Tatsache without ever mentioning Kant’s earlier deductions. Not a single sentence in the Encyclopedic Dictionary’s six volumes hinted at anything like a rupture between Groundwork III and the second Critique. All of this stands in contrast to the reversal reading prominent in contempor ary scholarship. So we must ask, When did commentators come to believe that Kant’s project of moral justification underwent a major turn? The answer is elu sive, largely because attitudes toward Kantian philosophy in the early nineteenth century were inconsistent. Not everyone was as enthusiastic as Reinhold and Fichte, for instance. Still, what we can say for certain is that by the early 1800s, twenty years after the publication of the second Critique, the positive reception of Kant’s ethics was on the wane. There was a growing cynicism among German scholars of the time, not only toward the general system of Kantian philosophy, but also toward the specific doctrine Kant’s early advocates cherished: the doc trine of the fact of reason. In his On the Spirit of Philosophy (1803), for example, Jakob Salat complained that Kant’s appeal to a Factum reduced the foundations of his practical philosophy to ‘mysticism’, anticipating Hegel’s famous remark that ‘cold duty is the final undigested lump left within the stomach, the revelation given to reason’.8 If there was a shared sentiment among scholars, it was that the doctrine of the fact of reason committed a gross error; although there was no consensus at the time over the exact nature of this error.
to revise Kant’s project of moral justification along the lines of a foundationalist strategy that fore shadows one dominant strand of Kantian ethics today. 6 In his review of the second Critique, published in the Allgemeine Literatur-Zeitung (1788), Hartknoch offered a unified summary of Groundwork III and the doctrine of the fact of reason, indi cating no change in their overall structures (1788, 346). Similarly, Snell explained the concept of a Factum by alluding to Kant’s own example of the ‘most malicious scoundrel’ (der ärgste Bösewicht) from Groundwork III (G 4:454). See his (1796, esp. 14–15). 7 See his (1800, esp. 174–175). 8 Quoted in Allison (1990, 281).
74 The Early Reception of Kant ’ s Project Nearly every aspect of Kant’s moral philosophy had become the object of c riticism by the early nineteenth century—as one finds in the work of Creuzer, Maimon, Schelling, and Hegel, among others. However, we do not find anything approximating the reversal reading until 1841, the year Arthur Schopenhauer published his second prize essay, On the Basis of Morals.9 Two features of Schopenhauer’s essay are worth highlighting. First, he joined Salat and Hegel in considering the moral law’s authority to be non-rational. The ‘categorical impera tive’, he wrote, ‘appears more and more as a hyperphysical fact [Tatsache], as a Delphic temple in the human soul, from whose dark sanctuary proceed oracles that infallibly declare, sadly not what will happen, but what ought to happen.’10 Second, and more importantly, Schopenhauer was convinced that Kant’s doctrine of the fact of reason contradicted his aim in the Groundwork to give morality a pure foundation. In the earlier work Kant wanted to cleanse moral philosophy of everything empirical and belonging to experience. By later appealing to a Tatsache of consciousness, Schopenhauer believed Kant violated his own method of separ ating morality from anthropology, i.e., from the propensities, capacities, and inner experiences unique to our psychological makeup. Kant, in his view, ended up mixing pure elements of moral cognition with empirical elements. Schopenhauer’s argument for why Kant committed this mistake is not clear. More than once he launched an ad hominem attack, blaming Kant’s alleged incon sistency in the second Critique to overconfidence. ‘When Kant wrote this’, he remarked, ‘he had, finally and late in life, achieved his well-deserved fame: that made him sure of limitless attention, and he allowed more scope to the garrulous ness of old age.’11 In other places, however, we find a more penetrating criticism. Schopenhauer suggested that Kant’s recourse to a fact was a product of his formalist approach in ethics. ‘Kant himself ’, he wrote, ‘must have been silently conscious of the inadequacy of a foundation of morals that consists solely in a couple of wholly abstract and contentless concepts.’12 In the second Critique, he continued, ‘the foundation of ethics very gradually alters its nature, almost forgets that it is a web of abstract concept-combinations, and seems to want to become more substantial’.13 That is why Schopenhauer believed ‘doctrines most heterogeneous to it’ emerged from the ‘strict, austere critical philosophy’.14 It was not simply the garrulity of old age, but the impossibility of pure morality itself, that led Kant to the doctrine of the fact of reason. When we step back to consider Kant’s many readers over the past two hundred years, it is not clear who was the first to articulate the reversal reading. From the review I have just provided, we can see that Schopenhauer’s criticism—that Kant’s 9 Beneke anticipated some of Schopenhauer’s objections, although he did not give them systematic expression. See his (1832, esp. 65). 10 Schopenhauer (1841/1988, 44). 11 Schopenhauer (1841/1988, 17). 12 Schopenhauer (1841/1988, 17). 13 Schopenhauer (1841/1988, 44). 14 Schopenhauer (1841/1988, 46).
Freedom and Obligation 75 doctrine of the fact of reason conflicts with his project of giving morality a pure foundation—is different from the view that Kant reversed the direction of his argument in the second Critique, proceeding from morality to freedom rather than from freedom to morality. Yet in the nineteenth-century reception of Kant’s ethics, there is no question that Schopenhauer cleared space for readers to think seriously about the nature of Kant’s intellectual development in the late 1780s. Schopenhauer’s official view was that the second Critique and the Groundwork differ only in style: ‘the latter has a more concise and strict form’, he wrote, ‘while the former does so with great breadth of execution and is interrupted by devi ations, and also, to heighten the impression, is padded by some moral rhetoric’.15 Despite this, it is fair to say that Schopenhauer paved the way for the present-day reversal reading by voicing a claim, relatively novel for the time, that Kant’s recourse to a ‘fact’ in the second Critique contradicted his earlier work. Seeds for the reversal reading were later sowed by Herbert Paton’s influential book, The Categorical Imperative: A Study in Kant’s Moral Philosophy, published in 1947. Curiously, however, Paton himself expressed a great deal of ambivalence on the topic of whether Kant changed his strategy of justification in the second Critique. On the one hand, he believed that Kant was seeking a strict proof in Groundwork III that would derive the validity of the moral law from ‘an inde pendent and necessary presupposition of freedom’, a deduction he subsequently gave up in 1788. Yet in the very next paragraph Paton said that we need not ‘exag gerate the difference between the two views’,16 adding: ‘it is doubtful how far he regards freedom in the Groundwork as a non-moral principle’.17 What is even more odd is that Paton cited the final paragraph of Groundwork III where Kant states the impossibility of comprehending the necessity of the moral law (G 4:463). Yet Paton took this as evidence that Kant was merely ‘beginning to see dimly’ that his strict proof was doomed to fail.18 Scholarship in the twentieth century did not acquire a uniform shape on this topic until 1960: the same year Dieter Henrich published his essay, ‘The Concept of Moral Insight and Kant’s Doctrine of the Fact of Reason’,19 and Lewis White Beck published his book, A Commentary on Kant’s Critique of Practical Reason.20 Each author reached the same general conclusion about Kant’s project of moral justification. As Beck put it, the argument of the second Critique ‘takes a truly
15 Schopenhauer (1841/1988, 17). 16 Paton (1947, 203) 17 Paton (1947, 204) 18 Paton (1947, 225). 19 The original 1960 essay, ‘Der Begriff der sittlichen Einsicht und Kants Lehre vom Factum der Vernunft’, was later translated in 1994. 20 Limits of space prevent me from examining the history of Kant interpretation from 1841 to 1960, but it is worth pointing out that many prominent scholars during this time saw no substantial change in the second Critique. To take two examples, both Green and Hägerström compared the Groundwork and the second Critique without ever mentioning a reversal. See Green (1886) and Hägerström (1902).
76 A Hidden Circle astonishing turn’.21 By having said ‘that the principle [of morality] needs no deduction, he apparently stands the argument of the Groundwork on its head. He uses the moral law, the fact of reason, as the prius to deduce something else, namely, freedom, which is its ratio essendi’.22 Henrich went further in his essay to explain the cause of this apparent reversal, suggesting that the argument of the Groundwork was still driven by Kant’s pre-critical project of deriving morality from theoretical reason. ‘The difference in the structure of the Groundwork and the Critique of Practical Reason can be understood’, he wrote, ‘only if we realize that Kant did not yet see clearly in the Groundwork that a deductive justification of ethics must necessarily turn out to be unsatisfactory and contradictory.23 What is remarkable, given the reception of Kant’s ethics I have just reviewed, is that the opinion Beck and Henrich expressed has gone largely unchallenged by scholars over the last sixty years. Taking up this challenge forms my critical task in this chapter.
3.3 A Hidden Circle I suspect that many readers find the reversal reading attractive because it pur ports to explain Kant’s worry about a ‘hidden circle’ in Groundwork III.24 At first all Kant says is that ‘one must freely confess that a kind of circle appears here from which, as it seems, there is no way out’ (G 4:450). Now if we assume that the threat of a circle arises from using a moral premise to establish our freedom, then we can agree that our only means to escape the circle would be to find non-moral evidence for thinking we are free. On this assumption it would be natural to hear Kant’s remarks in the second Critique as departing from his earlier approach—as when he says that the moral law, while admitting of no deduction, ‘conversely itself serves as the principle of the deduction of an inscrutable faculty’, that is, the faculty of freedom (KpV 5:47).25 Yet the naturalness of this reading rests on how we characterize Kant’s worry in Groundwork III. If the threat of a circle does not come from using a moral prem ise to establish our freedom, then it is far from obvious that our only means of escape is to find a non-moral route. Let us consider what Kant says more closely:
21 Beck (1960, 173). 22 Beck (1960, 172). 23 Henrich (1960, 81). Henrich appears to have changed his mind in a later essay where he main tains that Kant’s argument in Groundwork III has an irreducibly moral core. See Henrich (1975, 86). 24 For fuller treatments of the circle problem, see Schönecker (1999), Allison (2011), and Berger (2015). 25 While I do not have space to defend this claim here, I take Kant to be committed to a strong form of incompatibilist freedom. The literature on this issue is quite extensive. For recent treatments, see Pereboom (2006), Kohl (2020), and McLear (2020).
Freedom and Obligation 77 It must be freely admitted that a kind of circle comes to light here from which, as it seems, there is no way to escape. We take ourselves as free in the order of effi cient causes in order to think of ourselves under moral laws in the order of ends; and we afterwards think of ourselves as subject to these laws because we have ascribed to ourselves freedom of will. (G 4:450)
Upon reading this, it seems Kant is saying that a circle arises if we proceed from a moral starting point—for example, if we consider ourselves under moral laws in the order of ends. Consequently, it is tempting to think that our only hope of escaping the circle lies in finding a non-moral premise, one that would establish the validity of moral laws without begging the question. While the content of this premise is a topic of debate, the majority of scholars agree that Kant’s argument has this overall structure. They believe his aim is to find non-moral evidence for thinking we are free—either on the grounds that we must engage in deliberation or on the grounds that our reason is self-active.26 While tempting, I find this familiar portrait of Kant’s worry problematic. When we look at how he describes the circle more closely, it is by no means clear that our only hope of escape lies in finding a non-moral route. As we read further Kant suggests that our only hope of escape lies in finding a non-analytic or synthetic route. Consider what he says, for instance, in the remainder of the passage I quoted above: For, freedom and the will’s own lawgiving are both autonomy and hence recipro cal concepts, and for this very reason one cannot be used to explain the other or to furnish a ground [Grund] for it but can at most be used only for the logical pur pose of reducing apparently different representations of the same object to one single concept (as different fractions of equal value are reduced to their lowest expression). (G 4:450)
The problem of a circle emerges, then, because freedom, autonomy, and morality are all co-entailing concepts. Analyzing freedom leads us to the concept of auton omy, and analyzing autonomy leads us to the concept of morality. Yet the move ment from one to another is not a movement of justification: it is a movement of clarification. At no point have we established a ground that would reveal their synthetic connection. All we have shown, rather, is that apparently different representations reduce to a common source. And that is true regardless of our starting point. Whether we start from freedom and end up with the principle of morality, or whether we start from the principle of morality and end up with freedom, we cannot conceptualize our way to the validity of a concept.
26 I will discuss these two interpretive options in §3.8 and §3.9.
78 A Hidden Circle This begins to explain why Kant voices his worry in the first place. What commentators often overlook is that in Groundwork III Kant is trying to motivate the final ‘transition’ of the book, namely, from a metaphysics of morals to what he calls a critique of pure practical reason or a ‘critique of the subject’ (G 4:440). As we saw in Chapter 1, a metaphysics of morals is an effective procedure for clarify ing our common idea of duty. When we proceed along the analytic path, we can ascend to autonomy as the highest formula of the moral law. We would never have discovered this formula had we, like the popular philosophers of Kant’s day, adopted a mixed method of inquiry. And yet, as we learn by the end of Groundwork II, the method of a metaphysics of morals is limited. Since it does nothing more than clarify the supreme principle of morality, it leaves open the question of whether this principle is actually binding upon a human will (G 4:445). What we need then is a new path, the path of justification. We need a critical examination of our faculty of practical reason—separating its empirical and pure uses—to warrant our possession of the latter. Only then can we descend along the synthetic path and, recombining the empirical and pure uses of our reason, uncover the ground of the moral ‘ought’ in our own ‘will’. My suggestion, in short, is that Kant is highlighting a procedural issue at the beginning of Groundwork III. Speaking of a potential circle is his way of telling the reader that we need to go beyond a metaphysics of morals and examine our faculty of reason instead. To be clear, Kant is not saying that we must find a nonmoral route to freedom. He is saying that we must find a new method of justifica tion, one that would legitimate our claim to freedom, and it is left open whether Kant’s critique in Groundwork III has a moral or non-moral point of orientation. Contrary to the standard reading, I will argue that Kant’s critique appeals to a normatively thick premise, namely, the idea of an intelligible world, and that he uses this premise to secure the belief that we are free. One advantage of this read ing is that it makes sense of Kant’s opening remark in Groundwork III that he plans to offer a ‘deduction of the concept of freedom from pure practical reason, and with it the possibility of a categorical imperative’ (G 4:447).27 A deduction of the concept of freedom from pure practical reason amounts to a moral argument for freedom, one very similar to the moral argument of the second Critique.28 27 With the exception of Wolff (2009, 546, note 64) and Puls (2011, 543, note 19), the significance of this remark has been neglected in the literature devoted to Groundwork III. 28 Kant only refers to a ‘deduction’ (Deduktion) three times in the Groundwork: the first, just cited, refers to a deduction of the concept of freedom (G 4:447); the second refers to a deduction of the cat egorical imperative (G 4:454); and the third refers to a deduction of the supreme principle of morality (G 4:463). Allison (2011) takes the presence of this third reference to tip the scales in favor of his own reading. However, while Allison is correct to identify the supreme principle of morality with the moral law rather than its form as a categorical imperative (2011, 275), this passage does not offer sup port to identify a deduction of the moral law in Groundwork III. What Allison does not mention, although this is the all-important point, is that Kant says it is ‘no reproach’ against his foregoing deduction, but only a shortcoming inherent in human reason itself, that it cannot make the moral law’s ‘absolute necessity’ (absoluten Notwendigkeit) ‘comprehensible’ (begrieflich) (G 4:463). In this
Freedom and Obligation 79
3.4 The Intelligible World: A Moral Ideal After introducing the problem of a circle, Kant tells us that there is one way out, namely to try: whether when, through freedom, we think of ourselves as causes efficient a priori we do not take up a standpoint that is different from when we represent ourselves according to our actions as effects that we see before our eyes. (G 5:450)
This new standpoint leads Kant to the concept of an intelligible world, which we arrive at by abstracting from all conditions of the world of sense, including our needs, impulses, and inclinations. What is left over from this abstraction is a com plete pure concept or ‘Idea’, namely, the Idea of a world entirely governed by rational laws. Kant makes this point explicit in the second Critique when he char acterizes the intelligible world as the ‘archetypal world’, which he says is a ‘model [Muster] for the determinations of our will’ (KpV 5:43). What is more, in both texts Kant is clear that the force of this model is entirely practical, meaning that it gives speculative reason nothing to intuit or cognize.29 In a revealing passage near the end of Groundwork III, Kant writes: I might indeed revel in the intelligible world . . . but even though I have an idea of it, which has its good grounds, yet I have not the least cognizance of it . . . It signi fies only a ‘something’ that is left over when I have excluded from the determin ing grounds of my will everything belonging to the world of sense, merely in order to limit the principle of motives from the field of sensibility by circumscribing this field and showing that it does not include everything within itself but that there is still more beyond it . . . As for pure reason—which thinks this ideal—after its isolation from all matter, that is, cognition of objects, nothing is left for me but the form of it—namely the practical law of the universal validity of maxims—and to think of reason, conformably with this, with reference to a pure world of understanding as a possible efficient cause, that is, a cause determining the will. (G 4:462; emphasis added)30
context Kant’s third reference therefore denies, rather than supports, the view that Groundwork III contains a deduction of the moral law, for such a deduction would require precisely what Kant denies we can attain: namely, insight into the absolute necessity of the moral law as a principle valid for all rational beings. 29 For further discussion of the concept of an archetypal world and its connection to Kant’s theory of the will, see Ware (2010). 30 The second emphasis is mine. For another revealing passage, consider what Kant says about the ‘moral world’ in the first Critique: ‘I call the world as it would be if it were in conformity with all moral laws (as it can be in accordance with the freedom of rational beings and should be in accordance with the necessary laws of morality) a moral world. This is conceived thus far merely as an intelligible
80 The Intelligible World: A Moral Ideal The direction of Kant’s argument comes out clearly here. First, he is saying we can think of the law of an intelligible world—the ‘practical law of the universal valid ity of maxims’—and we can think of our reason in conformity with this law—as a ‘cause determining the will’. In doing so we recognize our capacity to act inde pendently of all motives coming from the world of sense. We recognize our fac ulty of reason as having a pure use beyond the satisfaction of our needs, impulses, and inclinations. Second, as the above passage makes plain, this recognition is not a product of speculative reason. Kant is saying that when we think of ourselves in conformity with the law of an intelligible world, i.e., the moral law, we have grounds to think of ourselves as free. This pre-empts the threat of a circle, then, not because it gives us a non-moral route to freedom, but because it gives us a non-analytic or synthetic route to freedom. We can go beyond the mere clarifica tion of concepts and actually claim to be free, but—and this is the crucial point I wish to emphasize—the basis of our claim comes from a moral Idea.31 Support for this reading comes from what Kant says about the ‘common’ pre supposition that we are free. The ‘legitimate claim even of common human rea son to freedom of the will’, he writes, ‘is founded on the consciousness and the granted presupposition of the independence of reason from merely subjectively determining causes’ (G 4:457). In the next paragraph Kant then links this con sciousness of independence to the ‘law’ (Gesetz) we give ourselves in the intelli gible world: That is why a human being presumes for himself a will that lets nothing belong ing merely to his desires and inclinations be put on its account, and on the contrary thinks possible—indeed even necessary—through himself actions that can be done only by setting aside all desires and sensuous stimulations. Their causality lies in him as an intelligence and in the laws of effects and actions world, since abstraction is made therein from all conditions (ends) and even from all hindrances to morality in it (weakness or impurity of human nature). Thus far it is therefore a mere, yet practical, idea, which really can and should have its influence on the sensible world, in order to make it agree as far as possible with this idea. The idea of a moral world thus has objective reality, not as if it pertained to an object of an intelligible intuition (for we cannot even think of such a thing), but as pertaining to the sensible world, although as an object of pure reason in its practical use’ (A808/B836). 31 One might be open to this claim but still wonder why such an ideal must be a recognizably moral one (as is the fact of reason) as opposed to a more generic ideal of rationality or agency. As an anonymous reviewer puts it: ‘Kant’s argument in Groundwork III might just be that once we appreci ate the connections between morality, autonomy, and rationality, we can also see that moral acts are the only possible fully rational or intelligible or justified ones, or those in which our agency is most fully realized.’ The reason why I wish to insist upon calling the concept of an intelligible world a moral ideal is that it specifies practical laws that, in contrast to technical or prudential imperatives, express absolute necessity. Moreover, Kant himself is clear in the first Critique that the intelligible world is a moral ideal, so my reading has good textual support. As he states, the intelligible world is nothing more than the idea of the ‘world as it would be if it were in conformity with all moral laws’ (KrV A808/B836; see also Refl 5086 and 6977 dated from the 1770s). For a helpful account of this passage in connection with Groundwork III, see Rauscher (2009, 2015).
Freedom and Obligation 81 according to principles of an intelligible world, of which he may well know nothing more than that solely reason, and indeed pure reason independent of sensibility, gives the law in it. (G 4:457)32
Our warrant for thinking we are free therefore comes not merely from our mem bership in an intelligible world, but from our consciousness of the law we give to ourselves in this world—the ‘moral law’ (Moralgesetz). No longer limited to the world of sense and the practical perspective it affords us, we can at last see the real possibility of determining our will to action on the basis of pure reason alone. And since we are no longer analyzing a relationship between concepts, the threat Kant used to motivate his discussion, the threat of a ‘hidden circle’, is now out of place.33 Of course, one might still wonder: Even if we are entitled to think of ourselves as free for the reasons just spelled out, what right do we have to speak of those moral laws governing the intelligible world? Do they not stand in need of justification as well? Kant’s answer is clear, but it only becomes explicit at the end of Groundwork III. Given the nature of human reason, he explains, we can only comprehend what happens or what ought to happen on the basis of prior conditions (G 4:463). Yet the laws of morality convey absolute necessity. So when it comes to those laws themselves, the most Kant says we can do is appreciate the fact that our philo sophical investigation has come to an end. That means we cannot secure a deduc tion of the moral law itself (or a genetic proof that would afford us insight into its source), but we can at least see why this is no fault of our own. ‘It is’, he writes, ‘no criticism of our deduction of the supreme principle of morality, but an accusation that would have to be brought against human reason in general, that it cannot make comprehensible, as regards its absolute necessity, an unconditional practical law (such as the categorical imperative must be)’ (G 4:463).
32 It is also worth pointing out that when Kant introduces the concept of an intelligible world in Groundwork III, he says that a human being has ‘two standpoints from which he can regard himself and cognize [erkennen] laws for the use of his powers’, by which he means to include our practical powers of will and volition (G 4:452). As members of the world of sense, Kant explains, we have a standpoint to consider ourselves ‘under laws of nature’, but as members of the intelligible world, we have a standpoint to consider ourselves ‘under laws which, being independent of nature, are not empirical but grounded merely in reason’ (G 4:452). Since laws ‘grounded merely in reason’ are none other than the laws of morality (cf., G 4:458 and 4:462), the implication of this passage is very similar to what Kant will assert in the second Critique: that our moral consciousness reveals our freedom of will. 33 At this ‘highest limit of moral inquiry’, what remains beyond our epistemic reach is not the moral law’s conceptual structure. Kant had already clarified that structure in Groundwork II using a method of analysis. What remains ‘incomprehensible’ then is the moral law’s necessity. And Kant repeats this point in the second Critique when he explains why the moral law, unlike the pure categor ies of the understanding, admits of no justification: ‘With the deduction [of the moral law], that is, the justification of its objective and universal validity and the discernment of the possibility of such a synthetic proposition a priori, one cannot hope to get on so well as was the case with the principles of the pure theoretical understanding’ (KpV 5:46). See Baum (2014) for further discussion.
82 The Intelligible World: A Moral Ideal Bringing these parts of our discussion together, we can organize Kant’s argument into the following steps: Step 1. When we think of ourselves as causes efficient a priori, we presuppose our membership in a world beyond our impulses and inclinations. Step 2. This merely intelligible world gives us nothing to intuit or cognize. But it gives us a model or Idea for thinking of our will in conformity with rational laws. Step 3. When we think of our will in this way, we have a basis for recognizing our positive freedom: our capacity to act by reason alone. This capacity is equiva lent to autonomy. Having reached this point, Kant is able to connect his deduction back to the for mula of morality he had established earlier in Groundwork II: Step 4. Autonomy is equivalent to the formula of morality: ‘[So act] that the maxims of your choice are also included as universal law in the same volition’ (G 4:440; cf., 4:431). Now where does this leave us in Groundwork III? After avoiding the threat of a circle, Kant turns to consider the question of how the moral law takes the form of an imperative for us in subsection four, titled ‘How is a Categorical Imperative Possible?’34 As we saw in Chapter 1, Kant’s argument is that the moral law takes the form of an imperative for us because it has its source within us: what we ‘would’ do were we only members of an intelligible world is what we ‘ought’ to do as beings who belong to the world of sense ‘at the same time’ (G 4:454). Even the malicious scoundrel recognizes this, Kant argues, when he attends to examples of virtuous conduct and ‘shifts standpoints’ to a world beyond his impulses and inclinations. Without repeating the details of this argument, all I want to empha size here is that Kant had stated his division of tasks in this exact order, i.e., to provide a ‘deduction of the concept of freedom from pure practical reason, and with it the possibility of a categorical imperative’ (G 4:447).35 So looking back this is just what we should have expected: a warrant for our consciousness of freedom, followed by a warrant for our consciousness of the moral ‘ought’. Keeping this 34 My understanding of Kant’s second step has been influenced by Timmermann (2007) and Stern (2011, 2015), who both argue—correctly, I believe—that in this section of Groundwork III, Kant is only trying to explain the possibility of moral bindingness (i.e., the imperatival form the moral law takes to finite rational beings like us), not the moral law itself. For a helpful discussion of this distinc tion, i.e., between the moral law and its imperatival form, see Willaschek (1992, 176). 35 This comment strongly indicates that Kant envisions the work of justification to consist of two deductions, one of freedom and another of the categorical imperative, or at least to consist of a deduc tion with two distinguishable steps. Whatever we make of the relationship of these deductions (either as two proofs or as one proof with two steps), it is clear that Kant aims to warrant the concepts of freedom and the categorical imperative, and it is also clear that this warrant will proceed ‘from pure practical reason’ (which is to say, not from purely theoretical or practically neutral premises).
Freedom and Obligation 83 distinction in mind, we now have enough material to pursue a closer comparison to Kant’s argument in the second Critique.36
3.5 The Argument of the Second Critique Those sympathetic to the reversal reading may wonder how I can explain Kant’s apparent confession in the second Critique that his earlier strategy of justification was a failure. In a much-discussed passage Kant writes that ‘the moral law cannot be proved by any deduction’, adding: But something different and quite paradoxical takes the place of this vainly sought deduction of the moral principle, namely that the moral principle con versely itself serves as the principle of the deduction of an inscrutable faculty which no experience could prove but which speculative reason had to assume as at least possible . . . namely the faculty of freedom, of which the moral law, which itself has no need of justifying grounds, proves not only the possibility but the reality in beings who cognize this law as binding upon them. (KpV 5:47)
As far as confessions go, this passage seems to present us with evidence that Kant changed his mind by the time he composed the second Critique. For what could this ‘vainly sought deduction of the moral principle’ refer to, if not the very deduction Kant had ventured in 1785?37
36 Many commentators agree that Kant’s appeal to the idea of an intelligible world plays a crucial role in Groundwork III, but what they overlook is that this idea has two functions. When we consider ourselves as members of an intelligible world, we have grounds to ascribe a negative sense of freedom to ourselves since we are presupposing our independence from the world to which our impulses and inclinations belong, the world of sense. Yet, on my reading, this is not yet the key step. The key step is that the intelligible world also affords us a standpoint for thinking of ourselves in accordance with a pure practical law, i.e., the moral law. Only now do we have grounds to ascribe a positive sense of freedom to ourselves—to think of ourselves as possessing a will—since the moral law is self-legislated. Note, too, that at G 4:458 Kant says that these two senses of freedom are ‘bound’ (verbunden) in the idea of an intelligible world, reiterating his earlier stipulation that synthetic propositions are ‘only pos sible by this, that both cognitions are bound [verbunden] together by their connection with a third, in which they are both to be found’ (G 4:447; cf., B162–5). 37 Timmermann, for example, has argued that Kant’s mention of a ‘vainly sought deduction’ (vergeblich gesuchten Deduction) is a thinly veiled reference to Groundwork III: ‘Of course, Kant does not explicitly say that he is the person who tried in vain to provide a deduction of the moral principle, but as Groundwork III contains precisely such a deduction, and proving the possibility of the categorical imperative was hardly fashionable amongst his philosophical colleagues, we can safely infer that Kant had come to reject his earlier justificatory attempt by the time he composed the second Critique’ (2010, 74). On closer inspection, however, this assumption is far from self-evident. Just before Kant says that ‘something different and quite paradoxical takes the place of this [dieser] vainly sought deduction’, he writes: ‘The objective reality of the moral law cannot be proved by any deduction, by any efforts of theoretical reason, speculative or empirically supported’ (KpV 5:47). This suggests Kant’s reference is quite general and does not necessarily imply his earlier argument. Given its placement in the text, the ‘vainly sought deduction’ is only a reproach against efforts to establish the validity of the moral law on theoretical grounds, i.e., the sort we find in the writings of Hobbes, Hutcheson, Wolff, or Crusius.
84 The Argument of the Second Critique This passage acquires a different meaning, however, when we read it alongside the closing statements of Groundwork III. In the final paragraph of the work, as we have seen, Kant explains that our faculty of reason faces a dilemma. On the one hand, our reason is driven to seek the condition of what happens or what ought to happen with the aim of finding something absolutely necessary, some thing that would end the regress of conditions. On the other hand, due to the very limitations of reason itself, we can only comprehend what happens or what ought to happen when we discover its underlying condition. So the very principle driving the activity of reason itself, namely, to find something unconditionally necessary, can never be satisfied. As Kant puts it, reason ‘restlessly seeks the unconditionally necessary and sees itself constrained [genöthigt] to assume it without any means of making it comprehensible to itself, fortunate [glücklich] enough if it can dis cover only the concept that is compatible with this presupposition’, that is, the concept of freedom (G 4:463; cf., KpV 5:93–4).38 This is why Kant thinks we should not hope to gain insight into the moral law’s necessity, such as a deduction or genetic proof would require of us, because its necessity is by definition absolute. For this reason a deduction of the moral law would be vainly sought since we cannot ‘comprehend the practical unconditional necessity [Notwendigkeit] of the moral imperative’ (G 4:463). When Kant makes this point explicit in the second Critique, he may have Groundwork III in mind, but not because he is repudiating his earlier approach. On the reading I favor, Kant may simply be wanting to draw the same lesson in each text, that human insight reaches an end as soon as we are dealing with fundamental laws or funda mental faculties (KpV 5:47). That is why seeking to comprehend the moral law would be futile. All we can do is comprehend its incomprehensibility, which is, as Kant says, ‘all that can fairly be required of a philosophy that strives in its prin ciples to the very boundary of human reason’ (G 4:463). On this point at least Groundwork III and the second Critique stand in agreement: a deduction or genetic proof of the moral law is beyond our epistemic reach. Granted, when Kant says that the moral law ‘cannot be proved by any deduc tion’, it is hard to rid ourselves of the impression that he is referring to his earlier strategy of argument. Yet further inspection reveals that this impression is deceiving, since the argument of Groundwork III refers explicitly to a deduction 38 Kant echoes various turns of phrase from Groundwork III when he writes: ‘But instead of the deduction of the supreme principle [obersten Princips] of pure practical reason—that is, the explan ation of the possibility of such a cognition a priori—nothing more could be adduced than that, if one had insight into the possibility of freedom of an efficient cause, one would also have insight into not merely the possibility but even the necessity of the moral law . . . But no insight can be had into the possibility of the freedom of an efficient cause, especially in the sensible world: we are fortunate [glücklich] if only we can be sufficiently assured that there is no proof of its impossibility, and are now constrained [genöthigt] to assume it and are thereby justified in doing so by the moral law, which postulates it’ (KpV 5:93–4).
Freedom and Obligation 85 of the categorical imperative, not to a deduction of the moral law (G 4:454). The ‘insight’ (Einsicht) Kant is seeking in the final section of the Groundwork is insight into how the moral law takes the form of an imperative for finite rational beings like us. Strictly speaking, he is seeking a genetic proof of the law’s normative ‘necessitation’ (Nötigung), not a proof of the law’s normative ‘necessity’ (Notwendigkeit). As we have seen, Kant denies that we could ever attain the latter given the subjective limitations of our reason (the fact that we only comprehend things on the basis of prior conditions). Accordingly, Kant’s remark about a ‘vainly sought deduction’ does not support the reversal reading, since the deduc tion he affirms in Groundwork III and the deduction he denies in the second Critique are not one and the same.39
3.6 Presupposing the Groundwork Interestingly, the only place where Kant makes a direct reference to the Groundwork is in the Preface to the second Critique, but what he says there is actually positive: I must leave it to connoisseurs of a work of this kind to estimate whether such a system of pure practical reason as is here developed from the first Critique of it has cost much or little trouble, especially so as not to miss the right point of view from which the whole can be correctly traced out. It presupposes, indeed, the Groundwork for the Metaphysics of Morals, but only insofar as this constitutes preliminary acquaintance with the principle of duty and provides and justifies [rechtfertig] a determinate formula of it; otherwise, it stands on its own. (KpV 5:8)
For decades after the publication of the second Critique, readers would refer to this passage to explain the relationship between Kant’s two books, as we saw with Mellin’s Encyclopedic Dictionary. Looking at this passage now, there is no trace of a rift in Kant’s conception of the path he took from 1785 to 1788. Instead of con fessing a change of mind, he underlines the continuity of his project, not only with his formulation of the moral principle (the task of Groundwork I and II), but also with his justification of it (the task of Groundwork III). What is more, the language of KpV 5:8 is strongly reminiscent of what Kant says in the Preface to the Groundwork: that the book offers ‘nothing more than the search for [Aufsuchung] and establishment of [Festsetzung]’ the highest principle of morality (G 4:392). When Kant later says that the second Critique ‘presupposes’ the Groundwork, he is not restricting his statement to the task of clarification, for in 39 See McCarthy (1982) and Benton (1978).
86 The Continuity Reading each place he speaks of justification (Rechtfertigung or Festsetzung).40 At the very least, there is no direct evidence here to support the reversal reading.41
3.7 The Continuity Reading All the pieces for my continuity reading are now in place. To begin with, Kant still accepts the first step of his previous deduction: Step 1. When we think of ourselves as causes efficient a priori, we presuppose our membership in a world beyond our impulses and inclinations.42 One difference is that in the second Critique Kant does not move directly to Step 2: Step 2. This merely intelligible world gives us nothing to intuit or cognize. But it gives us a model or Idea for thinking of our will in conformity with rational laws. Instead, Kant appeals to the Factum of reason: ‘It is therefore the moral law’, he writes, ‘that first offers itself to us and, inasmuch as reason presents it as a deter mining ground not to be outweighed by any sensible conditions and indeed quite independent of them, leads directly to the concept of freedom’ (KpV 5:29–30; cf., 5:4n). Nevertheless, it is clear that Step 2 of Groundwork III plays a role similar to the Factum of reason. In both cases Kant is appealing to our consciousness of pure practical laws to establish our freedom. In this respect both the arguments of Groundwork III and the second Critique employ a normatively thick premise to infer: 40 For an insightful discussion of these points, see Wood (2011). Schönecker defends a different reading, arguing that Kant is only speaking of justifying a ‘determinate formula’ of the moral principle, the project of Groundwork II (2013, 8, note 10). Yet on examination Kant’s remark proves to be ambiguous. When he speaks of justifying a ‘determinate formula’ of the moral principle, he could mean either (a) justifying the conceptual structure of the principle, or (b) justifying the principle’s application to us as a categorical imperative. I think (b) is a plausible reading, especially since Kant never characterizes the task of Groundwork I or II in terms of Rechtfertigung. That being said, even if Schönecker is right, Kant’s exclusion of Groundwork III from what the second Critique ‘presupposes’ does not offer direct support for the view that he abandoned his earlier strategy of justification. So even if we read Kant’s remark in terms of (a), further evidence is required to support the reversal reading. 41 Aside from the passage at KpV 5:8, there is only one other place in the second Critique where Kant alludes to his earlier strategy of justification. As rational beings, he writes, we can regard our selves as members of an intelligible order, ‘for it has been sufficiently proved elsewhere that freedom, if it is attributed to us, transfers us into an intelligible order of things’ (KpV 5:42; emphasis added). Here, Kant is referring to Step 1 of his deduction from Groundwork III, that if we think of ourselves as free from external influences, we must take up a ‘different standpoint’ than when we ‘represent ourselves in terms of our actions as effects that we see before our eyes’ (G 4:451). But we must ask: Why would Kant cite this step with approval in the second Critique? If he had reversed his structure of justifica tion, consistency would have forced him to reject this earlier means of escaping the circle between freedom, autonomy, and morality. Instead of rejecting this move, though, we find Kant speaking of the doctrine of the fact of reason and his earlier strategy in the very same paragraph, as if the two were entirely compatible (KpV 5:42). 42 Related to the previous note, Kant’s commitment to Step 1 comes out clearly when he writes: ‘it has been sufficiently proved elsewhere [i.e., in Groundwork III] that freedom, if it is attributed to us, transfers us into an intelligible order of things’ (KpV 5:42; emphasis added).
Freedom and Obligation 87 Step 3. When we think of our will in this way, we have insight into our positive freedom: our capacity to act by reason alone. This capacity is equivalent to autonomy. The parallels I am tracing do not end here, however. Later in the second Critique Kant explains that the law of autonomy ‘is the moral law, which is therefore the fundamental law of a supersensible nature and of a pure world of the understand ing’ (KpV 5:42). As we have seen, Kant makes a similar move in Groundwork III when he connects the deduction back to his formula of morality.43 So in both texts Kant is committed to: Step 4. Autonomy is the highest formula of morality: ‘[So act] that the maxims of your choice are also included as universal law in the same volition’ (G 4:440; cf., 4:431).44 Lastly, it is worth pointing out that Kant offers a brief explanation of how a moral ‘ought’ is possible in the second Critique, writing: Now this principle of morality . . . is, therefore, not limited to human beings only but applies to all finite beings that have reason and will and even includes the infinite being as the supreme intelligence. In the first case, however, the law has the form of an imperative, because in them, as rational beings, one can presup pose a pure will but, insofar as they are beings affected by needs and sensible motives, not a holy will, that is, such a will as would not be capable of any maxim conflicting with the moral law. Accordingly the moral law is for them an imperative that commands categorically because the law is unconditional; the relation of such a will to this law is dependence under the name of obligation, which sig nifies a necessitation [Nötigung], though only by reason and its objective law, to an action which is called duty. (KpV 5:32)
While the details are missing, we can see that Kant is working toward the same conclusion as Groundwork III. He is saying that moral imperatives are possible for us because they have their source within us: what we ‘would’ do were we only 43 Kant’s motive for making the doctrine of the fact of reason prominent in the second Critique was perhaps a reaction to Tittel (1786). As I mentioned in the Introduction, Kant replied to Tittel by saying: ‘A reviewer who wanted to say something censuring this work [i.e., the Groundwork] hit the mark better than he himself may have intended when he said that no new principle of morality is set forth in it but only a new formula. But who would even want to introduce a new principle of all morality and, as it were, first invent it? Just as if, before him, the world had been ignorant of what duty is or in thoroughgoing error about it’ (KpV 5:8n). By stressing that our consciousness of the moral law is something actual—a ‘fact’ that precedes abstract speculation—Kant likely wanted to prevent other readers from confusing the formula of morality with morality itself, as Tittel did. It is also telling that the footnote at KpV 5:8 appears just after Kant speaks of his earlier attempt to justify the formula of morality in Groundwork III. 44 Kant’s formula of the ‘fundamental law’ in the second Critique is arguably equivalent to the formula of autonomy: ‘So act that the maxim of your will could always hold at the same time as a principle in a giving of universal law’ (KpV 5:30).
88 Objections and Replies members of an intelligible world is what we ‘ought’ to do as beings who belong to the world of sense ‘at the same time’ (G 4:454). On the whole, then, there is remarkable consistency in the layout of each argument. Kant is not venturing a deduction of the moral law itself, but in both places he is providing a deduction of the concept of freedom from pure practical reason, followed by an account of how categorical imperatives are possible.
3.8 Objections and Replies There are at least three objections one could raise against the preceding sections of this chapter. One could argue that I have overlooked Kant’s solution to the circle problem in Groundwork III, and one could argue that I have overlooked evidence that Kant is seeking a non-moral route to freedom, both in the Groundwork and in his earlier writings.
3.8.1 The Problem of a Hidden Circle Kant gives us two characterizations of the circle problem in Groundwork III, and commentators often draw upon the second to support the view that he is seeking a non-moral argument. On this occasion, Kant writes: The suspicion that we raised above is now removed, the suspicion that a hidden circle was contained in our inference from freedom to autonomy and from the latter to the moral law—namely that we perhaps took as a ground the idea of freedom only for the sake of the moral law, so that we could afterwards infer the latter in turn from freedom, and that we were thus unable to furnish any ground at all for the moral law but could put it forward only as a petitio principii [or ‘question-begging’ claim] disposed souls would gladly grant us, but never as a demonstrable proposition. (G 4:453)
One might object that if my reading is correct, and Kant is appealing to a moral premise to establish our freedom of will, then he should be saying something else in this passage—that the suspicion of a circle has been confirmed. On my reading, it may seem that Kant is starting with the moral law (something ‘disposed souls would gladly grant us’) in a way that would be question-begging if directed to a skeptic (a ‘petitio principii’ but never a ‘demonstrable proposition’). When we return to the cited passage, however, Kant is clear in saying that he has established the moral law as a demonstrable proposition, and this suggests he was arguing to, rather than from, a moral premise, just as the standard reading tells us.45 45 Thanks to an anonymous reviewer for formulating this objection.
Freedom and Obligation 89 As compelling as this evidence may appear, let us first consider the meaning of Kant’s terminology more closely. On first glance, talk of a ‘demonstrable proposi tion’ seems to be precisely what Kant denies in the second Critique: that we can gain insight into the moral law’s necessity as a principle for all rational beings. And if this is what he has in mind in Groundwork III, there would be good reason to think his project of moral justification changes significantly in later works. Yet on further inspection it turns out that Kant’s reference at G 4:453 is open to inter pretation. The proposition he says requires demonstration could refer to his formula of the moral law, rather than to the necessity of the law itself (which Kant later emphasizes we cannot comprehend at G 4:463). This would be consistent with my reading because the ‘moral premise’ I have claimed Kant employs to establish our freedom of will (the pure practical law we cognize from the stand point of an intelligible world in Step 2) is distinct from the ‘moral conclusion’ he works toward (the formula of this law in terms of autonomy in Step 4). In light of this distinction, talk of a demonstrable proposition is open to interpretation, which is all I need to make the continuity reading a plausible alternative.
3.8.2 A Non-Moral Route in the Groundwork A second objection I would like to consider is that my reading of Groundwork III overlooks two passages where Kant appears to be offering non-moral evidence for freedom. The first concerns his remark that rational beings must act under the idea of freedom; the second concerns his remark that our faculty of reason is selfactive. Scholars have drawn upon these remarks to defend either non-metaphysical or metaphysical interpretations of Groundwork III. Yet both lines of interpretation encounter difficulties, as we shall see.
First Passage: The Idea of Freedom In the first passage, Kant writes: Now I say: every being that cannot act otherwise than under the idea of freedom is actually free, in a practical respect, precisely because of that; i.e., all laws that are inseparably bound up with freedom hold for it just as if its will had also been declared free in itself, and in a way that is valid in theoretical philosophy. (G 4:448)46
Given its placement in the text, it may look like Kant is advancing the following argument: (i) that freedom and morality are co-entailing concepts, (ii) that we must think of ourselves under the idea of freedom, and (iii) that morality is therefore valid for us. The problem with this interpretation, however, is that Kant 46 For further discussion of this passage, see Tenenbaum (2012), Saunders (2014), and Horn (2015).
90 Objections and Replies does not intend his remark about acting ‘under the idea of freedom’ to serve as a premise in his deduction. The remark occurs in subsection two where Kant is offering nothing more than ‘preparation’ for the deduction proper (G 4:447). Afterwards Kant explains that all he has done in subsection two is trace the ‘deter minate concept of morality back to the idea of freedom’, indicating that his dis cussion here still belongs to the analytic path of a metaphysics of morals, i.e., before he has made the final step in Groundwork III to the synthetic path of a cri tique. Kant even remarks that ‘we could not, however, prove freedom as some thing actual even in ourselves or in human nature; we saw only that we must presuppose it if we want to think of a being as rational and endowed with con sciousness of its causality with regard to its actions, i.e., with a will’ (G 4:447). In other words, Kant’s claim is only that we must attribute freedom to the concept of a rational being with a will, and the question of whether we must attribute free dom to ourselves remains open. Still, one might think we can reconstruct an argument for freedom on the basis of subsection two, regardless of how well it captures Kant’s intentions in Groundwork III. In fact, a number of contemporary philosophers have thought there is a deeper truth to Kant’s assertion that a rational being must act ‘under the idea of freedom’, since this treats freedom as a ‘necessary condition of playing the game of deliberation’47 or as a ‘fundamental feature of the standpoint from which decisions are made’.48 Along these lines it seems Kant could have said: 1. We must play the game of deliberation. 2. In playing this game we must think of ourselves under the idea of freedom. 3. A free will and a will under moral laws are co-entailing concepts. 4. Therefore, given (1)–(3), morality is valid for us. Even as a reconstruction, however, it is not clear whether this argument works, since the sense of ‘freedom’ in premise (2) appears too weak to support the infer ence to premise (3). To see why, suppose I am caught between two desires, say, between working the rest of the afternoon or going to the movies. We can agree that there is a sense of ‘freedom’ relevant for my decision making: when I give reasons for my action, whether to work or go to the movies, I am guiding myself by concepts and by doing so I am presupposing my distance from the two desires I feel at the moment. Granting all of this, we are still not committed to saying that I must regard myself as independent from all external influences when I play the game of deliberation, for I could very well be calculating the comparative degree of pleasure I will receive from one activity versus the other. Nor are we committed to saying that I must regard myself as autonomous in this game, i.e., as legislating myself by laws of reason, for the principles I appeal to in making up my mind may 47 Hill (1998, 265).
48 Korsgaard (1996b, 163).
Freedom and Obligation 91 be prudential and so dependent on my natural sensibility after all. In brief, the status of freedom as a practical assumption, even if a true characterization of our agency in general, falls short of both the negative and positive senses of freedom in premise (3), and so falls short of supporting the desired conclusion.
Second Passage: The Self-Activity of Reason Aside from whether the above reconstruction works, those interested in capturing Kant’s intentions in Groundwork III will no doubt press me on another passage where he appears to present non-moral evidence for freedom.49 Consider what he says, for example, about the faculty of reason: Now, a human being actually does find in himself a capacity by which he is distinguished from all other things, even from himself, in so far as he is affected by objects, and that is reason . . . [This faculty] under the name of the ideas shows a spontaneity so pure that thereby he goes far beyond anything that sensibility can ever afford him. (G 4:452; cf., A546/B574; Prol 4:345)
A few sentences earlier Kant had claimed that ‘what reaches our consciousness not by affection of the senses, but immediately’ exhibits a ‘pure activity’ within us, and that such activity indicates that we do not belong wholly to the world of sense. After identifying the source of this activity in our capacity to think ideas of reason, Kant then concludes: ‘On account of this a rational being must view itself, as an intelligence (thus not from the side of its lower powers), as belonging not to the world of sense, but to that of understanding’ (G 4:452). Along these lines it looks like Kant is advancing the following argument: 1. We are conscious of reason’s pure activity within ourselves. 2. This activity secures our membership in the intelligible world. 3. As members of this world we are negatively free. 4. The principle of a negatively free will is a principle of autonomy. 5. The principle of autonomy is equivalent to morality. 6. Therefore, given (1)–(5), morality is valid for us. But there is a missing step here.50 After all, the activity we display in thinking ideas of reason is a theoretical power, and so it does not yet speak to our freedom 49 Ameriks (2003) and Guyer (2009), for example, have pursued a reading of Groundwork III along these lines. 50 Another potential problem with this argument—noted first by Pistorius in his 1786 review of the Groundwork—is that Kant appears to overstep the limits he places on human knowledge by resting his deduction of the categorical imperative on a noumenal claim: ‘[Kant] helps himself to his problematic concept of freedom, transfers us from the world of sense to the world of understanding and brings over from this completely unknown world [völlig unbekannten Welt] the grounds of the possibility and necessity of his categorical imperative’ (Pistorius 1786/1975, 159; emphasis added); see also
92 Objections and Replies of will, i.e., our capacity to act on the basis of reason alone. Even if it is true that as members of the intelligible world we are unaffected by all external influences, this does not yet afford us insight into our capacity for self-legislation. In other words, the spontaneity we display in thinking ideas is not yet a causality; it is not yet a capacity to determine ourselves independently from our impulses and inclinations.51 The above interpretation is incomplete, then, not because the sense of freedom it secures is too weak, namely, freedom to play the game of deliberation, but because the freedom it secures is of the wrong kind, namely, freedom to think ideas.52
3.8.3 A Non-Moral Route Pre-1785 A third and final objection I would like to consider is that my reading lies in tension with the trajectory of Kant’s pre-1785 writings, all of which suggest that he was seek ing a non-moral route to freedom. In the larger context of these writings, it seems that Kant hoped to derive freedom from theoretical reason, either by appealing to the spontaneity of ‘I think’, or by appealing to the spontaneity of the understanding.
First Passage: The Spontaneity of ‘I Think’ The first piece of evidence is from Kant’s lectures on metaphysics during the mid-1770s: When I say, ‘I think, I act, etc.’ either the word ‘I’ is employed improperly, or I am free . . . But now I am conscious to myself that I can say: ‘I do’; therefore I am conscious of no determination in me, and thus I act absolutely freely . . . All practical objective propositions would have no sense if human beings were not free. (V-Met/Heinze 28:269) Pistorius (1786/2007, 16). For a helpful discussion of Pistorius’s charge and its potential impact on Kant, see Ludwig (2010, 2012). See also Schönecker (1999, 2006) and Guyer (2007) for recent refor mulations of this line of criticism. While this topic deserves fuller treatment, I would here like to make one preliminary remark. Even if we agree that Kant was sensitive to Pistorius’s review, it remains an open question whether his reaction was (i) to rewrite the inner structure of his argument in 1788, or (ii) to change its point of focus. The fact that Kant was influenced by Pistorius does not by itself resolve this issue. 51 Later Kant writes that the idea of freedom ‘is valid only as a necessary presupposition of reason in a being that believes itself to be conscious of a will, i.e., of a capacity distinct from a mere faculty of desire (namely, to determine itself to action as an intelligence, hence according to laws of reason inde pendently of natural instincts’ (G 4:459). What is noteworthy here is that Kant does not say the idea of freedom is valid for a being conscious of a capacity to think pure theoretical ideas. Nor does he say the idea is valid for a being conscious of a capacity to act according to prudential imperatives. Instead, Kant is careful to qualify his remark so that the idea of freedom is valid only for a being conscious of a capacity to act according to laws of reason, that is, according to moral laws. 52 One way to avoid this problem is to emphasize the practical law we cognize as members of the intelligible world, as I have in §3.4. However, if we follow this reading, we must give up the claim that Kant was seeking a non-moral route in Groundwork III.
Freedom and Obligation 93 The argument here seems to be (i) that by expressing ‘I think, I do’ we display a pure form of spontaneity, (ii) that this spontaneity proves our freedom from all external determination, and (iii) that we therefore we have a secure foundation for ‘practical objective propositions’, i.e., for moral laws. As Henrich has inter preted this passage, Kant is proceeding from a strictly theoretical starting point. ‘That moral obligation is something real’, Henrich explains, ‘is certain because the consciousness of the self in thought, which alone could doubt that obligation, can be conceived only as freedom. The principle of theoretical reason gives founda tion to the possibility of moral existence.’53 However, other passages from this same lecture show that Henrich’s interpret ation is less decisive than it first appears. At one point Kant says freedom is beyond our comprehension, but that we still have grounds to attribute freedom to ourselves because we are subject to practical imperatives (V-Met/Heinze 28:270). After claiming that ‘all practical objective propositions would have no sense if human beings were not free’, he adds: But now there are such imperatives according to which I should do something; therefore all practical imperatives, problematic as well as pragmatic and moral, must presuppose a freedom in me. (V-Met/Heinze 28:269)
This last point is especially important when we return to the passage Henrich cites, for we can see that Kant is using a distinctly practical premise—the con sciousness of our will under imperatives—to warrant our claim to ‘absolute spon taneity’ (V-Met/Heinze 28:268). What makes Kant’s argument in the 1770s different from his mature position, then, is not that he appeals to the spontaneity of thinking to infer our freedom in the practical sphere, as Henrich and others suppose. Rather, what makes his argument different is that he thinks we have access to our freedom via practical imperatives in general, not via moral impera tives in particular. In these early lectures Kant had yet to see clearly that only a robustly moral point of view affords us insight into the positive character of our freedom.
Second Passage: From Thinking to Willing Another place scholars have thought Kant was seeking a non-moral route to freedom is in his ‘Review of Schulz’ from 1783, around the same time he was composing the Groundwork. Near the end of this review Kant writes that even a strict fatalist
53 Henrich (1960/1994, 80–1). An evident problem with Henrich’s reading is that Kant is not oper ating from a cogito-like premise about the self-certainty of thought or the proposition ‘I think’. Rather, he is operating from a premise about the self-certainty of action expressed in the proposition ‘I do’ (V-Met/Heinze 28:269).
94 Objections and Replies has assumed in the depths of his soul that understanding is able to determine his judgment in accordance with objective grounds that are always valid and is not subject to the mechanism of merely subjectively determining causes . . . hence he always admits freedom to think, without which there is no reason. In the same way he must also assume freedom of the will in acting, without which there would be no morals. (RezSchulz 8:14)
Scholars have read this passage as saying that the spontaneity we exhibit in the theoretical sphere (where we govern our understanding) warrants the assump tion that we are also free in the practical sphere (where we govern our will). However, a few sentences above the cited passage Kant writes that the ‘most con firmed fatalist, who is a fatalist as long as he gives himself up to mere speculation, must still, as soon as he has to do with wisdom and duty, always act as if he were free, and this idea also actually produces the deed that accords with it and can alone produce it’ (RezSchulz 8:13; emphasis added). Kant’s point is that a fatalist must assume that he is free as soon as he becomes conscious of the demands of morality (i.e., ‘wisdom and duty’), for such demands presuppose his independ ence from sensible inclinations. Compared to his lectures on metaphysics, it is clear that by 1783 Kant had attained his mature position—that only moral laws provide the grounds for cognizing our freedom—since he must have realized that prudential imperatives do not presuppose our absolute spontaneity.54 If anything, the passages I have discussed above show that Kant’s commitment to finding moral evidence for freedom is much earlier than commentators have assumed. In addition to the ‘Review of Schulz’ from 1783,55 we find Kant arguing along these lines in the 1781 edition of the first Critique: I assume that there are really pure moral laws, which determine completely a priori (without regard to empirical motives, i.e., happiness) the action and omis sion, i.e., the use of the freedom of a rational being in general, and that these laws command absolutely (not merely hypothetically under the presupposition of other empirical ends), and are thus necessary in every respect. I can legitim ately presuppose this proposition by appealing not only to the proofs of the most enlightened moralists but also to the moral judgment of every human being, if he will distinctly think such a law. (A807/B835)56 54 The view that Kant is moving from the spontaneity of thinking to freedom of will is simply incorrect, since he is really making a claim of comparison, not a claim of inference. Kant is saying that we assume our freedom to think when we govern our understanding according to pure theoretical principles, in the same way that we assume our freedom to act when we govern our will by pure prac tical laws. See Schönecker (2006) for further discussion. 55 1783 is the same year Kant published his Prolegomena, and in this work he draws an explicit connection between freedom and the concept of an ‘ought’ (Prol 4:344–45). 56 Some passages in the 1781 edition indicate that Kant had yet to appreciate the distinctive status of moral imperatives. At times, for instance, he talks as if our faculty of reason displays absolute spon taneity in all ‘ought’ statements, including prudential ones: ‘Whether it is an object of mere sensibility (the agreeable) or even of pure reason (the good), reason does not give in to those grounds which are
Freedom and Obligation 95 By 1781, then, we have evidence that Kant believed a common standpoint confirms the existence of pure practical laws, and that such laws bear upon our freedom as rational beings.57 It is difficult to say exactly when Kant moved toward this view, since the lectures on metaphysics from the mid-1770s indicate that he had yet to see the special status of moral imperatives compared to prudential imperatives. From the documents we have of this time, signs foreshadowing Kant’s mature view begin to appear in the late 1770s, and it is striking how close they come to stating a thesis normally associated with the second Critique. In Reflexion 6849, for example, Kant writes that the ‘primary ought (original = absolute or the universal idea of duty) cannot be comprehended’, adding in Reflexion 6850: ‘The primary ought is a condition under which alone freedom becomes a capacity in accordance with constant rules that determine a priori.’ But even if we treat such literary fragments as inconclusive, the fact remains that Kant’s moral orientation shines through in the early 1780s. And we should not overlook the fact that in the Groundwork itself Kant states unequivocally that reason ‘first becomes aware that it can of itself also be practical’ through ‘the pure thought of duty and in general of the moral law, mixed with no foreign addition of empirical inducements’ (G 4:410; emphasis added). So whether we focus on the official or unofficial documents of this time, there is ample evidence to believe Kant was committed to a moral argument for freedom much earlier than the second Critique.
3.9 Real Differences Considering the evidence laid out in this chapter, we have good grounds to take Kant at his word, as readers did for many decades, when he claimed that the
empirically given, and it does not follow the order of things as they are presented in intuition, but with complete spontaneity it makes its own order according to ideas’ (A547). However, following this pas sage, Kant goes on to say that by virtue of its ideas, reason ‘declares actions to be necessary that yet have not occurred and perhaps will not occur, nevertheless presupposing of all such actions that rea son could have causality in relation to them’. And later, he clarifies that the projected ‘order’ reason fashions for itself is that of a moral world, that is, a world of rational beings ‘insofar as their free choice under moral laws has thoroughgoing systematic unity in itself as well as with the freedom of everyone else’ (A808/B836). In retrospect, Kant may have thought it was sufficient to point to practical impera tives in general to indicate a divide between what ‘is’ and what ‘ought’ to be, all the while knowing that only moral imperatives presuppose our complete spontaneity in the strict (i.e., transcendental) sense of the word. Thanks to Bernd Ludwig for pressing me on this issue. 57 ‘Now that this reason has causality, or that we can at least represent something of the sort in it, is clear from the imperatives that we propose as rules to our powers of execution in everything prac tical. The ought expresses a species of necessity and a connection with grounds which does not occur anywhere else in the whole of nature’ (A547/B575). In an excellent discussion of this passage, Patricia Kitcher (2011) writes: ‘Even in the [first] Critique Kant is thus fairly clear that the serious proof of freedom is provided by the agent’s awareness of the imperatives of morality. This segment [namely, A547/B575] also implies that readers err in following the tendency of the whole passage to assimilate the cases of theoretical and practical reason—since the imperatives of the latter involve a unique kind of necessity’ (2011, 246).
96 Real Differences second Critique ‘presupposes’ the Groundwork in its formulation and justification of the moral principle. In my understanding, the real difference between Kant’s two texts is a difference of focus: the second Critique says little about the formula of the moral law, and even less about how it appears to us as a binding imperative. What I shall now propose is that we can explain this difference of focus in terms of the two genres Kant consciously worked within: classical moral philosophy, on the one hand, and critical transcendental philosophy, on the other. We shall also see that while both the Groundwork and the second Critique unfold in a similar pattern, beginning with the analytic path and ascending to the highest formula of the moral law, the synthetic path of the second Critique is broader in scope. For, in the second Critique Kant proceeds to reveal the moral law’s necessary connec tion to our faculty of feeling—yielding a theory of moral sensibility—the details of which I shall save for the next chapter. Nonetheless, some scholars have argued that the very titles Kant uses in Groundwork III and the second Critique reflect an underlying shift in his project.58 In the former work, he speaks of a ‘critique of pure practical reason’, only to claim in the latter work that we have ‘no need to criticize the pure faculty itself in order to see whether reason is merely making a claim in which it presumptuously oversteps itself ’ (KpV 5:3). Now this would indeed signal a change of mind if Kant believed in Groundwork III that pure practical reason oversteps its boundar ies, and so requires a ‘critique’ in the negative sense of the word. But when we turn to the details of the text, we find Kant speaking of a critique in a positive sense as well. This distinction is implicit in Kant’s explanation of why he has introduced the Idea of an intelligible world: the Idea serves to ‘limit the principle of motives from the field of sensibility by circumscribing this field and showing that it does not include everything within itself but that there is still more beyond it’ (G 4:462; emphasis added). By curbing the pretensions of empirical practical reason (a critique in the negative sense), we can see that practical reason has a pure use as well (a critique in the positive sense). When Kant later speaks of only providing a critique of ‘practical reason’—in justifying the title of his 1788 book— he is clearly speaking of a critique in the former, negative sense. Thus the change in titles alone need not be taken as evidence of any shift. This brings us to my own view. In the Preface to the second Critique Kant tells us his goal is to show ‘that there is pure practical reason’ (KpV 5:3). This differs from his stated goal in the Groundwork, namely, to explain and establish ‘the supreme principle of morality’ (G 4:392). Since the second Critique ‘presupposes’ the Groundwork, as Kant tells us, we should not be surprised to discover abbrevi ated versions of the 1785 work turning up in the 1788 work. Bearing in mind Kant’s different aims, we can see why he would devote an entire section to explaining the moral law as a principle of autonomy in Groundwork II, only to compress this 58 See Bojanowski (2017).
Freedom and Obligation 97 discussion in the second Critique. In the latter work Kant bypasses his complex analysis of the formulas, and proceeds directly to what he calls ‘the fundamental law’ of pure practical reason: ‘So act that the maxim of your will could always hold at the same time as a principle in a giving of universal law’ (KpV 5:30). Similarly, we can also see why Kant would devote an entire subsection to estab lishing the principle of autonomy as a binding imperative in Groundwork III, only to compress this discussion in the second Critique. In the latter work Kant bypasses the details of his earlier deduction, and proceeds to say that the moral law takes the form of an unconditional ‘ought’ for us because we are, unlike holy wills, imperfectly rational (KpV 5:32). My hypothesis, simply stated, is that the Groundwork and the second Critique diverge in their presentations because they belong to genres with different aims.59 On the one hand, the Groundwork was written to prepare for a ‘future metaphys ics of morals’, a project Kant had conceived for many years but only got around to publishing in 1797. What this shows is that Kant’s Groundwork belongs to the tradition of classical moral philosophy: it aims to support a division of duties and a theory of virtue that can inform real decision-making. What Kant must have realized was necessary before the publication of such a system was a clearing of foundations—in particular, a clearing away of popular philosophy—in order to articulate the true method of moral inquiry. On the other hand, the second Critique was written to complete a project Kant had already begun in 1781 with his first Critique. What this shows in turn is that the second Critique belongs to the genre of Kant’s own critical philosophy: it aims to show that pure reason, though dialectical in the speculative sphere, has a legitimate use in the practical sphere. From the perspective of genres Kant’s brevity in the second Critique makes perfect sense. He did not feel compelled to offer a long analysis of the cat egorical imperative, nor a deduction of its possibility, because his aim was no longer to prepare for a future metaphysics of morals. His aim was to render the transition from pure speculative to pure practical reason more perspicuous, and to that end, his first objective was simply to show that pure practical reason exists. The real difference between the Groundwork and the second Critique also comes into focus when we turn our attention to Kant’s brief, but very informative, remarks about his methodology. It is worth recalling that Kant speaks of following an analytic path first, followed by a synthetic path (G 4:392), a distinction we find echoed in the Preface to the second Critique when Kant describes a ‘synthetic return’ (synthetische Wiederkehr) to what had ‘previously been given analytically’ (KpV 5:10). Let us pause to consider what Kant says in the second Critique more carefully: When it is a matter of determining a particular faculty of the human soul as to its sources, its contents, and its limits, then, from the nature of human cognition, 59 See Hill (1998).
98 Real Differences one can begin only with the parts, with an accurate and complete presentation of them. But there is a second thing to be attended to, which is more philosophic and architectonic: namely, to grasp correctly the idea of the whole [die Idee des Ganzen] and from this idea to see all those parts in their reciprocal relation [wechselseitigen Beziehung] to each other by means of their derivation from the concept of that whole in a pure rational faculty. (KpV 5:10)60
Kant adds, however, that this ‘examination and guarantee is possible only through the most intimate acquaintance with the system’, so that those ‘who find the first inquiry too irksome’ will never arrive at ‘the second stage, namely the overview, which is a synthetic return to what had previously been given analytically’ (KpV 5:10).61 Upon comparing this passage with Kant’s stated methodology in the Groundwork, the following points come to light. Both the Groundwork and the second Critique begin by ascending along the analytic path, culminating in autonomy as the highest formula of the moral law. In pursuing this path we see that Kant is careful to separate the faculty of practical reason into its pure and empirical uses, since it is only on the basis of this separation that we can discover the law of pure practical reason. This accords with Kant’s chemistry metaphor discussed in the previous chapter, whereby the philosopher begins by distinguish ing morality and happiness as if they were two unmixable compounds. And let us not forget that a similar procedure lies at the heart of Groundwork III when Kant offers us two ‘standpoints’ for viewing our faculty of will, one empirical (in rela tion to sensibility) and the other pure (in relation to reason). Such is the separ ation that even the scoundrel draws when he reflects upon examples of virtuous conduct; and such is the separation Kant illustrates for us again, in the second Critique, with the thought experiment of a man facing a difficult decision between duty and death. The result of each analysis, then, is that we are immediately conscious of the separation of morality and happiness and of the former’s normative authority. Yet
60 Kant makes a similar point in the first Critique, writing that it is ‘first possible for us to glimpse the Idea in a clearer light and to outline a whole architectonically, in accordance with the ends of rea son, only after we have long collected relevant cognitions haphazardly like building materials and worked through them technically with only a hint from an idea lying hidden within us’ (A 834–5/B 862–3). These systems, Kant goes on to say, appear to be ‘the mere confluence of aggregated concepts, garbled at first but complete in time, although they all had their schema, as the original seed, in the mere self-development of reason, and on that account are not merely each articulated for themselves in accordance with an idea but are rather all in turn purposively united with each other as members of a whole in a system of human cognition, and allow an architectonic to all human knowledge’ (A 834–5/B 862–3). 61 To this long passage Kant adds the following, if cryptic remark: that ‘it is no wonder that they [i.e., the critical reviewers of the Groundwork] find inconsistencies everywhere, although the gaps they suppose they find are not in the system itself but only in their own incoherent train of thought’ (KpV 5:10).
Freedom and Obligation 99 on further reflection it becomes clear that the arc of the analytic/synthetic paths differs in each text. The Groundwork has a longer way of ascent, since it is con cerned first and foremost with the correct articulation of the highest principle of morality (Sections I and II), before showing how this principle is actually binding for a human will (Section III). In presupposing this, the way of ascent in the sec ond Critique is relatively short: as I have shown, Kant gives us abbreviated ver sions of his arguments for why autonomy is the highest principle of morality and why this principle takes the form an ‘ought’ for us. By contrast, the way of descent in his later work turns out to be much longer and more elaborate than in Groundwork III, since Kant is not simply concerned to ‘recombine’ the two per spectives on our will (the empirical and the pure) in order to show how a categor ical imperative is possible. In fact, Kant’s path of justification in the second Critique is much broader in scope than commentators have traditionally thought. In this later text, the ‘synthetic return’ involves showing how all the parts of our faculty of practical reason, including our faculty of feeling, constitute a unity under the idea of that faculty as a whole.
3.10 Closing Remarks The far-reaching implication that is now open for us to explore is that Kant’s doctrine of the fact of reason from the first chapter of the second Critique is only the beginning of his project of moral justification. The first chapter performs the requisite separation of practical reason into its empirical and pure uses from a common standpoint, i.e., as a Factum prior to all speculation about its possibility. But the success of Kant’s project requires, not a further analysis of the faculty of practical reason into its parts, but rather a synthesis of those parts under the idea of that faculty as a whole. That is why, after Kant presents the fact of reason in the first chapter, he goes on to develop two further accounts, one devoted to judg ment (the second chapter), and the other devoted to feeling (the third chapter). It is the latter chapter in particular that reveals a ‘reciprocal interaction’ between what initially looked like two heterogeneous elements of practical reason, i.e., the reciprocal interaction of reason and sensibility in the moral feeling of ‘respect’ (Achtung). Looking ahead to the reading I plan to defend, Kant’s theory of moral sensibility from the third chapter plays an essential role in Kant’s project by showing that the moral law actually binds our ‘hearts’ and not just our ‘heads’— which is to say, it shows that the moral law is a real motivating incentive to action. If we are to understand the full scope of Kant’s project of moral justification, then, we must move our investigation into this new territory, which is my task for the next chapter.
4
Moral Sensibility Nobody can or ever will comprehend how the understanding should have a motivating power; it can admittedly judge, but to give this judgment power so that it becomes a motive able to impel the will to performance of an action—to understand this is the philosopher’s stone. – Lectures on Ethics (V-Mo/Collins 27:1428)
4.1 Introduction In the previous chapter I showed why there is no reason to think that Kant’s doctrine of the fact of reason abandons his earlier project. I argued that even in Groundwork III we find Kant invoking a normatively thick premise, the idea of the intelligible world, to secure our possession of a pure will. I then concluded by sketching the real differences between the Groundwork and the second Critique, focusing once again on Kant’s stated methodology in the prefaces of each text. What I proposed is that the Groundwork and the second Critique unfold along the same trajectory, ascending along the path of analysis and descending along the path of synthesis; but there is a contrast in their respective arcs. This now brings me to the next main thesis of the present study: the synthetic path of the second Critique is broader in scope, since Kant seeks to reveal a necessary connection between our consciousness of the moral law and our capacity to feel pleasure and displeasure. In this chapter my aim is to show how Kant’s theory of moral sens ibility, far from constituting an addendum to the project of justification in the second Critique, forms an essential step in its completion. For when we examine the details of this theory, we shall see that Kant is returning to the same problem of a natural dialectic that animated the Groundwork—the problem of our tendency to rationalize against the claims of morality—except now he is investigating that problem at a deeper level, tracing its origin to our development as rational and sensible beings. Apart from how one understands the relationship between Groundwork III and Kant’s doctrine of the fact of reason, whether in terms of reversal or continuity, my claim in this chapter bears significance for rethinking the status of Kant’s
Kant’s Justification of Ethics. Owen Ware, Oxford University Press (2021). © Owen Ware. DOI: 10.1093/oso/9780198849933.003.0005
Moral Sensibility 101 account of moral motivation. After providing a survey of how this account evolved over the course of Kant’s career—beginning with strong ties to the British sentimentalist school—I shall explain how the concept of moral ‘respect’ (Achtung) factors in the Groundwork. As we shall see, Kant observes in Groundwork III that moral motivation requires a ‘causality of reason to determine sensibility in conformity with its principles’, one in which reason ‘infuses’ a feeling of pleasure in the fulfillment of duty (G 4:460). Yet he claims that this goes beyond sense experience and so falls outside the scope of his project. In the second Critique, however, while Kant does not pretend to give such causal insight, he does offer a phenomenological study of what effects our consciousness of the moral law must have upon our capacity to feel. As part of the synthetic path, which aims to reveal the pure and empirical parts of practical reason in their reciprocal interaction, Kant’s new method serves to justify the moral law’s applic ability to our faculty of sensibility, thereby showing that the moral law binds our hearts and not just our heads.
4.2 Background: The ‘Philosopher’s Stone’ While it may be an exaggeration to say Kant went through a sentimentalist phase, in the ‘Prize Essay’ of 1762 he does claim that our judgments of the good derive from more basic, unanalyzable feelings; and he praises Francis Hutcheson for ‘providing us with a starting point from which to develop some excellent observations’ under the name ‘moral feeling’ (DG 2:300). Yet the essay ends on a tentative note. Kant says we do not yet know ‘whether it is merely the faculty of cognition, or whether it is feeling (the first inner ground of the faculty of desire) which decides its first principles’ (DG 2:300). Kant seems to lean toward a sentimentalist view in his Observations on the Feeling of the Beautiful and Sublime, published in 1764, when he says that virtue arises from a ‘consciousness of a feeling that lives in every human breast’, that is, ‘the feeling of the beauty and the dignity of human nature’ (GSE 2:217). But right before saying this he asserts that virtue must be based on ‘principles’. This indecisiveness runs throughout his other writings of the period.1 The only place where Kant appears to entertain a strong sentimentalist view is in an unpublished note
1 For example, in his ‘Announcement’ to the lecture program he designed for the Winter Semester of 1765–1766, Kant writes: ‘The distinction between good and evil in actions, and the judgment of moral rightness, can be known, easily and accurately, by the human heart through what is called sentiment, and that without the elaborate necessity of proofs’ (NEV 2:311). But this does not commit him to the view that sentiment grounds moral principles, only that the conclusions of such principles can be known ‘by the heart’. This may not be that different from Kant’s mature view that moral principles are implicit in common human reason.
102 Background: The ‘ Philosopher ’ s Stone ’ written sometime around 1764–68: ‘The rules of morality proceed from a special, eponymous feeling, upon which the understanding is guided’ (Refl 19:93). But even this is unclear. Does he mean moral feeling grounds those rules, or that it simply guides their execution? It is difficult to say how we should read these early writings. There are points of agreement between Kant’s early and mature ethical views, although I think it is important to highlight the uncertainty of his thinking prior to the 1780s. His praise of Hutcheson in the ‘Prize Essay’, to take one example, does not clarify his agreement with the sentimentalists. Nor do the other remarks I have cited. My view is that during the 1760s Kant saw himself building on the ideas of the British, rather than defending their views. He indicates as much in his ‘Announcement’ to the lecture series of 1765–66, where he says that Shaftesbury, Hutcheson, and Hume have ‘penetrated furthest in the search for the fundamental principles of all morality’, though their search is ‘incomplete and defective’. There he promises to supply in his lectures what the sentimentalists lack: ‘precision and completeness’ (NEV 2:311). However we read these remarks, it is evident that Kant changed his focus between the publication of the ‘Prize Essay’ and the Inaugural Dissertation (1770).2 In the latter work we do not hear any praise of the sentimentalists. Kant does not even mention Hutcheson (or Shaftesbury or Hume, for that matter), and his brief remark on the foundation of moral principles indicates a firm break from their thinking: ‘moral concepts’, he writes in section 7, ‘are cognized not by experiencing them but by the pure understanding itself ’ (MSI 2:395). Other reflections from the late 1760s foreshadow Kant’s emerging view that moral prin ciples are grounded in reason, and that moral feeling is an effect of those prin ciples, not their foundation.3 In a letter to Marcus Herz, dated toward the end of 1773, Kant makes his new position more clear: The highest ground of morality must not simply be inferred from the pleasant; it must itself be pleasing in the highest degree. For it is no mere speculative idea; it must have the power to move. Therefore, though the highest ground of morality
2 What brought about this shift? In his essay ‘Kant and Greek Ethics’, Klaus Reich (1939) gives us the following date: When Kant read the Phaedo, the dialogue in which Socrates famously denounces moral systems grounded in sense rather than reason, which Reich believes Kant must have read before embarking on the Dissertation. Reich speculates the Phaedo must have impressed Kant, not only because it criticized a sentimentalist leaning he may have felt at the time, but because Mendelssohn, in his introduction to the 1767 German translation, made a persuasive case that the Phaedo represents Plato’s own thinking. This is an interesting suggestion, but for lack of space I cannot pursue it here. 3 Also compare the following unpublished fragment dated from 1772: ‘Moral feeling succeeds the moral concept, but does not produce it; all the less can it replace it, rather it presupposes it’ (Refl 19:150).
Moral Sensibility 103 is intellectual, it must nevertheless have a direct relation to the incentives [Triebfedern]4 of the will. (Br 10:145)5
Kant would spend the next eleven years struggling to formulate this ‘direct relation’. Although the details of his view changed, Kant never departed from his nonsentimentalist view that moral feeling follows from moral judgment. As he puts it: ‘When I judge by understanding that the action is morally good, I am still very far from doing this action of which I have so judged. But if this judgment moves me to do the action, that is the moral feeling’ (V-Mo/Collins 27:1428). The under lying puzzle was to explain how a principle of reason can relate to the will as an incentive, that is, how it can move us to act without depending on ordinary feelings of pleasure or pain. For years this puzzle appeared to Kant as if under a veil of mystery, so much so that he is reported to have said, during lecture, that its solution would be the ‘philosopher’s stone’ (V-Mo/Collins 27:1428).
4.3 Moral Feeling in the Groundwork With the publication of the Groundwork, Kant appears to have abandoned the idea that moral feeling is necessary for moral motivation.6 Now the topic of action ‘from duty’ (aus Pflicht) takes on central importance. In Groundwork I, for example, Kant argues that our actions possess moral worth only if they are performed for the sake of duty, and from this he concludes that moral action must arise from a maxim that has ‘respect’ as its basis. Kant’s argument for how respect functions as a ‘feeling’ (Gefühl) is tucked away in a single footnote—at G 4:401n— and it is here that we find him briefly returning to the problem of the philosopher’s stone. In this footnote Kant begins by repeating a claim from his early writings, that moral feeling is an effect of our consciousness of the moral law and not its ground. ‘Respect’, he says, is ‘the effect of the law on the subject, not the cause of the law’ (G 4:401n). Working this claim further, Kant says that respect is not a feeling ‘received by means of influence’ but is rather ‘self-generated by means of a rational concept’: it arises when we apprehend a worth that ‘infringes upon’ our self-love 4 None of the English translations of ‘Triebfeder’ are satisfying. Stephen Engstrom makes a good case for ‘driving spring’ (2010, 92), but I will follow Mary Gregor in using ‘incentive’. 5 Kant makes a similar claim in an earlier note around 1769–1770. ‘The immediate good can be found only in freedom’, he writes. ‘For, because freedom is a capacity for action, even if it does not please us, freedom is not dependent upon the condition of a private feeling; however, it always refers only to that which pleases, so it has a relation to feeling and can have a universally valid relation to feeling in general’ (Refl 19:103). 6 Due to restrictions of space, I will not take my discussion of moral sensibility beyond the second Critique, although I think there are developments worth exploring in Kant’s writings from the 1790s. For discussion of these later texts, see Guyer (2010) and Grenberg (2001).
104 Moral Feeling in the Groundwork (G 4:401n).7 To this extent, ‘there is something that is regarded as an object neither of inclination nor of fear, though it has something analogous to both’. Clarifying this remark, Kant adds: The object of respect is therefore simply the law, and indeed the law that we impose upon ourselves and yet as necessary in itself. As a law we are subject to it without consulting self-love; as imposed upon us by ourselves it is nevertheless a result of our will; and in the first respect it has an analogy with fear, in the second with inclination. (G 4:401n)8
While this passage is open to interpretation, Kant may be saying that one side of respect is like fear because we regard anything that opposes our self-love with displeasure. This feeling would be ‘pathological’ (i.e., based in sensibility) because it would occur relative to our inclinations. Yet this answer is incomplete, as we can see, because a pathological feeling—the pain of having our self-love infringed upon—cannot explain what moves us to act from duty. Such pain is not related internally to our consciousness of the moral law itself. To what extent, then, can the feeling of respect make a positive contribution to the process of moral motivation? To make the problem clear, consider the following analogy. Suppose I am driving and feel the desire to speed, and suppose I see a police officer and decide to slow down.9 What can we say is motivating me in this scenario? From a prudential point of view, it may be that I would rather deny myself the pleasure of speeding than suffer the consequences of getting a ticket. In making this judgment, I will likely experience something like pain in having my desire thwarted, and for good reason: I am prevented from doing what I want. Notice, however, that my initial pain-like feeling is not related internally to my thought of wanting to avoid a traffic fine. I am motivated by a desire to avoid what I judge would be the greater displeasure of paying a fine, relative to the momentary pleasure of indulging a desire. Thus, the pain I feel in having to restrain my impulse cannot serve as my incentive to act (namely, to slow down) because it comes after I have already formed a decision about what to do. Such pain comes too late to explain the structure of non-moral motivation, i.e., motivation based on comparing degrees of pleasure and displeasure. 7 Werner Pluhar translates Abbruch tut as ‘impairs’ for symmetry with the violence invoked by ‘striking down’ (Nierderschlagung). I prefer ‘infringes upon’, but in §4.5 I will say why the Nierderschlagung of self-conceit is best understood as a kind of disillusionment. The battle metaphor is limited, and potentially misleading. 8 Kant contrasts the feeling of fear with an inclination; but as Paton observes, an inclination is not a feeling (1947, 64). When discussing the ‘matter’ of the faculty of desire in the second Critique, Kant speaks of ‘objects of inclination, whether of hope or fear’ (KpV 5:74). I believe this distinction is more consistent with Kant’s view in the Groundwork: hope refers to an object of pleasure, fear to an object of pain. 9 See Herman (2007, 19).
Moral Sensibility 105 Generalizing from this example, we can now ask whether our consciousness of the moral law is only ever like this, whether it only ever effects a pain-like feeling when we restrain our egoistic tendencies of choice. If so, it is unclear how the affective dimension of respect, even if in some sense inevitable, could play a positive role in moral motivation proper.10 To be sure, in the footnote from G 4:401 Kant also says that respect has an analogy with inclination: it is a pleasurable feeling we have when we discover the moral law is ‘imposed upon us by ourselves’ and hence ‘a result of our will’. But this remark only serves to invite the same question: Might such a pleasurable feeling not come too late to explain the structure of action from duty? The general worry here is not whether reason has an effect on sensibility—we can grant that it does—but whether it can establish the right kind of motivational link to choice and action. By the right kind of motiv ational link I mean one that speaks to an internal connection between our consciousness of the moral law (or of what we judge we ought to do for moral reasons) and the effect this judgment has on our capacity to feel. For if this latter effect is only ever rooted in pathological feelings of pleasure and pain, as the two examples just sketched indicate, it is not clear how pure practical reason and our faculty of sensibility could ever positively interact. Before going forward, it is important to see that this version of motivational skepticism under discussion differs from its classical version, famously voiced by Hume, that reason can never ‘immediately prevent or produce any action’.11 As I remarked earlier, Kant is clear in Groundwork III that skepticism about the causal efficacy of reason poses no genuine threat to his moral philosophy. ‘In order for a sensibly affected rational being to will that for which reason alone prescribes the ought’, he writes, ‘it is admittedly required that his reason have the capacity to infuse [einzuflößen] a feeling of pleasure or of delight in the fulfillment of duty, and thus there is required a causality of reason to determine sensibility in conformity with its principles’ (G 4:460). But this presupposes a causal connection entirely unknowable to us, a connection between a noumenal cause and a phenomenal effect in feeling. So it is impossible to see, Kant tells us, ‘how a mere thought which itself contains nothing sensible produces a feeling of pleasure or displeasure’ (G 4:460). Our ability to explain causal relations is restricted to objects of possible
10 Versions of this problem were voiced by Hegel and Schiller, and more recently by Williams and Henrich. To the best of my knowledge, Beck is the only commentator to recognize the significance of what I am calling motivational skepticism. He writes: ‘Though it displaces the inexplicable mystery of man by only one step, the Critique of Practical Reason [in the third chapter] does attempt at an explan ation, in psychological terms, of how the knowledge of the moral law can be effective in the determin ation of conduct . . . It is essential that this mystery be removed from the phenomenological surface, as it were, for the thing is so puzzling that doubts of its reality can have the actual effect of reducing the effectiveness of this incentive’ (1960, 210–11). However, Beck does not say anything beyond this to show how Kant succeeds in removing the mystery. 11 See Hume’s Treatise (1739/1978, §3.1.1). Christine Korsgaard also discusses this kind of skepticism in connection with Kant and Bernard Williams (Korsgaard 1996b).
106 Moral Feeling in the Second Critique experience, and that absolves us of any obligation to explain the noumenal causality of moral feeling. In saying this, however, Kant wants to avert any doubts we might have about the source of moral feeling. If we had to explain this source, from a third-person perspective, it seems we would be defenseless against those who deny reason can produce a motivationally effective feeling in us. Yet Kant’s point is just the reverse. We cannot know the source of moral feeling because it rests on the assumption of a noumenal cause, and that means our ignorance here is unavoidable. We are thereby entitled to what Kant elsewhere calls a ‘polemical method’: we can defend the possibility of moral motivation against those who presume to have deeper insight than we do into the nature of things (G 4:459).12 We can show that the skeptic is not entitled to reject the concept of moral feeling because our investigation has reached the ‘highest limit of all moral inquiry’.13 My point, however, is that a different kind of motivational skepticism surfaces when we consider the effects our consciousness of the moral law has on sensibility; for then we can ask, not how the cause of moral feeling is possible, but how the effects themselves are suited to motivate us from a deliberative point of view.
4.4 Moral Feeling in the Second Critique Before turning to Kant’s effort to answer this latter question, one point of clarification is in order. In the Preface to the second Critique, after dismissing the crit ical reviews of the Groundwork, Kant says ‘a further objection could have been raised’ (KpV 5:9n).14 Here he tells us one could have criticized the Groundwork for 12 ‘Now by the polemical use of pure reason I understand the defense of its propositions against dogmatic denials of them. Here the issue is not whether its own assertions might perhaps also be false, but only that no one can ever assert the opposite with apodictic certainty (or even only with greater plausibility). For in this case we do not hold our possession merely by sufferance if we have a title to it, even if not a sufficient one, and it is completely certain that no one can ever prove the unlawfulness of this possession’ (A739–40/B767–8). 13 It is worth bearing in mind that Kant’s central aim in the Groundwork is to clarify the principle of morality and establish its authority as a categorical imperative. Within this context it is not necessary to work out all the subtleties of a critique of practical reason, including the question of how reason can be practical for sensible beings like ourselves. So understood, the task Kant assigns to the footnote from G 4:401 is straightforward. He is speaking not so much to a critic who demands a fuller account of the feeling of respect, owing to the question of whether respect is suited to the role of moral motiv ation. Rather, he is speaking to someone who is liable to assume that feelings are not only obscure but also in every case ‘received by means of influence’ and so tied to inclination or fear. In this way Kant’s treatment of moral feeling in the Groundwork, while brief, serves its role for the book. 14 In the third Critique, Kant says explicitly that respect is the only feeling of which we have a priori cognition. ‘To be sure’, he explains, ‘in the Critique of Practical Reason we actually derived the feeling of respect (as a special and peculiar modification of this feeling, which will not coincide exactly either with the pleasure or with the displeasure that we obtain from empirical objects) from the universal moral concepts a priori. But there we could also step beyond the bounds of experience and appeal to a causality that rests on a supersensible property of the subject, namely that of freedom’ (KU 5:222). I will follow this clue in my reading of the second Critique to follow.
Moral Sensibility 107 not defining the concept of pleasure. In response, Kant says that pleasure is the ‘representation of agreement of an object or of an action with the subjective conditions of life’ (KpV 5:9n).15 As this definition makes clear,16 every feeling has an intentional object (perceived by the senses or imagined in thought) as well as an accompanying representation of the object’s agreeableness or disagreeableness to the subject.17 This makes every feeling self-reflexive in character: ‘pleasure’ is the representation of agreeableness-to-oneself; ‘pain’, the representation of disagreeableness-to-oneself. Now it is important to see that nothing about this definition commits us to saying that what we find agreeable or disagreeable bears upon our sensible nature. Kant even makes this explicit. ‘It is easily seen’, he writes, ‘that the question whether pleasure must always be put at the basis of the faculty of desire or whether under certain conditions pleasure only follows upon its determination, is left undecided from this exposition’ (KpV 5:9n; emphasis added).18 One noteworthy feature of Kant’s definition at KpV 5:9n is that feelings are treated as non-cognitive representational states. They are non-cognitive, on the one hand, because they do not subsume an object under a concept, as with judgments. On the other hand, feelings are still representational because they highlight the suitableness or unsuitableness of an object to the acting subject. In the practical domain, then, sensibility and understanding are distinct faculties because the former involves representations of what is ‘agreeable’ or ‘disagreeable’ and not full-fledged judgments of what is ‘good’ or ‘bad’. This distinction is important for showing the extent to which representations of agreeableness or disagreeableness are bound up with the perspective of the agent. Although it is common in the empiricist tradition to speak of feelings as ‘strong’ or ‘weak’ forces, from a third-person point of view, that is not the framework of explanation Kant 15 The complaint would be ‘unfair’, Kant thinks, because his account of these concepts in psych ology ‘could be reasonably presupposed’. Presumably Kant is referring to the chapter on psychology offered in his Lectures on Metaphysics. See, for example, V-Met/Mron 29:891. See also Wuerth (2013, 158) for further discussion of the footnote from KpV 5:9. 16 Kant calls it a transcendental definition in the first Introduction to the third Critique. He writes: ‘It is useful to attempt a transcendental definition of concepts which are used as empirical principles, if one has cause to suspect that they have kinship with the pure faculty of cognition a priori. One then proceeds like the mathematician, who makes it much easier to solve his problem by leaving its empir ical data undetermined and bringing the mere synthesis of them under the expressions of pure arithmetic’ (KU 20:230). He also says we can think of empirical concepts in this way if we recognize their ‘kinship’ or ‘affinity’ with reason (cf., KU 5:177n). See also Kant’s marginal note on transcendental definitions (at A538/B566) in his personal copy of the first Critique (pages 535–536 in the Guyer/ Wood translation). 17 This claim is not unique to the second Critique. Consider the following notes recorded by Kant’s students in the 1770s: ‘But what is a feeling? That is something hard to determine. We sense ourselves . . . The subjective representation of the entire power of life for receiving or excluding objects is the relation of satisfaction or dissatisfaction. Thus feeling is the relation of objects not to the representation, but rather to the entire power of the mind . . . Feeling consists in the relation of a representation not to the object, but to the entire subject. Pleasure and displeasure are not cognitions at all’ (LM 28:586; 28:247). For an excellent discussion of pre-Kantian views of feeling, see Thiel (1997). 18 For further discussion of Kant’s concept of feeling, see Ware (2010), Geiger (2011), Dewitt (2014), Newton (2017), Cohen (2020), and Tizzard (2018).
108 Moral Feeling in the Second Critique deploys in the second Critique. Feelings, for Kant, are essentially first-personal in character: the content of what is agreeable-to-oneself or disagreeable-to-oneself is only meaningful from the standpoint of the agent for whom objects either agree or disagree. Another way of expressing Kant’s position is to say that feelings are agentfocused and not object-focused. To take a simple case, suppose I find swimming pleasurable. Here I can say that my feeling represents the activity, the object of my feeling, as agreeable-to-me. If I then determine my will by this representation, i.e., by seeking to realize the activity, then I transform my feeling into a ‘desire’ (Begehren): namely, a desire to swim. On one level the activity of swimming must be an object of possible experience, which is to say that I must perceive the activity or represent it in thought; but my feeling of pleasure for it is subjective. This is why the capacity to feel is not like sensibility in general, that of receiving impressions through the mode of outer sense (space) or inner sense (time).19 Kant emphasizes this when he says that sensibility in the practical domain ‘is not regarded as a capacity for intuition at all but only as feeling (which can be a sub jective ground of desire)’ (KpV 5:90). A feeling of agreeableness, whether in relation to an activity like swimming or to the claims of morality, concerns me—how I represent these things—and not to anything outside of me.20 Since this definition of feeling is implicit in the analysis of respect from the Groundwork, Kant is right to say it would be ‘unfair’ to accuse him for not defining pleasure or related concepts like desire (KpV 5:9n). However, I think it is clear that the definition of feeling from the Preface of the second Critique allows Kant to approach the question of moral feeling from a new perspective. To see why, we need only turn to consider his opening statement in the third chapter: Nothing further remains than to determine carefully in what way the moral law becomes the incentive and, inasmuch as it is, what happens to the human faculty of desire as an effect of that determining ground upon it . . . What we shall have to show a priori is, therefore, not the ground from which the moral law in itself supplies an incentive but rather what it effects (or, to put it better, must effect) in the mind insofar as it is an incentive. (KpV 5:72; emphasis added)21 19 As Kant explains in the first Critique: ‘The pleasant taste of a wine does not belong to the object ive determinations of the wine, thus of an object even considered as an appearance, but rather to the particular constitution of sense in the subject that enjoys it’ (A28). 20 Kant is sensitive to the possibility that feeling may only have private validity for the subject. ‘If the object agrees only with the state of the subject, then it cannot pleasure universally, but rather according to the private satisfaction of the subject’ (LM 28:253–4). ‘Any feeling has a private validity only, and is not accessible to anyone else, and is also in itself pathological; if someone says he feels it so in himself, that cannot hold good for others, who do not even know how he is feeling, and once anyone appeals to a feeling he is giving up all grounds of reason’ (LE 27:276). At KpV 5:9n, however, Kant wants to leave open the possibility of feeling connecting to pure practical reason. 21 In the third chapter Kant uses the term ‘a priori’ to mean ‘in advance of ’ particular empirical experiences. So defined, his aim is to see the effects our consciousness of the moral law must have on
Moral Sensibility 109 On first glance this passage may seem to conflict with Kant’s claim from Groundwork III: that we cannot comprehend ‘how a mere thought which itself contains nothing sensible produces a feeling of pleasure or displeasure’ (G 4:460). Yet on closer inspection we can see that Kant is bracketing the question of reason’s causal efficacy, from a third-person perspective, focusing instead on the question of what the moral law must feel like, from a first-person perspective. In this way Kant is building upon the doctrine of the fact of reason discussed in Chapter 3; but rather than limit his analysis to our consciousness of moral necessity, as it manifests itself in common human reason, Kant’s concern in the third chapter of the second Critique is how this consciousness positively interacts with our faculty of sensibility. To be clear, this does not give us insight into the ground of moral feeling as a noumenal cause—that remains beyond our epistemic reach—but it does make room for what I am calling Kant’s phenomenological method.22 As we shall see, this method allows Kant to describe how reason and sensibility can be brought into reciprocal interaction with each other.
4.5 Self-Love and Self-Conceit The third chapter is titled ‘On the Incentives of Pure Practical Reason’, and it begins with what is now a familiar point. ‘What is essential to any moral worth of actions’, Kant writes, ‘is that the moral law determine the will immediately’ (KpV 5:71). The implications of this claim are relatively clear, although hardly unproblematic. Kant is saying that if any sensible incentive slips into the determining ground of my choice, then my ensuing action cannot possess inner worth. It does not matter whether my ensuing action outwardly conforms to the moral law, for that would only amount to a show of good will, what Kant calls ‘legality’. If my conduct is not wholly inspired by the thought of duty—if, for example, I am moved by hope of reward or fear of punishment—then we cannot say I have acted in the spirit of ‘morality’. These are debatable claims, of course, but for our purposes it only matters what Kant wants to draw from them. What he wants to draw from them, starting at KpV 5:73, is a number of a priori arguments that purport to explain the effects the moral law must have on our capacity to feel. Off hand this task might strike us as unusual, for when we speak of ‘pleasure’ or ‘displeasure’ we typically have in mind feelings that are unique to one person. We think the connection between feelings
our capacity to feel, irrespective of anyone’s particular psychological constitution. This is important for understanding the status of Kant’s phenomenological method, as I will discuss in §4.5. 22 As we shall see, focusing on this aspect of the third chapter opens up a framework of explanation missing from the debate between intellectualist and affectivist readings of Kant.
110 Self-Love and Self-Conceit and their objects is only ever contingent, arising from someone’s particular psychological or physiological characteristics. For Kant, however, it is crucial that we raise the question of moral feeling from a different point of view. For he wants to describe how the moral law operates as an incentive to action by studying the effects it must have on all human rational agents, irrespective of anyone’s particular sensibility.23 Guided by this question in the third chapter, Kant begins by advancing two claims. In the first place, he argues that when we become conscious of the moral law we must feel a limitation placed on the concern we have for our happiness (KpV 5:73). As a result, our consciousness of the moral law must ‘infringe upon’ self-love, restricting the importance we attach to the satisfaction of our sensible needs and inclinations. This is the first effect the moral law must have on our faculty of sensibility, but it is not the only effect. ‘Pure practical reason merely infringes upon self-love [Eigenliebe]’, Kant explains. But it also ‘strikes down’ the tendency we have to inflate the value of our needs and inclinations, what he calls ‘self-conceit’ (Eigendünkel) (KpV 5:73). In the second place, then, we must experience something like ‘humiliation’ (Demütigung) when we realize we have ascribed authority to our happiness, acting as though it were a lawgiving principle. Since neither of these feelings presuppose subjective traits peculiar to any one person, they are ‘a priori’ in Kant’s sense of the term, that is, they are feelings we can cognize independent of experience. To make this distinction clear, I will hereafter speak of the pain of self-love (‘painsl’) and the pain of self-conceit (‘painsc’).24 One difficulty, however, is that Kant does not say what he means by self-conceit or why it is a ‘propensity’ (Hang). But we can, I believe, fill in these missing details with the benefit of hindsight. In his 1793 book on religion Kant defines a ‘propensity’ as the ‘subjective ground of the possibility of an inclination (habitual desire, concupiscentia)’ (R 6:28; cf., MS 6:212; Anth 7:251). The Latin gloss indicates he is following a broadly Scholastic theory according to which concupiscentia is the capacity of a sinful desire, as when Augustine speaks of people who have a ‘weakness for wine’ (cf., R 6:28n). Of course, Kant is not speaking of self-conceit in the colloquial sense, as a character trait certain individuals display, like arrogance or pride. Rather, he thinks of it as a tendency of choice all rational human beings possess. The point of the Latin gloss, I take it, is that self-conceit is a ‘propensity’ in the sense that we author it prior to any physical deed, in the same way that one’s susceptibility to alcohol precedes one’s addiction to it.25 23 Remember what Kant says, that the task of the third chapter is ‘to determine carefully in what way the moral law becomes the incentive and, inasmuch as it is, what happens to the human faculty of desire as an effect of that determining ground upon it’ (KpV 5:72). 24 Kant’s concept of self-conceit is complex and I am only able to offer a sketch here (what I am calling the developmental theory). For two thoughtful and insightful treatments of this topic, see Moran (2014) and Russell (2020). 25 Kant repeats this point in the Metaphysics of Morals: ‘Concupiscence (lusting after something) must also be distinguished from desire itself, as a stimulus to determining desire. Concupiscence is
Moral Sensibility 111 This is somewhat confusing, however. Kant maintains that self-conceit is rooted in our power of rational choice, but he also says that what it aspires to is irrational, all things considered. In self-conceit we confer value onto the inclinations in such a way that the demands of our sensible nature appear to have the authority of law. As Kant puts it, ‘If self-love makes itself lawgiving and the unconditional practical principle, it can be called self-conceit’ (KpV 5:74). In this way self-conceit involves an illegitimate promotion of self-love, our natural tendency to care for the inclinations, to a position of first rank in the will. Yet this brings us to ask, Why do we have a tendency to do this? Why do we freely orient our power of choice in a way that makes the pursuit of happiness lawgiving? To begin a reply to this question, consider what Kant says in the following passage: A human being is a being with needs, insofar as he belongs to the sensible world, and to this extent his reason certainly has a commission [Auftrag] from the side of his sensibility which it cannot refuse, to attend to its interest and to form practical maxims with a view to happiness in this life and, where possible, in a future life as well. (KpV 5:61)
What this passage suggests is that both self-love and self-conceit emerge from the way practical reason develops within us as sensible rational beings. Before we are aware of the moral law, Kant is saying, all we have to attend to are sensible needs and inclinations. They present themselves to us as demands, and so seem to be objective. Once we form an idea of their total satisfaction, we not only think that our happiness should have preference to any temporary pleasure of the moment (self-love); we also think that our happiness should have preference to any other consideration for acting (self-conceit). This means we are prone to mistake a maxim of satisfying the inclinations for an unconditional principle of the will. Our subsequent ‘act’ of conferring authority onto the pursuit of happiness is tempting, and naturally so, because it rests on a deceptive illusion that arises when reason ‘cultivates itself ’, to use Kant’s language from the Groundwork (see §1.11).26 Support for this developmental view of self-conceit comes from what Kant has to say about the constitution of our faculty of desire:
always a sensible modification of the mind but one that has not yet become an act of the faculty of desire’ (MS 6:213). For more on Kant’s idea of a propensity, see Frierson (2005, 21). 26 As Herman (2005) explains, in taking up the task of happiness we begin a process of individu ation. ‘Given the variety of needs, desires, and interests that we have, work needs to be done with and among them—develop some, defer others, sublimate, repress—the whole battery of techniques, conscious and unconscious, by which we come to be the specific individuals we are’ (2005, 26).
112 Self-Love and Self-Conceit We find our nature as sensible beings so constituted that the matter of the faculty of desire (objects of inclination, whether of hope or fear) first forces itself upon us, and we find our pathologically determinable self, even though it is quite unfit to give universal law through its maxims, nevertheless striving antecedently to make its claims primary and originally valid, just as if it constituted our entire self [ganzes Selbst]. (KpV 5:74; emphasis added)
To contextualize what Kant is saying, consider first what it is like to have a desire. When I am thirsty, for example, my desire for water does not initially appear to me as a proposal. As Tamar Schapiro explains, my first experience of desire is more like a demand—‘Drink!’ or ‘Drink this!’ or ‘Drink water!’—where I do not yet realize the demand is internal to me.27 Only with the capacity to step back and reflect are we able to recognize that these responses have their source within us. And that is Schapiro’s point. Prior to this capacity, we experience our desires as demands because we cannot yet distinguish ourselves from objects that attract (or repel) us. From the standpoint of self-love, then, our desires have a forceful character: they speak (as Hobbes puts it) in the language of commands, ‘as do this, forbear that’.28 I believe Kant is making a similar point at KpV 5:74. As sensible beings our faculty of choice is first awakened by objects of desire and aversion, i.e., things in the world we represent as agreeable (with the hope of pleasure) or disagreeable (with the fear of pain). Notice, however, that in the second half of the passage Kant shifts attention to the standpoint we take in self-love before self-conscious reflection. Before self-conscious reflection objects of inclination appear demanding (‘as do this, forbear that’), and from the standpoint of self-love we take this appearance of authority to be objective.29 On the developmental view I am proposing, Kant is saying that we are prone to think the claims of our sensible nature should take priority in deliberation because they are the first in time to make demands on us, and they make demands on us repeatedly (we always have desires, even after developing our rational capacities). As a result, we are prone to think that pursuing our happiness, the sum-total satisfaction of our desires, should take priority in all matters of choice (cf., KpV 5:25, 5:73; G 4:399). To use a spatial metaphor, we can say the idea of our happiness appears large in our deliberative outlook, and we are prone to confuse this largeness for its authority.
27 Schapiro (2009, 246–7). 28 Hobbes (1651/1996, I.vi). Following Schapiro (2009), we might think of our pathologically determinable self as the motivational capacity we share in common with non-rational animals. Our animal self is agential (for it is responsive to practically salient features of objects, e.g., as desirable or aversive), but non-rational (for it does not offer grounds for acting one way or the other). In this respect, to find ourselves under the influence of desire is to experience our animal self ‘reaching out’ or ‘going for’ an object, just as creatures of instinct would act. 29 Here I am indebted to Darwall’s (2006) analysis of self-love.
Moral Sensibility 113
4.6 The Motivational Link There is, however, a problem not far from the surface. First, the cause of infringement discussed above is a limit the moral law poses to doing what we want. Insofar as this limit could have come from another object—such as a rain shower that thwarts one’s plans for the afternoon30—the link between this feeling and the law itself is not intrinsic. Notice too that the cause of humiliation is the deflation of one’s self-worth; but again, insofar as this deflation could have come from an object other than the moral law, we do not have an intrinsic link here either. So at this point in the third chapter Kant’s account is limited. He has shown why certain painful feelings necessarily arise when we are faced with the moral law (irrespective of anyone’s particular sensibility), but none of these effects, ‘infringement’ or ‘humiliation’, are suited to the role of a moral incentive, which raises the specter of motivational skepticism once again. Thankfully this is not where Kant ends his discussion in the third chapter. While it is true that we must experience a loss of self-worth in comparing ourselves to the moral law, Kant wants to show why this feeling has a positive aspect as well. The feeling of humiliation can only arise from our consciousness of something whose worth and authority is greater than what we, in our self-conceit, have bestowed upon our pursuit of happiness. For this reason the object of our humili ation can only be something we represent to ourselves as more worthy of our esteem than our sensible nature. Now since the moral law is the only principle that can command our respect in this way (that is, as something more valuable than the sum-total satisfaction of our desires), we can only feel humiliated when we compare ourselves to the moral law. ‘If something represented as a determining ground of our will humiliates us in our self-consciousness’, Kant writes, ‘it awakens respect for itself insofar as it is positive and a determining ground’ (KpV 5:74). But there is still a missing detail here. So far we have seen that the moral law must effect us negatively by infringing upon our self-love and striking down our self-conceit. We have also seen that humiliation relates necessarily to ‘respect’ for the law, because the law is the only principle of choice that can outweigh the sum of our inclinations. Yet from all of this it remains to be seen whether respect has a positive affective element, an element that would play an active role in our choice to adopt the moral law as our maxim. I believe we can find this missing detail in the concept of ‘self-approbation’ (Selbstbilligung) that Kant introduces later in the 30 I am borrowing this imagery from Engstrom (2010). As he writes, the pain of infringement ‘is not adequate to the moral law in the sense that it involves no distinctive reflection of it. Such a feeling would be produced just as well by any other obstacle, such as a rain shower that spoils one’s plans for a picnic. Indeed, the inclinations themselves can be such obstacles to one another. The feeling reflects the nature of the inclinations rather than the moral law; nothing has yet emerged that qualifies as moral feeling in any interesting sense’ (2010, 100).
114 The Motivational Link third chapter: ‘As submission to a law, that is, as a command’, he writes, there is no pleasure in action required by duty, ‘but instead, so far, displeasure’ (KpV 5:80). Yet Kant immediately goes on to speak of the pleasurable counterpart of this feeling: On the other hand, however, since this constraint is exercised only by the lawgiving of [the agent’s] own reason, it also contains something elevating [Erhebung], and the subjective effect on feeling, inasmuch as pure practical reason is the sole cause of it, can thus be called self-approbation. (KpV 5:80–1)
As I understand it, Kant’s point is that just as we must feel esteem for the law when we recognize its authority (as a principle greater than the sum of our inclin ations), we must also feel esteem for ourselves when we recognize our capacity to legislate the law. Consequently, the awareness of ourselves in this exalted position, giving the law rather than submitting to it, must awaken a feeling in us very much like pleasure.31 It is the effect of seeing ourselves rise above the vocation we had vainly ascribed to our pursuit of happiness in self-conceit, discovering in our autonomy of will an infinitely greater source of value. Kant’s battle metaphor is somewhat misleading, though. As I see things, our self-conceit is not so much struck down as it is painfully enlightened. Our selfconceit is based on the misperceived value of our sensible nature, and we must feel something like pain when this conception of ourselves is stripped away from us. When we see through the pretense of our self-conceit (our striving to make the claims of sensibility ‘primary and originally valid’), we must recognize our mistake in having treated those claims as a source of law. Yet—to repeat my point above—this feeling is different from that of infringement, and if our ego suffers a blow here it is only because we cast reproach on ourselves. The pain of humili ation is not a feeling of hindrance from doing what we please; it is a feeling of reproach, reproach for what we have already done. Moreover, in criticizing our tendency to overvalue the inclinations, we are aware of our capacity to act independently of pathological incentives, and this discovery elicits a kind of pleasure in us, what I shall call the pleasure of self-respect (‘pleasuresr’) for reasons that will soon be clear.
31 Kant’s brief discussion of interpersonal respect is also helpful here (KpV 5:77). On the one hand, he says, to feel respect for another human being is unlike a feeling of pleasure, since we often try our best to resist it. When a person of virtuous character elicits our respect, we try our best to find something to rid ourselves of this feeling, ‘some fault in him to compensate us for the humiliation that comes upon us through such an example’ (KpV 5:77). On the other hand, respect is unlike a feeling of pain, for once we have humbled our pride we feel esteem for the moral law itself, ‘and the soul believes itself elevated in proportion as it sees the holy elevated above itself and its frail nature’ (KpV 5:77). When we acknowledge a source of value and authority greater than our self-conceit, we experience something bordering on reverence—a feeling for the moral law’s ‘majesty’—and Kant’s point is that we experience ourselves elevated as a result.
Moral Sensibility 115 Reviewing the points covered so far, we can say that the value we recognize in our rational nature is greater than that of our inclinations. In this comparison, our inclinations—once so large in our deliberative outlook—now appear small.32 If this is right, then we can appreciate why Kant later relates the feeling of selfrespect to a feeling of the sublime, for the sublime is that against which everything else seems small (KU 5:250). As he writes: ‘The majesty of the law instills awe (not dread, which repels; and also not fascination which invites familiarity); and this awe rouses the respect of the subject toward his master, except that in this case, since the master lies in us, it rouses a feeling of the sublimity of our own vocation and enraptures us more than any beauty’ (R 6:23n).33 It may be this feeling, a feeling for the sublimity of our rational nature, which allows the moral law to become a real motivating incentive for us. For when we recognize our rational nature, we come to have a proper understanding of our relationship to the moral law itself, namely, as a law we give to ourselves.34 Now if we did not confer authority to the inclinations in self-conceit, we would not be in a position to criticize ourselves before the law. Nor would we be in a position to recognize a different source of value within ourselves. The lowering and uplifting moments of respect are connected, then, because they have the same object.35 That is, our free conferral of authority to the inclinations is the object of our humiliation, but this freedom to confer authority is also the object of our elevation. There is no room for motivational skepticism, therefore, because there is no gap between the moral law and its manner of influence on our faculty of sensibility. The pleasuresr we feel in respecting ourselves arises from the same awareness that underlies the painsc brought about by our self-reproach. It is the awareness of ourselves as rational beings with a vocation higher than the pursuit of happiness.36 32 Allison makes a similar point: ‘Kant’s basic claim is that the constraint or limitation of our sensuous nature and its pretensions, which supposedly results from the awareness of the authoritativeness of the moral law, is at the same time an elevation (Erhebung) of our rational nature and produces a feeling of self-approbation that is the positive counterpart of the negative feelings of pain and humili ation. As Kant makes clear, this is because the constraint is self-imposed, that is, imposed by one’s own legislative reason’ (1990, 125). I agree that the ‘unmasking’ of self-conceit, and the humiliating feeling that follows from it, must result from our awareness of the moral law’s authoritativeness. But Allison does not explain how a humiliating feeling connects to the uplifting discovery of our rational nature. 33 Herman aptly describes our initial encounter with the moral law as bringing about a ‘shock of self-consciousness’ (2007, 196). She writes: ‘We see that what we had taken to be first in the order of value—our satisfaction—is not. We are also revealed to ourselves to be persons of moral standing or dignity: our own rational nature as a source of value that has authority over all action and choice’ (2007, 196). Bagnoli makes a similar point about the feeling of humiliation (2012, 173–4). 34 For discussion of the similarities between the feeling of respect and the sublime, see Nahm (1957), Walker (2000), and Clewis (2009). 35 Although we arrive at this conclusion from different routes, my account is similar to Engstrom’s (2010), especially when he speaks of ‘the felt diminution of self-conceit . . . and the corresponding felt magnification of the moral law’ (2010, 117). 36 Speaking of ‘awareness’ is somewhat misleading. Yet we should bear in mind that, while every feeling has a relation to the subject, that relation is not one of cognition. This is essential for Kant’s account, for the simple reason that feelings are capable of motivating us, i.e., of relating to the faculty of desire, in a way that would not be possible for any form of theoretical self-awareness.
116 Our Higher Vocation
4.7 Our Higher Vocation Later in the third chapter Kant connects the feeling of self-respect to our discovery of autonomy in the activity of moral constraint. As he writes, ‘since this constraint is exercised only by the lawgiving of the agent’s own reason, it also contains something elevating [Erhebung], and the subjective effect on feeling, inasmuch as pure practical reason is the sole cause of it, can thus be called self-approbation [Selbstbilligung] with reference to pure practical reason’ (KpV 5:80-81). Kant then links this subjective effect to a pure moral ‘interest’ (Interesse): the agent ‘now becomes conscious of an altogether different interest subjectively produced by the law, which is purely practical and free’ (KpV 5:80–1).37 Considering Kant’s claim that an interest is always based on an incentive (KpV 5:80), it follows from the logic of his analysis that the awareness of ourselves as rational beings must have an effect on us like pleasure—what Kant now terms ‘self-approbation’ and what I am calling ‘pleasuresr’—and that this effect is the basis of our interest to make the law our highest maxim of choice. Kant returns to this point in speaking of the ‘constitution’ of the moral incentive. ‘This incentive’, he writes, ‘is none other than the pure moral law itself insofar as this law allows us to discern the sublimity of our own supersensible existence and subjectively brings about, in human beings, respect for their higher vocation [höhere Bestimmung]’ (KpV 5:88). To fill in the details of this remark, we can say that our consciousness of the moral law is elevating because it lets us see the connection of our rational nature to a higher vocation relative to the end we had assigned to ourselves in self-conceit.38 And that is why the spatial metaphor is fitting after all. In feeling respect for ourselves we come to see that the value of our happiness is infinitely small in comparison to the value of autonomy expressed in our rational nature.39 37 The notion of self-approbation is also central to Kant’s other major works in moral philosophy. In the Groundwork, for example, he says that we are only aware of our dignity when we recognize our capacity to give the law. ‘For there is indeed no sublimity in him insofar as he is subject to the moral law, but there certainly is insofar as he is at the same time lawgiving with respect to it and only for that reason subordinated to it. We have also shown above how neither fear nor inclination but simply respect for the law is that incentive which can give actions a moral worth. Our own will insofar as it would act only under the condition of a possible giving of universal law through its maxims—this will possible for us in idea—is the proper object of respect’ (G 4:439–40). We find a similar claim in the Metaphysics of Morals when Kant writes that ‘from our capacity for internal lawgiving and from the (natural) human being’s feeling himself compelled to revere the (moral) human being within his own person, at the same time there comes exaltation of the highest self-esteem, the feeling of his inner worth (valor), in terms of which he is above any price (pretium) and possesses an inalienable dignity (dignitas interna), which instills in him respect for himself (reverentia)’ (MS 6:436). 38 On this point my reading adds to Korsgaard’s (1996b), who also argues that our higher vocation constitutes an incentive to lead a life guided by the moral law (1996b, 176). 39 Reath notes that the metaphor of mechanical force is ‘consistently embedded in discussions in which the dominant theme is a struggle for authority, sovereignty, superiority, and so on, in which it is claims or pretensions that are being opposed to each other’ (2006, 30, note 26). Here I agree with Reath: ‘a close reading of this chapter of the second Critiqueshows that legal and political metaphors dominate’ (2006, 30, note 26).
Moral Sensibility 117 Support for this reading comes from what Kant says at the end of the third chapter when he connects the notion of duty to what he calls our ‘personhood’ (Persönlichkeit). Asking where the ‘noble descent’ of duty can be found, given that it never speaks in favor of the inclinations, he writes: It can be nothing less than what elevates [erhebt] a human being above himself (as a part of the sensible world), what connects him with an order of things that only the understanding can think . . . It is nothing other than personhood, that is, freedom and independence from the mechanism of the whole of nature, regarded nevertheless as also a capacity of a being subject to special laws— namely pure practical laws given by his own reason, so that a person as belonging to the sensible world is subject to his own personhood insofar as he also belongs to the intelligible world. It is then not to be wondered at that a human being, as belonging to both worlds, must regard his own nature in reference to his second and highest vocation only with reverence, and its laws with the highest respect. (KpV 5:86–7)
The following picture now begins to emerge in the second Critique. The moral law serves as a real motivating incentive when we see that its authority is greater than the sum of our inclinations, and so more worthy of esteem than the pursuit of happiness. Yet the basis of our interest in the moral law, i.e., our interest to make it our own highest maxim of choice, comes from the specific feeling of selfapprobation. This is because self-approbation is directed to our personhood, which we see places us infinitely above the demands of our sensible nature. Selfapprobation constitutes the basis of our interest to adopt the moral law as our life-governing policy—as a ‘ground for maxims of a course of life in conformity with it’—because it is the law of our higher vocation as persons.40 This begins to explain the peculiar status of moral feeling as a synthesis of reason and sensibility, of what is pure in our faculty of practical reason and what is empirical. For while self-approbation is something I feel, it is not merely subjective or reducible to some set of psychological features unique to me alone, because it does not arise from anything I want. The awareness of my rational 40 Kant is sensitive to the problem of moral ‘enthusiasm’ (Schwärmarei) as a tendency to overestimate our elevated dignity as rational beings. After a digression on the Biblical commandments of love, Kant explains that his interest is not so much to curb religious enthusiasm as it is to clarify ‘the moral disposition directly, in regard to our duties toward human beings as well, and to check, or where pos sible prevent, a merely moral enthusiasm which infects many people’ (KpV 5:84). To this end Kant says we must realize that complete perfection of will is beyond our reach. The possibility of violating the moral law will always be alive for us; and that is why we must view the moral law as a system of imperatives, as what we ‘ought’ to do. It would be foolish of us to think we could relate to it differently (KpV 5:84). Indeed, it would be nothing more than a fantasy of transcending human nature itself, ‘as if we could ever bring it about that without respect for the law, which is connected with fear or at least apprehension of transgressing it, we of ourselves, like the Deity raised beyond all dependence, could come into possession of holiness of will’ (KpV 5:82).
118 Kant on Confession nature (my capacity for self-legislation) is pleasurable, but the source of this feeling is interpersonal, a capacity I share with all finite rational beings. Once I feel respect for my higher vocation, then, it is possible for my consciousness of the moral law to be motivationally compelling, i.e., as an incentive for me to adopt the moral law as a maxim for living my life by its precepts.41 While it is painful and indeed humiliating to see that I vainly ascribed authority to my sensible nature in self-conceit, it is also pleasurable when I recognize a greater source of value within myself, my personhood. Self-approbation is a morally motivating feeling, as a result, because it underlies my choice to adopt the law of my personhood, to make it my own highest policy of choice.
4.8 Kant on Confession At this juncture, one might raise the following worry.42 It seems difficult to recall a moment in which we have felt ourselves humiliated by the moral law or uplifted by the discovery of our rational nature. It seems especially difficult to find a moment in which we have seen through the inclinations’ ‘largeness’ and recognized in our rational nature an infinitely greater source of value. To address this worry, we should bear in mind that Kant’s aim in the third chapter is not to present any kind of mere descriptive analysis. He is not trying to make empirical generalizations based on what moral experience is like for certain people, under certain conditions. Rather, Kant is concerned with what moral experience must be like for every rational being with the capacity to feel, regardless of his or her particular sensibility. That is why the status of his phenomenological method in the second Critique is a priori (cf., KpV 5:72).43 Even with this qualification in place, worries about the accuracy of Kant’s moral phenomenology are not unfounded. For if Kant’s account is successful, we should expect to see the humiliating-elevating aspects of moral feeling reflected in more specific, everyday cases. While settling this issue would take us beyond the scope of this book, it is worth noting that Kant supplies material for a reply later in the second Critique. One passage worth considering more closely occurs near the end of Part II where Kant connects the humiliating-elevating aspects of 41 Here one might worry that I have violated Kant’s claim that we cannot know our ourselves (G 4:407). But we should keep in mind that when Kant speaks of self-approbation in the second Critique, he is referring to a feeling we have upon discovering our capacity for self-legislation. This is distinct from any presumption of knowing our moral character. In short, awareness of a capacity is distinct from knowledge of a capacity’s use; and Kant, I take it, only denies our access the latter. The topic of self-knowledge and its limits will be central to Chapter 5. 42 I would like to thank John Dyck and Chris Herrera for pressing me on this point. 43 Kant is not providing a moral phenomenology in the (mostly deflationary) sense used today. There are, however, similarities worth exploring. See Mandelbaum (1955), Horgan and Timmons (2005), and Gill (2009).
Moral Sensibility 119 moral feeling to the phenomenology of confession. What is telling is that he relates this experience to the adoption of a standpoint that is initially painful yet subsequently exalting: In a case where I alone know that the wrong is on my side—although a free confession of it and an offer of satisfaction are strongly opposed by vanity, selfishness, and even an otherwise not illegitimate antipathy to him whose right I have detracted from—I am nevertheless able to disregard all these considerations; and this includes consciousness of an independence from inclinations, from circumstances, and of the possibility of being sufficient to myself, which is salutary to me in general and in other respects as well. And now the law of duty, through the positive worth that observance of it lets us feel, finds readier access through the respect for ourselves in the consciousness of our freedom. (KpV 5:161)
An act of confession must be humbling insofar as we face our wrongdoing and own up to the harm we have caused another person. Perhaps not every case of confession will be as elevating as Kant is describing here. But insofar as we recognize that it is the better part of ourselves that demands the confession in the first place, it is reasonable to suppose that our humility will change into esteem for ourselves. And if that is all Kant means when he says the moral law finds ‘readier admittance’ to sensibility, then it is also reasonable to suppose that our incentive to act in compliance with the moral law will manifest itself as a form of selfrespect. Humiliation must give way to self-respect, then, when we recognize that the demand to confess—to expose our wrongdoing—is one we freely impose upon ourselves.
4.9 The Affectivist View For readers whose only encounter with Kant’s ethics is through the Groundwork, the presence of a chapter in the second Critique devoted to the incentives of pure practical reason—which investigates the affective dimension of moral life—is likely to come as a surprise. Reactions to Kant’s third chapter have varied among scholars, ranging from lively approval (‘the most brilliant phenomenological study of the phenomenon of morality that we have from him’44) to frustration (‘the most repetitious and least well-organized chapter of the book’45). In recent years, however, the third chapter has become the site of two major disputes over the role of feeling in Kant’s ethics, to which I now turn.
44 Heidegger (1927/1988, 133).
45 Beck (1960, 219).
120 The Affectivist View Regarding the first dispute, a growing number of scholars have argued that respect is like a feeling in a very real sense: it has affective force, and that force is variable in degree of intensity. Richard McCarty is the main proponent of this reading, the affectivist view, and like others he opposes a reading of Kant that explains moral motivation in strictly cognitive terms, the intellectualist view.46 This dispute centers, as McCarty notes, on the question of ‘whether the affective component of respect plays any role in the mechanism of moral motivation’.47 As he writes: 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. Opposed to the intellectualist interpretation is what I shall call the affectivist view. Affectivists need not deny that Kantian moral motivation initially arises from an intellectual recognition of the moral law. Contrary to intellectualists, however, they maintain that it also depends on a peculiar moral feeling of respect for law, one consequent to the initial recognition or moral judgment the intellectualists emphasize exclusively.48
McCarty thus disagrees with Andrews Reath’s view that moral feeling plays no motivational role in Kant’s account. According to Reath, ‘it is the practical aspect [of respect] that is active in motivating moral conduct, while the affective side, or feeling of respect, is its effect on certain sensible tendencies’.49 I agree with McCarty that the intellectualist view faces a number of difficulties. One is that the third chapter of the second Critique contains many passages indicating that the affective or feeling-aspect of respect is necessary to motivate moral conduct. Consider the following: Thus the moral law, since it is a formal determining ground of action through practical pure reason . . . is also a subjective determining ground—that is, an incentive—to this action inasmuch as it has influence on the sensibility of the subject and effects a feeling conducive to the influence of the law upon the will. (KpV 5:75; emphasis added) 46 For those who land more clearly on the side of affectivism, see Herrera (2000), Grenberg (2001), Guyer (2000, 2008, 2012), Ameriks (2004), Zinkin (2006), Morrisson (2008), Engstrom (2010), and Frierson (2014). Reath is the most forceful proponent of intellectualism (2006), although a version of the view is found in Allison (1990). 47 McCarty (1993, 430). 48 McCarty (1993, 421). As Frierson summarizes this dispute, intellectualists ‘claim that morally good action is motivated solely by cognition or consciousness of the moral law, with a feeling (of respect) generally seen as an effect of moral motivation rather than its cause’ (2014, 117). By contrast, affectivists argue that ‘the feeling of respect is the immediate cause of moral motivation, the means by which an otherwise inert cognition of the moral law can give rise to an action’ (2014, 117). 49 Reath (2006, 10).
Moral Sensibility 121 Respect for the moral law must be regarded as also a positive though indirect effect of the moral law on feeling insofar as the law weakens the hindering influence of the inclinations by humiliating self-conceit, and must therefore be regarded as a subjective ground of activity—that is, as the incentive to compliance with the law—and as the ground for maxims of a course of life in conformity with it. (KpV 5:79; emphasis added)
Now if our consciousness of the moral law’s authority is necessary and sufficient to move us, why would Kant speak of reason’s ‘influence on the sensibility of the subject’ as effecting ‘a feeling conducive to the influence of the law upon the will’? As far as I can tell, it is difficult to make sense of this if we keep to an intellectualist view. A second difficulty for Reath’s interpretation is that it does not fit neatly into the structure of Kant’s argument from the second Critique. By the third chapter Kant is already assuming the moral law ‘determines the will objectively and immediately in the judgment of reason’ as part of the Factum we discussed in Chapter 2: our common consciousness of the moral law’s necessity. The question of how pure practical reason can interact with our faculty of sensibility is a question of how this objective judgment can also be effective as a real motivating incentive to action. Lastly, a further difficulty worth mentioning is that Reath’s interpretation only attends to the negative aspects of moral feeling, either the pain of having our self-love limited (painsl) or the humiliation of having our self-conceit struck down (painsc). Yet this ignores Kant’s effort to connect self-conceit’s humiliation to a positive feeling, what I called pleasuresr (see §4.6). The affectivist view defended by McCarty suffers from its own shortcomings. Whereas Reath is unable to explain the positive effect reason has on sensibility due to the priority he assigns to our intellectual consciousness of the moral law, McCarty is unable to explain this effect because he views moral motivation strictly under a ‘battle of forces’ model. The problem with this model is not that it stands in tension with Kant’s commitment to freedom of will (a point intellectualist commentators often make50), but that it only considers the process of moral motivation from a third-person perspective. As I argued in §4.4, while we can speak of feelings as ‘strong’ or ‘weak’ forces, that is not the standpoint we find in the second Critique. For Kant, the content of feelings, either a representation of agreeableness-to-oneself or that of disagreeableness-to-oneself, is only meaningful from the standpoint of the agent for whom objects in the world either agree or disagree. So far as I can tell, neither the intellectualist nor the affectivist provide
50 For example, see Reath (2006, 13) and Allison (1990, 126). The question of whether a battle of forces model ultimately conflicts with Kant’s account of free choice is a much larger topic than I can go into. Frierson, for example, argues for their compatibility (2005), and I am sympathetic with aspects of his account.
122 The Affectivist View conceptual resources to illuminate the affective yet first-personal dimension of feeling, the very dimension that underlies Kant’s phenomenology of respect. One might wonder where I stand in this dispute. If the awareness of our rational nature elicits a positive feeling in us, and if this feeling is what motivates us to act in compliance with the moral law, then have I not located the active factor of moral motivation in the affective rather than the cognitive side of respect? I am prepared to say yes, but only with the following qualification in place. Like the affectivist, I deny that our consciousness of the moral law is a sufficient condition for moral motivation, and that is why I cannot hold to an intellectualist view in any strict sense. But the point I want to make is that we can deny this without having to limit our account of moral feeling to a battle of forces model. That is to say, we can affirm that feeling and sensibility have a positive role to play in Kant’s moral psychology without limiting ourselves to the third-person perspective McCarty and others focus on exclusively.51 To sharpen this point of my disagreement, the problem with the traditional affectivist view is that it only considers respect as a feeling with varying strength, but as we have seen, the concept of strength (or weakness for that matter) does not apply directly to the first-person standpoint an agent takes up in deliberation. The question Kant is concerned with in the second Critique is not how moral feeling can overpower the inclinations in a competition of brute force. It is, rather, the question of how reason can interact with sensibility without that interaction reducing to a pathological feeling of pleasure or pain. Kant’s aim is to show how our consciousness of the moral law must effect our capacity to feel, with the positive effect (the analogue of sensible pleasure) serving as our interest to make the law our highest maxim of choice. But to understand this process, in my view, it will not help to consider feelings merely as forces that act upon the agent, even if we characterize those forces as noumenal in origin. We must instead develop our investigation from the agent’s point of view, the point of view where feelings have intentional content. We therefore need a phenomenological method, the method I have argued informs Kant’s approach in the third chapter. Of course, if someone were to ask us how reason can exercise causal efficacy and infuse a feeling of pleasure in the fulfillment of duty, all we can do is affirm
51 In a more recent work, McCarty (2009) shows sensitivity to the phenomenological perspective of the third chapter. He observes that respect is a kind of pleasure that presupposes ‘a prior, frustrating and therefore painful effect on inclinations’ (2009, 179). Yet he expresses uncertainty about how these moments relate: ‘[Respect] can be called moral pleasure, and our feeling it would always presuppose a prior, frustrating and therefore painful effect on inclinations. I think there is more than this in the phenomenology of respect, however. For Kant claimed that the moral law’s ‘negative effect on feeling (disagreeableness) is pathological’ (KpV 5:75). But he also characterized the feeling of respect as not pathological but ‘practical only’ (2009, 179). After this McCarty goes on to suggest that the elevating aspect of respect comes from ‘at least a vague sense of freedom from the limitations of sensible inclin ations’ (2009, 180). McCarty, however, presents no solution to the threat of motivational skepticism discussed in §4.6.
Moral Sensibility 123 our ignorance of such things and, if necessary, employ a polemical method of defense. And that is Kant’s final point in Section III of the Groundwork: the skeptic who denies reason can produce a motivationally effective feeling is really speaking about things she cannot claim to know. From a third-person perspective, then, the philosopher’s stone will always remain a mystery, since we cannot cognize causal relations beyond the sphere of possible experience. But when Kant takes up this question again in the second Critique, he does so from a different point of view. The approach we find in the third chapter, as I have shown, centers on a first-person perspective that applies to any being with reason and feeling, and from this perspective Kant seeks to study the effects our consciousness of the moral law must have on our capacity to feel. It is no doubt remarkable that in the second Critique we appear to gain insight into a special synthesis, the reciprocal interaction of reason and sensibility. But it bears repeating that this is a phenom enological insight, one that presupposes (without claiming to cognize) the nou menal causality of moral motivation.
4.10 The Heideggerian View Another dispute I would like to discuss concerns the relationship between the fact of reason (as Factum) and the feeling of respect (as Achtung). Many commentators believe that the fact of reason belongs to Kant’s project of justification in the first chapter, but that his phenomenology of respect in the third is a mere addendum to this project. For an example of this reading, consider the following remark by Allison: Fortunately [in the third chapter of the second Critique] it is also possible to ‘bracket’, as it were, the underlying presupposition of the validity of the moral law and to regard the discussion of respect essentially as a phenomenology of moral experience. Considered as such, it can be taken as complementing the account of moral worth offered in the Groundwork without denying the possibility that, for all we know, morality might be nothing more than a ‘phantom of the brain’ (Hirngespinst).52
In recent years, however, Allison’s view has come under criticism by Jeanine Grenberg and Dieter Schönecker who have worked to rehabilitate a view first defended by Martin Heidegger, according to which the fact of reason is only revealed to us through the feeling of respect.53 Moreover, unlike Allison, who claims that Kant’s theory of moral sensibility lies outside his project of justification,
52 Allison (1990, 121).
53 See Grenberg (2013) and Schönecker (2013).
124 The Heideggerian View Grenberg and Schönecker argue that the feeling of respect confirms the validity of the moral law, making it central to his project of justification.54 As Grenberg presents this reading, the fact of reason is ‘a felt, given phenomenological experience of categorical obligation’.55 On her account, which remains the most detailed in the literature to date, the Heideggerian view emerges from the following claims: (1) Facts force themselves upon us. (2) What is forced upon us is accessible only in our receptive faculty, i.e., in sensibility. (3) The fact of reason forces itself upon us. (4) Therefore, the fact of reason is accessible only in sensibility (through the feeling of respect). To see whether this line of interpretation is correct, however, let us start with the first premise of her argument. (1) Facts force themselves upon us. As Grenberg writes, ‘When I encounter something in my empirical experience of objects, things are indeed given to me in this way: I cannot choose whether to perceive whatever happens to be in my range of sensible perception.’56 The coffee cup I see on my desk, for example, is a fact in this sense. It appears to me in such a way that I cannot randomly decide its shape, color, texture, and so forth. What is more, I do not arrive at my knowledge of the coffee cup through a process of inferential or deductive reasoning. It presents itself to me, and in this respect it has a forceful character in my experience. When Kant introduces the fact of reason in the second Critique, it is true that he says it ‘forces [aufdringt] itself upon us of itself as a synthetic a prioripropos ition’ (KpV 5:31). Whatever this fact is, we know that it is not an ordinary fact like the coffee cup I see on my desk, since my cognition of the cup occurs in experience. The fact of reason cannot belong to empirical experience in this way; it cannot be something I grasp through sense perception. And I take it this is why Kant qualifies his remark by adding: 54 As Schönecker argues, ‘it is through feelings that we recognize the validity of the moral law’ (2013, 2). 55 Grenberg (2013, 206). It is not always clear whether Grenberg thinks that our access to the moral law in the fact of reason is limited to the feeling of respect, or whether the two are somehow identical. At times she appears to be making a stronger claim of identity: ‘The Gallows Man’s felt experience is indeed the felt experience of the fact of reason itself ’ (2013, 236); ‘Kant’s deduction of freedom finds its common starting point in the felt experience of the fact of reason we have discussed’ (2013, 261); ‘The beginning principle of the argument is, indeed, still the fact of reason, which is a felt, phenom enological fact’ (2013, 263). While I will sometimes speak in these terms below, my criticism is aimed primarily at Grenberg’s weaker claim of access. 56 Grenberg (2013, 193).
Moral Sensibility 125 However, in order to avoid misrepresentation in regarding this law as a given [gegeben], it must be noted carefully that it is not an empirical fact but the sole fact of pure reason which, by it, announces itself as originally lawgiving. (KpV 5:31)
Grenberg is aware of this qualification. The fact of reason, she writes, is a ‘synthetic a priori’ proposition, the source of which I can never apprehend through sense perception.57 Yet Grenberg still thinks that the forceful quality of empirical facts, the way they are given to consciousness, sheds light on the fact of reason. In particular, she claims, their forceful quality directs us to the only faculty that could make the moral law accessible to us. This brings us to the second step of her argument: (2) What is forced upon us is accessible only in our receptive faculty, i.e., in sensibility. This means a ‘forced fact has to be a felt fact’.58 As Grenberg puts it: ‘when we have something forced upon us, we are passive or receptive in relation to that which is doing the forcing. As such, explanation of how we take in this forced fact must involve some aspect of ourselves capable of being passive or receptive. But such a capacity is our capacity for sensibility.’59 Now if step (2) is correct, the rest of Grenberg’s argument follows. Once we acknowledge (3), that the fact of reason ‘forces itself upon us’, we must admit that we are subject to something unavoidably given in our consciousness. We must admit (4), that our experience of categorical obligation is accessible to us only through a feeling. However, I do not think we must accept this conclusion, since I do not think step (2) is true. Notice, first, that for the Heideggerian view to work, we must justify the claim that whatever is given to us—our perception of coffee cups, our consciousness of the moral law—can only appear through the faculty of sensibil ity. Yet this claim strikes me as incorrect. I think Kant would say that many representations have a forceful quality without linking to our sensible faculty at all.60 In the first Critique, as we saw in Chapter 2, Kant uses language nearly identical to his characterization of the fact of reason when he explains how concepts like 57 Grenberg (2013, 193). 58 Grenberg (2013, 143). 59 Grenberg (2013, 143; cf., 193). 60 This is a problem for Grenberg, since (in addition to the passage at B6 cited above) there are places where Kant speaks of something forcing itself upon us—using the same term as at KpV 5:31, ‘aufdringt’—without the mediation of feeling or sensibility. In the first Critique, for example, Kant writes: ‘The transcendental attempts of pure reason, however, are all conducted within the real medium of dialectical illusion, i.e., the subjective which offers itself to or even forces itself [aufdringt] upon reason as objective in its premises’ (A792/B820). Textual evidence aside, the claim that a forced fact has to be a felt fact strikes me as counter-intuitive. In performing an exercise in logic, I may conclude that something cannot (by logical necessity) be both A and not-A. However, that does not mean that my access to the Principle of Non-Contradiction is felt or limited to sensibility. Thanks to Oliver Sensen for this example.
126 The Heideggerian View substance ‘force themselves’ upon us. Kant says that we can display the pure use of our cognitive faculty as a fact, and to illustrate this he has us perform the following thought experiment: If you remove from your empirical concept of every object, whether corporeal or incorporeal, all those properties of which experience teaches you, you could still not take from it that by means of which you think of it as a substance or as dependent on a substance. (B6)
He then concludes: Thus convinced by the necessity with which this concept forces [again: aufdringt] itself on you, you must concede that it has its seat in your faculty of cognition a priori. (B6)
If we accept Grenberg’s claim that forceful cognitions are accessible only through our receptive faculty, we would be led to conclude from this thought experiment that the necessity attaching to the concept of substance is accessible to us only in sensibility. Yet there is no indication that Kant holds this view, nor is it clear this would make sense within the context of the first Critique. As we discussed in Chapter 1, the very separability of the categories (belonging to the faculty of understanding) and intuitions (belonging to the faculty of sensibility) is what motivates Kant’s chapter on the transcendental deduction (A90/B123).61 To make progress here, consider again Kant’s warning in the second Critique: ‘in order to avoid misrepresentation in regarding this law as a given, it must be noted carefully that it is not an empirical fact’ (KpV 5:31). Grenberg cites this passage, as we have seen, but she interprets Kant’s word of caution to mean only that the fact of reason does not ground ‘synthetic a posteriori knowledge claims’.62 What Grenberg wants to preserve from the analogy to empirical facts in turn is the sense in which we receive them passively. Yet in light of what Kant says in the first Critique, it seems clear that by speaking of a fact he means only to register the unique epistemic status of certain cognitions. On this reading, what we should take away from the analogy is not that empirical facts are received, but that they are actual. The point of Kant’s thought experiment is to show that, when we try to remove every quality from an object, we realize that we cannot remove the
61 In a footnote Grenberg even admits that in the first Critique Kant appeals ‘to first-personal experiences (for example, of watching the ship move from up stream to down stream in the Second Analogy’ [A192–3/B237–8]). But these, she notes, are ‘not felt experiences; and, because Kant is concerned to refer them to empirical objects of experience, they cannot be felt experiences’ (2013, 123, note 19). In saying this, however, Grenberg seems to be contradicting step (2), her claim that a ‘forced fact has to be a felt fact’ (2013, 143). 62 Grenberg (2013, 193).
Moral Sensibility 127 concept of substance (B6). We must then concede, as he says, that our consciousness of necessity must have come from a pure faculty of cognition, for no empirical faculty could have generated it within us. After Kant explains that our consciousness of freedom comes from the moral law, he turns to ask the next logical question: ‘But how is consciousness of that moral law possible?’ (KpV 5:30). To quote his reply yet again: We can become aware of pure practical laws just as we are aware of pure theoretical principles, by paying attention to the necessity with which reason prescribes them to us and to the setting aside of all empirical conditions to which reason directs us. The concept of a pure will arises from the first, as consciousness of a pure understanding arises from the latter. (KpV 5:30; emphasis added)
As we discussed in Chapter 3, reading this passage alongside the thought experiment in the first Critique is illuminating. In the theoretical sphere, Kant is saying, we can show that our faculty of cognition has a pure use by attending to the necessity of concepts such as substance. Our actual consciousness of this necessity is sufficient to show that our cognitive faculty is not empirically conditioned all the way down. In the passage at KpV 5:30, Kant is saying we can show that our faculty of desire has a pure use in the same way, i.e., by attending to the necessity of moral laws. Our consciousness of this necessity is also something actual—a Factum of the most common human understanding—and that is sufficient to show that our faculty of desire it not merely in the service of the inclinations. Interestingly, Grenberg cites the remark from KpV 5:30 twice: What we need to do with this consciousness of the moral law is pay attention to it: ‘we can become aware of pure practical laws . . . by attending to the necessity with which reason prescribes them to us.’63 In order to have ‘immediate consciousness’ of the moral law, we have to pay attention: ‘we become aware of pure practical laws . . . by attending to the necessity with which reason prescribes them to us.’64
Curiously, the ellipses in these passages hide Kant’s comparison to the theoretical sphere, where he says we can become aware of pure practical laws ‘just as we are aware of pure theoretical principles’ (KpV 5:30).65 Without this comparison, the 63 Grenberg (2013, 161; ellipsis in original). 64 Grenberg (2013, 179; ellipsis in original). 65 Schönecker does cite this stretch of text: ‘[2] But how is consciousness of that moral law pos sible? [3] [3.1] We can become aware of pure practical laws, [3.2] just as we are aware of pure theoretical principles, [3.3] by attending to the necessity with which reason prescribes them to us’ (cited in Schönecker 2013, 17; numbers in original). Oddly, Schönecker proceeds to offer a detailed commentary of every numbered point except (3.2), where Kant draws a direct parallel to the theoretical sphere. Clearly, then, this parallel is an obstacle for proponents of the Heideggerian view.
128 Further Issues idea that Kant wants us to reflect attentively on a felt experience looks plausible. But once we bring this comparison to light, the passage from KpV 5:30 takes on a different meaning. I understand Kant’s point to be that moral laws, like pure the oretical principles, express a species of strict necessity that could not have arisen from empirical experience alone. In the practical sphere, he is saying, we are actually conscious of the necessity of moral laws, and that is enough to show that pure reason is practical within us. So when Kant says that our consciousness of the moral law ‘forces itself ’ upon us, we need not take him to mean this consciousness is felt. Nor must we assume that our consciousness of the moral law operates through sensibility because it is ‘given’ like a matter of fact.
4.11 Further Issues What other support might we find for the Heideggerian view? There is one passage from the second Critique that I think deserves special attention. Kant writes that the ‘justification of moral principles as principles of a pure reason could also be carried out very well and with sufficient certainty by a mere appeal to the judgement of common human understanding’ (KpV 5:90). To explain this, he adds the following remark: Anything empirical that might slip into our maxims as a determining ground of the will makes itself known at once by the feeling of gratification or pain that necessarily attaches to it insofar as it arouses desire, whereas pure practical reason directly opposes taking this feeling into its principle as a condition. The dissimilarity of determining grounds [of the will] (empirical and rational) is made known by [the] resistance of a practically lawgiving reason to every meddling inclination, by a special kind of feeling. (KpV 5:91–2)
Grenberg cites this passage by way of criticizing scholars, like Allison, who ‘find feeling important for making sense of the incentive or interest we have to act in accordance with moral principles’ but maintain that ‘proof of the authority that those principles have over us must be accomplished entirely independently of appeal to feeling’.66 It seems clear from what Kant says that the moral law’s validity is revealed to us in a felt experience of conflict between empirical and rational grounds of choice, making Kant’s analysis of ‘respect’ central to his project of justification in the second Critique. As Grenberg notes, however, the text just prior to KpV 5:91-92 is open to interpretation. Here Kant reflects on the organization of topics in the Analytic,
66 Grenberg (2013, 140).
Moral Sensibility 129 observing that they proceed from ‘practical principles a priori’ (the first chapter), to ‘concepts of simply good and evil’ (the second chapter), and finally to ‘the relation of pure practical reason to sensibility’ (the third chapter) (KpV 5:90). Kant repeats this observation a few sentences later, writing that the division of the Analytic of pure practical reason must turn out like that of a syllogism, proceeding from the universal in the major premise (the moral prin ciple), through undertaking in a minor premise a subsumption of possible actions (as good or evil) under the former, to the conclusion, namely, the subjective determination of the will (an interest in the practically possible good and in the maxim based on it). (KpV 5:90)67
Looking at all the textual evidence before us, I believe the burden of proof lands on Grenberg’s side. I say this because scholars such as Allison only have to deal with one passage from the second Critique that seems to oppose their view, the passage where Kant links the justification of moral principles to the feeling of respect (at KpV 5:90–1). Scholars such as Grenberg, however, are forced to adopt a revisionary interpretation of the entire Analytic, questioning Kant’s division between the justification of moral principles in the first chapter and the phenomenology of moral feeling in the third.68 Moreover, for her reading to work, Grenberg needs to show that our access to the moral law is limited to a felt experience, but on closer inspection we can see that the passage from KpV 5:91–2 only states that the ‘dissimilarity’ between empirical and rational grounds of choice is revealed to us by ‘respect’. This in itself does not prove that Kant regards the fact of reason to be an affect, nor does it get Grenberg to the conclusion she seeks, that respect is the sole medium through which we gain access to the moral law.69 To be fair, Grenberg is aware of the passages that conflict with her view. Referring to KpV 5:90, she says it is ‘surprising that Kant claims he has accessed “the possibility of practical principles a priori” without appeal to sensibility’.70 But when Grenberg turns to support her alternative reading, she does not offer us
67 In this context Kant also explains that, in contrast to the Analytic of the first Critique, the Analytic of the second Critique had to begin ‘from the possibility of practical principles a priori’ (KpV 5:90). Only from these, he adds, ‘could it proceed to concepts of objects of a practical reason, namely, to the concepts of the simply good and evil . . . and only then could the last chapter conclude this part, namely the chapter about the relation of pure practical reason to sensibility and about its necessary influence upon sensibility to be cognized a priori, that is, about moral feeling’ (KpV 5:90). 68 Grenberg may be willing to accept the revisionary interpretation, but doing so would come at a cost. One of the unifying claims of her (2013) project is that the argument of Groundwork III suffers from a ‘phenomenological failure’, but that Kant overcomes this failure in the second Critique. Grenberg would have to abandon this narrative if she opted for a revisionary reading. 69 Nor should we overlook the fact that Kant, in the passage Grenberg draws upon to support her reading, speaks counterfactually about an argument that ‘could also be carried out very well’ (KpV 5:90), suggesting that this was not the argument he had previously developed in the second Critique. 70 Grenberg (2013, 143).
130 Further Issues new textual evidence. Instead she retreats (on p. 143) to what I have called step (2), the claim that ‘a forced fact has to be a felt fact’, and she concludes this paragraph by saying: ‘we must appeal to our capacity for receptivity in the sensible form of feeling when making sense of how finite rational beings access the fact of reason; there seems no other capacity by which sensibly affected beings could take something in as given’.71 Yet, as we have seen, Grenberg provides us with no further argument as to why the forcefulness of a cognition entails that our access to it is limited to the faculty of sensibility. Without this claim, however, I see no other way to support the Heideggerian view, and that is the substance of my worry here, since I am not convinced that step (2) is true. The question I now want to raise is this. If we move away from the Heideggerian view, as I recommend that we do, must we return to a narrow reading like Allison’s, which denies that moral feeling plays any role in Kant’s project of justification? I think Grenberg assumes there is no way of giving moral feeling its due in Kant’s moral philosophy if we separate the fact of reason (as Factum, our consciousness of the moral law’s necessity) from the feeling of respect (as Achtung, the effect this consciousness has on our sensibility). Yet I am not convinced that this assumption is well founded. It seems that a third alternative remains open to us, whereby we follow Allison in separating the Factum of reason from the feeling of Achtung without endorsing his specific view of their relationship. As we have seen, Allison claims that Kant’s analysis of moral feeling in the second Critique carries no justificatory weight: in his view, it ‘brackets’ the moral law’s validity.72 But following the interpretation I have defended in this book, we can reject this specific claim without having to collapse the distinction between moral consciousness and moral feeling. We can appreciate the justificatory weight of respect, in other words, without making our access to the fact of reason strictly affective.73 One advantage of this reading is that it makes sense of Kant’s organization of topics in the Analytic. After showing that there is only one fundamental law of pure practical reason, and that we are actually conscious of this law (the first chapter), Kant has two further tasks before him. He must show how the moral law can operate through our faculty of judgement (the second chapter), and he must show how the moral law can operate through our faculty of feeling (the third chapter). Here we can distinguish the fact of reason as our consciousness of the 71 Grenberg (2013, 143). 72 Allison (1990, 121) 73 For this reason I also disagree with Puls, who claims that a ‘human being has a feeling of respect for the moral law—and thereby has an immediate cognition of this law’ (2015, 269; cf., 9, 169, 172, 173). In Puls’s view, this is precisely Kant’s solution to the threat of a circle. For the reading to work, however, it must be the case (1) that the feeling of respect plays not only a moral-motivational role, but also a moral-epistemological role; (2) that the function of respect is to enable our Erkenntnis (cognition) of the moral law’s validity; and (3) that the feeling of respect so understood is identical (or at least very similar to) the fact of reason Kant draws upon in the second Critique. While I am inclined to see continuity in Kant’s project of moral justification—with Puls—I am not convinced that the feeling of respect has the argumentative task he assigns to it.
Moral Sensibility 131 moral law’s necessity and the feeling of respect as the effect this consciousness has on our sensibility, yet maintain that Kant’s phenomenology of respect is central to his project of justification. On my reading, we can say that the third chapter completes Kant’s project of justification in the second Critique by showing how pure reason can be an incentive for living one’s life by its precepts. My main worry is that the Heideggerian view ends up trivializing this last step, since it already assumes what Kant is trying to show in the third chapter, namely, that pure practical reason has a necessary connection to our faculty of sensibility. If we read the second Critique as I am suggesting here, then every chapter of the Analytic has a justificatory role to play, and this coheres well with Kant’s own description of the Analytic as a ‘syllogism’, proceeding from a major premise, to a minor premise, to a conclusion (KpV 5:90). Moreover, my focus on Kant’s analytic and synthetic procedures gives us resources to make sense of this puzzling remark. In the first place, we can see that the doctrine of the fact of reason belongs to the analytic path of the second Critique, where Kant separates the pure and empirical elements of practical reason in order to discover its highest law, the moral law as a principle of autonomy. Taking our common consciousness of this law as a Factum, which precedes all ‘subtle reasoning about its possibility’, Kant then descends along the synthetic path: he shows in the third chapter that what had previously been separated, reason and sensibility, form a real connection in the feeling of Achtung. Moreover, I have argued that Kant’s key to elucidating this connection takes the form of a phenomenological method. Indeed, this method allows Kant to describe the effects moral consciousness must have on our capacity to feel pleasure and displeasure, without that explanation sliding illicitly into the realm of noumenal causes. While the doctrine of the fact of reason establishes the actuality of our moral consciousness, thereby warranting our possession of pure practical reason, it does not yet secure our possession of a motivational faculty responsive to the claims of morality—a faculty of moral sensibility—which is precisely what Kant seeks to prove over the course of the third chapter. In this respect, attending to the arc of Kant’s synthetic path suggests that his project of justifying ethics must include a theory of moral feeling, for only such a theory can show that the moral law is applicable (and thus binding) upon the hearts of human beings.
4.12 Closing Remarks My goal for this chapter was to show that Kant’s theory of moral sensibility plays an essential role in his project of justification. But in light of Kant’s statement of method, I also wanted to supply further evidence to support my claim that Kant’s doctrine of the fact of reason is continuous with Groundwork III. This doctrine appears in the first chapter where Kant is still ascending along the analytic path,
132 Closing Remarks separating our higher and lower faculties of desire in order to discover the supreme law of practical reason. When Kant reaches this peak, he then appeals to our common consciousness of this law as a Factum, thus offering a genetic proof of the concept of freedom (in a way that parallels his ‘deduction of the concept of freedom from pure practical reason’ in Groundwork III). But the arc of Kant’s synthetic path in the second Critique is longer, I have argued, since it seeks to show that the moral law also connects to our faculty of sensibility as a subjective motiv at ing ground of action. Recall that all Kant says about this connection in Groundwork III is that it falls outside the scope of his project of justification, since he only broaches the question in terms of how reason can operate as a noumenal cause. The second Critique goes further, not by providing us insight into such a cause, but by giving us a phenomenological study of the effects that consciousness of the moral law must have on our capacity to feel. In this regard, the synthetic path of each text culminates in a special mode of cognition.74 The synthetic path of the Groundwork culminates in a special a priori insight, revealing the possibility of moral obligation (our cognition of a moral ‘ought’). And the synthetic path of the second Critique culminates in a special a priori insight, too, revealing the possibility of moral motivation (our feeling of respect for the moral law). Another way of presenting my claim here, independently of how we understand Kant’s distinction of analytic and synthetic paths, is to say that the theory of moral sensibility we find in the second Critique goes deeper in resolving a problem first introduced in the Groundwork: our tendency to rationalize against the moral law. In the Groundwork Kant frames this problem in terms of a natural dialectic that arises once we begin to cultivate our rational capacities. In our early stages of self-cultivation we become aware of reasons for action that do not speak to the promotion of our own happiness. These reasons appear to us as constraining imperatives: they command our attention and may even elicit a feeling of respect from us. However, we are liable to reassert our previous attachment to happiness (which was active in us prior to our emerging consciousness of duty), but now with the intention to privilege self-love over and above the claims of morality. This is the birthplace of what Kant in the second Critique calls ‘selfconceit’: our tendency to regard the sensibly affected part of our will as if it constituted our ‘entire self ’ (KpV 5:74). To be sure, Kant had worked to avert this self-deception in Groundwork III by tracing the origin of the moral law’s necessitation to the idea of our pure will. But my point is that the second Critique goes deeper into the perspective of a developing agent faced with the demands of its rational and sensible nature. In the second Critique Kant is able to show that our consciousness of the moral law not only infringes upon our self-love and strikes down our self-conceit—it also awakens in us an elevating feeling of self-worth when we recognize our higher vocation as persons. 74 In other words, the cognition, and not just the strategy of exposition, is synthetic in what Gave (2015) calls the ‘narrower sense’ of the word.
5
Self-Knowledge and Despair The depths of the human heart are unfathomable.
– Kant (MS 6:447)
5.1 Introduction In the foregoing chapters I have focused almost exclusively on the Groundwork and the second Critique, since these texts contain Kant’s most detailed arguments for the reality of human freedom and the normativity of the moral law. On the account I have tried to make compelling, what ties Kant’s project together and gives it unity over these two works is a concern with what we might call a prac tical conflict at the heart of common moral consciousness: a conflict between the claims of happiness and the claims of morality. Indeed, this explains why so much of Kant’s project of moral justification is devoted to an investigation of sources. After all, if we can gain insight into the source of moral obligation, and see that it lies within us—even in our very capacity to feel, as Kant argues in the second Critique—we would have sufficient grounds for trusting that our vocation lies higher than the pursuit of happiness. This would in effect rescue us, not from doubt (which is not at issue here), but from despair, for Kant’s account would vindicate what we are already committed to from the standpoint of common human reason. On the reading I have developed so far, Kant’s project of moral justification seeks to answer a general question, justifying our possession of a faculty of prac tical reason whose parts (the empirical and the pure) are necessarily connected. Yet this does not answer the further question of how we can act under our higher vocation, after we have acknowledged the authority of the moral law, felt uplifted in the awareness of our rational nature, and moved to adopt the law as our own life-governing maxim. For it is one thing to show that we are the kind of beings for whom morality applies, and quite another to show that our moral aspirations are on the right track. The latter raises a question of moral self-knowledge, since it asks how we as individuals can have assurance that our moral progress is genuine. And this brings a new form of despair to the foreground of Kant’s ethics, for without knowing whether our moral progress is genuine or not, he writes that ‘the unavoidable consequence of a rational estimate of our moral state is a feeling
Kant’s Justification of Ethics. Owen Ware, Oxford University Press (2021). © Owen Ware. DOI: 10.1093/oso/9780198849933.003.0006
134 Know Thyself of hopelessness’ (R 6:71). It is this question, I now want to argue, that impels Kant to develop a theory of ‘conscience’ (Gewissen) in his later writings, Religion within the Boundaries of Mere Reason and the Metaphysics of Morals.
5.2 Know Thyself One of the most striking features of Kant’s moral philosophy is his claim that we lack direct access to our motives, intentions, and dispositions—or what we might call the opacity thesis.1 Versions of this thesis play a crucial role in nearly every aspect of his thought from the early 1780s to the late 1790s.2 And yet, in his last major contribution to ethics, Kant advances what appears to be an opposing claim: that we stand under the obligation to Know Thyself (MS 6:441). But how can we reconcile these apparently inconsistent views? How can Kant affirm that self-knowledge is the ‘first command’ of all duties to myself but that direct cogni tion of the self is, strictly speaking, impossible?3 My first task in this chapter is to determine whether the concept of a ‘duty’ of self-knowledge is even coherent. One difficulty here is that Kant distinguishes between generic self-knowledge and particular self-knowledge, and it is not clear which of these is essential to the command ‘Know Thyself ’. Afterwards our task will be to determine whether the duty of self-knowledge has a practical role in an agent’s moral life, and whether that role is positive or negative. For Kant, the obli gation to know myself is a matter of knowing my moral character and my pro gress in meeting the demands of duty. But he argues that I cannot have introspective certainty of my character or progress, for two reasons. First, my character is beyond the limits of immediate consciousness. I cannot cognize my underlying disposition by ‘observing myself ’. Second, and more seriously, I am prone to deceive myself. I may be subjectively convinced that I acted virtuously, or that I reoriented my life to a higher vocation beyond the pursuit of happiness. But my subjective conviction may be mistaken. In truth I may continue to act from self-conceit. And my character may be the same as before. In response to such difficulties, Kant proposes a theory of inferential selfknowledge. If I have genuinely resolved to reorient my character, Kant claims, that resolve should be visible in my life conduct. I should be able to infer the 1 For examples of the opacity thesis in Kant’s moral theory, see: G 4:407, R 6:51, 6:71, 6:87–8, as well as MS 6:447. 2 Variations of this claim run throughout all three of Kant’s critiques. For one example, see the footnote at B xxiv in the Critique. 3 Kant’s formulation of the duty of self-knowledge plays a central role in his early lectures on ethics. The Herder notes show that Kant worked on this theme as early as 1762 (V-PP/Herder 27:43). In the Collins notes of 1784–1785, Kant speaks of ‘self-testing’ and ‘self-examination’ as ‘the primary duty to oneself ’ and disposing the agent so that ‘he may be capable of observing all moral duties’ (V-Mo/ Collins 27:348).
Self-Knowledge and Despair 135 moral status of my disposition, whether restored or unrestored, by way of my actions. But the problem with this theory is that actions are not judgment-neutral. We still need to evaluate, assess, and appraise our own deeds, which forces us back to the problem of self-deception. Who is to say I am a legitimate judge of my own life conduct? How can I possibly examine my actions sincerely and impar tially? I could easily turn a blind eye to my past exploits and thereby construct a false conception of myself. After all, it is within my interest to judge myself in a morally flattering or forgiving light. The question here, then, is whether inferen tial self-knowledge can ever be free from the threat of self-deception. After pre senting this problem, I will outline and assess Kant’s solutions to this threat. Finding none of his solutions entirely satisfactory, I will then present a new account based on Kant’s theory of conscience.
5.3 The Opacity Thesis As I have noted, the limits Kant places on self-knowledge are rather strict and wide-ranging. Not only does he limit the knowledge we can have of others, he also limits the knowledge we can have of ourselves. ‘Indeed’, Kant writes, ‘even a human being’s inner experience of himself does not allow him so to fathom the depths of his heart as to be able to attain, through self-observation, an entirely reliable cognition of the basis of the maxims which he professes, and of their purity and stability’ (R 6:63; emphasis added). I cannot know, for example, whether my particular actions arise from conformity with the moral law or from some hidden self-interest. As Kant argues in Groundwork II, ‘it is absolutely impossible by means of experience to make out with complete certainty a single case in which the aim of an action otherwise in conformity with duty rested sim ply in moral grounds’ (G 4:407). For even when I think—and Kant believes we have a tendency for such thoughts—that I am conforming my actions to the strict commands of duty, I may, on further reflection, detect that my actions have arisen from the ‘dear self ’, which Kant notes ‘is always turning up’ (G 4:407). The simplest version of the opacity thesis is that the ground of my maxims lies beyond my epistemic reach. I am opaque to myself to the extent that I can never know my disposition immediately by way of introspection. What Kant is rejecting here is the idea that I can catch a glimpse of my true character by way of some intuition, as when Nathaniel Hawthorne speaks of ‘one of those moments—which sometimes occur only at the interval of years—when a man’s moral aspect is faithfully revealed to his mind’s eye’.4 But Kant’s discussion from Groundwork II also introduces a second type of opacity. I can never be certain of the purity of my
4 Hawthorne (1850/2003, 150).
136 The Opacity Thesis actions because of my deeply selfish nature. I can never be entirely confident that ‘no covert impulse of self-love’ determined what I thought was my self-sacrificing deed. Kant observes that ‘we like to flatter ourselves by falsely attributing to our selves a nobler motive, whereas in fact we can never, even by the most strenuous self-examination, get entirely behind our covert incentives’ (G 4:407).5 We are therefore opaque to ourselves in two distinct senses. The ground of our maxims is opaque because unknowable, i.e., it falls outside our epistemic reach—or what I shall call type-1 opacity.6 But the purity of our maxims is also opaque because covert, i.e., we cannot know whether our motives have been corrupted by other, less praiseworthy, motives—or what I shall call type-2 opacity.7 It will soon become clear that the greater threat to the duty of self-knowledge is not the fact that I am incapable of perceiving myself directly, whatever that may mean. The greater threat, rather, is self-deception or type-2 opacity. For even if I avoid all forms of introspection by appraising my actions alone, I still need to judge and draw inferences from my actions. But therein lies the new problem of despair. For if I cannot trust my authority as a self-judge, then I have no reliable means to assess my moral progress, and the ‘unavoidable consequence’ of such opacity would be, as Kant says, a ‘feeling of hopelessness’ (R 6:71).8
5 Kant’s idea of the ‘dear self ’ suggests an asymmetry between our self-knowledge of virtuous and vicious motives. In general, he believes we reach the limits of introspection with respect to what we imagine is a good motive behind our deed, because he believes everyone is prone to inflate the moral value of his or her actions. On the other hand, he does not believe our bad motives are opaque to us, implying that one cannot possibly act on a covertly virtuous motive. 6 The Type-1 version of the opacity thesis has its origins in Kant’s critique of self-knowledge from the Paralogisms of the first Critique. In that text he argues against the idealist view that one’s internal thoughts provide indubitable certainty of one’s status as a thinking substance—e.g., Descartes’ ‘I am’. Kant points out that the empirical knowledge we have of ourselves is no more reliable than the empir ical knowledge we have of objects in the world (see Grenberg 2005). Objective self-knowledge is therefore beyond the scope of theoretical reason; and that is because introspection only ever reveals appearances of inner sense, just as sensation only ever reveals appearances of outer sense. See, for instance, A371. However, a fuller discussion on how Kant’s version of the opacity thesis in the first Critique relates to versions found in his later writings falls outside the scope of the present discussion. 7 These two types of opacity are related, although I am not committing myself to any further explanation of their relationship. Roughly, we can say the phenomenon of self-deception is only pos sible on the condition that I can never have objective knowledge of myself. In this sense, type-2 opacity is dependent on the type-1 variety. The nature of this dependency is, of course, a mystery for Kant, since he believes the root of deception (like the root of ‘radical evil’ more generally) is inscrutable. 8 The tension between the duty of self-knowledge and Kant’s opacity thesis often remains a periph eral issue in the philosophical literature. For example, Lara Denis’s study (2001) devotes only half a page to the duty of self-knowledge. Two notable exceptions are O’Neill (1996) and Grenberg (2005). However, I think O’Neill and Grenberg misrepresent the duty in terms of its ‘prescriptive efficacy’. For Kant, moral self-knowledge does not provide us with guiding knowledge of how to act. And so, in this sense, our duty is not to know, even minimally and humbly, the objective correlation between our actions and the principles upon which we think we act. As we will see, our duty is to examine our character, the ground of our maxims, and so to determine whether we have formed our moral judg ments with due care. Other places where expositors of Kant’s views note both the duty of self-know ledge and the difficulty of self-knowledge are Sullivan (1989), Wood (1999), Guyer (2000), Jacobs and Kain (2003), and Frierson (2003). For more recent discussions, see Schumski (2018), Papish (2018), and Berg (2020).
Self-Knowledge and Despair 137
5.4 The Duty of Self-Knowledge In view of the restrictions Kant places on self-knowledge, how are we to understand his claim that moral self-cognition is ‘the First Command of all Duties to Oneself ’ (MS 6:441)? As he tells us in the Metaphysics of Morals: This command is ‘know (scrutinize, fathom) yourself ,’ not in terms of your nat ural perfection (your fitness or unfitness for all sorts of discretionary or even commanded ends) but rather in terms of your moral perfection in relation to your duty. That is, know your heart—whether it is good or evil, whether the source of your actions is pure or impure, and what can be imputed to you as belonging originally to the substance of a human being or as derived (acquired or developed) and belonging to your moral condition. (MS 6:441)
Kant speaks of natural self-knowledge, knowledge of my ‘natural perfection’, and moral self-knowledge, knowledge of ‘my heart’. To complicate matters, he divides moral self-knowledge into two kinds: what is imputable to my moral condition either (i) substantially or (ii) derivatively. To begin with, what Kant may be saying is that I need to know what is good and evil of myself and what is also good and evil of everyone (where ‘evil’ designates our tendency to s elf-conceit).9 Substantial self-knowledge would then be knowledge of the good and evil imputable to me as a member of the human species. Derived self-knowledge would in turn be knowledge of the good and evil imputable to me as an individual, i.e., knowledge of my own idiosyncratic habits, propensities, and tendencies. On top of these divisions Kant makes the dramatic assertion that ‘only the descent into the hell of self-cognition can pave the way to godliness’ (MS 6:441). But given the different kinds of self-cognition outlined above, we must ask: Which hell? Must I descend into the hell of cognizing my generic moral condition, the good and evil I share in common with the human species? Or must I descend into the hell of cognizing my idiosyncratic condition, the good and evil I have cultivated within myself?
5.5 Generic Self-Knowledge In the first place, we should ask why Kant excludes knowledge of natural perfec tions from the duty of self-knowledge.10 The rejection is instructive because it 9 See §4.5. 10 In the Metaphysics of Morals Kant divides all self-regarding duties into perfect and imperfect duties. Perfect duties are what he calls ‘limiting (negative) duties’ (MS 6:419), which place specific constraints on our actions, either physically (e.g., against self-mutilation and suicide) or morally (e.g., against lying, avarice). On the other hand, imperfect duties are what Kant calls ‘widening (positive duties to oneself)’ (MS 6:419), which place us under obligation to adopt choices as ends, specifically,
138 Generic Self-Knowledge tells us that not every object of self-knowledge pertains to the first self-regarding duty. To cite the key passage, Kant says that the duty of self-knowledge is not a matter of knowing my ‘fitness or unfitness for all sorts of discretionary or even commanded ends’ (MS 6:441). I take it this is because natural perfections, or ‘gifts of nature’ as Kant calls them, lack intrinsic moral worth. I may have a naturally courageous character, but if my heart is corrupt I could employ that gift to vicious ends, say, by performing evil deeds with a steady hand. Kant therefore excludes knowledge of natural perfections from the duty of self-knowledge because I must know my heart (whether it is good or evil) before I can begin to fulfil the obliga tion of self-perfection.11 The alternative is clear: the only suitable object of selfknowledge is the ground of my maxims, my enduring moral character. I need to know my heart, whether I am committed to the moral law or to my own happi ness as my highest maxim of choice.12 This requirement raises a new set of difficulties. In the Religion, Kant argues that a human being adopts selfish maxims in such a manner that ‘he expresses at the same time the character of his species’ (R 6:21). And a little later he notes, ‘by the “human being” of whom we say that he is good or evil by nature we are entitled to understand not individuals (for otherwise one human being could be assumed to be good, and another evil, by nature) but the whole species’ (R 6:25). This does not mean, as Kant is careful to point out, that we can infer evil from the general concept of humanity (otherwise, evil would be a necessary quality of human nature). Nor is it an attempt to explain the ultimate origins of evil, which for Kant are inscrutable. His point, rather, is that we can ‘spare ourselves the for mal proof that there must be such a corrupt propensity rooted in the human being’ in light of the many examples of humanity’s evil, both inside and outside the boundaries of the so-called ‘civilized’ world (R 6:32). However we decide to survey human nature, experience leads us to the view of its corrupt propensity,
the end of self-perfection. For Kant, the duty of self-knowledge is neither strictly perfect nor imperfect but rather conditions the possibility of both, which is why, to repeat, he describes the duty of selfknowledge as ‘the First Command of all Duties to Oneself ’ (MS 6:441) 11 A possible exception to this rule would be the case of the good-hearted fool, someone who is naturally well intentioned but lacks practical judgment. One could argue that such an individual would have an obligation to know his good nature so that he could properly align his intentions with his judgments. Thanks to Steve Engstrom for this suggestion. 12 Denis’s passing comment that the duty of self-knowledge is imperfect (2001, 115) seems wrong for two reasons. (1) Imperfect duties, such as natural perfections, structurally require the agent’s moral self-cognition first. (2) If there were such a thing as an imperfect duty to know oneself, we could not specify what about the self one should know. As we will see, however, Kant argues that the duty of self-knowledge does require one to know specific aspects of one’s heart, whether it is good or evil, with respect to one’s generic and particular self-identity. I believe Kant’s claim that the duty of selfknowledge is the first of all self-regarding duties must be taken seriously; it means that without moral self-cognition (i) one could not act out of respect for one’s inner humanity, and (ii) one could not ethically pursue one’s natural perfections of mind, body, and spirit. Hence, the duty of self-knowledge precedes and conditions the possibility of both perfect and imperfect duties.
Self-Knowledge and Despair 139 that human nature ‘cannot be judged otherwise’ (R 6:32). Evil is thereby imputable to the substance of my moral condition, something that everyone brings upon his or herself, without exception. Kant thus reaffirms the words of Sir Robert Walpole: ‘Every man has his price, for which he sells himself ’, to which he supplements Romans 3:9: ‘None is righteous . . . no, not one’ (R 6:38). A curious implication now comes into view. If the notion we have of humani ty’s corruption arises inevitably from experience, something we perceive ‘in every human being, even the best’ (R 6:32), then we would not have a duty to know it. For Kant, a duty pertains to what I should do, in this case to what I should know, as opposed to what I already do or know. That is why he denies that we have a duty to our own happiness because happiness is an end that ‘everyone already wants unavoidably’ as sensibly affected beings. ‘Hence it is self-contradictory’, Kant argues, ‘to say that he is under obligation to promote his own happiness with all his powers’ (MS 6:286; cf., R 6:7n).13 Similarly, it does not make sense to say I am under obligation to know something that by virtue of experience I know or will know easily enough. The conclusion we can draw here is that the duty of selfknowledge cannot require me to know my generically evil heart. But does it make sense to say I am under obligation to know the good I share in common with the human species? In the Religion Kant argues that the command, that ‘we ought to become better human beings’, is one that ‘resounds unabated in our souls’ (R 6:45). So we may safely conclude, in light of such comments, that knowledge of my generically good heart is not difficult to attain either. As we have seen, Kant says in Groundwork III that no person ‘accustomed to use reason in other ways’ can fail to grasp the authority of the moral law, ‘not even the most malicious scoundrel . . . who, when one sets before him examples of honesty of purpose, of steadfastness in following good maxims, of sympathy and general benevolence (even combined with great sacrifices of advantage and comfort), does not wish that he might also be so disposed’ (G 4:454). The scoundrel perceives in the display of virtuous actions what his corrupt will ought to be like. Even he is conscious of his moral vocation, if only faintly. But clearly I cannot be under an obligation to know what is, by Kant’s own lights, obvious—even to scoundrels.14
13 It is not clear, according to Kant’s terminology, why the duty to happiness would be ‘self- contradictory’. To weaken the claim, I will say that the concept of a duty to happiness (or the duty to know one’s generically evil nature, for that matter) is simply vacuous. This accords with Kant’s state ment from the second Critique that a ‘command that everyone should seek to make himself happy would be foolish, for one never commands of someone what he unavoidably wants already’ (KpV 5:37; emphasis added). Thanks to Arthur Ripstein for pointing this out to me. 14 And so is anyone else who fulfils his duty on the most trivial level. An individual finds moral support in his dutiful action, Kant writes, ‘by the consciousness that he has maintained humanity in its proper dignity in his own person’, even if his action consisted of abstaining from telling a harmless lie (KpV 5:88).
140 Generic Self-Knowledge How then can we make sense of the claim that I have a duty to know my moral nature? One point to consider is that knowledge of the opposite, the evil imput able to the substance of a human being, can easily lead me to loathe humanity. I therefore have a duty to become aware of the noble predisposition to the good within myself, for otherwise I could become overwhelmed by the impression of humanity’s evil that experience forces upon me. Kant seems to have this point in mind when he notes: This moral cognition of oneself will, first, dispel fanatical contempt for oneself as a human being (for the whole human race), since this contradicts itself.—It is only through the noble predisposition to the good in us, which makes the human being worthy of respect, that one can find one who acts contrary to it contemptible (the human being himself, but not the humanity in him). (MS 6:441)
Kant observes that ‘our species, on closer acquaintance, is not particularly lovable’, but ‘hatred of them is always hateful’ (MS 6:402). Contempt for the whole human race commits the error of judging the human being solely on the grounds of his corrupt will, ignoring his personality which is predisposed to the good. On this reading, knowing the good I share in common with others could be an antidote to what Kant calls ‘another vice, namely that of misanthropy’ (R 6:34). A further possibility to consider is that generic self-knowledge of my moral nature could strengthen my feeling for the moral law, which would in turn strengthen my desire for self-improvement. In a passage I cited in the previous chapter, Kant remarks that the ‘majesty of the law . . . rouses a feeling of the sublimity of our own vocation that enraptures us more than any beauty’ (R 6:23n). Later in the same text, he writes: Often to arouse this feeling of the sublimity of our moral vocation is especially praiseworthy as a means of awakening moral dispositions, since it directly coun ters the innate propensity to pervert the incentives in the maxims of our power of choice. Thus it works, in the unconditional respect for the law which is the highest condition of all the maxims to be adopted . . . for the restoration to its purity of the predisposition in the human heart to the good. (R 6:50)
To be clear, Kant does not think we have a duty to acquire a feeling for the moral law, since he considers moral feeling, along with conscience, love of one’s neigh bour, and respect for oneself, as ‘subjective conditions of receptiveness to the con cept of duty’, meaning that such feelings are constitutive of our moral agency as such (MS 6:399). One’s duty regarding moral feeling is rather to ‘cultivate’ and ‘sharpen one’s attentiveness’ to it; hence, one’s duty is only indirect (MS 6:401). But even if moral feeling constitutes our receptiveness to duty, cultivating our view of humanity’s moral vocation could help strengthen that receptiveness.
Self-Knowledge and Despair 141 Reminding ourselves of our dignity, and the dignity of others, could thereby enliven our resolve to improve ourselves.15
5.6 The Particular Self At least as it stands, the theory of generic self-knowledge I have presented so far remains exposed to a potential criticism. One could argue that knowing what is imputable to me as a member of the human species does not bear upon me as a particular moral agent. Knowing that I have a propensity to self-conceit or a pre disposition to the good, for example, involves nothing more than knowing what I share in common with all others; but it says nothing about the details of my moral condition, about my motives, my habits, my traits, etc. A closer look at how Kant formulates the duty of self-knowledge will alleviate this concern. As he writes, the duty is to cognize not only what belongs ‘originally to the substance of a human being’, but also what belongs to my moral condition, which is ‘derived (acquired or developed)’ (MS 6:441). In the latter case I must ask myself, as one author puts it, ‘why did I act that way? Am I a generally sympathetic person? Stingy or open handed? Quick to anger or unduly self-effacing? What is most important to me? What attitudes, desires, and beliefs guide the over-all structure of my character and actions? Most importantly, have I really placed the pursuit of moral ends above my pursuit of self-love?’16 Questions like these force me to consider my unique position as a moral agent. I must seek particular knowledge of what kind of person I am, and whether my heart is good or selfish. But does it make sense to say I am under obligation to know my particular attitudes, desires, beliefs, habits, actions, and intentions? Surely I know these bet ter than anyone else does. Or do I? Experience brings me to the view of humani ty’s evil, of which I recognize exists generically in myself. But with respect to my particular heart, I might be inclined to judge myself more leniently than I do the human species as a whole. One reason for this is that I am just a single person, capable of limited wrongdoing, but the list of humanity’s misdeeds is of no com prehensible end. The human species brings to view more examples of evil than I could possibly produce on my own. Another reason is that I have, or at least think I have, a privileged outlook regarding my own intentions, which I of course lack when judging others. Kant observes that people often produce morally forgiving self-conceptions, which make them less inclined to accept or acknowledge their own shortcomings. All of this suggests that I ought to know the particulars of my 15 Kant touches on this idea in a later text, when he tells us that ‘reason, in representing the morally good by connecting its ideas with intuitions (examples) that have been imputed to them, can produce an enlivening of the will (in spiritual or political speeches to the people, or even in solitary speeches to oneself)’ (Anth 7:254). 16 Grenberg (2005, 226).
142 The Particular Self moral condition, my own failures, shortcomings, and misdeeds—especially since I am less inclined to acknowledge these things on my own. Kant even maintains that this kind of particular self-knowledge is a prelimin ary step to ‘develop the original predisposition to a good will’ (MS 6:441). But this opens up a more difficult issue. While I can easily cognize the general command that I ought to become a better person, how can I be certain that my own dispos ition has actually improved? How can I be certain of a genuine change toward the good in my own heart? I may be subjectively convinced that I have changed for the better, that I have become a ‘new person’—but how can I be certain of this? The problem is not that my underlying character is inscrutable (type-1 opacity), but that my resolution to change for the good could be another mani festation of the dear self (type-2 opacity). Kant is aware that ‘one is never more easily deceived than in what promotes a good opinion of oneself ’ (R 6:68). And what better opinion can one have than a belief in one’s newfound goodness, purity, or righteousness? Consider Oscar Wilde's character of Dorian Gray. Gray remained convinced that his act of sparing a girl the shame of social scandal was evidence of his new moral character. Only when he perceived the visible stains of sin on his portrait, elsewhere described in the novel as his ‘conscience’ and the ‘mirror of his soul’, did he come to realize the dark truth of the matter: Had it been merely vanity that made him do his one good deed? Or the desire for a new sensation, as Lord Henry had hinted, with his mocking laugh? Or that passion to act a part that sometimes makes us do things finer than we are our selves? Or, perhaps, all of these? . . . Had there been nothing more in his renunci ation than that? There had been something more. At least he thought so. But who could tell? . . . No. There had been nothing more. Through vanity he had spared her. In hypocrisy he had worn the mask of goodness. For curiosity’s sake he had tried the denial of self. He recognized that now.17
Dorian Gray was subjectively convinced in the reality of his moral conversion; he ‘felt’ the change, so to speak. In truth, his disposition remained the same, and what he considered his one virtuous deed was nothing more than what Kant would call a bit of ‘moral enthusiasm’. Moral enthusiasts like Gray take up the commands of duty not with an attitude of respect, but with a ‘frivolous, high-flown,
17 Wilde (1897/2004, 252–3). Moral enthusiasts thus mistake the motive of dutiful action in some empirical feeling rather than in the law itself. Someone who adopts the maxim of sympathy to others, for example, only perceives the moral worth of the maxim itself, and thereby fails to notice the path ology of his disposition. Kant therefore argues: ‘Actions of others that are done with great sacrifice and for the sake of duty alone may indeed be praised by calling them noble and sublime deeds, but only insofar as there are traces suggesting that they were done wholly from respect for duty and not from ebullitions of feeling’ (KpV 5:85).
Self-Knowledge and Despair 143 fantastic cast of mind’, as if the execution of duty were something meritorious and a call for celebration (KpV 5:85). Gray’s disillusionment is instructive: it shows we cannot rely on introspection to determine the purity of our maxims or our moral character in general. The case of Dorian Gray also shows us that the greater threat to particular selfknowledge is not type-1 opacity. For Kant, perfect self-knowledge is impossible to attain, but that in no way detracts from our duty to pursue moral self-cognition, however imperfect that cognition will be. Without doubt the greater threat is selfdeception (type-2 opacity), because it is essential that the agent appraise her moral worth and moral progress sincerely. If self-deception can undermine the sincerity of the agent’s inward appraisal, then the duty of particular self-knowledge would be impossible to approximate, and the duty itself would become contra dictory. We would have an obligation to know something we cannot know at all, which violates Kant’s formula that every duty assumes our ability to act on it—that ‘ought implies can’. So in light of Kant’s remark that ‘only the descent into the hell of self-cognition can pave the way to godliness’, we must now ask: How can I ever be certain that my ascent toward godliness is not, in truth, a plunge into moral enthusiasm? Another way of expressing the problem before us is to say that Kant’s formula, ‘ought implies can’, does not address how I can be certain of my particular change of heart. The ought of the moral law implies that I am capable of meeting its demands; yet it does not imply that I can cognize the success of my moral aspir ations. In light of these difficulties Kant writes: ‘All that a creature can have with respect to hope for this share [in the highest good] is consciousness of his tried disposition, so that, from the progress he has already made from the worse to the morally better and from the immutable resolution he has thereby come to know, he may hope for a further uninterrupted continuance of this progress’ (KpV 5:123). In a footnote to this passage he offers the following caveat: ‘Conviction of the immutability of one’s disposition in progress toward the good seems, never theless, to be in itself impossible for a creature’ (KpV 5:123n; emphasis added). Once again we find a tension with Kant’s ethics and the opacity thesis. The moral law commands us to restore the original goodness of our heart. But it seems that we can never know for certain whether this restoration has really occurred.
5.7 The Inferential View This tension reaches a peak in the Religion. Here Kant acknowledges, much more so than in his previous writings, the agent’s need to have ‘assurance of the reality and constancy of a disposition that always advances in goodness’ (R 6:67). I need to know that my change of heart is authentic and that it will not relapse into self-conceit. But Kant finds the traditional responses to this desired assurance
144 The Inferential View unsatisfactory, and for good reason. On one extreme end, there is the view that a supernatural power will sanctify my resolution if it is genuine (e.g., ‘His Spirit gives witness to our spirit’). On the other end, there is the view that I cannot have any degree of assurance in my changed disposition, so that I must live in a con stant state of ‘fear and trembling’ (R 6:68). The former view contradicts the limits of human understanding (type-1 opacity) and can quickly turn to a form of reli gious enthusiasm, whereby people claim to perceive the effects of grace in others or in themselves. The fear-and-trembling approach, if taken too far, leads to what Kant calls the ‘darkest enthusiasm’, by which he means a kind of obsessive selfscrutiny such as we find in the diaries of Pascal and Haller (Anth 7:133). The opacity thesis no doubt has an intended element of fear and trembling in it, but Kant admits that ‘without any confidence in the disposition once acquired, perseverance in it would hardly be possible’ (R 6:68). Thus he thinks that the con cept of assurance has a negative, but indispensable, function in moral life. That is to say, it does not contribute directly to my restoration (because the concept pre supposes that such restoration has already taken place), but it gives me confidence that my moral progress is authentic, thereby motivating me to continue on the path of virtue.18 What is at stake here is not merely my subjective commitment to self-improvement. For without that commitment my motivation to become a bet ter person would become deaf to the demands of duty, no matter how loudly those commands abate in my soul. Failing to account for assurance in our moral restoration, which itself gives one ‘hope of absolution’, would therefore have a crippling effect on the moral agent (R 6:76), leading one to what Kant calls ‘wild despair’ (R 6:71). Even in the second Critique, we find that Kant is sensitive to this problem: Someone who is aware of having persisted through a long portion of his life up to its end in progress to the better, and this from genuine moral motives, may very well have the comforting hope, though not certitude, that even in an exist ence continuing beyond this life he will persevere in these principles . . . In this progress which, though it has to do with a goal endlessly postponed, yet holds for God as possession, he can have a prospect of a future beatitude. (KpV 5:123n; emphasis added)
18 The concept of assurance from the Religion is in many respects similar to Kant’s discussion of contentment from the second Critique. He asks, ‘Have we not, however, a word that does not denote enjoyment, as the word happiness does, but that nevertheless indicates a satisfaction with one’s exist ence, an analogue of happiness that must necessarily accompany consciousness of virtue? Yes! This word is contentment with oneself [Selbstzufriedenheit], which in its strict meaning always designates only a negative satisfaction with one’s existence, in which one is conscious of needing nothing’ (KpV 5:117). The psychological state of ‘self-contentment’ would be a matter of having what we call in colloquial terms a clear conscience, i.e., knowing one has done all that is within one’s power.
Self-Knowledge and Despair 145 Kant develops the first half of this proposition further in the Religion, leaving questions of beatitude and grace to the parerga of rational religion (R 6:53). He argues that we can acquire confidence in our restored disposition ‘without delivering ourselves to the sweetness or the anxiety of enthusiasm, by comparing our life conduct so far pursued with the resolution we once embraced’ (R 6:68). The process involves ‘observing ourselves through actions’ (V-Mo/Collins 27:365)—or what I shall call the Inferential View. It consists of inferring our dis position, whether it is good or evil, from the quality of our past actions: [Take] a human who, from the time of his adoption of the principles of the good and throughout a sufficiently long life hence-forth, has perceived the efficacy of these principles on what he does, i.e., on the conduct of his life as it steadily improves, and from that has cause to infer, but only by way of conjecture, a fun damental improvement in his disposition. (R 6:68)
Kant suggests that an assessment of my life conduct from the time I have resolved to change for the good is sufficient to secure my confidence in the authenticity of that change. If the moral curve of my past actions shows steady and uninterrupted progress, I can envision a ‘boundless future’ which is ‘desirable and happy’ (R 6:69; cf., V-Phil-Th/Pölitz 28:1087). But if my past actions are inconsistently virtuous, sometimes slipping back into self-conceit, I can only envision a future of misery, one sinking deeper into corruption. Thus the image of a happy future, according to Kant, allows us to infer the authenticity of our restoration. Even here, however, Kant does not for a moment compromise the opacity thesis. He quickly adds to his discussion the warning that we ‘cannot base this confidence upon an immedi ate consciousness of the immutability of our disposition since we cannot see through to the latter but must at best infer it from the consequence that it has on the conduct of our life’ (R 6:71). Assurance in our restored disposition can never be a matter of perfect certainty. By focusing on moral conduct the Inferential View avoids the many difficulties we encountered with introspection. Action-based inferences, for example, are not dependent on the agent’s often-deluded inner states (‘I feel like a new person’; ‘I consider myself restored’). However, in the course of his argument Kant begins to detect a new set of problems with what I am calling the Inferential View. The first is that an assessment of the agent’s actions does not allow her to infer the strength or stability of her restoration (R 6:71). Someone may have a long history of good deeds behind her, but that does not allow us to conclude that if placed before a morally challenging situation, she would remain steadfast to her virtuous prin ciples. The second problem, which Kant does not address, is that it is extremely difficult to say what time period is ‘sufficiently long’ for the agent to legitimately infer a restored disposition from her actions. The character of Dorian
146 The Comparative View Gray shows us that one virtuous deed does not suffice. But where are we to draw the line? Would a month of virtuous deeds suffice? Two years? Half a lifetime? A related difficulty, which Kant only hints at, is that if we do require a long his tory of ‘empirical proofs’ of the agent’s change of heart, then an agent’s impending death would render the production of such proofs impossible. Suppose, for example, that I have undergone a genuine change of heart the same day I discover I will soon die from a terminal illness. According to the Inferential View even if I spent the rest of my allotted time performing virtuous deeds, I would lack a suf ficient amount of evidence to infer that I have become a new person. The issue of impending death also brings us back to the first difficulty discussed above. Elsewhere Kant addresses the scenario of an evildoer suddenly possessing an ‘honorable and upright disposition’ on his deathbed. Without certainty in his restored disposition, he would have no basis to tell if his moral conversion was real.19 Kant therefore advises, ‘a man always has to get to know himself in a grad ual fashion’ (V-Mo/Collins 27:365). But then we must ask, what is ‘gradual’?, which brings us back to the second difficulty of the Inferential View.
5.8 The Comparative View Any one of the difficulties we have outlined threatens to undermine the Inferential View, once again leaving the moral agent without assurance in his or her change of heart. Aware of these difficulties Kant modifies his account of inferential selfknowledge in the closing paragraphs of Section One, Book Two of the Religion, which I shall call the Comparative View.20 Instead of framing the question of the agent’s assurance in terms of the future life he could hope to lead in view of his past actions (boundless happiness and continued improvement, or boundless misery and continued evil), the Comparative View frames the question in terms of the verdict the agent could hope to receive were his whole life placed before a judge: 19 In general, Kant’s views concerning the possibility of a change of heart later in one’s life seem inconsistent. On some occasions he argues that such a change is hardly possible if one has led a long life of viciousness: ‘The human being who has always led a depraved life and wants to be converted in an instant cannot possibly get there, for it would be nothing short of a miracle for him to become in an instant the same as someone who has conducted himself well during his entire life and always thought upright thoughts’ (Päd 9:488). On other occasions, Kant argues for what appears to be the opposite view: that only after one becomes weary of an unstable life led by instinct does one decide to ground one’s character in reason. Kant even states that such revolutionary decisions are rare, even before the age of forty (Anth 7:294). 20 Kant hints at a third alternative in one of the Religion’s footnotes, what we might call the Transformative View. If an agent has genuinely restored his disposition, so the argument runs, he will gladly take on any punishments attributable to his old disposition. Thus, the agent transforms the meaning of ‘punishments’ into ‘so many opportunities to test and exercise his disposition for the good’ (R 6:75n). Here the ‘proof ’ of the agent’s restored disposition rests in his newfound pro-attitude toward the punitive consequences of his past deeds.
Self-Knowledge and Despair 147 Since he can derive no certain and definite concept of his disposition through immediate consciousness but only from the conduct he has actually led in life, he shall not be able to think of any other condition of being delivered to the verdict of a future judge . . . than that his whole life be one day placed before the judge’s eyes, and not just a segment of it, perhaps the last and to him still the most advantageous. (R 6:77)
Recall that the Inferential View only takes into consideration the agent’s moral conduct after her supposed restoration. One might think that assessing the agent’s whole life, including her old disposition, would negatively affect the agent’s hopedfor verdict. But this assessment can also have a positive upshot, since appraising the agent’s life as a totality allows the hypothetical judge to perceive the actual change that has taken place within her. If I have resolved to change for the good, one way to assess my resolution would be to reflect on the differences between my present and past moral conduct. By considering the agent’s old disposition, one could ‘examine what and how much of this disposition he has cast off, as well as the quality (whether pure or still impure) and the grade of the supposed new disposition for overcoming the old one and preventing relapse into it’ (R 6:77). The more perceived difference there is between the two, the more readily I can infer a change of heart that is both real and unwavering. The Comparative View thus overcomes the two general difficulties of the Inferential View. First, it allows us to infer the stability of the agent’s restored dis position (by way of comparison with the old one), something we cannot infer on the basis of improved moral conduct alone. Second, this view does not require a substantial time-period to legitimate the inference of the agent’s restoration. While some duration of time is necessary for the agent to exhibit her new moral character, the inference is grounded, not in time-duration, but in the perceived difference between the agent’s old and new ways of conduct. So impending death need not throw the agent into wild despair. Nevertheless, one problem still threatens to undermine the Comparative View, and this concerns the nature of the hypothetical judge. If we cannot establish the judge’s authority, we have no reason to accept its final verdict. For Kant, there is a deep connection between how we behave before this judge and how we represent it. He maintains, for example, that representing the judge of the agent’s whole life as another (i.e., God), ‘of whom news [of the agent’s restoration] will be had through sources of information elsewhere’, will have a detrimental effect on the agent’s moral conduct. For then the accused will have much with which to counter the judge’s severity under the pretext of human frailty; he will think he can get around him, whether by forestalling his punishment through remorseful self-inflicted torments that do not, however, originate in any genuine disposition toward improvement or by mollifying
148 Conscience: The Inner Court him with prayers and entreaties, even with incantations and self-proclaimed professions of faith. (R 6:77)
Simply by locating the judge outside the agent’s consciousness, Kant believes the accused will attempt to assuage the judge’s verdicts through an enthusiastic dis play of false piety and righteousness. On the other hand, if we represent the judge as oneself, what Kant calls the ‘judge within him’, he believes the agent will thereby ‘pronounce a stern judgment upon himself, for he cannot bribe his reason’ (R 6:77). Kant refers to the ‘judge within’ as one’s reason, but it is more precise to say that it is one’s ‘conscience’. As he writes in an earlier lecture, ‘conscience, that judge in us which is not to be bribed, will place before the eyes of each one the whole world of his earthly life and convince himself of the justice of the verdict’ (V-Phil-Th/Pölitz 28:1087). Still, introducing the concept of conscientious self-judgment raises more ques tions than it solves. We might first ask whether the notion of an ‘inner judge’ is even intelligible, for how can I truly condemn myself? If I am responsible for issu ing the verdict on my life as a whole, would I not be tempted to deceive myself, to render my life acquitted even if an impartial judge would render me guilty? It seems intuitive at least to identify a judge’s impartiality with his or her distance from the accused. And therein lies the puzzle. For at the most crucial juncture of his argument, when assurance in our restoration is at stake, Kant seems to have fallen victim to an odd form of optimism. He has effectively placed the question of the agent’s assurance in her own hands, so that she herself—and no one else— is responsible for judging her life as a whole. The Comparative View loses all legitimacy, however, if the agent can delude herself on the level of her own selfassessment. So the issue we must now seek to resolve is how conscience, or the ‘judge within’, is immune to the workings of self-deception (i.e., type-2 opacity).
5.9 Conscience: The Inner Court I believe we can find the beginning of a solution by moving ahead four years from the Religion to the Metaphysics of Morals, for only in the latter work does Kant offer a full treatment of the ‘inner judicial court’ called conscience.21 While a complete summary of Kant’s discussion falls outside the scope of this book, 21 For lack of space, I have refrained from exploring the historical sources of Kant’s idea of con science. It is common to see this idea in a Christian-Lutheran (specifically German Pietist) light, as many commentators of Kant do. However, I think this view can obscure other, equally important, historical sources. To mention a few, one can trace Kant’s idea of conscience to Rousseau’s Emile, and from there to British Sentimentalists such as Butler and Shaftesbury—all three of which draw heavily from Stoic sources (Epictetus’s Discourses in particular).
Self-Knowledge and Despair 149 I would like to look at the identity of conscience and its judicial functions. In what follows I will suggest that adding Kant’s theory of conscience to the Comparative View offers a solution to the problem of possible deception in self-judgment.22 In the Metaphysics of Morals, Kant addresses the apparent contradiction of identifying the ‘inner prosecutor’ and the accused, for ‘to think of a human being who is accused by his conscience as one and the same person as the judge is an absurd way of representing a court, since then the prosecutor would always lose’ (MS 6:438). Given this passage, I think it would be a mistake to construe Kant’s talk of ‘courtrooms’, ‘prosecutors’, ‘defence counsels’, and ‘final verdicts’ as nothing more than metaphors. The distinctions Kant wishes to establish within conscience are normative, not merely symbolic. He argues, for instance, that ‘one constrained by his reason’ must necessarily represent the accusations of conscience as the accusations ‘of another person’ (MS 6:438). He clarifies this idea in a footnote, pointing out that a human being who accuses and judges himself in conscience must think of a dual personality [zwiefache Persönlichkeit] in himself, a doubled self [doppelte Selbst] which, on the one hand, has to stand trembling at the bar of a court that is yet entrusted to him, but which, on the other hand, administers the office of judge that it holds by innate authority. (MS 6:438n)
The ‘doubled self ’ refers to the human being’s twofold membership in the intelli gible and sensible worlds. Having already discussed the import of this distinction in Chapter 3, what I want to emphasize at present is Kant’s claim that the agent experiences her conscience, not as her empirical and corrupted will (or who she is), but as her pure and perfected will (or who she ought to be). The authority of conscience arises from the fact that we necessarily represent it as our ‘true’ or ‘proper’ self, which is why Kant describes conscience as the ‘inner judge of all free actions’ (MS 6:438). In this same context Kant outlines the following process of moral deliberation: • Practical understanding provides me with the ‘rules’ or ‘principles’ of morality that constrain my range of choices. These principles allow me to assess what I ought and ought not to do.
22 Kant’s theory of conscience has attracted increasing attention within the recent philosophical literature. I have benefited from the excellent accounts presented by Hill (2002) and especially by Wood (2007). Other authors who discuss Kant’s views on conscience are Munzel (1999) and Howard (2004).
150 Conscience: The Inner Court • The faculty of judgment then determines two things: - generally, whether my past or projected action has the status of what Kant calls a ‘deed’, an action falling within the jurisdiction of the moral law; and - specifically, whether my past or projected action properly incorporates the judgments of practical understanding—whether I acted (or will act) on what I judge to be my duty. • Finally, conscience issues the verdict on my action: - either immoral (‘guilty’) or - moral (‘not guilty’). ‘All of this takes place’, Kant writes, ‘before a tribunal . . . an inner court in the human being’ (MS 6:438). Notice that I take an active role in assessing my duties, in deliberating what actions the moral law, by way of my understanding, compels me to pursue or avoid. But my action (before or after it occurs) is passive to the appraisal of conscience (MS 6:439; cf., MpVT 8:269n). This is one of the senses in which I experience my conscience as another, for my conscience condemns me ‘spontaneously’ or ‘instinctually’ if I fail to act on what I judge to be my duty. Practical understanding is prone to error, however. This leads to the second judi cial function of conscience. I can assess my actions, and act on what I ‘objectively’ judge to be right, but still fail to properly incorporate the rules of morality into my actions. People make wrong moral judgments all the time. But how are we to make sense of Kant’s claim that an ‘erring conscience is an absurdity’ (MS 6:401)? If I understand him correctly, Kant’s idea is that I cannot fail to believe whether I have submitted my actions to the appraisal of practical understanding. I cannot fail to believe, in other words, whether I have consciously examined my duties. As he writes, while I can indeed be mistaken at times in objective judgment as to whether something is a duty or not, I cannot be mistaken in my subjective judgment as to whether I have submitted it to my practical reason (here in its role as judge) for such a judgment. (MS 6:401)
This helps explain Kant’s statement from the Religion that conscience is ‘the moral faculty of judgment, passing judgment upon itself’ (R 6:186).23 By this he means conscience judges the agent’s awareness in having thoroughly appraised her duties. The second judicial function of conscience is thus a higher-order judg ment of the care the agent applies (or fails to apply) in the act of examining what she ought to do.24 In this case I stand guilty before the inner judge not by failing 23 Hill offers the important insight that Kant speaks of conscience’s ‘judgment’ in two different senses: ‘Metaphorically speaking, ‘judgment1’ (one sense of ‘judgment’) is what is responsible for appraising the act diligently, and ‘judgment2’ (a second sense of ‘judgment’) on ‘judgment1’ as to whether it has fulfilled that responsibility’ (2002, 302). 24 Kant remarks that it is one’s responsibility to ‘enlighten his understanding in the matter of what is or is not duty’ but as soon as he acts (or is about to act), conscience ‘speaks involuntarily and
Self-Knowledge and Despair 151 to act on what I believe to be my duty but by failing to scrutinize, under the lights of practical understanding, what my duty is. While the first function of conscience is to see whether the agent’s actions really do line up with the judgments of practical understanding, the second func tion is to see whether the agent really does perform a self-critical assessment of her judgments (before or after those judgments take effect in action). Conscience thereby displays the capacity we have to judge our own judgments, so that when we take ourselves to be responding to a particular moral demand we can ask our selves, ‘Have I carefully reflected on how I should act?’ which is different from, ‘Have I really acted on what I judge to be my duty?’ If it turns out that we have assessed the appropriateness of our action haphazardly—or not at all—then con science condemns us. So at least pertaining to its higher-order function, Kant is right: an erring conscience is an absurdity, for the simple reason that an agent cannot critically assess her duties unconsciously. The additional qualification of a ‘careful’ assessment of one’s duties is second ary to the basic awareness one has of performing the assessment itself. In this way, we are incapable of disavowing the higher-order verdicts of conscience—whether or not we have reflected on the appropriateness of our moral actions—because we are conscious of whether we have or have not done so. Kant describes this type of self-conscious awareness in terms of truthfulness, which is the reflexive stand point we take to own moral judgments, as opposed to truth, which bears upon the right-making or wrong-making properties of an action (i.e., its lawfulness or unlawfulness). By drawing this distinction, Kant’s point is that while we can never know the objective truth-value of our moral judgments we can still have ‘imme diate consciousness’ of holding these judgments to be true (MpVT 8:267), and that is precisely the level on which conscience operates. Here it may appear as though Kant is contradicting his opacity thesis. How can I have immediate consciousness of the verdicts of my conscience? How can I claim to know, without doubt, that I stand guilty or not before the inner judicial court? Kant’s answer is that conscience accuses me exactly where I am transparent: my sense of truthfulness.25 But this transparency is not in the order of knowledge.
unavoidably’ (MS 4:401). Conscience is phenomenologically distinct from practical understanding in that its judgments are immediate and spontaneous, not deliberative and thoughtful. Kant argues for this point in his early lectures on ethics, when he speaks of the ‘instinct’ of conscience: ‘Everyone has a faculty of speculative judgment, though that is at our discretion; there is, however, something in us which compels us to pass judgment on our actions. It sets the law before us, and obliges us to appear before the court. It passes sentence on us against our will, and is thus a true judge’ (V-Mo/Collins 27:297). Thus, I am passive to my conscience in two different senses. First, conscience is not developed or acquired (it is constitutive of my moral agency, like moral feeling). Second, as the inner voice of the moral law, I necessarily represent conscience as ‘another’. It is in this sense that I experience my con science as ‘external’ to myself, i.e., that I experience the reproaches of conscience as if issued from another (morally ideal) person. 25 Kant calls this ‘formal conscientiousness’ (formale Gewissenhaftigkeit) (MpVT 8:268).
152 Conscience: The Inner Court By Kant’s definitions, ‘knowing’ is a matter of having objectively sufficient grounds to hold something as true, which is why we characterize knowledge in terms of universal agreement (for example, something shareable by every rational being). Only knowledge yields certainty; but certainty is not the only mode of taking something to be true. Kant defines ‘believing’, for instance, as having subjectively sufficient grounds to hold something as true in the absence of objective grounds; and while these grounds do not yield ‘certainty’, they do yield what Kant calls ‘conviction’ (Überzeugung) (A822/B850).26 To say that I am convinced of the ver dicts of my conscience does not contradict the opacity thesis, then, for that thesis only constrains the objective grounds of my knowledge-claims. I can, for example, maintain conviction in the honesty of my declaration, even though its objective truth-value falls outside the limits of my understanding. Kant is emphatic on this point: ‘I can indeed err in the judgment in which I believe to be right, for this belongs to the understanding which alone judges objectively (rightly or wrongly); but in the judgment whether I in fact believe to be right (or merely pretend it) I absolutely cannot be mistaken’ (MpVT 8:268). Similarly, I can err in the judgment of what my duty is, but I cannot err in my general conviction of having properly appraised my duties. Again, this is why I cannot avoid accepting the final verdict of my conscience, guilty or not guilty, because I am conscious of holding this verdict within myself. I know whether I sincerely examined my actions, or whether I lied to myself or to others. These actions are transparent to me because I cannot rationally deny my own belief in having performed them. I therefore have privileged access to my conscience because that access is grounded in a reflexive grasp of my own beliefs, and this is why Kant is at pains to distinguish the subjective character of conviction from the objective character of certainty.27 In light of the identity of conscience and its two judicial functions, I believe we can complete Kant’s Comparative View developed in the Religion. To summarize, the Comparative View infers the agent’s change of heart on the basis of the per ceived difference between her old and supposedly new ways of moral conduct.
26 See Chignell (2007), as well as the Introduction to the logic lectures compiled by Jäsche. 27 Kant is not denying that I can attempt to ignore the accusations of my conscience, to distract myself from the stirrings of the inner judge. His point is that precisely by attempting to ignore my own guilt, I am testifying against myself, if you will, in favor of conscience. Kant further argues that the privileged access we have to our own beliefs prevents us from immunizing ourselves from the onslaughts of conscience. As he writes: ‘A human being may use what art he will to paint some unlaw ful conduct he remembers as an unintentional fault. . . and to declare himself innocent of it; he never theless finds that the advocate who speaks in his favor can by no means reduce to silence the prosecutor within him, if only he is aware that at the time he did this wrong he was in his senses, that is, had the use of his freedom; and while he explains his misconduct by certain bad habits. . . this can not protect him from the reproach and censure he casts upon himself ’ (KpV 5:98). For Kant, firstpersonal awareness is necessarily tied to basic beliefs of our own agency. To say, ‘I acted’ in such and such way is implicitly to say, ‘I was free of constraint’ in so acting. Presumably, the possibility of an ‘inner prosecutor’ arises from our own subjective belief in having acted freely.
Self-Knowledge and Despair 153 The agent must therefore ask herself what kind of verdict she could hope to receive were her whole life placed before a judge. Recall Kant’s striking claim that if the agent represents this judge as another, she will be tempted to deceive it through an empty display of moral righteousness. But if she represents the judge as herself, Kant believes, she will appraise herself firmly. As we noted above, the legitimacy of this final verdict remains questionable until we can determine the authority of this judge. What we might call the Comparative-Conscientious View finally answers this question. In the first place, Kant’s theory of conscience overcomes the problem that threatened to undermine the Comparative View, which is how I can con demn myself. The normative dualism of conscience explains how I can stand accused before myself, something that is contradictory if the judge of my moral conduct turns out to be the dear self. This also explains why I cannot bribe or even disagree with conscience’s final verdict. While the purity of my disposition, whether it is actually good or evil, is impossible to cognize directly, I am immedi ately conscious of whether I have examined my life conduct with due care. Kant’s point, as I discussed above, is that I cannot fail to be aware of my own honesty or dishonesty, whether this applies to the actions I submit to the appraisal of prac tical understanding, or more simply, to the sincerity of my testimony. Conscience alone does not establish assurance in my restored disposition, but rather allows me to trust the final sentence I pass on my life as a whole. While comparative self-knowledge is responsible for assessing the perceived difference between my old and new ways of life, conscience is responsible for condemning or acquitting me in my effort (or lack thereof) to examine this difference dili gently. Together, the Comparative-Conscientious View not only supplies me with confidence in the authenticity of my restoration, which is essential for the con tinual pursuit of virtue; it also gives me confidence in my own self-assessment. Only by adding conscience to the Comparative View can we account for the fun damental self-trustneeded in our effort to fulfil the duty of self-knowledge. Of course, this does not require us to have one of Hawthorne’s moments, those moments ‘when a man’s moral aspect is faithfully revealed to his mind’s eye’, for that would contradict type-1 opacity. Nor am I claiming to infer my disposition and progress toward the good by way of my actions alone, for actions are not judgment-neutral. I still need to exhibit my moral character through my lifeconduct; and to this extent, comparative self-knowledge is still inferential. But the deeper question here is how I can trust the inferences I draw from my conduct as a whole. The advantage of adding conscience to the Comparative View is that we can finally overcome the major threat to the duty of self-knowledge, namely, the dear self. All the dear self can do is obscure the empirical evidence I bring before the ‘inner court’, such as the list of past actions evincing my moral improvement; and it can do this because I will never have objective grounds to assess my moral disposition. But the dear self cannot corrupt my awareness of having appraised
154 The Depths of the Human Heart my life thoroughly. This entails that the truth of the evidence I bring before the inner court of conscience is fallible, because it can always be corrupted by type-2 opacity, but the truthfulness of my attempt to examine this evidence is beyond corruption. And that is why an erring conscience is an absurdity. Kant’s claim in the Religion that one should stand in judgment before oneself makes sense only if we replace ‘oneself ’ with ‘one’s conscience’, something he alludes to when speaking of the ‘judge within’. This explains why later in the Metaphysics of Morals Kant situates conscience, or the ‘human being’s duty to himself as his own innate judge’, before the duty of self-knowledge. In my under standing, this is because the authority of conscience, as the access we have to our own beliefs, is an essential requirement for the duty of self-knowledge. Now one might ask, Why is self-knowledge the first command of all duties to oneself? Should we not give primacy to the duty of conscience instead? Just as I cannot properly pursue my natural perfections until I know the purity of my heart, whether it is good or evil, I cannot properly judge my heart until I can trust my self-judgments. Is not self-knowledge only possible on the grounds of conscience, on our capacity to judge ourselves sincerely? Kant’s answer is affirmative: The duty of self-knowledge is grounded in con science. But, he argues, the concept of a duty to conscience is contradictory for the simple reason that ‘conscience is not something that can be acquired’ (MS 6:400). Rather, ‘every human being, as a moral being, has a conscience within him originally’ (MS 6:400). Conscience, like moral feeling, is thus constitutive of our identity as moral agents. And like moral feeling, our duty to conscience is only indirect. One’s obligation here, Kant maintains, is to ‘cultivate one’s conscience, to sharpen one’s attentiveness to the voice of the inner judge and to use every means to obtain a hearing for it’ (MS 6:401). Conscience conditions the possibility of self-knowledge, but self-knowledge is still the first duty to oneself.
5.10 The Depths of the Human Heart I argued earlier in this chapter that the duty to know my particular heart is impos sible if I can never have any degree of conviction in my restoration. Kant himself admits that without such conviction one would be led into hopelessness and despair. It would be rather disturbing if generic self-knowledge led me to the insight that I ought to become a morally better person but that from the stand point of my particular self I could never know whether my attempts to become good were genuine or fake. The duty of self-knowledge clearly requires the agent to apprehend the particular evils that stand in the way of her restoration, but for the sake of continued motivation and perseverance she must also be able to tell, if only slightly, that her self-improvement will not slip back into self-conceit. For Kant, it is not within the power of one’s conscience to grant rewards for good
Self-Knowledge and Despair 155 behavior, which is why the ‘comforting encouragement of one’s conscience is not positive (joy) but merely negative (relief from preceding anxiety)’ (MS 6:440). Perhaps the most we can hope for regarding the assurance of our moral restor ation is a clear conscience, the conviction that we have appraised our life-conduct cautiously and with due care. To live with a clear conscience would be to live free from anxiety, free from hopelessness and despair. And that is perhaps the most we can ask for, given that the depths of the human heart are, after all, unfathomable.
Conclusion What I hoped to have accomplished in this book is a plausible alternative to the standard story of Kant’s justification of ethics. After showing that this story is relatively new in the reception of Kant’s work, I suggested that readers have found it compelling because it purports to explain Kant’s worry about a hidden circle in Groundwork III. On my alternative reading, the idea of an intelligible world averts the circle, but only because our consciousness of its law, i.e., the moral law, legit imates our claim to freedom. Turning to the second Critique, I showed that Kant’s later references to the Groundwork are all positive, and his comment about a ‘vainly sought deduction’ of the moral law is consistent with the conclusion of Groundwork III. Indeed, as we saw, Kant’s early readers frequently cited his state ment from KpV 5:8 that the second Critique ‘presupposes’ the Groundwork to frame the relationship between these two texts. Nothing like the present-day reversal reading appears until Schopenhauer’s prize essay on morality; and noth ing like an established consensus in the literature appears until 1960. In this final set of reflections, I would like to return briefly to the legacy of Kant’s project— sketched in Chapter 3—focusing first on its early reception (in the work of Karl Reinhold, Leonard Creuzer, and Solomon Maimon), and then jumping ahead to contemporary Kantian arguments for the normativity of moral requirements.1 Surprisingly, we shall see that these contemporary arguments are closer, both in spirit and strategy, to those first post-Kantians who claimed to be revising or rejecting Kant’s position. Signs of dissatisfaction with how Kant framed the connection between free dom and morality are evident in Reinhold’s Attempt at a New Theory of Human Representation, first published in 1789. In this work Reinhold claims that ‘he who has not philosophized about freedom is as convinced about its actuality as his 1 Nagel’s (1970) strategy for justifying moral (what he calls ‘altruistic’) requirements of practical reason turns on appealing first to ‘more fundamental forms of practical reasoning and the conception of oneself to which they are related’ (1970, 143). Gewirth (1978) appeals to what he calls a ‘principle of general consistency’. A similar kind of foundationalist strategy underpins Korsgaard’s attempt to jus tify the normativity of moral reasons, either on the basis of one’s ‘practical self-identity’ (1996b) or on the basis of fundamental norms constitutive of action and agency as such (2008, 2009). Parallel strat egies for deriving moral normativity are found in Railton (1997), Velleman (1989, 2000, 2009), and Rosati (2003), although their aims are comparatively modest insofar as they are seeking to establish something less robust than a Kantian view of the moral law’s unconditional necessity. However, my reason for grouping them under the foundationalist approach to moral justification is that they share one thing in common: namely, an attempt to find a non-moral route to moral normativity. That is the aim these modern authors share in common with Kant’s early critics.
Kant’s Justification of Ethics. Owen Ware, Oxford University Press (2021). © Owen Ware. DOI: 10.1093/oso/9780198849933.003.0007
Conclusion 157 own existence’.2 Freedom, in other words, qualifies as a ‘fact’ (Tatsache). However, Reinhold is careful to distance this fact from the moral law. Freedom is a Tatsache, he writes, that one ‘knows from his inner experience’ and that one is ‘conscious of through self-feeling’.3 The epistemic ground of freedom is therefore independent of the concept of an absolute practical law. We have access to it, Reinhold argues, simply through an inner feeling of activity, of which only philosophers are in the habit of doubting. What is interesting is that Reinhold does not reject the status of the moral law as a fact; on the contrary, he argues that to ask, ‘Is there a cognitive ground of the moral law?’ amounts to the question, ‘Is there a moral law?’ which he says nobody, not even philosophers, sincerely call into question.4 Yet it is clear, both from this text and from the book version of his Letters on Kantian Philosophy (1792), that Reinhold accepts the status of the moral law as a fact without making it the basis for accessing our freedom. A similar position appears in Creuzer’s 1793 Skeptical Observations on Freedom of the Will, where he argues that judgments concerning what ‘happens’ and what ‘ought to happen’ are part of the most common human understanding. One need only ‘hear’ the moral law, Creuzer says, to ‘understand immediately what it is, namely, an unconditioned, unlimited, unchangeable, and universally valid norm of our actions’—a norm, he adds, that even the ‘greatest evildoer’ recognizes in his heart.5 For this reason Creuzer calls the moral law an ‘undeniable fact of human nature’. and he appears to side with Kant’s disclosure thesis in saying that one ‘cognizes himself as a member of the supersensible world’ through this fact.6 In fact, Creuzer claims that ‘independence from foreign laws and freedom are therefore inseparably bound with one another’ and that ‘consciousness of free dom is, like consciousness of the moral law, a fact of reason [ein Factum der Vernunft]’.7 However, Creuzer qualifies his position in a footnote, saying that he agrees with Kant in making freedom the essential ground of the moral law (its ratio essendi), but he disagrees with Kant in making the moral law the cognitive ground of freedom (its ratio cognoscendi).8 In Creuzer’s view we have no reason to accept the disclosure thesis because consciousness of freedom is ‘already active before the development of the moral law’.9 In line with Reinhold, Creuzer recom mends that we should seek to explain freedom as its own ‘original immediate consciousness’ apart from of our notions of duty, obligation, or law.10 Nor were Reinhold and Creuzer alone in advocating this separation. It set the backdrop against which Maimon would propose to ‘improve’ upon Kant’s moral philosophy, starting with his 1794 essay ‘Attempt at a New Presentation of the 2 Reinhold (1789, 91). 3 Reinhold (1789, 91–2). For a detailed study of this early post-Kantian period, with a focus on Fichte, see Ware (2020). 4 Reinhold (1789, 101). 5 Creuzer (1793, 3). 6 Creuzer (1793, 7). 7 Creuzer (1793, 8–9). 8 Creuzer (1793, 9n). 9 Creuzer (1793, 9n). 10 Creuzer (1793, 9n).
158 Conclusion Moral Principle and a New Deduction of Its Reality’. What is unique about Maimon’s contribution is that he criticizes Kant’s methodology for its ‘unscientific’ reliance upon ‘common human understanding’. Common human understanding, he says, is prone to error and illusion, and so there is no reliable way to tell if the concepts we develop from this standpoint rest on mere ‘psychological deception’.11 It may be an ‘immediate fact of consciousness’ that the moral law issues its com mands unconditionally, but that in itself tells us nothing about the moral law’s objectivity. Even if we feel ourselves under moral constraint, for instance, how do we know that the moral law is really binding upon our will? In reply, Maimon offers a new methodology that begins with a more primary ‘fact of consciousness’. The specific Tatsache he argues is more primary than our everyday consciousness of duty is our ‘drive for the cognition of truth’ (Trieb zur Erkenntnis der Wahrheit),12 a fact he says is not suspect of psychological deception. And this last point is crucial, since Maimon goes on to claim that as rational beings we neces sarily strive to meet a principle of ‘universal validity’ in our thoughts, the same principle, he contends, under which we necessarily strive to meet the ‘demands of duty’ in our actions.13 This link to universal validity—a general principle of rea son as such—is what secures the objectivity of the moral law. Or so Maimon argues. In this way Maimon rejects, much more clearly than his contemporaries, Kant’s claim that the moral law gives us knowledge of our freedom. By starting with our drive for the cognition of truth, Maimon does not presuppose Kant’s doctrine of the fact of reason, insofar as this doctrine operates from the standpoint of com mon human reason. For Kant, of course, the buck stops with our consciousness of the moral law, since there is no alternative means for accessing the concept of an absolute practical principle. That is why Kant says, in answer to the question, ‘How is consciousness of that moral law possible?’, that we need only attend to its necessity (KpV 5:30), as we saw in Chapter 2. While Reinhold and Creuzer seem to agree with this point, they both deny that the moral law reveals our freedom to us, either because we can access our freedom through self-feeling (Reinhold), or because our consciousness of freedom is active prior to the moral law (Creuzer). But neither of them went as far as Maimon in raising the skeptical possibility that Kant’s Factum might be a grand delusion, and neither of them went as far as Maimon in developing a foundationalist strategy for securing the moral law’s objectivity. Turning our attention to the contemporary landscape, it is nothing less than a remarkable twist in the history of ethics that a version of this foundationalist strategy would resurface on American soil nearly two hundred years later. For a common theme one finds in the work of Nagel, Korsgaard, Velleman, and others, is that the normativity of moral requirements can be derived from a more basic
11 Maimon (1794, 404).
12 Maimon (1794, 407).
13 Maimon (1794, 419).
Conclusion 159 conception of action or even from a more basic conception of theoretical ration ality. However, given everything we have discussed in this book, it seems to me that Kant would have been just as resistant to present-day foundationalism in ethics as he was to the revisionary strategies of Reinhold, Creuzer, and Maimon, all of whom were seeking to derive the concepts of ethics from non-moral premises. For Kant, nothing about such premises can reveal the necessity of the moral law to us, and only our consciousness of this necessity gives us knowledge of our freedom. Of course, it is difficult for us to evaluate this move today, if not in terms of Kant’s failure to secure a non-moral route to the moral law’s normativity. Foundationalism, or something like it, may seem like a more promising paradigm for moral justification. Yet we should not lose sight of the fact that Kant drew attention to the limits of his argument in Groundwork III, saying that it is a ‘reproach that must be brought against human reason in general, that it cannot make comprehensible as regards its absolute necessity an unconditional practical law (such as the categorical imperative must be)’ (G 4:463). Whether or not we agree with Kant on this issue, we should not assume in advance that a successful justification of morality is one that proceeds from an independent premise (either theoretical, or practical yet non-moral).14 We should not assume that foundationalism is our ideal proofstructure. Kant believed that moral consciousness itself is the only viable orienta tion a project of justification can take, because it is the only orientation that works within the bounds of common human reason. At this final juncture, then, it may seem that we are left with a standoff of intu itions about what the task of justification in ethics should accomplish. Those who think it should refute all forms of moral skepticism will likely find my reading of Kant here disappointing. For if I am right, Kant was not seeking to justify the absolute necessity of the moral law in any of his mature writings—indeed, that is precisely what he thought goes beyond our comprehension and thus beyond the scope of any deduction. Naturally those who insist that we should be seeking such a deduction will wonder what kind of justification of ethics Kant ends up with, if any. To call our consciousness of the moral law a Factum hardly seems satisfying, even if we follow Kant’s repeated claim that this fact is part of a shared, pretheoretical and hence common standpoint. However, it is worth repeating that Kant did not regard refuting the skeptic to be the measure of success for his moral philosophy. As I have argued in the foregoing chapters, he was very much speaking to those of us who are committed to morality, but who lack the self-understanding 14 Referring to recent Kantian arguments for moral normativity, Tenenbaum writes: ‘Impressive and interesting as many of these arguments certainly are, it is hard to avoid approaching them with the sense that someone is trying to sell us the Brooklyn Bridge. It might be the case that we should be grateful to Kant for leading us away from the attempt to derive the moral law from a thinner notion of freedom or rationality’ (2012, 589).
160 Conclusion required to make this commitment fully intelligible. That in itself does not settle the standoff of intuitions we have, but it does bring an important point to light. It shows that how we frame the task of justification in ethics comes down to where we locate the center of our discussion, whether in philosophical speculation or in ordinary life. In arguing from a moral point of view, Kant was at least staying faithful to his own conception of what the highest system of reason can strive to accomplish. And his advice to the ‘schools’ of eighteenth-century philosophy is perhaps something we would benefit from listening to today: The schools are instructed to pretend to no higher or more comprehensive insight on any point touching the universal human concerns than the insight that is accessible to the great multitude (who are always most worthy of our respect), and to limit themselves to the cultivation of those grounds of proof alone that can be grasped universally and are sufficient from a moral standpoint. (B xxxiii)
References Adams, Robert Merrihew. 1997. ‘Things in Themselves’. Philosophy and Phenomenological Research 57 (4): 801–25. Allison, Henry. 1990. Kant’s Theory of Freedom. Cambridge: Cambridge University Press. Allison, Henry. 2004. Kant’s Transcendental Idealism: An Interpretation and Defense. Rev. ed. New Haven: Yale University Press. Allison, Henry. 2007. ‘Comments on Guyer’. Inquiry 50 (5): 480–8. Allison, Henry. 2011. Kant’s Groundwork for the Metaphysics of Morals: A Commentary. Oxford: Oxford University Press. Ameriks, Karl. 1982. Kant’s Theory of Mind. Oxford: Oxford University Press. Ameriks, Karl. 2000. Kant and the Fate of Autonomy. Cambridge: Cambridge University Press. Ameriks, Karl. 2003. Interpreting Kant’s Critiques. Oxford: Oxford University Press. Ameriks, Karl. 2004. Kant and the Historical Turn: Philosophy as Critical Interpretation. Oxford: Oxford University Press. Ameriks, Karl. 2012. Kant’s Elliptical Path. Oxford: Oxford University. Aristotle. 1984. Nicomachean Ethics. In Complete Works of Aristotle, Volume 1 The Revised Oxford Translation, edited by Jonathan Barnes. Princeton, NJ: Princeton University Press. Austin, John L. 1961. ‘Unfair to Facts’. In Philosophical Papers, edited by James O. Urmson and Geoffrey Warnock, 102–22. Oxford: Oxford University Press. Bacin, Stefano. 2010. ‘The Meaning of the Critique of Practical Reason for Moral Beings: The Doctrine of Method of Pure Practical Reason’. In Kant’s Critique of Practical Reason: A Critical Guide, edited by Andrews Reath and Jens Timmermann, 197–215. Cambridge: Cambridge University Press. Bacon, Francis. 1605 [1857]. The Advancement of Learning. In The Philosophical Works of Francis Bacon, edited by James Spedding, Robert L. Ellis, and Douglas D. Heath. London: Houghton Mifflin. Bagnoli, Carla. 2012. ‘Emotions and the Categorical Authority of Moral Reason’. In Morality and the Emotions, edited by Carla Bagnoli, 62–81. Oxford: Oxford University Press. Baum, Manfred. 2014. ‘Sittengesetz und Freiheit. Kant 1785 und 1788’. In Kants Rechtfertigung des Sittengesetzes in Grundlegung III: Deduktion Oder Faktum?, edited by Heiko Puls, 209–26. Berlin: Walter de Gruyter. Beck, Lewis White. 1960. A Commentary on Kant’s Critique of Practical Reason. Chicago: University of Chicago Press. Bennett, Jonathan. 1966. Kant’s Analytic. Cambridge: Cambridge University Press. Benton, Robert J. 1978. ‘The Transcendental Argument in Kant’s Groundwork of the Metaphysic of Morals’. The Journal of Value Inquiry 12 (3): 225–37. Berg, Anastasia N. A. 2020. ‘Kant on Moral Self-Opacity’. European Journal of Philosophy (Early View): https://doi.org/10.1111/ejop.12520. Berger, Larissa. 2015. ‘Der “Zirkel” im dritten Abschnitt der Grundlegung – Eine neue Interpretation und ein Literaturbericht’. In Kants Begründung von Freiheit und Moral in ‘Grundlegung III’: neue Interpretationen, edited by Dieter Schönecker, 9–82. Münster: Mentis.
162 References Bittner, Rüdiger. 1989. ‘Das Unternehmen einer Grundlegung zur Metaphysik der Sitten. Zur Vorrede der Grundlegung’. In Grundlegung zur Metaphysik der Sitten. Ein kooperativer Komentar, edited by Otfried Höffe, 13–30. Frankfurt: Vittorio Klostermann. Bojanowski, Jochen. 2017. ‘Kant on the Justification of Moral Principles’. Kant-Studien 108 (1): 55–88. Boyle, Robert. 1661/1772. ‘A Proëmial Essay, wherein, with some Considerations touching Experimental Essays in General’. In The Works of the Honourable Robert Boyle. In Six Volumes. To Which Is Prefixed the Life of the Author, edited by Thomas Birch. London: J. and F. Rivington. Brandt, Reinhardt. 2007. Die Bestimmung des Menschen bei Kant. Hamburg: F. Meiner. Brastberger, Gebhard Ulrich. 1792. Untersuchungen über Kants Kritik der practischen Vernunft. Tübingen: Cottaischen Buchhandlung. Breazeale, Daniel. 2013. Thinking Through the Wissenschaftslehre: Themes from Fichte’s Early Philosophy. Oxford: Oxford University Press. Budick, Sanford. 2010. Kant and Milton. Cambridge, MA: Harvard University Press. Butler, Joseph. 1736/1896. The Analogy of Religion, Natural and Revealed, in The Works of Joseph Butler, edited by William E. Gladstone. Oxford: Oxford University Press. Calvino, Italo. 1982. ‘Why Read the Classics?’ In The Uses of Literature, translated by Patrick Creagh, 125–34. New York: Harvest Books. Callanan, John. 2019. ‘Kant on Misology and the Natural Dialectic’. Philosophers' Imprint 19 (47): 1–22. Chignell, Andrew. 2007. ‘Belief in Kant’. The Philosophical Review 116 (3): 323–60. Choi, Yoon. 2019. ‘Spontaneity and Self-Consciousness in the Groundwork and the B-Critique’. Canadian Journal of Philosophy 49 (7): 936–55. Clewis, Robert. 2009. The Kantian Sublime and the Revelation of Freedom. Cambridge: Cambridge University Press. Cohen, Alix. 2020. ‘A Kantian Account of Emotions as Feelings’. Mind 129 (514): 429–60. Conant, James. 2004. ‘Varieties of Skepticism’. In Wittgenstein and Skepticism, edited by Denis McManus, 1–73. London: Routledge. Creuzer, Leonard. 1793. Skeptische Betrachtung über die Freiheit des Willens mit Hinsicht auf die neuesten Theorien über dieselben. Gießen: Heyer. Darwall, Stephen. 2006. The Second-Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Deligiorgi, Katerina. 2017. ‘The Philosopher as Legislator: Kant on History’. In The Palgrave Kant Handbook, edited by Matthew Altman, 683–705. London: Palgrave Macmillan. Denis, Laura. 2001. Moral Self-Regard: Duties to Oneself in Kant’s Moral Theory. London: Routledge. DeWitt, Janelle. 2014. ‘Respect for the Moral Law: The Emotional Side of Reason’. Philosophy 89 (1): 31–62. Dyck, Corey W. 2014. Kant and Rational Psychology. Oxford: Oxford University Press. Engstrom, Stephen. 1994. ‘The Transcendental Deduction and Skepticism’. Journal of the History of Philosophy 32 (3): 359–80. Engstrom, Stephen. 2002. ‘Introduction’. In Critique of Practical Reason, edited by Stephen Engstrom, xv–liv. Indianapolis: Hackett. Engstrom, Stephen. 2009. The Form of Practical Knowledge: An Essay on the Categorical Imperative. Cambridge, MA: Harvard University Press. Engstrom, Stephen. 2010. ‘The Triebfeder of Pure Practical Reason’. In Kant’s Critique of Practical Reason: A Critical Guide, edited by Andrews Reath and Jens Timmermann, 90–118. Cambridge: Cambridge University Press. ‘Fact, n’. OED Online. Oxford University Press.
References 163 Ferrarin, Alfredo. 2015. The Powers of Pure Reason: Kant and the Idea of Cosmic Philosophy. Chicago: University of Chicago Press. Forster, Michael. 2008. Kant and Skepticism. Princeton, NJ: Princeton University Press. Franks, Paul W. 2000. ‘Transcendental Arguments, Reason, and Skepticism: Contemporary Debates and the Origins of Post-Kantian Idealism’. In Transcendental Arguments: Problems and Prospects, edited by Robert Stern, 111–45. Oxford: Oxford University Press. Franks, Paul W. 2005. All or Nothing: Systematicity, Transcendental Arguments, and Skepticism in German Idealism. Cambridge, MA: Harvard University Press. Franks, Paul W. 2008. ‘Ancient Skepticism, Modern Naturalism, and Nihilism in Hegel’s Early Jena Writings’. In The Cambridge Companion to Hegel and Nineteenth-Century Philosophy, edited by Frederick Beiser, 52–73. Cambridge: Cambridge University Press. Frierson, Patrick. 2003. Freedom and Anthropology in Kant’s Moral Philosophy. Cambridge: Cambridge University Press. Frierson, Patrick. 2005. ‘Kant’s Empirical Account of Action’. Philosophers’ Imprint 5 (7): 1–34. Frierson, Patrick. 2014. Kant’s Empirical Psychology. Cambridge: Cambridge University Press. Gardner, Sebastian. 1999. Kant and the ‘Critique of Pure Reason’. London: Routledge. Gava, Gabriele. 2015. ‘Kant’s Synthetic and Analytic Method in the Critique of Pure Reason and the Distinction between Philosophical and Mathematical Syntheses’. European Journal of Philosophy 23 (3): 728–49. Geiger, Ido. 2011. ‘Rational Feelings and Moral Agency’. Kantian Review 16 (2): 283–308. Gewirth, Alan. 1978. Reason and Morality. Chicago: University of Chicago Press. Gill, Michael. 2009. ‘Moral Phenomenology in Hutcheson and Hume’. Journal of the History of Philosophy 47 (4): 569–94. Giovanni, George di. 2005. Freedom and Religion in Kant and His Immediate Successors: The Vocation of Humankind, 1774–1800. Cambridge: Cambridge University Press. Giovanni, George di. 2011. ‘The Year 1786 and Die Bestimmung des Menschen, or Popularphilosophie in Crisis’. In Studies in German Idealism, edited by Renier Munk, 217–34. New York: Springer. Green, Thomas Hill. 1886. Lectures on the Philosophy of Kant. In The Works of Thomas Hill Green, Vol. II. Philosophical Works, edited by R. L. Nettleship. London: Longmans, Green, and Co. Grenberg, Jeanine. 2001. ‘Feeling, Desire and Interest in Kant’s Theory of Action’. KantStudien 92 (2): 153–79. Grenberg, Jeanine. 2005. Kant and the Ethics of Humanity: A Story of Dependence, Corruption, and Virtue. Cambridge: Cambridge University Press. Grenberg, Jeanine. 2013. Kant’s Defense of Common Moral Experience: A Phenomenological Account. Cambridge: Cambridge University Press, 2013. Guyer, Paul. 2000. Kant on Freedom, Law, and Happiness. Cambridge: Cambridge University Press. Guyer, Paul. 2007. ‘Naturalistic and Transcendental Moments in Kant’s Moral Philosophy’. Inquiry 50 (5): 444–64. Guyer, Paul. 2008. Knowledge, Reason, and Taste: Kant’s Response to Hume. Princeton, NJ: Princeton University Press. Guyer, Paul. 2009. ‘Problems with Freedom: Kant’s Argument in Groundwork III and its Subsequent Emendations’. In Kant’s Groundwork of the Metaphysics of Morals: A Critical Guide, edited by Jens Timmermann, 176–202. Cambridge: Cambridge University Press. Guyer, Paul. 2010. ‘Moral Feeling in the Metaphysics of Morals’. In Kant’s Metaphysics of Morals: A Critical Guide, edited by Lara Denis, 130–51. Cambridge: Cambridge University Press.
164 References Guyer, Paul. 2012. ‘Examples of Moral Possibility’. In Kant and Education, edited by Klas Roth and Chris W. Surprenant, 126–38. New York: Routledge. Guyer, Paul. 2017. ‘The Proof-Structure of the Groundwork and the Role of Section III’. In The Virtues of Freedom: Selected Essays on Kant, 127–45. Oxford: Oxford University Press. Hägerstrom, Axel. 1902. Kants Ethik im Verhältnis zu seinen erkenntnistheoretischen Grundgedanken. Upsala: Almqvist & Wiksells. Hahmann, Andree. 2012. ‘Ist ‘Freiheit die Wahrheit der Notwendigkeit’? Das Ding an sich als Grund der Erscheinung bei Kant’. In Sind wir Bürger zweier Welten? Freiheit und moralische Verantwortung im transzendentalen Idealismus, edited by Mario Brandhorst, Andree Hahmann, and Bernd Ludwig, 135–54. Hamburg: Meiner. Hawthorne, Nathaniel. 1850/2003. The Scarlet Letter: A Romance. New York: Penguin Books. Heide, Dai. 2020. ‘A Mereological Argument for the Non-Spatiotemporality of Things in Themselves’. European Journal of Philosophy 28 (1): 48–76. Heidegger, Martin. 1927/1988. The Basic Problems of Phenomenology, translated by Albert Hofstadter. Bloomington: Indiana University Press, 1988. Henrich, Dieter. 1960. ‘Der Begriff der sittlichen Einsicht und Kants Lehre vom Faktum der Vernunft’. In Die Gegenwart der Griechen im neueren Denken. Festschrift für HansGeorg Gadamer zum 60. Geburtstag, edited by Dieter Henrich, Walter Schulz, and KarlHeinz Volkmann-Schluck, 77–115. Tübingen: Mohr Siebeck. Henrich, Dieter. 1975. ‘Die Deduktion des Sittengesetzes: Über die Gründe der Dunkelheit des letzten Abschnittes von Kants “Grundlegung zur Metaphysik der Sitten” ’. In Denken im Schatten des Nihilismus. Festschrift für Wilhelm Weischedel zum 70. Geburtstag, edited by Alexander Schwan, 55–112. Darmstadt: Wissenschaftliche Buchgesellschaft. Henrich, Dieter. 1989. ‘Kant’s Notion of a Deduction and the Methodological Background of the First Critique’. In Kant’s Transcendental Deductions: The Three Critiques and the Opus Postumum, edited by Eckart Förster, 29–46. Stanford: Stanford University Press. Henrich, Dieter. 1994. ‘The Concept of Moral Insight and Kant’s Doctrine of the Fact of Reason’. In The Unity of Reason: Essays on Kant’s Philosophy, edited by Richard L. Velkley, 55–88. Cambridge, MA: Harvard University Press. Herman, Barbara. 1989. ‘Justification and Objectivity: Comments on Rawls and Allison’. In Kant’s Transcendental Deductions: The Three Critiques and the Opus Postumum, edited by Eckart Förster, 131–44. Stanford, CA: Stanford University Press. Herman, Barbara. 2005. ‘Transforming Incentives: Feelings and the Making of the Kantian Moral Agent’. In Philosophical Aspects of the Emotions, edited by Asa Carlson, 17–44. Stockholm: Thales. Herman, Barbara. 2007. Moral Literacy. Cambridge, MA: Harvard University Press. Herrera, Larry. 2000. ‘Kant on the Moral Triebfeder’. Kant-Studien 91 (4): 395–410. Hill, Thomas E. 1998. ‘Kant’s Argument for the Rationality of Moral Conduct’. In Kant’s Groundwork of the Metaphysics of Morals: Critical Essays, edited by Paul Guyer, 249–72. New York: Roman & Littlefield. Hill, Thomas E. 2002. ‘Four Conceptions of Conscience’. In Human Welfare and Moral Worth: Kantian Perspectives, 277–309. Oxford: Oxford University Press. Hill, Thomas E. 2012. Virtues, Rules, and Justice: Kantian Aspirations. Oxford: Oxford University Press. Hobbes, Thomas. 1651/1996. Leviathan, edited by Richard Tuck. Cambridge: Cambridge University Press. Hogan, Desmond. 2009. ‘Noumenal Affection’. Philosophical Review 118 (4): 501–32. Horgan, Terry, and Timmons, Mark. 2005. ‘Moral Phenomenology and Moral Theory’. Philosophical Issues 15 (1): 56–77.
References 165 Horn, Christoph. 2015. ‘Das Bewusstsein, unter dem moralischen Gesetz zu stehen – Kants Freiheitsargument in GMS III’. In Kants Begründung von Freiheit und Moral in ‘Grundlegung III’: Neue Interpretationen, edited by Dieter Schönecker, 137–56. Münster: Mentis. Howard, Jason. 2004. ‘Kant and Moral Imputation: Conscience and the Riddle of the Given’. The American Catholic Philosophical Quarterly 78 (4): 609–27. Hume, David. 1739/1978. A Treatise of Human Nature. 2nd ed. Oxford: Oxford University Press. Indregard, Jonas Jervell. 2018. ‘A Gradual Reformation: Empirical Character and Causal Powers in Kant’. Canadian Journal of Philosophy 48 (5): 662–83. Insole, Christopher J. 2013. Kant and the Creation of Freedom. Oxford: Oxford University Press. Jacobs, Brian and Kain, Patrick (eds). 2003. Essays on Kant’s Anthropology. Cambridge: Cambridge University Press. Johnston, David. 2004. ‘The Natural History of Fact’. Australasian Journal of Philosophy 82 (2): 275–91. Kain, Patrick. 2006. ‘Realism and Anti-Realism in Kant’s Second Critique’. Philosophy Compass 1 (5): 449–65. Kain, Patrick. 2010. ‘Practical Cognition, Intuition, and the Fact of Reason’. In Kant’s Moral Metaphysics: God, Freedom, and Immortality, edited by Benjamin Lipscomb and James Krueger, 211–32. Berlin: Walter de Gruyter. Keller, Pierre. 2010. ‘Two Conceptions of Compatibilism in the Critical Elucidation’. In Kant’s Critique of Practical Reason: A Critical Guide, edited by Jens Timmermann and Andrews Reath, 119–44. Cambridge: Cambridge University Press. Kelsey, Matthew. 2014. ‘Kant’s Diagnosis of the Unity of Skepticism’. Philosophers’ Imprint 14 (14): 1–21. Kitcher, Patricia. 2011. Kant’s Thinker. New York: Oxford University Press. Kleingeld, Pauline. 2010. ‘Moral Consciousness and the “Fact of Reason.” ’ In Kant’s Critique of Practical Reason: A Critical Guide, edited by Jens Timmermann and Andrews Reath, 55–72. Cambridge: Cambridge University Press. Klemme, Heiner. 2010. ‘The Origin and Aim of Kant’s Critique of Practical Reason’. In Kant’s Critique of Practical Reason: A Critical Guide, edited by Andrews Reath and Jens Timmermann, 11–30. Cambridge: Cambridge University Press. Kohl, Markus. 2020. ‘Kant on Cognizing Oneself as a Spontaneous Cognizer’. Canadian Journal of Philosophy 50 (3): 395–412. Korsgaard, Christine. 1989. ‘Morality as Freedom’. In Kant’s Practical Philosophy Reconsidered, edited by Yirmiyahu Yovel, 23–48. Dordrecht: Kluwer Academic. Korsgaard, Christine. 1996a. Sources of Normativity. Cambridge: Cambridge University Press. Korsgaard, Christine. 1996b. Creating the Kingdom of Ends. Cambridge: Cambridge University Press. Korsgaard, Christine. 2008. Self-Constitution: Agency, Identity, and Integrity. Oxford: Oxford University Press. Korsgaard, Christine. 2009. The Constitution of Agency: Essays on Practical Reason and Moral Psychology. Oxford: Oxford University Press. Lavater, Johann Kaspar. 1772. Geheimes Tagebuch von einem Beobachter seiner selbst. Leipzig: Weidmanns and Reich. Locke, John. 1689/1975. An Essay Concerning Human Understanding, edited by P.H. Nidditch. Oxford: Oxford University Press. Louden, Robert B. 2000. Kant’s Impure Ethics: From Rational Beings to Human Beings Oxford: Oxford University Press.
166 References Ludwig, Bernd. 2010. ‘Die “consequente Denkungsart der speculativen Kritik.” Kants radikale Umgestaltung seiner Freiheitslehre im Jahre 1786 und die Folgen für die Kritische Philosophie als Ganze’. Deutsche Zeitschrift für Philosophie 58 (4): 595–628. Ludwig, Bernd. 2012. ‘Was weiss ich vom Ich? Kants Lehre vom Faktum der reinen praktischen Vernunft, seine Neufassung der Paralogismen und die verborgenen Fortschritte der Kritischen Metaphysik im Jahre 1786’. In Sind wir Bürger zweier Welten? Freiheit und moralische Verantwortung im transzendentalen Idealismus, edited by Mario Brandhorst, Andree Hahmann, and Bernd Ludwig, 155–94. Hamburg: Meiner. Ludwig, Bernd. 2015. ‘Die Kritik der reinen Vernunft hat die Wirklichkeit der Freiheit nicht bewiesen, ja nicht einmal deren Möglichkeit’: Über die folgenreiche Fehlinterpretation eines Absatzes in der Kritik der reinen Vernunft’. Kant-Studien 106: 398–417. Ludwig, Bernd. 2018. ‘Über drei Deduktionen in Kants Moralphilosophie – und über eine vierte, die man dort vergeblich sucht. Zur Rehabilitierung von Grundlegung III’. KantStudien 109: 47–71. Łuków, Paweł. 1993. ‘The Fact of Reason: Kant’s Passage to Ordinary Moral Knowledge’. Kant-Studien 84 (2): 204–21. Maimon, Salomon. 1794. ‘Versuch einer neuen Darstellung des Moralprincips und Dedukzion seiner Realität’. Berlinische Monatschrift 24: 402–53. Mandelbaum, Maurice. 1955. The Phenomenology of Moral Experience. Glencoe, IL: The Free Press. Marshall, Colin. 2010. ‘Kant’s Metaphysics of the Self ’. Philosophers’ Imprint 10 (8): 1–21. McCarthy, Michael H. 1982. ‘Kant’s Rejection of the Argument of Groundwork III’. Kant-Studien 73 (1): 169–90. McCarty, Richard. 1993. ‘Kantian Moral Motivation and the Feeling of Respect’. Journal of the History of Philosophy 31 (3): 421–35. McCarty, Richard. 2009. Kant’s Theory of Action. Oxford: Oxford University Press. McLear, Colin. 2020. ‘On the Transcendental Freedom of the Intellect’. Ergo: An Open Access Journal of Philosophy 7 (2): 1–70. Mellin, Georg Samuel Albert. 1800. Encyclopädische Wörterbuch der Kritischen Philosophie. Leipzig: Frommann. Mendelssohn, Moses. 1783. Jerusalem oder über religiöse Macht und Judentum. Berlin: Friedrich Maurer. Menn, Stephen. 2002. ‘Plato and the Method of Analysis’. Phronesis 47 (3): 193–223. Michaelis, Christian Friedrich. 1796–97. Ueber die sittliche natur und bestimmung des menschen: ein versuch zur erläuterung über I. Kant’s Kritik der praktischen vernunf. Leipzig: J.G. Beigang. Moran, Kate. 2014. ‘Delusions of Virtue: Kant on Self-Conceit’. Kantian Review 19 (3): 419–47. Morrisson, Iain. 2008. Kant and the Role of Pleasure in Moral Action. Columbus: Ohio University Press. Moyar, Dean. 2008. ‘Unstable Autonomy: Conscience and Judgment in Kant’s Moral Philosophy’. Journal of Moral Philosophy 5 (3): 327–60. Munzel, Felicitas. 2012. Kant’s Conception of Pedagogy: Toward Education for Freedom. Evanston: Northwestern University Press. Nagel, Thomas. 1970. The Possibility of Altruism. Princeton, NJ: Princeton University Press. Nahm, Milton. 1957. ‘Sublimity and the Moral Law in Kant’s Ethics’. Kant-Studien 48 (1): 502–24. Neiman, Susan. 1994. The Unity of Reason: Rereading Kant. New York: Oxford University Press. Newton, Alexandra. 2017. ‘Kant on Animal and Human Pleasure’. Canadian Journal of Philosophy 47 (4): 518–40.
References 167 O’Neill, Onora. 1989. Constructions of Reason: Explorations of Kant’s Practical Philosophy. Cambridge: Cambridge University Press. O’Neill, Onora. 1996. ‘Kant’s Virtues’. In How Should One Live? Essays on the Virtues, edited by Roger Crisp, 77–98. Oxford: Oxford University Press. Papish, Laura. 2018. Kant on Evil, Self-Deception, and Moral Reform. Oxford: Oxford University Press. Paton, Herbert. 1947. The Categorical Imperative: A Study in Kant’s Moral Philosophy. London: Hutchinson University Library. Pereboom, Derk. 2006. ‘Kant on Transcendental Freedom’. Philosophy and Phenomenological Research 73 (3): 537–67. Philonenko, Alexander. 2008. ‘Introduction’. In Fondements de la métaphysique des moeurs, 7–59. Paris: Vrin. Pistorius, Herman Andreas. 1786/1975. ‘Rezension der Grundlegung zur Metaphysik der Sitten’. In Materialien zu Kants Kritik der praktischen Vernunft, edited by Rüdiger Bittner and Konrad Cramer. Frankfurt: Suhrkamp. Pistorius, Herman Andreas. 1786/2007. ‘Rezension von J. Schulzes Erläuterungen zur Kritik der reinenVernunft’. In Kants vergessener Rezensent. Die Kritik der theoretischen und praktischen Philosophie Kants in fünf frühen Rezensionen von Hermann Andreas Pistorius, edited by Bernward Gesang. Hamburg: Meiner. Plato. 1997. Plato: Complete Work, edited by John Cooper. London: Hackett. Printy, Michael. 2013. ‘The Determination of Man: Johann Joachim Spalding and the Protestant Enlightenment’. Journal of the History of Ideas 74 (2): 189–212. Proops, Ian. 2003. ‘Kant’s Legal Metaphor and the Nature of a Deduction’. Journal of the History of Philosophy 41 (2): 209–29. Puls, Heiko. 2011. ‘Freiheit als Unabhängigkeit von bloß subjektiv bestimmenden Ursachen – Kants Auflösung des Zirkelverdachts im dritten Abschnitt der Grundlegung zur Metaphysik der Sitten’. In Zeitschrift für Philosophische Forschung 65 (4): 534–62. Puls, Heiko. 2014. ‘Quo errat demonstrator – warum es in der Grundlegung eine FaktumThese gibt. Drei Argumente gegen Dieter Schöneckers Interpretation’. In Kants Rechtfertigung des Sittengesetzes in Grundlegung III: Deduktion Oder Faktum?, edited by Heiko Puls, 15–32. Berlin: Walter de Gruyter. Puls, Heiko. 2016. Sittliches Bewusstsein und Kategorischer Imperativ in Kants Grundlegung: Ein Kommentar zum dritten Abschnitt. Berlin: De Gruyter. Railton, Peter. 1997. ‘On the Hypothetical and the Non-Hypothetical in Reasoning about Belief and Action’. In Ethics and Practical Reason, edited by Garrett Cullity and Berys Gaut, 53–79. Oxford: Oxford University Press. Rauscher, Frederick. 2002. ‘Kant’s Moral Anti-Realism’. Journal of the History of Philosophy 40 (4): 477–99. Rauscher, Frederick. 2009. ‘Freedom and Reason in Groundwork III’. In Kant’s Groundwork for the Metaphysics of Morals: A Guide, edited by Jens Timmermann. Cambridge: Cambridge University Press, 203–23. Rauscher, Frederick. 2015. Naturalism and Realism in Kant’s Ethics. Cambridge: Cambridge University Press. Rawls, John. 1989. ‘Themes in Kant’s Moral Philosophy’. In Kant’s Transcendental Deductions: The Three Critiques and the Opus Postumum, edited by Eckart Förster, 81–113. Stanford: Stanford University Press. Rawls, John. 2000. Lectures on the History of Moral Philosophy. Cambridge, MA: Harvard University Press, 2000. Reath, Andrews. 2006. Agency and Autonomy in Kant’s Moral Theory. Oxford: Oxford University Press.
168 References Reath, Andrews. 2012. ‘Kant’s Moral Philosophy’. In The Oxford Handbook on the History of Ethics, edited by Roger Crisp, 443–64. Oxford: Oxford University Press. Reich, Klaus. 1939. ‘Kant and Greek Ethics’. Mind 48 (191): 446–63. Reinhold, Karl. 1789. Versuch einer neuen Theorie des vorstellungsvermögens. Prague: Widtmann and Mauke. Reinhold, Karl. 1786–97. Briefe über die Kantische Philosophie. Vol. 1. Leipzig: Göschen. Ress, Johann Heinrich. 1777. Die Auferstehungsgeschichte Jesu Christi gegen einige im vierten Beytrage zur Geschichte und Litteratur aus den Schätzen der Herzoglichen Bibliothek zu Wolfenbüttel gemachte neuere Einwendungen vertheidiget. Braunschweig: Verlag der Fürstl. Rohden, Valerio. 2012. ‘An Experiment with Practical Reason’. In Kant in Brazil, edited by Frederick Rauscher and Daniel O. Perez, 98–109. Rochester: University of Rochester Press. Rosati, Connie. 2003. ‘Agency and the Open Question Argument’. Ethics 113 (3): 490–527. Ross, William David. 1954. Kant’s Ethical Theory: A Commentary on the Grundlegung Zur Metaphysik Der Sitten. Oxford: Oxford University Press. Rousseau, Jean-Jacques. 1762/1979. Emile, or On Education, translated by Harold Bloom. New York: Basic Books. Russell, Francey. 2020. ‘Kantian Self-Conceit and the Two Guises of Authority’. Canadian Journal of Philosophy 50 (2): 268–83. Russell, Paul. 2004. ‘Butler’s “Future State” and Hume’s “Guide to Life” ’. Journal of the History of Philosophy 42 (4): 425–48. Sade, Marquis de. 1791/1999. The Misfortunes of Virtue and Other Early Tales, translated by David Coward. Oxford: Oxford University Press, 1999. Saunders, Joe. 2014. ‘Kant, Rational Psychology and Practical Reason’. Kant Yearbook 6 (1): 115–35. Schneewind, Jerome B. 1998. The Invention of Autonomy: A History of Modern Moral Philosophy. Cambridge: Cambridge University Press. Schopenhauer, Arthur. 1841/1988. Über die Grundlage der Moral. Edited by Peter Welsen. Hamburg: Felix Meiner. Schapiro, Tamar. 2009. ‘The Nature of Inclination’. Ethics 119 (2): 229–56. Schönecker, Dieter. 1997. ‘Gemeine sittliche und philosophische Vernunfterkenntnis: Zum ersten Übergang in Kants Grundlegung’. Kant-Studien 88 (3): 311–33. Schönecker, Dieter. 1999. Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs. Freiburg/München: Karl Alber-Verlag. Schönecker, Dieter. 2006. ‘How is a Categorical Imperative Possible?’ In Groundwork for the Metaphysics of Morals, edited by Christoph Horn and Dieter Schönecker, 301–23. Berlin: De Gruyter. Schönecker, Dieter. 2013. ‘Kant’s Moral Intuitionism: The Fact of Reason and Moral Predispositions’. Kant Studies Online 1 (1): 1–38. Schönecker, Dieter. 2014. ‘Warum es in der Grundlegung keine Faktum-These gibt. Drei Argumente’. In Kants Rechtfertigung des Sittengesetzes in Grundlegung III: Deduktion Oder Faktum?, edited by Heiko Puls, 1–15. Berlin: Walter de Gruyter. Schumski, Irina. 2018. ‘Kant on Self-Opacity and Self-Cultivation’. In Ethics and SelfCultivation: Historical and Contemporary Perspectives, edited by Matthew Dennis and Sander Werkhoven, 47–67. London: Routledge. Serjeantson, Richard. 1999. ‘Testimony and Proof in Early-Modern England’. Studies In History and Philosophy of Science 30 (2): 195–236. Shapin, Steven. 1984. ‘Pump and Circumstance: Robert Boyle’s Literary Technology’. Social Studies of Science 14 (4): 481–520.
References 169 Shapin, Steven, and Schaffer, Simon. 1985. Leviathan and the Air-Pump: Hobbes, Boyle, and Experimental Life. Princeton, NJ: Princeton University Press. Shapiro, Barbara. 1994. ‘The Concept ‘Fact’: Legal Origins and Cultural Diffusion’. Albion: A Quarterly Journal Concerned with British Studies 26 (2): 227–52. Shapiro, Barbara. 2000. A Culture of Fact: England, 1550–1720. Ithaca: Cornell University Press. Shapiro, Barbara. 2002. ‘Testimony in Seventeenth-Century English Natural Philosophy: Legal Origins and Early Development’. Studies in History and Philosophy of Science 33 (2): 243–63. Spalding, Johann Joachim. 1748/1997. Die Bestimmung des Menschen, edited W.E. Müller. Waltrop: Hartmut Spenner. Stang, Nicholas. 2016. Kant’s Modal Metaphysics. Oxford: Oxford University Press. Stark, Cynthia. 1998. ‘The Rationality of Valuing Oneself: A Critique of Kant on SelfRespect’. Journal of the History of Philosophy 35 (1): 65–82. Starck, Johann August. 1779. Geschichte der Christlichen Kirche des Ersten Jahrhunderts. Berlin and Leipzig: Decker. Stattler, Benedict. 1788. Anti-Kant, Bd. 1–2. München: Lentner. Stern, Robert. 1999. ‘On Kant’s Response to Hume: The Second Analogy as Transcendental Argument’. In Transcendental Arguments: Problems and Prospects, edited by Robert Stern, 47–66. Oxford: Oxford University Press. Stern, Robert. 2011. Understanding Moral Obligation: Kant, Hegel, Kierkegaard. Cambridge: Cambridge University Press. Stern, Robert. 2015. Kantian Ethics: Value, Agency, and Obligation. Oxford: Oxford University Press. Stratton-Lake, Philip. 2000. Kant, Duty, and Moral Worth. London: Routledge Sticker, Martin. 2014. ‘How Can Common Rational Capacities Confirm the Correctness of the Deduction in Groundwork III—and Why Does It Matter?’ Hegel Bulletin 35 (2): 228–51. Sticker, Martin. 2015. ‘Educating the Common Agent: Kant on the Varieties of Moral Education’. Archiv für Geschichte der Philosophie 97 (3): 358–87. Strawson, Peter. 1966. The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason. Methuen & Co. Ltd. Sullivan, Roger. 1989. Immanuel Kant’s Moral Theory. Cambridge: Cambridge University Press. Sussman, David. 2008. ‘From Deduction to Deed: Kant’s Grounding of the Moral Law’. Kantian Review 13 (1): 52–81. Tenenbaum, Sergio. 2012. ‘The Idea of Freedom and Moral Cognition in Groundwork III’. Philosophy and Phenomenological Research 84 (3): 555–89. Thiel, Udo. 1997. ‘Varieties of Inner Sense: Two Pre-Kantian Theories’. Archiv für Geschichte der Philosophie 79 (1): 58–79. Timmermann, Jens. 2007. Kant’s Groundwork of the Metaphysics of Morals: A Commentary. Cambridge: Cambridge University Press. Timmermann, Jens. 2010. ‘Reversal or Retreat? Kant’s Deductions of Freedom and Morality’. In Kant’s Critique of Practical Reason: A Critical Guide, edited by Jens Timmermann and Andrews Reath, 73–89. Cambridge: Cambridge University Press. Tippmann, Caroline. 2011. Die Bestimmung des Menschen bei Johann Joachim Spalding. Leipzig: Evangelische Verlagsanstalt. Tittel, Gottlob August. 1786. Ueber Herrn Kant’s Moralreform. Frankfurt: Pfähler. Tizzard, Jessica. 2018. ‘Kant on Space, Time, and Respect for the Moral Law as Analogous Formal Elements of Sensibility’. European Journal of Philosophy 26 (1): 630–46.
170 References Vanzo, Alberto. 2012. ‘Kant on Experiment’. In Rationis Defensor: Essays in Honour of Colin Cheyne, edited by James Maclaurin, 75–96. Dordrecht: Springer. Velkley, Richard L. 2002. Being After Rousseau: Philosophy and Culture in Question. Chicago: University of Chicago Press. Velleman, David. 1989. Practical Reflection. Princeton, NJ: Princeton University Press. Velleman, David. Velleman, David. 2000. The Possibility of Practical Reason. Oxford: Oxford University Press. Velleman, David. 2006. Self to Self: Selected Essays. Cambridge: Cambridge University Press. Velleman, David. 2009. How We Get Along. Cambridge: Cambridge University Press. Walker, Ralph. 2000. ‘Achtung in the Grundlegung’. In Grundlegung zur Metaphysik der Sitten: Ein kooperativer Kommentar, edited Edited by Dieter Höffe, 97–116. Frankfurt and Main: Vittorio Klosterman. Ware, Owen. 2009. ‘The Duty of Self-Knowledge’. Philosophy and Phenomenological Research 79 (3): 671–98. Ware, Owen. 2010. Kant, Skepticism, and Moral Sensibility. Dissertation Thesis, University of Toronto. Ware, Owen. 2014a. ‘Kant on Moral Sensibility and Moral Motivation’. Journal of the History of Philosophy 52 (4): 727–46. Ware, Owen. 2014b. ‘Rethinking Kant’s Fact of Reason’. Philosophers’ Imprint 14 (32): 1–21. Ware, Owen. 2015. ‘Accessing the Moral Law through Feeling’. Kantian Review 20 (2): 301–11. Ware, Owen. 2016. ‘Skepticism in Kant’s Groundwork’. European Journal of Philosophy 24 (2): 375–96. Ware, Owen. 2017. ‘Kant’s Deductions of Morality and Freedom’. Canadian Journal of Philosophy 47 (1): 116–47. Ware, Owen. 2020. Fichte’s Moral Philosophy. New York: Oxford University Press. Watkins, Eric. 2005. Kant and the Metaphysics of Causality. Cambridge: Cambridge University Press. Wilde, Oscar. 1897/2004. The Picture of Dorian Gray. New York: Random House. Williams, Bernard. 1985. Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press. Williams, Terence Charles. 1968. The Concept of the Categorical Imperative: A Study of the Place of the Categorical Imperative in Kant’s Ethical Theory. Oxford: Oxford University Press. Willaschek, Marcus. 1991. ‘Die Tat der Vernunft: Zur Bedeutung der Kantischen These vom “Factum der Vernunft” ’. In Akten des Siebenten Internationalen Kant-Kongresses, edited by G. Funke, 455–66. Bonn: Bouvier. Willaschek, Marcus. 1992. Praktische Vernunft: Handlungstheorie und Moralbegründung bei Kant. Weimar: Metzler. Wolff, Michael. 2009. ‘Warum das Faktum der Vernunft ein Faktum ist. Auflösung einiger Verständnisschwierigkeiten in Kants Grundlegung der Moral’. Deutsche Zeitschrift für Philosophie 57 (4): 511–49. Wolff, Robert Paul. 1963. Kant’s Theory of Mental Activity a Commentary on the Transcendental Analytic of the Critique of Pure Reason. Harvard, MA: Harvard University Press. Wood, Allen. 1999. Kant’s Ethical Thought. Cambridge: Cambridge University Press. Wood, Allen. 2007. Kantian Ethics. Cambridge: Cambridge University Press. Wood, Allen. 2011. ‘Preface and Introduction (3–16)’. In Immanuel Kant: Kritik der praktischen Vernunft, edited by O. Höffe. Berlin: Akademie Verlag.
References 171 Wright, Cripsin. 1991. ‘Scepticism and Dreaming: Imploding the Demon’. Mind 100 (1): 87–116. Wuerth, Julian. 2013. ‘Sense and Sensibility in Kant’s Practical Agent: Against the Intellectualism of Korsgaard and Sidgwick’. European Journal of Philosophy 21 (1): 1–36. Zammito, John H. 2002. Kant, Herder, and the Birth of Anthropology. Chicago: University of Chicago Press. Zande, J. V. D. 1995. ‘In the Image of Cicero: German Philosophy between Wolff and Kant’. Journal of the History of Ideas 56 (3): 419–42. Zinkin, Melissa. 2006. ‘Respect for the Law and the Use of Dynamical Terms in Kant’s Theory of Moral Motivation’. Archiv für Geschichte der Philosophie 88 (1): 31–53.
Index For the benefit of digital users, indexed terms that span two pages (e.g., 52–53) may, on occasion, appear on only one of those pages. Affectivism 119–23 Agency Rational agency or nature 30, 90–1, 115–18, 122, 133–4, 156n.1 Standpoint of agency 37–8 Allison, Henry 2n.6, 10n.24, 11n.28, 16–17, 17n.3, 26n.15, 34n.27, 38n. 35, 52n.28, 54n.32, 66, 80n.31, 115n.32, 123–4, 130 Ameriks, Karl v–vi, 3n.7, 5n.17, 6n.18, 10n.24, 44, 45n.4, 65n.54 Analytic method 20–1 Architectonic method 97–8 Aristotle 2n.4, 21 Assurance (see trust) 143–54 Autonomy (see also agency, freedom) Autonomy of the will 8, 68n.58, 77, 114, 116, Principle or Formula of autonomy 20, 30–2, 36–8, 65–8, 77–8, 82, 87–8, 91, 96–8, 131 Bacon, Francis 46–7, 47n.11, 53n.30 Bagnoli, Carla 115n.35 Beck, Lewis White 48, 75–6 Bennett, Jonathan 1n.2 Berger, Larissa 78n.27 Bindingness (see necessitation) Bittner, Rüdiger 11n.25 Bojanowski, Jochen 10n.24, 98n.61 Boyle, Robert 46–8 Brandt, Reinhardt 49n.19 British experimentalists 47, 50, 64–5 Budick, Sanford 62n.48 Butler, Joseph 48–50, 148n.21 Categorical Imperative (see also morality) 30–1, 35–9, 68–9, 76–86, 93n.53, 97, 106n.13 Categories 1–2, 32–4, 36–7, 38n.35, 54n.32, 55n.36, 83n.36 Causality (see also necessity, necessitation) 21–3, 63, 80–1, 89–92, 97n.59, 98n.60 Character 110, 114n.31, 134–7, 141–3 Choice (see also agency) 52–64, 105, 109–18, 128–9, 137–41, 137n.10
Circle 76–83, 88–9, 130n.73, 156 Clewis, Robert 115n.34 Cohen, Alix 107n.18 Common human reason 6–15, 18–23, 64–5, 80, 95, 158–60 Conant, James 17n.8 Confession 118–19 Conscience 142, 144n.18, 148–54 Constructivism 56n.38 Continuity reading 86–8 Conversion 5, 12–13 Creuzer, Leonard 74, 157–9 Critique Critique of pure practical reason 22, 26n.15, 35, 78, 96 Critique of the subject 78 Darwall, Stephen 10n.24, 112n.29 Dear self (see also self-conceit) 28–9, 135, 136n.5, 142, 153–4 Death 7–8, 62n.48, 98, 146–7 Deduction Apogogic 9n.21 Genetic 9–11, 32–3, 38–9, 42–6, 51, 84–5, 131–2 Legal 32, 64–5 Ostensive 8–9, 32 Deligiorgi, Katerina 6n.19 Denis, Lara 136n.8, 138n.12 Despair 14, 133–4, 144, 147, 154–5 DeWitt, Janelle 107n.18 Dignity 14, 30, 101–2, 115n.33, 116n.37, 117n.40, 139n.14, 140–1 Disclosure thesis 63–4, 157 Duties to oneself 137–41 Duty (see also necessitation) 18–41, 57–70, 85–8, 94–8, 106–23, 133–54, 157 Egoism (see also self-love, self-conceit) 27, 114 Engstrom, Stephen 27n.17, 113n.30, 115n.35, 120n.46 Experimental method 7–11, 45–7, 52, 53n.30, 57–60
174 Index Fact of Reason 11–12, 44, 71–6, 89n.46, 99, 123–31, 157–8 Facts 46–57, 62–5, 69–70 Factum (see fact of reason) Faculty of desire (see also practical reason) 52, 54, 54n.32, 57, 65, 67–8, 106–8, 111–12, 127 Feeling (see also sensibility) Capacity to feel 8, 11–12, 14, 105, 108, 108n.21, 109–10, 118, 122–3, 131 Faculty of feeling 95–6, 98–9, 130–1 Feeling of pleasure and displeasure 105–9, 114n.31, 122–3, 131 Ferrarin, Alfredo 6n.19 Formula of Humanity 30 Formula of Universal Law (see also categorical imperative) 25–6, 30 Forster, Michael 16–18 Foundationalism in ethics 76, 158–9 Franks, Paul W. 11n.28, 14n.35, 46n.5 Freedom (see also disclosure thesis) 18, 37n.33, 62–8, 71, 103n.5, 106n.14, 119, 121–2, 127–32, 156–9 Frierson, Patrick 110n.25, 120n.48 Gardner, Sebastian 33n.26 Garve, Christian 21, 35n.29, 68–9 Gava, Gabriele 11n.26 Gewirth, Alan 156n.1 Giovanni, George di. 49n.19 God 4–5, 49, 137, 143–4 Great reversal (see reversal reading) Green, Thomas Hill 76n.21 Grenberg, Jeanine 57n.39, 123–31, 136n.8 Guyer, Paul 16–18, 22n.12, 35n.29, 37n.33, 38n.35, 54n.33 Hägerstrom, Axel 76n.21 Hahmann, Andree 10n.24 Happiness (see also self-love) 7–14, 23–7, 34–42, 98, 110–17, 132, 134–5 Hawthorne, Nathaniel 135–6, 153–4 Heide, Dai 3n.10 Heidegger, Martin 12n.29, 123–31 Henrich, Dieter 32n.25, 44–5, 64–5, 75–6, 93–4 Herman, Barbara 2, 49n.20, 111n.26, 115n.33 Herrera, Larry 120n.46 Herz, Marcus 102 Heteronomy 67 Hill, Thomas E. 17–18, 41, 150n.23 Hobbes, Thomas 17, 112 Hogan, Desmond 3n.10 Hope 104n.8, 109, 112, 143–4 Hopelessness (see despair)
Horgan, Terry 118n.43 Horn, Christoph 91n.49 Howard, Jason 149n.22 Hume, David 16–17, 51, 102, 105–6 Humiliation 110, 113–15, 119, 121 Hypothetical Imperative 41–2 Imagination 9, 25, 27–8, 27n.19, 31, 33–4, 39, 41–2 Incentive 11–12, 27–9, 99, 103, 108–10, 113, 116–21, 128 Inclination 24–42, 58–9, 79–83, 86, 91–2, 94, 103–22, 127–8 Indregard, Jonas Jervell 3n.10 Inner judge (see conscience) Insole, Christopher J. 3n.10 Intellectualism 120–2 Jacobs, Brian 136n.8 Johnston, David 48 Justification 4–23, 33, 44–6, 65, 72–7, 81, 83–6, 123–31, 158–60 Kain, Patrick 55n.35, 56n.38 Keller, Pierre 58n.42 Kelsey, Matthew 5n.15 Kitcher, Patricia 98n.60 Kleingeld, Pauline 45n.2, 55n.37 Klemme, Heiner 37n.32 Kohl, Markus 78n.28 Korsgaard, Christine 16, 76, 105n.11, 158–9 Lavater, Johann Kaspar 49n.20 Laws Laws of freedom 18 Laws of nature 4, 8, 82n.35 Locke, John 46–52 Louden, Robert B. 56n.38 Ludwig, Bernd 10n.24, 93n.53 Łuków, Paweł 10n.24 Maimon, Salomon 72n.5, 73, 157–60 Mandelbaum, Maurice 118n.43 Marshall, Colin 3n.10 Matters of fact (see fact) 46–57, 62–4 Maxims 20n.9, 30–1, 40, 54–7, 79, 111, 128, 135–43 McCarty, Richard 119–23 McLear, Colin 78n.28 Mellin, Georg Samuel Albert 72–3, 85–6 Mendelssohn, Moses 21, 27n.18, 49, 102n.2 Menn, Stephen 21n.11 Metaphysics 2–6, 13
Index 175 Michaelis, Christian Friedrich 72–3 Moral consciousness (see also fact of reason, respect, necessitation) 8, 14, 53n.31, 54–7, 59–60, 63–5, 68–9, 76, 82n.35, 130, 133, 158–9 Moral education 57n.39, 61–2 Moral enthusiasm (see also self-conceit) 117n.40, 142–3 Moral feeling (see also respect, moral sensibility, moral motivation) 103–13, 118–31, 140–1, 154 Moral phenomenology 118–19 Moral sensibility 11–12, 12n.29, 14, 95–6, 100 Moral weakness 27–8 Morality (see also categorical imperative, autonomy) 16–34, 41–2, 63–8, 71–89, 98, 101–3 Moran, Kate 110n.24 Motivation 103–6, 120–3, 131–2, 144, 154–5 Motives 54, 60–4, 67–8, 87, 94–6, 134–5, 141–3 Moyar, Dean 57n.39, 60n.45 Munzel, Felicitas 49n.19, 149n.22 Nagel, Thomas 16, 76, 156n.1, 158–9 Nahm, Milton 115n.34 Necessitation 9–10, 39, 54n.33, 84–5, 87, 132 Necessity 9–11, 19, 30–7, 51–69, 81–9, 98n.60, 126–31, 156–60 Neiman, Susan 10n.24 Newton, Alexandra 107n.18 Normativity 1–4, 76, 156–60 Noumenal 8, 68n.58, 93n.53, 105–9, 122–3, 131–2 Obligation (see also duty, necessitation) 9–14, 19, 35–9, 65–7, 71, 125, 131–2, 134–54 Opacity thesis (see also self-deception) 25, 134–7, 143, 151–2 Ought (see also duty, obligation, necessitation) 8–9, 39–43, 78, 82–8, 95–9, 96nn.58–59, 139, 141–3, 149 Paton, Herbert 75, 104n.8 Pereboom, Derk 78n.28 Perfectionism 38–9, 117n.40, 137–41, 154 Persons 4–6, 117, 132 Personhood 117–18 Phenomenal 8, 105–6 Phenomenology of respect (see also moral phenomenology) 121–3, 130–1 Phenomenological method 8, 109, 122, 131 Philosophy as defense 5n.17 Philosopher’s stone 100–6, 122–3 Pistorius, Herman Andreas 72n.5, 93n.53
Plato 21, 25n.14, 40n.36, 102n.2 Pleasure (see also happiness) 103–31 Polemical method 8, 106, 122–3 Popular philosophy 8 Practical laws 36, 53–5, 63–4, 82n.34, 86, 95, 117, 127–8 Practical reason (see also will) Empirical practical reason 96 Pure practical reason 54n.33, 55nn.37–38, 78–88, 96–8, 110, 114–21, 128–32 Printy, Michael 49n.19 Progressive method (see synthetic method) Proof (see deduction) Proops, Ian 57n.39, 64–5 Puls, Heiko 130n.73 Purposiveness 24–5 Railton, Peter 76, 156n.1 Ratio cognoscendi (see fact of reason, disclosure thesis) Rauscher, Frederick 54n.33 Rawls, John 1–2, 5n.17, 16–18, 35n.29, 56n.38 Reath, Andrews 116n.39, 120–2, 120n.46 Regressive method (see analytic method) Reich, Klaus 102n.2 Reinhold, Karl 49–50, 72–3, 156–9 Respect 26, 58–60, 103–31, 140–1 Ress, Johann Heinrich 49n.20 Reversal reading 72–6, 83–6, 156 Rohden, Valerio 58n.41 Rosati, Connie 76, 156n.1 Ross, William David 10–11 Rousseau, Jean-Jacques 5–6, 13 Russell, Francey 110n.24 Russell, Paul 48n.16 Sade, Marquis de 59n.43 Saunders, Joe 91n.49 Schopenhauer, Arthur 45n.4, 53n.31, 74–6 Schapiro, Tamar 112 Schönecker, Dieter 12n.29, 22n.12, 37n.33, 87n.43, 123–4, 127n.65 Schaffer, Simon 47n.11 Schneewind, J. B. 68n.58 Schumski, Irina 136n.8 Scoundrel 10–14, 20, 39–41, 73n.6, 82–3, 98, 139 Self-conceit 26, 109–19, 134, 137, 141, 145, 154–5 Self-deception 132, 134–43, 148 Self-legislation (see also autonomy) 91–2, 117–18 Self-love 19–20, 26, 27n.19, 28–9, 34–6, 103–4, 109–14, 121, 132, 135–6
176 Index Self-approbation 113–18 Self-reproach 8–12, 115 Self-respect 59–64 Self-trust (see also trust) 153 Sensibility 8–14, 33–9, 41–2, 80–1, 90–100 Sentimentalism 101–3 Serjeantson, Richard 47nn.10–11 Shapin, Steven 47n.11 Shapiro, Barbara 47nn.9–10 Skeptical method 7, 17–18, 22–3, 31–2 Skepticism 5, 12–13, 16, 68, 105–6 Soul 40, 61–2, 74, 94, 97–8, 114n.31, 142–4 Spalding, Johann Joachim 48–9, 62 Spontaneity 91–5 Stang, Nicholas 3n.10 Standpoints First person 2–3, 53, 122–3 Third person 106–9, 121–3 Stattler, Benedict 72n.5 Stern, Robert 1–2, 37n.33, 83n.37 Sticker, Martin 11n.25 Sublime 115, 142n.17 Sullivan, Roger 136n.8 Synthetic method/path 11, 20–2, 41–3, 78, 89–90, 131–2 Tenenbaum, Sergio 40n.36, 91n.49, 159n.14 Testimony 46–7, 59, 153 Thiel, Udo 107n.17 Thought experiments 44–6, 58–62 Timmermann, Jens 16–17, 37n.33, 83n.37, 86n.40 Timmons, Mark 118n.43 Tippmann, Caroline 49n.19 Tittel, Gottlob August 89n.46 Tizzard, Jessica 107n.18
Transcendental idealism (see phenomena, noumena) Trust 14–15, 153–5 Truthfulness 58, 60, 151–4 Universal validity 55n.36, 79–81, 157–8 Vanzo, Alberto 53n.30 Velleman, David 1–2, 76, 158–9 Vocation 4–15, 23–42, 49, 116–18, 134–41 Walker, Ralph 115n.34 Ware, Owen 72n.2, 81n.32, 107n.18, 157n.3 Watkins, Eric 3n.10 Wilde, Oscar 142 Will 7–10, 23–7, 38–43, 53–69, 72–99, 102–14, 120–2, 128–32, 149 Williams, Bernard 13–14, 105nn.10–11 Willaschek, Marcus 46n.5 Williams, Terence Charles 10n.24 Wish 40–1, 50n.26, 61–2 Wolff, Michael 10n.24, 79n.30 Wolff, Robert Paul 1–2 Worlds Archetypal world 79 World of sense 36–41, 79–83, 87–8, 91 World of understanding 37–8, 79, 93n.53 Moral world 81n.33, 97n.59 Intelligible world 37–43, 79–83, 86–9, 91–6, 117, 156 Wood, Allen 16–17, 87n.43, 103n.4 Wright, Cripsin 17–18 Wuerth, Julian 3n.10, 107n.15 Zammito, John H 49n.19 Zande, J. V. D 27n.18 Zinkin, Melissa 120n.46