Immanuel Kant's <i>Groundwork for the Metaphysics of Morals</i>: A Commentary 9780674736214

A defining work of moral philosophy, Groundwork for the Metaphysics of Morals has been highly influential and famously d

163 93 1MB

English Pages 228 [248] Year 2015

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Contents
Preface
1. Kant’s Preface: The Metaphysics of Morals and the Strategy of the Groundwork
1.1 The Task, Method, and Transitions of the Groundwork
1.2 The Concept and Idea of a Metaphysics of Morals
1.3 Summary
2. Section I of the Groundwork: The Good Will, Duty, and the Derivation of the Categorical Imperative
2.1 The Structure and Argument of GMS I
2.2 The Good Will
2.2.1 The Conditioned Worth of Gifts of Nature
2.2.2 The Conditioned Worth of Gifts of Fortune
2.2.3 The Efficacy and Effects of the Good Will
2.2.4 The Teleological Argument
2.3 Duty and Respect
2.3.1 From the Concept of the Good Will to the Concept of Duty
2.3.2 The Three Propositions Regarding Duty
2.4 The Derivation of the Categorical Imperative
2.5 Summary
3. Section II of the Groundwork: Practical Reason, Imperatives, and Kant’s Formulas
3.1 The Structure and Basic Argument of GMS II
3.2 The Practical Faculty of Reason and the Division of Imperatives
3.2.1 Imperatives as Objective Rational Principles for Beings That Are at Once Sensuous and Rational
3.2.2 Categorical and Hypothetical Imperatives
3.3 The Possibility of Hypothetical Imperatives
3.4 The Formulations of the Categorical Imperative and Kant’s Examples
3.4.1 Enumerating the Various Formulas
3.4.2 The Derivation of Particular Duties: Kant’s Formulas and His Examples
3.4.3 The Categorical Imperative and the Relationship of Its Formulas to One Another
3.5 Summary
4. Section III of the Groundwork: The Deduction of the Categorical Imperative
4.1 The Structure and Task of GMS III
4.2 Freedom and Morality
4.2.1 The Thesis of Analyticity
4.2.2 The Presupposition of Freedom
4.3 The Deduction of the Categorical Imperative
4.4 Summary
5. Bibliography
5.1 Kant
5.2 Secondary Literature
Index
Recommend Papers

Immanuel Kant's <i>Groundwork for the Metaphysics of Morals</i>: A Commentary
 9780674736214

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Immanuel Kant’s Groundwork for the Metaphysics of Morals A Commentary

Dieter Schönecker and Allen W. Wood

Translated by Nicholas Walker

Cambridge, Massachusetts / London, England / 2015

Copyright © 2015 by the President and Fellows of Harvard College All rights reserved Printed in the United States of America First Printing Originally published in German as Immanuel Kant Grundlegung zur Metaphysik der Sitten: Ein einfürhrender Kommentar, copyright © 2002 Verlag Ferdinand Schöningh GmbH & Co. KG, Paderborn. The translation of this text was made possible by the generous fi nancial support of The University of Siegen Library of Congress Cataloging-in-Publication Data Schönecker, Dieter. Immanuel Kant’s groundwork for The metaphysics of morals : a commentary / Dieter Schönecker and Allen W. Wood. pages cm Includes bibliographical references and index. ISBN 978- 0- 674-43013- 6 (alk. paper) 1. Kant, Immanuel, 1724–1804. Grundlegung zur Metaphysik der Sitten. English. 2. Ethics. I. Title. B2766.Z7S27 2014 170—dc23 2014007806

For Rega und Rica

Contents

1

2

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Kant’s Preface: The Metaphysics of Morals and the Strategy of the Groundwork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

1.1 The Task, Method, and Transitions of the Groundwork . . . . . . . . . .

1

1.2 The Concept and Idea of a Metaphysics of Morals . . . . . . . . . . . . . .

14

1.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

Section I of the Groundwork: The Good Will, Duty, and the Derivation of the Categorical Imperative . . . . . . . . . . . . . . . . . . . . .

32

2.1 The Structure and Argument of GMS I . . . . . . . . . . . . . . . . . . . . . .

32

2.2 The Good Will . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34

2.2.1 2.2.2 2.2.3 2.2.4

The Conditioned Worth of Gifts of Nature . . . . . . . . . . . . . . . . . . . . . The Conditioned Worth of Gifts of Fortune . . . . . . . . . . . . . . . . . . . . The Efficacy and Effects of the Good Will . . . . . . . . . . . . . . . . . . . . . The Teleological Argument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3 Duty and Respect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 2.3.2

From the Concept of the Good Will to the Concept of Duty. . . . . . . The Three Propositions Regarding Duty . . . . . . . . . . . . . . . . . . . . . . 2.3.2.1 The First Proposition Regarding Duty as Acting out of Respect . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2.2 The Second Proposition Regarding Duty: The Objective Law. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2.3 The Third Proposition Regarding Duty . . . . . . . . . . . . . . .

34 40 41 47

50 51 54 58 76 78

viii

3

Contents

2.4 The Derivation of the Categorical Imperative . . . . . . . . . . . . . . . . .

89

2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

93

Section II of the Groundwork: Practical Reason, Imperatives, and Kant’s Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

95

3.1 The Structure and Basic Argument of GMS II . . . . . . . . . . . . . . . .

95

3.2 The Practical Faculty of Reason and the Division of Imperatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

97

3.2.1

Imperatives as Objective Rational Principles for Beings That Are at Once Sensuous and Rational. . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 3.2.2 Categorical and Hypothetical Imperatives . . . . . . . . . . . . . . . . . . . . . . 107

3.3 The Possibility of Hypothetical Imperatives . . . . . . . . . . . . . . . . . . 115 3.4 The Formulations of the Categorical Imperative and Kant’s Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 3.4.1 Enumerating the Various Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2 The Derivation of Par tic u lar Duties: Kant’s Formulas and His Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2.1 The Formula of Universalizability and the Formula of the Law of Nature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2.2 The Formula of Humanity as End in Itself . . . . . . . . . . . . . 3.4.2.3 Autonomy and the Realm of Ends . . . . . . . . . . . . . . . . . . . . 3.4.3 The Categorical Imperative and the Relationship of Its Formulas to One Another . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123 125 125 141 156 164

3.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

4

Section III of the Groundwork: The Deduction of the Categorical Imperative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 4.1 The Structure and Task of GMS III . . . . . . . . . . . . . . . . . . . . . . . . . 175 4.2 Freedom and Morality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 4.2.1 The Thesis of Analyticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 4.2.2 The Presupposition of Freedom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

4.3 The Deduction of the Categorical Imperative . . . . . . . . . . . . . . . . . 205 4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

5

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 5.1 Kant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 5.2 Secondary Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

Preface

Kant’s Groundwork for the Metaphysics of Morals is probably the most influential work of moral philosophy since classical antiquity. It is usually regarded as the principal source of several ideas that are fundamental to the modern moral outlook: the equal dignity of all human beings as ends in themselves; the conception of every rational being as a free and selfgoverning agent, owing obedience to no moral duties except those of which it can regard its own will as the author; and the idea of all rational beings as constituting an ideal community, bound only by those moral laws on which they might have rationally agreed. For a work that occupies this position, the Groundwork is surprisingly brief: only about seventy-five printed pages in the Academy Edition. It is also famously difficult; Kant was never a successful popu larizer of his philosophical ideas, and the Groundwork was devoted exclusively (as its title tells us) to laying the foundations of moral philosophy, not to answering the kinds of practical questions that most people might bring to a work on morals. For these reasons, the aims, theses, and arguments of the Groundwork are perplexing to many on first approach and also easy even for professional moral philosophers to misunderstand. The secondary literature generated in the last two and a quarter centuries by the study of Kant’s moral philosophy, and of the Groundwork in particular, is massive beyond description. But there are relatively few direct commentaries on it, and fewer still that, without sacrificing scholarly accuracy or philosophical rigor, are intended for the beginning student as

x

Preface

possible aids to the initial study of this important work. The present commentary is one with that intent, although we certainly hope that scholars may also profit from it. It is a slightly revised translation of our commentary in German, first published in 2001, which has gone through several editions since then (the fourth in 2011). We are gratified that it seems to have helped many readers of the original text understand it better, and we hope that an English version will enable many more readers to profit from reading the Groundwork. We intend our commentary to be relatively brief. Therefore, resisting the temptation to venture into scholarly controversy, we focus on the text of the Groundwork itself in a way that will be illuminating to introductory students of it (though we hope not only to them). There are many passages in the Groundwork that might be illuminated by considering Kant’s other ethical writings, in which he addresses many of the nonfoundational issues that are at the forefront of concern for most of his readers. The Groundwork has also been the basis for many creative interpretations of Kantian ideas, both in the voluminous secondary literature and in many important works of moral philosophy that Kant’s thought has influenced. We do from time to time mention materials from other Kantian writings when we think that they can be used briefly and effectively to shed light on what is said in the Groundwork. But our aim of keeping this commentary brief and useful in par tic u lar to introductory readers led us to limit our references to Kant’s other writings and virtually to exclude any discussion of secondary literature. Obviously, this does not mean that we have not benefited from the secondary literature; quite the opposite is true, and there is a great deal of discussion of it in our other writings that deal with the Groundwork. We also have included a selective bibliography of works, especially recent ones, that we recommend as further reading on the Groundwork and on Kant’s moral philosophy more generally. We have tried to remain very close to the text, to discuss specific passages, and always to orient our discussions to what Kant explicitly says in the Groundwork. However, we took pains not to turn our commentary into a paraphrase. Our attempt at every point is instead to explain Kant’s aims and to reconstruct his system of propositions and his arguments for them. We tried to keep our discussion centered on Kant’s texts rather than on what has been, or might be, made out of them by other philoso-

Preface

xi

phers (including ourselves). Nevertheless, since our aim is to help students of the text read it philosophically, we have tried to offer them a model of the way this can be done accurately and critically. We are constantly concerned not only with the accurate presentation of Kant’s arguments but also with their cogency, and we have not avoided criticizing Kant’s claims and arguments where we think this might be warranted. Further, we have also tried to indicate the points at which scholarly controversies and criticisms might arise, so that students may learn that, as with any great work of philosophy, including this one, there may be more than one way in which a text may be read and interpreted. The intent of our commentary is to make it easier for each reader of the Groundwork to arrive at her or his own informed and intelligent interpretation of this difficult text. The idea of a shared commentary grew out of our personal communication at Yale University while Schönecker was coediting a new edition of the Grundlegung zur Metaphysik der Sitten for Felix Meiner Verlag (published in 1999), and Wood was preparing a new English translation of it for Yale University Press (published in 2002). The contents of our commentary are of course nonaccidentally related to Wood’s discussion of the first two sections of the Groundwork in chapters 1–5 of Kant’s Ethical Thought (1999) and Schönecker’s discussion of the third section in Kant: Grundlegung III; Die Deduktion des kategorischen Imperativs (1999). But this commentary is a work independent of the earlier writings of either of us. (It is certainly not the case that Wood authored the commentary on the first two sections of the Groundwork [GMS I–II], while Schönecker wrote the commentary on the third section [GMS III].) As the native German speaker of the two of us, Schönecker drafted the entire text of the German commentary, which we then discussed and modified in countless ways until we had a text on which we could agree. Both of us similarly reviewed and discussed this English translation of the German text of the commentary, ably, accurately, and colloquially rendered in a draft by Nicholas Walker, and again, a number of modifications were made until the text suited both authors. Special thanks go to Larissa Berger and Theresa Specht, who in numerous ways helped us edit the text, and to Harvard University Press for the publication.

1 Kant’s Preface The Metaphysics of Morals and the Strategy of the Groundwork

The first part (1.1) of this chapter examines the central task and methodology of the Groundwork and the nature of the ‘transitions’ between the different sections of Kant’s text. The second part (1.2) investigates Kant’s concept and project of a ‘metaphysics of morals.’ The final part (1.3) summarizes the basic points and conclusions of our analysis. The discussion in this chapter may appear rather dry and abstract. However, it should afford a helpful view of the basic structure of the Groundwork for the Metaphysics of Morals (GMS),1 allowing us to concentrate in what follows on the fundamental arguments in Kant’s text.

1.1 The Task, Method, and Transitions of the Groundwork Kant did not begin his career as a moral philosopher. His concern with philosophy initially sprang from his specific interest in the natural sciences, and although he did engage repeatedly with questions of ethics and

1. See the Bibliography for an explanation of the various abbreviations appearing in the text.

2

Immanuel Kant’s Groundwork for the Metaphysics of Morals

aesthetics during the 1760s and the 1770s, we can still probably say that it was the field of theoretical philosophy that constituted the basic focus of his philosophical work until the publication of the Critique of Pure Reason in 1781. The Groundwork for the Metaphysics of Morals, published in 1785, is the first work that Kant dedicated exclusively to questions of moral philosophy. It was succeeded by further significant and substantial contributions to this field, such as the Critique of Practical Reason (1788) and the Metaphysics of Morals (1797–1798), and by a series of shorter but exceedingly important writings that also belong to practical philosophy.2 But even if Kant’s first publications were principally concerned with issues of theoretical philosophy, it is also clear that ethical questions occupied him from early on. It was in his Inquiry Concerning the Distinctness of the Principles of Natural Theology and Morality, completed in 1762 and published in 1764, that Kant first openly engaged with fundamental moral questions; this text already effectively involves the distinction between ‘hypothetical’ and ‘categorical’ imperatives.3 And in Kant’s correspondence from the middle of the 1760s on we find repeated references to what he calls a ‘metaphysics of morals.’ In his letter to Herder of 9 May 1768 Kant even says, “I hope to be finished with this work this year” (AA 10:74). About two years later Kant repeats his intentions in this regard. And it is clear in this connection that Kant already associates the envisaged ‘metaphysics of morals’ with an idea that he will never subsequently relinquish: that a metaphysics of morals involves an a priori ethics. Thus on 2 September 1770 he tells Johann Heinrich Lambert that he has “resolved to put in order and complete my investigations of pure moral philosophy, in which no empirical principles are to be found, as it were the Metaphysics of Morals” (AA 10:97; o.e. instead of Kant’s here). But if Kant had originally hoped to combine his theoretical and practical philosophy in a single work— one that he titles ‘The Limits of Sensibility and Reason’ in his let2. For example, Idea for a Universal History with a Cosmopolitan Aim (1784), An Answer to the Question: What Is Enlightenment? (1784), Conjectural Beginning of Human History (1786), On the Common Saying: That May Be Correct in Theory, but It Does Not Work in Practice (1793), Religion within the Boundaries of Mere Reason (1793–1794), and Toward Perpetual Peace (1795). 3. However, Kant does not yet deploy this specific terminology; cf. DG:298.

Kant’s Preface

3

ter to Marcus Herz of 21 February 1772—he soon abandoned this plan. In fact, over the next few years we hear nothing more from Kant about the idea of this ‘metaphysics of morals.’ 4 In his remarks at the beginning of the Preface (387–388), which are principally concerned with the ‘architectonic’ character of philosophy, Kant describes ‘ethics’ (die Ethik) as the science of the objects and laws of freedom. As a form of pure or a priori philosophy, this ethics is also described as a ‘metaphysics of morals.’ We shall examine the notion of ‘metaphysics’ and ‘the a priori’ that is involved in Kant’s ethics in some detail later, but in order to avoid unnecessary confusion, we should already clearly note that Kant employs the expression ‘metaphysics of morals’ in different ways within the text of GMS. In the most general sense, Kant understands it to signify that part of ethics that attempts to set out and to ground moral laws independently of experience (i.e., a priori). The Groundwork for the Metaphysics of Morals, as the name suggests, lays the ground or foundation (Grundlegung) of this project. But since GMS itself already proceeds in an a priori manner and must even answer the “principal question” (392:8) of the whole enterprise, the text itself is also already a metaphysics of morals. Kant also announces a ‘metaphysics of morals’ that he intends to provide “someday” (391:16) and whose task will be to present a systematic and complete “division of duties” (421:31). And Kant describes this as a “future metaphysics of morals” (421:32).5 In addition, we should note that Kant also bestowed the title ‘metaphysics of morals’ on that part of GMS to which he provides a transition in the second section of his text and beyond which in turn he moves in the third section. The concept of a ‘metaphysics of morals’ therefore has a threefold meaning: first, it is a general concept that captures Kant’s whole enterprise of a priori ethics, to which GMS also belongs (MS1 ); second, it also describes 4. For more on the historical background and specific development of GMS, cf. the introduction to the Felix Meiner edition of Kant’s text, edited by Bernd Kraft and Dieter Schönecker (Hamburg, 1999, pp. vii–xiii). 5. Whether the work that Kant envisaged as a ‘future’ metaphysics of morals in 1785 was actually realized in the Metaphysics of Morals that he published in 1797 is impossible to determine; it is also not easy to decide whether Kant kept to his original idea of a ‘division of duties.’

4

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Kant’s ‘future’ doctrine of right and virtue that is yet to be written (MS2 ); third, it designates one specific part of GMS itself (MS3 ).6

The Task of GMS Kant’s project of a ‘metaphysics of morals’ therefore falls into two parts: the ‘future’ metaphysics of morals (the doctrine of right and the doctrine of virtue proper: MS2) and the ‘foundational’ part: GMS itself. If we ignore Kant’s remarks regarding the method of GMS for the moment, then the basic structure of the work, at least, is easily grasped: GMS I and II are concerned with “the search” (392:3 ) for the categorical imperative (henceforth often abbreviated CI), while GMS III is directed toward the “establishment” (392:4) of the categorical imperative as the supreme principle of morality. What is the significance of this twofold division of GMS? In the first two sections of GMS Kant concentrates on analyzing the meaning of fundamental ethical concepts. What does it mean to call something ‘good’? What are we to understand by ‘the practical faculty of reason’? What does it mean for something to function as a ‘moral law’ in the fi rst place? How can we actually fi nd ethically acceptable principles? It is with questions such as these that GMS I and II are essentially concerned. Kant defi nes this task as “the mere analysis of the concepts of morality” (440:29; o.e.).7 Kant repeatedly emphasizes that, at least in the first instance, he is concerned simply with the meaning or ‘significance’ of our fundamental moral concepts. As he puts it himself: “Thus we have established at least this much: that if duty is a concept that is to contain significance and actual legislation for our actions, then this duty could be expressed only in categorical imperatives, but by no means in hypothetical ones” (425:1; o.e.); at this point we may “leave unsettled whether in general what one calls ‘duty’ is an empty concept” (421:11; o.e.) as long as we can at “least indicate what we are thinking in the concept of duty, and what this concept means” (421:12; o.e.). This analysis of our concepts is thus not 6. Of course, MS2 and MS3 belong to the whole project of a metaphysic of morals (MS1 ), but the point here is simply to clarify the use of the term ‘metaphysics of morals’ within the text of GMS. 7. Cf. also 397:1–10; 403:34–404:10; 406:5– 8; 412:15–25; 420:18–23; 421:9–13; 425:1– 11; 440:16–32; and 444:35–445:15.

Kant’s Preface

5

concerned with whether in fact there are any moral laws. It is concerned merely with what can properly be understood as the content of moral and ethical thought (namely, categorical imperatives). But at this stage it is by no means excluded that morality as a whole might simply turn out to be “a chimerical idea without truth” (445:6) or a mere “fiction of the mind” (445:8).8 It is only in GMS III that Kant addresses the question of the truth and reality of the concepts that have been analyzed in GMS I and II and attempts the ‘establishment’ of the CI. But these two parts of the text are not really separate from each other after all. For the principal result of the ‘search’ for the supreme principle of morality is the recognition that moral action is action for the sake of morality. Kant’s fundamental idea is that ethics and morality are essentially concerned with what we ought to do, and that when it comes to doing what is morally right, we should choose to do it, if need be, quite independently of any contingent interests or inclinations that might be presupposed on our part. The faculty or capacity that allows us to act independently of our subjective interests is what Kant calls pure practical reason. Kant believes that there is such a faculty, and he thereby directly challenges a tradition of thought for which the idea of rational action without any reference to presupposed interests is inconceivable and unintelligible. To this day the name of Kant basically stands for one (rationalistic) tradition, and that of David Hume for the alternative (empiricist) tradition that Kant specifically criticizes in GMS.9 If it is essential to the very meaning of ‘morality’ that it involves action independently of subjective interests, then Kant must also show that such action is possible in the first place, or, in other words, that the faculty of pure practical reason actually exists, or, in other words again, that we are free. Kant shows this only in GMS III (although, as we shall see, he prepares the way for it in GMS II). But even if Kant is able to show that we can in fact act morally (i.e., without regard to our desires or inclinations), he has not yet shown us that or why we ought to act morally. This is also the task of GMS

8. Cf. also 394:36; 421:11–12; 423:36–37; 429:35; and 440:20–28. 9. Some writers speak of ‘externalism’ and ‘internalism’ in this connection, but the use of such terminology is unfortunately rather inconsistent and easily leads to confusion.

6

Immanuel Kant’s Groundwork for the Metaphysics of Morals

III. Indeed, this is the principal task of what Kant calls the “deduction” (454:21) of the categorical imperative. GMS thus has two parts: it analyzes and it grounds the categorical imperative. GMS as a whole is in turn the first part of the metaphysics of morals (MS1 ) and answers the ‘principal question’ of the latter. The second part of this metaphysics of morals is then “the application of the same principle to the entire system” (392:8), namely, the systematic derivation of the duties of right and the duties of virtue from this highest moral principle.

The Method of GMS There is no doubt about the twofold division of GMS, but it is difficult to reconcile this actual division with Kant’s own methodological remarks, which are found in two specific places in GMS. At the end of the Preface, Kant tells us that he has chosen a “method” (392:17) that he believes is “the one best suited if one wants to take the way analytically from common cognition to the determination of its supreme principle and then, in turn, synthetically from the testing of this principle and its sources back to common cognition, in which its use is encountered” (392:18). And at the end of the second section, Kant tells us that GMS I/II have been “merely analytical” (445:8), and that we cannot venture on “a possible synthetic use of pure practical reason” (445:11) without a critique of the faculty of pure practical reason that is to be accomplished in GMS III. Kant thus seems to distinguish, methodologically speaking, between an ‘analytical’ and a ‘synthetic’ part of GMS, and this, in turn, would seem to correspond to the actual distinction between the conceptual-analytical task that is undertaken in GMS I/II and the task of grounding that is addressed in GMS III. Nonetheless, the matter is not so simple. In the Prolegomena (1783 ) Kant stresses that the analytical method is “something quite different from an aggregate of analytic propositions” (PM:276, footnote; o.e.).10 Since the conceptual-analytical part of GMS is ultimately nothing but such ‘an ag-

10. In this passage he also emphasizes that although the expressions ‘analytical’ and ‘synthetic’ have now become ‘classical’ in the sense of being traditionally established, he wants to propose a “new and more appropriate application” of the terms in question (PM:276, footnote).

Kant’s Preface

7

gregate of analytic propositions,’ the ‘analysis of concepts’ that is conducted in GMS I/II must be ‘quite different’ from the analytical method (and so too the synthetic method must be ‘quite different’ from the synthetic treatment of concepts in GMS III). But that implies that the actual twofold division of GMS is not meant to be understood in the same way as the twofold distinction of method that is described in the Prolegomena. In that case, Kant’s remarks on method, and especially his use of the term “method” (392:17), should not be interpreted in the light of the Prolegomena either. So when Kant deploys the predicates ‘analytical’ and ‘synthetic’ in GMS, he is not referring to the analytical or synthetic method; rather, he is thereby referring to the conceptual-analytical part, namely, GMS I/II (the ‘analysis of concepts’), and to the part concerned with grounding the central concepts, namely, GMS III (the ‘deduction’). And the expression ‘method’ also has a quite different significance here. Kant is thereby referring to the transitions that are specifically accomplished in the three sections of GMS. Kant’s ‘method’ in this work, therefore, consists in taking what he calls “common rational moral cognition” (392:23 ) as a point of departure and going on to distinguish different levels of rational moral cognition proper—the first drawn from ordinary human cognition (or ‘common sense’), the second from a philosophical presentation of rational moral cognition. It is in the context of these different levels that Kant provides both a conceptual analysis and a deduction of the relevant moral concepts.11

The Relationship among the Transitions in the Text Kant begins the Groundwork with a discussion of our ‘common rational moral cognition.’ Before examining this in more detail, we should initially clarify the relationship among the various transitions in the text. It should be clear that the first “transition from common rational moral cognition to philosophical rational moral cognition” (392:23 ) is actually accomplished in GMS I. It is quite true that right at the end of GMS I

11. GMS I and II are thus indeed ‘merely analytic.’ GMS III is ‘synthetic’ insofar as it undertakes to ground the categorical imperative as a synthetic a priori principle; Kant’s talk about a ‘possible synthetic use of pure practical reason’ is to be understood in this sense too.

8

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Kant tells us that we can now “step into the field of practical philosophy” (405:23 ) and “seek help in philosophy” (405:32). But that is certainly not meant to suggest that we are not already engaged with ‘philosophical rational moral cognition’ in GMS I itself.12 In the first place, GMS I is specifically titled ‘Transition from Common Rational Moral Cognition to Philosophical Rational Moral Cognition,’ and why would Kant describe this section in such terms if the transition in question were not in fact accomplished here? In the second place, Kant subsequently tells us quite explicitly that he has actually taken the step from “common moral judgment . . . to philosophical [judgment]” (412:15–17).13 The third transition “from the metaphysics of morals to the critique of pure practical reason” (392:27) clearly begins where the second transition fi nishes, namely, with the ‘metaphysics of morals’ (MS3 ). But the second transition (“from popular moral philosophy to the metaphysics of morals,” 392:25) does not begin from the end of the preceding (first) transition. Thus ‘philosophical rational moral cognition’ is not identical with ‘popular moral philosophy.’ Kant begins his second transition with “popular moral philosophy” (406:2) because he wants to demonstrate once again that we can derive neither the meaning nor the validity of moral laws from the domain of experience. To summarize: in the three sections of the text Kant passes not from A to B, from B to C, and from C to D, but from A to B (from common to philosophical rational moral cognition), from C to D (from popular moral philosophy to metaphysics of morals, MS3 ), and from D to E (from metaphysics of morals to the critique of pure practical reason). It will soon become clear that C (‘popular moral philosophy’) and D (MS3 ) can initially be understood as rival versions of B (‘philosophical rational moral cognition’), although D is ultimately exhib-

12. This ‘philosophical rational moral cognition’ should therefore not be identified with ‘practical philosophy’ proper. There is a transition from common rational moral cognition to philosophical rational moral cognition, but given the ‘dialectical’ problems that threaten to arise in this connection, we still require a transition to practical philosophy in the narrow sense (as we shall shortly explain). 13. Cf. Kant’s retrospective remarks in a subsequent footnote: “In the First Section we have seen that with an action from duty it is not the interest in an object that has to  be looked to, but merely the action itself and its principle in reason (the law)” (414:34–36; o.e.).

Kant’s Preface

9

ited as the only meaningful option. In a certain respect, therefore, C actually stands in a close relationship to A (namely, to ‘common rational moral cognition’).

The Starting Point: Common Rational Moral Cognition Kant begins the series of transitions with ‘common rational moral cognition,’ and if his remarks in the Preface are to be believed, he ends with this as well.14 He frequently emphasizes that GMS I contains, in principle, nothing more than what ‘common rational moral cognition’ already knows. In the Preface Kant already asserts that his claims regarding the intrinsic necessity and universality of moral laws are something that “everyone must admit” (389:11). He is convinced that “moral human reason even in the most common understanding, can easily be brought to great correctness and completeness” (391:21). Kant also believes, therefore, that the concept of an intrinsically good will “dwells already in the naturally healthy understanding” (397:2), and that the latter knows what the CI enjoins: common human reason agrees, indeed, perfectly with this—namely, with Kant’s claims regarding the ‘mere lawfulness’ of the CI—“in its practical judgment, and has the principle just cited always before its eyes” (402:13 ). Hence, after the successful transition to ‘philosophical rational cognition,’ Kant can write: “Thus in the moral cognition of common human reason we have attained to its principle, which it obviously does not think abstractly in such a universal form, but actually has always before its eyes and uses as its standard of judgment” (403:34). And at the beginning of GMS II Kant looks back and tells us once again that “we have thus far drawn our concept of duty from the common use of our practical reason” (406:5). But what precisely are we to understand by ‘common rational moral cognition’? And why does Kant begin his investigation by taking this as 14. In the Preface Kant writes that he will proceed “in turn from the testing of this principle and its sources back to common cognition, in which its use is encountered” (392:19; o.e.). This movement ‘back’ is to be found, if it is to be found at all, in Section 4 of GMS III: “The practical use of common human reason confi rms the correctness of this deduction” (454:20). For Kant’s use of the expression ‘common human reason,’ cf. 389:9–23; 394:34–35; 397:2–4; 402:13–15; 403:34–405; 406:5– 6; 412:15–22; and 454:20–455:9.

10

Immanuel Kant’s Groundwork for the Metaphysics of Morals

his point of departure? It seems plausible to interpret the moral cognition of common reason15 simply as moral common sense, that is, as the basic moral convictions that most human beings appear to share. For, as we have already seen, Kant emphatically believes that there is no fundamental difference in principle between the moral knowledge explored by philosophical ethics and what he calls our “common moral judgment” (412:15): “It might even have been conjectured in advance that the acquaintance with what every human being is obliged to do, hence to know, would also be the affair of everyone, even of the most common human being” (404:7; o.e.).16 The task of ‘philosophical rational moral cognition’ is to “enlighten” (397:3 ) and “develop” (397:6) the ethical (moral) knowledge that our ‘common human reason’ already contains. In this sense, as Kant observes, common reason possesses the relevant moral principle, “which it obviously does not think abstractly in such a universal form [as philosophical rational moral cognition does], but actually has always before its eyes and uses as its standard of judgment” (403:35; o.e.). Thus ‘common moral judgment’ is “worthy of great respect” (412:16) and differs from philosophical ethics solely on account of its lesser degree of abstractness.17 But it would be misleading to interpret ‘common rational moral cognition’ as nothing more than a fundamental, though ultimately contingent (i.e., culturally and historically conditioned), understanding of morality. It is quite clear that some of the ethical conceptions that Kant ascribes to ‘common human reason’ differ significantly from other traditional ethical notions. Thus other cultures and ethical systems have certainly entertained the idea that happiness is the highest good (which Kant denies), and also that particular characteristics such as ‘moderation’ or ‘courage’ are unconditionally good (which Kant also denies). Furthermore, the

15. And the other words and expressions that Kant uses: ‘understanding,’ ‘human reason,’ ‘human understanding,’ ‘judgment,’ and ‘cognition.’ 16. Cf. also KpV:8: “A reviewer who wanted to say something censuring this work [GMS] hit the mark better than he himself may have intended when he said that no new principle of morality was 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, fi rst invent it? Just as if, before him, the world had been ignorant of what duty is or in thoroughgoing error about it.” 17. Cf. also 390:14–16 and 409:12–14.

Kant’s Preface

11

claim that certain ethical conceptions concur with those of most human beings or even an entire culture is, of course, an empirical claim that may or may not be correct. GMS offers no empirical evidence or support for an empirical claim regarding ‘common moral rational cognition.’ Hence it looks as though what Kant calls ‘common rational moral cognition’ should not simply (merely) be understood as a form of abstract knowledge, nor merely as a sort of layman’s ethical theory, for it also represents a quite elementary form of knowledge with regard to action and conduct, one that Kant believes is already involved and presupposed in our actual moral practice in a prereflective way. Wisdom, as Kant points out, “consists more in deeds and omissions than in knowledge” (405:2; o.e.). Kant wishes to underline that, pathological cases apart, every human being, as a morally acting subject, is quite capable, independently of her or his specific level of educational attainment and certainly of any philosophical knowledge or expertise, of understanding moral principles and acting in accordance with such principles. Kant is assuming (what many philosophers and social scientists might well question) that every human being has the capacity to think for himself or herself and also the capacity to recognize some important moral truths that may be at odds with that human being’s contingent education or social conditioning, or any moral doctrine or theory that might have been imposed in these ways. That is why, Kant tells us, a human being “needs no science or philosophy to know what one has to do to be honest and good, or indeed, even wise or virtuous” (404:5). Thus Kant thinks that if we are to hold human beings to be morally responsible, we must also ascribe to them the (same) cognitive capacities that are required in order to have a ‘good will.’18 Moral philosophers have enjoyed no special monopoly on the concept of virtue, and in fact the ‘common understanding’ can enjoy “just as good a hope of getting things right as any philosopher might promise to do; indeed it is almost more secure in this even than the latter” (404:23 ). 18. Kant’s claims regarding the conditioned value of natural abilities and gifts of fortune should also be interpreted from this perspective. This is also why Kant does not simply speak of ‘common rational moral cognition’ or moral ‘judgment’ in this connection. In the relevant passage at the beginning of GMS II Kant says that we have drawn the concept of duty “from the common use of our practical reason” (406:5; o.e.); cf. also 405:31.

12

Immanuel Kant’s Groundwork for the Metaphysics of Morals

But if ‘common rational moral cognition’ already knows all this in principle, would it not “be more advisable in moral things to stay with the judgment of common reason” (404:29)? Kant’s answer to this question (404:37–405:35) rests on the anthropological thesis that human beings are massively exposed to the corrupting influences of their ‘inclinations’ (about which we shall have more to say shortly). From this constant conflict between the moral claims of reason and the urge to satisfy our needs (to fi nd happiness) there “arises a natural dialectic, that is, a propensity to ratiocinate against those strict laws of duty . . . and, where possible, to make them better suited to our wishes and inclinations” (405:13 ).19 It is precisely the task of philosophy to counter this ‘propensity.’ That is why, as Kant argues here and in other places, there is not merely a theoretical need for a metaphysics of morals but also a practical one. This is not (really) contradicted by the fact that in the Preface to GMS Kant justifies the title Groundwork for the Metaphysics of Morals (“instead of the term critique of pure practical reason” [391:31]) by observing, among other things, that ‘common human reason’ is not particularly dialectical in relation to ethical questions (391:20–24). In the first place, Kant does not say in the Preface that there is absolutely no danger of such a dialectic here. He simply says that this danger is not especially great in this context (as distinct from the case of theoretical reason that is “entirely dialectical” [391:24; o.e.]), and that in the field of moral philosophy, therefore, a critique of pure practical reason is “not of such utmost necessity” (391:20; o.e.). In the second place, the ‘natural dialectic’ in question does not merely lie in the fact that human beings feel within themselves a “powerful counterweight” (405:5) against the requisite moral laws. It is quite true that this dialectic ultimately “arises” (405:13; o.e.) from the polarized force field of reason and sensibility. But the dialectic “ensues” or unfolds “in common practical reason” (405:30; o.e.), once it is “cultivated” (405:31; o.e.) and subjected to ‘ratiocination.’ Kant takes both human civilization and the development of our

19. We should note that this is a ‘natural dialectic.’ It arises both out of our natural inclinations as animal beings and through the process through which our faculties become cultivated (and also corrupted) in society by historical processes that Kant regards as guided by a natural teleology (see I:17–22).

Kant’s Preface

13

rational faculties not only to provide us with the capacity to think correctly about morality but also with distorting influences that may ‘dialectically’ corrupt that same thinking. And this dialectic is cultivated, among other things, by a ‘popular moral philosophy’ that also pays heed to human nature and cannot produce more than “half-reasoned principles” (409:31). It still remains to ask why Kant chooses to begin with ‘common rational moral cognition’ at all rather than broaching the problem in entirely abstract terms from the start or simply starting directly with the discussion pursued in GMS II. There are two reasons for this. First, Kant tells us explicitly that in GMS I and II he is solely concerned with analyzing concepts. In order to counter the possible objection that he is simply loading the concepts in question with everything that he wishes subsequently to derive (GMS I/II) and to justify (GMS III), Kant claims that this conceptual explication is simply the explication of the very meanings that ‘common rational moral cognition’ also connects with our moral concepts. Kant is therefore referring and appealing, so he claims, to our actual use of language and to our elementary knowledge of human conduct, that is, to the moral and ethical conceptions with which he fi nds himself presented, and, moreover, is referring and appealing to the use of practical reason itself. It is precisely through this connection that Kant hopes to lend further plausibility to the conceptual explication provided in GMS I/II. Second, Kant also wants to counter the ‘charge of obscurity’ that had been raised against him on all sides ever since the publication of the Critique of Pure Reason.20 He wishes to show that a ‘popular’ style does not even have to be attempted here since we are dealing with universally recognized knowledge anyway, and he specifically wishes to show, at the beginning of GMS II, where a mistaken notion of ‘popularity’ can actually lead.21 20. Kant addresses this problem in some detail in the Prolegomena, a work that can be seen as a direct response to this charge of obscurity (cf. the preface and the “Remarks on Method” in par ticu lar). 21. On this question of how ‘popu lar’ in character philosophical works can or should be— a highly topical and controversial issue in Kant’s time— cf. 391:34–36 and, particularly, 409ff.

14

Immanuel Kant’s Groundwork for the Metaphysics of Morals

1.2 The Concept and Idea of a Metaphysics of Morals After this preliminary overview of the basic tasks and method of GMS, we must now consider the very concept and idea of a metaphysics of morals. Kant certainly did not invent this particular expression.22 But what precisely does it mean here?

The Apriority of the Metaphysics of Morals It is impossible to overlook the fact that Kant regarded the clarification of the concept of metaphysics as an absolutely fundamental task. In GMS Kant repeatedly returns to this concept and contrasts his own ethical theory, as a metaphysics of morals, from other types of ‘practical philosophy.’ Apart from his remarks in the Preface (387–392), Kant specifically describes what he means by a metaphysics of morals at a number of crucial points in the text itself: at the beginning of the second section, where he expressly contrasts it with ‘popular ethical philosophy’ or ‘popular moral philosophy’ (406–412:14); immediately before the transition to the metaphysics of morals (MS3 ) in GMS II (425–427:18); in the contrast that he draws between the fundamental concepts of autonomy and heteronomy (441:25–444:34); and finally, briefly, in Section 2 of GMS III (447:28– 448:4), where he accomplishes the transition from the metaphysics of morals (MS3 ) to the critique of pure practical reason.23 Now these programmatic passages can also be dangerously misleading because they can easily give the impression that Kant is essentially concerned with an ethics that can be successfully pursued without recourse to any empirical (psychological, social, or historical) knowledge about human beings, and in which the concrete consequences of our moral conduct can simply be regarded as irrelevant. At first sight, indeed, this interpretation hardly appears unjustified. We read, for example, that “it is of the utmost necessity to work out once a pure moral philosophy which is fully cleansed of everything that might be in any way empirical and belong to anthropology” (389:7); that all moral philosophy “rests entirely on

22. The philosophical school that derived from Christian Wolff already recognized a discipline called ‘metaphysica moralis.’ 23. Cf. also MS:214ff.

Kant’s Preface

15

its pure part, and when applied to the human being it borrows not the least bit from knowledge about him” (389:27); that we are concerned above all with “a pure rational cognition abstracted from everything empirical” (409:16); that the principles of morality must be discovered “in pure concepts of reason, fully a priori, free from everything empirical” (410:11); and that there cannot be “too many warnings against this negligent or even base way of thinking that seeks out the principle among empirical motivations and law” (426:12).24 These passages and others like them have often distorted the common image of Kant’s ethical thought to the point of caricature. But such distortions serve only to conceal the authentic core of Kant’s thoroughly rationalist concerns here. The fact that Kant refers repeatedly to the “practical importance” (411:18) of a moral theory that acknowledges ethical dispositions should already have alerted us in this regard, for in these decisive passages Kant explicitly emphasizes that we have more than strictly theoretical reasons for endorsing an a priori ethical perspective. In Kant’s view, a metaphysics of morals is also required “because morals themselves remain subject to all sorts of corruption as long as that guiding thread and supreme norm of their correct judgment is lacking” (390:2). And the metaphysics of morals is “not only an indispensable substrate of all theoretical cognition of duties which is securely determined, but it is at the same time a desideratum of the highest importance for the actual fulfilment of its precepts” (410:22). Only an a priori ethics of duty is capable of withstanding the corrupting influence of human nature, and indeed we require philosophical “science not in order to learn from it but in order to provide entry and durability for its precepts” (405:3 ). Moral actions must be performed for the sake of morality itself. Mere conformity to duty is not enough “because the unmoral ground will now and then produce lawful actions, but more often actions contrary to the law” (390:7), and “a mixed doctrine of morals, composed from incentives of feelings and inclinations and simultaneously from concepts of reason, must make the mind waver between motivations that cannot be brought under any principle,

24. Kant repeatedly warns against the danger of ‘mixing’ pure principles with empirical (or any other sort) of principles: cf. 388:27; 390:13; 409:16; 409:30; 410:8; 410:16; 410:22; 410:26; and 411:3.

16

Immanuel Kant’s Groundwork for the Metaphysics of Morals

and can lead us only very contingently to the good, but often also to the evil” (411:3 ). It is clear that an empirical theory of human nature lies behind these claims. According to this theory, our natural needs and inclinations by no means naturally or spontaneously accord with moral demands. No inclination or natural feeling (in contrast to rational feeling) can produce what reason is able to produce, namely, reliable moral dispositions (and actions). This is a claim that may be true or false (it was highly controversial in Kant’s own time and is probably still so today), but it is unambiguously empirical in nature. Thus whatever the a priori character of Kant’s ethical metaphysics may be, it cannot signify that Kant excludes empirical knowledge altogether. But then what is it precisely that allows us to characterize Kant’s moral philosophy as ‘pure’ and ‘a priori’ in character? What is it precisely that makes it a ‘metaphysics of morals’? The general answer to this question, and thus the basic idea of a metaphysics of morals, can be framed as follows: if we are to speak meaningfully of moral principles at all, then these principles must be binding for all rational beings (universally binding) unconditionally and without exception (necessarily binding). Kant makes this quite clear in the lectures on moral philosophy that he presented around the time he began thinking about GMS: “Morality cannot be constructed on empirical principles, for this could supply merely conditioned rather than absolute necessity—yet morality says: You must do this without condition or exception” (MM II:599; o.e.).25 Thus necessity and universality belong to the very concept of morality. This is what morality ‘says,’ and in this sense Kant undertakes to “analyze” the language of morality in GMS I/II. In GMS Kant repeatedly employs the concepts of ‘necessity’ and ‘universality’ to describe the character of moral laws, and it is quite clear that he wishes to establish a parallel with theoretical philosophy in this regard. For Kant, the CI is itself a “synthetic-practical a priori proposition” (420:14). We should note, of course, that Kant speaks of a synthetic-practical a priori proposition here, as distinct from synthetic-theoretical a priori 25. Kant delivered these lectures (MM II) in the winter semester of 1784/1785; GMS was written in 1784 and published in 1785.

Kant’s Preface

17

propositions. In GMS Kant himself refers specifically to “propositions of this kind in theoretical cognition” (420:15; o.e.). This is why, in the Preface, he speaks of “laws” (387:13 ) in the most general terms, which are specified and distinguished as laws of nature and laws of freedom only later. This is also why in the Preface he starts by asking “how much pure reason could achieve in both cases” (388:37; o.e.) before proceeding to ‘limit’ the task of GMS to an examination of the “pure” part of ethics (389:6). Nothing indicates this parallel between theoretical and practical philosophy more clearly than the fact that Kant expresses the categorical imperative in a formula that specifically compares morality to a law of nature. Laws of nature (a priori) possess universality and necessity, and it is precisely these formal properties, so Kant claims, that also characterize moral laws, or at least must characterize them if they are to have any meaning and significance: “Because the validity of the will as a universal law [o.e.] for possible actions has an analogy with the universal connection of the existence of things in accordance with universal laws, which is what is formal in nature in general [o.e.], the categorical imperative can also be expressed thus: Act in accordance with maxims that can at the same time have themselves as universal laws of nature for their object” (437:13 ). In the case of these imperatives we are talking about “the lawfulness of actions generally similar to an order of nature” (431:26).26 In spite of this ‘analogy’ that Kant himself draws between laws of nature and laws of morality, the ‘universality’ and ‘necessity’ we encounter in the realm of practical philosophy are different from those encountered in the realm of theoretical philosophy. Even in the ‘practical’ context Kant does not always employ these concepts in a univocal fashion. But it is nonetheless quite clear how and why Kant develops and understands his ethical philosophy specifically as a metaphysics of morals. 1. Necessity as Categorical Validity. If we understand ‘subjectivist’ ethical theories as those that trace the meaning and validity of normative claims back solely to subjective preferences (interests), then Kant’s ethical theory is emphatically anti-subjectivist. Expressed in positive terms, we must say 26. For the concept of ‘analogy’ here, cf. also 438:23. We shall examine the relevant concept of nature later in our analysis.

18

Immanuel Kant’s Groundwork for the Metaphysics of Morals

that Kant defends an ethical objectivism. The fundamental significance of ‘necessity’ here consists in this very simple but far-reaching thought: moral laws are necessarily binding insofar as they command us to perform or avoid certain actions quite irrespective of our interests or inclinations. A moral principle, therefore, must be one that “grounds itself on no interest and hence it alone among all possible imperatives can be unconditioned” (432:16). This also implies that moral laws must be binding without “exception” (424:20), for exceptions arise precisely from attempting to avoid the categorical validity of the moral law “for the advantage of our inclination” (424:19).27 The most helpful clarification of the meaning of ‘necessary’ in this connection is probably that provided in one of Kant’s footnotes. It is no accident that he also uses this occasion to elucidate precisely what he means by a ‘synthetic-practical proposition’: “I connect the deed a priori with the will, without a presupposed condition from any inclination, hence necessarily” (420:29; o.e. instead of Kant’s here). In this sense ‘necessary’ means nothing more than ‘categorical’: “The categorical imperative would be one which represented an action as objectively necessary for itself, without any reference to another end” (414:15).28 ‘Necessity’ in the categorical sense also implies apriority. This is quite clear from the decisive passages regarding the concept of a metaphysics of morals that we have already indicated. Thus the Preface already presents the fundamental idea that the actions we describe as morally valuable in the strict sense are actions that are done from duty: “For as to what is to

27. This does not mean, however, that Kant allows no exceptions whatsoever. 28. It is important to recognize that Kant also has a concept of ‘necessity’ that relates to nonmoral principles of reason. This will become clearer when we discuss the ‘hypothetical’ imperatives, but it is naturally irrelevant here where we are examining the idea of a ‘metaphysics of morals.’ With regard to the categorical character of the moral law, Kant employs the terms ‘necessary’ or ‘necessity’ in the following passages: 401:32; 415:4; 416:22; 416:28; 419:14; 419:21; 420:10; 420:28; 420:30; 421:4; 424:24; 426:22; and 427:37. He speaks of the associated necessity of performing an action on the basis of duty in the following passages: 400:18; 403:30; 412:32; 413:3; 414:19; 418:30; 425:16; 429:15; 429:29; 434:16; and 449:21. The fact that moral laws are binding without exception also follows directly from their categorical character. In this regard, cf. 408:17; 421:34; 424:20; and especially 424:15–37. Kant also understands ‘necessity’ to refer to the validity of the moral law in general, which he addresses explicitly in GMS III; cf. 444:36; 445:11; 449:26; 449:30; 463:25; and 463:29.

Kant’s Preface

19

be morally good, it is not enough that it conform to the moral law, but it must also happen for the sake of this law” (390:4). Kant defines a will that is determined by purely moral motives as a “pure will” (390:26). This is a will that is “determined without any empirical motives fully from principles a priori” (390:25). Shortly afterward, Kant once again distinguishes “motives that are represented as such fully a priori merely through reason, and are properly moral, from empirical ones” (391:4). It is therefore the principles that determine action without regard to interests and inclinations that are ‘pure’ and ‘a priori.’ And since these inclinations and interests can be known only on the basis of experience, Kant also describes them as ‘empirical.’ Thus an ethics that excludes such inclinations and interests from any role in grounding and determining the moral law is properly described as an a priori ethics.29 This basic thought, along with its conceptual context, is repeated and developed in more detail at other crucial points of the text and particularly in the critique of “popular moral philosophy” (407–412). Kant starts by emphasizing that the concept of duty cannot be treated as a “concept of experience” (406:7) (which also implies that ethics, at least as regards the discovery and grounding of its fundamental principle, cannot legitimately be pursued in empirical terms). Kant presents three arguments in favor of his thesis, one of which is explicitly connected with the principle of universality (which we shall examine more closely later). The two other arguments can be considered here. 1. If the concept of duty were empirical, we should be able to know from experience that there are such things as actions performed out of duty. But Kant maintains that this is impossible, for we cannot be sure that actions that seem to others and also to ourselves to be motivated in purely moral terms do not actually spring from the principle 29. In Kant’s discussion of his ‘second proposition’ regarding the concept of duty, he says that the will stands “at a crossroads, as it were, between its principle a priori, which is formal, and its incentive a posteriori, which is material” (400:10). Cf., in addition, Kant’s distinction between ‘formal’ and ‘material’ rational principles in the second clarification of the faculty of practical reason (427:19–428:2). For his identification of material incentives with ‘empirical’ ones, cf. 389:22; 410:26; 411:1; 418:6; 426:7; 426:14; 442:33; 450:11; 460:35; and 462:26.

20

Immanuel Kant’s Groundwork for the Metaphysics of Morals

of self-love.30 Whether there have ever been any actions performed from duty is a question that Kant regards as irrelevant (“We are not talking here about whether this or that happens,” 408:1; o.e.). All that counts here is the theoretical question concerning the validity of the claim that “reason commands, for itself and independently of all appearances, what ought to happen” (408:2).31 Thus although Kant propounds the strong thesis that there could be purely moral actions (i.e., actions performed without regard to any already-given subjective wishes), he himself excludes the possibility of empirically establishing that such actions really occur or proving that they actually exist. 2. In order to recognize a particular action as an example of morally right or required action in general, we must already possess a concept of what morally right or required action is. This is why ethical concepts themselves cannot be derived from examples.32 This argument is directly connected with Kant’s distinction between the principium diiudicationis (standard) and the principium executionis (incentive), a distinction that is bound up in turn with his critique of the theory of ‘moral sense.’33 Kant admits moral feeling (properly understood) as a motive for action, but he insists that only reason can know what is morally right or wrong. Human beings do actually take an interest in moral laws, “the foundation of which in us we call ‘moral feeling,’ which is falsely given out by some as the standard of our moral judgment, since it has to be regarded rather as the subjective effect that the law exercises on the will, for which reason alone provides the objective 30. Cf. 407–408:11. The same argument is presented again, in a particularly clear formulation, in GMS II (419:16–35). 31. The term ‘appearances’ here refers to “desires and inclinations” (cf. 453:24). Kant repeats this last point in the third crucial passage of the text (immediately before the transition to the metaphysics of morals itself, or MS3 ) when he writes that “what is to be established are not grounds for what happens, but laws for what ought to happen, even if it never does happen, i.e., objectively practical laws” (427:1). 32. Cf. 408:28–409:8. 33. Kant was well informed about the philosophical tradition that appealed to ‘moral sense,’ ‘moral feeling,’ or ‘moral sentiment’ (as represented by Anthony Ashley Cooper, Third Earl of Shaftesbury, Adam Smith, Francis Hutcheson, and David Hume). Cf. Kant’s explicit reference to Hutcheson in the footnote in GMS II (442).

Kant’s Preface

21

grounds” (460:2).34 But Kant is not merely claiming that reason is in this sense the principium diiudicationis, for he also claims that pure (moral) reason can itself become practical without regard to subjective interests, namely, that it can be the principium executionis. It is true that it does so through a kind of feeling (the feeling of respect). But he takes this feeling to be self-wrought through pure reason and hence solely on its account; this illustrates that pure reason can motivate our actions. That reason is capable of doing so is what Kant then goes on to show in GMS III. 2. Necessity as Universality. The notion of necessity as universality involves a simple, though philosophically challenging, thought: moral laws spring from reason and are also therefore valid or binding for all rational beings (not only for human beings).35 Kant’s remarks regarding the range of those addressed by moral laws are quite unambiguous: “It is self-evident from the common idea of duty and of moral laws” (389:10), he claims, that “the command ‘You ought not to lie’ is valid not merely for human beings, as though other rational beings did not have to heed it” (389:13 ); the moral law is “of such extensive significance that it would have to be valid not merely for human beings but for all rational beings in general” (408:14); “Moral laws are to be valid for every rational being in general” (412:2); and duty “must be valid for all rational beings (for only to them can an imperative apply at all)” (425:17). The question therefore is this: “Is it [the moral law] a necessary law for all rational beings?” (426:22). The universality of the moral law consists precisely in the way its principles “are to be valid for all rational beings without distinction” (442:7). But what exactly 34. For the distinction in question, cf. Menzer (44) and Refl:5448; for Kant’s critical observations on the idea of ‘moral feeling,’ cf. GMS, 442:22–443:2, and MM II:625– 626. 35. This idea is also presented in the crucial passages to which we have already referred: 389:13–23; 408:12–27; 410:3–18; 411:22–412:8; 425:12–31; 426:22–427:18; 442:6– 12; and 447:28–448:4. Kant uses the terms ‘universality’ and ‘universal’ (and other variations) very frequently throughout GMS, although in the specific sense we have just described rather rarely. One of the few places in which he does so is the following passage, where he says that the moral law (in regard to the human being as an end in himself) possesses “universality, since it applies to all rational beings in general” (431:2). For other examples of this use, cf. 408:21; 427:36; 437:37; 438:22; and 442:7.

22

Immanuel Kant’s Groundwork for the Metaphysics of Morals

is Kant’s argument for the thesis that moral laws must be valid and binding for all rational beings? And what is the connection between this universality of the moral law and its ‘necessity’? The answer to the second question already contains the answer to the first. Kant argues that the only principles that deserve to be described as ‘moral’ are those that are categorically valid and binding, that is, without regard to our interests and inclinations (and therefore without exception). If it were impossible to act rationally without such rational actions being motivated by subjective interests, then there could be no such thing as moral actions. There could also be no such thing as universal moral rules because human beings have all sorts of different interests (and if moral rules had to be bound and connected to these interests, then they too would be dependent, as far as content and justification were concerned, on these individual interests). Hence it is the intrinsically good will alone, according to Kant, that is morally valuable in an authentic and unqualified sense. But this good will is nothing other than respect for the moral law, and this respect in turn is nothing other than the effect of pure practical reason, which is, independently of all inclinations, both principium diiudicationis and principium executionis. Neither nature nor indeed God is the origin or ground of the binding force of the moral law. For this ground is pure practical reason alone. Hence Kant concludes his extensive reflections on the metaphysics of morals as follows: “From what we have adduced it is clear that all moral concepts have their seat and origin fully a priori in reason” (411:8). This is why he repeatedly emphasizes (as he already does in the Preface) that he is essentially concerned with the ‘pure will,’ which is nothing other than ‘reason that is practical in its own right.’ In the metaphysics of morals, therefore, “we are talking about objectively practical laws, hence about the relation of a will to itself insofar as it determines itself merely through reason, such that everything that has reference to the empirical falls away of itself; because if reason for itself alone determines conduct (the possibility of which we will investigate right now36), it must necessarily do this a priori” (427:13 ).37

36. Here ‘right now’ refers to the metaphysics of morals as part of GMS II (MS3 ). 37. Since Kant understands moral laws to be necessary rational principles, these laws also possess ‘universality’ with respect to their ‘form.’ Kant uses the terms ‘uni-

Kant’s Preface

23

This thought—moral laws must possess categorical validity (necessity); they can do so only as rational laws; therefore, moral laws are also valid as rational laws for all rational beings (universality)— can be confirmed through a close reading of the text. First, it is clear that Kant constantly emphasizes the connection between the necessity and the universality of moral principles.38 The argument proceeds in two directions: The demand for universality results from the demand for necessity: “For duty ought to be the practically unconditioned necessity of action; thus it must be valid for all rational beings . . . , and must only for this reason be a law for every human will” (425:15; o.e. here, Kant’s emphasis omitted). But the necessity of the moral law depends entirely on reason (and thus on universality): “Empirical principles are everywhere unsuited to having moral laws grounded on them. For the universality, with which they are to be valid for all rational beings without distinction, the unconditioned practical necessity, which is thereby [i.e., by ‘universality’] imposed on these beings, drops out if the ground of these principles is taken from par ticular adaptations of human nature or from the contingent circumstances in which it is placed.” (442:6; o.e. here, Kant’s emphasis omitted). Sometimes one may have the impression that Kant is directly, as it were, concerned to emphasize that all rational beings belong to the sphere of those addressed by the moral law.39 But he is simply attempting to reveal that moral laws must be nonempirical rational principles. The “extensive significance” (408:14) of the moral law with regard to the range of those addressed by it thus springs simply from the fact that this law is a rational law: “For since morality serves as a law for us merely as rational beings, it must also be valid for all rational beings” (447:30; the second emphasis is ours). versality’ or ‘universal’ in the sense of possible universality or ‘universal validity’ (438:2; 449:32; 458:15; 460:28; 461:27; 462:14) in relation to a maxim or law in the following passages: 402:6– 9; 403:7–11–16–22–25; 421:2– 8–14–16–19; 422:3–7–12–28– 29–31; 423:8–12–23–29; 424:2–5–8–17–18–23–24–30; 426:24; 430:32; 431:11; 432:11; 433:19; 434:12; 436:15–17–27–32; 437:8–10–13–18; 438:11–22; 441:23; 444:32; 447:4–12; 449:10; 458:30; and 460:22. Kant also speaks in this context of a ‘universally legislating will’: 431:17–20; 432:13–16–21–30–31; 433:13; 434:13–15–26; 435:31–37; 438:10; 439:2–25; 440:9–11–20; 441:4; 449:11; and 432:3 (‘universally legislative will’). We shall discuss these two aspects later. 38. In par tic u lar, cf. 389:11–16 and 408:12–18. 39. Cf. 408:19–24 and 412:2–3.

24

Immanuel Kant’s Groundwork for the Metaphysics of Morals

The a priori character of the metaphysics of morals thus springs from the necessity (the categorical feature) and the universality (the rational feature) of the moral law. But this by no means already exhausts all aspects of Kant’s concept of the a priori. The notion of the a priori is also intimately connected with the whole conceptual network involving ‘freedom,’ ‘spontaneity,’ and ‘autonomy,’ as well as the theme of ‘enlightenment.’ 3. Apriority and Freedom. We shall have to examine the question of freedom in much greater detail, of course, when we come to interpret GMS III. But the Preface already makes clear the importance of this concept. Here Kant already characterizes moral laws specifically as laws of “freedom” (387:14). In GMS III this aspect plays a central role because the concept of freedom is regarded as “the key to the definition of the autonomy of the will” (446:6). And moral laws, Kant tells us in the same context, “must be derived solely from the quality of freedom” (447:32). In the first place, and quite generally, Kant understands freedom as a capacity or faculty of spontaneity, that is, the capacity to do something or produce something entirely of its own accord, without any internal or external determining factors. In the context of theoretical philosophy, in addition to the passive faculty of receptivity, we have the active faculty of spontaneity as the second fundamental faculty of human cognition. This is the faculty that allows reason and the understanding to produce concepts, judgments, and inferences. The basic idea of Kant’s theory of knowledge is that all human cognition involves certain a priori elements. This already holds for space and time as the two a priori forms of intuition, and it emphatically holds for the process of thinking, which actively (i.e., a priori) brings the material of sensuous intuition, given to us in space and time, into specific forms of relation and context. It is important for Kant’s theory of knowledge that we clearly distinguish such a priori elements from the earlier notion of innate ideas. What is innate has been “given” to us from birth (planted within us by God, for example, or provided by genetic inheritance), and indeed independently of our experience and the exercise of our cognitive faculties alike. What is a priori, on the other hand, is produced by ourselves through the exercise of our own faculties. It is not difficult to extend this idea to the field of practical philosophy. Only when we grasp moral laws as produced a priori by ourselves and

Kant’s Preface

25

our practical reason can we do proper justice to the content of this law. If the idea of the moral law is to make any sense at all, Kant claims, it can do so only on condition that such laws are valid and binding in a universal and necessary way. But they can be valid and binding in this way only if they are prescribed and imposed by ourselves, that is, autonomously, rather than prescribed and imposed heteronomously, whether by God, by our environment with its various authorities and traditions, or by our natural endowments or genetic predispositions. For if these moral laws are heteronomous, then they are always inevitably bound up with some interest that we take in an action, and this contradicts the idea of disinterested action that is performed from duty.40 Our practical faculty of reason does not consist merely in the fact that we can somehow react or respond to our desires and inclinations in a specific way. On the contrary, it is we ourselves, through our own reason and independently of empirical interests, who are the source of the a priori principles in accordance with which we act. Moral laws spring from the freedom and spontaneity of our practical reason. In this sense too, these laws are a priori, and so too is the philosophy that presents and explicates them (the metaphysics of morals). 4. Apriority and Enlightenment. It is a fundamental concept of Kant’s theoretical and practical philosophy that we can think and act autonomously. It is also a characteristic expression of the age that, perhaps, Kant most powerfully represents—the Enlightenment. Kant tells us that “the maxim according to which we should always think for ourselves is the maxim of enlightenment” (O:146). We cannot pursue this idea in further detail here, but we should not forget that Kant’s celebrated call for us “to make use of our own understanding” (WA:35) has profound philosophical roots that lie in the thought of apriority itself.41 40. This thought is expressed with par ticu lar force at 432:25–433:11; we shall come back to this question later. 41. The idea of ‘thinking for oneself’ and the allusion to the famous phrase of Horace, Sapere aude (cf. WA:35), can already be found in the very early representatives of Enlightenment thought. The notion of thinking-for-oneself (Selbstdenken as a substantive) actually occurs only once in GMS (388:27, where Kant refers to ‘Selbstdenker,’ or ‘independent thinkers’), and then in a rather ironic and negative sense.

26

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Anthropology and the Metaphysics of Morals In GMS Kant expressly contrasts his ethics of autonomy with an ethics of heteronomy, but he also emphatically contrasts his form of ethics with what he calls ‘anthropology.’ 42 He introduces anthropology as the “empirical part” (388:31) of ethics, but “ahead” of this (388:36) we require a metaphysics of morals that is quite “independent” (412:5) of anthropology. The metaphysics of morals must therefore be “fully cleansed of everything that might be in any way empirical and belong to anthropology” (389:8) and must be “mixed with no anthropology” whatsoever (410:19). This contrast between the metaphysics of morals and anthropology can be understood, in part, as one aspect of the distinction between empirical and a priori ethics in general. The empirical conception of ethics (an ethics based on the idea of happiness, for example) inevitably appeals to the ‘nature’ of human beings, and this in turn, in a certain sense, is the appropriate object of anthropology as an empirical science. Since anthropology plays no further direct role in the essential development of the argument of GMS (as a ‘science’ in its own right), there is no need to examine any further Kant’s precise conception of anthropology as a science within his system as a whole. But we should note that in the Preface Kant presents anthropology as the empirical part of ethics that certainly also considers the moral laws “in accordance with which everything ought to happen” but also reckons “with the conditions under which it often does not happen” (388:1). That rather sounds as if the task of anthropology were to discover the conditions that allow us to set forth moral laws for human beings in the first place. This reading would seem to be supported by the comparison that Kant draws between anthropology as applied ethics— applied “namely to human nature” (410:32)— and applied mathematics and logic (410, footnote). It is quite true that he explicitly says here that “moral principles are not grounded on the peculiarities of human nature, but must be subsistent a priori for themselves” (410:34, footnote). But then from these principles in turn “practical rules must be derivable” (410:36) in relation to human nature, and that would obviously be the task of anthropology. But if we were to understand Kant’s remarks in this way, 42. The concept of ‘anthropology’ is mentioned in the following passages: 388:13; 388:35; 389:9; 389:28; 410:20; and 412:5.

Kant’s Preface

27

it would remain unclear, in the first place, just what these rules, supposedly derived through application to human nature under empirical conditions, could possibly consist in. And, in the second place, Kant already clearly emphasizes in the Preface that “all moral philosophy rests entirely on its purest part, and when applied to the human being it borrows not the least bit from knowledge about him (anthropology)” but “gives him as a rational being laws a priori” (389:26; o.e.). In fact, it is quite possible to understand the remarks in Kant’s footnote (410) and the earlier passage from the Preface (387) in a different way. Kant’s reference to applied logic is particularly instructive in this connection. He is alluding here to the distinction between general pure logic and general applied logic as this was commonly drawn in his time and as he himself also presents it in the Critique of Pure Reason (A52ff./B77ff.). ‘Pure logic’ roughly corresponds to what we would now understand as ‘formal logic’ (which was limited at this period, of course, to the theoretical analysis of concepts, judgments, and different forms of inference). ‘Applied logic,’ on the other hand, as Kant says, is really a form of psychology rather than logic,43 for it investigates not how we ought to think but only how in fact we do so under specific empirical conditions and circumstances (in relation to memory, intellectual habits and tendencies, etc.). It is interesting to note that in the Critique of Pure Reason Kant actually compares the distinction between pure and applied logic with the distinction between pure and applied ethics. He claims that the two logical disciplines stand “in the same relation as pure ethics, which contains only the necessary moral laws of a free will in general, stands to the doctrine of the virtues strictly so called—the doctrine which considers these laws under the limitations of the feelings, inclinations, and passions to which men are more or less subject” (KrV:A54–55/ B79). Comparable in this regard to applied logic, applied ethics investigates not how we ought to act, but how in fact we act under specific empirical conditions and circumstances (in relation to inclination, tradition, etc.). This fits plausibly with the fact that in GMS Kant also understands anthropology essentially as “knowledge” about human beings (389:28). If we understand anthropology in this sense as an empirical discipline, it becomes quite clear what the task of such an anthropology is and precisely why it differs from a metaphysics of morals: “In a practical philosophy, where what is to 43. Cf., for example, KrV A53/B77, JL AA 9:18.

28

Immanuel Kant’s Groundwork for the Metaphysics of Morals

be established are not grounds for what happens, but laws for what ought to happen, even if it never does happen, i.e., objectively practical laws, there we do not find it necessary to institute an investigation into the grounds why something pleases or displeases, how the gratification of mere sensation is to be distinguished from taste, and whether the latter is distinct from a universal satisfaction of reason; on what the feelings of pleasure and displeasure rest, and how from them arise desires and inclinations, and from these, again, through the cooperation of reason, maxims arise; for all that belongs to an empirical doctrine of the soul” (427:1).44 Kant also mentions anthropology once again in his later exposition of the metaphysics of morals (MS2 ). There he stresses that we “often have to take as our object the par ticular nature of human beings, which is cognized only by experience, in order to show in it what can be inferred from universal moral principles. But this will in no way detract from the purity of these principles or cast doubt on their a priori source.—That is to say, in effect, that a metaphysics of morals cannot be based upon anthropology but can still be applied to it” (MS:217). Apart from clearly indicating that for Kant a moral philosophy can have empirical elements even though this ‘in no way detracts from the purity of moral principles,’ it is still unclear what Kant means by trying to ‘show’ these principles in relation to human nature. What Kant then immediately goes on to describe as “moral anthropology” (ibid.) corresponds precisely to applied ethics (‘practical anthropology’ in the sense we have already described): “The counterpart of a metaphysics of morals . . . would be moral anthropology, which, however, would deal only with the subjective conditions in human nature that hinder people or help them in fulfilling the laws of a metaphysics of morals. It would deal with the development, spreading, and strengthening of moral principles (in education in schools and in popular instruction), and with other similar teachings and precepts based on experience” (ibid.). We should also note that in addition to this ‘practical’ or ‘moral’ anthropology, Kant also engaged with what was called ‘pragmatic anthro44. It admittedly seems problematic that this empirical doctrine of the soul, as Kant says, should constitute “the second part of the doctrine of nature, if one considers it as philosophy of nature insofar as it is grounded on empirical laws” (427:10). In the Preface to GMS anthropology is understood as the empirical part of ethics and is specifically distinguished from the ‘doctrine of nature.’

Kant’s Preface

29

pology’ and frequently delivered lectures on this subject.45 The precise status of this pragmatic anthropology is also difficult to determine. In a lecture transcript (MM II:599) that dates from the same period as GMS, practical anthropology is presented as part of pragmatic anthropology, and the latter is concerned, among other things, with the human being as a free being.46 What all of this signifies is of no further concern to us here, but we should never forget that Kant returned repeatedly to these anthropological questions (even if he remained skeptical with regard to the ‘scientific’ status of this subject). We have already pointed out that in spite of the emphatic ‘purity’ of the metaphysics of morals, GMS is not presented completely without reference to our empirical knowledge regarding human beings, and Kant never claimed that it could be so presented. If we understand this ‘purity’ in the sense already suggested (necessity as categorical validity and necessity as universality), then the character of the CI, the ground of its validity, and the range of those addressed by it are essentially a priori. But that does not imply that moral ends can be set and moral laws (duties) can be derived without recourse to empirical knowledge. We can argue about whether Kant’s claims in GMS regarding the purpose and function of anthropology can or cannot be understood in a way that excludes all anthropological knowledge from the concept of moral legislation. The only thing that is relevant for our purposes is that it is quite impossible to understand the various derivations of our moral duties without reference to empirical context; in this sense anthropology certainly belongs to practical rather than theoretical philosophy. We do not employ experience solely to determine the particular means for particular ends. Experience is also employed to establish (moral) ends, as we shall see more clearly when we discuss Kant’s famous ‘examples.’47

45. It is only as of 1998 that the relevant student transcripts of Kant’s lectures on anthropology have been reliably edited (AA 25). It should be pointed out that the early lectures in this field are never described as ‘pragmatic anthropology.’ 46. Cf. Anth.:119. 47. The anthropological and historical dimensions of Kant’s philosophy as a whole would also serve to bring out what one could call, if we wish, the realistic and empirical character of his ethical thought.

30

Immanuel Kant’s Groundwork for the Metaphysics of Morals

1.3 Summary48 1. GMS has a twofold task: GMS I and II of the text analyze the basic concepts of ethics and introduce the CI as the fundamental principle of morality. In GMS III the grounds are given why we may take ourselves to be free, and why the moral law is valid. 2. The distinction between analysis and deduction should not be identified or conflated with the distinction between the analytical and the synthetic method. It is true that Kant describes the conceptualanalytic part of the text (GMS I/II) as ‘analytic’ and the ‘grounding’ part of the text (GMS III) as ‘synthetic.’ However, the concept of ‘method’ in GMS refers not to these descriptive terms but to the procedure involved in the transitions. 3. With regard to these transitions, we should observe that the fi rst one is actually accomplished, and the point to which the transition is made should not be identified or confused with the starting point of the second transition. 4. Kant repeatedly emphasizes the fundamental accordance between ‘common rational moral cognition’ and ‘philosophical rational cognition.’ What he wishes to show here is that moral action requires no special capacities or particular knowledge, and that his conceptual analyses accord with our actual use of language. The task of philosophy is to elucidate more precisely the moral-ethical knowledge harbored by our ‘common reason’ and (thereby) to guard against the danger of a certain ‘natural dialectic’ (that of conceptual ratiocination). 5. Kant uses the expression ‘metaphysics of morals’ in a threefold sense: first, as a general concept for a priori ethics itself (MS1 ); second, for the metaphysics of morals that Kant has yet to write (MS2 ); and third, for that part of GMS to which the transition is made in Section II of the text (MS3 ). 6. Kant draws a parallel between the universality and necessity of moral laws and the universality and necessity of natural laws; it is from this

48. With regard to this summary and especially those that conclude the following (longer) chapters, it should be noted that we merely present the principal and most important conclusions of the preceding discussion.

Kant’s Preface

31

idea, along with the concept of freedom, that the a priori character of the metaphysics of morals derives. ‘Necessity’ here signifies that moral laws must be valid and binding independently of our inclinations, and therefore categorically valid and binding without exception; in addition, we can never know from experience whether there are such things as acts performed out of duty, and we can never learn from examples what is morally required of us. The ‘universality’ of moral laws derives from the fact that they spring from human reason itself (rather than from our ‘nature’) and are therefore also valid and binding for all rational beings; and it is only as rational laws that they can also be ‘necessary.’ As laws of freedom, moral laws are at the same time autonomous laws that are produced by ourselves (and in this sense are a priori); in a political and general historical perspective this approach is thus also intrinsically bound up with the Enlightenment idea of freely thinking and acting for oneself. 7. Kant repeatedly insists on the nonempirical status of the metaphysics of morals. This easily creates the impression that he wishes to banish all empirical knowledge from the field of ethics proper. But that cannot really be the case since the practical relevance of the metaphysics of morals— only an ethics based on reason can produce reliable dispositions—rests on the anthropological thesis that human beings are massively exposed to the corrupting influence of nature. 8. Considered as a systematic science and an empirical part of ethics, anthropology plays no role whatsoever in GMS, but that does not mean that empirical knowledge plays no role in the formulation and derivation of concrete duties. What Kant understands by anthropology as a specific discipline in relation to the pure metaphysics of morals itself remains rather unclear in GMS (and not only there).

2 Section I of the Groundwork The Good Will, Duty, and the Derivation of the Categorical Imperative

We shall begin with a brief overview of the structure and argument of the first section of the Groundwork (2.1). The structure of our analysis and commentary will then essentially follow the basic structure of the text: we shall examine the concept of a will that is good in itself (2.2), proceed to Kant’s analysis of the concept of duty (2.3 ), and conclude with a discussion of Kant’s first derivation and formulation of the categorical imperative (2.4). Finally, we provide a brief summary of the entire argument (2.5).

2.1 The Structure and Argument of GMS I It is not difficult to grasp the overall structure and organization of Section I. Kant begins by examining the concept of the good will (393–396). It should be noted here that rather than telling us precisely what the good will is, he simply tells us what an unqualified good cannot be. It is only with the analysis of the concept of duty that the concept of the good will is specifically “developed” (397–401), and it is on the basis of these analyses that Kant then derives his initial formulation of the categorical imperative (402–403:33 ). He concludes with some remarks on the “natural dialectic” of common human reason (403:34–405). Since we have already discussed this last part of Section I in our interpretation of Kant’s Preface, there is

Section I of the Groundwork

33

no need to examine it in any further detail here, and it contributes nothing in particular to the fundamental argument of GMS I (indeed, that last part comes after the derivation of the categorical imperative). We shall therefore concentrate specifically on the concept of the will, the concept of duty, and the derivation of the categorical imperative. The structure of the corresponding passages in the text is also relatively easy to grasp. The analysis of the concept of a will that is good in itself falls into two main parts: (1) Kant shows that what he calls ‘gifts of nature’ and ‘gifts of fortune’ can claim only a conditional or qualified worth (393:5–394:12), he also argues in this connection that considerations of ‘efficacy’ and ‘effects’ play no part in determining the worth of a good will (394:13–394:31); and (2) he provides a teleological argument (394:32–396:37) to show that it is the task of reason to produce a good will rather than happiness. In the following paragraph of the text (397:1–397:10) Kant advances from the concept of a will that is good in itself to the concept of duty. He understands the analysis of the concept of duty here as a ‘development’ of the concept of the good will, where this is now explicitly understood as the will of a being that is a moral as well as a rational being. The analysis of the concept of duty is presented in three ‘propositions,’ although the first one is not expressly formulated as such (397:11–399:34). The second proposition is clearly formulated and elucidated (399:35–400:16), as is the third (400:17– 401:2). Finally, Kant once again summarizes his thesis regarding the absolute worth of the will that is good in itself (of acting from duty) (401:3– 401:40) and adds a footnote intended to clarify the concept of respect. It is only the good will and, with regard to human beings, respect for the moral law that can claim to be an unlimited or unqualified good. But for all the apparent clarity of the structure of GMS I, it is one of the partic u lar difficulties of this section of the Groundwork that the question regarding precisely what law it is that demands respect is raised and answered only after the analysis of the concept of the will and of duty, namely, in the discussion at 402:1–403:33. We can provide the following summary overview: 393–396: The Concept of a Will That Is Good in Itself 393:5–394:12: The conditioned worth of gifts of nature and gifts of fortune

34

Immanuel Kant’s Groundwork for the Metaphysics of Morals

394:13–31: The efficacy and effects of the good will 394:32–396:37: The teleological argument 397–401: Analysis of the Concept of Duty 397:1–397:10: Advance to the concept of duty 397:11–399:34: The first proposition regarding duty 399:35–400:16: The second proposition regarding duty 400:17–401:2: The third proposition regarding duty 401:3–401:40: Summary of the argument 402–403:33: Derivation of the Categorical Imperative 402:1–402:15: The derivation 402:16–403:33: An example of the categorical imperative 403:34–405: Common Rational Moral Cognition and Its Natural Dialectic

2.2 The Good Will The basic question of GMS I is this: what does it mean to describe something as morally good, and indeed as morally good in an unqualified sense? What are we referring to when we say that something is morally good or bad? We react with approval or censure when we pass moral judgment on others or, indeed, on ourselves (in this case we speak of self-respect or shame). But what precisely are we passing judgment on here? Are we talking about the individual acts of human beings? Are we talking of acts in relation to their consequences? Or is it the motives and intentions of the human beings who act that we are passing judgment on? And is it human beings as such, as persons, or is it specific virtues and attributes that we are referring to when we judge in this way? We can distinguish different kinds of moral theory in terms of how they answer these questions. Kant’s answer is that it is the moral quality of the will that is decisive here.

2.2.1 The Conditioned Worth of Gifts of Nature The opening sentence of GMS I expresses the principal thesis of the first three paragraphs: “There is nothing it is possible to think of anywhere in the world, or indeed anything at all outside it, that can be held to be good without limitation, excepting only a good will” (393:5). By this Kant does not mean that all goods are in principle good only insofar as they depend on

Section I of the Groundwork

35

a morally good will. That cannot be so, given that there are many situations in which certain actions are morally neutral, where something is good for something, and also good in an unlimited sense. There are many things that are goods quite independently of the good will and that have worth as such. Kant’s thesis that such goods possess “no inner unconditioned worth” (394:1) can mean only that the value of things of that kind is conditional. Thus if Kant claims in this context that the gifts of nature and fortune are “in some respects good and to be wished for” (393:10) or “good for many aims” (394:5), this does not necessarily mean that they are ‘good’ only when they are connected with a morally good will.1 That is true only when ‘good’ is interpreted as something that is ‘morally good’ or of ‘absolute worth.’ Things and attributes can also be described as ‘good’ in the broader context of means-ends relations independently of the good will.2 Kant by no means wishes to claim that the good will alone is intrinsically good (good in itself). If we understand something of intrinsic worth as a good that we pursue (de facto) for its own sake, then Kant is far from claiming that there are no such goods to be found among gifts of nature and gifts of fortune. On the contrary, he explicitly says, for example, that we pursue happiness “in accordance with a natural necessity” (415:32). It is not Kant’s thesis, therefore, that the gifts of nature, and happiness in particular, are not things of intrinsic value in this sense. To say that something is intrinsically good, for Kant, is not the same as saying that it is good without limitation or qualification. Likewise, the fact that the good will alone possesses an unconditioned inner worth does not imply that there are no intrinsic goods except the good will.

1. The German dictionary compiled by the brothers Grimm distinguishes (in eighteenth-century usage) between the word Absicht in the sense of intentio and the phrase in Absicht auf (and Absicht plus the genitive) in the sense of ‘in relation to’ or ‘with regard to’ and the like. On pages 393 and 394 of the text Kant uses the word Absicht rather than in Absicht auf, which indicates that he is indeed thinking of intentio or ‘aim’ here (in English translations Absicht is thus often incorrectly rendered as ‘respect’ or ‘regard’). Absicht as intentio is best understood in this context as a ‘pursued end or aim’; see, however, notes 7 and 60 in this chapter. 2. In this sense, on pages 413–414, the concept of ‘good’ is also explicitly applied to means-ends relations.

36

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Maybe the best way to understand Kant’s thesis regarding that which is good without qualification (a will that is good in itself) is to rephrase the formula “good without limitation” by replacing the word ‘good’ with ‘valuable’ or ‘having worth.’ Then we see that Kant is claiming that the good will alone ‘is valuable without limitation’ (although it is quite true that Kant does not employ this expression himself). Kant repeatedly speaks of “value” or “worth” [Wert], and specifically in these fi rst three paragraphs of the text.3 Kant contrasts “the absolute worth of the mere will” (394:32) with certain things that it would be difficult to describe as “good without limitation” (394:7). A person deserves moral recognition not simply because she possesses certain ‘talents of mind’ and ‘qualities of temperament’ as “natural gifts” (393:12) or “gifts of fortune” (393:13 ); these ‘gifts’ are not considered good on the whole under just any and all conditions. Kant’s thesis is that these talents and qualities are only conditionally good; in contrast to the will that is good in itself, they possess “no inner unconditioned worth” (393:26).4 Kant here understands ‘gifts of nature’ to mean those more or less permanent characteristics of an individual that belong to his or her natural endowment: “talents of the mind” (393:8) and “qualities of temperament” (393:9). Kant specifically mentions understanding, wit,5 and the power of judgment as examples of talents of the mind and courage, resoluteness, persistence in an intention, moderation in affects and passions, self-control, and sober reflection as examples 3. Cf. 394:1; 394:6; 394:26; 394:27; and 394:31. 4. In these fi rst three paragraphs Kant himself does not employ the expression ‘conditioned,’ but he does speak of “unconditioned worth” (394:1) and refers to the praise that was ‘unconditionally’ (394:7) accorded to traditional virtues by “the ancients.” For a similar use of ‘unconditioned,’ cf. also 400:5; 401:10; and 436:3. Kant also explicitly refers back to the opening sentence of GMS I at 437:5–20, where he speaks of the “unconditionally good will” (437:6) rather than of the will that is good ‘without limitation,’ so that the expression ‘without limitation’ can be identified with “unconditionally” in this context. Cf. also 428:11, where Kant says that “all objects of inclinations have only a conditioned worth” (o.e.); in this regard cf. also 428:18 and 428:31 and Kant’s repeated reference to ‘conditions’ in his discussion of hypothetical imperatives and inclinations as themselves ‘conditions.’ 5. In a later passage Kant also mentions “wit” (435:9) as one of those qualities that have only an ‘affective price’ and cannot claim ‘absolute worth’ and, therefore, ‘dignity’: “by contrast, fidelity in promising, benevolence from principle (not from instinct) have an inner worth” (435:10).

Section I of the Groundwork

37

of qualities of temperament. The German expression usually rendered as ‘gifts of fortune’ [Glücksgaben] clearly reflects those gifts’ contribution to ‘happiness’ or ‘good fortune’ [das Glück]. But it also looks as though Kant wishes to emphasize that these gifts, unlike those of nature, are not (for the most part) naturally given but socially acquired. First of all, let us consider Kant’s observations regarding ‘gifts of nature.’ His argument here is simple and convincing: if such gifts of nature really enjoyed unconditioned, morally intrinsic worth, then we would approve their use even with a bad will and for bad ends. But we do not. Kant may also be suggesting that in exercising moral judgment we should have to describe someone who possesses all or some of these gifts as a person of ‘good character.’ But there are undoubtedly human beings who possess at least some of these attributes, but to whom we would certainly not ascribe a good character precisely because their intentions (and actions) are reprehensible. Thus these qualities and attributes cannot claim unconditioned worth, for their worth is limited or qualified by the context in which they are actually employed. Therefore, the gifts of nature are not a possible candidate for something that can properly be regarded as good ‘without limitation,’ for there is something (the will) that “limits” (394:3 ) the worth of these gifts insofar as this depends on the worth of the will that makes “use” of them (393:12; o.e.).6 Kant therefore understands the relation between something unconditionally good (the will that is good in itself) and something conditionally good (self-control, for example) as a relation between the unconditionally good qua condition and the conditionally good qua conditioned. The unconditionally good is the condition of the moral goodness of the conditionally good; it is the good will as unconditionally good that limits or qualifies the moral goodness of the conditionally good. The conditionally good or potentially evil gifts of nature are good or “also” (393:11) evil only in relation to the “intention” [Absicht] (393:10; 394:5) of the will in question (this is what defi nes their

6. Cf. MM II:599: “Understanding, bodily strength, prudence are good things, but extremely harmful when allied with an evil will.— Health, enjoyment, satisfaction, a constantly cheerful disposition, this is good only on condition that a human being has a good will that can make use of them in the fi rst place” (o.e.).

38

Immanuel Kant’s Groundwork for the Metaphysics of Morals

conditioned character). These gifts are “in some respects” good (393:10; 394:5) and in some respects evil. Here ‘in some respects’ really means ‘under some conditions’; it is only a good will that is good “in every respect and in all circumstances” (MM II:599; o.e.).7 If the gifts of nature are put to use by a good will, then they are themselves (conditionally) good; if such gifts are put to use by a will that is “not good” (393:13 ), then “they” (393:10; o.e.) are not good either. Thus the gifts of nature can “become” (393:11; 394:9) conditionally good or evil depending on the will that uses them and the kind of “principles” [Grundsätze] (394:8) that it follows. And it is because these kinds of things can become good or evil that they are neither good nor evil in themselves. Hence they possess no “inner” worth (394:1) but “always presuppose a good [or bad] will” (394:1) that determines their worth. “Among all the things that we describe as good, most of them are good in a conditional sense, and nothing is good without limitation except the good will” (MM II:599). Kant’s example of the “cold-bloodedness of a villain” (394:9) makes this quite clear.8 7. One of us (Dieter Schönecker) is not entirely happy with the translation of Absicht here (393:10; 394:5) as ‘respect.’ First, Absicht in German just does not mean Hinsicht (respect) but Absicht in the sense of intention (Lat. intentio); cf. note 1. Second (and here both of us are in agreement), at least in 393, Kant’s point seems to be that (for instance) a quality of temperament such as courage or resoluteness is not good merely with regard to a certain (fortunate) outcome but instead in its dependence on the intention with which it is used. For courage and resoluteness to be good, the intention and not merely the result must be good. 8. Kant’s concept of ‘character’ (393:13 or 398:37) implies more than the combination of a par tic u lar range of natural and acquired gifts and capacities in a specific individual. Rather, character is a “peculiar constitution” (393:12) of the will itself that must (‘should’) make ‘use’ of those gifts and capacities in the fi rst place. The good will (resp. its ‘constitution’) is “therefore” (393:13; o.e.) described as character, and not because it simply unites a number of virtues within itself. Kant describes this constitution as ‘peculiar’ [eigentümlich] because it is precisely the property of the will to make ‘use’ of the gifts of nature and of fortune (this is proper or peculiar to the will, and hence its ownmost property). In Anthropology from a Pragmatic Point of View Kant claims indeed that “a man of evil character (like Sulla)” (Anth:293 ) is repellent to us, but also that such an individual is “an object of admiration” at the same time for the firmness with which he holds to his maxims, monstrous though they are (ibid.). Since Kant’s concept of character plays no further role in the argument of GMS, we shall not pursue it in any further detail here.

Section I of the Groundwork

39

The second paragraph simply repeats and clarifies the argument of the first. Once again Kant invites us to ask: are these things indeed qualities or attributes that we would always and in all circumstances regard as praiseworthy? Kant expressly distances himself from the notions of virtue and the Good that were entertained by “the ancients” (394:8), who specifically “praised” (394:8) certain ‘qualities of temperament,’ and indeed praised them precisely as “unconditionally” (394:7; o.e.) good qualities of character (i.e., as virtues).9 Kant introduces these reflections as follows: “Some qualities are even conducive to this good will itself and can make its work much easier” (393:25). But it would be a misunderstanding to infer from this remark that only some qualities could be conducive to the good will.10 Kant’s argument is precisely that all gifts, of nature and of fortune alike, are capable of serving a morally determined will; the will can make ‘use’ of all of them, which is why Kant regards them as morally indifferent in themselves. But some of them (“even,” 393:25) manifest a particular propinquity to the good will. Kant thereby does justice to the fact that these qualities are typically ascribed to persons who possess a good will and thus also a good character. In this sense,11 they belong to “the inner worth of a person” (394:6), since a person who possesses a good will does show such qualities. Indeed, such a person must show them, for the good will itself cannot be “a mere wish” (394:23 ) but is something that involves “the summoning up of all the means” required (394:23). These 9. When Kant speaks of ‘the ancients,’ he is generally referring to the Greek and Roman philosophers. In MM II:599ff. he specifically mentions Diogenes, Epicurus, Plato, Zeno, and the Stoics in general; in his later Doctrine of Virtue Kant repeatedly mentions and strongly criticizes Aristotle (cf. MS:404 and 434 in par tic u lar). In the second paragraph we are discussing here, it is easy to recognize three of the four ‘cardinal virtues’ of the tradition: temperantia (moderation in affects and passions), fortitudo (courage), and prudentia (practical judgment). (In MM II:599 Kant also speaks of “prudence” [Klugheit] rather than “judgment” [Urteilskraft]; cf. the close connection that he establishes between judgment and prudence at MS:411 and 432–43 ). The missing virtue is justitia ( justice), which of course plays a very significant role among the cardinal virtues. 10. It is true that Kant speaks only of “qualities” here (393:25), but the examples and the specific context show that we are talking not only about ‘qualities of temperament’ (moderation in affects and passions, self-control) but also about ‘talents of the mind’ (such as sober reflection). 11. This is why Kant uses the word “seem” (394:5).

40

Immanuel Kant’s Groundwork for the Metaphysics of Morals

qualities are therefore ‘means’ of which the good will “ought” (393:12) to make use because they can “make its work much easier” (393:26).12 Thus Kant also emphasizes that we quite “rightly” (394:2) esteem these qualities. Nonetheless, this cannot change the fact that, for the reasons already given, they are still only conditionally good.

2.2.2 The Conditioned Worth of Gifts of Fortune It is important to recognize that Kant’s argument is not limited to the ‘talents of the mind’ and the ‘qualities of temperament,’ but applies equally to the ‘gifts of fortune’ (power, wealth, honor, health, satisfaction, selfcontentment in life).13 Kant unambiguously says that the situation is precisely “the same” (393:14; o.e.) with the gifts of fortune as it is with the gifts of nature. Thus the will makes ‘use’ of these too. This is immediately obvious with regard to power as a gift of fortune, for example, for this can clearly be exercised by a good as well as an evil will. It seems less obvious with regard to happiness (as a whole). Kant himself says that we seek happiness naturally and for its own sake (cf. 415, for example). But why should we not regard happiness as something of unlimited worth? Kant offers two arguments against this position: 1. Nothing guarantees that happiness produces morally acceptable acts; on the contrary, happiness exerts a corrupting influence, according to Kant, insofar as it can provoke a certain “courage [Mut] and thereby often also arrogance [Übermut]” (393:16).14 Thus we can see, right at the beginning

12. Kant also makes this point admirably clear in Anth:254. There he refers to the phlegmatic quality of character as a “gift of nature” and even speaks approvingly of a “fortunate kind of phlegm (in the moral sense).” An individual who enjoys this par ticu lar gift of nature cannot “indeed on that account yet be called wise, but he is certainly favored by nature, so that it is easier for him than for others to become such” (o.e.). 13. These ‘gifts of fortune’ must be understood as “the elements which belong to the concept of happiness” (418:5; o.e.). Thus happiness does not belong among the gifts of fortune, but all the aforementioned gifts of fortune together are components of happiness. Cf. also 395:8– 9, where Kant mentions the “preservation, . . . welfare—in a word, . . . happiness” of the human being, along with other aspects of the concept of happiness. 14. The word Übermut, in Kant’s time, could often signify ‘hubris’ or extreme ‘presumption.’

Section I of the Groundwork

41

of GMS I, how Kant alludes to anthropological considerations: If human beings acquire and enjoy certain gifts of fortune (if they become richer, more powerful, or more renowned than their fellow human beings), they have an innate tendency to believe that they have also deserved this happiness or good fortune, and this presumptuous self-deception often leads them into unwise and even immoral course of action. And since it is ‘the same’ with the gifts of fortune as it is with the gifts of nature, we may also expect these gifts of fortune to exercise such a corrupting influence. 2. If happiness could claim complete worth in itself and under all circumstances, then we would inevitably react to it with an expression of unreserved esteem. But Kant doubts that we do. In the context of our ‘common moral rational cognition’ Kant claims that “a rational impartial spectator can never take satisfaction even in the sight of the uninterrupted welfare of a being, if it is adorned with no trait of a pure and good will” (393:19). Once again Kant encourages us to analyze our evaluative responses. The point is not (as a eudaimonistic consequentialist would see things, for example) that we disapprove of the happiness of a particular individual because it has been acquired at the expense of a greater happiness on the part of a greater number of individuals. We disapprove of it, according to Kant, because it is the happiness of a bad person. (It is already evident that the evaluative approach of classical utilitarianism is wholly incompatible with Kant’s approach, for as far as hedonistic utilitarianism is concerned, it is happiness alone, or the greatest happiness of the greatest number, that is good; but in that case the good will cannot be such a good).

2.2.3 The Efficacy and Effects of the Good Will It is a standard view of Kant’s moral philosophy that it is an intentionalistdeontological moral theory that is primarily concerned with deontological concepts, and thus particularly with the concept of duty itself and with establishing that intentions are decisive in the moral context. This kind of ethical theory is then rigorously distinguished, above all, from utilitarianism and consequentialism (which is also sometimes described as a teleological moral theory). If we ignore the fact that these concepts and typologies are actually anything but clear-cut, for our purposes we can initially describe

42

Immanuel Kant’s Groundwork for the Metaphysics of Morals

the basic idea of consequentialism as the claim that actions and types of action are to be evaluated in terms of the results (consequences) that they produce. The moral quality of an action or type of action is thus determined by the value of the ensuing results: if the results are good, then the action or type of action is good; if they are bad, then the action or kind of action is bad. Consequentialism is thus also a form of value ethics because it obviously makes sense to speak of ‘good’ or ‘bad’ results only if there are values and value criteria that allow us to classify par ticular results as ‘good’ or ‘bad.’ But this does not necessarily mean that every form of value ethics is consequentialist (in fact, as we shall see, Kant’s moral theory is a form of value ethics). If we go beyond this basic idea, it does not make much sense to talk about ‘consequentialism’ as a single position. In the fi rst place, the notion of ‘results’ or ‘consequences,’ and with it the associated distinction between an action and a type of action, can be interpreted in a number of ways; in the second place, consequentialism can assume an intentionalist form; and in the third place, the concept of the value (of a consequence) is also susceptible to different interpretations. It is not our task here to examine the various forms of consequentialism or to assess their relative strength or weakness. We simply have to clarify Kant’s position with regard to the basic idea of consequentialism. It is essential to do this because Kant repeatedly emphasizes that with actions performed out of duty the ends or purposes being pursued in any given case are not relevant. This thesis is crucial to Kant’s derivation of the categorical imperative itself. Our only question is to identify precisely what this thesis means and precisely where and how Kant attempts to justify it. In this connection it will be helpful to examine Kant’s critique of consequentialism, or how it might be developed (even if Kant could not have known the classic form of utilitarianism or consequentialism that finds expression in Jeremy Bentham, for example). It has been claimed that Kant’s critique of consequentialism is presented in the third paragraph of GMS I (394:13–31). In order to substantiate this, it is necessary to make certain distinctions that affect how consequentialism is to be understood. We shall then have to return, and return repeatedly, to the question whether Kant does criticize consequentialism, and, if he does, precisely how he does so.

Section I of the Groundwork

43

Willing and Wishing We can begin our analysis of the third paragraph with a point that is easy to clarify and that Kant therefore simply indicates in brackets: namely, that the good will is not to be understood as “a mere wish” but as “the summoning up of all the means insofar as they are in our control” (394:23 ). What Kant is trying to emphasize, quite clearly, is that the good will is good precisely through willing what is morally demanded: “good . . . only through its willing, i.e., good in itself ” (394:15).15 But this is not to be understood, “to be sure” (394:23; o.e.), as if we might somehow will the good but not the action that is required or the means that belong to that action. For to will the morally good means to will the realization of the relevant action and to will the adoption of the possible means that are necessary for its realization (and thus for the realization of the good). Kant does sometimes suggest that a good will can be combined with a lack of virtue (lack of moral strength), so that although a person might will what is good, he or she could still fail to carry it out.16 But here he seems to be suggesting that anyone who claims to possess a good will without acting accordingly, although the corresponding action is indeed possible, is simply pretending something to oneself or at least to others. Anyone who genuinely wills the good cannot help but will the action that is required as well, for that person already wills them; the willing of the good contains or implies the willing of the appropriate action (and the pertinent means). Anyone who claims that she or he has the good will to help rescue a person in a situation of extreme danger but is not prepared to run any significant risk (personal injury for example) or incur any disadvantage (such as missing an important appointment) cannot be said to have a good will. But it is also important to recognize that the ‘summoning up of all the means,’ where this does transpire, is no guarantee that a person actually has a good will (since, of course, we never know whether someone is indeed acting from a good will rather than from some other reason). On the other hand, it is certainly possible to determine whether someone is acting in conformity with duty.

15. In fact this is the only positive thing we learn about the good will in the fi rst three paragraphs of GMS I. 16. Cf. MM II:408.

44

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Actual and Accomplished Consequences of Action It is possible to hold the consequentialist position that the essential thing in evaluating an action is whether the intended ends of the action (the aims) were realized in fact; on this view, an action would be good if the aim being pursued was actualized and the action in question was thus successful. But one can also hold the thesis that what is essential in evaluating actions is the consequences that are actually produced (on this view, consequences that are quite unintended are also taken into consideration). Kant rejects both these versions of consequentialism in the third paragraph. In the light of Kant’s argument they seem to be manifestly false, although in fact they are still defended in ethics today. Let us fi rst consider the context of this discussion. Kant has argued that the gifts of nature and the gifts of fortune have no unconditioned moral worth. If moral worth and our corresponding moral evaluation were indeed bound to these things, then human beings who were favored by nature with only few or only poor ‘talents of the mind’ or ‘qualities of temperament’ would never be able to receive any moral recognition on our part. It is much the same as far as the gifts of fortune are concerned. It is quite true that power, wealth, and honor—unlike the gifts of nature—may be won or acquired. But whether they are actually acquired depends first, once again, on gifts of nature over which one has no control; and, second, on the acquisition of these goods, which is conditioned by social and other contingent factors over which we also have little or no  control. Human beings who enjoyed only few or only poor gifts of nature or of fortune would thus be excluded from the realm of moral action and moral recognition or at least would fi nd it extremely difficult to participate in it. Kant emphatically rejects this idea. Whether we are able to act morally, he says, cannot depend on the casual distribution of social goods or advantages or on the capricious favor of nature: “Even if through the peculiar disfavor of fate, or through the meager endowment of a  stepmotherly nature, this [good] will were entirely lacking in the resources to carry out its aim, if with its greatest effort nothing of it were accomplished, and only the good will were left over (to be sure, not a mere wish, but as the summoning up of all the means insofar as they are in  our control): then it would shine like a jewel for itself, as something

Section I of the Groundwork

45

that  has its full worth in itself. Utility or fruitlessness can neither add to nor subtract anything from this worth” (394:18; o.e.). As was already evident from his starting point in ‘common rational moral cognition,’ Kant therefore endorses a strict moral principle of equality that is supposed to guarantee justice: all human beings, in their capacity as free and rational beings, are able to act morally. Later, Kant will argue that this is why human beings possess dignity and an absolute moral worth (and not because of any gifts of nature or fortune they may happen to enjoy). But in light of these considerations, both variants of consequentialism that we have outlined may be refuted. If someone strives to perform a particular action in conformity with duty for its own sake and with the ‘greatest effort’ does everything in her power to accomplish it, but fails to do so on under circumstances for which she cannot be held responsible, she deserves esteem rather than censure. Thus whether an individual actually succeeds in realizing her intended purposes is irrelevant to the question of moral worth or esteem as long as it is clear that she has tried all she possibly can to successfully realize them (and has not simply contented herself with mere ‘wishes’). This argument applies even more obviously to the actual but unintended results of an action: the consequences (of an action) that we could not foresee or for which we are not responsible cannot play any role with regard to the question of what is morally worthy. If they did, there could be no responsible subject of moral action, and the concept of morality then would lose its meaning.

Intended Consequences of Action There is a third version of consequentialism that merits serious consideration, namely, the consequentialism that focuses not on individual actions but on types of action (or the rules governing them) and that also acknowledges the moral value of intentions. We can describe this position as intentionalist rule consequentialism. On this view we may describe particular ends as ‘good,’ but the moral evaluation of human actions does not depend on whether they succeed in actually realizing these ends or on what actual (unintended) consequences the actions have. What is morally evaluated is only the intentions (and willing rather than wishing, as in Kant), and the intentions are good if their object is good, that is, if good

46

Immanuel Kant’s Groundwork for the Metaphysics of Morals

ends are pursued in accordance with our conscience and our best knowledge of the situation. What have we learned from Kant’s third paragraph so far? Neither the “disfavor of fate” (gifts of fortune) nor the ‘meager endowment of a stepmotherly nature’ (gifts of nature) and thus neither the actually accomplished nor the factually resulting consequences of action should exercise any influence on how we evaluate human action (or human beings). The question at issue is this: does Kant’s argument in the third paragraph also apply to intentionalist rule consequentialism? The whole paragraph consists of only four sentences. The last three sentences tell us that the achieved and the factual consequences of action must be considered irrelevant as far as moral evaluation is concerned. Thus the sentence “Utility or fruitlessness can neither add to nor subtract anything from this worth” (394:26) must also be read in this light rather than, for example, as an indication that Kant is here rejecting the idea of taking the consequences of action as a criterion. (The reference to ‘utility’ is thus a critique of the thesis that ‘useful’ actions alone can be considered morally good. It is not a critique of the form of intentionalist rule consequentialism in which the willing of a good, that is, useful, type of action is good in itself, even if the corresponding action is not realized.) And the first part of the first sentence also says the same: “The good will is good not through what it effects or accomplishes, not through its effi cacy for attaining any intended end, but only through its willing, i.e. good in itself ” (394:13; o.e.). Here too we are talking about a will whose moral quality does not depend on its ‘efficacy’ (which can be diminished by the ‘disfavor of fate’ or a ‘meager endowment of stepmotherly nature’). That the will is not to be evaluated in terms of what it ‘effects or accomplishes’ is thus, again, a critique not of intentionalist rule consequentialism but of the other versions of consequentialism that we have mentioned. It is only the second part of the first sentence that may suggest a critique of this version of consequentialism: “The good will is to be estimated far higher than anything that could be brought about by it in favor of any inclination, or indeed, if you prefer, of the sum of all inclinations” (394:15). The satisfaction of the ‘sum of all inclinations’ is happiness.17 Now it is quite true that consequentialist 17. Cf., for example, 399:7–13.

Section I of the Groundwork

47

positions typically place a very high value on happiness. But in the fi rst place, Kant has already criticized the appeal to happiness (in the first paragraph); and in the second place, this does not itself yield an argument against intentionalist rule consequentialism. The third paragraph, therefore, does not contain a critique of intentionalist rule consequentialism. This is hardly surprising, for we have not yet heard anything about intentions in the sense of acting out of duty. Kant still has to show that the good will consists in the pursuit of morality for its own sake. But this becomes clear only when we have made the transition to the concept of duty in the eighth paragraph. Before we proceed to examine this in detail, we must first briefly consider the teleological argument that Kant presents here.

2.2.4 The Teleological Argument In the first few paragraphs of the text Kant has already endeavored to show that happiness is not an unlimited or unreservedly valuable good. Now he tells us that happiness is not the “real end” (395:10) of human nature. Kant’s teleological argument seems somewhat out of place here. In the first place, it presupposes certain fundamental teleological reflections without justifying or even discussing them. In the second place, this argument, as Kant himself concedes (394:32–36), is not one that can be derived from an analysis of ‘common rational moral cognition’ but is presented in order to counter certain doubts relating to and involved in this ‘common rational moral cognition’ itself. In the third place, in contrast to Kant’s other arguments in this context, it is not particularly convincing and also plays no further role in the rest of GMS. In any event, the argument in question can be summarized as follows: 1. Organized living beings possess ‘instruments’ (i.e., organs) with their own distinctive ends, which they fulfi ll in optimal fashion. 2. Organized living beings are not only directed toward the ends of these organs (qua partial functions of the organism as a whole), but their existence also has an overall end. 3. There are organized living beings that possess reason and a will, that is, reason as a practical faculty.

48

Immanuel Kant’s Groundwork for the Metaphysics of Morals

4. Reason does not fulfill the function of securing happiness in optimal fashion (it is instinct that would fulfill this function in optimal fashion). 5. Happiness therefore is not the overall end of the existence of an organized living being that possesses reason as a practical faculty. 6. The function of reason as a practical faculty is to determine the will as good in itself.18 There are several obvious points at which this argument could be questioned. Premise 1 seems dubious if it is taken simply as an empirical claim. Kant himself merely says that we must “assume” this principle (395:5) and offers no support for it in GMS. Elsewhere, however, we learn that he understands it as a heuristic principle of natural teleology (or what he later calls ‘a regulative principle of reflective judgment’) whose aim is not to report observable facts but to guide our theorizing about the structure of organisms and our assignment of teleological functions to organs.19 In the context of our post-Darwinian biology, however, we may no longer have reason to accept it. Premise 2 is also a claim associated with Kant’s conception of an organized being or natural end (it is quite similar to Aristotle’s claim in his ‘function’ argument in Nicomachean Ethics, book 1, chapter 6). But its connection with premise 1 seems obscure. Even if it is true that all the organs of a living being fulfill their respective functions in optimal fashion, it does not follow that the “existence” (396:10) of such a living being also has an overall end, or a “real end of nature” (395:10).20 A purposively organized living being may well have organs that serve to maintain life and that fulfill their respective ends in optimal fashion, but this does not imply that, over and beyond these respective ends, all these ends serve and belong to some overall end. Furthermore, it may not be clear what an overall end of the existence of a living being is in the first place. Kant’s premise 4 that as far as “the project of happiness and the means of attaining it” (395:24) are concerned, reason is “weak and deceptive”

18. We can fi nd similar reflections in MM II:640 and I (19–20). 19. Cf. KU:376–383. 20. Kant speaks of the “aim [Absicht]” of their existence (396:10), but Absicht is clearly synonymous here with ‘end.’

Section I of the Groundwork

49

(395:20), whereas instinct seems much “more certainly” (396:16) suited to accomplish this end (and thus possesses the suitably adapted organs as required by premise 1), is also questionable on more than external grounds. It might even seem incompatible with the thesis presented in GMS II, namely, that the will can be “guided” by reason (396:16; cf. 395:21) even “on behalf of inclination” (413:35)— since “reason furnishes only the practical rules as to how the need of inclination is to be supplied” (413:36)— and the “satisfaction of all our needs” (396:15) is nothing other than happiness (cf. 405:8). For Kant himself, reason in its “practical use” (395:22) thus also at least has the “vocation” (396:20) of serving the “preservation” and “welfare” (395:8– 9) of human beings. But then even in the teleological argument presented in GMS I, it seems that Kant does not entirely wish to deny this. For he writes that reason, regarded as an instrument for procuring the satisfaction of our needs, is not “suffi ciently effective” (396:14; o.e.); that is why, as we have already noted, reason is capable of guiding the faculty of desire only in a “weak and deceptive” (395:20) rather than ‘safe’ manner. And if it is the “true” (396:20) and “highest” vocation (396:32) of reason to produce a good will, then there must also be a false and subordinate vocation, which would consist precisely in guiding the faculty of desire. Thus Kant seems caught in the following dilemma: either the vocation of reason is not to guide the will with regard to the satisfaction of needs, which contradicts his analysis of reason as a practical faculty in GMS II; or, in addition to the production of a good will, reason also fulfills the function, if in a ‘weak and deceptive’ manner, of guiding “the will safely in regard to its objects and the satisfaction of all our needs” (396:14), which contradicts premise 1 of his teleological argument here, according to which the ‘instruments’ fulfill their own function in optimal fashion. A possible way out of this dilemma would be to understand premise 4 as an endorsement of Rousseau’s idea that the development of our rational faculties does not directly make us happy, but that our happiness would be better cared for if, as with other animals, it were left to natural instinct. This may be for many of us a paradoxical claim, as Rousseau surely intended it to be. But there is no inconsistency in Kant’s maintaining that our happiness is an end of reason and morality while also holding that the development of rational faculties in a creature is not the best way for nature to promote this end. Thus on a more sympathetic reading Kant’s

50

Immanuel Kant’s Groundwork for the Metaphysics of Morals

argument could be understood as suggesting that since happiness cannot be the natural end of the faculty of rational volition, and since we have already seen that a good will is something we regard as an unlimited good, it is a plausible hypothesis that the true natural function of practical reason is to enable us to have a good will. But even if we could safely claim that it is not the function of reason to promote the satisfaction of our needs, we should still have to ask what its genuine function is. The mere fact that human beings possess “reason as a practical faculty” (396:18) by no means implies that it is the ‘true’ vocation of this faculty to produce a “will good in itself ” (396:21). It is therefore clear that the inference from premise 5 to premise 6 21 would be formally valid (and perhaps even sound) only if we supply the additional premise that practical reason can have only two conceivable vocations: either that of creating happiness or that of producing a will good in itself. Only then can we move from the (alleged) fact that the first vocation (the creation of happiness) does not belong to reason to the necessity of recognizing the other one (the production of a will good in itself). But even if Kant did tacitly presuppose this additional premise, he would still owe us an argument to support it. Kant has therefore clearly failed to demonstrate the purported conclusion of his argument in these paragraphs (the purpose of which, we should note, was not to demonstrate the reality of the good will itself, for that is attempted only in GMS III).

2.3 Duty and Respect In contrast to the classical tradition, Kant doubts that we can ascribe unconditioned worth either to the virtues or to the idea of happiness as the ultimate end of all human striving (he also argues against the idea of human nature as setting or implying ends, as we shall see in more detail shortly). But this tradition itself emphasizes that only those who also pursue good ends can be described as virtuous. One is regarded as genuinely virtuous only if, at the same time, one possesses a good will (and the idea of happiness is also related back to the good will); and the virtues are also 21. Kant writes that the true vocation of reason “must therefore” (396:20; o.e.) be to produce a will good in itself.

Section I of the Groundwork

51

seen as all connected with, and even as identical with, the character of the individual. How does Kant’s perspective differ from this? As we have pointed out, Kant has no doubt about the (limited) worth of virtues such as courage and self-control or of the idea of happiness and even concedes that the ‘inner worth of the person’ is also generally bound up with them. Nonetheless, according to Kant, the good will can be regarded as something that is in principle wholly independent of (or even in contradiction with) any specific gifts of nature or fortune precisely because such gifts may always be misused. The good will is not something that springs spontaneously or unreflectively from the feelings or predispositions of a sound and healthily constituted person. It must rather be understood as the will of a reflective subject that must struggle with its desires in order to act rightly in a moral sense. But this will become really clear only once we have examined the concept of duty.

2.3.1 From the Concept of the Good Will to the Concept of Duty Kant says that he wishes to “develop” (397:6) the concept of the good will. In order to accomplish this, we must “put before ourselves the concept of duty, which contains that of a good will, although under certain subjective limitations and hindrances” (397:6). To begin with, one should beware of a possible grammatical misunderstanding here: Kant does not mean to say that these “subjective limitations and hindrances” determine the way in which the concept of duty ‘contains’ that of the good will. On the contrary, they qualify the good will itself, or the concept of such a will. The passage must therefore be read as follows: ‘If we wish to put before ourselves the concept of duty, which contains [the concept of] a good will, although [the concept of a good will that stands] under certain subjective limitations and hindrances.’ By ‘subjective limitations and hindrances’ Kant is undoubtedly referring to our various natural (or, as Kant would also say, ‘empirical’ or ‘sensuous’) inclinations, desires, or interests, which indeed do not necessarily obstruct the demands of morality but certainly can and often do obstruct them. Later (in GMS II) Kant will say that the moral law is an imperative for us human beings only because, as creatures that are both sensuous and rational, we do not always act as

52

Immanuel Kant’s Groundwork for the Metaphysics of Morals

reason commands: “But if reason for itself alone does not sufficiently determine the will, if the will is still subject to subjective conditions (to certain incentives) which do not always agree with the objective conditions, in a word, if the will is not in itself fully in accord with reason (as it actually is in human beings), then the actions which are objectively recognized as necessary are subjectively contingent, and the determination of such a will, in accord with objective laws, is necessitation” (412:35). It is precisely to these ‘subjective conditions (certain incentives)’ that Kant is referring when he speaks about “subjective limitations and hindrances” in the passage where he moves on to the concept of duty.22 But in what sense does the concept of duty ‘contain’ the concept of the good will, and why does the former contain the latter rather than the other way around? The reason for this claim lies in Kant’s distinction between the concept of a perfectly rational being and that of an imperfectly rational being. This distinction is extremely important for GMS as a whole, and we shall frequently and explicitly return to it in the course of the argument. For the moment, the following remarks should suffice: Perfectly rational beings are beings that are subject to no ‘subjective limitations or hindrances,’ or beings that are viewed from a perspective that entirely abstracts from the notion of limitation. Such beings therefore always act rationally and thus always morally or are considered as always acting rationally and morally: “If reason determines the will without exception, then the actions of such a being, which are recognized as objectively necessary, are also subjectively necessary, i.e. the will is a faculty of choosing only that which reason, independently of inclination, recognizes as practically necessary, i.e. as good” (412:30). Imperfectly rational beings, by contrast, are those that are subject to the aforementioned ‘subjective limitations and hindrances.’ Hence they do not always act rationally and do not always act morally. For a purely rational being, the moral law is both objectively and subjectively necessary, whereas for a sensuous22. In this sense cf. also Kant’s use of the word ‘limitation’ at 434:5; a little later he says explicitly that the moral ‘Ought’ would be identical to volition if “reason were practical in him [i.e., in the human being] without any hindrances” (449:18; o.e.). Kant frequently uses the word ‘hindrance’ in this sense. Cf. MM II:605– 606; KpV:79; MS:380; and KrV:B 29. Cf. also Kant’s concept of the ‘moral world’ in KrV (A 808– 809/B 836– 837).

Section I of the Groundwork

53

rational being it is only objectively necessary. Hence duty and necessitation can be said to exist only for imperfectly rational beings and not for perfectly rational beings: “The dependence of a will which is not absolutely good on the principle of autonomy (moral necessitation) is obligation. Thus the latter cannot be referred to a holy [perfectly rational] being. The objective necessity of an action from obligation is called duty” (439:30). That there is no such thing as duty for perfectly rational beings also means that the moral law is not an imperative for such beings, for the CI is precisely “the universal imperative of duty” (421:17; o.e.): “A perfectly good will would thus stand just as much under objective laws (of the good), but it would not be possible to represent it as necessitated by them to lawful actions, because of itself, in accordance with its subjective constitution, it can be determined only through the representation of the good. Hence for the divine will, and in general for a holy will, no imperatives are valid; the ought is out of place here, because the volition is of itself already necessarily in harmony with the law. Hence imperatives are only formulas expressing the relation of objective laws of volition in general to the subjective imperfection of the will of this or that rational being, e.g., to the human being” (414:1). It is very important to recognize that when Kant talks about perfectly rational beings, he is not merely thinking of beings that are not in fact or indeed could not be subject to any subjective limitations or hindrances (such as God or angels). Kant is going to suggest later that we think of sensuous-rational beings, which are capable of acting morally, as members of an intelligible world (the world of things in themselves). If human beings were only members of this world, then they would always act in accordance with morality: “As a mere member of the world of understanding, all my actions would be perfectly in accord with the principle of the autonomy of the pure will” (453:25). Thus human beings (in terms of the idea) are also perfectly rational beings insofar as we simply regard them as members of the intelligible world. Human beings are sensuous-rational beings, but from this perspective they are considered solely as rational beings: this is “the idea of freedom [that] makes me a member of an intelligible world, through which, if I were that alone, all my actions would always be in accord with the autonomy of the will” (454:6). Having clarified this background, we can now return to the question why and in what sense the concept of duty ‘contains’ the concept of the

54

Immanuel Kant’s Groundwork for the Metaphysics of Morals

will rather than vice versa. A perfectly rational being (one that has a good will) does not act out of duty, since for such a being there is no such thing as duty; it is only for a being whose will is not perfectly good that duty can exist at all. Given this distinction, the proposition ‘Every action done out of duty is the action of a good will’ is correct, whereas the proposition ‘Every action of a good will is an action done out of duty’ is false precisely because the action of a perfectly good will is not an action done out of duty. Again, we can correctly formulate the proposition ‘Every action that is morally good without limitation and is performed by an imperfect being is an action out of duty.’ It is only in this way that we can explain why Kant says in the passage under discussion here that the concept of duty ‘contains’ the concept of the will, ‘although under certain subjective limitations and hindrances.’ The qualification ‘although’ makes it clear that there is a good will that stands under ‘certain subjective limitations and hindrances,’ but that there is also the possibility of a good will that does not. Only if we acknowledge this distinction can we also understand why Kant specifically says in the opening sentence of GMS I that there “is nothing it is possible to think of anywhere in the world, or indeed anything at all outside it, that can be held to be good without limitation except a good will” (393:5; o.e.). With this allusion to the possibility of a good will ‘outside’ the sensuous-rational world Kant is referring to the perfectly good will of purely rational beings. The concept of a will that is good in itself is Kant’s initial answer to the question of what alone can possess moral worth (or, more precisely, what has no specifically moral worth). The concept of duty is the answer to the same question, but expressly with a view to the fact that the willing that possesses moral worth is the willing of an imperfect being.

2.3.2 The Three Propositions Regarding Duty Before we proceed to examine Kant’s three propositions regarding duty in close detail, we must first ask what these three propositions effectively amount to. Kant specifically refers to a “second” (399:35) and a “third” (400:17) proposition and even says that the third proposition should be regarded as a “consequence of the first two” (400:17). Thus there is no doubt

Section I of the Groundwork

55

that there must be a first proposition regarding duty. But what exactly is this first proposition? Kant never directly mentions a ‘fi rst proposition’ in the text itself. What possibilities are there? Well, he might be talking about the very first sentence, or proposition, of GMS I (393:5–7), for it seems plausible to think that perhaps Kant does not explicitly refer to the ‘first proposition’ as such because he could have already interpreted this very fi rst sentence as the proposition in question. But Kant might also be thinking of some specifi c proposition that is actually to be found in GMS I. Finally, Kant might be referring simply to an implicit proposition that would express the essential point of his reflections without it actually being formulated. The first two suggestions here can quickly be ruled out. The very first sentence of GMS I (393:5) cannot be the ‘fi rst proposition’ we are seeking because the transition to the concept of duty is accomplished only later (397:1–10). If Kant were really alluding to the very first sentence, then he would be in the absurd position of formulating the ‘third proposition’ regarding duty as a ‘consequence’ of two propositions, the first of which was developed before the analysis of duty itself. As for the second suggestion, it has to be recognized that no such ‘first proposition’ can in fact be identified in written form in the relevant part of the text (397:11–399:34). This is confi rmed once it becomes clear what the ‘fi rst proposition’ is actually about. Many readers of GMS think that the (implicit) ‘first proposition’ is something like this: “Any act having authentic moral worth must be done from duty,” since this seems to be the common theme of the discussion at 397–399. This is certainly one suggestion that needs to be considered; it is also close to—although not quite the same as—the suggestion we are about to offer. Kant claims that the ‘third proposition’—“Duty is the necessity of an action from respect for the law” (400:18)— is a ‘consequence of the fi rst two.’ Thus we would expect the essential conceptual components of this ‘third proposition’ (‘the necessity of an action’ and ‘respect for the law’) to be contained in ‘the fi rst two.’ This is indeed the case, though not explicitly. The concepts of ‘necessity’ and ‘respect,’ like that of ‘law,’ are employed only in the ‘third proposition.’ Thus not only are we confronted

56

Immanuel Kant’s Groundwork for the Metaphysics of Morals

with the problem that the fi rst proposition is not expressly pointed out as such, but also there is the difficulty that the content of the second proposition cannot directly be subsumed under either of these two essential components of the third. The key to this problem lies in Kant’s comprehensive formulation that “nothing is left over for the will that can determine it except the law as what is objective and subjectively pure respect for this practical law” (400:31). 23 If an action, objectively considered, is to be morally good, it must be determined by the moral law: an action that is contrary to duty cannot be good. But even if an action is objectively determined by the moral law (is in conformity with duty), it is not yet thereby (morally) good without limitation, for such an action must be done with subjective constraint, out of respect for the moral law. If an action is to be an action out of duty, therefore, it must be not only objectively good but also subjectively good. Or, as Kant says, it must, subjectively considered, be determined by respect for the law. It is our view, therefore, that the ‘third proposition’ is a ‘consequence of the fi rst two’ insofar as it specifically connects this objective aspect (the necessity of the moral law) with the subjective aspect (respect for the law). Almost all commentators hold that the term Folgerung (400:17) cannot be understood in a strictly deductive sense as a ‘consequence.’ However, on the basis of our interpretation, such a reading seems possible: 1. All actions from duty are actions from respect for the moral law. 2. Subjectively considered, all duties are actions from duty. Therefore: Subjectively considered, all duties are actions from respect for the moral law. 1. All necessary actions according to a formal principle a priori are actions according to the necessity of the moral law. 2. Objectively considered, all duties are necessary actions according to a formal principle a priori.

23. Cf. KpV:81: “The concept of duty, therefore, requires of the action objective accord with the law but requires of the maxim of the action subjective respect for the law, as the sole way of determining the will by the law.”

Section I of the Groundwork

57

Therefore: Objectively considered, all duties are actions according to the necessity of the moral law. The third proposition (P3 ) then combines these: 1. If all duties are actions from respect for the moral law and according to the necessity of the moral law, then duty is the necessity of an action from respect for the moral law. 2. All duties are actions from respect for the moral law and according to the necessity of the moral law. Therefore: Duty is the necessity of an action from respect for the moral law. Thus the ‘conclusion’ in a strict sense is P3.24 Before we proceed to examine the individual propositions in detail, it may be helpful here to provide an overview of the argument so far. Kant begins with the thesis that the good will alone can be good without limitation. But this thesis is not actually grounded as such, inasmuch as Kant shows us what cannot be ‘good without limitation.’ It is only in a second step that Kant develops this thesis more precisely, specifically in relation to the concept of a being whose will is not perfectly good. This immediately raises the question how an imperfect will must be determined if it is to be good. The answer is this: first, it must be objectively determined by the moral law; second, it must be subjectively determined by respect for this law. The subjective aspect of determination is captured in the ‘first proposition’: P1: An action done out of duty is an action done out of respect for the law. The objective principle of determination is expressed in the ‘second proposition’: 24. Another meaning of Folgerung goes back to the Latin corollarium; we leave this aside.

58

Immanuel Kant’s Groundwork for the Metaphysics of Morals

P2: An action out of duty follows a maxim required by the moral law and is thus an action that is necessarily required by the objective moral law. The CI must be the principle of a morally worthy action; this is the ‘objective’ condition for claiming that an action has genuine moral worth. The ‘subjective’ condition consists in respect for this moral law. It is only when respect as the determining ground indeed determines the will that the action becomes an action that is not merely in conformity with duty but is an action performed out of duty. Thus the ‘third proposition’ reads: P3: Duty is the necessity of an action out of respect for the law. The ‘consequence’ in question consists therefore simply in connecting both of these aspects: if duty, subjectively considered, consists in respect for the law, and if duty, objectively considered, consists in the necessity of an action that the law requires, then duty must consequently consist in ‘the necessity of an action out of respect for the law.’ We must now justify this reading and this way of relating the three propositions more precisely.

2.3.2.1 The First Proposition Regarding Duty as Acting out of Respect The “question” (397:13 ) Kant wishes to answer is this: what are actions done “from duty” (397:13 )? Kant’s answer is extremely terse and has led to a host of misunderstandings and substantive problems, but his basic thesis is actually simple and convincing: our greatest moral esteem is bestowed on actions that are performed purely from duty, that is, purely out of respect for the moral law, where no subjective inclinations or interests in favor of the action are present.25 This does not mean that it is only such actions that possess value from the moral point of view; and it certainly 25. Kant speaks repeatedly of the “worth of action” (397:4; 398:13–14; 398:26–27; 399:26; 399:35–36; 400:6; 401:3 ). But we have already seen in our interpretation of the third paragraph of GMS I that whether an action is realized in fact is ultimately irrelevant to the question of what is morally worthy; what is decisive is the willing of the good (which is not to be confused with a ‘mere wish’ for the good).

Section I of the Groundwork

59

does not mean that one must act in opposition to inclinations if one’s actions are to possess moral worth. It is true that Kant’s specific examples have particularly encouraged these misinterpretations. But these same examples, properly understood, clearly indicate what Kant is trying to show. We must differentiate between various distinctions and questions here. Kant immediately excludes actions contrary to duty as possible candidates for actions from duty. The question thus concerns actions that are in conformity with duty, and here Kant also immediately excludes those in which some indirect inclination is involved (as in the shopkeeper example). There are two other possible combinations, roughly speaking, to be considered here: actions in conformity with duty that are performed purely out of duty and actions in conformity with duty that are performed out of duty and out of immediate inclination. Thus the decisive questions are these: Does Kant think that there can be a connection of the two motives ‘out of duty’ (from respect) and ‘out of immediate inclination,’ so that an action can be determined at one and the same time out of duty and out of immediate inclination? Or does he think that an action out of duty is an action performed purely out of duty without any accompanying inclination? Or does Kant even perhaps think that an action done from duty must be performed in opposition to an immediate or indirect inclination?

Actions Contrary to Duty Kant begins with a preliminary observation that is anything but selfevident. He says that he will “pass over all actions that are already recognized as contrary to duty, even though they might be useful for this or that aim; for with them the question cannot arise at all whether they might be done from duty” (397:11). But precisely who is it that recognizes that the action is contrary to duty? The agent who recognizes that an action is contrary to duty but nonetheless performs it, is obviously not acting from duty since the action ‘confl icts’ with duty. But how would we judge a case where one to the best of one’s knowledge and in full accordance with one’s conscience thinks that a par tic u lar action is in conformity with duty (is thus a matter of duty) and then acts out of respect for the supposed objective necessity of this duty, although the action in fact is objectively contrary to duty? The answer can only be this: such an action is not an action done from duty, for actions done from duty must

60

Immanuel Kant’s Groundwork for the Metaphysics of Morals

indeed be determined subjectively by respect for the moral law; objectively, however, they must be determined by the law itself; so an action done from duty has two components, and it cannot be reduced simply to the subjective component.26 An agent who is objectively mistaken regarding the determination (derivation) of a particular duty or even regarding the identification of the moral law itself may indeed be determined subjectively by the feeling of respect, but the action is still objectively contrary to duty and thus cannot be described as an action done from duty. Kant claims in this connection that we certainly need “no well-informed shrewdness” (403:19) in order to determine what our duty is. The CI is like a ‘compass’ with which we know very well “in all the cases that come before” us how to “distinguish what is good, what is evil, what conforms to duty or is contrary to duty” (404:2). Given this premise, actions that are objectively contrary to duty obviously cannot be actions done from duty. We may well doubt whether, as Kant appears to claim here, our duty can really be so easily and securely (objectively) determined ‘in all the cases that come before’ us. But in any case Kant has identified a fundamental condition that an action done from duty must fulfill: the action must be in conformity with duty.

Actions in Conformity with Duty on the Basis of Indirect Inclination Just as he had with actions contrary to duty, Kant also decides to “set aside” (397:15) actions in conformity with duty where an indirect inclination is involved. These are actions that “are actually in conformity with duty, for which, however, human beings have immediately no inclination, but nevertheless perform them because they are driven to it through another inclination” (397:15). An inclination is immediate if it is expressly directed toward the action in question. Thus a merchant would presumably have an immediate inclination to treat his customers honestly if he were well acquainted with all of them and, as Kant says in his third example, were the sort of person who could take delight “in the contentment of others” (398:11). Now Kant construes the example in such a way that we cannot presume that the merchant does have such immediate inclinations (this “is not to be assumed”

26. There are nonetheless a number of unambiguous contexts where the formula ‘from duty’ means exactly the same as the formula ‘from respect’; this is repeatedly the case in the context of Kant’s specific examples.

Section I of the Groundwork

61

here; 397:30). We know that it is in the long-term interest of the merchant to treat his customers honestly, and we thus presume that he has this reason for treating them honestly and does not need to constrain himself to do so by the thought of duty. The merchant has an interest in retaining (an inclination to retain) his customers, in not being prosecuted, etc. One means of satisfying this interest is to be honest in his dealings. Thus his honesty results “neither from duty nor from immediate inclination” (397:31), but from an indirect interest in this honesty. Actions done from duty cannot therefore be actions in conformity with duty that result from an indirect inclination.27

Moral Motivation It is easy to see what Kant is trying to show with his shopkeeper example, but his other examples are much more difficult in this regard. The difficulty springs, among other things, from the fact that Kant introduces them in the context of the shopkeeper example as follows: “For there it is easy to distinguish whether the action in conformity with duty is done from duty or from a self-seeking aim. It is much harder to notice this difference where the action is in conformity with duty and the subject yet has beside this an immediate inclination to it” (397:17). When Kant says ‘there,’ he is referring to the shopkeeper example, where someone has indeed no immediate inclination but does have “another” (397:17) inclination, namely, an indirect inclination, to perform the action; this indirect inclination is then also described as a “self-seeking” (397:18) or “self-serving” (397:32) aim.28 The “question” (397:13 ) is whether a given action is done from duty. In the shopkeeper example the objective component we have mentioned is fulfilled: the action is in conformity with duty. With regard to the subjective component, the question is then

27. In his shopkeeper example (397:21–32), Kant speaks only of ‘immediate inclination,’ but thus it seems fair to speak of ‘indirect’ inclinations as well. 28. It is clear that the shopkeeper example refers back the two following sentences: ‘I also set aside the actions’ and ‘For there it is easy to distinguish.’ It is true that between these two sentences and the shopkeeper example we also fi nd the sentence ‘It is much harder to notice the difference.’ But it is only the following examples that refer back to this, which is why the fi rst of these examples begins with the qualifying phrase “By contrast” (397:33 ). We shall not examine these examples individually in detail, but we shall return to them repeatedly from various perspectives.

62

Immanuel Kant’s Groundwork for the Metaphysics of Morals

‘whether the action in conformity with duty is done from duty or from a self-seeking aim.’ The example is supposed to show that in the case of the shopkeeper it is not hard to identify the difference and decide the question. As things stand, we are to assume that the shopkeeper acts not from respect for the moral principle of honesty but from his own interest. But in another case things are clearly ‘much harder’ to decide. The question is only this: in what case, exactly? What does ‘this difference’ refer to, what is this ‘difference,’ and what precisely does it mean to say that the action is in conformity with duty and the subject also has an immediate inclination to perform this action? Several interpretations have been proposed in this regard, but only one of them is truly convincing. Before we examine these interpretations, we should briefly recall the context in which the examples are presented. Kant is trying to analyze what actions done from duty properly are. But the analysis also serves a broader purpose. Kant employs this analysis of moral motivation (understood as constraint from respect for law) in order to derive the CI (402). He must therefore construct the examples in such a way that the decisive point appears as strongly and evidently as possible: that actions done from duty are something different from actions that are accomplished on the basis of an interest, and that the moral law also therefore abstracts from all ‘matter’ (as Kant puts it). We misunderstand Kant’s first proposition regarding duty if we fail to recognize the specifically constructed character of Kant’s examples. In the situations of action that are described here we are not talking about particularly ordinary actions but about extraordinary and perhaps even extreme cases of moral action. Kant begins his analysis of duty by indicating that he wants to distinguish the concept of the good will from the concept of duty and, at the same time, by acknowledging the ‘subjective limitations and hindrances,’ to “elevate it by contrast and let it shine forth all the more brightly” (397:9; o.e.). It is in this light that his examples must be interpreted. 1. The first interpretation: the ‘difference’ in question is the difference between an action done from duty and an action done from a self-seeking aim. This interpretation is based on the grammatical fact that Kant himself initially distinguishes an action in conformity with duty that is done ‘from duty’ from an action that is done from ‘a self-seeking aim’ and then with the demonstrative pronoun ‘this’ refers back to this distinction. The

Section I of the Groundwork

63

following examples (cases of suicide, benevolence, happiness, and love) would thus be concerned with the difference between action from duty and an action from a self-seeking aim. But although this interpretation is grammatically convincing, it cannot be accepted in substantive terms. Three arguments strongly speak against it: 1. None of the examples are actually concerned with the opposition between the motive of duty and self-seeking aims and purposes. The individual who is tempted to suicide is motivated by adversity and grief rather than by self-seeking inclination; the individual who was once a friend of humanity has to try to overcome not self-seeking aims but a sense of hopelessness and inability to participate in the fate of others that has been produced by misfortune and distress; and the individual afflicted with gout who ignores his prescribed diet is motivated by an inclination that contradicts his own interest. What these examples are designed to show, rather, is that duty itself ought to motivate us, even if there is no inclination (whether a self-seeking one or not) that motivates us, and that there is a special kind of moral worth that attaches to actions performed under these adverse conditions. 2. Kant writes that the action of the merchant was done “neither from duty nor from immediate inclination, but merely from a self-serving aim” (397:31; o.e.), and in the example of the ‘friend of humanity’ too we are expressly told that this person acts “without any other motive of vanity or utility to self ” (398:9; o.e.). It follows from this that actions done from immediate inclination are not necessarily ‘selfseeking’ or ‘self-serving.’ Indeed, Kant explicitly distinguishes selfseeking actions not only from actions done from duty but also from actions done from immediate inclination. Thus the only action that is specifically described here as self-seeking is one that springs from an indirect inclination; Kant does not say that all inclinations are self-seeking in character.29 29. It is rather problematic, however, that in the Critique of Practical Reason (theorem 2) Kant subsumes all nonmoral motives under the “principle of self-love or one’s own happiness” (KpV:22); cf. also Kant’s reference to “a crossroads” (400:12). The question is then whether we would not also, for example, have to subsume the benevolence of the friend of humanity under the principle of self-love.

64

Immanuel Kant’s Groundwork for the Metaphysics of Morals

3. If all immediate inclinations were self-seeking and could also immediately be recognized as such, then indeed it would not be ‘hard to notice’ the difference between this type of motive and the motive of acting from duty. It can only be ‘hard to notice’ the difference if the motive of duty and the motive of inclination somehow resemble each other (we shall return to this), and that again would be impossible if all inclinations were simply lumped together as self-seeking. 2. The second interpretation: The ‘difference’ in question is the difference between an action done from duty and an action done from immediate inclination in those cases where both inclination and duty (can) serve as incentives for the same action (in conformity with duty). In the light of this interpretation, it looks ‘hard’ to decide whether we are faced with an action done from duty or an action done from immediate inclination (given that duty and inclination are both present), and which of the two in any given case is the ‘real motive.’ On this interpretation, therefore, it is possible that a person can be motivated from duty to perform an action, even if immediate (nonmoral) inclinations in favor of the action are also present at the same time. This kind of approach readily suggests itself if we accept the premises of an empirical theory of action that understands an action to be the causal result of desires, wants, and passions. The motive of ‘acting from duty’ is then understood as one of the forces that move us and thereby produce human actions in a sort of parallelogram of psychological forces.30 This interpretation can also be supported from a grammatical point of view. Kant’s talk of the ‘distinction’ seems to presuppose that ‘the subject yet has beside this an immediate inclination to an action in conformity with duty.’ If it were necessary, for an act to result from duty, that no accompanying inclination were also present, then it would not be ‘difficult’ at all (on this interpretation) to determine whether an action results from 30. The fact that we have previously employed the rather empirical terminology of ‘motive’ and ‘moral motivation’ should not therefore encourage any misunderstanding here. For Kant, the moral motive is not one motive among several others that exerts or fails to exert a causal effect. It is rather a ground for action, one to which we relate in a free and rational manner ( just as in principle we rationally relate, or are capable of rationally relating, to the impulses that underlie our actions).

Section I of the Groundwork

65

duty, for we can already presuppose, with regard to the relevant action, that the subject yet has beside this an immediate inclination to it. This interpretation also appears to receive support from a well-known objection that has been raised against a third interpretation (the one that we endorse)—namely, that it inevitably leads to the implausible claim that the only thing that can possess moral worth is a will that acts purely from duty and is devoid of inclination. But this second interpretation cannot be substantively supported either and must, therefore, like the first, be rejected in spite of the philological evidence already mentioned. The three following arguments speak strongly against it (and thereby also strongly in favor of our own third interpretation: that in fact only actions from pure respect possess an unlimited moral worth, and that ‘the difference is much harder to notice’ insofar as actions done from immediate inclination so closely resemble those done from duty). 1. The second interpretation is challenged, in the first place, by the simple fact that all Kant’s examples are framed in such a way that the acting individuals here either do “not act from inclination” (398:6; 398:19; 399:1; 399:25; o.e.) or act “without any inclination” (398:26; o.e.). Kant wishes to emphasize that in all these cases we may assume that the human beings in question act not from inclination “but from duty” (398:6; 398:19; 399:2; 399:25; o.e.). They act, as he also says, “solely from duty” (398:26; o.e.).31 But in that case we may not assume that human beings who act from duty also act at the same time from other motives (from immediate inclinations). Kant chooses his examples to present us specifically with individuals and actions where it is not to be assumed that they have an immediate inclination to perform the corresponding action. Thus he presupposes, for example (398:20–399:2),32 that there is a human being who is ‘clouded over 31. Cf. Kant’s use of the word ‘solely’ [lediglich] at 403:31 and 407:3. In the context of the third proposition regarding duty Kant thus writes accordingly: “Now an action from duty is supposed entirely to abstract from the influence of inclination, and with it every object of the will” (400:29; o.e.); in this connection, cf. also Kant’s retrospective remark at 440:5–7. 32. Cf. the sentence beginning “Thus suppose the mind of that same friend of humanity” (398:20).

66

Immanuel Kant’s Groundwork for the Metaphysics of Morals

with grief’ and afflicted with a ‘deadly insensibility,’ who could experience ‘little sympathy’ and possesses ‘a temperament cold and indifferent to the sufferings of others,’ a human being whom ‘no inclination any longer stimulates’ to act benevolently. Such a human being acts not from immediate inclination but merely from duty.33 If the ‘difference’ we are talking about is actually the difference between an action done ‘from duty’ and an action done ‘from inclination’ (and formulated in such a general way, this is certainly correct), then the action we are looking for cannot result from both duty and inclination (when they are present together as incentives). Perhaps we might think that it is possible to perform an action from duty that is accompanied by an existing inclination that in fact exerts no influence. But it is rather difficult to imagine how, for example, one could love a human being (understanding love as an inclination in this connection) while helping that person in a particular situation from duty rather than from inclination, even though an inclination (love) is present. 2. This last point is part of an extensive discussion of what has sometimes been called the problem of ‘motivational overdetermination.’ If we opt for the second interpretation of the ‘difference’ at issue, then we must also ask what Kant would say about cases in which the action seems to be brought about through duty as well as through a nonmoral inclination. Is such an action even possible? Is such an action still an action done from duty? And what does it imply with regard to the moral goodness of the will? The concept of ‘motivational overdetermination’ already shows how complex such reflections turn out to be because this expression can mean different things: first, it may indicate actions that result from both duty and inclination, but

33. One should note that in the second example Kant is effectively speaking of two different persons: fi rst, of the friend of humanity whose benevolence is initially described (398:8–20), and who has then supposedly been ‘clouded by grief’ and so on (398:20–27), but then also of “this or that person” (398:27), i.e., of another person who, in contrast to the friend of humanity, is not originally attuned by nature to benevolence at all (“if nature has not really formed such a man into a friend of humanity”; 398:34). This second case is supposed to make the ‘difference’ under discussion even clearer (“Even more”; 398:27).

Section I of the Groundwork

67

where either motive on its own is or would be sufficient; second, it may indicate ‘compounded’ or dual-determined actions that also result from both duty and inclination, but where neither motive on its own is sufficient, and both are required to produce the action; or, third, it may indicate actions that also result from duty and inclination, but where one of the motives on its own is sufficient or insufficient as the case may be. If we ignore the third case for the moment, it seems clear that for Kant, neither overdetermined nor dual-determined actions can be regarded as actions done from duty. Dual-determined actions could never come about at all without the effective participation of inclinations, in which case they are precisely not actions done from duty, in the sense in which Kant here means the phrase ‘from duty.’ Actions done from duty are actions done from respect for the law. We must exclude the possibility that they would not come about if respect for the law were all that was present (and thus that respect alone would be an insufficient motive, which for Kant would simply mean that there is no moral motivation here at all). And as far as overdetermined actions are concerned, we have to say that the meaning of actions done from duty consists precisely in the fact that they result ‘from duty’ (instead of ‘from inclination’). In fact, Kant shows no interest in such a (potential) theory of ‘motivational overdetermination’ either in GMS or in his other writings.34 The

34. There is indeed a passage in Religion within the Boundaries of Mere Reason where Kant seems to concede that someone may be determined at once from duty and from inclination (cf. also TP:279 and KpV:72). In this text he describes the human “propensity to adulterate moral incentives with immoral ones” (R:29; o.e.). This “impurity (impuritas, improbitas) of the human heart consists in this, that the maxim . . . is not purely moral, i.e. it has not, as it should be [the case], adopted the law alone as its sufficient incentive, but, on the contrary, often (and perhaps always) needs still other incentives besides it in order to determine the power of choice for what duty requires; in other words, actions conforming to duty are not done purely from duty” (R:29–30; o.e., Kant’s emphasis omitted). Even if we disregard the fact that we are dealing with GMS here, rather than Kant’s ethical theory in its entirety, this passage itself clearly shows that, ‘as it should be,’ the law alone must serve as ‘sufficient incentive’ without inclinations having to play any additional role. Thus even if Kant would say here that an action can be determined from duty and from inclination, he would not thereby mean

68

Immanuel Kant’s Groundwork for the Metaphysics of Morals

whole passage in question (397–399) is evidently not concerned with a theory of psychological causality or with cases in which one and the same action may have more than one impulse or motivation. This is simply not the issue here. 3. If Kant wanted to say that a good will is present only where an action is done from duty, then he would say that we can have a good will only where there are no nonmoral motivations for doing our duty; and then perhaps he would even say that one acts from duty only when one has to overcome recalcitrant or opposing inclinations. From this point of view, anyone who strives to act from duty would also have to strive to act without any nonmoral motivations or even to have or develop inclinations that are contrary to duty (simply in order to overcome them); and this highly implausible assumption is, as we have seen, one reason that the second interpretation has so often been held.35 But it hardly needs saying that Kant naturally never claimed that we should have to hate our friends or cultivate any other inclinations that are contrary to duty simply to create an opportunity for resisting them (as if one should deliberately place a person dear to us in some danger in order to have the opportunity of demonstrating one’s commitment to duty by saving this person). On the contrary, Kant expressly says elsewhere that we actually have a duty to cultivate love, sympathy, and other feelings since they make it easier, and indeed first make it possible, for us to fulfill our duties (we shall say more about this shortly).36 that such actions represent the highest moral worth (although it should be noticed that he does not say that they have no moral worth). 35. This is also the thought behind Schiller’s famous lines from his Xenia: “Scruples of Conscience: Gladly do I serve my friends, / But more’s the pity I do it from inclination, / So I am often gnawed by the thought / That I am lacking in virtue. Decision: There is no alternative—you must seek to despise them / And to do with abhorrence what duty commands you to do” (Friedrich Schiller, Werke, Nationalausgabe, vol. 1 [Weimar, 1943], p. 357). 36. Cf. MS:402 and 456–457 and ED:337–338. In a brief rejoinder to Schiller’s treatise On Grace and Dignity Kant emphasizes that acting from duty by no means involves a “Carthusian frame of mind” (R:23, footnote) and calls on us to have “a heart joyous in the compliance with its duty” (R:24, footnote). In fact there is no substantive contra-

Section I of the Groundwork

69

3. The third interpretation: The ‘difference’ in question is the difference between an action conforming to duty that is done from duty and an action conforming to duty that is done from an immediate inclination to perform this action conforming to duty; accordingly, it is the distinction between moral esteem for actions conforming to duty that are done from duty and moral appreciation or recognition of actions conforming to duty that are done from immediate inclination. We may begin by clarifying once again the basic thought associated with the third interpretation here. Kant is asking precisely what it is that we value most highly in the moral sense, and to which we accord our greatest respect. His answer is: the willing of the morally good for the sake of the morally good itself and for no other reason.37 Thus it belongs to an action done from duty, first, that the action in question is objectively in conformity with duty; second, that the action in question is done for the sake of duty (that is, because it is in conformity with duty); and third (given the meaning of the second condition), that there is no further motive for performing the action in conformity with duty. In addition to the arguments we have already marshaled against the second interpretation, the following considerations also speak in favor of the third (our own) interpretation: 1. In a given action that is in conformity with duty it is ‘hard’ to see whether it is done from duty or from an immediate inclination because the motive of immediate inclination and the motive of duty may resemble each other. Kant explicitly says that reason must “instill a feeling of pleasure or satisfaction in the fulfilment of duty” (460:10). Respect therefore has an “analogy” (401:34) with inclination, and that diction between Kant and Schiller, in spite of the polemical Xenia we have cited, and Kant himself expressly says that he cannot recognize “any disagreement” in principle here (R:23, footnote). 37. Kant himself already uses this formulation in the Preface: “For as to what is morally good, it is not enough that it conform to the moral law, but it must also happen for the sake of this law” (390:4). In MM II:624 he writes: “Duty is not what I do for the sake of my own advantage, but what I do for the sake of the law.” Cf. the formulations Kant offers in KpV:72 (“for the sake of the law”) and, especially, KpV:81, where he says that “moral worth must be placed solely in this: that the action takes place from duty, that is, for the sake of the law alone” (o.e.).

70

Immanuel Kant’s Groundwork for the Metaphysics of Morals

is why it is difficult to distinguish it from other feelings and motives.38 Kant’s examples are meant to meet this difficulty. They are atypical examples that are constructed in order to resolve the difficulty by proposing cases where it seems we must exclude the presence of any inclination. These are cases where we have reason to suppose that there is no longer any danger of confusing the motive of duty with that of inclination. Here we may assume that respect rather than an inclination is in fact the motive that determines action, even though respect and certain inclinations (sympathy or love, for example) closely resemble one another. And if we recognize that the examples are indeed extreme cases, there is no need to ascribe any special normative character to them. With these examples, therefore, Kant is not describing how we ought to act. Thus when he says in one example (398:8–399:2) that an admirable friend of humanity, one of those who feel an ‘inner gratification in spreading joy around them,’ is less worthy of esteem than the cold and emotionally indifferent individual, and that the beneficence of the man (‘by temperament cold’) who acts from duty has “a far higher worth” (398:35) than that of the man (of ‘good-natured temperament’) who acts from inclination, this is certainly an appropriate description (a good example) of what it means to act from duty in the strict sense. But it is not to be understood in such a way as to suggest that one would have to place oneself in corresponding situations in order to be able to act morally at all, and it certainly does not imply that Kant would depreciate feelings such as love or sympathy in principle. There is no question about one thing: a beneficent action done from duty has a special kind of moral worth that an action done from natural sympathy does not. That does not mean, however, that Kant places little value on sympathy or would even encourage us to strive to perform actions from duty when they can also spring from sympathy. The resemblance (‘analogy’) between actions done from duty and actions done from immediate inclination also means that actions done from duty are by no means performed reluctantly or ‘against our will.’ We will to accomplish both actions (the resemblance results precisely from this and from the ‘pleasure’ connected with it). If we 38. This problem is also discussed in KpV (116–117).

Section I of the Groundwork

71

understand duty merely as a social expectation or perhaps as a form of coerced legal obligation, then the Kantian idea of an action done from duty is easily distorted or caricatured to suggest that a person who acts from duty is someone who fulfills his duty against her own will (because he is forced by the will of those around her or by legal coercion). But the very opposite is the case. For Kant, ‘duty’ refers to what we most rationally will to do for its own sake. It may well be that one accomplishes an action done from duty by overcoming strong natural inclinations (and it may then become clear, as in Kant’s examples, that this is an action done from duty). But an action done from duty is an action that as a rational being I myself will and also perform with ‘pleasure’ and ‘satisfaction.’ The point is simply— and this point is directed against a philosopher like Hume—that respect is not a natural feeling of desire that one does or does not have, and that comes on one passively, but is rather a “self-effected feeling” (401:21). It is an effect that the moral law itself exercises on the mind, something that springs from our own reason and freedom. Actions that result from respect are autonomous rather than products of an alien agency. This point also clearly emerges from Kant’s example, for with regard to the friend of humanity who has been ‘clouded over with his own grief,’ we read that “he tears himself out of this deadly insensibility and does the action without any inclination, solely from duty” (398:24; o.e.). In other words, this is something that he does himself. His action would be impossible if it were not also willed, albeit in a specific (i.e., rational) way. In this connection one might be tempted to challenge the third interpretation by recourse to the following argument: If an action done from duty fundamentally resembles an action done from immediate inclination, then we could never know whether an action done from duty occurs at all. The examples in GMS I then make no sense at all, for we cannot possibly know that an immediate inclination is not present. But there are two ways in which we can counter this argument: In the first place, we can point out, as Kant himself does, that we never can in fact be certain that the motive of duty alone has determined an action (407–408 and 419). But if this is a central thesis of Kant himself, then with these examples Kant cannot be concerned with providing sure and certain examples of actions

72

Immanuel Kant’s Groundwork for the Metaphysics of Morals

done from duty. It is necessary, therefore, to attend to the context of the examples, which is that Kant has now begun to analyze the concept of duty (397). He wishes to show what an action from duty specifically consists in, and for this purpose he stipulates that we are dealing with an action done from duty (something that he thinks in practice we could never know with certainty). Second, he uses this procedure in order to show how we, as possessors of common rational moral cognition, react to these examples. The examples are designed to clarify how and why we distinguish approval [Billigung] from esteem [Hochschätzung] in an ethical context. (The German word for ‘esteem’ etymologically means ‘high estimation.’) For this purpose, it is not relevant whether we can know whether the examples in question could actually occur. 2. Throughout GMS, I Kant relates his argument to our ‘common moral rational cognition.’ This cognition already distinguishes between the mere approval we accord to an action done from inclination and the moral esteem we accord to an action done from duty.39 We approve actions that are in conformity with duty, whether they occur on the basis of immediate or of mediate inclination, but we refuse to accord our highest moral evaluation to such acts, even if we certainly regard acts that are in conformity with duty and are done from immediate inclination— such as the beneficent activities of the friend of humanity— as truly “amiable . . . and worthy of honor” (398:14); they rightly deserve “praise and encouragement, but not esteem” (398:17). The difference (so ‘hard to notice’) between our ethical reactions (approval versus esteem) certainly goes back to the difference (also ‘hard to notice’) between an action in conformity with duty that is done from inclination and an action that is done from duty. But it is important to see that in terms of the strategy of GMS I, this distinction (action from duty versus action from inclination) is specifically (and philosophically) acknowledged because it is deeply rooted in ‘common rational moral cognition.’ And we must recognize, in contrast, that anyone endorsing a (more or less) Humean 39. For Kant’s references to the notions of ‘esteem’ and ‘estimation’ [Schätzung], cf. 394:2–16; 397:1; 397:4; 398:18; 403:28; 430:8; 435:24; 436:5; 437:30; and 442:31.

Section I of the Groundwork

73

picture of (moral) motivation would have to explain away this distinction (no easy task) that is already acknowledged in common rational moral cognition. In other words, the claim that respect for the moral law implies that there is something such as genuinely rationalmoral motivation (and not merely the ‘passions’ and ‘desires’ invoked by Hume) already looks highly plausible because we should otherwise be incapable of explaining why the distinction between actions done from duty and actions done from inclination (described in the examples) is so firmly rooted in common moral rational cognition. The burden of proof here lies not with Kant but with his philosophical opponents.40 3. The last point also explains why we must not suppose that Kant thinks that only actions done from duty are morally valuable in any way. There is no doubt that Kant repeatedly says that only an action from duty has moral worth. But what he writes, precisely speaking, is that only an action done from duty possesses “true moral worth” (398:14; o.e.) or “authentic moral worth” (398:27 and 399:26; o.e.). That is, the highest moral worth belongs in fact and exclusively to actions that alone spring from the moral motive of respect and nothing else. But this also implies that actions in conformity with duty, and particularly those that derive from immediate inclination, do not simply fall out of the context of moral judgment and reaction. But why do we show so much esteem for actions done purely from duty? Why are we not satisfied merely with actions that are in conformity with duty? The answer is initially very simple and is not susceptible to further analysis: because actions done from duty, exhibiting a good will, have a worth that is without limitation. The way we evaluate them is a reflex of the fact that such actions are morally worthy in themselves, and that all other candidates (gifts of nature, gifts of fortune, immediate inclinations toward actions in conformity with duty, actions that are in conformity with duty) have proved to be either less worthy or not worthy at all. In addition, Kant frequently emphasizes that only a purely moral motivation can guarantee a certain moral stability (reliability, consistency, regularity). It is quite 40. In this regard, cf. Kant’s remarks at 411:31–38.

74

Immanuel Kant’s Groundwork for the Metaphysics of Morals

true that nonmoral motives (mediate or immediate inclinations) may sometimes lead to actions in conformity with duty, but they do not always do so.41 4. In the last paragraph before we are presented with ‘the second proposition regarding duty’ Kant provides a brief discussion of love (399:27–34). He distinguishes ‘love as inclination,’ which he also calls ‘pathological’ love, from ‘beneficence solely from duty,’ which he identifies with ‘practical’ love.42 The decisive point is that love, as a natural (sensuous) feeling, is something that ‘cannot be commanded,’ whereas practical love is something that certainly can. In the ‘pathological’ sense, one either loves or does not love a particular human being. We cannot consciously produce this natural feeling or at least fi nd it exceedingly difficult to do so (as we easily understand when we recognize that we cannot fall in love on command). If only those whom we love from inclination had a right to our beneficence, it would hardly bode well for all those we do not love in this sense— which, of course, means most people. As free and rational beings, we can resolve to treat other human beings with benevolence and beneficence, as it is our duty to do (and this is what Kant calls ‘practical love’). When we do so, and do so without inclination and perhaps even in opposition to a ‘natural and invincible disinclination,’ then we do so from duty.43 If we do not do our duty, we must take responsibility for this. The responsibility in question, and the notion of accountability that goes with it, is possible only if the possibility of

41. Cf. 390:4– 8 and 411:3–7. 42. The fact that Kant speaks of ‘pathological’ love naturally does not mean that he is here contrasting a diseased or morbidly excessive love of some kind with what he has called ‘practical’ love. The term ‘pathological’ here simply means this is ‘natural’ love rather than love produced by the law itself (cf. also KrV:A534/B562 and A802/B830, where Kant uses ‘pathological’ to signify the influence of ‘sensuous impulses’). In this sense Kant also distinguishes between ‘practical interest’ and ‘pathological interest’ (413, footnote). Thus in MS Kant describes moral feeling as ‘pathological’ insofar as it “precedes the representation of the law” (MS:399), that is, insofar as it is not produced by or through the law. 43. Of course the disinclination is ‘invincible’ only in the sense that we cannot alter the fact that we have such an aversion; it can certainly be conquered in the sense that we can overcome it and act out of duty despite it.

Section I of the Groundwork

75

performing the action that is commanded ‘lies in the will and not in the propensity of feeling, in the principles of action and not in melting sympathy.’ If we confi ne our consideration of Kant’s moral philosophy solely to GMS, we can easily receive the impression that Kant has very little interest in the emotional side of moral action. But, in the first place, we have already seen that the motivation for acting from duty, and thus from respect, is itself a (positive) feeling or something accompanied by a feeling (even by a certain ‘pleasure’ in fulfilling our duty). What is decisive here, however, is that this feeling is not some natural feeling over which we can exercise no control and which for that reason cannot be commanded either. The feeling in question is produced by reason and the moral law that belongs to it. In the second place, Kant (or at least the later Kant) is well aware that there are “subjective conditions of receptiveness to the concept of duty” (MS:399). Kant describes these conditions as “natural predispositions of the mind” (ibid.) that one cannot command us to have. On the one hand, as predispositions, they are not something that can be commanded since they are themselves already presupposed if duty is to command us (we can only be commanded to cultivate them). On the other hand, there is no need for them to be commanded anyway since “every” (ibid.) human being as a moral being has them in any case. Benevolence, as love of humanity, also belongs among those predispositions that allow us to be “affected” (ibid.) by the moral law in the first place. If we apply this thought to Kant’s second example in GMS, we can say that ‘the friend of humanity’ whose mind was ‘clouded over with his own grief ’ and the other individual in whose heart ‘nature had put little sympathy at all’ are both acting from practical love. But we can also say that a ‘natural predisposition of the mind’ (MS), namely, love of humanity,44 as a self-effected rational 44. Here it should be emphasized once again that one cannot simply assemble different passages from the range of Kant’s ethical writings. One must of course attempt to read them as far as possible in relation to one another, but since Kant’s ethical ideas clearly evolved over time, one must exercise due caution in this regard. Moreover, in the Metaphysics of Morals Kant thematizes the phenomenon of philanthropy—‘love of humanity’—in two different contexts: on the one hand as a feeling, namely, as a

76

Immanuel Kant’s Groundwork for the Metaphysics of Morals

feeling, is what grounds the possibility of acting from duty in this way in the first place.

2.3.2.2 The Second Proposition Regarding Duty: The Objective Law “The second proposition is: an action from duty has its moral worth not in the aim that is supposed to be attained by it, but rather in the maxim in accordance with which it is resolved upon; thus that worth depends not on the actuality of the object of the action, but merely on the principle of the volition, in accordance with which the action is done, without regard to any object of the faculty of desire” (399:35). We argued earlier that the ‘second proposition’ regarding duty (P2) thematizes the objective dimension of the concept of duty, namely, the moral law. At first sight, this thesis does not look very convincing, for P2 seems to claim that the worth of an action from duty lies ‘in the maxim.’ Since a maxim is then specifically defined in a note a little further on as “the subjective principle of the volition” (400:34; o.e.), one might be tempted to think that P2 expresses the subjective dimension of the concept of duty.45 But this impression would be quite wrong. To begin with, we should briefly consider the concept of ‘maxim’ here; it appears repeatedly throughout GMS and is particularly relevant to the derivation of concrete duties later. In the footnote in question Kant says that a maxim is ‘the subjective principle of the volition.’ Now the subjectivity of maxims consists simply in the fact that they are rules in accordance with which (individual) subjects do in fact act, or will to act, but do not unconditionally have to act; maxims per se cannot claim universal validity. disposition toward morality (Section XII of the Introduction to the Doctrine of Virtue: MS:399ff.), and on the other hand as a rational maxim of benevolence (in the context of duties of love toward others: MS:448ff.; this corresponds to what Kant describes in GMS as ‘practical love’). It is not easy to bring these two passages into a perfectly coherent relationship to each other, nor is it easy to interpret the fi rst of them unproblematically in its own right (with regard to both the feeling of philanthropy itself and the feeling of respect, the treatment of which differs markedly from the analysis furnished in GMS and KpV). 45. This impression is strengthened by the sentence to which the note refers, which claims that the good will is determined by the ‘maxim.’

Section I of the Groundwork

77

Maxims can certainly coincide with objective rules— as they do when one makes an objective (for example, a moral) law into a maxim for one’s action—but in the first instance they are rules in accordance with which a subject undertakes to act: a maxim “contains the practical rule that reason determines in accord with the conditions of the subject (often its ignorance or also its inclinations)” (420:37). Thus although a maxim cannot claim any universal validity, it is, for the person who adopts it, nonetheless a principle and thus something objective. A maxim expresses what a person wills to do—that is why it is a principle of willing—if that person is in a particular situation, namely, to accomplish a particular action.46 To return to the discussion of P2, we should observe that it effectively consists of two parts, the second of which elucidates the content of the first.47 If, to begin with, we ignore the part of the sentence that recapitulates P1 and states what the worth of an action from duty does not consist in,48 and concentrate on the elucidatory part of the sentence, then what Kant is saying is this: ‘The worth of an action from duty depends on the principle of the volition’ (and in fact this ‘principle of the volition’ is specifically emphasized in the text). What then is this ‘principle of the volition’? After stating once again what the worth of such an action cannot consist in and then asking what it does consist in, Kant repeats his thesis: the worth lies “in the principle of the will” (400:9).49 Since Kant then says that the will is determined either through a “principle a priori” (400:11) or through an “incentive a posteriori” (400:11), and this latter is excluded by his thesis regarding the aims and effects of action (which has yet to be elucidated), the good will can be determined only through a ‘principle a priori,’ or, as Kant says, “through the formal principle in general of the 46. This does not, however, exclude the fact that maxims can be formulated in very specific ways, something to which we shall return in connection with the idea of universalizability. The subjectivity of maxims also involves a further aspect that we shall discuss later in the context of hypothetical imperatives, namely, that we do not simply have maxims but rather make them for ourselves. 47. The ‘thus’ at 399:37 is clearly to be read in an elucidatory or clarifying sense. 48. Since Kant returns to this issue in connection with P3, we shall discuss it then and consider what Kant says about it in the context of P2. But we should already note that this ‘negative’ aspect in the context of P2 is not part of the claim formulated in P2 itself; Kant expressly says that this is “clear from the preceding” (400:6). 49. The use of the words ‘volition’ [Wollen] and ‘will’ [Wille] makes no difference here.

78

Immanuel Kant’s Groundwork for the Metaphysics of Morals

volition” (400:13). In this way, the ‘principle of volition,’ the ‘principle of the will,’ the ‘principle a priori,’ and the ‘formal principle of volition’ are all identified with one another. But since such an a priori formal principle of the will is, in Kantian terminology, undoubtedly the moral law, P2 can be expressed briefly as follows: ‘The worth of an action from duty depends on the moral law.’ This is exactly what Kant repeats in the next paragraph: what determines the good will (the will that motivates an action from duty) is ‘objectively the law’ (and subjectively pure respect for this practical law’).50 Thus the fact that, according to P2, the worth of an action from duty lies ‘in the maxim’ should not cause any undue difficulty.51 It is quite true that a maxim is a subjective principle of volition. But this subjective principle can coincide with the objective principle (that is, with the moral law). It then certainly remains subjective insofar as a subject makes it the principle of his or her action; but it is the objective law that nonetheless determines the action in question. And Kant’s thesis, of course, is that an action, considered in its objective aspect, is an action done from duty only if the maxim underlying it is in fact the moral law.

2.3.2.3 The Third Proposition Regarding Duty In the third proposition itself (P3) there is basically nothing that has not already been said. It is simply a ‘consequence of the fi rst two’ in the sense that it connects the subjective and objective aspects in question. But in P3 we do apparently encounter concepts that have not been introduced earlier. This is also true of the concept of ‘respect,’ and to conclude this section of our analysis we should say something more in this regard. This will also serve to reveal a new argument for Kant’s thesis that the worth of an action from duty does not lie in the ‘effect to be expected from it.’

50. The subjective aspect (respect) is discussed in P1; we shall come back to this directly. 51. It is also worth noting that the passage “but rather in the maxim in accordance with which it is resolved upon” was absent from the fi rst edition of 1785, where P2 involved no mention of maxims at all, and Kant simply wrote “and it [the worth].”

Section I of the Groundwork

79

The Concept of Respect It is the motive of acting from duty, that is, from respect, that bestows unlimited moral worth on our actions. Although this thesis is undoubtedly the content of P1, there is no mention of ‘respect’ in the corresponding paragraphs (397–399); the concept in question is deployed explicitly for the first time in P3. Nonetheless, this difficulty is easily resolved because this concept has been present implicitly already: to act from duty, subjectively considered, is simply to act from respect. Kant later makes this unmistakably clear in retrospect: “Also we have shown above how neither fear nor inclination, but solely respect for the law, is the incentive that can give the action its moral worth” (440:5; o.e.). But it is not just this and other similar passages in which Kant uses the expressions ‘from duty’ and ‘from respect’ interchangeably that make this clear. In the context of his summary of P3 Kant introduces the concept of respect in a footnote (401:17–40). Since there is absolutely no difficulty in applying these determinations of the concept to the examples in P1, this is another indication that P1 is indeed concerned with the motive of respect. If we take the individual points made in the footnote together, the picture we have drawn so far is strengthened even further. Respect is a self-effected feeling that reflects an objective worth: 1. Respect is a feeling. We have already pointed out that for Kant all actions, including moral actions, must proceed, motivationally speaking, from a feeling of pleasure.52 Reason must have motivational power, and it can do so only if it can produce the volition of moral action without having to appeal to an already-existing subjective repertoire of wishes and interests on the part of the agent. As principium diiudicationis, reason must, if it is to be principium executionis as well, be able to produce ‘a feeling of pleasure or satisfaction in the fulfillment of duty.’ Respect is thus a “feeling” (401:19). The friend of humanity whose mind is ‘clouded over with his own grief’ still ‘tears himself out of this deadly insensibility’; the place of such insensibility is now taken by a capacity, a sensibility, for respect. 52. Cf. 460.

80

Immanuel Kant’s Groundwork for the Metaphysics of Morals

2. Respect is self-effected. Kant draws a distinction between feelings that are ‘received through influence’ (feelings ‘that may be reduced to inclination or fear’) and feelings that are ‘self-effected’ (such as ‘respect’). Kant’s claim that respect is ‘specifically distinguished’ from those other feelings is based (‘hence’) on this distinction. This notion of self-effection involves the following points: (i) The feeling of respect arises independently of natural conditions whose ‘influence’ is outside our control. (ii) This feeling, by contrast, is a ‘self-effected’ one. It is quite true that Kant says that the respect is the ‘effect of the law on the subject.’ But since this moral law is established ‘by ourselves’ in our autonomy (as self-legislation), it is we who are capable of producing the feeling of respect in us through this law.53 (iii) Since respect is a feeling, it cannot itself be directly commanded (any more than love can be commanded). But what can be commanded are reflection, the application of practical reason, and insight into the validity of the moral law. In assuming this rational practical attitude we experience the feeling of respect—as an ‘effect,’ as Kant says—which then functions as a subjective incentive for moral actions. But it is reason rather than feeling that stands at the source. (We have already seen how this dimension of autonomy plays a particular role in the example of beneficence and practical love.) 3. Respect reflects an objective worth. It is a deep part of Kant’s theory of action, and of his theory of human nature, that feelings have pragmatic significance, and that they implicitly register judgments of value. The feeling of respect is for Kant essential to the way in which we, as rational-sensible beings, become aware of the objective authority of the moral law.54 The fact that respect is an ‘effect’ of

53. When Kant says that respect is not to be regarded as the “cause” of the law (401:27), he means that respect is not itself the source of the moral law, as the proponents of the ‘moral sense’ tradition had more or less argued. 54. In the Critique of Practical Reason, respect is described as a cognitive feeling by which the validity of the categorical imperative becomes evident to us, i.e., is cognized such that no further justification or warrant is necessary; in this sense, the moral law is a ‘fact of reason.’ In GMS Kant also does not deny that there is such a thing that deserves to be called ‘consciousness of the moral law’; there, however, this consciousness does not suffice in and of itself to justify the moral law.

Section I of the Groundwork

81

the law but is nonetheless ‘self-effected’ is also evident from Kant’s observations on the object of respect. In the main text (400:17– 401:16) and in the relevant footnote (401:31), we are told that the object of respect is the moral law. For this claim, Kant offers an extended reflection that is formulated in two ways (400:19–401:2 and 401:6–16): a. One can have respect neither “for the object, as an effect of my proposed action” (400:19) nor “for inclination in general” (400:22).55 One can have respect only for a “ground” (400:25) and an “activity of a will” (400:21), that is, for “the law for itself ” (400:28). Kant then takes this thesis as an argument for the claim (which we have already discussed) that it is ‘the law as something objective and subjectively pure respect’ that must determine the will. Since, as far as action from duty is concerned, the objects of inclination and inclination itself cannot play any role, “so” (400:31; o.e.) it is only respect for the law, and thus this law itself, that can determine action for a good will. b. This reflection from paragraph 15 is presented again in another form in the next paragraph (this is clear from the elucidatory use of “thus” at 401:3 ).56 Kant repeats his claim that the moral worth of an action cannot lie in its ‘expected effects.’ He then continues: “For all these effects (agreeableness of one’s condition, indeed even the furthering of the happiness of others) could be brought about through other causes, and for them the will of a rational being is therefore not needed; but in it alone the highest and unconditioned good can nevertheless be encountered. Nothing other than the representation of the law in itself, which obviously occurs only in the rational being insofar as it, and not the hoped for effect, is the determining ground of the will, therefore constitutes that so preeminent good which we call moral, which is already 55. This thesis is specifically repeated at 428:11–17; cf. also 403:28–30. 56. That we are not dealing with an inferential ‘thus’ is clear from the fact that Kant at fi rst simply repeats his thesis about the irrelevance of ‘the expected effect” (401:3– 6) of an action. Then follows the other formulation of the reflection in question; the thesis regarding the irrelevance of the effect is clearly introduced with a justificatory “for” (401:6).

82

Immanuel Kant’s Groundwork for the Metaphysics of Morals

present in the person himself who acts in accordance with it, but must not first of all be expected from the effect” (401:6). There is far more to be learned from these two passages than first meets the eye. The main idea of the first passage (400:19–401:2) clearly lies in the concept of the “activity of a will” (paraphrased a few lines further on as a ‘ground’). Kant’s principal thesis is that this activity alone permits and deserves respect. Kant repeats this thesis in the second passage (401:6–16) when he says that ‘the highest and unconditioned good can be encountered’ only in ‘the will of a rational being.’57 This is simply a repetition of that thesis since the ‘activity’ of the ‘will of a rational being’ consists precisely in producing moral action through the moral law. Kant says indeed that the ‘representation of the law’ constitutes the morally good. But this moral goodness ‘obviously occurs only in the rational being,’ or, as he puts it immediately before this, it can be encountered ‘alone’ in the will of a rational being. As we have already suggested, these seemingly rather incidental remarks of Kant, which are not provided with any further grounding here, will later prove to be central elements of the argument (in GMS II and III), and, as we shall shortly see, they also already play a major role in the derivation of the CI in GMS I.58 What properly possesses worth and dignity, we learn in GMS II, is the free and rational human being in his or her capacity as an autonomous being to establish the moral law and to subject himself or herself freely and independently to this law. The human being is a free, that is, spontaneous or self-active being and deserves respect

57. The fact that one’s own happiness, or the happiness of others, could also ‘be brought about through other causes’ is not by itself an argument for the thesis that the unlimited good can be found in the rational alone, for the thesis that happiness can be brought about by causes other than that of a rational will does not necessarily imply that only the rational will is good. But the indirect reference to Kant’s teleological argument is evident here, for this argument claims indeed that happiness is more easily produced by other causes, namely, through instinct, than it is through reason. 58. With reference to respect and the worth that is bound up with it, Kant says in GMS I that we still “do not have any insight into that on which it is grounded (which the philosopher may investigate)” (403:26); cf. the remarks that follow shortly after this (405:23–27). Kant himself thus says that the thesis that one must have respect for the moral law has not yet been grounded as such. This grounding is presented only in GMS III.

Section I of the Groundwork

83

in this regard; nothing but this is what is meant by the ‘activity of the will’ and the respect that is owed to the person in question. That is why Kant says that the object of respect is the ‘law’ or the ‘representation of the law.’ But since this law is the product of our own reason, he can also subsequently say: “Our own will, insofar as it would act only under the condition of a possible universal legislation through its maxims, this will possible to us in the idea, is the authentic object of respect” (440:7; o.e.).59 This point is also present in the examples provided in P1. The man who is ‘by temperament cold and indifferent’ nonetheless fi nds “a source within himself to give himself a far higher worth than that which a good-natured temperament might have” (398:35; o.e.). For just here “begins the worth of character, which is moral and the highest without any comparison, namely that he is beneficent not from inclination but from duty” (398:37; o.e.).

Aims and Effects Let us now take a step back and consider what Kant says about ‘aims’ 60 and ‘effects’ in the context of P2 and P3. 1. P2 does not merely tell us that the worth of an action done from duty lies (objectively) in the moral law. In fact, P2 (or the ensuing elucidation up to 400:16) also tells us, or rather tells us again, what the worth of such an action does not consist in. It consists (i) ‘not in the aim that is supposed to be attained by it’; it depends (ii) ‘not on the actuality of the object of the 59. The very same close-knit points can be found again in paragraph 19, which summarizes the argument (403:18–33 ). First, Kant says that “for this [possible universal] legislation reason extorts immediate respect from me” (403:25). Then, in complementary relation to this ‘respect,’ Kant speaks of the “estimation of a worth which far outweighs everything whose worth is commended by inclination” (403:28). In turn, at the end of the sentence, this ‘worth’ is ascribed not to the moral law itself but rather to “a will that is good in itself, whose worth surpasses everything” (403:33 ). Cf. similarly 435:20–22; 436:2– 6; and 439:3–7. Correspondingly, at 401:3–16, the ‘highest and unconditioned good’ or also the ‘preeminent good’ is described fi rst as the ‘representation of the law’ but then also as the rational will itself. Kant obviously sees no contradiction here, for this representation springs from the rational being itself; that is also why we read that the representation of the law ‘obviously occurs only in the rational being.’ In the context of Kant’s teleological argument, which we have already discussed, the good will is also described as “the highest good” (396:25). 60. Absicht is sometimes translated into English as ‘intention’ (cf. note 1 in this chapter), but also sometimes as ‘aim’ (when it seems to mean something like Zweck).

84

Immanuel Kant’s Groundwork for the Metaphysics of Morals

action’; thus the source of worth here is neither (iii) ‘the aims that we may have in actions’ nor (iv) ‘their effects, as ends and incentives of the will,’ and thus (v) not ‘the effect hoped for.’ Kant summarizes all this by saying that in an action done from duty ‘every material principle has been withdrawn.’ This thesis is not actually substantiated in the context of P2. Rather, Kant says that it is “clear from the preceding” (400:6; o.e.). But since, as we have shown, the third paragraph of GMS I makes no mention whatsoever of ‘effects’ in the sense of a ‘material principle,’ this reference to ‘the preceding’ can only mean P1 and the elucidations that are connected with it.61 P1 is thus supposed to show that the worth of an action done from duty does not lie in a “material principle” (400:15) and ‘the effects that are aimed at’ (as we may summarize the preceding formulations). P1 essentially implies that only an action done from duty is worthy in an unlimited moral sense, for only such an action transpires (without any immediate or indeed indirect inclination) for the sake of the law itself. This can be negatively expressed as follows: one who acts from duty does not act because in this action one wants to satisfy a subjective interest. It is true that the moral person also pursues an end (for according to Kant himself there is no action and no maxim in the absence of an end).62 But the person in question does not act because moral action procures him something (as the indirect inclination of the merchant, for example, is satisfied by his acting in conformity with duty) or because he is moved to act in this way by immediate inclination (as in the case of the friend of humanity). Kant describes this very thought later in the text when he specifically distinguishes between hypothetical and categorical imperatives. Hypothetical imperatives command actions as means to an end; “The categorical imperative would be that one which represented an action as objectively necessary for itself, without any reference to another end” (414:15; o.e.). In this connection Kant then adds a note (413:26ff.) that is also particularly important because Kant here offers a retrospective summary of what has been achieved in GMS I. One who acts merely “from interest” (413:32)

61. Nonetheless, the fi rst three paragraphs of GMS I do already contain an argument against ascribing an unconditioned goodness to happiness. 62. Cf. 436:13–22.

Section I of the Groundwork

85

does not act morally because he or she is concerned only with satisfying an inclination. On the other hand, one who takes an interest in an action itself has a ‘practical interest’ (as distinct from a ‘pathological’ one): “In the first case [practical interest] the action interests me, in the second [pathological interest] the object of the action (insofar as it is agreeable to me). In the first section of the Groundwork we have seen that with an action from duty it is not the interest in an object that has to be looked to, but merely the action itself and its principle in reason (the law)” (413:37; o.e.). The moral person acts solely out of respect for the validity of the law and pursues no related ‘aims and effects,’ no ‘pathological interest.’ This is thus the point that is ‘clear from the preceding.’63 One could interpret this point as a critique of the position in presentday moral philosophy known as contractualism, at least in its egoistic form (which is sometimes called ‘contractarianism’). The basic idea behind contractarianism (perhaps formulated most clearly in the history of philosophy by Thomas Hobbes, but more recently by David Gauthier) can be expressed as follows: All human beings require cooperation and protection in order to be able to satisfy their individual interests. Therefore, they also have self-interested reasons for seeing that rules are established, maintained, and observed to facilitate such cooperation. This is accomplished by means of a contract that governs these rules in a regular and binding way. According to contractarianism, the only reason for establishing and observing this contract is that it maximizes the overall advantage and thus also the advantage of the individual. There is a range of problems associated with this position (such as the so-called free-rider problem or the ‘prisoner’s dilemma,’ which shows that cooperative behavior does not necessarily secure optimal advantage in individual cases).

63. In connection with P3 Kant once again emphasizes and recapitulates this point. In the sentence we have already repeatedly cited regarding the subjective and objective aspect of the concept of duty, Kant expressly says that “an action from duty is supposed entirely to abstract from the influence of inclination, and with it every object of the will” (400:30; o.e.), so that nothing but the law and respect for this law are left to determine the will: “The moral worth of the action thus lies not in the effect to be expected from it; thus also not in any principle of action which needs to get its motive from this expected effect” (401:3; o.e.). The worth lies rather in “the maxim of complying with such a law, even when it infringes all my inclinations” (400:33 ).

86

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Kant’s objection is simply that contractarianism has nothing whatever to do with what we understand by ‘morality.’ It is at best a sort of ‘quasimorality’ because it transforms the system of morality as a system of categorical imperatives into a system of hypothetical imperatives that simply and exclusively serve the purpose of satisfying subjective interests. One might object here that what we actually understand by ‘morality’ is different from the way in which we should understand ‘morality,’ and that contractarianism gives us a rational reconstruction of morality as we should understand it. Up to a certain point, Kant would acknowledge this objection, for he has not yet claimed that there is something such as ‘morality’ at all (it might, after all, simply be a ‘figment of the mind’). But Kant claims that we not only actually understand the concept of morality in terms of categorical imperatives but also should understand it in this way. Kant’s reference to ‘common rational moral cognition’ is supposed to show that this is indeed how we do understand morality, and—by referring to this understanding as ‘cognition’—he implies that this understanding is correct, even though it will not be until GMS III that he thinks that he is in a position to vindicate it philosophically. This feature of common rational moral cognition is precisely what the examples in P1 are supposed to demonstrate. We could not grasp or appreciate these examples if we did not actually make a distinction between actions in conformity with duty and actions that are done from duty. It is only in the metaphysics of morals proper (MS3 ), and ultimately only in GMS III, that we can fully understand this point: to act morally is to recognize that there is something that is absolutely worthy, something that in the dignity that thereby defines it escapes the influence of action that is determined by subjective interests. 2. Now (classical) utilitarianism would also already emphatically reject the idea of reducing morality to hypothetical imperatives. According to this position, it is the duty of every individual to act in such a way that the greatest happiness of the greatest number is realized. The utilitarian position does not claim that if we want to realize the greatest happiness of the greatest number, then we must perform particular actions (such as keeping one’s promises); rather, we ought to realize the greatest happiness of the greatest number whether we actually want to or not. Thus utilitarianism is by no means an ethics for egoists, and when Kant critically observes

Section I of the Groundwork

87

that morality cannot simply be concerned with one’s own individual interests, a utilitarian could endorse this perspective as well. But Kant does not simply criticize the tendency to orient ethical criteria toward one’s own interests. He also challenges the idea that ethics is concerned with the interests of everyone and their ( justly distributed) happiness at all. And the “furthering of the happiness of others” (401:7; o.e.) is also irrelevant here. In hypothetical imperatives one finds nothing but the “necessity of action from a certain interest. Now this might be one’s own interest or someone else’s” (433:6; o.e.). Thus Kant can then say explicitly: “Thus, e.g., I should seek to promote someone else’s happiness, not as if its existence mattered to me . . . but merely because the maxim that exludes it cannot be comprehended in one and the same volition as a universal law” (441:19; o.e.). Kant’s arguments against treating happiness as the highest criterion of ethics are not confi ned to the reflections contained in the first three paragraphs of GMS I. Kant is basically attempting to develop a different theory of worth or value. The question is simply whether this theory inevitably leads to the derivation of the CI, as Kant claims and as he undertakes to demonstrate in a preliminary way in GMS (402). 3. For intentionalist rule consequentialism, the willing (of an action) is good if the willing (and acting) person wills to act (and acts) in accordance with a rule that, if it were universally observed, would have good consequences. The principal thesis of Kant, as we have seen, consists in the contrary claim that we must value morality for its own sake. Kant is essentially concerned with the “motivating ground”64 [Bewegungsgrund] of action: why do human beings act morally, and why should they act morally? As we shall see later, the question ‘How are categorical imperatives possible?’ is also equivalent, among other things, to the question ‘Why should I be moral?’ But what precisely does it mean to act ‘for the sake of morality’? Critics would presumably agree with Kant that actions that he describes as done ‘from duty’ are different from actions that transpire ‘from inclination.’ But the egoistic contractarian would argue that they differ only on the surface: actions that look like purely moral actions are not in fact such at all; and even if altruism appears in a particular case, 64. In the immediate context of the three propositions regarding duty; cf. 398:10; 401:5; and 403:32.

88

Immanuel Kant’s Groundwork for the Metaphysics of Morals

this is ultimately to be interpreted as a form of behavior that is concerned with optimizing one’s own interest. According to a second objection, even if moral action is not to be understood in egoistic contractarian terms as self-interested action, it is still ultimately always a question of the consequences of action. From a certain perspective the position of rule consequentialism can also bestow extremely high value on a rule in and of itself, but its value or worth still ultimately lies in the consequences that it produces as a rule that is generally or universally observed. But it is misleading to see the decisive difference between Kantian ethics and consequentialism in the fact that the former, in contrast to the latter, pays no attention to consequences. The decisive difference here is rather that the two positions are based on quite different theories of value or worth. For a consequentialist the worth of moral laws lies in the function they perform in attaining something else (general happiness, the fulfilment of preferences). It is something other than the law itself that is worthy or valuable; and it is something about human beings that is worthy or valuable, rather than human beings themselves as autonomous beings in their own right.65 Consquentialists thus in a certain sense externalize worth and dignity, whereas Kant places worth in the moral law itself and the legislator of this law. But this decisive point is not made clear in GMS I. Apart from the vague references to the ‘activity of the will,’ one cannot find a theory of value or worth in GMS I at all; this theory can be found, as we have indicated, only within the metaphysics of morals that forms part of GMS II (MS3 ). Only when we get there do we read: “But suppose there were something whose existence in itself had an absolute worth, something that, as end in itself, could be a ground of determinate laws; then in it and only in it alone would lie the ground of a possible categorical imperative, i.e. of a

65. In the lectures on moral philosophy (MM II), which Kant delivered around the time of the composition of GMS, he distinguishes between the ‘condition’ of a person and the ‘person’ as such (MM II:599ff., 612– 613, 620). Kant reproaches ‘the ancients’ for not clearly distinguishing between ‘virtue’ and ‘happiness,’ “for the one shows the worth of the person, whereas the other the worth of the condition” (MM II:600). Cf. also GMS 450:12: we ought to consider ourselves “as free in acting and thus nevertheless take ourselves to be subject to certain laws in order to fi nd a worth merely in our person, and that this could compensate us for the loss of all the worth procured for our condition” (o.e.).

Section I of the Groundwork

89

practical law” (428:3 ). This end in itself is “rational nature” (429:2). In GMS I it is not Kant’s intention to illuminate this ‘ground’ in any detail. He simply wants to show that even at the level of ‘common rational moral cognition’ we already distinguish between relative and absolute worth, and that this distinction is captured by that between actions done from duty and actions done from inclination.

2.4 The Derivation of the Categorical Imperative There is a sense in which P3 embodies the central claim of GMS I: in actions that possess moral worth, the crucial thing is that they transpire for the sake of the law itself. But up to this point Kant has not yet explained what this ‘moral law’ actually is. He has certainly clarified what the ‘subjective’ aspect of the concept of duty consists in (i.e., respect). But it is only after P3 has been formulated that Kant asks what that law properly consists in: “But what kind of law can it be, whose representation, without even taking account of the effect expected from it, must determine the will, so that it can be called good absolutely and without limitation? Since I have robbed the will of every impulse that could have arisen from the obedience to any law, there is nothing left over except the universal lawfulness of the action in general which alone is to serve the will as its principle, i.e. I ought never to conduct myself except so that I could also will that my maxim become a universal law. Here it is mere lawfulness in general (without grounding it on any law determining certain actions) that serves the will as its principle, and also must so serve it, if duty is not to be everywhere an empty delusion and a chimerical concept; common human reason, indeed, agrees perfectly with this in its practical judgment, and has the principle just cited always before its eyes” (402:1).66 It is here, therefore, that the CI is first formulated and derived.67 How precisely are we to understand this ‘derivation’? First of all, we should briefly consider the example that Kant appends to the derivation of the CI (402:16–403:17): “Let the question be, e.g.: When

66. One can fi nd a parallel passage at 420:24–421:5, and there is another parallel passage in Section 1 of GMS III (although this passage is not often recognized as such). 67. This ‘derivation’ [Ableitung] of the categorical imperative is not to be confused with its ‘deduction’ [Deduktion] in GMS III.

90

Immanuel Kant’s Groundwork for the Metaphysics of Morals

I am in a tight spot, may I not make a promise with the intention of not keeping it?” (402:16). Kant ascribes two meanings or significations to this question:68 in the first place (402:17–403:2), it is concerned (once again) with the distinction between an action done from duty and one done from inclination; in the second place (403:2–17), it is concerned with what is often called the ‘test of universalizability.’ Since we shall discuss the second aspect of this ‘question’ (whether a maxim is ‘universalizable’) more closely when we consider the examples in GMS II, there is no need to examine it in detail here. The first part of this example requires no further examination either, for Kant is here simply repeating his principal point that in the case of an action done from duty we are not concerned with “the worrisome consequences” (402:31), for “to be truthful from duty is something entirely different from being truthful out of worry over disadvantageous consequences” (402:32). The fact that Kant explains what the moral law consists in only after the formulation of P3 clearly reveals his strategy: he wants to derive the moral law itself and its content from an analysis of moral motivation, that is, from an analysis of what it means to act from duty. An action done from duty may not contain any motive by which the action serves the realization of a subjective end; thus Kant argues that there is nothing but ‘mere and universal lawfulness in general’ that can constitute the content of the law itself. But clear though Kant’s basic strategy may be, this derivation is highly problematic (at least on the basis of the argument presented in GMS I). For up to this point all that Kant has claimed about the feature of ignoring all material interests is simply that this is a feature of the motive for obeying the moral law; he has certainly not claimed that it is a feature of the law itself. As yet, he has provided no argument for the thesis that this feature implies anything about the character of the law itself. Even if Kant had shown that the ‘effects’ may not play any relevant role (if he thus possessed an argument against a defender of rule consequentialism),

68. Kant himself expressly points out the two distinct meanings here: with regard to the fi rst, he says that he is making a “distinction in the signifi cation the question can have” (402:18; o.e.); he then introduces the second meaning as follows: “Meanwhile, to inform myself . . . in regard to my answer to this problem” (403:2; o.e.).

Section I of the Groundwork

91

it would not follow from this that ‘mere lawfulness’ is the sole criterion of the law, because this on its own would not exclude the possibility of understanding the moral law in a supernatural manner as the law of God, for example. But perhaps we can understand Kant’s reflections in a more appropriate and plausible fashion if we consider them in the light of his (admittedly rather sparse) remarks on the idea of autonomy. Kant’s principal thesis in GMS I is this: only the good will is good without limitation. As far as human beings are concerned, this means that only actions done from duty possess unconditioned moral worth. Actions done from duty are, subjectively considered, actions done from respect. This respect is ‘selfeffected’ (i.e., springs from reason) and relates in a fundamental sense to itself (to the ‘activity of the will’). What deserves respect is the human being in his or her capacity to give the moral law to himself or herself; it is only autonomy that deserves respect. Kant then reasons as follows: if what is decisive about an action done from duty is that it transpires for the sake of the law itself, and if the respect intrinsically involved relates to the dignity of the human being in his or her autonomy rather than to inclinations (and ‘effects’), then the law itself cannot be characterized by any relation to such inclinations either, but only by the fact that it is a law  as such (‘mere lawfulness in general’), for there is nothing else “left” (402:6). Kant obviously mixes the derivation here with reflections drawn from the later sections of GMS. In the first place, we already have Kant’s idea that the binding character (the validity) of the moral law can be derived only from the autonomy of the human being: “One saw the human being bound through his duty to laws, but it did not occur to one that he was subject only to his own and yet universal legislation. . . . For if one thought of him only as subject to a law (whatever it might be), then this would have to bring with it some interest as a stimulus or coercion, because as a law it did not arise from his will, but rather this will was necessitated by something else to act in a certain way in conformity with the law. Through this entirely necessary consequence, however, all the labor of fi nding a supreme ground of duty was irretrievably lost. For from it one never got duty, but only necessity of action from a certain interest. Now this might be one’s own interest or someone else’s. But then the imperative always

92

Immanuel Kant’s Groundwork for the Metaphysics of Morals

had to come out as conditioned, and could never work at all as a moral command” (432:28).69 In the second place, the idea of autonomy already affords a glimpse of the third derivation of the CI, which is properly presented only in Section 1 of GMS III. There Kant argues that the freedom of the will has to be understood as a particular form of causality. Since it involves independence from the laws of nature, and since, like any form of causality, it must possess a law-like character, this freedom of the will can be nothing other than “autonomy, i.e. the quality of the will of being a law to itself [.] But the proposition: ‘The will is in all actions a law to itself,’ designates only the principle of acting in accordance with no other maxim than that which can also have itself as a universal law as its object. But this is just the formula of the categorical imperative” (447:1). Here too, therefore, the derivation of the moral law arises immediately from the autonomy of the will. But even if we allow that Kant’s argumentation is based on a specific theory of value or worth that is fully developed only in GMS II and III, the derivation here is still defective. From the (alleged) fact that the human being as a legislating being is an end in himself or herself and thus possesses absolute worth, it by no means follows that ‘mere lawfulness’ (universalizability) is precisely what constitutes the moral law. We could equally well conceive of a moral law that was aimed at protecting the absolute dignity of the human being— as is indeed the case in the formula that speaks of human beings as ends in themselves. This involves far more than ‘mere lawfulness,’ quite apart from the fact that Kant does not actually say what we are to understand by ‘mere lawfulness’ and also says nothing further about the claim ‘that I could also will that my maxim become a universal law’; all this is elucidated only in GMS II. Thus it cannot be said that Kant has presented a compelling argument against intentionalist rule consequentialism in GMS I. He also needs to supplement this argument with some reason for thinking that ‘what I can will’ should play a role in determining the content of the moral law.

69. Cf. Kant’s formulation in the passage presenting the derivation at 402: “that could have arisen from the obedience to any law” (402:4) and “(without grounding it on any law determining certain actions)” (402:10).

Section I of the Groundwork

93

2.5 Summary 1. The gifts of nature and the gifts of fortune possess no unconditioned moral worth because they can be used either morally or immorally; they acquire a conditioned worth only insofar as they depend on a good will that makes use of them. Kant emphasizes that happiness too is good only in a limited sense since it is capable of exercising a corrupting influence and cannot be regarded as good if it is enjoyed by a bad human being. 2. A good will involves the summoning up of all the means at its disposal; to will the good involves more than simply wishing for it. 3. The moral quality of a will does not depend on whether it is successful in realizing the aims it has pursued, for the initial and circumstantial conditions required for this do not lie in the field of responsibility of the agent (and this holds even more obviously for the actual and unintended consequences). GMS I does not present a critique of intentionalist rule consequentialism. 4. Kant supports his thesis regarding the unlimited worth of the good will by appeal to a teleological argument: the organs that belong to organized living beings spontaneously fulfill their purpose in an optimal way; if it were the task of reason to produce happiness as the ultimate end, it would not fulfill its function in an optimal way; this is therefore not the task of reason; rather, its task is to determine the will as something that is good in itself. For a variety of reasons, this argument is untenable. 5. The introduction of the concept of duty derives from the difference between beings that act in a perfectly rational (and thus moral) way and beings that are both sensuous and rational. It is only for beings that have subjective inclinations and interests that the moral law presents itself as a duty. Consequently, every action done from duty is an action done from a good will, though not vice versa. 6. The first proposition regarding duty (P1), though not explicitly identified by Kant, is this: an action done from duty is an action done from respect for the law; this describes the subjective aspect of the concept of duty. The objective aspect of this concept is thematized in

94

Immanuel Kant’s Groundwork for the Metaphysics of Morals

the second proposition (P2): an action done from duty is an action that observes a maxim that is demanded by the moral law, and is thus an action that is necessarily commanded by the objective moral law. The third proposition (P3 ) connects both of these aspects: duty is the necessity of an action from respect for the law. 7. In the context of P1 an action done from duty is described as an action in conformity with the law that transpires, without any inclination, simply for the sake of the law. Although Kant is not concerned with offering a theory of ‘motivational overdetermination’ here, it is clear that an action done from duty cannot simultaneously be determined or accompanied by any inclination. Kant’s examples are deliberately constructed and artificial; they are designed to indicate clearly how an action done from duty and an action done from immediate inclination differ, which is necessary precisely because these two types of action resemble each other. They resemble each other because in both cases the action is something that we ourselves will; an action done from duty cannot be said to transpire against one’s will. 8. Respect is a feeling that reflects an objective worth. As a feeling, respect can motivate us to an action, as inclinations also can. As a selfeffected feeling, respect is not dependent on the contingent nature of the agent but rather on the freedom of reason that belongs to the agent. Respect ultimately relates— as Kant indicates only in passing in GMS I—to that activity of the autonomous will that allows it to establish moral laws; respect relates to something that possesses objective worth. The decisive difference between consequentialism and Kantian ethics thus also lies in the different conceptions of value or worth at issue here. 9. The analysis of moral motivation also serves to determine the content of the moral law. However, Kant’s fundamental idea that only the form of the law itself is relevant, since subjective interests are not permitted to play any role in actions done from duty, is not convincing; the derivation of the CI here is a failure, or at least it remains incomplete and needs to be supplemented.

3 Section II of the Groundwork Practical Reason, Imperatives, and Kant’s Formulas

We begin with an overview of the structure and basic argument of GMS II (3.1). Then we shall address the issue of reason as a practical faculty and  the division of rational principles into categorical and hypothetical imperatives (3.2). The question regarding the possibility of hypothetical imperatives merits separate discussion in its own right (3.3 ) before we examine Kant’s various formulas for the categorical imperative and his attempt to derive different duties from these formulas (3.4). Finally, we summarize the main results (3.5).

3.1 The Structure and Basic Argument of GMS II The opening pages of GMS II constitute one of the key passages where Kant explicitly distinguishes the metaphysics of morals from ‘popular moral philosophy’ (406–412:26). Before he makes the transition to the metaphysics of morals (MS3 ) (425–427:18) and when he comes to analyze and contrast the fundamental concepts of autonomy and heteronomy (441:25–444:34), Kant repeats and expands on these general observations. Since we have already discussed this question in connection with the Preface, it does not require any further analysis here.

96

Immanuel Kant’s Groundwork for the Metaphysics of Morals

It is only when Kant analyzes “the practical faculty of reason” (412:23 ) that substantively new reflections are introduced. His analysis is centered on the concept of an imperative that is defi ned as a necessitating principle of reason and then divided into categorical and hypothetical imperatives (412:26–417:2). He immediately asks how such imperatives are possible (417:3– 6). He provides an answer in relation to hypothetical imperatives (417:7–419:11) but postpones the answer to the question how categorical imperatives are possible (419:12–420:23 ). He proceeds instead with his conceptual analysis and presents a formula, or rather the basic formula, of the CI (420:24–421:13 ). He then employs the alternative formula, which speaks of a law of nature, to furnish an initial division of duties (421:14– 424:37). But after this he once again postpones the question regarding the possibility of a CI (425:7–11). He instead presents the transition to the metaphysics of morals (425:1–427:18). Within this metaphysics of morals (MS3 ) Kant thematizes some of the fundamental concepts of his ethical theory in terms of the corresponding formulations of the CI by reference to an end in itself (427:19–430:27), to autonomy (430:28–433:11), and to the realm of ends (433:12–436:7). Kant then provides a summary (436:8– 440:13 ) and an overview of the distinction between autonomy and heteronomy (440:14–444:34). Finally, at the end of GMS II (444:35–445:15), Kant once again emphasizes that the question how a CI is possible at all can be answered only in GMS III. So the basic structure of GMS II can be outlined as follows: Popular moral philosophy and the metaphysics of morals (406:1–412:25) The practical faculty of reason and the division of imperatives (412:26–417:2) The possibility of imperatives (417:3–420:23) The basic formula of the categorical imperative (420:24–421:13) The law-of-nature formula and the division of duties (421:14–424:37) The transition to the metaphysics of morals (425:1–427:18) The concept of an end in itself (427:19–430:27) Autonomy (430:28–433:11) The realm of ends (433:12–436:7) Summary (436:8–440:13)

Section II of the Groundwork

97

Autonomy and heteronomy (440:14–444:34) Transition to GMS III (444:35–445:15) If we disregard the recapitulatory passages that we have already mentioned, the overall structure of the argument is clear, but it is also surprising. Kant begins with an analysis of the different kinds of imperatives and immediately asks how such imperatives are possible. But the answer to the question regarding the possibility of the CI itself is postponed, and indeed postponed twice (before he provides the basic formula of the CI and the alternative formula of the law of nature, and before the transition to the metaphysics of morals); this task, he tells us, “no longer lies within the boundaries of the metaphysics of morals” (444:36). Two things are particularly noteworthy here: (1) Although the basic and the law-of-nature formulas unquestionably belong among the various formulations of the CI—and are specifically included by Kant among the three fundamental ways of formulating it (436)—Kant introduces and discusses them before the metaphysics of morals (MS3), whereas the other formulations are discussed within the metaphysics of morals.1 (2) The normative question why the CI is valid and binding on us, and what is presupposed by this validity, is already thematized by Kant very early on in the investigation. Although he specifically postpones the answer to this question, the metaphysical reflections in MS3 must be understood as constituent elements of the subsequent deduction of the CI (this deduction is the answer to that normative question).

3.2 The Practical Faculty of Reason and the Division of Imperatives Kant introduces the properly systematic part of GMS II by saying that he now intends “to follow and distinctly exhibit the practical faculty of reason” (412:23).2 Although this part of GMS is of central significance for Kant’s ethics, it is very brief. But since it is also fairly easy to grasp, we can

1. At the same time, it is clear that all the formulations belong to MS1. 2. The metaphysics of morals (MS3 ) begins with another brief analysis of the practical faculty of reason (427:19–428:2); we shall not discuss that passage of the text in any detail.

98

Immanuel Kant’s Groundwork for the Metaphysics of Morals

keep our discussion relatively brief here. The first thing we have to do is clearly identify the essential features of what Kant means by ‘imperatives’ (3.2.1). Then we can proceed to elucidate his division of the field into categorical and hypothetical imperatives (3.2.2).

3.2.1 Imperatives as Objective Rational Principles for Beings That Are at Once Sensuous and Rational “All imperatives are expressed through an ought and thereby indicate the relation of an objective law of reason to a will which in its subjective constitution is not necessarily determined by that law (a necessitation)” (413:12). This sentence furnishes a brief summary of Kant’s theory of reason as a practical faculty. ‘All’ imperatives (including, therefore, hypothetical ones) are ‘objective laws of reason,’ that is, are rules for action that possess objective validity for all rational beings; and ‘all’ imperatives involve an ‘ought.’ Taken literally as a linguistic statement, this last claim is, of course, not true. Imperatives as Kant understands them can take many linguistic forms; ‘imperative’ is not being used here as a grammatical term.3 As we will see presently, one canonical way in which Kant thinks that an imperative to do an action A might be expressed is “A is practically good.” The point is rather that these laws are addressed to sensuousrational beings who do not necessarily observe these laws, and for whom therefore the latter assume the form of ‘necessitation’ or (self-)constraint. An imperative (in the sense meant here) is anything that serves to express the idea that the agent is rationally constrained regarding an action or an end. Let us examine these defining features of imperatives in more detail.

Reason, Will, and Ends In contrast to his lectures and also his later writings, in GMS Kant does not furnish an explicit doctrine of the human faculties or offer a precise 3. The concept of a categorical imperative is recognized in substance already in KrV, but there Kant calls it (as he does repeatedly later as well) a “pure practical law” (e.g., KrV:A800/B828) or a “pure moral law” (e.g., KrV:A807/B835). The term ‘imperative’ already occurs in KrV (A547/B575, A802/B830), but the technical term ‘categorical imperative’— a neologism of Kant’s (drawn from its use in logic and grammar)—is found in his published writings for the fi rst time in GMS (414).

Section II of the Groundwork

99

and detailed discussion of the fundamental concepts involved in the theory of action. This is partly responsible for the fact that Kant’s use of concepts such as ‘reason,’ ‘will,’ ‘inclination,’ and so on is not always as precise as it could be and at first sight can sometimes even appear rather contradictory.4 There is no need here to examine the way in which Kant divides the “faculties or capacities of the soul” (KU:177) into the faculty of cognition, the feeling of pleasure and displeasure, and the faculty of desire or to explore the way in which such fundamental action-theoretical concepts are deployed in Kant’s work as a whole. We must simply bear in mind that Kant does not in fact always employ the central concepts of ‘desire,’ ‘inclinations,’ ‘will,’ and ‘reason’ in a univocal fashion, but that it is nonetheless generally quite clear what he wants to say.5 Thus Kant identifies the will with practical reason: “The will is nothing other than practical reason, namely the capacity to act in accordance with principles” (412:29). But he appears to contradict himself in the very next sentence when he says: “If reason determines the will without exception” (412:30), and again, a little further on, when he writes: “But if reason for itself alone does not sufficiently determine the will” (412:35). If practical reason is itself ‘nothing other’ than the will, how can it be said to determine the will? In fact, in an earlier passage, reason is defi ned as “a practical faculty, i.e., as one that ought to have influence on the will” (396:18). For our purposes here, it is enough to note that practical reason is that faculty that establishes and grounds practical principles (hypothetical and categorical imperatives), and the faculty of acting through the representation of such principles is what Kant calls the ‘will.’ But he also distinguishes between the pure and the empirical will, and that is why (as in 412) he sometimes identifies the (pure) will with practical (moral) reason without thereby abandoning the usual sense of the will or willing—what we will can be rational or contrary to reason, as the case may be. If we ignore more detailed and precise distinctions, what Kant understands by ‘inclinations’ is roughly what we would understand today as subjective interests or wishes 4. In KpV (9, footnote) Kant himself admits as much and then proceeds to explicate and defi ne concepts such as ‘life,’ ‘faculty of desire,’ and ‘pleasure’; in this regard; cf. also the relevant discussion in KU (176ff.). 5. Kant’s well-known distinction between will [Wille] and choice [Willkür] (cf. MS:213–214) plays no role (at least terminologically speaking) in the context of GMS.

100

Immanuel Kant’s Groundwork for the Metaphysics of Morals

(preferences) that are bound up with pleasure or displeasure.6 This is why Kant also describes happiness as the idea in which “all inclinations are united in a sum” (399:9; cf. 405:7). The decisive point for the present discussion is that inclinations are understood as part of our sensuous constitution (cf. 457:36). Here too, as in the sphere of theoretical philosophy, we are confronted with an opposition between reason and sensibility. And here too, as in the context of theoretical philosophy, the task is to determine how these two stems of cognition and action are connected with one another. Although Kant speaks of “the derivation of actions from laws” (412:28)— and in this sense the “derivation” (423:37) of various duties from the CI plays an important role later—when he speaks of “reason” here, unlike in the context of theoretical philosophy, he is not referring to the faculty of inference. As we already made clear in our discussion of the metaphysics of morals (MS1 ), Kant understands practical reason (here) as the faculty or capacity for establishing objective and necessary principles that are justified for all rational beings. We say that a rule or a judgment is ‘objective’ if it holds without regard to subjective interests and perspectives or is valid precisely for all the parties involved on the basis of reasons that can be recognized by everyone. It is in this precise sense that Kant also speaks repeatedly, in the few pages under discussion here (412:26– 414:11), of the objectivity and necessity of rational principles.7 Rational 6. For the way in which Kant relates and combines these and similar expressions, cf., for example, 405:7; 405:16; 413:27; 425:25; 434:21; 434:35; 453:24; 453:28; 454:27; 457:26; and 457:36; at 427:7– 9 Kant explicitly says that “desires and inclinations” rest on “the feeling of pleasure and displeasure.” However, ‘inclination’ is not equivalent to ‘desire,’ since Kant recognizes desires that arise from reason rather than empirical impulses; these latter desires must be present if action is to occur in response to moral imperatives. He says so quite clearly in GMS: “In order for a sensibly affected rational being to will that which reason alone prescribes the ‘ought,’ there obviously must belong to it a faculty of reason to instill a feeling of pleasure or satisfaction in the fulfillment of duty” (460; cf. KpV 5:9n, MS 6:212–213). Inclinations are also habitual empirical desires, involving the choice of a maxim and the self-determination of the agent (MS:212; Anth:251). 7. The actions that reason prescribes on the basis of these principles are described by Kant as “objectively necessary” (412:32; o.e.). In the pages in question (412–416) Kant speaks of the “objectivity” (of laws) at 412:37; 413:2; 413:4; 413:5; 413:9; 413:13; 413:19; 414:1; 414:9; 414:30; 415:4; and 416:21; he speaks of “necessity” at 412:34; 413:3; 414:13; 414:17; 414:19; 414:24; 415:4; 415:8; 415:34; 415:35; 416:22; 416:24; and 416:28.

Section II of the Groundwork

101

(objective) principles are rules for action that reason recognizes and establishes “independently of inclination” (412:34). Such rules are binding “objectively, i.e. from grounds that are valid for every rational being as such” (413:19). An objective principle is therefore objective because it is “valid for everyone” (413:24) or could be recognized as valid. Kant consequently distinguishes such “grounds of reason” (413:7) from merely “subjective causes” (413:19; 413:22). It is also in precisely this sense, as we saw earlier, that an objective rational principle is distinguished from a maxim, for in the first instance a maxim is merely subjective (inasmuch as it is the maxim of a subject). Maxims are certainly also rules, but they lack, or rather can lack, the objectivity that makes them binding for all beings. What Kant says about the objective character of rational rules for action holds for all imperatives, not only for categorical ones. This is also evident from the fact that the concept of the ‘good’ in this context is not initially limited to that of the morally good. Thus when Kant writes that the will is the faculty of “choosing only that which reason, independently of inclination, recognizes as practically necessary, i.e. as good” (412:33), he is not referring solely to categorical imperatives (as the footnote at 413 might perhaps suggest). It is quite true that categorical imperatives are indeed in every respect ‘independent of inclination’ (unlike hypothetical imperatives, which are based on particular subjective ends whose cause is to be found in our inclinations). But the formulation and pursuit of hypothetical imperatives is a rational and objective act that is also valid for all rational beings. The end that an individual proposes to pursue in terms of hypothetical imperatives is purely subjective; but the “practical rule as to how the need of inclination is to be supplied” (413:36) has an objective character and is valid for all rational beings. This is why Kant can write of the good in general as follows: “Practical good, however, is that which determines the will by means of the representations of reason, hence not from subjective causes, but objectively, i.e. from grounds that are valid for every rational being as such” (413:18). Immediately after he has introduced the distinction between categorical and hypothetical imperatives, Kant therefore directly brings together the universal and the dual character of the good: “Because every practical law represents a possible action as good, and therefore as necessary for a subject practically determinable by reason, all imperatives are formulas of the determination of action, which

102

Immanuel Kant’s Groundwork for the Metaphysics of Morals

is necessary in accordance with the principles of a will which is good in some way [o.e.]. Now if the action were good merely as a means to something else, then the imperative is hypothetical; if it is represented as good in itself, hence as necessary, as the principle of the will, in a will that in itself accords with reason, then it is categorical” (414:18). There is another important point that needs to be clarified briefly here. We have indicated that maxims are subjective principles of willing. But Kant seems to emphasize that we do not simply have maxims. Rather, they are something that we ‘make’ for ourselves. Kant refers repeatedly to this process of ‘making’ in his specific examples of maxims, which is why he also defines them (“i.e.”) as “self-imposed rules” (438:24; o.e.). This is particularly important in the case of hypothetical imperatives. In contrast to what Kant describes as ‘animal volition’ (tierische Willkür), human beings do not act immediately or mechanically on the basis of an incentive. Where rational beings are concerned, when they are rational, wishes, desires, interests, or inclinations (or however we choose to describe these underlying sensuous incentives) only give rise to actions once the relevant motivational force has been incorporated into a maxim that has an envisaged end and a corresponding means of securing that end as its content. At least this is what Kant claims in his book on religion when he says: “Freedom of the power of choice has the characteristic, entirely peculiar to it, that it cannot be determined to action through any incentive except so far as the human being has incorporated it into his maxim (has made it into a universal rule for himself according to which he wills to conduct himself)” (R:23–24). Naturally it would be quite misleading to think that someone who acts in accordance with a maxim must in fact at the same time be conscious of this maxim in every par ticular action; and indeed it is not the case that the underlying maxim would have to receive a clear and invariant formulation. Human beings can act in accordance with principles (rules) without actually being conscious of these rules (in the performance of the action), indeed even without ever consciously having learned these rules or ‘incorporated’ them in this sense. The way we learn a language is a good example of this: we do not learn our mother tongue by grasping, practicing, or memorizing grammatical rules in a conscious way. On the contrary, it all transpires unconsciously. We follow these rules although we do not do so consciously and would probably be inca-

Section II of the Groundwork

103

pable of articulating the rules we follow if we were asked to do so (and most certainly not those encoded in ‘deep grammar’). Maxims can work in a similar way: it is certainly true that human beings absorb maxims in the course of socialization and education without being aware of the process itself, but we do not for that reason deny them all freedom of choice (although we might accuse them of not reflecting sufficiently on their maxims). All the same, the passage we have just cited derives not from GMS but from the much later text Religion within the Boundaries of Mere Reason of 1793. The thesis as it is expressed there presupposes that action deriving from the representation of hypothetical imperatives is already a case of free action. But Kant is notoriously unclear about this point. It poses repeated problems for the interpretation of Kant (and not only of GMS) that for a long time (at least between the fi rst and second editions of the first Critique) Kant himself remained unclear about what form of freedom is required for the rational activity of setting ends (for action in accordance with hypothetical imperatives). Is it transcendental freedom that is involved here, or is it some weaker form of freedom?8 We shall see that in GMS III Kant identifies freedom in what he calls the ‘positive sense’ with autonomy, and he understands the latter explicitly as the capacity that allows us to act morally. In this sense, the status of the human being as an end in itself is grounded in autonomy rather than in the capacity that allows us to set rational ends. Hypothetical and categorical imperatives have another feature in common: they are both bound up with the setting of ends. Although hypothetical imperatives are distinguished from categorical imperatives by the fact that they are based on subjective ends, it would be quite wrong to think that someone who acts in accordance with categorical imperatives is not pursuing any ends. Kant expressly emphasizes that with all rational actions— that is to say, with action in accordance with maxims— ends are indeed pursued. “All maxims have . . . a matter, namely an end” (436:13).9 8. Thus in KrV (A801ff./B829ff.), for example, what Kant calls “practical freedom” is even described as “empirical.” 9. Cf. 427:19–20 and especially 437:21–30; cf. in addition MS:384–385; KpV:58–59; TP:279, footnote; and R:4. In due course we shall explore the significance of the purposive character of all action for categorical imperatives in par tic u lar.

104

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Imperatives and Necessitation One might almost think that the general observations regarding ‘imperatives’ that we have just presented are entirely self-evident as far as Kant is concerned. Instead of elucidating the objective character of rational principles and the concept of reason itself in any detail, Kant concentrates on describing what constitutes the specifically imperative character of imperatives. Immediately after his rather brief elucidation of the concepts of ‘will’ and ‘reason,’ Kant goes on to explain the imperatival character of rational laws. We have already pointed out, in connection with the transition from the concept of will to the concept of duty (GMS I:397), just how important this notion is for Kant. The basic idea is quite simple: Rational principles of action are objective rules for action in accordance with which beings that are purely rational or possess a ‘pure will’ always act or would act. Like all beings susceptible to the influence of inclinations and interests, human beings do not always act rationally. For such beings, therefore, those objective rules for action are of an essentially imperatival nature. They possess a normative character; they demand something in the face of potential resistance; they prescribe something: “All imperatives are expressed through an ought and thereby indicate the relation of an objective law of reason to a will which in its subjective constitution is not necessarily determined by that law (a necessitation)” (413:12). The importance of this idea for GMS as a whole can hardly be overestimated. We should therefore specifically consider all the relevant passages of the text in this regard: 1. “A maxim is the subjective principle of the volition; the objective principle (i.e. that which would serve all rational beings also subjectively as a practical principle if reason had full control over the faculty of desire) is the practical law” (400:34, footnote; second emphasis ours). 2. “If reason determines the will without exception, then the actions of such a being, which are recognized as objectively necessary, are also subjectively necessary, i.e. the will is a faculty of choosing only that which reason, independently of inclination, recognizes as practically necessary, i.e. as good. But if reason for itself alone does not sufficiently determine the will, if the will is still subject to subjective con-

Section II of the Groundwork

3.

4. 5.

6.

7.

8.

105

ditions (to certain incentives) which do not always agree with the objective conditions, in a word, if the will is not in itself fully in accord with reason (as it actually is with human beings), then the actions which are objectively recognized as necessary are subjectively contingent, and the determinations of such a will, in accord with objective laws, is necessitation, i.e. the relation of objective laws to a will which is not thoroughly good is represented as the determination of the will of a rational being through grounds of reason to which, however, this will in accordance with its nature is not necessarily obedient” (412:30). “A perfectly good will would thus stand just as much under objective laws (of the good), but it would not be possible to represent it as necessitated by them to lawful actions, because of itself, in accordance with its subjective constitution, it can be determined only through the representation of the good. Hence, for the divine will, and in general for a holy will, no imperatives are valid; the ought is out of place here, because the volition is of itself necessarily in harmony with the law” (414:1). “The will whose maxims necessarily harmonize with the laws of autonomy is a holy, absolutely good will” (439:28). “Our own will, insofar as it would act only under the condition of a possible universal legislation through its maxims, this will possible to us in the idea, is the authentic object of respect” (440:7). “This ‘ought’ is really a volition that would be valid for every rational being, under the condition that reason were practical in him without any hindrances; for beings, such as we are, who are also affected through sensibility, as with incentives of another kind, with whom what reason for itself alone would always do does not always happen, that necessity of action is called only an ‘ought,’ and the subjective necessity is different from the objective” (449:16). “As a mere member of the world of understanding, all my actions would be perfectly in accord with the principle of the autonomy of the pure will” (453:25). “The idea of freedom makes me into a member of an intelligible world, through which, if I were that alone, all my actions would always be in accord with the autonomy of the will” (454:6).

106

Immanuel Kant’s Groundwork for the Metaphysics of Morals

9. “The moral ‘ought’ is thus his own necessary volition as a member of an intelligible world” (455:7).10 From all this it is clear at once how important the imperatival character of rational laws is for Kant, and how this imperative character is connected with the objectivity of such laws. What we ought to do is what rational beings will; for purely rational beings, therefore, there can be no imperative at all. But that also implies that what we ought to do is what as rational beings, which we also are, we will to do; if we were purely rational rather than sensuous-rational beings, then rational laws would not be imperatives for us either. We could also elucidate this thought as follows: The moral law describes how a purely rational being acts. The actions of a purely rational being are therefore always such that the underlying maxims of the actions are universalizable (to use this formulation here); beings of this kind and their acts can reliably be described by means of this predicate. As a categorical imperative, by contrast, the moral law prescribes for sensuous-rational beings how they ought to act: act in such a way that the maxim of your action is universalizable. (We shall soon see that this is connected with Kant’s characterization of the CI as synthetic a priori in character.) One might think that the ‘necessitation’ bound up with the ‘ought’ involved in moral laws is simply a feeling that arises when such laws are observed.11 But that cannot be right, for then the observance of an objec10. In his lectures on natural law of 1784 (at the time when he was preparing GMS) Kant explored the concept of a perfect will in some detail: “If the will of a being is good in itself, then the objective laws of its will are not distinguished from subjective laws.—The human will is not of such a kind that the subjective grounds of volition accord with the objective ones. Now the objective rule of volition, applied to a will whose subjective rules do not accord with the objective ones, is called imperative” (NF: 1323 ). In the Mrongovius lectures on moral philosophy, which date from around the same period, Kant writes that “imperatives are derived from the idea of a perfect will and are binding as rules for my imperfect will; duty is an idea of a perfect will as a norm for an imperfect will” (MM II:606; o.e.). With regard to the concept of the ‘ought,’ and thus the concept of duty, the significance of which is central to GMS I, Kant writes in the same lectures: “The ‘ought’ signifies the notion that a possible free act of my own would necessarily transpire if reason completely controlled our will” (MM II:605; o.e.). 11. For the concept of ‘necessitation,’ cf. 413:4; 413:10; 413:15; 414:3; 416:6; 417:5; 425:30; 434:17; and 439:31.

Section II of the Groundwork

107

tive law without a sense of reluctance or ‘unwillingness’ (if one may put it this way) would not involve such necessitation. But for Kant, objective rational principles as imperatives are always ‘necessitating.’12 From a certain perspective necessitation cannot be distinguished from necessity (in the sense of the binding character of a law) as far as sensuous-rational beings are concerned. That is why in a moral context Kant understands “obligation” (439:31) to mean “the dependence of a will that is not absolutely good on the principle of autonomy (moral necessitation)” (439:30; o.e.).13 The “objective necessity of an action from obligation is called duty” (439:33), and duty itself is thus even identified (“i.e.”) as “practical necessitation” (434:17). With regard to purely rational beings, rational laws have the character of necessity as well. But for beings of this kind, actions “which are recognized as objectively necessary are also subjectively necessary” (412:31; o.e.). For sensuous-rational beings, “the actions which are objectively recognized as necessary are subjectively contingent” (413:2; o.e.). It is simply this aspect that Kant expresses with the concept of necessitation. Hence when he introduces this concept, he does not defi ne it directly but merely describes it as a “relation” (413:4).

3.2.2 Categorical and Hypothetical Imperatives After Kant has explained what imperatives as such are (412:26–414:11), he proceeds to a closer examination and division of them (414:12–417:2): “Now all imperatives command either hypothetically or categorically” (414:12). Kant then further divides hypothetical imperatives into what he calls imperatives or rules of skill (415:13; 416:19) and imperatives or counsels of prudence (416:2; 416:19). He also describes the rules of skill as

12. But in the absence of ‘subjective necessity’ the constraining character of moral laws is obviously important. Moral necessitation is a “determination to actions however reluctantly they may be done” (KpV:80); it is a kind of “self-constraint, that is, inner necessitation to what one does not altogether like to do” (KpV:83– 84). Cf. also MSV:489 and MC:256. 13. In this sense Kant writes elsewhere that the “ ‘ought’ . . . expresses objective necessitation to the action” (KpV:20). Cf. also KpV:32, where we read that obligation “signifies” a necessitation; later Kant basically identifies necessitation and obligation or moral necessity (KpV:81).

108

Immanuel Kant’s Groundwork for the Metaphysics of Morals

problematical (415:1) or technical (416:29) imperatives, and the counsels of prudence as assertoric (415:1) or pragmatic (417:1) imperatives. Kant also characterizes the CI in turn as the law of morality [Sittlichkeit] (416:13; 416:20), which is also described as apodictic (415:4) or moral (417:1).

Categorical-Moral Imperatives of Morality We have already indicated the basic idea of categorical imperatives in our initial presentation of Kant’s metaphysics of morals (MS1 ) and in our analysis of what it means to act from duty. To describe imperatives as categorical simply means that their binding and obligatory character is independent of our subjective interests and inclinations; and in fact on pages 412–417 Kant tells us nothing more than this, namely, that the CI “is not limited by any condition” (416:27).14 What is new, however, is Kant’s thesis that the CI is a synthetic-practical a priori proposition. This too is connected with the fundamental distinction between purely rational beings and sensuous-rational beings: “I connect the deed a priori with the will, without a presupposed condition from any inclination, hence necessarily (although only objectively, i.e. under the idea of reason, which would have full control over all subjective motivations). This is therefore a practical proposition that does not derive the volition of an action analytically from any other volition already presupposed (for we have no such perfect will), but is immediately connected with the concept of the will of a rational being, as something not contained in it” (420:29, footnote).15 Whatever a purely rational being may will, its volition or willing will never stand in contradiction to the moral law. For a being of this kind, therefore, the moral law is neither an imperative nor a duty but simply an analytic-descriptive proposition. For sensuous-rational beings, the situation is quite different. For them, the moral law is a CI and thus, according to Kant, a synthetic a priori proposition. The moral law is an imperative because it is an objective law that involves a necessitation for those whom it addresses. This imperative is

14. Later we are told that “the withdrawal of all interest in the case of volition from duty . . . is indicated as the specific sign distinguishing the categorical from the hypothetical imperative” (431:35; o.e.). 15. Cf. 440:24; 440:26; 444:35; 447:10; 447:14; 454:11; and 454:17.

Section II of the Groundwork

109

categorical because it commands an action as something absolutely necessary without thereby presupposing a subjective interest in the action; the CI is a priori on account of its categorical validity and necessity. The imperative is synthetic because it first “connects” (420:35; o.e.) what is not ‘analytically’ contained elsewhere; that is, it connects the will of an imperfect being with the moral law. The synthetic a priori character of the CI thus consists simply in the fact that sensuous-rational beings that do not always will the morally good ought to will it, and that without being allowed to presuppose any inclination or interest in this regard. Thus in this sense the will of a sensuous-rational being is ‘connected’ with the law of morality through the CI. The law of morality, on the other hand, as a moral law as such, is already connected with the will of a perfectly rational and free being. It belongs to the concept of a perfectly rational being that its inclinations and desires, insofar as it has any, do not determine its actions; it is the representation of the moral law alone that is relevant to the action of such a being. For such a being, the moral law is not a synthetic but an analytic a priori proposition.16

Hypothetical Imperatives The basic idea of hypothetical imperatives is also easy enough to grasp: they “represent the practical necessity of a possible action as a means to 16. One must therefore bear in mind that this talk of the synthetic a priori character of the CI can be understood only in an analogical sense. This merely analogical way of speaking is particularly evident in two of the passages where Kant speaks specifically of the a priori synthetic character of the CI (440:18–24 and 447:10–25). What is ‘connected’ here is the “will” (420:29), namely, the “will affected through sensible desires” (454:12) as a “member of the world of sense” (454:10), and the “deed” (420:30, footnote), namely, the “volition of an action” (420:32, footnote) as the volition of “a will . . . belonging to the world of understanding, a pure practical will, practical for itself ” (454:13 ). This is the ‘connection’ at issue. But all these passages also make it clear that we are not talking about a predicate that is supposed to be connected with a subject. Kant himself says that the will is “bound” to the practical rules (440:22). This act of synthesis therefore is something that transpires with the CI rather than something contained in it. The CI itself is connected with the will of a sensuous-rational being; this imperative requires the connection between the moral law contained in it and a will of this kind. Perhaps Kant was also well aware of the analogical character of this language: after all, on two occasions he specifically speaks of synthetic-practical a priori propositions (420:14 and 444:35).

110

Immanuel Kant’s Groundwork for the Metaphysics of Morals

attain something else that one wills (or which it is possible that one might will)” (414:13). Just like categorical imperatives, hypothetical imperatives are also objective and necessitating obligatory laws of reason. They are objective because they are based on rational knowledge concerning the relation between a particular end and the corresponding means. Kant expressly emphasizes that the determination of the means is not identical with the hypothetical imperative.17 Even if one does not pursue a hypothetical imperative because the end for which it prescribes the means is not pursued at all, the underlying proposition regarding the relation of ends and means is correct. The proposition ‘The division of a line into two equal parts is accomplished by the construction of two arcs at the endpoints of this line’ is true and objectively valid irrespective of whether I actually intend to divide a line into two equal parts. Hypothetical imperatives are obligatory laws because, like all imperatives, they are addressed to agents who do not always do and will to do what they rationally ought to do and will. Someone who pursues the aim of playing the piano well must practice. But not everyone who wants to play the piano well also (always) wants to practice. Imperatives are hypothetical because they command an action that one must accomplish if and only if one in fact pursues the end that is to be attained thereby: “I ought to do something because I will something else” (441:10; cf. 444:11). A hypothetical imperative is thus hypothetical insofar as it is subject to a condition: “The imperative is conditioned, namely: if or because one wills this object, one ought to act thus or so” (444:3). Kant’s conception of hypothetical imperatives is best understood when we distinguish between the formula of a hypothetical imperative in general and the specific (concrete) imperative. A concrete hypothetical imperative is a directive for action, for example, ‘Construct two arcs at the endpoints of a line.’ The corresponding hypothetical imperative commands an action or the volition of an action,18 where the action consists in adopting the means that are necessary in order to realize the end that has 17. “Synthetic propositions belong to determining the means themselves to a proposed aim, but they have nothing to do with the ground, with making the act of the will actual, but rather with how to make the object actual” (417:15). 18. For the (unproblematic) distinction here, cf. 414:32; 417:4; 417:14; 418:29; 418:29; 419:9; and 444:4.

Section II of the Groundwork

111

already been willed and are available to the one who is addressed by the imperative. The formula of a hypothetical imperative can be expressed as follows: It must first be clear that only certain particular means permit the realization of a quite particular end; this determination of the meansend relation lies at the basis of the hypothetical imperative. Then it must also be clear that the particular means are actually available to the person in question; the means required of someone must be such that they are “in his control” (417:31). Finally, and this is the decisive point, it must be clear that the end in question is actually pursued by someone. If therefore (1) the relation between the end p and the means q is such that only the adoption of q permits the realization of p, (2) one is able to adopt q, and (3) one wills p, then one ought to adopt q; we can describe this as the formula of the hypothetical imperative. Thus (1) if it is true that “the division of a line into two equal parts is accomplished by the construction of two arcs at the endpoints of this line,” (2) if this construction is possible for the addressee, and (3) if the addressee wills the division of the line (pursues this as an end), then the concrete directive regarding the means directly follows: ‘Construct two arcs at the endpoints of a line.’ The concrete directive regarding the means can thus be derived from the formula of the hypothetical imperative and the conditions that are actually fulfi lled.19 In order to simplify the structure of the argument and bring out the decisive point, we can abstract (with Kant) from both initial conditions and frame the formula of the hypothetical imperative as follows: ‘If one wills a particular end, one ought to will (adopt) the necessary means.’ This abbreviated formula captures what Kant calls a ‘hypothetical imperative.’

Imperatives of Skill and Imperatives of Prudence An assertoric-pragmatic imperative of prudence is distinguished from a problematic-technical imperative of skill “only in this, that with the latter 19. It is easy to see that the hypothetical imperative is described as ‘hypothetical’ solely on account of this conditioned necessity and not on account of its semantic or logical form; the same is also true of categorical imperatives. Categorical imperatives can be expressed in a hypothetical form (“If you have given a promise, you must keep it!”) and hypothetical imperatives in a categorical form (“Practice at the piano!”), although this would provide only an incomplete formulation for such hypothetical imperatives.

112

Immanuel Kant’s Groundwork for the Metaphysics of Morals

the end is merely possible, but with the former it is given” (419:5). This initially appears to be a perfectly clear distinction: technical imperatives are ‘problematic’ insofar as the particular end at issue in each case is merely ‘possible’ (‘discretionary’ [415:22]). Not all human beings pursue the corresponding end (being able to play the piano, for example), and it is only for those who do pursue this end that the corresponding hypothetical imperative or the concrete directive regarding the means is relevant.20 Imperatives of prudence, on the other hand, are assertoric insofar as the end in question can be presupposed in all human beings (all human beings, according to Kant, strive after happiness). This seems plausible enough, but assertoric-pragmatic imperatives actually create certain difficulties. Kant understands happiness to mean “a maximum of welfare . . . in my present and in every future condition” (418:8), where this ‘welfare’ arises from the “satisfaction” (399:12; cf. 405:8) of all our needs and inclinations. Happiness is thus no momentary satisfaction of need. It is a “sum” (399:9; 399:12; cf. 394:17), a united satisfaction of all needs and inclinations, and indeed ‘in my present and in every future condition’; all present and future needs must be included in the calculation. This defi nition of ‘happiness’ is unambiguous. But precisely because happiness is understood in this way, the concept itself, as soon as one tries to fill it out, remains “an indeterminate concept” (418:2) and “a wavering idea” (399:16).21 The problem is not that one cannot define what happiness means as an “ideal” (418:36) (as we have seen, it is the satisfaction of all present and future needs). The ‘indeterminacy’ of the concept lies in the fact that one cannot possibly indicate which ‘elements’ of happiness, that is, which means, to which extent, and in which relation to one another, relative to the actual wishes of an individual and over time, can facilitate a balanced and enduring satisfaction of needs.22 For this reason Kant writes that “although

20. Later Kant expressly rejected the use of the term ‘problematic’; cf. the “First Introduction” to the Critique of the Power of Judgment (AA 20:200). 21. Kant frequently emphasizes this point, especially in relation to what we will call the question of ‘possibility’; cf. 399:10–26 and 417:27–419:11, but also the teleological argument (394:32–396:37). 22. Cf. Kant’s examples at 418 (wealth, knowledge, longevity, health). Thus, according to Kant, the “private prudence” (416:31) at issue here also consists in someone’s ability “to unite all these [his] aims to his own enduring advantage” (416:33; o.e.).

Section II of the Groundwork

113

every human being wishes to attain it [happiness], he can never say, determinately and in a way that is harmonious with himself, what he really wishes” (418:2; o.e.) because to do so would require “omniscience”(418:24).23 Even if all human beings shared more or less the same ideas in this regard, it would still be impossible to determine in any precise way the “means to happiness” (419:4). But human beings also have their own individual ideas. For these reasons, therefore, it is also impossible to set up imperatives of prudence as universal laws for all rational beings (418:28–419:2). It is impossible to provide anything more than “counsels” (416:19) backed by “experience” (418:6; 418:27) that “most promote welfare on the average” (418:27; o.e.). Thus despite Kant’s repeated talk of ‘imperatives of prudence,’ the fact is that there are no such imperatives, strictly speaking; general rules of prudence “to speak precisely, cannot command at all, i.e. cannot exhibit actions objectively as practically necessary; that they are sooner to be taken as advising (consilia) than as commands (praecepta) of reason; that the problem of determining, certainly and universally, what action will promote the happiness of a rational being, is fully insoluble, hence no imperative in regard to it is possible” (418:29). Assertoric imperatives are therefore not imperatives at all. But are they assertoric? We said that with problematic imperatives the end is ‘merely possible,’ whereas with assertoric imperatives it is ‘given.’ According to Kant, “All human beings always have of themselves the most powerful and inward inclination to happiness” (399:7; o.e.). The aim of happiness is one regarding which we “can safely presuppose” that sensuous-rational beings “do have it in accordance with a natural necessity” (415:31; o.e., Kant’s emphasis omitted); happiness is a “natural end” (430:19), an “aim that one can presuppose safely and a priori with every human being, because it belongs to his essence” (415:37; second emphasis ours).24 That certainly seems plausible, as we said, for who would not strive after his or her own happiness? But what precisely does this mean? It is certainly true that all human beings strive after certain “gifts of fortune” (393:13) or “elements” of happiness

23. Cf. 418:11 (“what he really wills here”; o.e.) and 418:23 (“what will make him truly happy”; o.e.). 24. In the fi rst edition of GMS (1785) Kant wrote “nature” rather than “essence” here.

114

Immanuel Kant’s Groundwork for the Metaphysics of Morals

(418:5). Kant is not interested in this trivial fact, but neither can he think that all human beings in fact always consider and hold up happiness as a ‘sum’ before them when they act, for that is certainly (empirically) false. In the example of the man suffering from gout (399) it is even assumed that “the general inclination to happiness does not determine his will” (399:21; o.e.). Rather, he ought to “secure” (399:3) and “promote” (399:25) his happiness; this is a “law” (399:24) and his (indirect) “duty” (399:3).25 But then what is the meaning of the claim that he would, like all other human beings, strive after happiness in accordance with a ‘natural necessity’? To begin with, it is not the case that the sufferer from gout in the given situation does not strive after happiness at all. Rather, he “reckons” (399:18; 399:23) with the whole situation; that is, he makes a rough cost-advantage calculation and decides to choose the “enjoyment of the present moment” (399:20; o.e.) over the ‘wavering idea’ of happiness.26 The point is not that the gout sufferer does not will his happiness at all, nor that his sense of instrumental rationality fails him (for he does not choose the wrong means for attaining happiness; rather, he chooses the right means for satisfying his momentarily presupposed end). The point is that he does not will his happiness strongly enough but sacrifices it to a momentary satisfaction of need. Happiness for Kant is a normative a priori principle of prudential reason. Reason commands us to set happiness as an end, and whoever fails in this regard is responsible for a failure of his or her prudential reason. Prudence does not (simply) consist in choosing the right means to attain one’s happiness. It is quite true that Kant defi nes prudence as “skill in the choice of means for one’s own happiness” (416:1), and certainly we would not describe a person as prudent if that person always acts in a clumsy and inappropriate way in the choice of means (difficult as they are to determine) to attain his or her own happiness. But this choice makes sense only if a rational, normatively guided striving for happiness is presupposed here. Only then does the skill in question have an aim,

25. Cf. MS:387–388. 26. One should note that it is because (daher, 399:13 ) of the lack of certainty regarding the determination of the means in imperatives of prudence that the gout sufferer privileges a par tic u lar satisfaction of need over his happiness in general.

Section II of the Groundwork

115

and prudence must consist fi rst and above all in firmly envisaging one’s own happiness as an end encompassing life as a whole and in evaluating par tic u lar ends with a view to this encompassing aim. One who strives after happiness acts in such a way that the various ends and the means associated with them should yield a good “sum” (of happiness). Those who show skill in choosing the means to their own happiness basically do just this. They act in such a way that all their ends and all the means that they adopt in realizing their par ticu lar ends, taken as a whole, yield happiness. Kant’s reference to the ‘natural necessity’ with which we strive for happiness is rather misleading and remains somewhat problematic. It should really be understood as the claim that we all do in fact strive after happiness, but that at the same time on account of the weakness of our nature we ought to strive after it. Happiness is an aim that reason sets for us.

3.3 The Possibility of Hypothetical Imperatives On page 417 Kant asks: “How are all these imperatives possible?” (417:3). First, Kant addresses the question how hypothetical imperatives are possible. This question—let us call it the question of possibility—is concerned with the validity of hypothetical imperatives (we shall see later that the parallel question, namely, how a categorical imperative is possible, also asks, among other things, why categorical imperatives are valid or binding). The first question asks how we are to “think the necessitation of the will that the imperative expresses in the problem” (417:4; o.e.). Since the ‘necessitation’ that belongs to an imperative simply signifies its necessity for sensuous-rational beings, and since necessity is synonymous with validity, the question thus amounts to this: how can the necessity (the validity, the obligatory character) of the hypothetical imperative be thought? Alternatively, how can we ground the claim that someone who wills an end ought also to will and to adopt the means that are necessary for the realization of this end?

The Basic Idea behind Kant’s Response Before we examine Kant’s response to this question in detail, we should note a few points that can be textually substantiated without further ado.

116

Immanuel Kant’s Groundwork for the Metaphysics of Morals

1. The argument turns on a proposition that Kant twice describes as “analytic” (417:11; 417:23). Kant provides three formulations (which he clearly regards as equivalent) of what we shall call the end-means proposition (EMP): EMP1: “Whoever wills the end, also wills the means (insofar as reason has decisive influence on his actions)” (417:8). EMP2: “But that if I know that the specified effect can occur only through such an action, then if I completely will the effect, I would also will the action that is required for it—that is an analytic proposition” (417:21). EMP3: “Whoever wills the end, also wills (necessarily in accord with reason) the sole means to it in his control” (417:30). The basic form of the proposition is the same: whoever wills the end also wills the means. But the qualifi cation concerning rationality that is incorporated into all three formulations is differently expressed in each case (EMP1: ‘insofar as reason has decisive influence on his actions’; EMP2: ‘if I completely will’; EMP3: ‘necessarily in accord with reason’). For the sake of simplicity, we shall assume that these three formulations do mean the same thing. We shall therefore take EMP1 as the basic formulation (EMP for short).27 It is obvious that the EMP cannot possibly be identical with the hypothetical imperative. The hypothetical imperative commands something, while the EMP does not. The EMP tells us that whoever wills an end also rationally wills the means. The hypothetical imperative, on the other hand, as we have cited it earlier, tells us that ‘I ought to do something because I will something else.’ 2. Kant describes hypothetical imperatives, as well as the EMP, as “analytic” (417:29; 419:10).28 We do not need to clarify, at this point, what Kant means by this in order to see that he regards hypothetical imperatives as analytic because the EMP is analytic, for the thesis that hypothetical

27. The qualifying remark in EMP2 that reads ‘If I know that the specified effect can occur only through such an action’ is not part of the qualification concerning rationality; we shall ignore this qualifying remark concerning ‘knowledge’ here. 28. Cf. 417:29.

Section II of the Groundwork

117

imperatives are analytic is unambiguously grounded by reference to the EMP (417:30: “For here, as there” [Denn]). 3. The EMP is introduced directly after Kant’s claim that the question concerning the possibility of technical imperatives “needs no par ticular discussion” (417:7). It is thus in relation to EMP and its analytic character that Kant identifies the analytic character of hypothetical imperatives as the reason why the question of their possibility poses “no difficulty” (419:11). Thus the answer to the question ‘how are hypothetical imperatives possible?’ is: because they are analytic.

Analysis and Criticism of Kant’s Response to the Question of Possibility Thus Kant’s three basic theses here are these: (1) Whoever wills the end also wills the means (insofar as reason has decisive influence on that person’s actions); this proposition (the EMP) is analytic. (2) Since the EMP is analytic, the hypothetical imperative is also analytic. (3) Since the hypothetical imperative is analytic, the question concerning the possibility of such imperatives raises no problems. Let us now examine these theses individually in more detail. Ad 1: Kant grounds his thesis that the EMP is analytic twice with the same argument. The EMP is analytic, according to Kant, “for in the volition of an object, as my effect, is already thought my causality as an acting cause, i.e. the use of means” (417:11; o.e.), and “for to represent something as an effect possible through me in a certain way and to represent myself, in regard to it, acting in this same way—those are entirely the same” (417:24; o.e.).29 At first it seems clear that in both arguments Kant has the familiar logico-semantic meaning of the term ‘analytic’ in mind. In that case, the predicate of an analytic proposition merely expresses what the subject of the proposition already contains: analytic propositions “express nothing in the predicate but what has already been thought in the concept of the subject, though not so clearly and with the same consciousness” (PM:266). In this sense, Kant seems to be saying that ‘in’ the volition of the end the volition of the means is ‘already thought,’ or even that the volition of the

29. One should not be concerned by the fact that in both of these justificatory remarks (introduced by “for”) we are talking only about volition in general; the clause referring to rationality is naturally implied here as well.

118

Immanuel Kant’s Groundwork for the Metaphysics of Morals

end and the volition of the means are ‘entirely the same.’ But if the volition of the means is contained in the volition of the end and is merely analytically explicated in the EMP, then it actually seems to belong to the meaning of volition of the end that volition of the means is already thought in it. But precisely how is that to be understood? It might seem plausible to reply that Kant already wants us to understand volition in the EMP as ‘rational volition.’ In that case, we can talk about the volition of an end at all only when the one who wills the end also wills the required means. Someone who claims, for example, that he wills to learn how to play the piano but refuses to will the practicing involved does not on this interpretation really (truly, seriously) will to learn how to play the piano at all. On this view, the true preferences are in a certain sense revealed only subsequently; it is only when the means are truly and seriously adopted that one sees who actually wills or has willed what. Since the one who in willing the end also wills the required means is rational, the volition in the EMP would thus have to be understood as rational volition; volition is rational, or it is not volition at all (but merely a wish). From this perspective one would formulate the EMP as follows: whoever rationally wills the end also wills the means; and this would indeed be an analytic proposition, for ‘to rationally will an end’ would also be to will the required means. Although the analytic character of the EMP would easily be explained in this way, this reading is nonetheless problematic both philosophically and interpretatively. There are two fundamental problems with it. First, it is central to Kant’s theory of rational volition and action that human beings do not always will or act rationally. If this were not the case, there would be no imperatives at all. Thus in the context of technical and pragmatic rationality too it must be possible to act irrationally, or indeed to will irrationally. The question how hypothetical imperatives are possible is concerned specifically with answering why one ought to act rationally in this context. If it were impossible to will an end without willing the required means, the question why one ought to will (and adopt) the means would be meaningless, for then one would not really will (not will at all), and the question of possibility would not arise in the fi rst place. A hypothetical imperative also indicates “which action possible through me would be good, and represents the practical rule in relation to a will which does not directly do an action because it is good, in part because the subject does not always know that it is good”—we shall

Section II of the Groundwork

119

ignore this case here—“in part because if it did know this, its maxims could still be contrary to the objective principles of a practical reason” (414:26). On this view, therefore, it is possible to will an end without willing the required means. For anyone who claims that volition is identical with rational volition, there can be no irrationality in volition at all, which is highly implausible. Second, if we were simply to identify ‘volition’ with ‘rational volition,’ not only would there no longer be any irrationality in the practical sense, but also there would no longer be any imperatives. For if the volition of the means were actually analytically contained (in the logico-semantic sense) in the volition of the end, there could be no such thing as a hypothetical imperative, which represents this volition of the means as something that one fi rst ought to will, for one wills it already because one already wills the end (and the volition of the end is already presupposed in the case of hypothetical imperatives). But according to Kant’s own words, the volition of the means is precisely what the imperative commands in the fi rst place. As Kant writes: “The imperative that commands the volition of the means for him who wills the end is . . . analytic” (419:8; o.e.). We should thus assume that in the EMP Kant does not simply define ‘volition’ as rational volition. Not only would this fail to cohere with his overall conception, but it would also contradict his formulations of the EMP. In EMP1 Kant specifically refers to the ‘decisive influence’ of reason, which implies that there must also be another kind of influence (i.e., that of ‘incentives’); and to speak of ‘completely willing’ something (as he does in EMP2) makes sense only if an incomplete willing or volition is also possible. That is why Kant also says in EMP1 that anyone who wills the end, ‘insofar as reason has decisive influence on his actions,’ also wills the means. What the hypothetical imperative commands is to be rational in one’s willing, that is, to will (and so to act) in such a way that in willing the end one also wills the means; but that presupposes that one can also act in an irrational way. Kant does not claim that volition is, by defi nition, rational. What he says, in the EMP, is simply that if someone wills an end and this person is rational, then he or she also wills the end. But what precisely does the ‘decisive influence of reason’ mean here? It is necessary to recognize that the hypothetical imperative is already incorporated into the EMP through the qualifying references to rationality, for one who sets an end thereby simultaneously sets oneself the norm and

120

Immanuel Kant’s Groundwork for the Metaphysics of Morals

the obligation to pursue the corresponding means as well. The act of setting the end thus simultaneously contains the normative self-established act of willing and having to adopt the corresponding means as well. One who sets oneself an end (wills the end) but does not adopt the corresponding means does not act in accordance with the norm that he or she has set for himself through the very act of setting ends; that person acts irrationally. The ‘decisive influence’ of reason lies in successfully commanding the hypothetical imperative. The EMP describes a person who wills something (that is, an end), who thereby sets himself or herself a norm (that is, of willing the means), and whose reason actually exercises the ‘decisive influence,’ so that this person also wills the means. One could thus formulate the EMP as follows: someone who wills the end, if that person obeys the hypothetical imperative, also wills the end. And this is certainly an analytic proposition. Ad 2: This brings us to the second point, namely, Kant’s thesis that hypothetical imperatives are analytic. We have said that one who sets an end thereby simultaneously sets oneself the norm of acting in such a way that the end is realized. Someone who says, “I will this end,” thereby at the same time says to oneself, “I ought to act in such a way that I can attain this end”; and since this action essentially consists in adopting particular means, one who sets oneself an end commands oneself to adopt the corresponding means. It thus belongs to the meaning of the act of setting oneself an end that one thereby commands oneself to will the corresponding means as well. That does not mean that one wills the means when one wills an end; it means, rather, that one ought to will the means. The hypothetical imperative is thus analytic in the sense that a normative principle— namely, the hypothetical imperative itself—is inseparably connected with the activity of setting ends. Ad 3: We can therefore also now accept Kant’s thesis that hypothetical imperatives are possible because they are analytic, for if the act of setting oneself an end is analytically connected with setting oneself the norm of adopting the necessary means, this also grounds why one ought to adopt the means. This is what Kant means when he writes, “The [hypothetical] imperative extracts the concept of actions necessary for this [given] end out of the concept of a volition of this end” (417:13; o.e.), or again, “Since, however, both [types of hypothetical imperatives] merely command the

Section II of the Groundwork

121

means to that which it is presupposed that one wills as an end, then the imperative that commands the volition of the means for him who wills the end is in both cases analytic” (419:7; o.e.). The ‘since . . . then’ structure indicates that Kant is giving us the reason that hypothetical imperatives are ‘in both cases analytic.’ This reason lies solely in the presupposed end and the normative act connected with it. This fits with the fact that in addressing the question regarding the possibility of the categorical imperative, Kant writes that the latter is “without doubt the sole question in need of a solution, since it is not all hypothetical and thus the necessity, represented as objective, cannot be based on any presupposition, as with the hypothetical imperatives” (419:12; o.e., Kant’s emphasis omitted). This implies that the validity of hypothetical imperatives and thus the answer to the question of their possibility is ‘based’ precisely on such a ‘presupposition’ (that an end is willed). Thus the answer to why one who wills the end ought to will the corresponding means as well is this: because a self-given norm is involved in the act of setting ends. A resolute skeptic might protest here: “I concede that if one is rational, one observes the rule of adopting the means when one wills an end. But if I inquire into the validity of the hypothetical imperative, what I want to know is specifically why I should do what I ought to do; why ought I to be rational?” But this question is meaningless (as meaningless as the question why one ought to do what one ought to do), for if the skeptic admits that it is rational to adopt the corresponding means to an end, presupposing that one wills the end, then that skeptic already has the very reason she or he is seeking. That it is rational to act in a particular way simply means that one has a good reason for so acting, and then it no longer makes sense to ask for a reason. It may be that some people (irrationally) refuse to respond to the reasons they have to act in a certain way (morally, prudently, or with instrumental rationality) by being motivated to do so; and they may express their (irrational, self-deceptive) state of mind by asking: “Why should I do what is rational?” Further, the correct response to their question—that it makes no sense—may not give them the motivation they lack. But that is no defect in the response. The only defect is in them. Kant’s thesis that the EMP is analytic is correct. His thesis that the hypothetical imperative is analytic is also correct, and to that extent the question of possibility is unproblematic. However, Kant is wrong in one

122

Immanuel Kant’s Groundwork for the Metaphysics of Morals

respect: hypothetical imperatives are not analytic simply because the EMP is analytic. The central thought behind Kant’s argument is that setting ends is a normative act. The hypothetical imperative is prescriptive: it commands something. The EMP, on the other hand, only describes, through the qualifying references to rationality, what happens if the normative act connected with the setting of ends— namely, a hypothetical imperative—is in fact successful: if a person wills an end and that person is rational, then he or she also wills the means. Thus the EMP describes what a rational being does in a purposive context (i.e., willing the means if one wills the end). This presupposes the hypothetical imperative, the analytic character of which cannot therefore be derived from the analytic character of the EMP. We have discussed the hypothetical imperatives in such detail here because they already clearly show that practical reason accomplishes more than simply determining the means for ends that are already presupposed. According to a Humean model of practical rationality, actions essentially arise from wishes and beliefs: wishes motivate us to a par ticular action with regard to a particular end, and reason merely supplies the beliefs that identify which action one ought to take to accomplish this end. On this model of practical rationality the task of reason is limited to providing theoretical knowledge regarding the relations of ends and means, so that a wish (for a certain end) gives rise to another wish (for the means of attaining this end). But in fact, at least according to Kant, instrumental and prudential reason is already normative. Its task is not merely the theoretical one of determining means. Rather, it provides direction for willing insofar as it instructs us to pursue our self-given ends in a rational way and keep them steadfastly in view. And categorical imperatives show that practical reason is normative in a much stronger sense and accomplishes much more than simply determining relations of means and ends, for practical reason, according to Kant, must also be recognized as pure practical reason.

3.4 The Formulations of the Categorical Imperative and Kant’s Examples Kant’s various formulations of the CI and the examples of it he provides have given rise to much discussion. We shall begin with a brief overview

Section II of the Groundwork

123

of these different formulations and their plurality (3.4.1). Then we shall look at Kant’s attempt to derive particular duties from these formulations, and thus also at his par tic u lar examples (3.4.2). Kant illustrates this derivation solely by reference to the formula of universalizability and the formula in terms of a law of nature (3.4.2.1) and the formula of ends in themselves (3.4.2.2). We undertake to examine and evaluate these two formulas in close detail. Then we proceed to a discussion of further elements of the metaphysics of morals: autonomy and the realm of ends (3.4.2.3). Finally, we offer a concluding overview of the various formulations and their relationship with one another (3.4.3).

3.4.1 Enumerating the Various Formulas Kant’s references to the CI, his various formulas, and their relationship with one another are scattered throughout the entire text of GMS. Let us briefly recall and clarify the structure of GMS II: when Kant distinguishes the different types of imperatives (412:26–417:2), he speaks quite generally and almost invariably of the CI.30 But he also speaks of moral “commands” or “laws” (416:20), as well as of moral “imperatives” (417:1).31 That need not worry us unduly; these expressions simply refer to the various duties that can be derived from the CI.32 One should not confuse, however, the difference between the imperative (singular) and the imperatives (plural) with the difference between the general formula of the CI (which is still to be clarified) and the various formulas (plural) that are connected with it. The CI is first formulated at 421, although it is immediately supplemented with a formula that speaks in terms of a law of nature. After the transition to the metaphysics of morals (427), Kant introduces the formula that speaks of ends in themselves. At 431 Kant provides his first summary and, by including the formula of autonomy, speaks of three formulas of 30. Cf. 414:15 (“the,” singular); 414:25 (“it”); 415:2 (“the,” singular); 416:7–14 (“a,” “the,” ‘singular’ etc.); and 416:26 (“the,” singular). In subsequent passages as well Kant almost always refers to the CI in the singular. 31. Kant speaks of moral imperatives in the plural at 421:9; 425:3; 431:25; 454:4; and 454:6; he also speaks correspondingly of moral laws in the plural (as well as of ‘the’ moral law). 32. Cf. 421:9–10.

124

Immanuel Kant’s Groundwork for the Metaphysics of Morals

the CI. The formula of autonomy then gives rise to the formula of the realm of ends (433). When he offers an overview at 436–437, Kant speaks once again of three formulas before finally again describing their relationship to one another (437ff.). The most important passages are 431:9–18 and 436:6–437:4. It is clear that at 431 Kant enumerates at least three different formulas: the formula of the law of nature (FLN), the formula of humanity as end in itself (FH), and the formula of autonomy (FA).33 At 436–437 too Kant initially enumerates three formulas, but he has here replaced the FA with the formula of the realm of ends (FRE); thus it seems that we are already talking about four formulas. But Kant also says at 436 that the FLN, the FH, and the FRE are ‘only so many formulas of precisely the same law.’ He then encapsulates this underlying law as the ‘universal formula of the categorical imperative’ (FU), so that we now have five formulas. The thesis that the FU underlies the other formulas effectively means that there is really only one single law that can be expressed in the universal formula (FU) but also in these other formulas. Indeed, Kant also speaks of just such a ‘single’ law in another passage: “The categorical imperative is thus only a single one, and specifically this: Act only in accordance with that maxim through which you at the same time can will that it become a universal law” (421:6; o.e., Kant’s emphasis omitted). But this is also called the formula of universal law (FUL). The question is whether the FU is in fact identical with the FUL or rather with another formula (perhaps with the FA), or whether perhaps we must even acknowledge a new formula here; in that case, we should have to speak of six formulas of the categorical imperative. Kant distinguishes the ‘concept,’ the ‘formula,’ and the ‘proposition’ of the CI (420:18–20), although it is not entirely clear what these differences are supposed to involve. But it is certainly clear that the term ‘formula’ does not signify a distinction between a fundamental principle and its various formulations, for since the FU itself is also a formula, one cannot argue that the various formulas would be formulations of an underlying law that could not for its part be described as a ‘formula’ (in fact, the German term Formel derives from the Latin formula, which can mean, among other things, ‘norm,’ ‘rule,’ or also ‘formulation’). 33. At 432:2, picking up on 431, Kant speaks of the FA as the “third formula.”

Section II of the Groundwork

125

3.4.2 The Derivation of Particular Duties: Kant’s Formulas and His Examples Kant’s different formulations of the CI have generated an enormous amount of interpretation and discussion. Defenders and opponents of Kant’s ethics alike have largely concentrated their attention on the FUL (or the FLN). While some are mistaken in thinking that the FUL provides a moral algorithm that allows us to test our maxims and derive our duties, others err in believing that the FUL is itself the all-decisive formula. We shall see that the FUL does not actually allow us to derive duties from it. The criticism that has been directed at the FUL would indeed be devastating if it were the only formula that Kant provides, and it would also be devastating if this formula were really the decisive one among the others. But there are these other formulas, and even if Kant himself believed that the FUL is identical with the FU and thus represents the most fundamental formula, it will become apparent that the FH and the FA essentially have more substance (even if they are not entirely without problems of their own).

3.4.2.1 The Formula of Universalizability and the Formula of the Law of Nature The Fresh Derivation of the Categorical Imperative Kant postpones the answer to the question how a CI is possible to his later discussion in GMS III. Instead of addressing this problem here, he says that he “will first try to see whether perhaps the mere concept of a categorical imperative does not also provide us with a formula, containing the proposition which alone can be a categorical imperative” (420:18).34 Kant derives, by way of contrast, this ‘mere concept’ of the CI from the idea that in the case of a hypothetical imperative one knows what it commands only when 34. After this has been accomplished with regard to the FUL and the FLN, Kant says: “Thus we have established at least this much: that if duty is a concept that is to contain significance and actual legislation for our actions, then this duty could be expressed only in categorical imperatives, but by no means in hypothetical ones; likewise, which is already quite a bit, we have exhibited distinctly and for every use the content of the categorical imperative which would have to contain the principle of all duty (if there is such a thing at all)” (425:1–7).

126

Immanuel Kant’s Groundwork for the Metaphysics of Morals

one knows what the presupposed end in question is. In the case of a CI, on the other side, “I know directly what it contains. For since besides the law, the imperative contains only the necessity of the maxim, that it should accord with this law, but the law contains no condition to which it is limited, there remains nothing left over but the universality of a law in general with which the maxim of the action is to be in accord, and this accordance alone is what the imperative really represents necessarily” (420:27); Kant then identifies this imperative as the FUL (and its variant the FLN). If we remove all sensuous-material incentives from a rational will, then the basic thought here is that ‘there remains nothing left over but the universality of a law in general.’ In principle this is the same derivation of the CI (as the FUL) as the derivation provided for its preliminary formulation at 402. It is also vulnerable to the same objections, for it simply tells us that our maxims must accord with some laws or other, but it does not tell us anything about how to discover these laws, nor, therefore, does it imply that the maxims do accord with these laws if they pass the test that is set by the FUL and the FLN. But, above all, it by no means follows from the concept of the CI alone that the will of a rational being (rather than that of a divine being, for example) plays a role in determining the content of these laws. It is already clear that the FUL and the FLN alone are insufficient formulations of the CI, for the rationality and purposiveness that belong to a rational and autonomous being are things that are captured only by the FH and the FA.

FUL and FLN Let us now look at the FUL and the FLN more closely. Kant presents both formulas on several occasions and in a variety of ways. It is useful to provide an initial overview of Kant’s various formulations of the FUL: 1. As a preliminary formula in GMS I: “I ought never to conduct myself except so that I could also will that my maxim become a universal law” (402:7; cf. 403:21: “Can you also will that your maxim should become a universal law?”). 2. As a basic formula in GMS II: “Act only in accordance with that maxim through which you at the same time can will that it become a universal law” (421:7).

Section II of the Groundwork

127

3. “The question is therefore this: Is it a necessary law for all rational beings to judge their actions always in accordance with those maxims of which they themselves can will that they should serve as universal laws?” (426:22). 4. In the context of the summary at 431: “The ground of all practical legislation, namely, lies objectively in the rule and the form of universality, which makes it capable of being a law (at least a law of nature)” (431:9). 5. As the formula of a legislation “through which alone a realm of ends is possible” (434:8): “Do no action in accordance with any other maxim, except one that could subsist with its being a universal law” (434:10). 6. As the FU, that is, as the “universal formula of the categorical imperative: Act in accordance with that maxim which can at the same time make itself into a universal law” (436:30). 7. In the context of the summary at 437–438: “That will is absolutely good which cannot be evil, hence whose maxim, if it is made into a universal law, can never conflict with itself. This principle is therefore also its supreme law: ‘Act always in accord with that maxim whose universality as law you can at the same time will’ ” (437:6). 8. Again in the context of the summary at 437–438: “Act in accordance with a maxim that at the same time contains its own universal validity for every rational being” (437:36); Kant then elucidates this formula as “that I ought to limit my maxim in the use of means to every end to the condition of its universality as a law for every subject” (438:1). 9. In relation to the realm of ends: “Act as though your maxim should serve at the same time as a universal law (for all rational beings)” (438:21). 10. In GMS III: “The principle of acting in accordance with no other maxim than that which can also have itself as a universal law as its object” (447:3). 11. As a synthetic proposition in GMS III: “An absolutely good will is that whose maxim can always contain itself considered as universal law” (447:10). 12. In GMS III: “That the subjective principles of actions, i.e. maxims, have always to be taken so that they can also be valid objectively, i.e.

128

Immanuel Kant’s Groundwork for the Metaphysics of Morals

universally as principles, hence serve for our own universal legislation” (449:8). 13. “Why the universal validity of our maxim as a law has to be the limiting condition of our actions” (449:32). 14. “So to act that the principle of the actions is in accord with the essential constitution of a rational cause, i.e. the condition of the universal validity of the maxim as a law” (458:13). 15. The “principle of the universal validity of all its maxims as laws” (461:27).35 Some of these passages may in fact be formulations not of the FUL but of the FA. We should distinguish between the formulas that offer only a test of permissibility for maxims—that they can be thought and willed as universal laws— and formulas that command us positively to act according to those maxims that actually are (or are rationally willed as) universal laws. As we shall see presently (3.4.2.3), the latter is sometimes Kant’s way of expressing the FA—the idea that the rational will itself may regard itself as the legislator of the moral law (namely, by actually willing its maxims to be universal laws). Strictly speaking, then, one would really have to look at each of these formulations in detail. That is not possible here, and since it is obvious that these various formulations are supposed in each case to be expressing the same thing, we shall capture them with a standard formulation. We can thus encapsulate the FUL as follows: FUL: Act only in accordance with maxims of which you can at once think and will that they become universal laws It is important here that the universalizability (and thus what one ‘can will’ and ‘can think’) results solely from the avoidance of contradictions in thinking or willing. The shift from the FUL to the FLN is accomplished by reference to the concept of universality: “Because the universality of the law in accordance with which effects happen, constitutes that which is really called nature in the most general sense (in accordance with its form), i.e. the ex35. Cf. Kant’s remarks here on universality and universal validity, e.g., “universality of the maxim as a law, hence morality” (460:22); cf. also 461:27.

Section II of the Groundwork

129

istence of things insofar as it is determined in accordance with universal laws” (421:14), one can also, so Kant claims, express the FUL in terms of the FLN.36 As we shall soon see, Kant’s talk of “analogy” (437:16) is not supposed to suggest that the universality involved in FUL is ‘similar’ to that involved in natural laws.37 Rather, with the FLN Kant wants to illustrate the universal character of the FUL and the associated test of maxims by calling on us to envisage normative laws as natural laws that it is causally impossible for us to violate. We can provide the following overview of the various formulations of the FLN: 1. As a basic formula in GMS II: “So act, as if the maxim of your action were to become through your will a universal law of nature” (421:18). 2. As the variant “One must be able to will that a maxim of our action should become a universal law: this is the canon of the moral judgment of this action in general” (424:1). 3. In the context of the summary at 431: “The ground of all practical legislation, namely, lies objectively in the rule and the form of universality, which makes it capable of being a law (at least a law of nature)” (431:9); “Imperatives represented in the above way, namely of the lawfulness of actions generally similar to an order of nature” (431:25). 4. In relation to the form of the maxim: “That the maxims must be chosen as if they are supposed to be valid as universal laws of nature” (436:16). 5. In the context of the summary at 437–438: “Act in accordance with maxims that can at the same time have themselves as universal laws of nature for their object” (437:17). 6. “When we carefully conduct ourselves in accordance with maxims of freedom as though they were laws of nature” (462:36). Through FLN maxims we should really represent ourselves in terms of strict natural laws, as part of a nature in which rational beings such as 36. For this concept of nature, cf. (apart from the examples adduced for the FLN) 412:26 and 455:21; for the analogy between the FUL and the FLN, cf. above all 437:13– 20; for the related concept of the realm of ends, cf. 436, footnote; 438:23–29; and 438:35. 37. Thus Kant is not operating with the quite general concept of nature presented in § 14 of the Prolegomena.

130

Immanuel Kant’s Groundwork for the Metaphysics of Morals

ourselves act with natural necessity and without exception in accordance with those maxims.38 In this sense we can formulate the FLN as follows: FLN: Act in such a way that you can think and will your maxims as universal laws of nature Let us now turn to Kant’s specific examples.

The Examples for the FLN Kant divides the field of duties into perfect and imperfect duties that relate to others or to ourselves. Perfect duties are those that command particular actions or the avoidance of particular actions (the prohibition against suicide, for example); imperfect duties are those that do indeed command particular ends (that of benevolence, for example) but leave open how far or how intensively one pursues these ends and which means must be adopted in order to realize them. There are various tests that correspond to this division of duties. Kant claims that maxims suspected of violating perfect duties should be tested whether a contradiction in conception is involved (we can call this the CC test); the contradiction consists in the fact that a “maxim cannot even be thought without contradiction as a universal law of nature” (424:4). If a maxim can be thought as a universal law of nature but cannot be willed without contradiction, then we have a contradiction in willing; in this case it is “impossible to will that their maxims should be elevated to the universality of a natural law” (424:7). Kant applies this test (which we can call the CW test) to maxims suspected of violating imperfect duties. (We shall return later to the correspondence thesis, namely, the claim that perfect duties always correspond to the CC test, while imperfect duties always correspond to the CW test, and show that it is actually untenable.) It should be noticed that Kant uses the FLN in these four examples to illustrate the point he made at the end of GMS I: Human beings are susceptible to a ‘natural dialectic’—“a propensity to ratiocinate against those strict laws of duty and to bring into doubt their validity, or at least their strictness and purity, and where possible, to make them better suited to 38. Cf. the same thought expressed in the “typic of the pure practical judgment” in the Critique of Practical Reason (KpV:67ff.).

Section II of the Groundwork

131

our wishes and inclinations” (405). In all four examples, the agent is aware of a duty but is tempted to exempt himself from it by appealing to his needs and inclinations— either (in the first two examples) because of the exceptionally difficult circumstances in which he finds himself with regard to their satisfaction or (in the second two examples) because he thinks that the special constitution of his inclinations gives him a reason to ignore the duty. These restrictions on the use of the FLN should be taken into account in considering the aims and scope of Kant’s arguments from it. The CI and its various formulas serve to show how one ought to act. Since actions are forms of conduct in accordance with principles, and principles in the first instance are subjective maxims, one must test whether these maxims of action are morally acceptable. Thus the relevant maxim must first be identified. Kant formulates the maxim in his first example as follows: M1: If in the longer term my life threatens more ill than it promises agreeableness, then from self-love I make it a principle to shorten that life. The question then is whether this maxim “could become a universal law of nature” (422:3; 422:7; cf. 422:29). As we have indicated, this strictly implies whether it is possible to conceive a nature in which beings acted according to this maxim in a manner causally determined by laws of nature. Before the maxim is subjected to the CC test, it must therefore first be brought into a universalized form: LN1: It is a universal law of nature that all beings from self-love make it a principle to shorten their lives if in the longer term that life threatens more ill than it promises agreeableness. It is important to note that LN1 contains no contradiction. Thus the universalized maxim is not directly contradictory in itself, but only when it is brought into relation to the following natural-teleological principle: NT1: If F is a feeling whose natural end lies in producing P, then it is self-contradictory to assume a nature in which there was a law that F under particular circumstances would always produce the opposite of P.

132

Immanuel Kant’s Groundwork for the Metaphysics of Morals

The feeling in question is that of self-love.39 This feeling, in the context of nature as presupposed here, has the function of maintaining and promoting life. Now if this very feeling, under particular circumstances, had the function of terminating life, this would involve a contradiction in terms of NT1. It would be impossible to think such a nature, so Kant, assuming a nature in which the feeling of self-love promotes life and assuming NT 1, concludes that M 1 cannot be thought as a law of nature without contradiction and consequently fails the CC test. In the context of the Preface we said that although Kant certainly developed his ethics as an a priori metaphysics of morals, this does not imply that it can do without reference to empirical knowledge. The fi rst example here shows this very clearly, for that the function of self-love in nature is to promote life is undoubtedly an empirical (and probably false) claim. And even if it were true, it would describe only the contingent fact that self-love has this function in nature as it now exists; Kant does not say that in all possible worlds natures are so constituted that self-love always has this function within them. Without reference to contingent, empirical knowledge, therefore, the CC test cannot be applied at all. And as with all empirical knowledge, this alleged knowledge can be challenged on good grounds. As we already saw with Kant’s teleological argument, one can above all challenge the notion of natural purposiveness (which is very important for Kant).40 But even if we accepted this notion, it would still be unclear exactly how the characterization of selflove as something that ‘promotes life’ is to be understood and assessed. (Thus the ‘promotion of life’ might be understood, for example, as the promotion of a maximally pleasurable life or as the promotion of the life of the species rather than of the individual; in either case Kant’s argu39. From the moral point of view it is irrelevant to the test whether someone acts from self-love (with regard to the universalizability of maxims it is irrelevant what moral motive is involved). But it is decisive for the maxim test in relation to the FLN that this self-love is understood as part of a teleological system of nature. Grammatically speaking, it is possible that the formulation that speaks of “the same feeling” (422:9) may not refer to self-love. Kant might be thinking of a feeling that is not explicitly mentioned here, such as a ‘feeling of displeasure in conditions of deprivation,’ but this makes no (essential) difference to the argument. 40. In this connection we should note that Kant developed the principles of a teleology of nature in a precise way only in his later Critique of the Power of Judgment.

Section II of the Groundwork

133

ment against M 1 would be unconvincing.) One must also make the critical point that on Kantian assumptions M1 is untenable even without universalization (i.e., without reference to LN1), for if it is true that selflove always has the function of promoting life, then M1 cannot be a universal law of nature because it directly contradicts this premise, and no universalization is required to expose this contradiction. The same problem emerges in the case of the third example, so we can say that with duties toward oneself, universalization actually plays no role in the context of the FLN. In Kant’s second example41 the maxim can be put as follows: M 2: If I am in fi nancial distress,42 I shall borrow money that I promise to pay back, although I have no intention of keeping the promise.43 Kant universalizes the maxim as follows (abstracting from the par ticular situation of financial distress, 422:31–33): LN2: It is a universal law of nature that everyone who is in distress makes a promise with the aim of escaping from this distress and the intention of not keeping the promise. 41. We cannot go into Kant’s other rather similar example of the ‘deposit’ (KpV:27; cf. AA 8:286–287); for the injunction not to lie, cf. also KpV:30; KpV:87– 88; MS:238; MS:403; MS:429ff.; Theo:267; VRL; MC:444ff.; AA 11:332; and AA 23:267. 42. The German term Geldnot may be alluding (in the agent’s rationalization of his false promise) to the idea of a Notrecht (right of necessity or emergency), applied to exceptional cases in which it is sometimes held that what would ordinarily be wrong is permissible in this case owing to a special emergency or a condition of extreme need or distress. The agent in the example may be trying to pretend to himself that his urgent desire for money (represented in his dishonest imagination as a special emergency) could make his false promise a permissible exception to the moral duty to make only promises one means to keep. The FLN is then being used to expose this attempted rationalization as the patent sophistry it is, since it shows that if this were a permissible exception, then it would abolish the basic obligation imposed by promises in general. 43. Whether we are talking about an intention not to repay the money or about a knowledge that one will be unable to repay it is not entirely clear from the formulation of the maxim itself (“although I know,” 422:23 ); it is only in its universalized form that Kant speaks of an “intention” (422:33 ).

134

Immanuel Kant’s Groundwork for the Metaphysics of Morals

But LN2 could not possibly be a law of nature, according to Kant, for the universality of such a law “would make impossible the promise and the end one might have in making it, since no one would believe that anything has been promised to him, but rather would laugh about every such utterance as vain pretense” (422:33). Thus LN2 must “necessarily contradict itself ” (422:31). But what does the contradiction consist in? At first it seems clear that Kant (in contrast with his fi rst example) is not employing a principle of natural teleology here. He does not say that the purpose of promising, as given by nature, consists in facilitating trust and cooperation, and LN2 would contradict this. That interpretation would be not only substantively quite implausible but also false, for this is simply not the argument that Kant presents (at least there is no talk of this kind in the relevant passage of the text). Kant’s first example is misleading in this regard because the FLN is merely supposed to be a more vivid variant of the FUL, and the FUL is concerned with testing maxims to see whether they are possible as universal normative laws without making any assumptions regarding the teleology of nature in the context of this test. (It is thus entirely concerned with the test that is familiar to us in everyday life and that we use in particular to encourage children to engage in moral reflection: Just imagine if everyone acted in this way! And here, if we do not allow ourselves to be distracted by considerations involving the idea of a teleology of nature, it is helpful to think of the corresponding maxim as one that determines people’s actions under these circumstances without exception, as happens with a law of nature.) Although Kant formulates LN2 in a way that may not correspond particularly well to the basic idea of the FLN,44 we must proceed from the assumption that LN2 also really describes how human beings necessarily behave in particular situations (namely, in such a way that if they are in 44. That everyone who is in fi nancial distress “could promise whatever occurred to him with the intention of not keeping it” (422:32; o.e.) does not perhaps adequately bring out that in the context of the FLN a maxim must be understood as a strict law of nature. One might read the ‘could’ here as ‘may,’ and then one would perhaps have to envisage a law according to which anyone in fi nancial distress may—is permitted to—make false promises. LN2, however, states that anyone in fi nancial distress does make false promises.

Section II of the Groundwork

135

distress, they always make promises with the purpose of escaping this distress and the intention of not keeping them). Now one can distinguish between a ‘practical’ and a ‘logical’ interpretation of the resulting contradiction here. On the practical interpretation, we presuppose that a universal rule according to which one may make false promises in situations of financial distress would result in human beings losing all trust in the practice of promising, so that nobody would lend money anymore on the basis of such promises, and consequently the maxim of making false promises when in fi nancial distress would make it impossible to realize the desired end. The practical contradiction then consists in the fact that it is impossible to will an end (the loan) and at the same time will the adoption of means (i.e., M 2 as LN2) that make the realization of the end impossible. One would be willing something (the universalization of one’s own maxim) that contradicts one’s own willing (that of escaping distress through a loan). This is the form of contradiction that in the context of hypothetical imperatives we have described as a form of practical irrationality: one wills an end and at the same time wills means that make it impossible to accomplish the end. That may be a good argument against making false promises, but it is hardly Kant’s argument in the context of the FLN, for it seems obvious that it applies a test that reveals a contradiction in willing (a CW test), and honesty in promising would then be an imperfect duty, which is certainly not what Kant claims. M 2 is supposed to be subjected to the CC test (which indicates that LN2 cannot even be thought as a law of nature). And we read that false ‘promising itself ’ would become ‘impossible’ if it were universalized as a law of nature. This interpretation reveals a logical contradiction: if LN2 were binding, false promises would be impossible, and since LN2 specifically involves the making of false promises, LN2 cannot meaningfully be thought as a universal law of nature. That false promises would be impossible through LN2 does not mean that de facto with such a law nobody would any longer trust a promise made in fi nancial distress, and that therefore de facto nobody would any longer make one (this would be merely an empirical consequence). Rather, Kant has a logical problem in view: a false promise is a lie; one can lie only on the assumption that the other party is unaware of being lied to, and one can

136

Immanuel Kant’s Groundwork for the Metaphysics of Morals

be successfully lied to only if one assumes that the other party is not lying.45 If LN2 were really a universal law of nature, neither presupposition could be fulfilled; making false promises in a situation of fi nancial distress would be impossible. If we bear in mind that according to LN2 human beings in a situation of financial distress always lie (or always would lie) and that thus in such situations they always make (or would make) false promises, although such false promises are not possible at all, then in such situations no promises at all are possible because in these situations only false promises would be possible. (Kant does not say that under LN2 no promises at all would any longer be possible; he says that under LN2 no false promises like this one—promises in situations of fi nancial distress, and consequently no promises at all in situations of fi nancial distress— would any longer be possible.) In the logical interpretation, the thought is that a nature where in situations of financial distress false promises are given already makes no sense; on this interpretation, therefore, the contradiction lies in LN2 itself. In the practical interpretation, on the other hand, the contradiction first arises from what one wills and from the means that one wills in order to realize the end (or would have to will if M 2 were a universal law). We shall not discuss the third example here.46 But we still need to clarify how the CW test is accomplished. We also find such a test in the fourth example, which is concerned with benevolence and readiness to help others. The relevant maxim can be expressed as follows:

45. Cf. also the thought experiment Kant offers in his Anthropology, where he imagines “that on another planet there were rational beings that could not help thinking out loud, that is, whether sleeping or dreaming, whether alone or in society, could entertain no thoughts which they did not at the same time openly express” (Anth:332). It is clear that under such conditions the very possibility of lying would be excluded; that one can lie at all thus depends, among other things, on the natural (and contingent) fact that human beings do not have to think out loud. 46. There are simply three points that we want to mention in this connection: fi rst, it is clear that in this example too the test depends on empirical and contingent facts; second, as we noted earlier, one cannot will the corresponding maxim (here, that of neglecting one’s talents), irrespective of whether it is universalized; and third, the thesis that one cannot will the neglect of one’s talents “as a rational being” (423:13 ) is not grounded at all (something that requires reference to the FH, which only confi rms that the FLN is merely a preliminary formula).

Section II of the Groundwork

137

M4: I do not harm others or rob them of their property, but I do not help them in situations of distress or participate benevolently in their aims unless a strict right compels me to this assistance. If M4 is conceived as a universal law of nature, we obtain the following: LN4: It is a universal law of nature that nobody does harm to others or robs them of their property, but no one helps others in situations of distress or participates benevolently in their aims unless a strict right compels them to this assistance. Kant claims that M4 passes the CC test but not the CW test: “But although it is possible that a universal law of nature could well subsist in accordance with that maxim, yet it is impossible to will that such a principle should be valid without exception as a natural law. For a will that resolved on this would conflict with itself, since the case could sometimes arise in which he needs the love and sympathetic participation of others, and where, through such a natural law arising from his own will, he would rob himself of all the hope of assistance that he wishes for himself ” (423:28). This argumentation is easy to misunderstand in a number of ways. If we examine the possible misunderstandings here, we shall see more clearly just how Kant himself intends the argument to be understood: 1. The argument is not motivated by self-interest. The point of testing the maxim is not to discover which actions transpire from duty but which actions are morally acceptable, that is, in accordance with duty. Thus it is quite true that the reflection involves an empirical premise with regard to what rational beings (as rational beings) will on the basis of self-interest (help in conditions of distress, among other things). But it is not true that the ground or reason that M4 is not in accordance with duty, and that one must in principle be prepared to help others, is one’s own self-interest. 2. One might object that Kant is wrong to infer the impossibility of willing M4 as a universal law of nature already right now from the fact that someone might perhaps one day require the help of another

138

Immanuel Kant’s Groundwork for the Metaphysics of Morals

person. But Kant is asking whether it can be rational for me at any time to deprive myself for all time of the help of others. And his thesis is that this—under the real and empirically given conditions of human existence—has no sense. 3. But is it not conceivable that some people could indeed will M4 as LN4, namely, those individuals who, on the basis of their particular empirical circumstances, can incur the risk of not calling on the help of others because the disadvantage attaching to this risk is less than that which arises from offering help to others? Kant is well aware of this problem. Indeed, that is why he describes a person who upholds M4 as a human being for whom things are “going well” (423:17), and also why he thinks that M4 cannot be excluded by invoking the socalled Golden Rule.47 The premise that Kant deploys is (once again) empirical: it asserts that all human beings without exception are profoundly dependent—with regard to their life, their ends, and their general well-being—on the voluntary cooperation and potential assistance of their fellow human beings. Even the most comfortable, secure, and prosperous human beings, in order to lead a satisfying life, require the love, friendship, and benevolence of other human beings and cannot depend only on what others owe them by strict right; in view of this, it would be simply irrational to want to adopt a radically individualistic form of life in which people offered one another no voluntary aid, love, or sympathy. This is not necessarily to say that such an individualistic life could not possibly be satisfying. But it is to say that it would be irrational to rely on that possibility; the premise is thus not only empirical but also (prudentially) normative. But even on this interpretation Kant’s argument might appear unsatisfactory, for a CW test conceived along these lines could not exclude more narrowly formulated maxims that are obviously not supposed to pass the test. Thus a different maxim, one of refusing a quite specifi c kind of help that I can be quite sure I shall never require myself, could certainly be willed by me even in its universalized form. This problem could be re47. Cf. 430:34, footnote: “For many would gladly acquiesce that others should not be beneficent to him, if only he might be relieved from showing beneficence to them.”

Section II of the Groundwork

139

solved only by assuming, in a hypothetical situation, a certain awareness of the fundamental neediness of human existence but no knowledge about one’s own circumstances or resources (as with the ‘veil of ignorance’ made famous through its use by John Rawls). Given this premise of a fair and disinterested perspective, it would be impossible to will either M4 or a more narrowly specified maxim.

Further Problems of Universalization There are thus many problems with Kant’s examples, and there are also other fundamental problems with the FUL (or the FLN). These could be serious if we think that the FUL may indeed be identical to the FU and suppose that it is being offered as a universal permissibility test for all maxims. It is true, in any case, that the FUL has very often been taken as Kant’s crucial formula. As we shall see, however, there are good substantive and interpretive reasons for regarding the FUL (and the FLN) simply as preliminary formulas and seeing their use as limited to the specific kind of situation Kant is considering (namely, one where an agent suspects that his or her maxim may violate a recognized duty and needs to have the suspicion confirmed). If, however, we assume that the FUL and the FLN are meant as permissibility tests for maxims in general and even as ways of grounding all our moral duties, then there are four substantive problems: (1) the correspondence thesis bound up with the FUL is untenable; (2) connected with the fi rst point, at most one single positive duty can be derived from the FUL; (3) the universalizability tests admit of false positives (maxims that pass the tests although intuitively they are impermissible); and (4) they admit of false negatives (maxims that fail the test although intuitively they are permissible or even morally admirable). Ad 1: The correspondence thesis implies that perfect duties can be identified by reference to the CC test and imperfect duties by reference to the CW test. This thesis is untenable. First, there are maxims that if followed could imply the infringement of both kinds of duties (for example, “I shall leave no insult unavenged”; KpV:19), so that from the fact that such a maxim fails (or passes) one of the two tests we cannot infer that the corresponding duty belongs exclusively to the kind of duty correlated with the relevant test. Second, one can conceive a maxim that can be thought without contradiction as a universal law of nature but nonetheless confl icts

140

Immanuel Kant’s Groundwork for the Metaphysics of Morals

with duties that at least prima facie one would regard as perfect duties. Thus, third, the maxim that one should kill other people if that were a reliable and effective means of promoting one’s self-interest can pass the CC test (and would then be a perfect duty) but not the CW test (so that its opposite, the maxim not to kill people, would certainly be a duty, but only an imperfect one). The third reason that counts against the correspondence thesis provides a further point of criticism with regard to the FUL. Ad 2: The FUL and the CC and CW tests that are bound up with it are supposed to tell us how we ought to act, but they fail to accomplish this (which is also why no taxonomy of duties can be derived from the FUL, as required by the correspondence thesis). This is simply because the CC and CW tests can show only that it is permissible or impermissible to act in accordance with a par tic u lar given maxim, but not how, in positive terms, we ought to act in a par tic u lar way; these tests show only how we are and are not permitted to act, but they do not directly show that we have a positive duty to act in a particular way. We might also object that it does not follow from the impermissibility of a maxim (such as M2) that we are required to take its contradictory equivalent as our maxim (thus to try to make promises). From the fact that we are not permitted to act according to one specific maxim (M2) it does not directly follow that we must act according to another.48 Ad 3: From early on in the reception of GMS Kant was confronted with the false-positives objection: that particular actions that everyone regards as forbidden can nonetheless be presented as acceptable on the basis of the FUL. Here is a relevant example: The maxim of making a false promise to a person named Hildreth Milton Flitcraft on Tuesday, 25 August, in order to borrow some money from him is one that everyone naturally regards as reprehensible. But it would produce no consequences, even if it were universalized (or became a universal law of nature), that would lead this maxim to fail the CC test. The more specifically the formulation of a maxim involving a false promise incorporates circumstantial conditions under which the promise is made (even without the use of proper names

48. In fact, as Kant argues at KpV:34–35 and MS:393, only a single positive duty follows from the FUL (and then only in the context of further considerations): we ought to make the happiness of others a duty for ourselves.

Section II of the Groundwork

141

or indexical terms), the less likely it is to fail the CC test, for then there are no relevant universal consequences (indeed, it is conceivable that one could formulate a maxim that produces absolutely no effect when it is universalized). Naturally, maxims like this are rather odd or unusual. But they are still maxims, and a test that is supposed to tell us which maxims are reprehensible must be in a position to do so even with regard to maxims like this. Ad 4: An analogous and even more evident problem of false negatives arises in the case of maxims that everyone regards as permissible although they fail to pass the universalizability tests. Here are three simple examples. The maxim “I shall buy antique clocks, but I shall never sell any” would destroy itself as a universal law (and thus fail to pass the CC test), for if nobody ever sold antique clocks, nobody could buy them either. The maxim “I shall always play tennis very early on Sunday mornings because the courts are hardly used then” would also destroy itself, for if everyone acted on this maxim (if it were thus a universal law), then the courts would be completely full early in the morning, and the end I intend to pursue with my maxim would be canceled precisely through its universalization. And the maxim “I will give a larger percentage of my income to charity than the average person does” seems permissible, perhaps even admirable, but it would (obviously) be logically impossible for everyone to follow it. The problems of false positives and false negatives are each quite sufficient to show the inadequacy of the FUL or the FLN as a universal permissibility test for all maxims.

3.4.2.2 The Formula of Humanity as End in Itself To this day Kant’s ethics has typically been described as a deontological ethics. If it is essential for deontological ethics that substantive (intrinsically existing) ends or values can play only a subordinate role, then we must say that Kant’s ethics is not only not deontological but is even decisively antideontological, for Kant holds that it is impossible either to determine the content of moral duties or to ground their validity without reference to a substantive concept of value or worth. Rational beings as autonomous beings that are capable of setting moral ends have an absolute value or worth (dignity). It is this, rather than the idea of the formal

142

Immanuel Kant’s Groundwork for the Metaphysics of Morals

universalization of maxims, that constitutes the central thesis of Kant’s ethics. That is also why it is not deontological concepts (concepts such as ‘forbidden,’ ‘commanded,’ or ‘permitted’) that are ultimately decisive for Kant, but rather concepts of worth. As we have already seen, Kant begins the metaphysics of morals proper (MS3 ) not with the FUL or the FLN but only with the FH. Only at this point do we have both the form and the content of the CI, and hence the entire moral law, before us. The fi nal or definitive formulation of the law, when both sides are brought together, occurs only with the introduction of the FA. This does not alter the fact that the FH too is problematic. There are two reasons: first, it is not clear how (or even that) Kant actually grounds the unconditioned value or worth of rational beings; and second, it remains unclear how maxims can be qualified as ethically permissible by reference to the FH.

Kant’s Theory of Value: Worth, Dignity, and Ends in Themselves We saw in GMS I that Kant employs the concept of value or worth in connection with the concept of respect. His thesis is that one cannot properly have respect either for an object of inclination or for an inclination itself (400). Kant specifically repeats this thesis at 428:11–17. Again, as in GMS I, he claims that the only thing that merits respect is the activity of a rational being. But it is only now that he provides the grounding for this claim: “Rational nature exists as end in itself ” (492:2). What does he mean by this? And what precisely is this ‘activity’? Kant begins the transition to the metaphysics of morals (MS3 ) by offering a further brief analysis of rational action (427–428). The concept ‘end’ is central here. One is tempted to understand the word as signifying some condition or object that does not as yet exist, although we wish and strive for its existence. On this view, an end would always be an end for someone, and if this were indeed the only meaning that can be attached to the concept of end, the concept of an end in itself would be meaningless.49 But 49. How confusing Kant’s terminology can be is clearly revealed by some remarks in Religion within the Boundaries of Mere Reason. There Kant initially defi nes ‘end’ in general terms as the object of an inclination that one would like to possess, but in the very next sentence he introduces the concept of an objective end that is furnished by reason alone (R:6–7).

Section II of the Groundwork

143

one can also generally understand an end as that for the sake of which something happens, and that is also how Kant understands the word here: “Now that which serves the will as the objective ground of its selfdetermination is the end” (427:21). In this sense, an end in itself is also an end without its being an end for someone and without this end necessarily being a condition or object (of our interests). If there is a CI that commands particular actions as necessary, and if all actions (including these necessary actions) pursue an end, then there have to be ends that must necessarily be pursued or taken into account by everyone in her or his actions. And such ends are ends in themselves. Kant ascribes various attributes to the concept of end. He speaks of ‘subjective,’50 ‘objective,’51 ‘material,’52 ‘relative,’53 and ‘arbitrary’54 ends, of an end ‘to be effected’55 and of an ‘independent’56 end, and frequently of an end ‘in itself.’57 These various concepts of end are associated with various concepts of value or worth. Relative ends (subjective, material, and arbitrary ends and ends to be effected) are those whose existence “has a worth for us” (428:26), that is, not in themselves, and not necessarily for all rational beings. By contrast, objective (independent) ends (in themselves) are those whose existence, independently of subjective interests and means-end relations, has a value or worth in itself. Correspondingly, Kant distinguishes between the “relative worth” (428:20; 435:3) of relative ends and the “absolute worth” (428:4; 428:30) of objective ends. Something that has only relative worth has a “price” (434:31), whereas something that has absolute worth possesses “dignity” (434:32). Thus rather than speaking of the “absolute worth” of objective ends, Kant also speaks of the “inner worth, i.e. dignity” (435:4) of objective ends, and since objective ends are nothing other than ends in themselves, Kant’s thesis is this: ends

50. 427:28; 427:31; 428:25; 431:5; 431:8. 51. 427:29; 428:27; 431:6. 52. 427:34. 53. 427:34; 428:2; 436:21. 54. 436:22. 55. 437:25. 56. 437:27. 57. This concept is mentioned at least twenty-five times in the crucial pages 428– 438 alone.

144

Immanuel Kant’s Groundwork for the Metaphysics of Morals

in themselves have an absolute worth or, specifically, have dignity.58 What are just “things” (428:21; 429:21) have a price, whereas “persons” (428:22) are ends in themselves. But who or what is an end in itself? Who or what is a person? The principal thesis is undoubtedly that rational beings are ends in themselves: “Now I say that the human being and in general every rational being exists as end in itself ” (428:7).59 This is precisely what the concept of person is meant to capture, for only “rational beings . . . are called persons, because their nature already marks them out as ends in themselves” (428:21). But this does not yet tell us what a ‘person’ or precisely a ‘rational being’ is. We make use of reason to cognize things (taking reason in the broadest sense that includes the faculties of the understanding, of reason in the narrower sense, and of the power of judgment). Or we can act in a rational way. And to act rationally is to act in accordance with the representation of hypothetical or categorical imperatives. A rational being could therefore already consider itself as an end in itself inasmuch as it can set and rationally pursue ends; or again, a rational being could be an end in itself because it can act autonomously. What is Kant’s position? He says explicitly that every human being and every rational being must necessarily understand its existence as an “end in itself ” (429:3–7), but why it must do so is unclear and an object of dispute among scholars. In GMS II, moreover, the thesis that all rational beings do and must understand themselves as ends in themselves is said to be merely a ‘postulate’ and is expressly grounded only in GMS III.60 We shall discuss this question in detail in due course, but the argument can be summarized roughly here: In its very capacity to cognize and judge things, the human being must already understand itself as a free and spontaneous being because the

58. Kant speaks of the dignity of ends in themselves (that is, as we shall see, of rational-autonomous beings) at 434:29, 434:32, 434:34, 435:4, 435:8, 436:6, 438:12, 439:4, 440:1, and 440:11; but he also speaks of the dignity of moral dispositions themselves. 59. For this use of the concept of ‘rational being’ as identical with the concept of ‘end in itself,’ cf. 429:2; 430:6; 430:28; 431:13; 431:27; 433:22; 438:8; 438:16; and 439:4. 60. As Kant writes in a footnote: “This proposition [the thesis that rational beings must understand themselves as ends in themselves] I here set forth as a postulate. In the last section one will fi nd the grounds for it” (429:35, footnote).

Section II of the Groundwork

145

judgment that one is not free in thinking and judging presents itself, and can present itself, precisely only as a well-grounded and therefore free judgment (and is thus self-contradictory). As such a thinking being, the human being understands itself as an ‘intelligence’ and thus as part of the world of understanding, and on the basis of the unity of theoretical and practical reason the human being understands itself at the same time as a practically free being. As an intelligence, the human being understands itself as an ‘authentic self’ that possesses a par tic u lar and more elevated ontological status. One might suppose, therefore, that Kant understands a rational being that is simply able to think already as an end in itself. But not only is this unsupported by the text, but also the thesis that the human being is an end in itself precisely (and already) on account of its capacity to set and rationally pursue ends cannot be sustained, at least not in relation to GMS.61 This is already clear from the fact that Kant does understand ‘persons’ as rational beings, but specifically as autonomous rational beings. Autonomy is the capacity to establish and to observe moral laws in a free and self- determined fashion (this is the ‘activity’ to which Kant refers). It is this and no other capacity or characteristic that makes the human being into an end in itself and thus into a being that possesses dignity: “Autonomy is thus the ground of the dignity of the human and of every rational nature” (436:6); “The dignity of humanity consists precisely in this capacity for universal legislation, although with the proviso that it is at the same time itself subject to this legislation” (440:10).62 At the very beginning of GMS III Kant declares that freedom is “the key to the definition of the autonomy of the will” (446:6), but in GMS II he basically speaks of freedom only in passing. He tells us only that the

61. The situation may be different with other texts; cf., for example, Religion within the Boundaries of Mere Reason, where Kant specifically distinguishes between animality, humanity, and personality (R:26ff.), or his remarks in the Critique of the Power of Judgment (KU:427). 62. Cf. 438:14–16, and Kant’s reference there to “every other rational being as a universally legislative being (which is why they are also called ‘persons’)” (o.e.). Cf. also 434:23–25, where Kant says that “the will of one rational being must always at the same time be regarded as universally legislative, because otherwise the rational being could not think of the other rational beings as ends in themselves.”

146

Immanuel Kant’s Groundwork for the Metaphysics of Morals

realm of ends— that is, the realm of autonomous beings as ends in themselves— is “possible through freedom” (434:2).63 But we are fortunate enough to possess a transcript of a lecture that Kant delivered in 1784 (the year in which GMS was composed). This lecture makes it clear that it is freedom of the will that confers dignity on the human being: “The freedom of the human being is the condition under which the human being himself can be an end” (NF:1320). Or again: “I must presuppose the freedom of this being if it is to be an end in its own eyes. Such a being must therefore have freedom of the will” (NF:1322). It is also quite clear from the lecture that a being is not an end in itself simply because it is rational and capable of setting ends: “If rational beings alone can be ends in themselves, they cannot be so because they have reason, but because they have freedom. Reason is merely a means.—Through reason the human being could produce in accordance with universal laws of nature, without freedom, what the animal produces through instinct” (NF:1321–1322). Only if a rational being is free in the positive and emphatic sense that this freedom is “a law for itself ” (NF:1322) is such a being an end in itself. In its capacity as a rational and free being it then gives itself moral laws, and only then does it possess dignity: “The inner worth of the human being rests upon his freedom, upon the fact that he has a will of his own” (NF:1319); this “inner worth” is also defined here as “dignity” (ibid.).64 Thus an end in itself is a rational being that is capable of autonomously establishing and observing moral laws. Since this autonomy is rooted in freedom, it is ultimately freedom that makes a rational being into an end in itself. That rational beings must represent themselves as ends in themselves is merely ‘postulated’ in GMS II. In GMS III we do find an argument to explain why rational beings must in fact regard themselves as free in their thinking and willing. But no grounding is provided for why ratio-

63. Cf. also the passing use of the attribute ‘free’ at 435:35. 64. We have already pointed out that Kant wavers constantly (and not merely in GMS) with regard to the question whether action based on the representation of hypothetical imperatives should be understood as truly free action. If transcendental freedom must already be presupposed for rational action in general, and transcendental freedom is understood as the attribute of rational beings as ends in themselves, then in view of the argument of GMS III, rationally acting beings in general, rather than solely autonomous beings, must be understood as ends in themselves.

Section II of the Groundwork

147

nal beings are therefore regarded as enjoying absolute worth and possessing dignity. ‘Worth’ and likewise ‘dignity’ are axiological concepts with a normative content (as we shall soon see, Kant derives various duties from the claim that human beings are ends in themselves and thus possess value or worth). But Kant never really developed a genuine theory of value. He does not tell us either what exactly values in an ethical context are, or how we recognize or identify such values. Neither, therefore, does he ground why it is autonomous rational beings that specifically possess absolute value or worth (dignity). Why cannot thinking beings as such be regarded as possessing absolute worth? Why could we not say this of beings endowed with consciousness? Or of beings capable of feeling or sensation (such as animals)? Why not of unconscious but living beings (such as plants)? Kant says explicitly that in the concept of an end in itself and “in it alone” can we fi nd “the ground of a possible categorical imperative, i.e. of a practical law” (428:5): “The ground of this principle is: Rational nature exists as end in itself ” (429:2).65 But he does not tell us why this is so.

65. The passage in question (428:34–429:9) is easily misinterpreted. The “ground of this principle” (429:2; o.e.) is obviously the ground of the categorical imperative, for the ‘principle’ is clearly identified as the CI in the opening sentence of this paragraph (“If, then, there is supposed to be a supreme practical principle, and in regard to the human will a categorical imperative,” 428:34; o.e.). The ‘ground’ of the CI is thus the fact, as Kant claims it is, that rational nature exists as end in itself. With respect to this fact, Kant postulates that the human being represents his existence to himself in this way (as end in itself), although every other rational being also does the same. In this connection Kant distinguishes between the “subjective principle” (429:4) and the “objective principle” (429:7). It is important here to note that with this principle he is referring not to the previous proposition emphasized in the text (‘Rational nature exists as end in itself ’) but still to the ‘highest practical principle,’ i.e., to the CI; the neuter pronoun “it” at 429:4 and 429:7 thus refers to the CI. This highest practical principle is a ‘subjective principle’ insofar as “the human being” (429:3; o.e.) represents his existence to himself as end in itself; it is an ‘objective’ principle because “every other rational being” (429:5; o.e.) also represents its existence in this way. This interpretation is also confi rmed by the fact that after formulating the FH and the associated examples, Kant simply continues with the sentence beginning “This principle of humanity” (430:28; o.e.). The only intelligible referent for the demonstrative pronoun here is the ‘principle’ mentioned at 429:4 and 429:7 (identified by the ‘it,’ as we have just indicated). Since in 430:28–431:18 it is clearly the moral principle, not the proposition that ‘rational nature exists as end in itself,’ that is alluded to, it can only be the ‘highest practical principle,’ and thus the moral law, that is being referred to at 429.

148

Immanuel Kant’s Groundwork for the Metaphysics of Morals

It remains unclear (at least as far as GMS is concerned) how and why any normative validity can specifically be derived from the fact that human beings are capable of acting morally. In terms that we might use today, we can say that Kant is a moral realist. He is convinced that our moral judgments possess a cognitive character (i.e., can be true or false) and relate to moral facts. The reality in question is the fact that autonomous beings are ends in themselves. The ‘absolute worth’ of this fact is not something that human beings produce, something that they bestow, or something that could somehow be derived from an already-assumed de facto recognition of values; Kant is neither a constructivist nor a subjectivist. That is why he also distinguishes autonomous beings as ends in themselves from “every end to be effected” (437:25). An end in itself is nothing that we could produce but is rather a “selfsuffi cient end” (437:27). Such an end is not “an object that one actually from oneself makes into an end, but . . . an objective end” (431:5; o.e.). Such an end is thus, in the sense of moral realism, something with objective being, an “existence” [Dasein], as Kant says (428:3; 428:27; 429:4; 429:6), something that “exists” [existiert] (428:8; 429:3). It is only for this reason that such an end is “necessarily an end for everyone, because it is an end in itself ” (428:36; o.e., Kant’s emphasis omitted). But no (convincing) grounding is provided for these claims here.66 Nor do we find (at least in GMS) anything to suggest whether we might perhaps have to distinguish different levels of rationality in this connection. It is obvious that persons are more or less rational, but that is not the problem. The problem is that autonomy too may present itself in different degrees, or aspects, or even in merely potential ways (one has only to think of embryos, the newly born, or those who are severely mentally handicapped). A Kantian solution to this problem will be no less complex than the response to the question how in the context of Kantian ethics 66. It is true that in GMS III Kant grounds the validity of the CI by reference to the ontological status and the superiority of the human being as an ‘intelligible being.’ But quite apart from the fact that this thesis is unconvincing on both internal and external grounds, we fi nd almost nothing here regarding worth or values, or how we might have knowledge of them (in fact, the concept of an end in itself no longer plays any role in GMS III).

Section II of the Groundwork

149

(at least) the required protection of animals and natural objects can properly be grounded. But it is clear in any case that (for Kant) it follows directly from the thesis of autonomous beings that the worth or value of all autonomous beings is identical and absolute. Rational beings possess dignity insofar as they are autonomous, irrespective of whether or to what extent they make use of their autonomy, or whether they are good and moral human beings. For all the contempt or disdain we may feel for human beings such as Hitler, and for all the punishment they rightly require, they still possess and retain worth and dignity. Their moral status and the rights that derive from this remain untouched by their actual transgressions and can be neither earned nor forfeited.67

The Derivation of Duties from the FH As in the case of the FUL and the FLN, it is helpful to begin with an overview of the various formulations that Kant provides for the FH: 1. The human being “in all its actions, those directed toward itself as well as those directed toward other rational beings, . . . must always at the same time be considered as an end” (428:9). 2. As a basic formula in GMS II: “Act so that you use humanity in your own person as in the person of every other, always at the same time as end and never merely as means” (429:10). 3. As a formula in the context of the realm of ends: “For rational beings all stand under the law that every one of them ought to treat itself and all others never merely as means, but always at the same time as end in itself ” (433:26).68

67. Cf. also MS:463. It is indeed Kant’s thesis that the good will alone is good in an unlimited sense, but it does not follow from this that only beings who act morally are ends in themselves. The end in itself is indeed supposed to be the ‘ground’ of the necessary binding character of the CI, and it could not be so if it consisted in something whose existence on Kant’s own view were quite doubtful (i.e., the existence of a will that is actually good in itself). 68. At 433:15 Kant describes the concept of the realm of ends as ‘depending’ on the concept of autonomy; but at 433:28 he says that the concept of the realm of ends arises ‘from this,’ that is, from the FH.

150

Immanuel Kant’s Groundwork for the Metaphysics of Morals

4. In relation to the matter of the maxim “That the rational being, as an end in accordance with its nature, hence as an end in itself, must serve for every maxim as a limiting condition of all merely relative and arbitrary ends” (436:19). 5. In the context of the summary at 437–438: “I.e. never to be acted against [that is to say, against the self-sufficient end], which therefore has to be estimated in every volition never merely as means but always at the same time as end” (437:28). 6. Again, in the context of the summary at 437–438: “Act in reference to every rational being (to yourself and others) so that in your maxim it is always valid at the same time as an end in itself ” (437:34); this formula is elucidated in turn: “The subject of ends, i.e. the rational being itself, must be made the ground of all maxims of actions never merely as means, but as the supreme limiting condition in the use of all means, i.e. always at the same time as end” (438:4). As in the case of the FUL or the FLN, so too with regard to the FH Kant says that “all laws of the will must be able to be derived” (429:8) from it. He formulates the FH and then returns to and discusses the same examples in relation to the new formula (429:14–430:27); once again, therefore, Kant is concerned with the derivation of perfect and imperfect duties with regard to oneself and to others.69 In this connection Kant speaks not 69. Before we examine this more closely, we must make another important point. Kant begins his discussion of the various formulas by indicating that all duties “can be derived” (421:10; o.e.) from the FUL. After offering the four examples for the FLN (as a variant of the FUL), Kant continues as follows: “Now these are some of the many actual duties, or at least of what we take to be duties, whose derivation from the single principle just adduced clearly meets the eye” (423:36; o.e., reading Ableitung here rather than Abteilung). There is no doubt that the “single principle” in question is the “one imperative” (421:9) introduced just before the examples, namely, the FUL. It seems evident, therefore, that after the examples Kant is repeating what he had announced earlier: he had claimed that all duties ‘can be derived’ from the FUL, and after giving the examples he says that this ‘derivation’ was successful (that ‘the derivation from the single principle . . . clearly meets the eye’); correspondingly, he says once again, before the FH, that the duties ‘must be able to be derived’ from the FUL. But the problem is that in both original editions of 1785 and 1786 the text in 423:37 has Abteilung (partition) rather than Ableitung (derivation). And since in the examples Kant also undertakes to provide a division of duties into perfect and imperfect duties

Section II of the Groundwork

151

only of the ‘person’ but also of ‘humanity.’ In contrast to his later writings (KpV, R, and MS), Kant does not here introduce these terms in a systematically differentiated way. In those writings, humanity is the predisposition to set ends and pursue happiness, while personality is the capacity for autonomous self-legislation and moral accountability.70 But here both concepts are introduced in connection with the capacity to set ends according to reason and thereafter might be thought to refer indiscriminately to rational nature in both aspects; at least sometimes, Kant relates ‘humanity’ specifically to our moral vocation. “Thus morality and humanity, insofar as it is capable of morality, is that alone which has dignity” (435:7; o.e.); hence Kant also speaks of the “idea of humanity as an end in itself ” (429:17).71 Kant’s FH may sound to us like a good, noble, and beautiful ideal, but one that is too abstract to be helpful or relevant. Kant thought quite otherwise. As we have seen, he even regards the FH as particularly

toward oneself and others, one might suppose that Abteilung should be read in the sense of Aufteilung (division). But other considerations, in addition to the textual context we have indicated and Kant’s general use of the term ‘derivation,’ speak against this and in favor of a corresponding substitution of Ableitung for Abteilung. Kant also explicitly describes the “division” (Einteilung) (421:31) he proposes as “discretionary” (421:33, footnote), and he would hardly wish, a few pages later, to claim that the Abteilung (in the sense of this Einteilung) is something that ‘from the single principle clearly meets the eye.’ Finally, it should be pointed out that the expression Abteilung aus is not to be found anywhere else in Kant’s work. 70. Cf. R:26–28 and MM:6, 223. 71. Cf. 430:11; 430:16; 430:23; 430:28; and 431:4. In this context, therefore, ‘man’ or ‘humanity’ may not designate primarily human beings as an animal species but rather may refer to the intelligible character of human beings as members of the intelligible world (cf. MS:239; at KpV:87 Kant describes this as ‘personality’ in contrast to ‘person’ qua sensuous being). In view of Kant’s repeated remarks on the a priori character of the metaphysics of morals, it is clear that the status of an end in itself must be ascribed to all autonomous beings as such; thus Kant can also say, for example, “This principle of humanity and of every rational nature in general as end in itself ” (430:28; o.e., Kant’s emphasis omitted) is a universal and a priori principle, and that precisely because “it applies to all rational beings in general” (431:2; o.e.), “Autonomy is thus the ground of the dignity of the human and of every rational nature” (436:6; o.e., Kant’s emphasis omitted). So although Kant speaks of ‘humanity,’ he means rational nature rather than our biological species; he cannot be charged with arrogantly privileging the human biological species over everything else.

152

Immanuel Kant’s Groundwork for the Metaphysics of Morals

compelling (436–437), and it has also become abundantly clear that the concept of end implied in the FH is central to Kant’s ethics in an intrinsically substantive and normative sense: in the later-published Metaphysics of Morals (MS2 ) almost all our duties are derived by reference to the FH rather than to the FUL.72 And when Kant introduces the same four examples in relation to the FH (429–430) that he did in relation to the FLN (421–424), he uses the FH not to disqualify par ticular maxims as contrary to duty but instead to ground the general positive duties that provided the assumed background of his discussion at 421–424. The thought implied in FH could be expressed as follows: FH: Respect both in your own person and that of others the absolute worth of autonomous beings as ends in themselves and also treat them for this reason never merely as means but at the same time always also as ends in themselves. Actions that are oriented to the FH always therefore express recognition of and respect for an existing absolute value. There exists something that is absolutely valuable or worthy, and rather than damaging, violating, or destroying something of absolute worth, one respects and acknowledges it. Thus it is really only with the FH that the decisive difference between Kantian ethics and consequentialist approaches to ethics properly emerges. In consequentialist ethics we are always essentially concerned with producing desirable states of affairs in an optimal fashion, so that in principle anything could be permitted as a means to this end. In Kant’s ethics, oriented as it is to the conception of worth or value, we are not (primarily) concerned with producing some state or condition but rather with respecting an existing absolute worth or value. Worth or value of this kind sets limits that may not be infringed even with the prospect of achieving the greatest amount of happiness. In all four of Kant’s examples it is immediately evident that the basic underlying structure is the same: certain actions are commanded as duties because they express respect for the absolute worth of persons, while 72. Cf. in par tic u lar MS:237; 389; 423; 425; 427; 436; 444; 451; 453–454; 456; 459; and 462.

Section II of the Groundwork

153

other actions are forbidden because they fail to express such respect. Without going into the interpretive details, we can understand the derivation of duties in the examples in the following manner: 1. Any action that fails to respect in one’s own person or in every other person the absolute worth of autonomous beings as ends in themselves, and thus fails to treat such beings never merely as means but at the same time always also as ends in themselves, is morally prohibited. 2. Action x fails to respect in one’s own person or in every other person the absolute worth of autonomous beings as ends in themselves and thus fails to treat such beings never merely as means but at the same time always also as ends in themselves. Therefore: action x is morally prohibited.73 If we apply this pattern to the fi rst example, we have the following argument: 1. Any action that fails to respect in one’s own person or in every other person the absolute worth of autonomous beings as ends in themselves, and thus fails to treat such beings never merely as means but at the same time always also as ends in themselves, is morally prohibited. 2. The action of suicide fails to respect in one’s own person or in every other person the absolute worth of autonomous beings as ends in themselves and thus fails to treat such beings never merely as means but at the same time always also as ends in themselves. Therefore: the action of suicide is morally prohibited. We have already pointed out that Kant does not actually ground the fi rst premise (i.e., the FH). He says little about what values or instances of worth really are (ontologically speaking), about how they can be known 73. The derivation of positive duties follows a corresponding pattern.

154

Immanuel Kant’s Groundwork for the Metaphysics of Morals

or recognized (epistemologically speaking), or about how and why the normative question concerning the validity of moral laws can be answered in relation to such values (ethically or metaethically speaking). This lack of clarity is also particularly evident in the derivation of our duties. Even if we accept the FH, this still does not tell us which action would contradict the FH; thus it is unclear whether the second premise in each case applies. How do we know whether a particular action expresses or fails to express respect for the absolute value or worth of autonomous beings? Thus it is conceivable, for example, that a proponent of actively assisted suicide would argue that we would show respect for the dignity of the human being precisely by permitting and even helping someone end his or her life if he or she were suffering from terrible and untreatable pain without any prospect of a future cure.74 An argument of this kind would specifically challenge what Kant claims in his example, namely, that someone who commits suicide “makes use of a person merely as a means” (429:19). It is true that such a person uses himself or herself as a means, but not “merely” (429:21) as a means, and the FH does not forbid using persons as means as long as they are also (and not less) respected as ends in themselves. Indeed, we will see presently that the FRE requires rational beings to treat one another reciprocally as both ends and means. In the second example, there would seem to be greater consensus regarding the second premise: 1. Any action that fails to respect in one’s own person or in every other person the absolute worth of autonomous beings as ends in themselves, and thus fails to treat such beings never merely as means but at the same time always also as ends in themselves, is morally prohibited. 2. The action of making a lying promise fails to respect in one’s own person or in every other person the absolute worth of autonomous beings as ends in themselves and thus fails to treat such beings never merely as means but at the same time always also as ends in themselves. 74. In certain cases Kant himself seems to regard suicide as permitted or even commanded; cf. MC:370–371.

Section II of the Groundwork

155

Therefore: the action of making a lying promise is morally prohibited. It seems simply obvious that mendacious promising harms other human beings and their ability to set ends. But Kant has a specific argument. He claims that someone who is lied to (deceived) by me “cannot possibly be in harmony [einstimmen] with my way of conducting myself toward him” (429:34). This might seem to provide an argument that could solve the problem of derivation (how do we know which actions contradict the status of rational beings as ends in themselves?): an action (or a kind of action) contradicts the FH, and thus the status of rational beings as ends in themselves, if not all who are affected by the action can be in harmony or agree with the action. In this perspective, the FH calls on us to treat rational beings in such a way that these beings can also express agreement with the way in which they are treated. Interestingly enough, a little later in the text, Kant claims that the FH and the FUL (and perhaps he means the FA as well) are “fundamentally the same” (438:1; o.e.), and that the one formula says “just as much” (438:4) as the other. Now the FUL (like the FA) involves the thought of universality: moral laws are laws are that are valid for everyone without exception. That they are valid for everyone, however, also means that they can be willed (and, if we consider the FA, that they are willed) by all who are addressed by and affected by these laws. But this concept of ‘can be willed’ is problematic, for either it assumes the actual willing, and then it is clear that one person wills this and another wills that, without the reference to the willing of those affected helping us resolve the question of what actual willing is morally permitted or deserves to be taken into account; or the willing is already conceived as rational willing, but then this assumes what is still to be shown (namely, that a specific case of willing is rational, i.e., is permitted). It is not sufficient to say that rational beings cannot agree or be ‘in harmony’ with particular types of actions, for the question is precisely which types of actions these are.75

75. It is no accident, with regard to the second example in which he speaks of ‘harmony,’ that Kant adds a footnote on the so-called Golden Rule (“What you do not will that another shall do to you,” etc.). He describes the deficiencies of the Golden Rule here but fails to make clear why these deficiencies do not apply to the FH as well if the latter is read in term of the ‘harmony’ in question.

156

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Thus it is still difficult to see how concrete duties can be derived from the FH. The actions that are relevant in each case in the second premise need to be interpreted. It is doubtful that, in providing that interpretation, we can make do without referring to something that Kant would prefer as far as possible to exclude from the task of defi ning duties (although not that of obeying them), namely, the role of feelings. Kant explicitly acknowledges feelings (especially, of course, respect) as a principium executionis. Thus the accusation that Kant reduces human beings simply to the dimension of their rationality certainly cannot be sustained, for reason can come to determine our actions only through a feeling, namely, respect, and in his later writings in particular, as we have seen, Kant strongly emphasizes the indispensability of such emotional dispositions as love, conscience, respect, and moral feeling. Reason, on the other hand, is supposed to serve as the sole principium diiudicationis. But that there are values is something that we can understand only if we experience them. The assertion that ‘human beings have absolute worth or value’ would be entirely meaningless for us if we did not experience values (noncognitively) or somehow know what they are. It would be as meaningless as attempting to explain the meaning of the word ‘yellow’ to someone completely blind.

3.4.2.3 Autonomy and the Realm of Ends We now offer a brief examination of the FA and the FRE (the formula of the realm of ends). In our interpretation of GMS III we shall return in some detail to the concept of autonomy and its connection with Kant’s theory of freedom. The basic idea of autonomy, insofar as it functions as the ground of validity, is this: If we understand morality as Kant does, then moral laws and moral adherence to these laws (not simply conformity with them) cannot be bound to the idea that one is thereby pursuing any interest, for that would transform categorical imperatives into hypothetical ones.76 But if no such interest is permitted in this regard, what can 76. “Now it is no wonder, when we look back on all the previous efforts that have ever been undertaken to bring to light the principle of morality, why they all had to fail. One saw the human being bound through his duty to laws, but it did not occur to one that he was subject only to his own and yet universal legislation, and that he was obligated only to act in accord with his own will, which, however, in accordance with its

Section II of the Groundwork

157

motivate us to act morally? Such a thing is conceivable only if what we ought to do is in a certain sense something that we ourselves will to do. Only what we ourselves already will can be willed without having to fi nd some additional interest to explain why we ought to will it, for one already wills it; the interest is already there (even if it can be overlaid or obscured by other empirical interests). Now Kant claims that we human beings, insofar as we are rational, do “take an interest” in morality (413:31, footnote).77 We do so because the moral law springs from that free reason that every single one of us possesses as a free and rational being. In this sense the moral law is a law that we give ourselves (this is what ‘auto-nomy’ means), and that can also therefore motivate us. Hence Kant also thinks that with the FA, in contrast to the FLN and the FH, something else is “indicated” (432:1), namely, “the withdrawal of all interest in the case of volition from duty . . . as the specific sign distinguishing the categorical from the hypothetical imperative” (431:35; o.e.). In order to follow Kant’s thought here, let us first review his formulations with regard to the FA: 1. As the “idea of the will of every rational being as a will giving universal law. [new paragraph] All maxims are repudiated in accordance with this principle which cannot subsist together with the will’s own universal legislation” (431:16). 2. As the “idea of the will of every rational being as a universally legislative will” (432:3); the FA involves the “principle of every human will as a will legislating universally through all its maxims” (432:12). The categorical natural end, is a universally legislative will. For if one thought of him only as subject to a law (whatever it might be), then this would have to bring with it some interest as a stimulus or coercion, because as a law it did not arise from his will, but rather this will was necessitated by something else to act in a certain way in conformity with the law. Through this entirely necessary consequence, however, all the labor of fi nding a supreme ground of duty was irretrievably lost. For from it one never got duty, but only necessity of action from a certain interest. Now this might be one’s own interest or someone else’s. But then the imperative always had to come out as conditioned, and could never work at all as a moral command” (432:25). 77. Kant distinguishes between the ‘pathological’ interest that is produced by our inclinations and the pure interest in moral laws that is produced by reason (‘practical’ interest); cf. 413–414, footnote; 449–450; and 459–460.

158

3.

4.

5.

6.

7.

8. 9.

Immanuel Kant’s Groundwork for the Metaphysics of Morals

imperative in the FA “can command only that everything be done from the maxim of its will as a will that at the same time have as its object itself as universally legislative” (432:20). It is concerned with “the concept of every rational being that must consider itself as giving universal law through all the maxims of its will in order to judge itself and its actions from this point of view” (433:12). At 434 Kant connects the FA with the FLN through a “hence” (434:12): “Do no action in accordance with any other maxim, except one that could subsist with its being a universal law, and hence only so that the will could through its maxim at the same time consider itself as universally legislative” (434:10). In the context of the summary at 437–438: “Act in accordance with a maxim that at the same time contains its own universal validity for every rational being” (437:36). Kant then elucidates this formula by saying “that I ought to limit my maxim in the use of means to every end to the condition of its universality as a law for every subject” (438:1). With reference to the realm of ends: “Act as though your maxim should serve at the same time as a universal law (for all rational beings)” (438:21). “The principle of autonomy is thus: ‘Not to choose otherwise than so that the maxims of one’s choice are at the same time comprehended with it in the same volition as universal law’ ” (440:18); “The suitability of its maxims for its own universal legislation” (441:3). “Autonomy, i.e. the suitability of the maxim of every good will to make itself into a universal law, is itself the sole law that the will of every rational being imposes on itself ” (444:30). In GMS III: “An absolutely good will is that whose maxim can always contain itself considered as a universal law” (447:3). As a synthetic proposition in GMS III: “That the subjective principles of actions, i.e. maxims, have always to be taken so that they can also be valid objectively, i.e. universally as principles, hence serve for our own universal legislation” (447:10).78

78. As we pointed out earlier, formulations 4 and 5, as well as 8–10 (at least), can be understood in the sense both of the FA and the FUL.

Section II of the Groundwork

159

All these formulations are concerned with the idea of a universally legislative will. Kant grounds the ‘withdrawal from all interest’ that we have just mentioned as follows: “It is impossible for a will that is itself supremely legislative to depend on any interest; for such a dependent will would need yet another law, which limited the interest of its self-love to the condition of a validity for the universal law” (432:7). In other words, to represent to oneself a rational legislator who establishes universal laws for everyone and thus also for herself implies the idea of a legislator who abstracts in this legislation from her subjective interests, for if she did not abstract from the latter, she could not as such establish any universal laws. The meaning of ‘universal’ here, it seems, is determined by the FUL. In fact, as we shall see, Kant understands the FA as a principle that embraces the FUL and the FH within itself (he even says that the FA “follows” [431:14] from the FUL and the FH). One might therefore think that the FA cannot be distinguished from the FUL as far as the idea of ‘universality’ is concerned, and this impression is only strengthened by the fact that Kant feels no need to provide any examples to elucidate the FA “since those that first elucidated the categorical imperative and its formula can all serve here for precisely that end” (432:35).79 But even if Kant did think that this was the case, one can readily see that the FA involves a stronger criterion than the FUL as far as the concept of legislation is concerned. The FA calls on us to represent our own will as a will that is in fact (positively) ‘universally legislative.’ But such a will not only can prescribe how we are not permitted to act but also must positively prescribe how we ought to act. But it is quite conceivable that a maxim M can be thought and willed without contradiction (and thus pass the test of the FUL) without it following from this and from M itself that one actually wills M as a universal law. Let us consider the following maxim as an example: M2a: If I am in financial distress, I shall borrow money on the promise to repay it. There is no doubt that one can think and also will this maxim without contradiction as a universal law, and, of course, the corresponding action is 79. On page 432 the Academy Edition has jumped a line after line 34.

160

Immanuel Kant’s Groundwork for the Metaphysics of Morals

permitted (the maxim thus passes both the CC and the CW test). But from the fact that in financial distress one may borrow money (on the promise of repaying it) it does not follow that one can or even must will that everyone who is in financial distress must borrow money— one could also acquire the money, for example, by working for it oneself, selling a portion of one’s property, or getting someone simply to give one the money. There are many maxims that express morally acceptable ways of responding to financial distress, and M2a is just one of them. The FUL and the FLN, if anything, merely involve tests for assessing the permissibility of maxims, but they do not tell us the maxims in accordance with which we ought to act. Kant says that the FA “leads to a very fruitful concept depending on it, namely that of a realm of ends” (433:15). This involves a further formula, that of the realm of ends (the FRE): 1. In relation to the complete determination of all maxims: “That all maxims ought to harmonize from one’s own legislation into a possible realm of ends as a realm of nature” (436:24). 2. “Accordingly, every rational being must act as if it were through its maxims always a legislative member in a universal realm of ends” (438:18). 3. “Act in accordance with maxims of a universally legislative member for a merely possible realm of ends” (439:1). But what do we mean by a ‘realm of ends’? As we shall see more precisely later, Kant introduces the concept of the ‘world of understanding’ in GMS III. He uses this concept to refer quite generally to the world of ‘things in themselves’ (in contrast to the world of sense as the world of ‘appearances’). But human beings also belong to the world of understanding insofar as one regards them as transcendentally free and in the strict sense as rational (autonomous) beings. In this sense Kant speaks of the “intelligible world” (458:27), which is “the whole of rational beings as things in themselves” (458:28), that is, a “world of intelligences” (462:2). Now one might think that this concept of an ‘intelligible world’ is precisely what Kant means by a ‘realm of ends,’ for the concept of the ‘intelligible world’ is indeed described as an “ideal” (462:12), and Kant also understands the realm of ends as such an “ideal” (433:32; 462:34). But the world of understanding is an

Section II of the Groundwork

161

ideal only insofar as there can be no empirically based explanation and no theoretical proof of this world.80 Kant leaves us in no doubt that there is such a world, and in a certain sense he even offers a proof. The realm of ends, by contrast, is an ‘ideal’ in the sense that it does not really exist, although it could exist and, above all, ought to exist. A realm of ends is simply a “practical idea to bring about that which does not exist but which can become actual through our deeds and omissions and what we are to bring about in accord with precisely this idea” (436:36, footnote; o.e.).81 Kant defi nes the realm of ends as “a whole of all ends—(of rational beings as ends in themselves, as well as of their own ends, which each may set for himself) in systematic connection” (433:21). Thus autonomous beings and their subjective ends belong to the realm of ends.82 But the capacity for moral legislation alone does not make a rational being into a member of the realm of ends. It is immediately evident that this is the case, for otherwise this realm would not be something that should be brought about. The ‘world of intelligences’ is thus the necessary but not sufficient presupposition of the realm of ends because this world comprises the beings that are capable of acting autonomously (that is, are ends in themselves), and membership in the realm of ends is therefore something for which a rational being “by its own nature was already destined” (435:33; o.e.), that is, precisely by its autonomous nature.83 Thus it seems that Kant is saying that those autonomous beings alone are members of the realm of ends that are actually governed in their actions by laws that they 80. In this sense Kant also describes it (along with freedom) as an “idea” (458:26; 462:3; 462:18; 462:30; 462:32). 81. This is why Kant speaks repeatedly of a merely ‘possible’ realm of ends; cf. 435:6; 436:24; 438:17; 438:23; 438:36; and 439:2. 82. In the same paragraph (433:17–25), however, Kant also writes that in the realm of ends one must abstract “from the personal differences between rational beings, as likewise from every content of their private ends.” It is very difficult to see how this is supposed to be compatible with the thought that the realm of ends also includes ‘their own ends which each [rational being] may set for himself ’ (after all, everything that has a price also belongs to the realm of ends [cf. 434:31]). Naturally we must exclude all ends that have no chance whatsoever of being universally accepted; cf. 450: “Good ends are those which are necessarily approved by everyone, and which are also capable at the same time of being everyone’s ends.” 83. In this sense cf. the intimate connection Kant establishes between the world of understanding and the realm of ends at 462:29–463:2.

162

Immanuel Kant’s Groundwork for the Metaphysics of Morals

themselves establish and that make possible the moral community of all members in the realm of ends.84 Such a realm of ends, Kant claims, “would actually be brought about through maxims, the rule of which is prescribed by the categorical imperatives of all rational beings, if they were universally followed” (438:29).85 In other words, only autonomous beings as ends in themselves can build this realm of ends at all, but in order to build it these beings must steadfastly follow their own vocation of acting morally.86 Kant clearly regards the FRE as a variant of the FA. At 431 he expressly describes the FLN (the FUL), the FH, and the FA as three principles. Since a few pages later Kant speaks in retrospect of the “three ways mentioned of representing the principle of morality” (436:8; o.e.), he can be 84. There is no doubt that members of the realm of ends are beings that are also capable of acting immorally, for the moral law is still a “duty” (434:18), as Kant emphasizes, as far as members of the realm of ends are concerned. What Kant describes, on the other hand, as the “supreme head” of the realm of ends is “a fully independent being, without need and without limitation” (434:5). It is thus unclear whether we can simply identify the realm of ends with the “idea of a moral world” (A 808/B 836) that Kant describes in the KrV. On the one hand, this moral world is understood as a world “insofar as it may be in accordance with all moral laws (and this is what by means of the freedom of the rational being it can be, and what according to the necessary laws of morality it ought to be)” (ibid.). On the other hand, Kant emphasizes that in this world we “abstract from all conditions (ends), and even from all hindrances to morality here (the weakness or impurity of human nature)” (ibid.; o.e.; cf. A 809/B 837); Kant describes this world correspondingly as “a corpus mysticum” (ibid.), although he subsequently emphasizes that “we are necessarily constrained by reason to represent ourselves as belonging to such a [moral] world” (A 811/B 839). But how is this abstraction supposed to be possible and necessary if the moral world ‘according to the necessary laws of morality’ is something that ‘ought to be’? This ‘ought’ makes sense only if we are talking about beings that are also capable of acting immorally. 85. Cf. also 435:29: “And now, what is it that justifies the morally good disposition or virtue in making such high claims? It is nothing less than the share that it procures for the rational being in the universal legislation, thereby making it suitable as a member in a possible realm of ends, for which it by its own nature was already destined, as end in itself and precisely for this reason as legislative in the realm of ends” (o.e., Kant’s emphasis omitted); and 462:35: a realm of ends “to which we can belong as members only when we carefully conduct ourselves in accordance with maxims of freedom as though they were laws of nature” (o.e.). 86. In fact, this thought sometimes tempts Kant into formulations that even suggest that only those rational beings that in fact act morally can possess dignity itself. He says, for example, that the “suitableness of its [the rational being’s] maxims for the universal legislation designates it as an end in itself ” (438:11); Cf. similarly 435:2– 9 and 439:7–17.

Section II of the Groundwork

163

referring only to the aforementioned FLN, FH, and specifically FA; and if in the renewed listing at 436 Kant then mentions the FLN, the FH, and the FRE, this obviously signifies that the FRE is a variant of the FA.87 The idea of a realm of ends yields an important new aspect that was already evident in the FA: the “systematic combination of rational beings through communal laws” (433:17). It might initially seem plausible to understand this realm politically as an ideal state. But that cannot be correct, for the concept of the political state, for Kant too, essentially involves the notion of right, and the notion of right in turn involves the permissibility of coercion. Legality merely requires actions in conformity with duty, and one can be coerced in relation to such actions. But Kant explicitly emphasizes that it is only the “morally good disposition” (435:29) that makes rational beings into members of the realm of ends; they must will morality for the sake of morality. But the ‘systematic combination’88 of rational beings also does not consist in the fact that these beings have maxims or ends that are merely compatible with the maxims and ends of others. The realm of ends “arises” (433:29), as Kant says, from the command expressed by the FH to respect other rational beings as ends in themselves and to respect their ends, for the realm of ends is defined, as we already saw, as a whole of ends in themselves and the ends that they set themselves. The realm of ends is organized through “communal laws” (433:18; o.e.), as the laws of an (ideal) community of rational beings. These laws “have as their aim the reference of these [rational] beings to one another as ends and means” (433:31). From this, and from the “analogy” (438:23) that Kant draws between the realm of ends and a teleologically organized “realm of nature” (436:25; 436:34; 438:24–29–35), it emerges that the realm of ends must be thought of as an organized community of beings that harmoniously organize and mutually support one

87. This is also evident from the fact that the idea of autonomy is contained in the formulation of the FRE at 436: “That all maxims ought to harmonize from one’s own legislation into a possible realm of ends” (436:24; o.e.). We also read that the realm of ends can arise “from this” (433:28; o.e.), i.e. through the validity of the FH. 88. For the idea of ‘systematic combination,’ cf. also 433:23 and 433:29. The reference to the “complete determination of all maxims” (436:23) is also to be understood in this sense, for here we must recognize that “all maxims ought to harmonize from one’s own legislation into a possible realm of ends” (436:24; o.e.).

164

Immanuel Kant’s Groundwork for the Metaphysics of Morals

another and the ends they set themselves. Since all maxims (actions) involve the pursuit of ends, and the laws of the realm of ends are ‘communal laws,’ these laws must also refer to a ‘communal’ end. This end is precisely the harmonious and reciprocally supportive community, or the realm of ends itself. Here we see once again that the validity of moral laws cannot be grounded in the concept of self-interest in a contractualist manner. The self-interest of rational beings is always only one’s own interest, and there is no way of moving from one’s own subjective interest to the interest of other beings and respecting or even supporting the interest of other beings unless these other beings and their interests (ends) are understood as something that has objective worth or value. It is only if one’s own human existence is understood as something objectively worthy or valuable, as an end in itself—as something that does not have worth or value in every case merely for a subjective existence but possesses worth or value in itself—that something like a realm of ends can meaningfully be conceived at all.89

3.4.3 The Categorical Imperative and the Relationship of Its Formulas to One Another Let us conclude our analysis by looking once again at the specific function of Kant’s various formulas and their relationship to one another. There are various aspects and passages to consider here: 1. The singular character of the categorical imperative: Kant begins directly by claiming that there is only a “single” CI (421:6).90 Even when Kant moves on to discuss the other formulas, he constantly emphasizes that we are still dealing in each case with another formula of

89. Kant did not in fact undertake, either in GMS or in KpV, or indeed in MS, to derive par tic u lar duties by reference to the FA or the FRE (although there are some attempts in this direction in his lectures; cf., for example, RP:1100). We shall not attempt an analysis of this question here. 90. Three lines later (421:9) Kant once again describes this ‘single’ imperative as the ‘one’ imperative. After he has identified this ‘single’ imperative with the FLN, he subsequently speaks of the “derivation [of duties] from the one principle just adduced” (423:37; o.e.); later in the same paragraph he speaks once again of the “one principle” (424:14).

Section II of the Groundwork

165

the same principle. Thus he writes that “the universal imperative of duty can also be stated as follows” (421:17; o.e.) and then immediately presents the FLN. When he moves on to the FH, he says that if there is “a supreme practical principle, and in regard to the human will a categorical imperative” (428:34; o.e.), it can also be expressed in relation to the idea of an end in itself: “The practical imperative will thus be the following” (429:9; o.e.); and the FH then follows. After the four examples of the FH, we come to that first key passage at 431. This passage may initially suggest that Kant is not distinguishing three formulas for expressing the same principle but rather three different principles (he actually speaks of the ‘fi rst,’ ‘second,’ and ‘third’ principle). But then he describes the FA as the “third formula of the principle” (432:2; o.e.) and in the following remarks speaks once again of ‘the’ CI (expressed in the FA, of which the FH allegedly is only a variant, as we saw). In the overview that Kant himself provides at 436–437 it seems fi nally clear that the FLN, the FH, and the FA are “fundamentally only so many formulas of precisely the same law” (436:9; o.e.), and that this one law can be expressed in “the universal formula of the categorical imperative” (436:30; o.e.), that is, in the FU. In Kant’s concluding elucidations we read accordingly (as cited earlier) that the FH and the FUL (alternatively, the FA) are “fundamentally the same” (438:1; o.e.), for one formula says “just as much” (438:4) as the other. Many commentators infer from Kant’s claim that the formulas are formulas of ‘precisely the same law’ that they must be equivalent. But Kant never says any such thing, and we have already seen that the formulas are assigned different functions: the FUL and the FLN are used to test particular maxims for permissibility in cases where they are suspected of violating an already-assumed duty, whereas the FH is used to ground these same duties. This seems a perfectly straightforward sense in which these formulas are not equivalent. The commentators who claim that they are seem to be assuming that a moral principle must operate by identifying certain specific actions as required, forbidden, and so on, and two formulas would then be equivalent if and only if they assigned the same moral properties to the same actions; but this way of looking at the relation of a fundamental

166

Immanuel Kant’s Groundwork for the Metaphysics of Morals

principle to duties and actions is not Kant’s way. In MS he grounds on the moral principle (mainly the FH) a whole system of distinct ethical duties and holds that these are applied to particular actions only through a faculty of judgment that cannot be reduced to rules or a deductive procedure (MS 6:411). So it looks as if Kant declines to claim that the formulas are equivalent because even the question whether they are equivalent presupposes a model of moral philosophy that he rejects. 2. The function of the formulas: The various formulas, according to Kant, all express ‘fundamentally the same law.’ But he also says (and this explains the qualification ‘fundamentally’ here): “Nonetheless, there is a variety among them, which is to be sure more subjectively than objectively practical, namely, that of bringing the idea of reason nearer to intuition (in accordance with a certain analogy) and through this nearer to feeling” (436:10). After he has then presented the three formulas by reference to the distinction between ‘form’ (the FLN), ‘matter’ (the FH), and ‘complete determination’ (the FRE), Kant continues as follows: “But one does better in moral judging always to proceed in accordance with the strict method and take as ground the universal formula of the categorical imperative: Act in accordance with that maxim which can at the same time make itself into a universal law. But if one wants at the same time to obtain access for the moral law, then it is very useful to take one and the same action through the three named concepts and thus, as far as may be done, to bring the action nearer to intuition” (436:29). Thus Kant’s thesis is clear: the function of the various formulas is solely to bring the moral law (the CI as expressed in the FU) intuitively closer and ‘thus’ to bring out its affective dimension and facilitate ‘access’ for it.91 On the other hand, where ‘moral judging’ is concerned—that is, when we need to evaluate our maxims of action, as Kant clearly illustrates with his celebrated examples92—he recommends that we should al91. For this function with regard to ‘access,’ an important matter for Kant, cf. also 389:32; 405:4; and 409:24. 92. This evaluation of maxims in terms of a test also permits the derivation and relevant division of duties; cf. 424:3, where Kant describes the FUL (alternatively, the FLN) as a “canon of moral judgment” for our maxims (or actions). For ‘judgment’ in

Section II of the Groundwork

167

ways draw on the FU.93 The question whether this ‘intuitive’ presentation in terms of the formulas is actually successful need not detain us here. It is more important to see whether in fact the FU must be drawn on for testing maxims. It is nonetheless worth noting that Kant himself also entrusts the ‘moral judgment’ of maxims to the FH (429–430); and as we have seen—in clear contrast with Kant’s own view—the FH is more promising than the FUL, at least with regard to such ‘judgment.’ 3. What is the FU? Kant formulates the FU at 436–437. Since he clearly separates this universal formula from the three others we have mentioned (the FLN, the FH, and the FRE), it cannot be identical with any one of these formulas. Now it is most implausible to assume that Kant is here introducing a further quite new formula; there is no suggestion to this effect in the text, and indeed at 436–437 he simply summarizes what has already been said. But since, in addition to the FLN, the FH, and the FRE, only the FUL and the FA have been mentioned, it is clear that the FU can be identical only with one of these two. Thus the decisive question is this: is the FU identical with the FUL or with the FA? It is not so easy to answer it, however. The claim that the FU is identical with the FUL can be supported by the following arguments (whether the supposed identity is actually convincing, or even compatible with the prominent position accorded to the FUL, is another question):94 a. Kant introduces the FLN at 421 and the FH at 427–428. Then he provides a preliminary summary of the argument at 431, where he this sense (as the evaluation of maxims), cf. also 390:3; 403:37; 404:17 (“power of judgment”); 404:11 (“faculty of judgment”); 424:3; 426:24; 433:14; 436:30; and 460:4. 93. In this connection there is an interesting passage in the Critique of Practical Reason. Kant is addressing the objection raised against GMS “that no new principle of morality is set forth in it but only a new formula” (KpV:8, footnote). Kant explicitly distinguishes between a principle and a formula and continues: “But whoever knows what a formula means to a mathematician, which determines quite precisely what is to be done to solve a problem and does not let him miss it, will not take a formula which does this with respect to all duty in general as something that is insignificant and can be dispensed with.” 94. Schönecker proposes that the FU should be identified with the FUL, whereas Wood argues for its identification with the FA.

168

Immanuel Kant’s Groundwork for the Metaphysics of Morals

b.

c.

d.

e.

formulates the FLN and the FH again and says: “From this now follows the third practical principle” (431:14; o.e.), namely, the FA. When a few lines later Kant speaks of the “three ways mentioned of representing the principle of morality” (436:8; o.e.), he is obviously referring, as we pointed out, to the formulas that he has already ‘mentioned’ (the FLN, the FH, and the FA). Since, as we showed, the FU cannot be identical with the FRE, while the FRE is a variant of the FA and is also presented as such in Kant’s own threefold enumeration, it would be highly implausible to identify the FU with the FA (this would make sense only if the FA were significantly different from the FRE, which is not the case). The ‘universal’ imperative is therefore the FUL. At 420 Kant sets himself the ‘task’ of determining ‘the’ CI. He accomplishes this, and the formula through which he presents the CI is the FUL (421). And it is this CI—formulated in the FUL—that is described as a ‘single’ one. Not only is this substantively consistent with the assumption that this ‘single’ CI (that is, the FUL) is also the ‘universal’ CI (that is, the FU), but Kant also says so himself, for when he introduces the FLN at 421, he refers unambiguously to the FUL and says that the (this) “universal imperative of duty” (421:17; o.e.) can ‘also’ be expressed as the FLN. The FUL is used (at 421–424) to test (or judge) particular maxims; this might be the very function of the ‘strict method of judgment’ to which the FU is also assigned. The singular character of the CI and its identity with the FU and the FUL are also evident from the fact that the FUL is the most frequently deployed of the formulas; and thus we also find that in the very first derivation of the moral law (at 402) this law (the CI) is likewise presented in terms of the FUL. Last but not least, the way in which the FU itself is formulated might suggest that it is identical with the FUL. Thus the FU reads: ‘Act in accordance with that maxim which can make itself at the same time into a universal law.’ It is difficult to see how this formula is supposed to differ in any essential way from the FUL:

Section II of the Groundwork

169

“Act only in accordance with that maxim through which you can at the same time will that it become a universal law.” Naturally there are some differences here, but such differences can be found even in the various formulations that clearly belong to one formula. The decisive expression as far as the FU and the FUL are concerned is that of ‘universal law,’ which they both have in common. On the other hand, the following arguments can be adduced in support of the claim that the FU is identical with the FA: a. The FU is presented in the same paragraph that is also dedicated to the third formula (in its FRE variant). The claim that we should judge according to the ‘strict method’ can be understood to mean that we should employ the ‘less intuitive’ version of the third formula, namely, the FA. b. The FU tells us that we ought to act in accordance with a maxim that can make itself into a universal law. But this can be taken as equivalent to the FUL only if the expression ‘can make itself into a universal law’ means the same as ‘can be willed as a universal law (without contradiction in thinking or willing).’ This reading is implausible, for there is no good reason, as we saw above, to suppose that every maxim that can be thought or willed without contradiction as a universal law is at the same time a maxim that in fact could also be a universal law; many such maxims could be incompatible with other universal laws. Even less should we suppose that every such maxim could make itself into a universal law. This follows simply from the fact that the universalization tests furnished by the FUL and the FLN show only that certain maxims are permissible; all we learn from a maxim that passes such tests is that one may act in accordance with it, not that we must act in accordance with it. But it would not show that such a maxim could be made into a universal law or could make itself into one. c. There is no difficulty in understanding the idea of a maxim that can ‘make itself into a universal law’ as equivalent to the idea

170

Immanuel Kant’s Groundwork for the Metaphysics of Morals

of a will that “could through its maxim at the same time consider itself as universally legislative” (434:13). The latter formulation is employed as an elucidation of the FA and is closely connected with the thought that is central to the FA, namely, that ‘a will gives itself a universal law through all its maxims.’ Kant also expresses similar thoughts in the two following formulations (which are all concerned with the FA): “Act in accordance with maxims that can at the same time have themselves as universal laws of nature for their object” (437:17); and “Not to choose otherwise than so that the maxims of one’s choice are at the same time comprehended with it in the same volition as universal law” (440:18). (This last formulation is explicitly linked to the thought of autonomy.) 4. The relationship between the formulas at 431 and 436: 95 After he has described the FUL (alternatively, the FLN) and the FH as principles of morality (at 431), Kant continues as follows: “The ground of all practical legislation, namely, lies objectively in the rule and the form of universality, which makes it capable of being a law (at least a law of nature) (in accordance with the fi rst principle [i.e., the FUL, alternatively the FLN]), but subjectively it lies in the end; but the subject of all ends is every rational being as an end in itself (in accordance with the second principle [i.e., the FH]): from this now follows the third practical principle of the will” (431:9), namely, the FA. The fact that Kant speaks of ‘objectivity’ and ‘subjectivity’ in this connection, and especially his claim that the FA ‘follows’ here, immediately recalls the derivation of the concept of duty in GMS I (400). There the third proposition regarding duty (P3) was explicitly 95. There are various other places where Kant alludes to the relationship between the formulas; for reasons of space we cannot discuss these here. (Thus if we assume that Kant is talking about the FUL and the FH at 437:34–438:7, these two formulas are practically identified with one another as “fundamentally the same” [438:1].) In another passage Kant also seems to identify the FUL with the FA. The “principle” mentioned at 434:10 appears to be the FUL, which is treated—through a “hence only so” (434:12)— as equivalent to the FA. Finally, Kant describes the FUL as the “formal principle” (438:21) of maxims as part of a realm of ends, insofar as this realm has an “analogy” (438:23) with a realm of nature; and this analogy is rooted in the close relationship between the FUL and the FLN.

Section II of the Groundwork

171

understood as a “consequence” (400:17), and indeed as a consequence of two propositions, one of which (P1) concerns the ‘subjective’ dimension, while the other (P2) concerns the ‘objective’ dimension. As we remember, what determines the will is ‘objectively the law, subjectively pure respect for this practical law’; ‘objective’ here refers to the substantive determination of duty, while ‘subjective’ refers to the motive of the action. The observations at 431 can also be read in this sense: the FLN emphasizes the substantive aspect of the determination of duty through a law, while the FH emphasizes the ends that must be involved in morally motivated actions, just as they are in every action. The FA ‘follows,’ as Kant says, ‘from this’ (i.e., from the FUL/FLN and the FH) in the same way in which P3 follows as a ‘consequence’ of P1 and P2, which suggests that we are talking not about a logical derivation here, but about a connection or combination of two aspects.96 And in fact this makes sense, for the FA combines the thought of the FH that ‘every rational being as end in itself ’ is the subject of all ends with the central thought of the FUL that every maxim as a law must possess the ‘form of universality.’ The FA contains the “idea of the will of every rational being as a will giving universal law” (431:16). The ‘subjective’ and the ‘objective’ aspect are united in the FA because the autonomous will is nothing other than the ‘end in itself’ for the sake of which we act; and the law that this will establishes, and that ought to determine its actions, is the CI. We find the same or at least a very similar thought at 436–437. Kant says that every maxim has a ‘form’ and a ‘matter,’ and that all maxims are capable of a ‘complete determination’;97 and he relates the FLN, the FH, and the FRE to all these aspects. In this context he makes the rather enigmatic observation that the FLN, the FH, and the FRE are not merely a series but involve a certain “progression” (436:26), one that is specifically described in analogy with the

96. Recall, however, that a reconstruction in terms of a logical deduction proper seems possible. 97. Kant here alludes to concepts developed in the context of his theoretical philosophy; cf. KrV:A 266–267/B 322ff. and A 571ff./B 599ff.

172

Immanuel Kant’s Groundwork for the Metaphysics of Morals

categories of quantity (unity, plurality, and allness or totality).98 Since Kant claims in the KrV that “the third category in each class always arises from the combination of the second category with the first” (B 110), one could then understand the ‘progression’ in this way: the FRE (thus, effectively, the FA) arises from the ‘combination’ of the FLN and the FH (in the sense of the ‘follows’ at 431:9); and his claim with regard to the FLN, the FH, and the FRE that “one [of them] of itself unites the other two in itself ” (436:10) would mean that the FRE (the ‘one’) unites the FLN and the FH (‘the other two’) ‘in itself’; but it does not say (as some English translations of this passage inaccurately do) that each of the formulas ‘unites’ the other two ‘in itself.’

3.5 Summary 1. Imperatives are objective laws of reason that are valid for all rational beings without distinction. Since rational beings that can be influenced by sensuous factors do not always act rationally, imperatives express an ‘ought’ in relation to such beings (i.e., involve a kind of ‘necessitation’). 2. Categorical imperatives command actions without regard to the subjective interests of the agents. In this sense the CI is a priori. It is synthetic insofar as the will of a being at once rational and sensuous does not always will the morally good but must be ‘connected’ with the latter. 3. It is necessary to distinguish between a concrete specification of means (Do x!) and the formula for the hypothetical imperative (If x is a necessary means for z, and you can do x, and you actually will z, then do x!). Imperatives of skill relate to some possible end that is presupposed, while imperatives of prudence relate to happiness as an end which can actually be presupposed for all human beings (happiness is understood as the lasting and harmonious satisfaction of all our inclinations; as a presupposed end, happiness is actually to be understood as a normative command). Hypothetical imperatives are 98. Kant himself speaks of a “certain analogy” (436:12) here; cf. also the analogy suggested at 454:15–19 (“approximately in this way”).

Section II of the Groundwork

173

analytic (and thereby ‘possible’) insofar as the act of setting oneself an end is bound up with setting oneself the norm of adopting the corresponding means. 4. Kant himself speaks of a ‘single’ universal CI (the FU) but also of ‘three’ formulas that are supposed to present the moral law in a vivid and intuitive manner. In fact, we are offered at least five formulas: the universalization formula (the FUL), the law-of-nature formula (the FLN), the end-in-itself formula (the FH), the autonomy formula (the FA), and the realm-of-ends formula (the FRE). The precise relationship among these formulas ultimately remains rather unclear (although it looks as if the FA connects the FU, alternatively the FLN, with the FH). It remains a matter of particular controversy whether the FU should be identified with the FUL or with the FA. 5. The basic testing procedure of the FUL/FLN is supposed to involve bringing maxims, as subjective rules of action, into a properly universalized form (according to Kant, this is best accomplished if such maxims are envisaged as laws of nature that would determine the action of the agents) in order to see whether the universalized maxim can be thought and willed without contradiction. Here we must specifically distinguish between a contradiction involved in thinking and a contradiction involved in willing. Kant believes that the fi rst of these tests corresponds to perfect duties, while the second corresponds to imperfect duties. But none of Kant’s examples are convincing. The correspondence thesis cannot be substantiated anyway, for only a single positive duty can be derived from the FUL/FLN, and it turns out that the universalizability tests in the FUL/FLN admit of false positives and false negatives. Thus the attention that is almost always accorded to the FUL (or the FLN) is problematic from the substantive point of view and unconvincing from an interpretive perspective since Kant himself lays considerable weight on the other formulas. 6. It is FH that expresses the core idea of Kant’s ethics: autonomous beings (i.e., persons) are ends in themselves; in contrast to the price accorded to things, they possess an absolute worth and dignity. But Kant hardly grounds this thesis (which should be read as a kind of moral

174

Immanuel Kant’s Groundwork for the Metaphysics of Morals

realism) as such. It follows from the character of autonomous beings as ends in themselves, according to Kant, that they must never be treated merely as means but invariably also as an end. Maxims should be examined in terms of whether they express respect for the dignity of autonomous beings. But there is no criterion that enables us to recognize if and when they do so. The FA is stronger than the FUL and the FH, for this formula not only tells us which maxims can be thought or willed but also which maxims in fact ought to be binding on us. The FRE, which is bound up with the FA, involves very considerable difficulties of interpretation.

4 Section III of the Groundwork The Deduction of the Categorical Imperative

We shall begin by looking at the structure and the task of GMS III (4.1); it is particularly important to do this because significant misunderstandings can easily arise here. Among other things, this final section is centrally concerned with freedom and its relationship to morality (4.2). The other, quite decisive point concerns Kant’s demonstration of the validity of the CI, or what he calls its ‘deduction’ (4.3). Finally, as before, we offer a brief summary of the argument (4.4).

4.1 The Structure and Task of GMS III The Meaning of the Question “How Is a Categorical Imperative Possible?” In GMS III the question regarding the possibility of the CI furnishes the title of Section 4 of the text, and there is no doubt that the “deduction” mentioned in the text (447:22; 454:21; 463:21) is intended to answer this specific question;1 the second paragraph of Section 4 begins accordingly

1. When we speak of the ‘question of deduction’ here, we refer to this question how categorical imperatives are possible. It is clear that with this specific formulation Kant is alluding to the parallel question already posed in the context of theoretical philosophy (namely, how are theoretical synthetic a priori propositions possible?). In his preparatory studies for the Prolegomena Kant had already noted a “remarkable similarity”

176

Immanuel Kant’s Groundwork for the Metaphysics of Morals

as follows: “And thus categorical imperatives are possible through the fact that” (454:6; o.e.). But at the same time, in Section 5, we are also presented with a detailed explanation of why the question of deduction can be answered only in part: “Thus the question, ‘How is a categorical imperative possible?’ can be answered to this extent” (461:7; o.e.). But to what extent precisely can it be answered, and in what respects can it not be answered? Section 5 is concerned, as its title says, with “the uttermost boundary of all practical philosophy.” This limit or boundary lies in the fact that the question how pure reason can possess a practically motivating power cannot be answered.2 Kant’s basic thesis is that we cannot possibly explain how pure practical reason itself can actually bring about our obedience to the moral law. But this impossibility depends solely on the specifically Kantian concept of ‘explanation.’ Kant does not say that we cannot entertain any concept of the motivating power of practical reason; indeed, he says that we possess the relevant concept (namely, the concept of respect). But it remains true that “we can explain nothing unless we can trace it back to laws the object of which can be given in some possible experience” (459:3; o.e.). Thus the “subjective impossibility of explaining the freedom of the will is the same as the impossibility of bringing to light and making comprehensible an interest that the human being could take in moral laws” (459:32). Thus it is ‘to this extent’ that the question concerning the possibility of the CI is incapable of being answered. But this question can be answered insofar as we are concerned with the validity of the CI and with freedom as the metaphysical presupposition of the latter: Thus the question, “How is a categorical imperative possible?” can be answered to this extent: one can state the sole presupposition under which alone it is possible, namely the idea of freedom, and to the extent that one can have insight into the necessity of this presupposition, which is sufficient for the practical use of reason, i.e., for the conviction of the validity of this imperative, hence also of the moral law; but

(AA 23:60) between these two questions. We shall not undertake to examine this parallel in detail here. 2. Cf. Kant’s formulations at 458:37; 459:34; 460:10; 461:25; and 461:32.

Section III of the Groundwork

177

how this presupposition itself is possible, no insight into that can be gained through any human reason. (461:7; o.e., Kant’s initial emphasis omitted in the second part) Thus the question ‘How is a categorical imperative possible?’ really involves three questions: (1) Why is the CI valid? (2) How can we meaningfully conceive of freedom, and why are we entitled to regard ourselves as free? (3) How can pure reason bring about an interest in the moral law? As we have seen, Kant thinks that the third question cannot be answered. He answers the second question in Sections 2 and 3 of the text: freedom and morality are analytically connected with each other, and we may regard ourselves as free on the basis of our spontaneity. The first question is answered in Section 4. It amounts to the question ‘Why be moral?’ Since ‘being moral’ for Kant is equivalent to ‘acting from duty (for the sake of morality),’ the meaning of this question cannot lie in the usefulness of obeying moral laws; for that would directly contradict the meaning of ‘being moral.’ A moral skeptic who asks why he should act as he ought to act is ignoring Kant’s dictum that the moral law “does not have validity for us because it interests us” (460:24). All that we can ask, if we accept Kant’s analysis of the meaning of moral concepts in GMS I, is whether the moral law is valid in fact. Until this has been shown, it still seems possible that morality might be a mere “figment of the mind” (407:17; 445:8). We saw that Kant was already concerned with the ground of the obligatory character 3 of the CI in the Preface. Kant also speaks of the “reality” 4 of the CI, of its actuality,5 its validity,6 its rightness,7 and its objective necessity;8 he says that it is,9 that it actually is [wirklich stattfinde],10 and that the human being 3. For the concept of obligation, cf. 389:12; 389:16; 391:11; 432:31; 439:31; 439:33; and 448:34. 4. Cf. 425:14 and 449:26. 5. Cf. 406:15 and 420:1. 6. Cf., in par tic u lar, 389:12; 389:14; 403:7; 408:18; 412:3; 424:35; 425:18; 442:8; 447:32; 448:6; 448:32; 449:29; 460:25; 461:3; and 461:12. 7. Cf. 392:13. 8. Cf., in par tic u lar, 442:9; 449:26; and 449:30. 9. Cf. 419:18. 10. Cf. 425:9.

178

Immanuel Kant’s Groundwork for the Metaphysics of Morals

is subject 11 to it. All these concepts and expressions can be subsumed under the subsequent formula of the “validity of this imperative” (461:12). And just as he employs this multiplicity of concepts and expressions, Kant speaks not only of a ‘deduction’ but also of a ‘demonstration’ (and so on) that is to be accomplished.12 It is the validity of the CI, therefore, that is to be demonstrated, and the presupposition under which the validity of the CI is possible, namely, the idea of freedom, must also be justified. How does the question of deduction come to acquire this twofold meaning? Kant claims that action that possesses a morally unlimited worth consists in action that is done from duty. To act from duty is to act without regard to our own subjective interests, to act for the sake of the law. Kant must therefore show (in contrast to the Humean model) that human beings are free in a negative and also a positive sense, that is, are capable of acting without being directly or ineluctably determined in their action by subjective interests. He must show that we are able to act morally (independently of interests). But that does not yet suffice to demonstrate that we also ought to act morally, which is why the deduction has two tasks to fulfill.

The Structure of GMS III Kant begins in Section 1 with the thesis of analyticity: “Thus if freedom of the will is presupposed, then morality follows from it, together with its principle, from the mere analysis of its concept” (447:8; o.e.). Now if Section 2 shows (as its title promises) that freedom, as a property of the will of every rational being, must in fact be presupposed, then at the end of this subsection (or at least in Section 3) ‘morality, together with its principle,’ must have already ‘followed.’ This is exactly how Kant’s argument has repeatedly been construed: a free will is a will that stands under the moral law; freedom must be presupposed as a property of the will of rational beings; therefore, the human will stands under the moral law, and the CI is valid. If this reading were the whole story, however, Section 4 would be entirely superfluous, even though Kant answers the question of deduction only here in this subsection.

11. Cf. 449:12. 12. Cf. 392:4; 392:13; 403:27; 412:2– 8; 425:8; 425:15; 427:17; 431:33; 440:20–28; 445:1; 447:30–448:4; and 449:27.

Section III of the Groundwork

179

This interpretation principally derives from a misunderstanding of the thesis of analyticity by which it is taken to mean that with the freedom of the human will the validity of the categorical imperative for the human being is thereby also demonstrated; if this were so, then this validity would already be demonstrated by the end of Section 3. But Kant explicitly denies this. Something must therefore still be missing at the end of this subsection. In Section 1 Kant shows that a perfectly free and rational will is a will that always acts morally (the thesis of analyticity). Section 2 shows that a rational being must take itself to be theoretically and practically free on the basis of the spontaneity of its reason. In Section 3, by appeal to the distinction between the world of understanding and the world of sense, Kant shows that human beings may take themselves to be free in this sense too. But showing that the human being is free does not suffice to prove the validity of the CI, for the thesis of analyticity indicates only that perfectly free and rational beings act morally. Human beings, however, are beings that are at once sensuous and rational. For beings like these, which do not always act as they ought to act, it must still be shown why they ought to act as they are able to act. This is what Kant intends to accomplish in Section 4.

4.2 Freedom and Morality We shall now examine the thesis of analyticity and its justification as presented in Section 1 (4.2.1). Then we shall proceed to an analysis of Sections 2 and 3 (4.2.2). A key question in this connection will be the suspicion of circularity in Kant’s argument.

4.2.1 The Thesis of Analyticity The principal thesis advanced in Section 1 is the following: Thus a free will and a will under moral laws are the same. (447:6) This is what we shall call Kant’s thesis of analyticity. It needs to be read in the light of Kant’s repeated observation that for purely rational beings the moral law is not an imperative because such beings always necessarily act according to the moral law anyway. The thesis of analyticity basically says

180

Immanuel Kant’s Groundwork for the Metaphysics of Morals

nothing else; the only addition is that the notion of freedom is also in play here. In this regard Kant is still concerned with the ‘analysis of concepts.’13 What is ‘analyzed’ here is the concept of freedom. And all Kant wants to show is this: if there is any will that is free, then it is a will under the moral law. Whether there is in fact a free will is a question that is answered only later.

The Meaning of the Thesis of Analyticity The thesis of analyticity cannot mean that sensuous-rational beings that possess a free will always act morally, for that is manifestly not the case. But the thesis as such also cannot (and does not) mean that the free will of such a being would stand “under” (447:7) the moral law as a CI insofar as the CI is valid for such a being. Of course, it is (later) Kant’s claim that the free will of a sensuous-rational being is under the CI (which is to say that the CI is valid for such a being); however, this is not what he claims with the thesis of analyticity in Section 1, for in Section 3 Kant explicitly says that this validity has not yet been demonstrated: “But why ought I to subject myself to this [moral] principle” (449:11)? In Section 3 Kant is still asking “from whence the moral law obligates” (450:16); up to this point he has still provided “no satisfactory answer” (450:2) to the question of the deduction. Indeed, this validity of the moral law as an imperative is demonstrated only in Section 4. The free will of a sensuous-rational being and a will under moral laws are therefore not ‘the same’; that they ought to be the same is precisely what is still to be shown in the deduction of the CI. In the context of the thesis of analyticity, the will therefore cannot simply or directly be understood as the will of a sensuous-rational being. It must be understood as the will of a purely rational being whose free will is always a good will, or else as the will of a sensuous-rational being whose will is at the same time a member both of the world of sense and of the world of understanding and whose pure will as a member of the world of understanding is “the idea of reason, which would have full control over all subjective motivations” (420:31; o.e.). It is in precisely this sense that Kant says in the deduction that the “categorical ‘ought’ represents a synthetic proposition a priori by the fact that to my will affected through 13. Cf. 440:23; 440:29; 447:9; and 447:12.

Section III of the Groundwork

181

sensible desires there is also added the idea of precisely the same will, but one belonging to the world of understanding, a pure will, practical for itself ” (454:11; second emphasis ours). Correspondingly, at the point where Kant first explicates the synthetic character of the CI, he says that the CI is “a practical proposition that does not derive the volition of an action analytically from any other volition already presupposed (for we have no such perfect will), but is immediately connected with the concept of the will of a rational being, as something not contained in it” (420:32; o.e.). The volition of a good action is thus ‘not contained’ in the will of a sensuous-rational will and cannot be ‘derived’ from it. The second paragraph of Section 1 ends with the thesis of analyticity (‘Thus a free will and a will under moral laws are the same’). In the next (third) paragraph Kant continues: Thus if freedom of the will is presupposed, then morality follows from it, together with its principle, from the mere analysis of its concept. (447:8) Since the ‘thus’ (or the German also) can have two possible meanings—it can be used both in an explicative sense and to mark a consequence— there are two plausible ways to read this sentence (both of which confi rm that Kant argues for a thesis of analyticity in the way we read it): On the assumption that the also is merely explicative, this sentence just reformulates the thesis of analyticity: ‘If the concept of the freedom of the will is presupposed, then the concept of morality together with its principle (i.e., the moral law) follows from that concept by mere analysis.’ Again, the moral law does not ‘follow’ from the ‘mere analysis’ of the concept of the freedom of an imperfect will. It follows only in the case of a perfect or pure will. Hence for a perfectly rational being the moral law is an analytic, or a merely descriptive, proposition. It is analytic because it follows from the analysis of the concept of the freedom of a purely rational will that this latter wills the good; it is descriptive because the proposition ‘A perfectly rational will always wills the good’ just describes such a purely rational will. The thesis of analyticity thus implies the following: a will that is characterized solely by reason and freedom is a will that always wills the good and in this sense is analytically connected with the moral law.

182

Immanuel Kant’s Groundwork for the Metaphysics of Morals

On the assumption that the also indicates that there is some sort of conclusion rather than a mere explication, the reading goes like this: Kant states the thesis of analyticity (‘A free will and a will under moral laws are the same’). He then points out that if that is true, then once we know that there really is a free will, we know that it is under the moral law. So what Kant says on this reading is this: ‘Thus if the reality of the freedom of the will is presupposed, then the reality of moral law follows from this reality from the mere analysis of the concept of freedom.’ There is no contradiction between these two readings; as a matter of fact, the second reading just as well confirms our interpretation of the thesis of analyticity. One may conclude from the reality of a free will that the moral law also is real only if there is an analytic relation between the concept of freedom and the concept of morality; but there is one, says Kant, because according to the thesis of analyticity the mere analysis of the concept of ‘freedom of the will’ implies analytically the concept of morality (the concept of ‘being under the moral law’). Two points speak in favor of this second interpretation: First, once we know that there is such a thing as a free will (or at least that we must presuppose that there is such a thing), we know that it is under the moral law; and then Kant argues in Section 2 (as we will see soon in more detail) that— as the title of Section 2 already says—indeed ‘freedom must be presupposed as a quality of the will of all rational beings’ (o.e.); and he then demonstrates in Section 3 that the human being also must understand itself as a being with a free will. Second, this fits with Kant’s diagnosis at the end of Chapter 2 that the CI could still be a ‘figment,’ that is, not real. Immediately after stating the thesis of analyticity, Kant continues: Nonetheless, the latter is always a synthetic proposition: an absolutely good will is that whose maxim can always contain itself considered as a universal law, for through analysis of the concept of an absolutely good will that quality of the maxim cannot be found. (447:10) So far Kant has argued only for this: if we presuppose the concept of a perfectly free will or the will of a sensous-rational being considered just as a rational being, then the very concept of such a free will implies that it is under the moral law; hence (also) once we know that a being really has a

Section III of the Groundwork

183

free will, we know that it really is under the moral law. However, by this argument Kant has not yet shown that the moral law as a CI is real. This is why Kant says ‘nonetheless’: he is distinguishing the synthetic character of the CI from the analytic character of the ‘principle of morality’ (mentioned in the previous sentence); thus one must read: ‘Nonetheless, for sensuous- rational beings the moral law [the latter] is always a synthetic proposition.’ Whereas the ‘principle’ in question (the moral law as a nonimperatival, descriptive law) follows through conceptual analysis from the presupposition of the freedom of the will, it does not follow through analysis of the concept of an absolutely good will that every maxim of such a will ‘can contain itself considered as a universal law.’ This in turn means that the proposition after the colon (in the quotation above: ‘an absolutely good will . . .’) is a synthetic proposition, that is, it is the CI; indeed, Kant explains why analysis does not suffice here, which explanation specifically implies that the proposition is synthetic rather than analytic. Also, note that Kant then continues: “Such synthetic propositions are possible only” (447:14). Since such ‘synthetic propositions’ are very briefly thereafter identified with the “categorical imperative” (447:24), not only must that sentence after the colon (‘an absolutely good will is that whose maxim can always contain itself considered as a universal law’) be understood as a synthetic proposition, but it also becomes evident that Kant does not mean to identify here the moral law as a nonimperatival law but as a CI, for the question how synthetic-practical moral laws are possible refers only and exclusively to the possibility of the CI and not to the moral law as a nonimperatival law. Furthermore, one must recognize that the concept of an ‘absolutely good will’ here does not mean the same as the will of a perfectly rational being or the intelligible will, for if we ‘analyze’ the will of such a being, we specifically see that its maxim can always be a universal law.14

14. Kant’s talk of an ‘absolutely good will’ at this point has contributed to the misunderstanding of the thesis of analyticity because Kant also uses the same expression with reference to the perfect (‘holy’) will (cf. 439:29–34). But he also describes an imperfect will as ‘absolutely good’: “That will is absolutely good which cannot be evil, hence whose maxim, if it is made into a universal law, can never confl ict with itself ” (437:6). At this point Kant clearly does not speak of a perfectly good (holy or purely rational)

184

Immanuel Kant’s Groundwork for the Metaphysics of Morals

The general argument of GMS III further supports this interpretation of the thesis of analyticity. After Kant has dealt with the potential charge of circularity, he concludes Section 3 with the following observation: For now we see that if we think of ourselves as free, then we transport ourselves as members into the world of understanding and cognize the autonomy of the will, together with its consequence, morality; but if we think of ourselves as obligated by duty, then we consider ourselves as belonging to the world of sense and yet at the same time to the world of understanding. (453:11) As in Section 1 (447:8: “morality follows”), autonomy and morality are understood as a ‘consequence’ that follows from (the reality) of the freedom of the will. But the will is grasped here merely as a ‘member of the world of understanding’ (by ‘member of the world of understanding’ Kant alludes to the status either of a purely rational being with a perfect will or else to the status of a being that is considered in its capacity as free and rational). The parallel character of Kant’s formulation here is highly instructive: He is not contrasting the expression ‘if we think of ourselves as free’ with something like ‘if we think of ourselves as unfree.’ He is contrasting it with the expression ‘but if we think of ourselves as obligated by duty.’ Freedom is juxtaposed not with unfreedom but rather with duty. Since, on Kant’s view, morality for sensuous-rational beings is inconceivable without freedom, the latter, understood as a contrast to the concept of duty, in a certain sense cannot be entirely identical with that freedom that is the presupposition and enabling ground of morality for sensuous-rational beings. It must be freedom as the capacity of a being that is purely rational or is considered as a purely rational being. But a being that thinks of itself ‘as obligated by duty’ must also be free. Hence ‘thinking of ourselves as obligated by duty’ certainly means ‘considering ourselves as belonging to the world of sense,’ but it also means ‘thinking of ourselves at the same time as belonging to the world of understanding.’15 The first part of the quotation

will. He is referring to the will that is described as ‘absolutely good’ only if its maxim can be universalized; cf. also 426:10; 437:24; 437:32; and 444:28. 15. Cf. the parallel passages at 454:9; 455:9; and 462:31.

Section III of the Groundwork

185

above (‘For now we see . . . morality’) is a restatement of the thesis of analyticity, but the clause after the semicolon (‘but if we think . . . to the world of understanding’) makes it clear why the moral law is an imperative and thus a duty for a sensuous-rational being. Such a being certainly thinks of itself as ‘belonging to the world of understanding’ and thereby as free. But it also thinks of itself at the same time as belonging ‘to the world of sense’ and thereby thinks of itself as obligated by duty; that is, it thinks the moral law as an imperative. The thesis of analyticity also reappears in the actual deduction provided in Section 4: “As a mere member of the world of understanding, all my actions would be perfectly in accord with the principle of the autonomy of the pure will” (453:25; o.e.). Accordingly, a little further on we read: “And thus categorical imperatives are possible through the fact that the idea of freedom makes me into a member of an intelligible world, through which, if I were that alone, all my actions would always be in accord with the autonomy of the will” (454:6; initial emphasis ours). At the very end of this subsection Kant writes: “The moral ‘ought’ is thus his own necessary volition as a member of an intelligible world and is thought by him as an ought only insofar as he at the same time considers himself as a member of the sensible world” (455:7; o.e.). And again: “Under the presupposition of freedom of the will of an intelligence, its [the will’s] autonomy, as the formal condition under which alone it [the will] can be determined, is a necessary consequence” (461:14). By an ‘intelligence’ Kant means a rational being that considers itself solely as a member of the world of understanding. From this perspective, the human being also understands its will merely as a free or pure will, and such a will always wills the good (as a ‘mere member of the world of understanding all its actions would always be in accord with the principle of autonomy’). That is the meaning of Kant’s thesis of analyticity. But what is the precise argument for this thesis?

Kant’s Argument for the Thesis of Analyticity Kant says that a “positive concept of freedom” (446:14) flows “from” (446:14; o.e.) the ‘negative definition’ of freedom; and this positive concept then reveals itself as the concept of “autonomy” (447:1). It is clear that in speaking of a ‘negative definition’ of freedom here Kant is alluding to the theory of

186

Immanuel Kant’s Groundwork for the Metaphysics of Morals

freedom presented in the Critique of Pure Reason.16 It is helpful to recall the basic thought behind that theory here: the world as it is understood in terms of natural science is a completely determined world; in this world every event has a cause that itself has a cause, and so on. But we must distinguish the sensible world from the world of things in themselves (the world of understanding). If human beings were simply part of the world of appearances (the sensible world), then they would be considered unfree. But human freedom can at least be conceived without contradiction insofar as we consider the human being as a member of the world of understanding. In this connection Kant understands transcendental freedom as the capacity to produce an effect “entirely of itself ” (KrV:A534/B562), a capacity that Kant also describes as spontaneity or self-activity. In relation to human beings this freedom can be called practical freedom. Thus although an action, insofar as it is part of the sensible world, can be explained entirely by reference to appearances, it is also conceivable that it has an intelligible cause. Freedom is understood here as a property of reason, and reason is understood as a faculty that is not subject to any determination through the causality of nature but is capable of producing effects ‘of itself ’ on account of reasons and its own grounds. Freedom is therefore “the property of a thing in itself ” (PM:345). The rational subject in turn, which acts on the grounds of reason, is considered in this activity as a “thing in itself ” (A539/B567). In this sense, actions that transpire by reference to the CI (actions done from duty) are free actions from a transcendental-practical point of view.17 The distinction between positive and negative freedom is also important here: in the negative sense, practical freedom signifies independence from all natural determinations, while in the positive sense it signifies the capacity to initiate something ‘entirely of itself.’ It is central to Kant’s theory that transcendental-practical freedom can be neither proved nor explained by reference to experience. Kant also repeats this point again and again in GMS III, and it ultimately lies behind his thesis that we cannot answer the question how pure reason can be practical. 16. Cf. in par ticu lar the ‘third antinomy’ (A444ff./B472ff.) and the ‘solution’ that Kant proposes (A532ff./B560ff.). 17. Whether actions performed by reference to hypothetical imperatives must be included in the domain of free actions is not always clear in Kant, as we have seen; in KU, at least, he appears to deny this.

Section III of the Groundwork

187

There are two further theses from the KrV that appear directly in Section 1 of our text. First, Kant claims that there are only two forms of causality. If something is a ‘cause’ (or ‘has causality’), then it is either a free cause (is unconditioned or independent of the causality of nature) or a natural cause (a cause that is in turn conditioned by another cause, ‘and so on’).18 Second, Kant claims that the property of being a cause must always exhibit a law-like character: “Since the concept of a causality carries with it that of laws in accordance with which must be posited, through that which we call a cause, something else, namely its result; therefore freedom, even though it is not a quality of the will in accordance with natural laws, is not for this reason lawless, but rather it has to be a causality in accordance with unchangeable laws, but of a particular kind; for otherwise a free will would be an impossibility” (446:15).19 Kant does not ground these two theses in any way here, but his argument cannot function without them. This conceptual distinction of ‘negative’ and ‘positive’ freedom actually appears in these terms in only a single passage of KrV, and there it is applied to reason itself and its freedom.20 The decisive point here is that the negative concept of freedom that belongs to reason (‘independence of empirical conditions’) cannot be separated from the positive concept of freedom (‘action out of itself’): it is both aspects together that constitute the practical concept of freedom. Hence practical freedom is the capacity “to determine itself from itself [the positive aspect] quite independently of necessitation through sensuous impulses [the negative aspect]” (KrV:A534/ B562; o.e.). Kant defi nes his terms in such a way that the negative concept of freedom already rules out any role for the prudential or instrumental function of reason here. The negative freedom of reason consists precisely in the fact that its ‘causality’—its capacity to produce actions in accordance with rules—is free of any “alien causes determining it” (446:9). These ‘alien causes’ that can determine us are interests and inclinations. Thus the negative concept of freedom guarantees that “practical reason (will) does not merely administer some other interest” (441:17). But if reason exercises

18. Cf. Kant’s conceptual elucidations in KrV (A419/B447). 19. Exactly the same thought is to be found in KrV (A539/B567). 20. CF. KrV (A553ff. /B581–582).

188

Immanuel Kant’s Groundwork for the Metaphysics of Morals

either an instrumental-prudential function or a moral function, and if the negative concept of freedom saves it from ‘merely administering some other interest,’ then the positive concept of freedom can be synonymous only with autonomy. That is why the positive understanding of freedom also ‘flows from’ the negative definition of freedom, and why the negative concept of freedom is also “the key to the definition of autonomy of the will” (446:6; o.e.): Natural necessity was a heteronomy of efficient causes; for every effect was possible only in accordance with the law that something else determined the efficient cause to causality; what else, then, could the freedom of the will be, except autonomy, i.e. the quality of the will of being a law to itself? But the proposition: “The will is in all actions a law to itself,” designates only the principle of acting in accordance with no other maxim than that which can also have itself as a universal law as its object. (446:21) Kant speaks rather as if it were already obvious that freedom of the will must be understood as autonomy (‘what else, then, could the freedom of the will be, except autonomy?’), and indeed autonomy understood as giving the moral law to oneself. But why is it not possible for perfectly free practical reason to will what is morally wrong? The answer to this question, for Kant, can be only that the root of all wrong or evil lies in the heteronomy of the will. Heteronomy means alien determination through the causality of nature; that is why the actions of a being that is simply a member of the sensible world “would have to be taken as entirely in accord with the natural law of desires and inclinations, hence with the heteronomy of nature” (453:28). In the opening paragraph of Section 1 Kant describes the dependence on alien determining causes as ‘natural necessity’; and a few lines later he describes this natural necessity as a ‘heteronomy of efficient causes.’ If the will is free of this natural necessity, it cannot be said to will what is wrong or evil, for this must be traced back entirely to heteronomy; then it can will only the good. What is more, there can be no such thing as some alternative ‘law of evil.’ The negative concept of freedom makes it clear that if practical reason is free in the negative sense, it is thereby free of ‘desires and inclinations.’ But what follows, then, is this: “Since I

Section III of the Groundwork

189

have robbed the will of every impulse that could have arisen from the obedience to any law, there is nothing left over except the universal lawfulness of the action in general which alone is to serve the will as its principle, i.e. I ought never to conduct myself except so that I could also will that my maxim become a universal law” (402:4; o.e., Kant’s emphasis omitted).21 Of course, there can be such a thing as morally reprehensible maxims, that is, maxims that cannot be universal laws. But the whole point is that they cannot become such. Maxims that are reprehensible can always be traced back to desires and inclinations, that is, to natural causes. If reason can function as a cause, it must act through the power of laws; there can be no law of evil; practical reason is independent of natural laws; but mere lawfulness—that which characterizes reason as reason—leads to the moral law; free practical reason, according to Kant, is therefore reason that stands under the moral law.

Diffi culties with the Thesis of Analyticity Kant’s thesis of analyticity and his attempt to ground it invite a number of possible objections. It seems, we might say, to involve considerable metaphysical ballast. Two problems in particular are evident here: 1. If a free will is an autonomous will and thereby stands ‘under moral laws,’ how can such a will ever will something that is immoral? Either what is wrong or evil can spring from the free will, in which case the latter cannot be synonymous with autonomy, for autonomy just is the moral self-determination of the will; or what is wrong or evil cannot spring from the free will, in which case wrong or evil actions cannot be thought of as free actions, and we can no longer assign any meaning at all to the concept of responsibility. Kant renders the problem even more acute insofar as he understands heteronomy precisely as an act of alien or extraneous determination: “The will does not give itself the law but the object through its relation to the will gives the law to it” (441:7; o.e.). If an object (of inclination) determines the will, the latter “never determines itself immediately through the representation of the action, but only through the incentive” 21. Cf. 420:24–421:5. We have already discussed this ‘derivation’ of the CI above.

190

Immanuel Kant’s Groundwork for the Metaphysics of Morals

(444:9); the heteronomous “will does not give the law to itself, but rather an alien impulse gives it by means of the subject’s nature, which is attuned to the receptiveness of the will” (444:25; o.e.).22 2. The whole problem of wrong or evil action in GMS can ultimately be traced back to the problem that Kant conceives the free human will as a member of the world of understanding and the evil human will as a member of the world of sense, and that he never adequately explains how we are to understand the relationship of these two ‘wills’ to each other and their relationship in each case to the human person as a whole. Kant describes freedom as the property of a will that indeed on the one hand belongs entirely to the world of understanding but which, on the other hand, is at the same time the will of a sensuous-rational being. Thus the question is this: to whom or to what exactly is the quality or property of freedom properly ascribed? The will as a member of the world of understanding, it seems, ever wills only the good; for such a will, the moral law is not an imperative at all. But for the whole person, the one to whom this will belongs, the moral law is an imperative. But what does it really mean to say that for the intelligible will of a sensuous-rational being the moral law is not an imperative, whereas for the human person that is precisely what it is? If the free will qua free will always wills the good, then what is the relationship between this free will and the sensuousrational person whose will it is, and for whom the moral law presents itself as a command? Is the freedom of the person, who decides for or against the moral law, distinct from the freedom of the intelligible will of this same person? This free will wills the morally good, but the person whose will it is does not always will the good. Nonethe22. That is also why Kant specifically emphasizes (as we have already seen) that the question how from “desires and inclinations, and from these, again, through the cooperation of reason, maxims arise . . . belongs to an empirical doctrine of the soul, which constitutes the second part of the doctrine of nature, if one considers it as philosophy of nature insofar as it is grounded on empirical laws” (427:8). Even if reason can cooperate in the case of heteronomous actions, such actions are a matter for the philosophy of nature, according to GMS. In this sense, in GMS III, Kant effectively identifies heteronomy with natural necessity; cf. 446:22; 452:28; 453:29; 458:33; and 460:25. That is why the “scoundrel” (454:21) is also explicitly described as a “member of the sensible world” (455:5; o.e.).

Section III of the Groundwork

191

less, the CI is addressed to the person rather than the will, for it is from this will that the law springs. But how is the freedom of the person, which allows us to decide for or against the moral law, to be understood if the will alone is free, the will whose law this moral law is? Kant provides no answer to these questions in GMS.23

4.2.2 The Presupposition of Freedom Sections 2 and 3 of our text are also highly compressed and difficult to interpret. In the general context of our commentary we can do little more than provide a rough outline of the discussion. The central question is whether in GMS III Kant possesses an argument for the freedom of the will that does not already presuppose the validity of the moral law.

23. This is particularly evident at 457:25–458:5. We should note, above all, the formulation that speaks of a “will that lets nothing be put to its account that belongs merely to its desires and inclinations” (457:25). The refusal to let something be ‘put to its account’ means that one is not responsible for something, that one cannot be called to account in this regard. This means in turn that the will is not responsible for the desires and inclinations that belong to the human being. And indeed in the next sentence the term “human being” (457:34) is reserved for the person qua ‘appearance’; this fits with the fact that desires and inclinations are expressly understood as “appearances” (453:24) in Section 4. This implies that the will cannot be held responsible for the desires and inclinations that as appearances belong entirely to the realm of nature (as the world of appearances). But after Kant has indicated that the human being as an authentic self with his intelligible will is not responsible for his inclinations and impulses, he continues by observing that he is responsible for ‘the indulgence’ that he ‘would like to bear toward them, if, to the disadvantage of the rational laws of the will, he were to concede them influence on its maxims.’ The crucial question, for which Kant provides no answer whatsoever in GMS, is this: who is the ‘he’ that figures as the subject in this sentence? Who is it that exercises ‘indulgence’ here? It is impossible to furnish a precise grammatical reconstruction in this regard, and this indicates that Kant himself was by no means clear just how the ‘human being’ is to be understood in relation to the ‘authentic self.’ In the context of GMS it is not possible to fi nd a satisfactory answer to the question of who exercises such indulgence. It can hardly be the intelligible will itself that does so, for qua member of the intelligible world and authentic self, this will simply wills the good, which is indeed its law. In order to appreciate the problem here, one only needs to ask whether the act of indulgence toward the inclinations and impulses is or is not itself a free act.

192

Immanuel Kant’s Groundwork for the Metaphysics of Morals

A Transcendental Argument for the Freedom of the Will In Section 2 Kant begins by reminding us that freedom is only an idea and cannot be proved on the basis of experience (this thought is repeated in some detail in Section 5). But the thesis that directly follows, namely, that regardless of the character of freedom as an idea, the moral laws inseparably combined with it “are valid” (448:6), is not specifically grounded as such by Kant, and we shall not pursue this any further here (especially since it can be understood as a reformulation of the thesis of analyticity). What is much more important here is Kant’s claim regarding freedom: ‘To every rational being that has a will, thus to every being whose reason is practical, i.e. possesses causality in regard to its objects, we must necessarily lend the idea of freedom under which alone it would act.’24 Kant grounds this claim as follows: Now one cannot possibly think a reason that, in its own consciousness, would receive steering from elsewhere in regard to its judgments; for then the subject would ascribe the determination of its power of judgment not to its reason but to an impulse. It must regard itself as the author of its principles independently of alien influences, consequently it must, as practical reason or as the will of a rational being, be regarded by itself as free, i.e. the will of a rational being can be a will of its own only under the idea of freedom and must therefore with a practical aim be attributed to all rational beings. (448:13) We are talking about an argument that can already be found in Epicurus and has regularly been advanced up to this day in direct opposition to the claims of determinism. The basic idea here is this: reason, as a capacity for thinking (and judging), cannot be unfree, for anyone who claims this thereby (in and with this claim) appeals and necessarily appeals to the freedom that he contests; ‘consequently’ this same reason, as a capacity to act, cannot be unfree. That Kant is basically adopting this argument is also clearly evident from a parallel passage in his Review of Schulz’s Attempt at an Introduction to a Doctrine of Morals (1783). Here Kant criticizes Schulz’s “universal fatalism . . . 24. This is how we can paraphrase the two sentences at 448:9–10.

Section III of the Groundwork

193

where he transforms all human deeds into nothing but a marionette play” (RS:13). He states his argument against Schulz’s determinism as follows: Although he would not admit it, he 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, which could subsequently change; 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, when— as I have no doubt—he wants to proceed in his righteous conduct in conformity with the eternal laws of duty and not to be a plaything of instincts and inclinations, though at the same time he denies himself this freedom. (RS:14)25 The determinist Schulz claims that determinism is true. He has ‘presupposed,’ nolens volens, that with his own understanding he can make a judgment ‘in accordance with objective grounds’ rather than merely through ‘subjectively determining causes.’ He has thereby in fact already assumed the freedom of thinking when he defends his determinism, and he cannot defend his determinism unless he presupposes this freedom. There is thus a contradiction between what determinism asserts and the act of advancing this assertion (a ‘performative contradicton,’ as we might call it today). Given this reasoning, it is clear for Kant that every rational being must consider itself free in advancing its appropriately grounded judgments. But for Kant this means that a rational being must therefore regard itself as a member of the intelligible world. This is what he undertakes to show in Section 3.

25. We can fi nd a similar thought in ML1:268–269; cf. also KrV:A354 and Kant’s Reflections, 4338, 4495, and 5540. Although in ML1 Kant certainly provides an argument for the freedom of the subject and even speaks of ‘proofs’ in this regard (ML1:268, line 35), he recognizes the specific limits of what the argument can show: “Though for all that the spontaneitas absoluta cannot be grasped; but nor indeed can it be refuted” (ML1:268). In the same way, Kant presents an argument in GMS III for the thesis that although freedom must be “presupposed,” we cannot “prove it in its theoretical intent” (448, footnote).

194

Immanuel Kant’s Groundwork for the Metaphysics of Morals

The World of Understanding and the World of Sense In the first four paragraphs of Section 3 Kant once more repeats the position that has been reached so far; it is clear, above all, that the validity of the CI has not yet been demonstrated. In this connection Kant says that “a kind of circle shows itself here” (450:18), although it is one that, properly understood, serves to clarify the real task. In the Critique of Practical Reason Kant claims that whereas freedom is “the ratio essendi of the moral law, the moral law is the ratio cognoscendi of freedom” (KpV:4, footnote). In GMS, however, Kant does not infer the ‘can’ (freedom) from the ‘ought’ (morality). Here it is freedom of thinking (spontaneity) that is the ratio cognoscendi of the world of understanding and thus ultimately also the ratio cognoscendi of the validity of the moral law. Thus Kant must also show that one can argue for the reality of human freedom on the basis of the distinction between the world of understanding and the world of sense without already presupposing the validity of morality. But even if human freedom has been shown, in the second place the validity of the categorical imperative still remains an open question in spite of the thesis of analyticity. This question is answered only in Section 4 by reference to the ontological superiority of the intelligible I. The distinction between appearances and things in themselves is one of the most famous theoretical contributions of Kant’s thought. Its interpretation is an ongoing matter of controversy, which it cannot be our task here to settle. Like Kant’s theory of freedom in general, we must assume it without further justification here in relation to the particular purposes of our interpretation (in a sense as a “crude” distinction [451:18]). We suggest that it be understood as follows: Objects appear to us by means of our senses; since space and time are simply a priori forms of intuition, the attributes we ascribe to these sensuously given objects regarded as appearances cannot be the attributes of these objects regarded as things in themselves. Only if we regard representations qua representations in turn as entities does it make sense to distinguish different worlds here (appearances regarded as representations are of course not numerically identical with things in themselves). The distinction is therefore to be understood in epistemological rather than in ontological terms. Kant’s self-interpretation in the Opus postumum might serve as our guide here: “The distinction of what is called the object in itself from the object

Section III of the Groundwork

195

in appearance . . . does not signify an actual thing standing over against the object of the senses” (AA 22:24). This suggests that appearances and things in themselves are not two sets of entities, occupying two different metaphysical realms, but rather the same entities thought about in two different ways. (They may still constitute two different worlds, since a world is a whole constituted by a plurality of things, and the same things may belong to two different worlds if the two wholes are thought of differently.) Kant also claims that appearance is always the appearance of something. This ‘something’ is apparently the thing in itself of which we have no knowledge, and of which we can never have any knowledge, since we are acquainted with it only insofar ‘as it affects us,’ and if it affects us, then it is an object of sensibility and its subject-dependent forms of intuition. In this sense, Kant says that “behind appearances things in themselves (although hidden) must ground them” (459:28; o.e.; cf. 451:12). Kant also uses this kind of formulation, namely, that things in themselves ‘lie at the ground’ of appearances, in his deduction of the CI. There he says that “the world of understanding contains the ground of the world of sense, hence also of its laws” (453:31). But this is consistent with thinking of appearances as the same things as things in themselves if things regarded the latter way contain the ground of, or lie ‘behind,’ things regarded the former way. The world of things regarded as appearances is what Kant calls the world of sense; the same world regarded as a world of things in themselves is also called the world of understanding. There are two points to be noticed here in relation to GMS. First, the distinction of the world of sense from the world of understanding holds for “all things” (451:37; o.e.) (inasmuch as they can appear at all). Second, in GMS Kant also refers to the world of understanding as the world of intelligences (as things in themselves), and he further speaks of the ‘intellectual world’ or the ‘world of intelligences’ without distinguishing in any strict terminological sense between them.26 Thus he defines the “concept of an intelligible world” (458:27) as “the whole of rational beings as things in themselves” (458:28), and thus as a “world of intelligences” (462:2).

26. For the ‘intellectual world,’ cf. 451:35; for the ‘intelligible world,’ cf. 452:29; 452:31; 454:7; 455:8; 457:31; 458:28; 462:2; 462:18; and 462:28. We have already pointed out that Kant also mentions the concept of the “realm of ends” (462) in this connection.

196

Immanuel Kant’s Groundwork for the Metaphysics of Morals

Especially in GMS III there is a strong temptation (certainly not without textual evidence) to take Kant’s talk about two worlds here as committing him to a species of metaphysics that his own critical epistemological theory would condemn as ‘transcendent’ and must regard as lying beyond the bounds of our cognition. Kant’s language, however, might also admit of an interpretation according to which the ‘world of sense’ and ‘world of understanding’— and our membership in both—represent not a distinction between unknowable metaphysical realities but only two ways of thinking about matters with which we are directly acquainted: our own sensible feelings and inclinations and our intellectual awareness of ourselves as agents. To the extent that this might involve metaphysical commitments, their unknowability must always be kept in mind as part of Kant’s doctrine. This may help explain why Kant in the end regards our freedom as unknowable and incomprehensible (even if it must be assumed for practical purposes) and even (as we have seen) thinks that we cannot fully answer the question how a CI is possible. When he speaks of the world of sense and of understanding, Kant should be understood not as a supernaturalist metaphysician but rather as a philosopher who holds that there are basic questions about ourselves and our moral condition that we will never be able to answer.

Understanding, Reason, and Intelligence: The Human Being as Thing in Itself We should therefore approach Kant’s doctrines by seeing them not as metaphysical talk about transcendent realities but as a way of thinking about our faculties of knowing and acting with which we are acquainted. The world of understanding, as we have seen, is said to lie ‘at the ground’ of the world of sense, and we must assume things in themselves ‘behind’ the appearances. Now since the human being has some “acquaintance” (451;22) with himself as appearance, he must, “as is only fair” (459:27; o.e.), ‘behind’ his I as appearance also assume “his I, however that may be constituted in itself ” (451:30; o.e.). In Section 3 Kant connects the distinction between the world of understanding and the world of sense to the distinction between receptivity (sensibility) of the human cognitive subject and its spontaneity (self-activity). There is a variation or difference “be-

Section III of the Groundwork

197

tween the representations that are given to us from somewhere else, in which we are passive, and those which we produce solely from ourselves, and thus prove our activity” (451:9; o.e.). The different senses of selfconsciousness here (the I as appearance and the I as transcendental subject) rest on this distinction between receptivity and spontaneity. The human being belongs “in regard to mere perception and the receptivity of sensations,” that is, in relation to sensibility, “to the world of sense,” “but in regard to whatever in him may be pure activity,” that is, in relation to the spontaneity of understanding and reason, he belongs to “the intellectual world”; and Kant tells us here that the latter activity “attains to consciousness not through the affections of the senses but immediately” (all citations from 451:31–35). This ‘pure activity’ and our ‘immediate consciousness’ of it are clarified in the next and second major passage of this subsection: they are the selfactivity of the transcendental subject as manifest in understanding and reason (452:7–22).27 The reasoning runs as follows: inasmuch as we are conscious of ourselves as beings that can exercise understanding and reason as faculties of pure self-activity (the spontaneity of thought), we grasp ourselves as intelligences and thereby as members of the intelligible world;28 and Section 2 has already shown how the claim that we possess spontaneity (freedom) cannot be contested without self-contradiction. In this context Kant understands this spontaneity as transcendental freedom, so that understanding and reason, insofar as they are understood as spontaneous faculties, are transcendentally free.29 Kant further distinguishes the spontaneity of reason from the spontaneity of the understanding. He says that reason is “elevated” (452:10) above the understanding since in contrast with the latter, reason actively

27. There is a parallel passage at 457:4–458:5, although it is very difficult to interpret. 28. For precisely the same thought, cf. KrV (A546–547/B574–575). 29. For the spontaneity of understanding and reason, cf. KrV:A50/B74 and B130, B162 (footnote), and B428; in the context of the ‘third antinomy’ and its ‘resolution’ Kant basically identifies spontaneity and transcendental freedom (cf. A418/B446 and A445/B473). The identification of theoretical spontaneity and transcendental freedom is also particularly clear in ML1 (cf. especially pp. 268–269); some of Kant’s reflections are also instructive in this regard (Refl:4220, 5413, 5441).

198

Immanuel Kant’s Groundwork for the Metaphysics of Morals

produces “ideas” (452:18) without immediately relating to sensuously given objects.30 As a logical faculty, reason is the faculty of inferences. In the process of inference, reason moves from premises to conclusions. The premises can themselves in turn be understood as conclusions derived from other premises, and so on. In this way reason strives to find the ultimate premises, to discover the ‘unconditioned.’ This striving to find the ultimate conditions (premises) gives rise to ideas, and specifically, on analogy with the three different kinds of rational inference (categorical, hypothetical, and disjunctive), to the three ideas that were traditionally addressed in three specific disciplines: the soul (psychologia rationalis), the world and freedom (cosmologia rationalis), and God (theologia rationalis). When Kant talks of “ideas” at 452:18, therefore, he is referring to the metaphysical ideas rather than the moral ideas. After he has drawn attention to the spontaneity of reason and the understanding, Kant continues: “On account of this, a rational being has to regard itself as an intelligence (thus not from the side of its lower powers), as belonging not to the world of sense but to the world of understanding” (452:23).31 Insofar as the human being is affected through the senses, she is acquainted with herself as appearance; insofar as she has consciousness of the spontaneity of her understanding and her reason, she thinks herself, as Kant says a little later, “as intelligence, also as thing in itself ” (459:22; o.e.).32 Thus the human being recognizes her membership of the intelligible world and thereby also recognizes her freedom, independently of the moral law.

The I That Thinks and the I That Wills The argument against determinism derives from that freedom of thinking that the determinist himself cannot meaningfully contest. But in Sec-

30. Reason, in its pure self-activity, is ‘elevated’ above the understanding, not because it is a pure self-activity in contrast with the mere activity of the understanding (cf. Kant’s “in the respect that” at 452:10). 31. In GMS III Kant makes more frequent use of the concept of intelligence (in this specific sense) than almost anywhere else in his published work: 452:23; 453:17; 453:35; 457:9; 457:11; 457:22; 457:30; 457:34; 457:37; 458:24; 459:12; 459:22; 461:3; 461:15; 462:3; and 462:30; cf. also ML1:224; MM:878; and KrV:A354; A682/B710; and A742/B770. 32. In his ‘solution’ to the third antinomy (KrV:A539/B567) Kant describes spontaneity in terms of “its character as thing in itself.”

Section III of the Groundwork

199

tion 2 Kant uses this freedom to argue for the freedom of the will. He says that theoretical reason must regard itself as the author of its principles independently of alien influences, consequently it must as practical reason or as the will of a rational being, be regarded by itself as free. (448:17; o.e.) But how are we to understand the ‘consequently’ here?33 The transition here appears plausible only if the reason whose freedom cannot be contested without self-contradiction is the same reason that also becomes practical, so that then (‘consequently’) the will too, which is nothing but pure practical reason, must be regarded as free. This thought is also supported by a remark of Kant’s from the Preface. There he writes that “it can in the end be only one and the same reason that is distinguished merely in its application” (391:27; o.e.). Kant describes this as the “unity” (391:25) of theoretical and practical reason “in a common principle” (391:26). It is only when this unity of theoretical and practical reason has been completely demonstrated that we can attempt a critique of pure practical reason. Since Kant does not believe, when he is writing GMS, that he can bring this task to such “completeness” (391:29), he does not try to offer such a critique but describes his book as a groundwork for the metaphysics of morals instead. In fact, interestingly enough, in GMS III we are presented with a transition to the “critique of pure practical reason,” if not with any “completeness,” then at least in its “principal features” (445:15). Thus we must also assume that the ‘unity’ in question plays a role in this critique. And indeed it clearly does, albeit in a highly unsatisfactory form. Kant writes that he has not yet brought the matter to completeness “here” (391:29; o.e.), but in fact he would never achieve this completeness: the supposed demonstration of the unity of theoretical and practical reason never advanced beyond the stage of a program. Thus how exactly this ‘unity’ is to be understood, and how and why Kant moves from theoretical to practical freedom and from the concept of an intelligence to the concept of an

33. In a parallel passage in his review of Schulz (RS:14) we fi nd ‘likewise’ at the relevant point rather than ‘consequently’; this is a distinction that it is impossible to investigate here.

200

Immanuel Kant’s Groundwork for the Metaphysics of Morals

intelligence endowed with will—with regard to all these questions, GMS has almost nothing to say (indeed, it provides very little regarding the basic argument for freedom). The transition from the I that thinks to the I that wills through the ‘consequently’ in the central argument of Section 2 is thus not actually explicated or grounded in the text at all. Even if we accept the indirect argument against determinism (and we may well have doubts here), this would still show us very little with respect to the freedom of the will as long as it remains unclear whether the freedom of theoretical reason is indeed also identical with the freedom of practical reason.

The Suspicion of Circularity The ‘suspicion of circularity’ in Section 3 has constantly caused problems for Kant’s interpreters. It would be a laborious and time-consuming task to offer a precise reconstruction and evaluation of the entire text, so we shall have to forgo such an examination here. We shall simply try to bring out the principal points and thus clarify the development of the argument. The ‘circle’ in question is formulated twice: 1. “In the order of efficient causes we assume ourselves to be free in order to think of ourselves as under moral laws in the order of ends, and then afterward we think of ourselves as subject to these laws because we have attributed freedom of the will to ourselves” (450:19). 2. “Now the suspicion has been removed that we aroused above, that there was a hidden circle contained in our inference from freedom to autonomy and from the latter to the moral law, namely, that we perhaps took the idea of freedom as a ground only for the sake of the moral law in order afterward to infer the latter once again from freedom, hence that we could not offer any ground for the former, but rather only as begging a question, which well-disposed souls might concede to us, but which we could never set up as a provable proposition” (453:3). In referring to this suspicion of circular reasoning, Kant is drawing attention to two problems. First, at the beginning of Section 3 we have still received no plausible justification for the claim that the human being may

Section III of the Groundwork

201

indeed regard herself or himself as free. Section 1 was concerned with purely conceptual elucidations, and Section 2 showed only that rational beings must consider themselves free, but not that human beings may understand themselves in the strict sense as such rational beings and therefore as free beings. Kant accomplishes this by means of the distinction between the world of understanding and the world of sense and the concept of ‘intelligence’ that is associated with it. Second, one might be tempted to misinterpret the thesis of analyticity as if the conceptual claim that if a being acts purely rationally and freely, then it always also acts morally, also already proved the validity of the CI for human beings. This is not the case, and the suspicion of circular reasoning also makes this clear. Let us begin with the first problem. Kant’s own allusion to a ‘circle’ initially suggests that he was thinking of a circulus in probando (A because B, B because A).34 The circle would then assume this form: ‘We are free because we are subject to the law, and we are subject to the law because we are free.’ But Kant does not actually describe the problem in this way. Rather, he writes that we think of ourselves as free ‘in order to think of ourselves as under moral laws,’ and that of course is something quite different from the thesis that we are free because we are subject to the moral law. The situation is similar with the second passage we quoted: the problem lies in the notion that ‘we perhaps took the idea of freedom as a ground only for the sake of the moral law.’ Thus in fact Kant identifies the logical error as that of “begging a question” (453:9). But such a petitio principii does not involve what today we would call a vicious circle (i.e., a circulus in probando), but rather the following: “If a conclusion is derived from premises which are as uncertain as the conclusion itself, then the proofs are begged (petitio principii seu quaesiti)” (AA 16:774).35 What is missing (and in that sense ‘begged’) is evidently the justification for assuming the idea of freedom that might be thought impossible if the being who is free is also part of a deterministic mechanism of nature. And in fact Kant attempts throughout his discussion to dispel the

34. For Kant’s understanding of the circulus in probando, cf. AA 16:774 (in a footnote to Refl:3314) and JL (AA 9:135). 35. This formulation derives from G. F. Meyer’s handbook on logic, Auszug aus der Vernunftlehre, which Kant used in his lectures; cf. again JL (AA 9:135).

202

Immanuel Kant’s Groundwork for the Metaphysics of Morals

suspicion of circularity primarily by showing how the human being must understand himself as a free intelligence; he thereby supplies the missing justification for the proposition ‘We are free’ and thus removes the suspicion of circularity. As we have already seen, there is, at best, something like a transcendental argument in Section 2: one cannot consistently deny freedom of thinking (and hence freedom of the will) because by forming the judgment ‘There is no freedom,’ one is presupposing that in order to form this very judgment one must be free. However, there is no proof based on concepts that refer to intuitions. To say that freedom is an ‘idea’ is not to say that we are not theoretically justified or warranted in believing in freedom; it is to say that freedom is not a thing we would know about as we know, say, about the law of gravity. In this sense, freedom is unprovable, and in this respect, one can show only that the concept of freedom may be legitimated, saved from the threat of contradiction, as Kant had argued in KrV (A530–558/B558–585). In this context, Kant’s fi nal conclusion is that freedom and natural necessity “do not contradict each other” (A557/B585); his aim, he says, has been only to show that “nature at least does not confl ict with causality through freedom—that was the one single thing we could accomplish, and it alone was our sole aim” (A558/B586). The final sentence of Section 3 alludes specifically to the second problem (that the thesis of analyticity has not yet shown anything regarding the validity of the CI): For now we see that if we think of ourselves as free, then we transport ourselves as members into the world of understanding and cognize the autonomy of the will, together with its consequence, morality; but if we think of ourselves as obligated by duty, then we consider ourselves as belonging to the world of sense and yet at the same time to the world of understanding. (453:11) It is striking that this sentence, although it begins with ‘For’ and is supposed to show why the suspicion of circularity has now been ‘removed,’ does not even address the problem that the possibility of human freedom has not been proved. We have already pointed out that the first half of this sentence (up to the semicolon) recapitulates the thesis of analyticity. The

Section III of the Groundwork

203

second part of the sentence, on the other hand, emphasizes that the human being must be understood not merely as free but as obligated. For the human being as a sensuous-rational being, the ‘standing under laws’ established in the thesis of analyticity is not an analytic consequence of freedom but a subjection under the moral law as the CI, and one that must still be shown as necessary in its own right. The fi nal sentence of Section 3 thus alerts us to the possible misunderstanding that for human beings too the moral law is ‘inseparably connected’ with the idea of freedom, that is, connected analytically. If we look back again over the argument, we can see that this second problem also already appears in the formulations of the circularity issue. In the first formulation (450) it looks as if we are exclusively concerned with the ‘question-begging’ assumption of freedom (in the first part: ‘In the order . . . in the order of ends’) and as a result with the question begging of the moral law (in the second part: ‘and then afterward . . . to ourselves’). But we can also see from the second part that even if the possibility of freedom had already been established, the proof of the validity of the CI has still to be provided. For this, we need only to accentuate it slightly differently, namely, as ‘We think of ourselves as subjected to moral laws because we have attributed freedom of the will to ourselves.’ Thus the mistake lies in regarding ourselves as already ‘subjected’ to the moral law as an imperative merely because this law is analytically connected with the idea of the freedom of a will that is considered exclusively as a member of the world of understanding. This is a mistake because we with our will are ‘at the same time’ also members of the world of sense. According to the second passage (453), the suspicion of circularity lies in the thought ‘[a] that there was a hidden circle contained in our inference from freedom to autonomy and from the latter to the moral law, namely, [b] that we perhaps took the idea of freedom as a ground only for the sake of the moral law in order afterward to infer the latter once again from freedom, hence [c] that we could not offer any ground for the former, but rather only as begging a question, which well-disposed souls might concede to us, but which we could never set up as a provable proposition.’ The ‘question’ (or ‘principle’—Prinzip) in (c) is clearly identified through the second relative pronoun ‘which’ as a ‘proposition’ (Satz). If we employ the logical terminology of demonstrative proposition (Beweissatz), premise

204

Immanuel Kant’s Groundwork for the Metaphysics of Morals

(Vordersatz), and conclusion (Schlusssatz) in order to identify the petitio principii, there is certainly at least one proposition, maybe more, in the petitio principii, namely, ‘We are free’ and ‘We are subject to the moral law.’ But these propositions do not appear as such in the entire passage (453). The only ‘proposition’ that is actually formulated as such (albeit in compressed form) is the ‘conclusion’ (a). What is this conclusion (‘from freedom to autonomy and from the latter to the moral law’)? Naturally, one might immediately think of the thesis of analyticity here, and Kant himself recalls the thesis in the fi rst passage on the circle when he refers to “reciprocal concepts” (450:24). But what sense would it make to suspect a ‘hidden circle’ somewhere ‘in’ this thesis, and indeed specifically as a mistaken proof? The thesis of analyticity makes no existential claims and in this sense is not intended to prove anything. We could speak of a mistake only if, fi rst, in the context of an attempted proof of the validity of the CI for human beings, the thesis of analyticity and thus also the deduction (the freedom of a purely rational being) are presupposed as starting points without being grounded, and if, second, the inference ‘to the moral law’ is understood as an inference to the CI. It is only on the basis of transcendental idealism that we see that we too may regard ourselves as free and as members of the world of understanding (which obviates the danger of a petitio principii), and even that we must do so, a compulsion that means that as members of the world of sense, which at the same time we are, we cannot derive the moral law analytically (as a ‘consequence’) from the freedom of our will (which obviates the danger of misusing the thesis of analyticity to try and deduce the moral law as an imperative). In GMS II Kant had already warned against the notion that we could achieve anything simply by the process of conceptual analysis. The validity of the CI, as he said there, “cannot be proven through the mere analysis of the concepts occurring in it, because it is a synthetic proposition; one would have to advance beyond the cognition of objects and to a critique of the subject, i.e. of pure practical reason” (440:22). And it is in GMS III that we ‘advance’ to this critique of pure practical reason. The question of circularity marks the precise point of the transition from the conceptual analysis that belongs to the metaphysics of morals (MS3 ) to the ‘critique of the subject.’

Section III of the Groundwork

205

4.3 The Deduction of the Categorical Imperative After Section 3, one thing is clear: the human being can act morally. But why ought the human being to do so? It is evident that we are not asking about some nonmoral motive here. But the question still remains why the human being ought to act morally, even if she knows that she can do so. Kant wishes to show that in a certain sense the answer to the question why I ought to act morally, even if I can do so, lies in the fact that I can act morally. He can accomplish this only if the ability to act morally can be revealed as a unique and authentic will to act morally, a willing that possesses a higher value than that sensuously determined willing that leads us to the difference between will and obligation in the first place.

The Argument of the Deduction Kant begins Section 4 by recapitulating his previous reflections: The rational being counts himself as intelligence in the world of understanding, and merely as an efficient cause belonging to this world does it call its causality a will. From the other side, however, it is conscious of itself also as a piece of the world of sense, in which its actions, as mere appearances of that causality are encountered, but whose possibility from the latter, with which we have no acquaintance, is something into which we can have no insight, but rather in place of that we have to have insight into those actions as determined through other appearances, namely desires and inclinations as belonging to the world of sense. As a mere member of the world of understanding, all my actions would be perfectly in accord with the principle of the autonomy of the pure will; as a mere piece of the sensible world, they would have to be taken as entirely in accord with the natural law of desires and inclinations, hence with the heteronomy of nature. (The former would rest on the supreme principle of morality, the second on that of happiness.) (453:17) Kant thus starts by resuming the results from the preceding subsections. The first premise of the deduction contains the thesis that the human being must understand himself as a rational being:

206

Immanuel Kant’s Groundwork for the Metaphysics of Morals

(D1): The human being discovers the faculty of reason within himself, which as an epistemic faculty is a form of his self-activity. We saw how the argument in Sections 2 and 3 begins in each case with a consideration of reason as a theoretical faculty. The human being cannot deny without self-contradiction that he or she does possess reason. The activity of reason is pure self-activity, spontaneity, and freedom, and these attributes allow the human being to understand himself as an ‘intelligence.’ As intelligence, in turn, the human being must regard himself as a member of the world of understanding. (D2): As a rational (self-active) being the human being must regard himself as an intelligence, and thus, from this perspective, as a member of the world of understanding. It is only as intelligence or insofar as he is an intelligent being that the human being may regard himself as a member of the world of understanding. He regards his reason as free, thereby transports himself as intelligence into the world of understanding, and from this perspective grasps his will too as a member of this world of understanding and thereby as a free will. (D3): As a rational being, and consequently as one that belongs to the world of understanding, the human being can only think the causality of his own will as standing under the idea of freedom. At this point of the argument it is very clear how important it is to understand the thesis of analyticity correctly. The claim that as “a mere member of the world of understanding, all my actions would be perfectly in accord with the principle of the autonomy of the pure will” (453) just is the thesis of analyticity; and the analytic relationship of freedom and morality is valid only insofar as the human being ideally understands his will too as a member of the world of understanding. (D4): Since the moral law is analytically connected with freedom as the quality of the will of a rational being, which is a member of the intel-

Section III of the Groundwork

207

ligible world of understanding, the human being, if and insofar as he understands himself as such a being, also cognizes autonomy and the moral law as the law of his rational willing. But with this the validity of the CI has not yet been demonstrated. The question, therefore, is why, as a being that belongs to the world of understanding and to the world of sense, he ought to subject himself to the law of the world of understanding. The answer to this question, and thus the deduction itself, is provided in the following sentence, which presents what we may call Kant’s ontoethical principle (OP): But because the world of understanding contains the ground of the world of sense, hence also of its laws, hence is immediately legislative in regard to my will (which belongs wholly to the world of understanding), and hence must also be thought of wholly as such, therefore as intelligence I will cognize myself, although on the other side as a being belonging to the world of sense, as nevertheless subject to the laws of the first, i.e. to reason, which in the idea of freedom contains the law of the understanding’s world, and thus to autonomy of the will; consequently I must regard the laws of the world of understanding for myself as imperatives and the actions that accord with this principle as duties. (453:31) This sentence is extremely difficult from both a grammatical and a substantive point of view, and there is no room for a more detailed and precise analysis here. It is clear from the ‘because . . . therefore’ structure that the sentence falls into two principal parts, where the first part (‘But because . . . wholly as such’) contains the ground for a conclusion that is drawn in the second (‘therefore . . . this principle as duties’). Without offering further justification here we can reconstruct the argument of the first part as follows: ‘But because the world of understanding contains the ground of the world of sense, because it consequently also contains the ground of the laws of the world of sense, because it is thus immediately legislative in regard to my will, which belongs wholly to it, and because in regard to my will it must thus also be thought of as a world of understanding

208

Immanuel Kant’s Groundwork for the Metaphysics of Morals

which contains the ground of the world of sense and the ground of the laws of that world, therefore.’ In the second part of the sentence the human is being regarded in his twofold determination. That is, the human being cognizes himself ‘as intelligence, . . . although on the other side as a being belonging to the world of sense,’ as ‘subject’ to the law of the world of understanding and thus to the autonomy of the will. In the first part of the ‘therefore’ section of the sentence the claim is this: ‘I cognize myself as a being that regards itself at the same time as a member of the world of understanding (as intelligence) and as a member of the world of sense as being subject to the law of the world of understanding, and hence to reason, which contains the law of the world of understanding in the idea of freedom, and thus to the autonomy of the will.’ The rest of the sentence is then easy to construe. It makes it clear that the moral law, which as a descriptive proposition captures the subjectively and objectively necessary volition of purely rational beings, is a synthetic imperative for sensuous-rational beings: ‘I must regard the laws of the world of understanding as imperatives for me and the actions in accord with this principle as duties.’ The OP can be reconstructed as follows: OP: But because the world of understanding contains the ground of the world of sense, because it consequently also contains the ground of the laws of the world of sense, because it thus in regard to my will, which belongs wholly to it, is immediately legislative, and because it must thus in regard to my will also be thought as a world of understanding that contains the ground of the world of sense and the ground of the laws of that world, therefore I will cognize myself as a being that regards itself at the same time as a member of the world of understanding (as intelligence) and as a member of the world of sense as being subject to the law of the world of understanding, and thus to reason, which contains the law of the world of understanding in the idea of freedom, and thus to autonomy of the will, and consequently I must regard the laws of the world of understanding as imperatives for me and the actions in accord with this principle as duties. Thus Kant provides the answer to the question of the deduction in a single sentence. Then he summarizes the argument once again:

Section III of the Groundwork

209

And thus categorical imperatives are possible through the fact that the idea of freedom makes me into a member of an intelligible world, through which, if I were that alone, all my actions would always be in accord with the autonomy of the will, but since I intuit myself at the same time as member of the world of sense, they ought to be in accord with it, which categorical “ought” represents a synthetic proposition a priori by the fact that to my will affected through sensible desires there is also added the idea of precisely the same will, but one belonging to the world of understanding, a pure will, practical for itself, that contains the supreme condition of the fi rst in accordance with reason. (454:6) But what is the decisive thought expressed in this ontoethical principle? Kant claims that in his consciousness of understanding and reason as self-activity the human being is conscious of himself as ‘intelligence’ and thus as his “authentic self ” [eigentliches Selbst] (457:34; 458:2; 461:4). As an intelligence in this sense, the human being regards himself as a member of the world of understanding and thus as a thing in itself; his will qua intelligence is the ‘authentic self ’ of praxis.36 It is this thought, namely, that the will (regarded) as an intelligible faculty is the “authentic” self, as distinct from the human being insofar as he is “only” (457:35; o.e.) an appearance of himself (as “phenomenon in the world of sense,” 457:13), that constitutes the core of the OP. There is no way around it: Kant grounds the validity of the CI on the superior ontological status of the world of understanding. The moral law is the law of the will as a member of the world of understanding; this will is the ‘authentic self ’ of the human being. The human being as a thing in itself (and thereby as ‘authentic self ’) enjoys a higher ontological significance than the human being as appearance, and that is why the law of the world of understanding is valid as an imperative for the human being who is at the same time a member of the world of sense.37 36. In this sense Kant could already speak in KrV (A492/B520) of “the authentic self as it exists in itself, or the transcendental subject.” 37. The metaphysics of morals (MS3 ) itself had already recognized the “ground of a possible categorical imperative” (428:5) in something “whose existence in itself has an absolute worth” (428:3 ), and that is the free and rational being as end in itself. When

210

Immanuel Kant’s Groundwork for the Metaphysics of Morals

In a later reformulation of the OP Kant specifically emphasizes (evident in the original from the unusually large spacing of the type) that the moral law is valid for us as human beings, since [sic] it has arisen from our will as intelligence, hence from our authentic self; but what belongs to the mere appearance is necessarily subordinated by reason to the constitution of the thing in itself. (461:2) The law of the world of appearance is the “natural law of desires and inclinations” (453:28) that as “appearances” (453:24; o.e.) determine human actions. But this natural law belongs precisely to ‘the mere appearance’ and is for that reason ‘subordinated to the constitution of the thing in itself ’; thus Kant confirms once again his thesis of the ontological superiority of the world of understanding. And there is another passage where Kant places the weight of the argument entirely on the ontological superiority of the ‘authentic self.’ The human being as a rational being is a member of the world of understanding, and hence Kant argues that since in that world he himself only as intelligence is the authentic self (as human being, by contrast, only appearance of himself), those laws [of the world of understanding] apply to him immediately and categorically. (457:33; o.e.)

Kant distinguishes in MS3 between ‘price’ and ‘worth’ (434–435) and speaks of the ‘absolute worth’ of a person as an end in itself and thus as a member of the world of understanding (428), he is preparing the conceptual ground for the subsequent deduction, for it is precisely this “worth” (454:37) of the person as a member of the world of understanding, or this metaphysical status, that is the ultimate ground of the validity of the CI. This conception of worth or value is thus also explicitly reflected in the section on the deduction when Kant says with reference to common human reason that even “the most wicked scoundrel” (454:21) transports himself through the thought of his pure will into “entirely another order of things” (454:31) and fi nds there “a greater inner worth of his person” (454:36): “This better person, however, he believes himself to be when he transports himself into the standpoint of a member of the world of understanding” (454:37). Now we can also see how in GMS III Kant grounds the ‘postulate’ from GMS II (429) that all rational beings must understand themselves as things in themselves.

Section III of the Groundwork

211

We can thus summarize the OP and thereby the argument for the validity of the CI (while providing a further formulation of the OP) as follows: (D1):

The human being discovers in himself the faculty of reason, which as an epistemic faculty is a form of his self-activity. (D2): As a rational (self-active) being, the human being must regard himself as an intelligence and thus, from this perspective, as a member of the world of understanding. (D3): As a rational being, hence as one belonging to the intelligible world, the human being can think the causality of his own will only as standing under the idea of freedom. (D4): Since the moral law is analytically connected with freedom as the quality of the will of a rational being that is a member of the intelligible world of understanding, the human being also cognizes, if and so far as he understands himself as such a being, autonomy and the moral law as the law of his rational volition. (OP*): The world of sense is ontologically subordinated to the world of understanding and thereby also to the will as a member of this intelligible world, and thus the law of this world (the moral law) is also valid as a law (as the CI) for beings that are at the same time members of the world of sense and of the world of understanding; what I ought to do as a sensuous- rational being is what as a rational being and thereby as authentic self I myself will to do. Thus the moral law is valid because it springs from the pure will as the authentic self of the human being. What is important here is the insight that this pure will is no alien will but rather one’s own. If the commanding authority were some alien will (however this be understood) or an alien law rather than one’s own law, then it would be an alien volition or an alien form of law, and once again we should have to find some nonmoral motive in order to will (and accordingly do) what one does not in any original way will at all. Thus in the context of MS3 Kant has already announced that the human being is “subject only to his own and yet universal legislation” (432:29): “For if one thought of him only as subject to a law (whatever it might be), then this would have to bring with it some interest

212

Immanuel Kant’s Groundwork for the Metaphysics of Morals

as a stimulus or coercion, because as a law it did not arise from his will, but rather this will was necessitated by something else to act in a certain way in conformity with the law” (432:32). As long as I am or rather am supposed to be subject only to a command that I have not given to myself, the question arises why I ought to subject myself to this command. Anyone who undertakes to demonstrate the validity of a law that is not a self-given law is therefore immediately confronted with the problem that some ‘interest as a stimulus or coercion’ must then be conceived as well. If, on the other hand, I observe a law that I have given to myself, there is no need to introduce such an interest, for then it is my own law, that is, what I myself will, and it would make no sense to ask for a reason why I ought to will what I will—for it is already what I will. Thus for Kant the attempt to demonstrate the validity of the moral law involves showing that what one ought to do is always also what one wills to do. That is why he describes the ‘ought’ that is contained in the CI as an unlimited volition of our own: “For this ‘ought’ is really a volition” (449:16; o.e.), and “The moral ‘ought’ is thus his own necessary volition as a member of an intelligible world and is thought of by him [i.e., by the human being] as an ‘ought’ only insofar as he at the same time considers himself as a member of the sensible world” (455:7; o.e.). My will as a member of the intelligible world wills the morally good. Insofar as I am at the same time a member of the sensible world, I am subject to this self-given law, and indeed because it is given by myself—by my higher ‘authentic self’— or precisely because it is my own volition: “The good is always that which every human being wills, and he would also always do this, if only it was not difficult for him to perform it, and if our nature were so constituted that we always acted in accordance with the concept of the good, then we would be really free” (MM:903). The moral insight Kant wishes to produce in those to whom his demonstration is addressed is the awareness that they themselves will the ‘ought.’ The addressee thus realizes that he himself exercises the coercion that the moral law undoubtedly exercises because it is he himself who establishes this law. Thus we see that what was possibly not willed at all— the morally good as what we ought to do—is now recognized and understood as what is authentically willed, so that ultimately what is authentically willed is what is really willed. One who can show that what we ought to do is what we authentically already will is the one who best accomplishes that

Section III of the Groundwork

213

the ‘ought’ is also in fact willed— at least this is what Kant seems to believe.38

Critical Observations on the Deduction As with the thesis of analyticity earlier, there are also some evident difficulties with the deduction. These derive above all from a misapplication of Kant’s distinction between ‘appearance’ and ‘thing in itself’: 1. It is not possible to provide a really convincing reconstruction of the ‘grounding relation’ in the first part of the OP. This first part involves three specific assertions: (i) ‘The world of understanding contains the ground of the world of sense’; (ii) ‘Hence the world of understanding also contains the ground of the laws of the world of sense’; (iii) ‘In regard to my will, which belongs wholly to the world of understanding, this world is immediately legislative and must be thought in such a way that it contains the ground of the world of sense and the ground of the laws of that world.’ The general principle that the world of understanding is the ground of (the laws of) the world of sense is obviously supposed to be applied to the par ticular case of the will. But what precisely does this general principle consist in? It is plausible to read (i) in terms of the elemental relationship between the world of understanding and the world of sense (‘that behind appearances things in themselves (although hidden) must ground 38. Kant’s strategy is also clearly revealed by his observation that respect “has something analogical” (401:30; o.e.) with inclination and fear: “The object of respect is thus solely the law, and specifically that law that we lay upon ourselves and yet also as in itself necessary. As a law we are subject to it without asking permission of self-love; as laid upon us by ourselves, it is a consequence of our will, and has from the first point of view an analogy with fear, and from the second with inclination” (401:31; o.e., Kant’s emphasis in part omitted). By ‘the fi rst point of view’ Kant is referring to the fact that we are ‘subject’ to the moral law. The analogy with fear is illuminating since the fact that we are subject to the law also always means that the moral law has a compulsory character. But the ‘analogy with inclination’ is clear and instructive as well, for insofar as we ‘lay the law upon ourselves’ it is a ‘consequence of our will,’ or, in other words, something that we ourselves will. The analogy with inclination lies in the fact that an inclination is always something, or rather is always directed at something, ‘that we ourselves will.’ Respect is thus a ‘feeling of pleasure or satisfaction,’ albeit one of a distinctive kind; it is directed toward what we ourselves authentically will.

214

Immanuel Kant’s Groundwork for the Metaphysics of Morals

them’); the world of sense is “grounded” (451:21; 459:29) in the world of understanding. But then how does it stand with (ii)? One might suppose that Kant is referring here to the principal claim of his theory of knowledge: that the transcendental-logical I is the ground of the laws of nature. But in the first place this does not follow, as the “hence” (453:32) suggests, from the fact that in the elemental sense already mentioned the world of understanding contains the ground of the world of sense. And in the second place it is completely unclear how that general principle can actually be applied to the will. 2. The most serious difficulty with the deduction lies in the par ticular way in which Kant employs his distinction between appearance and thing in itself. As we have suggested earlier, these terms need not refer to (numerically) distinct worlds but rather to different standpoints or perspectives. One and the same thing can be characterized as thing in itself or as appearance. In a certain sense this does mean that the world of appearances in fact is not more than mere appearance, for whatever may appear in space and time cannot, according to Kant, actually exist (as thing in itself) in the way in which it is described as appearance, because space and time are nothing but mere a priori forms of intuition, to which nothing in the world of things in themselves corresponds. But only if one completely abandons the entire Kantian system could this be taken to mean that inclinations “as belonging to the world of sense” (453:24) are not actual in the narrower (rather than the phenomenal) sense; they must be more than mere representational items or contents. An inclination must be understood as something that, like other entities, also exists at the same time as thing in itself and as appearance, even if it can be experienced only qua appearance. But Kant assumes, in an almost vulgarly platonic sense, that inclinations possess, not only in an epistemological but also in an ontological sense, an entirely different status from things in themselves; as we have seen, he deploys the idea that whatever belongs to the realm of appearance is ‘subordinated’ to that of the thing in itself. But in the context of the Kantian system itself, that makes little sense. 3. If, on the other hand, we understand the distinction between appearance and thing in itself in its genuine and original Kantian

Section III of the Groundwork

215

sense, then the question of the deduction remains without an answer. The human being is certainly a being that is at once rational and sensuous, a being whose sensibility (interests and inclinations) can be understood in terms of appearances. But it does not follow from this that these appearances are in any sense subordinated to reason. For reason also appears,39 and inclinations are neither more nor less ‘actual’ than reason. But if this is so, why should the human being, as a being that belongs at once to the world of understanding and to the world of sense, subject himself to the laws of the world of understanding? The question of the deduction (in the sense ‘Why be moral?’) is answered by the claim that the human being as member of the world of understanding is his ‘authentic self.’ But that is really no answer at all, for the original question immediately reappears here in a new form: why should the human being subject his sensuously motivated volition to the restrictions of his purely rationally motivated volition? The only answer to this question is this: because the human being as a rational being is an end in himself and possesses worth. And this is indeed the answer that Kant gives. But instead of clarifying what it means to say that something possesses value or worth, and how value or worth can be recognized and identified, he persists in upholding a philosophically implausible ontological claim that is inappropriate in the context of his own thought. It is little wonder that few even of Kant’s most devoted admirers have ever been satisfied with the deduction of the moral law he offers in GMS III.

4.4 Summary 1. The question how a CI is possible involves three specific questions: Why is the CI valid? How can we meaningfully conceive of freedom, and why should we regard ourselves as free? And how can pure reason produce an interest in the moral law? The third question cannot be answered. 39. This is one of the principal theses of the ‘solution to the third antinomy’ in the KrV; cf. especially A538ff. /B566ff.

216

Immanuel Kant’s Groundwork for the Metaphysics of Morals

2. The thesis of analyticity tells us that a will that is characterized by reason and freedom alone is a will that always wills the good. For beings of this kind—for purely rational beings, or for sensuous-rational beings in their capacity as members of the world of understanding— the moral law is an analytic proposition rather than an imperative. The argument for the thesis of analyticity runs as follows: since any causality must possess a law-like character, and since there can be no law of evil and practical reason is independent of the laws of nature, the only thing that is left to determine practical reason is mere lawfulness itself; free practical reason is thus a reason that stands under the moral law. 3. The freedom of thinking (spontaneity) cannot meaningfully be contested, since the act of contesting this freedom is itself an act of freedom (namely, the freedom of judging in accordance with objective reasons and considerations). 4. Kant connects the thought of spontaneity with the distinction between the world of understanding (the world of things in themselves) and the world of sense (the world of appearances). Human beings can justifiably regard themselves as free on the basis of the spontaneity of their reason and understanding, and thereby also as members of the world of understanding; they recognize their I as intelligence and thus as thing in itself. On the basis of the supposed unity of theoretical and practical reason Kant infers practical freedom from theoretical freedom. 5. The suspicion of circularity that arises in relation to the question of freedom is not that of a vicious circle (i.e., Kant does not infer morality from freedom and freedom from morality). It is a question of two petitiones principii: the suspicion of circularity makes it clear that up to this point (until after Section 2) the freedom of the human will has not yet been shown; and it makes it clear that the validity of the moral law as an imperative by no means follows even from the connection between such an accomplished demonstration and the thesis of analyticity. Kant must therefore show that the human being is free and is therefore able to act morally; but he must also show that the human being ought to act morally.

Section III of the Groundwork

217

6. Kant grounds the validity of the CI in the ontoethical principle that the moral law of the world of understanding, on the basis of the ontological superiority of this world, is also valid as a categorical imperative for sensuous-rational beings that are members of the world of understanding and of the world of sense. What the human being ought to do is what he already wills to do as a rational being (as his ‘authentic self’). 7. The thesis of analyticity, the deduction of freedom, and the deduction of the CI are all open to serious criticism. Three points above all should be mentioned in this connection: (i) Kant employs the distinction between the world of understanding and the world of sense in an ontologizing way that does not correspond to the (epistemologically motivated) way in which he originally introduces this distinction. (ii) Kant understands the good will as part of the world of understanding and the bad or evil will as part of the world of sense. But the relationship between these two worlds remains unclarified; hence it also remains unclear whether bad or evil acts may be ascribed to the free will (if so, then the thesis of analyticity is incorrect; if not, then there can be no responsibility for bad or evil acts). (iii) The unity of theoretical and practical reason is affirmed but is not itself provided with justification; the transition from the freedom of the I that thinks to the freedom of the I that wills thus remains unsatisfactory.

5 Bibliography

5.1 Kant Kant’s Groundwork for the Metaphysics of Morals (GMS) is quoted in the translation by Allen W. Wood (New Haven: Yale University Press, 2002). The page numbers and line numbers given in brackets without the abbreviation (for example, 406:25) are page and line numbers from the Groundwork edition of Akademie-Ausgabe (1902–) (AA). The Critique of Pure Reason is cited according to the original pagination (A and B), for example, B 374. All other writings by Kant are cited according to the abbreviations given below and the pagination of AA. English translations can be found in the appropriate volumes of the Cambridge Edition of the Writings of Immanuel Kant (Ca), where the AA pagination is given in the margins. Anth

Anthropologie in pragmatischer Hinsicht, AA 7 Anthropology from a Pragmatic Point of View, Ca: Writings on Anthropology, History, and Education

DG

Untersuchung über die Deutlichkeit der Grundsätze der natürlichen Theologie und Moral, AA 2 Inquiry Concerning the Distinctness of the Principles of Natural Theology and Morality, Ca: Theoretical Writings, 1755–1770

ED

Über eine Entdeckung, nach der alle neue Kritik der reinen Vernunft durch eine ältere entbehrlich gemacht werden soll, AA 8 On a Discovery Whereby Any New Critique of Pure Reason Is to Be Made Superfluous by an Older One, Ca: Theoretical Writings after 1781

220

Bibliography

GMS

Grundlegung zur Metaphysik der Sitten, AA 4

I

Idee zu einer allgemeinen Geschichte in weltbürgerlichen Absicht, AA 8 Idea for a Universal History with a Cosmopolitan Aim, Ca: Writings on Anthropology, History, and Education

JL

Immanuel Kants Logik: Ein Handbuch zu seinen Vorlesungen, ed. G. B. Jäsche, AA 9 The Jäsche Logic, Ca: Lectures on Logic

KpV

Kritik der praktischen Vernunft, AA 5 Critique of Practical Reason, Ca: Practical Philosophy

KrV

Kritik der reinen Vernunft, AA 3 (B) and 4 (A) Critique of Pure Reason, Ca: Critique of Pure Reason

KU

Kritik der Urteilskraft, AA 5 Critique of the Power of Judgment, Ca: Critique of the Power of Judgment

MC

Moralphilosophie Collins, AA 27 Collins’s Lecture Notes, Ca: Lectures on Ethics

Menzer Paul Menzer: Zur Vorlesung Kants über Ethik, Berlin, 1924. ML1

Metaphysik L1, AA 28 Metaphysik L1, Ca: Lectures on Metaphysics

MM

Metaphysik Mrongovius, AA 29 Metaphysic Mrongovius, Ca: Lectures on Metaphysics

MM II Moral Mrongovius II, AA 29 Mrongovius’s Second Set of Lecture Notes, Ca: Lectures on Ethics MS

Metaphysik der Sitten, AA 6 Metaphysics of Morals, Ca: Practical Philosophy

MSV

Metaphysik der Sitten Vigilantius, AA 27 Vigilantius’s Lecture Notes, Ca: Lectures on Ethics

NF

Naturrecht Feyerabend, AA 27

O

Was heißt: Sich im Denken orientieren?, AA 8 What Does It Mean to Orient Oneself in Thinking?, Ca: Writings on Religion and Rational Theology

PM

Prolegomena zu einer jeden künftigen Metaphysik, die als Wissenschaft wird auftreten können, AA 4 Prolegomena to Any Future Metaphysics, Ca: Theoretical Philosophy after 1781

R

Die Religion innerhalb der Grenzen der bloßen Vernunft, AA 6 Religion within the Boundaries of Mere Reason, Ca: Writings on Religion and Rational Theology

Refl

Reflexionen, AA 16–19 Ca: Notes and Fragments

Bibliography

221

RP

Religionslehre Pölitz, AA 28 Lectures on the Philosophical Doctrine of Religion, Ca: Writings on Religion and Rational Theology

RS

Recension von Schulz’s “Versuch einer Anleitung zur Sittenlehre,” AA 8 Review of Schulz’s “Attempt at an Introduction to a Doctrine of Morals for all Human Beings Regardless of Different Religions,” Ca: Practical Philosophy

Theo

Über das Mißlingen aller philosophischen Versuche in der Theodicee, AA 8 On the Miscarriage of All Philosophical Trials in Theodicy, Ca: Writings on Religion and Rational Theology

TP

Über den Gemeinspruch: Das mag in der Theorie richtig sein, taugt aber nicht für die Praxis, AA 8 On the Common Saying: That May Be Correct in Theory, but It Does Not Work in Practice, Ca: Practical Philosophy

VRL

Über ein vermeintes Recht aus Menschenliebe zu lügen, AA 8 On a Supposed Right to Lie from Philanthropy, Ca: Practical Philosophy

WA

Beantwortung der Frage: Was ist Aufklärung?, AA 8 An Answer to the Question: What Is Enlightenment?, Ca: Practical Philosophy

In addition, the following abbreviations will be used: “GMS I,” “GMS II,” and “GMS III” refer to the fi rst, second, and third sections of GMS. The subsections of GMS III are called Section 1, Section 2, and so on (thus “Section 1” indicates the subsection “The Concept of Freedom Is the Key to the Defi nition of Autonomy of the Will”). The abbreviation “o.e.” means “our emphasis.”

5.2 Secondary Literature The literature on Kant’s ethics and on Groundwork for the Metaphysics of Morals is so great that it can hardly be surveyed. The following list is a narrow selection of contributions, and with few exceptions only those in which GMS or at least large passages from it are treated, as well as collections of essays (published before 2011). For a comprehensive bibliography, see the two volumes edited by Margit Ruffi ng, Kant- Bibliographie, 1896–1944 and Kant- Bibliographie, 1945–1990, (Frankfurt, 2007), and the current bibliographies published in Kant- Studien. Albrecht, Michael (1994): Kants Maximenethik und ihre Begründung. KantStudien 85, 129–146. Allison, Henry E. (1990): Kant’s Theory of Freedom. Cambridge. ——— (1996): On the Presumed Gap in the Derivation of the Categorical Imperative. In Idealism and Freedom: Essays on Kant’s Theoretical and Practical Philosophy. Cambridge, 143–154.

222

Bibliography

——— (2011): Kant’s “Groundwork for the Metaphysics of Morals”: A Commentary. Oxford. Alquié, Ferdinand (1985): Présentation des “Fondements de la Métaphysique de Mœurs.” Vol. 2. Paris. Ameriks, Karl (2000): Kant’s Theory of Mind: An Analysis of the Paralogisms of Pure Reason. 2nd ed. Oxford. ——— (2003): Kant’s Deduction of Freedom and Morality. In Interpreting Kant’s “Critiques.” Oxford, 161–192. Aune, Bruce (1979): Kant’s Theory of Morals. Princeton. Bailey, Tom (2011): Analysing the Good Will: Kant’s Argument in the First Section of the Groundwork. British Journal for the History of Philosophy 18(4), 635– 662. Baker, Judith (1988): Counting Categorical Imperatives. Kant-Studien 79, 389–406. Baron, Marcia (1995): Kantian Ethics Almost without Apology. Ithaca, NY. Baumanns, Peter (1982): Kants kategorischer Imperativ und das Problem der inhaltlichen Pflichtbestimmung. In Herta Nagl-Docekal (ed.): Überlieferung und Aufgabe: Festschrift für Erich Heintel zum 70. Geburtstag, vol. 2. Vienna, 165–179. Bittner, Rüdiger (1974): Maximen. In Gerhard Funke (ed.): Akten des 4. Internationalen Kant-Kongresses. Berlin, 485–498. ——— (1980): Hypothetische Imperative. Zeitschrift für philosophische Forschung 34, 210–226. ——— (2000): Wer frei ist, ist gebunden: Kants Argument aus dem dritten Abschnitt der Grundlegungsschrift. In Philosophiegeschichte und logische Analyse / Logical Analysis and History of Philosophy, vol. 3, Philosophie der Neuzeit / From Descartes to Kant. 209–221. Brandt, Reinhard (1988): Der Zirkel im dritten Abschnitt von Kants Grundlegung zur Metaphysik der Sitten. In Hariolf Oberer / Gerhard Seel (eds.): Kant: Analysen— Probleme—Kritik. Würzburg, 169–191. Brinkmann, Walter (2003): Praktische Notwendigkeit: Eine Formalisierung von Kants Kategorischem Imperativ. Paderborn. Burri, Alex / Freudiger, Jürg (1990): Zur Analytizität hypothetischer Imperative. Zeitschrift für philosophische Forschung 44, 98–105. Cicovacki, Predrag (2001): Zwischen gutem Willen und Kategorischem Imperativ. In Hans-Ulrich Baumgarten / Carsten Held (eds.): Systematische Ethik mit Kant. Freiburg, 330–354. Dalbosco, Cláudio Almir (2008): “Círculo vicioso” e idealismo transcendental na Grundlegung. Studia Kantiana 6/7, 207–235. Dean, Richard (2009): The Formula of Humanity as End in Itself. In Thomas E. Hill (ed.): The Blackwell Guide to Kant’s Ethics. Oxford, 83–101. Delfosse, Heinrich P. (2000): Stellenindex und Konkordanz zur “Grundlegung zur Metaphysik der Sitten.” Vol. 15 of Kant-Index. Stuttgart. Duncan, Alistair R. C. (1957): Practical Reason and Morality. A Study of Immanuel Kant’s “Foundations for the Metaphysics of Morals.” London. Esteves, Julio (2003): A dedução do imperative categórico na Fundamenção III. Studia Kantiana 5, 79–104.

Bibliography

223

Freudiger, Jürg (1993): Kants Begründung der praktischen Philosophie: Systematische Stellung, Methode und Argumentationsstruktur der “Grundlegung zur Metaphysik der Sitten.” Bern. Galvin, Richard (2009): The Universal Law Formulas. In Thomas E. Hill (ed.): The Blackwell Guide to Kant’s Ethics. Oxford, 52–82. Grenberg, Jeanine M. (2009): The Phenomenological Failure of Groundwork III. Inquiry 52, 335–356. Guyer, Paul (ed.) (1998): Kant’s “Groundwork of the Metaphysics of Morals”: Critical Essays. Totowa, NJ. ——— (2002): The Derivation of the Categorical Imperative: Kant’s Correction for a Fatal Flaw. Harvard Review of Philosophy 10, 64– 80. ——— (2007): Kant’s “Groundwork for the Metaphysics of Morals”: A Reader’s Guide. London. Henrich, Dieter (1975): Die Deduktion des Sittengesetzes: Über die Gründe der Dunkelheit des letzten Abschnittes von Kants Grundlegung zur Metaphysik der Sitten. In Alexander Schwan (ed.): Denken im Schatten des Nihilismus: Festschrift für Wilhelm Weischedel zum 70. Geburtstag. Darmstadt, 55–112. Herman, Barbara (1993): The Practice of Moral Judgment. Cambridge, MA. Hill, Thomas E., Jr. (1992): Dignity and Practical Reason in Kant’s Moral Theory. Ithaca, NY. Hill, Thomas E., Jr. / Zweig, Arnulf (2002): Editors’ Introduction: Some Main Themes of the Groundwork; Analysis of Arguments. In Immanuel Kant (1785): Groundwork for the Metaphysics of Morals, translated by Arnulf Zweig and edited by Thomas E. Hill Jr. and Arnulf Zweig. Oxford, 19–108 and 109–177. Höffe, Ottfried (ed.) (2010): Kants “Grundlegung zur Metaphysik der Sitten”: Ein kooperativer Kommentar. 4th ed. Frankfurt am Main. Högemann, Brigitte (1980): Die Idee der Freiheit und das Subjekt: Eine Untersuchung von Kants “Grundlegung zur Metaphysik der Sitten.” Königstein/Ts. Holtman, Sarah (2009): Autonomy and the Kingdom of Ends. In Thomas E. Hill (ed.): The Blackwell Guide to Kant’s Ethics. Oxford, 102–117. Horn, Christoph / Mieth, Corinna / Scarano, Nico (2007): Immanuel Kant: Grundlegung zur Metaphysik der Sitten. Kommentar von Ch. Horn, C. Mieth und N. Scarano. Berlin. Horn, Christoph / Schönecker, Dieter (eds.) (2006): Groundwork for the Metaphysics of Morals. Berlin. Huber, Herbert (2010): Immanuel Kant: Grundlegung zur Metaphysik der Sitten. In Herbert Huber (ed.): Klassische Werke zur philosophischen Ethik: Ein Studienbuch für Philosophie- und Ethiklehrer. Freiburg, 171–199. Johnson, Robert (2009): Good Will and the Moral Worth of Acting from Duty. In Thomas E. Hill (ed.): The Blackwell Guide to Kant’s Ethics. Oxford, 19–51. Kaplan, Shawn D. (2008): Bringing the Moral Law Closer to Intuition and Feeling: An Interpretive Framework for Kant’s Groundwork II. In Valerio Rohden / Ricardo R. Terra / Guido A. de Almeida / Margit Ruffi ng (eds.): Recht und Frieden in der Philosophie Kants, vol. 3. Berlin, 161–171.

224

Bibliography

Kaulbach, Friedrich (1988): Immanuel Kants “Grundlegung zur Metaphysik der Sitten”: Interpretation und Kommentar. Darmstadt. Kerstein, Samuel J. (2002): Kant’s Search for the Supreme Principle of Morality. Cambridge. Kim, Halla (2002): Has Kant committed the Fallacy of Circularity in Foundations III? Journal of Philosophical Research 27, 64– 81. Kirchmann, Julius Hermann von (1874): Erläuterungen zur “Grundlegung der Metaphysik der Sitten.” Leipzig. Köhl, Harald (1990): Kants Gesinnungsethik. Berlin. Korsgaard, Christine M. (1996): Creating the Kingdom of Ends. Cambridge. ——— (1998): Introduction to Groundwork of the Metaphysics of Morals, by Immanuel Kant. Translated and edited by Mary Gregor. Cambridge. Kraft, Bernd / Schönecker, Dieter (1999): Introduction to Grundlegung zur Metaphysik der Sitten, by Immanuel Kant. Edited by Bernd Kraft and Dieter Schönecker. Hamburg, vii–xxxix. Kupperman, Joel J. (2002): A Messy Derivation of the Categorial Imperative. Philosophy 77, 484–502. Liddell, Brendan E. A. (1970): Kant on the Foundation of Morality: A Modern Version of the “Grundlegung.” Translated with commentary. Bloomington. Ludwig, Bernd (2008): Was wird in Kants Grundlegung eigentlich deduziert? Über einen Grund der vermeintlichen Dunkelheit des “Dritten Abschnitts.” In B. Sharon Byrd / Joachim Hruschka / Jan C. Joerden: Jahrbuch für Recht und Ethik 16, 431–463. McCarthy, Michael H. (1976): Analytic Method and Analytic Propositions in Kant’s Groundwork. Dialogue (Canada) 15, 565–582. ——— (1979a): Kant’s Application of the Analytic/Synthetic Distinction to Imperatives. Dialogue (Canada) 18, 373–391. ——— (1979b): Paton’s Suggestion that Kant’s Principle of Autonomy Might be Analytic. Kant- Studien 70, 206–224. ——— (1982): Kant’s Rejection of the Argument of Groundwork III. Kant- Studien 73, 169–190. ——— (1984): Kant’s Groundwork Justification of Freedom. Dialogue (Canada) 23, 457–473. ——— (1985): The Objection of Circularity in Groundwork III. Kant- Studien 76, 28–42. Melches Gibert, Carlos (1994): Der Einfluß von Christian Garves Übersetzung Ciceros “De offi ciis” auf Kants “Grundlegung zur Metaphysik der Sitten.” Regensburg. Melnick, Arthur (2002): Kant’s Formulations of the Categorical Imperative. Kant- Studien 93, 291–308. Nell (O’Neill), Onora (1975): Acting on Principle. New York. O’Neill, Onora (1989): Constructions of Reason: Explorations of Kant’s Practical Philosophy. Cambridge. Onof, Christian (2009): Reconstructing the Grounding of Kant’s Ethics. KantStudien 100(4), 496–517.

Bibliography

225

Pasternack, Lawrence (ed.) (2002): Immanuel Kant: Groundwork of the Metaphysics of Morals. London. Paton, Henry James (1958): The Aim and Structure of Kant’s Grundlegung. Philosophical Quarterly 8, 112–130. Paton, Herbert James (1971): The Categorical Imperative: A Study in Kant’s Moral Philosophy. Philadelphia. Pistorius, Hermann Andreas (1787 [2007]): Rezension der Grundlegung zur Metaphysik der Sitten. In Bernward Gesang (ed.): Kants vergessener Rezensent: Die Kritik der theoretischen und praktischen Philosophie Kants in fünf frühen Rezensionen von Hermann Andreas Pistorius. Hamburg, 26–38. Porcheddu, Rocco (2009): Zweck an sich selbst und Subjektivität: Ein Versuch zu Kants Grundlegung zur Metaphysik der Sitten. In Christoph Asmuth (ed.): Kant und Fichte— Fichte und Kant. Amsterdam, 177–188. Puls, Heiko (2011): Freiheit als Unabhängigkeit von bloß subjektiv bestimmten Ursachen: Kants Auflösung des Zirkelverdachts im dritten Abschnitt der Grundlegung zur Metaphysik der Sitten. Zeitschrift für philosophische Forschung 65(4), 534–562. Quarfood, Marcel (2001): Kant’s Practical Deduction of Moral Obligation in Groundwork III. In Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (eds.): Kant und die Berliner Aufklärung, vol. 3. Berlin, 72–79. Rawls, John (2000): Lectures on the History of Moral Philosophy. Edited by Barbara Herman. Cambridge, MA. Ross, Sir William David (1954): Kant’s Ethical Theory: A Commentary on the “Grundlegung zur Metaphysik der Sitten.” Oxford. Schneewind, Jerome B. (2009): Why Study Kant’s Groundwork? In Essays on the History of Moral Philosophy. Oxford, 239–247. Schönecker, Dieter (1997): Die “Art von Zirkel” im dritten Abschnitt der Grundlegung. Allgemeine Zeitschrift für Philosophie 22, 189–202. ——— (1999): Kant: “Grundlegung” III: Die Deduktion des kategorischen Imperativs. Freiburg. ——— (2001): What Is the ‘First Proposition’ Regarding Duty in Kant’s Grundlegung? In Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (eds.): Kant und die Berliner Aufklärung, vol. 3. Berlin, 89– 95. ——— (2009): The Transition from Common Rational Moral Knowledge to Philosophical Rational Moral Knowledge in the Groundwork. In Karl Ameriks / Otfried Höffe (eds.): Kant’s Moral and Legal Philosophy. Cambridge, 93–122. ——— (2011): Kants Grundlegung über den bösen Willen: Eine kommentarische Interpretation von GMS 457.25–458.5. Studi Kantiani 24, 73– 91. ——— (2012a): “A Free Will and a Will under Moral Laws are the Same”: Kant’s Concept of Autonomy and His Thesis of Analyticity in Groundwork III. In Oliver Sensen (ed.): Kant on Moral Autonomy. Cambridge, 225–245. ——— (2012b): Once Again: What Is the “First Proposition” in Kant’s Groundwork? Some Refi nements, a New Proposal, and a Reply to Henry Allison. Kantian Review 17(2), 1–16.

226

Bibliography

Schossberger, Cynthia (2008): The Kingdom of Ends and the Fourth Example in the Groundwork II. In Valerio Rohden / Ricardo R. Terra / Guido A. de Almeida / Margit Ruffi ng (eds.): Recht und Frieden in der Philosophie Kants, vol. 3. Berlin, 369–377. Scott, J. W. (1924): Kant on the Moral Life: An Exposition of Kant’s “Grundlegung.” London. Sedgwick, Sally (2008): Kant’s “Groundwork of Metaphysics of Morals”: An Introduction. Cambridge. Sensen, Oliver (2009): Kant’s Conception of Human Dignity. Kant- Studien 100, 309–331. ——— (2010): Dignity and the Formula of Humanity (ad IV 429, IV 435). In Jens Timmermann (ed.): Kant’s “Groundwork of the Metaphysics of Morals”: A Critical Guide. Cambridge, 102–118. ——— (2011): Kant on Human Dignity. Berlin. Siep, Ludwig (2009): What Is the Purpose of a Metaphysics of Morals? Some Observations on the Preface to the Groundwork of the Metaphysic of Morals. In Karl Ameriks / Otfried Höffe (eds.): Kant’s Moral and Legal Philosophy. Cambridge, 77– 92. Staege, Roswitha (2002): Hypothetische Imperative. Kant- Studien 93, 42–56. Stattler, Benedikt (1788 [1968]): Anhang zum Anti-Kant in einer Widerlegung der Kantischen “Grundlegung zur Metaphysik der Sitten.” Munich. Reprint in the series Aetas Kantiana. Brussels. Steigleder, Klaus (2002): Kants Moralphilosophie: Die Selbstbezüglichkeit reiner praktischer Vernunft. Stuttgart. Stern, Robert (2013): Kant, Moral Obligation, and the Holy Will. In Mark Timmons / Sorin Baiasu (eds.): Kant on Practical Justifi cation: Interpretative Essays. Oxford, 125–152. Sullivan, Roger J. (1989): Immanuel Kant’s Moral Theory. Cambridge. ——— (1994): An Introduction to Kant’s Ethics. Cambridge. Tenenbaum, Sergio (2012): The Idea of Freedom and Moral Cognition in Groundwork III. Philosophy and Phenomenological Research 84, 555–589. Timmermann, Jens (ed.) (2004): Immanuel Kant: Grundlegung zur Metaphysik der Sitten. Göttingen. ——— (2007): Kant’s “Groundwork of the Metaphysics of Morals”: A Commentary. Cambridge. ——— (ed.) (2009): Kant’s “Groundwork of the Metaphysics of Morals”: A Critical Guide. Cambridge. Tugendhat, Ernst (1993): Vorlesungen über Ethik. Frankfurt am Main. Van Erp, Herman (2007): Mogelijkheid en geldigheid van de categorische imperatief: Kants bewijsvoering in de Grundlegung zur Metaphysik der Sitten. Tijdschrift vor Filosofie 69, 299–324. Wenzel, Uwe Justus (1992): Anthroponomie: Kants Archäologie der Autonomie. Berlin. Wilde, Leo Henri (1975): Hypothetische und kategorische Imperative: Eine Interpretation zu Kants “Grundlegung zur Metaphysik der Sitten.” Bonn.

Bibliography

227

Willaschek, Marcus (1992): Praktische Vernunft: Handlungstheorie und Moralbegründung bei Kant. Stuttgart. Wimmer, Reiner (1980): Universalisierung in der Ethik. Analyse, Kritik und Rekonstruktion ethischer Rationalitätsansprüche. Frankfurt a. M. Wolff, Robert P. (ed.) (1969): Foundations of the Metaphysics of Morals. Translated by Lewis White Beck, with critical essays edited by Robert Paul Wolff. New York. ——— (1973): The Autonomy of Reason: A Commentary on Kant’s “Groundwork of the Metaphysic of Morals.” New York. Wood, Allen W. (1999): Kant’s Ethical Thought. Cambridge. ——— (ed. and trans.) (2002): Immanuel Kant, Groundwork for the Metaphysics of Morals. With essays by J. B. Schneewind, Marcia Baron, Shelly Kagan, and Allen W. Wood. New Haven. ——— (2005): Kant. Malden, MA. ——— (2008): Kantian Ethics. New York. ——— (2014): The Free Development of Each: Studies on Freedom, Right and Ethics in Classical German Philosophy. Oxford. Wyrwich, Thomas (2011): Moralische Selbst- und Welterkenntnis: Die Deduktion des kategorischen Imperativs in der Kantischen Philosophie. Würzburg.

Index

Absicht, 35n1, 37, 38n7, 48n20, 83n60. See also motives and intentions action. See also imperatives; moral good and morality from duty (see duty) ends and, 37, 42, 44–50, 101–103 (see also end-means relation; humanity as end in itself ) from inclinations (see inclinations) analytical method (conceptual-analytical method), 4–7, 13, 16, 30, 204. See also thesis of analyticity the ancients, 36n4, 39n9, 50–51, 88n65 anthropology, 26–29, 41 appearances, 20, 160, 186, 194–195, 198, 210, 214–215. See also causality and determinism; world of sense approval, 37, 40n12, 41, 71, 72, 161n82 apriority. See also necessity; universality CI and, 7n11, 16–19, 29, 108–109, 180–181, 209 duty and, 56 enlightenment and, 25 freedom and, 24–25, 31 happiness and, 113–115 of metaphysics of morals, 2–3, 14–24, 31, 132, 151n71 principle of the will and, 19, 77–78 Aristotle, 48 Auszug aus der Vernunftlehre (Meyer), 201n35 authentic self, 145, 191n23, 209, 210–213, 215, 217

autonomy. See also categorical imperative formulas, FA CI and, 96 defi ned, 145 degrees of, 148–149 duty and, 107 freedom and, 53, 103, 105, 145–146, 184–186, 188, 189 heteronomy versus, 14, 24, 95, 96, 97, 187–188 moral law and, 25, 91– 92 rational beings and, 144–145, 151n71 realm of ends and, 156–164 respect and, 71, 91 beneficence and benevolence, 36n5, 63, 70, 71, 74, 76n44, 130, 137–139 categorical imperative (CI). See also categorical imperative deduction; categorical imperative formulas; moral law absolute worth and, 88– 89 categorical validity of, 18, 22–24, 29 (see also necessity) common rational moral cognition and, 9, 80n54 consequentialism and, 42 deduction of, 175–215 derivation of, 4–5, 34, 62, 89– 93 ends and, 84 example of, 34 grounding of, 144–145 hypothetical versus, 2, 102, 157

230

categorical imperative (CI) (continued) imperatives and, 123 maxims and, 166 necessity and, 18 objective determination of duty and, 60 possibility of, 87, 125, 175–178 realm of ends and, 156–164 second proposition regarding duty and, 76–78 singularity of, 164–165 as synthetic, 7n11, 16–18, 109, 127, 158, 180–183, 204, 208–209 universal form of, 17, 29, 89, 92, 101–102, 124, 127, 165–166 (see also categorical imperative formulas, FU and FUL; universality of moral law) universalizability and, 92, 106, 125–141 validity of, 18n28, 29, 177–178 categorical imperative deduction. See also thesis of analyticity ‘analytic’/’synthetic’ and, 7 argument of, 185, 205–213 circularity and, 194, 200–205 critical observations of, 213–215 freedom of will and, 178–180, 192–193 human beings as thing in itself and, 196–198 ought and, 6 unity of reason and, 198–200 worlds of sense/understanding and, 194–196 categorical imperative formulas FA (of autonomy), 124, 126, 128, 155, 156–164, 161–162, 164n89 FH (of humanity as end in itself), 92, 126, 136n46, 141–156, 159, 161–162, 163 FLN (of law of nature), 96, 97, 126–134, 158, 160, 161–162, 164n90 FRE (of realm of ends), 124, 149, 156–164 FU (of universality), 124–125, 127, 139, 165–169, 173 FUL (of universal law), 17, 32, 124–141, 150n69, 155, 158n78, 159–162, 165 function of, 166–167 metaphysics of morals and, 96, 97 overview, 122–125 relationships to one another, 164–172 causality and determinism, 186–187, 188, 192–193, 198–199, 201–202, 205, 206. See also heteronomy; law-like character character, 10, 36–40, 83. See also virtues coercion, 71, 91, 157n76, 163, 212

Index

common rational moral cognition. See also approval; esteem; estimation absolute worth and, 89 analytical method and, 6–13 correctness of, 86 on happiness, 41 inclinations and, 12–13, 58–59, 69–73 natural dialectic and, 34 teleological argument and, 47 universality and, 21 world of understanding and, 210n37 consequences (aims) (effects) (results). See also end-means relation Absicht and, 38n7 actual and accomplished, 44–45, 46, 58n25 good will and, 33, 34, 41–47 intended/unintended, 42, 44, 45–47 moral law versus, 81– 82, 90– 91 worth and, 77, 83– 89 consequentialism, 41–47, 87– 89, 90, 92, 152 contingency, 10–11, 12n19, 52. See also empirical factors contractualism (contractarianism), 85– 86, 164 contradiction in conception/will tests, 130, 132, 135–141, 160 correspondence thesis, 130, 139–140, 173 corruption, 12–13, 15–16, 31, 40, 41, 93 deontological moral theories, 41–42, 141–142. See also consequentialism determinism. See causality and determinism dialectic, natural, 8n12, 12–13, 30, 32, 34, 130–131. See also corruption dignity. See also humanity as end in itself; respect FH and, 173–174 free, rational beings and, 45, 82, 151n71, 162n86 interests and, 86 qualities of temperament and, 36n5 suicide and, 154 worth and, 88, 141–149 duty. See also deontological moral theories; respect acting contrary to, 59– 60 acting from, 8n13, 25, 45, 55, 56, 58–76, 84– 86, 90– 91, 178 acting in conformity to, 15, 42–43, 60– 65, 86, 137

Index

acting in conformity to in addition to inclination, 59, 64–74, 84 analysis of concept of, 4–5, 34, 54– 89 apriority and, 15, 18–21, 56 CI and, 4, 42, 123, 125–164, 171 common rational moral cognition and, 9 common use of practical reason and, 11n18 concept not empirical, 19–20, 23, 25, 31 correspondence thesis and, 130, 139–140, 173 derivation of, 6, 150n69, 155, 164n90 division of, 3, 3n5, 96 empirical knowledge and, 29, 31 feeling and, 69–71, 74–76 FH and, 149–156 fi rst proposition regarding, 54–76 freedom versus, 184–185, 186, 202–204 good will and, 32–33, 43, 47, 51–54, 62 inclinations and, 51–52, 58–59, 64– 68 interest and, 91– 92, 137 necessity/universality and, 21–23, 55, 56–58, 107 perfect versus imperfect, 130, 140, 150–151 respect and, 50–51, 55–56, 57, 67 second proposition regarding, 76–78 subjective and objective determination of, 52–53, 56–58, 59– 60 (see also respect) third proposition regarding, 78– 89 toward oneself and toward others, 60, 63, 66, 76n45, 81, 87, 130, 137, 140n48, 152 effects, 33, 34, 41–47, 83– 89. See also consequences empirical factors. See also causality; contingency; ethics, applied versus pure; heteronomy; human nature; inclinations; nature; sciences benevolence and, 137–138 common rational moral cognition and, 11 duty and, 19–20 FLN and, 132 freedom and, 186 of Kant’s philosophy, 29n47 moral philosophy and, 14–16 objectively practical laws and, 28 promising and, 136n36 purity of GMS and, 29 empiricist tradition, 5

231

end in itself. See humanity as end in itself end-means proposition (EMP), 116–122, 135 end-means relation, 35, 40, 43, 84, 110–111, 152–155, 163. See also consequences; hypothetical imperatives; realm of ends end-setting, 103, 115, 122 Enlightenment, 25 esteem, 41, 45, 58, 69, 70, 72, 73 estimation, 46, 72n39, 83n59 ethics, applied versus pure, 27–28 euthanasia, 154 evil, 15, 37, 38, 40, 41, 127, 183n14, 188–191 exceptions, 16, 18, 22, 133n42 experience. See anthropology; apriority; categorical imperative; metaphysics of morals feelings, 16, 20–21, 74–75, 79– 81, 166. See also pleasure; respect formulas of the categorical imperative. See categorical imperative formulas freedom. See also autonomy; causality and determinism; evil; spontaneity; thesis of analyticity analysis of concept, 180–181 anthropology and, 29 apriority and, 24–25, 31 argument for freedom of will, 192–193 categorical imperative deduction and, 200–204, 215–217 duty versus, 184–185, 186, 202–204 evil and, 188–191 the good and, 212 humanity as end in itself and, 146–147 as idea, 161n80 intelligence and, 144–145 as law-like causality, 92, 129–130, 187 laws of, 3, 17, 31 laws of nature and, 17, 92, 129–130, 146, 162n85, 186–190 maxims and, 102–103, 129, 162n85 morality and, 5, 177 moral law and, 24, 109, 146, 157, 206–207, 211 moral motivation and, 64n30 negative versus positive, 185–189 ought and, 106n10, 194, 205 presupposition of, 176–183, 191–204 realm of ends and, 162nn84– 85 reason and, 144–145, 186–187, 192–193, 197–198 respect and, 71, 82, 94

232

freedom (continued) spontaneity and, 24, 144–145, 177, 179, 186, 194, 196–198, 206, 216 world of understanding and, 145, 160, 206–209 Gauthier, David, 85 gifts of fortune and of nature, 11n18, 33–41, 44–45, 46, 51, 73 Golden Rule, 138, 155n75 good will, 9, 11, 22, 32–54, 62, 158, 183. See also autonomy; contradiction in conception/will tests; duty; respect; virtue; will; wishing Groundwork for the Metaphysics of Morals (Kant) history of, 1–3 structure and argument of, 32–34, 95– 97, 175–179 task, method and transitions of, 1–13, 175–179 happiness. See also inclinations; teleological argument; utilitarianism common rational moral cognition and, 10, 12 conditional worth of, 40–41, 84n61 empirical conception of ethics and, 26 FH and, 152 FUL and, 140n48 gifts of fortune and, 37, 40n13 good will versus, 33 hypothetical imperatives and, 112–115, 172–173 intentional rule consequentialism and, 46–47 intrinsic value of, 35 of others, 81, 86– 87 reason and, 47–50, 82n57 sensible world and, 205 virtue versus, 88n65 Herder, J. G., 2 Herz, Marcus, 3 heteronomy, 25, 95, 96, 97, 187–190, 205. See also causality and determinism hindrances to pure will, 52–54 Hobbes, Thomas, 85 human beings, 53. See also humanity as end in itself humanity as end in itself, 96, 103, 142–149, 151, 164, 173, 209n37, 215. See also categorical imperative formulas, FH and FRE; dignity

Index

human nature, 13, 23, 26, 28, 37, 51–52. See also anthropology; corruption; happiness; inclinations Hume, David, 5, 20n33, 71–72, 122, 178 Hutcheson, F., 20n33 hypothetical imperatives. See also imperatives analyticity of, 116–122 categorical versus, 2, 102, 157 characterized, 107–108, 109–115 freedom and, 186n17 inclinations and, 36n4 means-and-ends relations and, 84 possibility of, 96, 115–122 ideal, 160–161, 163 imperatives, 53, 97–107, 98, 106n10, 112–115, 123, 190. See also categorical imperative; hypothetical imperatives; maxims inclinations. See also dialectic, natural; empirical factors; happiness; interests, subjective acting from, 52, 59– 60, 81, 90 acting from in addition to duty, 59, 64–74, 84 acting in opposition to, 59, 74, 85n63 as appearances/thing in itself, 20n31, 191n23, 214 applied ethics and, 27 characterized, 99–100 common rational moral cognition (esteem versus appreciation) and, 12–13, 58–59, 69–73 conditional worth of, 36n4 contingency and, 51–52 corruption and, 12, 15–16 exceptions and, 18 freedom and, 187, 191n23 indirect, 59, 60– 62, 63, 84 love as, 74–75 maxims and, 77, 100n6 reason and, 49, 52 respect and, 142, 213n38 self-seeking versus, 62– 64 innate ideas, 24 instincts, 48, 49, 82n57, 146, 193 intelligence, 145, 185, 197–200, 216. See also worlds of sense and understanding intentions. See motives and intentions interests, subjective. See also appearances; contractualism (contractarianism);

Index

end-means relation; heteronomy; inclinations; motives and intentions CI and, 108–109, 156–157, 159 dignity and, 86 duty and, 8n13, 83– 84, 91– 92, 137 freedom and, 178, 187 in moral law, 177 moral rules and, 22 pathological versus practical, 85, 157n77 reason and, 79 subjectivist ethical theories and, 17 will and, 211–212 ‘internalism,’ 5n9 I that thinks and I that wills, 198–200 justice, 39n9, 45 Kant, Immanuel, ethical writings, 1–4, 75n44 Anthropology from a Pragmatic Point of View, 38n8, 136n45 Critique of Practical Reason, 2, 80n54, 130n38, 167n93, 194 Critique of Pure Reason, 13, 27, 98n3, 99n4, 103n8, 107n13, 140n48, 162n84, 187n19, 193n25, 198n32, 209n36 Critique of the Power of Judgment, 112n20, 132n40, 145n61, 186n17 Groundwork for the Metaphysics of Morals, 2, 3, 12 Inquiry Concerning the Distinctness of the Principles of Natural Theology and Morality, 2 Lectures on Anthropology, 29n45 Lectures on Moral Philosophy, 16, 88n65 Lectures on Natural Law, 106n10 ‘The Limits of Sensibility and Reason,’ 2–3 Metaphysics of Morals, 3n5 Mrongovius lectures on moral philosophy, 106n10 Prolegomena, 6, 7, 13n20, 129n37, 175n1 Refl ections, 193n25, 197n29 Religion within the Bounds of Mere Reason, 67n34, 103, 122n49, 145n61 Review of Schulz’s Attempt at an Introduction to a Doctrine of Morals, 192–193, 199n33 killing, 140 knowledge, 11, 13, 59, 118 Kraft, Bernd, 3n4

233

Lambert, Johann Heinrich, 2 law-like character, 17, 89– 92, 129–130, 187. See also causality and determinism; universality of moral law laws of nature, 17, 28n44, 30–31, 123, 214. See also categorical imperative formulas, FLN logic, 26, 27 love, practical, 74–76 lying. See promising material versus formal rational principles, 19n29, 84, 143, 171–172 maxims. See also imperatives admiration and, 38n8 of benevolence, 76n44 CI and, 126, 131, 166 empirical knowledge of, 28 ends and, 84, 103 of enlightenment, 25 freedom and, 102–103 inclinations and, 100n6 moral law and, 58, 67n34, 78, 83, 94 objective principles and, 118–119 practical reason and, 118–119 rational beings and, 150, 152, 157–158, 163 realm of ends and, 160, 162, 163–164 as self-imposed rules, 102–103 as subjective principle of volition, 76–77, 104 subjectivity of, 56n23, 77n46, 101, 104 universality and, 23n37, 76–77, 102 universalizability and, 87, 89, 92, 105, 106, 124 universalization of, 129–142, 159–160 metaphysics of morals. See also ethics, applied versus pure anthropology and, 26–29 apriority of, 2–3, 14–25, 151n71 CI formulas and, 96, 97 divisions of, 4– 6, 30 empirical knowledge and, 132 general, 3–4 need for, 12 popu lar moral philosophy versus, 8, 13, 14, 19, 95, 96 Meyer, G. F., 201n35 moral good and morality, 4– 6, 9, 15–17, 34–40, 157n76, 177, 212. See also categorical imperative; common rational moral cognition; evil; good will; moral law; worth

234

moral law. See also categorical imperative; duty; necessity; universality of moral law as analytic a priori, 109 “for the sake of,” 19, 69n37, 84, 89– 92, 94, 178, 200–203 freedom and, 24, 109, 146, 157, 206–207, 211 justification of, 80n54 rational beings and, 82 singular versus plural, 123 moral philosophy, 2–3, 8, 12, 14–17, 27, 166 motivational overdetermination, 66– 68 motives and intentions. See also deontological moral theories; happiness; inclinations; interests, subjective; respect acting from duty and, 61–76, 85n63, 90– 91 to act morally, 205, 211 a priori, 19 CI formulas and, 132n39, 171 gifts of nature and, 37–38 reason as, 20–21 nature, 17, 22, 28n44, 129n37, 134. See also causality and determinism; world of sense necessity. See also apriority; causality and determinism; universality apriority and, 16, 18–19, 24, 30–31 CI and, 18, 108–109 conceptual-analytical method and, 16–19 duty and, 23, 52–53, 55, 56–58, 60, 81, 107 ends and, 143 hypothetical imperatives and, 110, 115 imperatives and, 98, 104–107 interest and, 157n76 moral law and, 17–18, 18n28, 23, 106–107 nonmoral principles of reason and, 18n28 and the ought and, 105, 106–107, 162n84 subjective versus objective, 52–53, 105 as universality, 21–24 will and, 104–105 Nicomachean Ethics (Aristotle), 48 objective principles, 100–101, 104 On Grace and Dignity (Schiller), 68n36 onto-ethical principle (OP), 207–211, 213

Index

the ought CI as synthetic and, 109, 209 CI deduction and, 5– 6, 178 duty and, 23, 63 empirical factors and, 20, 26, 27–28 freedom and, 106n10, 194, 205 hindrances and, 52n22 hypothetical imperatives and, 110–111, 116, 118–121 interests and, 177 necessity and, 105, 106–107, 162n84 pleasure and, 100n6 reason and, 98, 99, 104, 106, 185, 217 sensual-rational beings and, 172, 209, 211–213 will and, 157, 180, 212–213 persons, 34, 39, 88n65, 144–145, 151–155, 190–191, 210. See also humanity as end in itself; rational beings; worlds of sense and understanding philosophical rational moral cognition, 7–10, 30 philosophy, moral, 2–3, 8, 12, 14–17, 27, 166 philosophy, practical, 8n12, 12, 14 pleasure, 28, 69–71, 74–75, 79, 99–100, 213n38. See also feelings popu lar moral philosophy, 8, 13, 14, 19, 95, 96 postulates, 144, 146, 147n65 practical freedom, 103n8, 145, 179, 186, 187 practical judgment, 9, 39n9, 130n38 practical law, 104 practical philosophy, 8n12, 14 practical reason. See also will appearances and, 20, 215 common rational moral cognition and, 13 defi ned, 5 empirical factors and, 16 end-means proposition and, 116, 117n29, 118 feeling and, 75, 79, 94, 156 freedom and, 144–145, 179, 186–187, 192–193, 197–198 happiness and, 47–50 imperatives and, 96–107 inclinations/interests and, 5, 49–50 maxims and, 28 metaphysics of morals and, 8– 9, 12, 14 moral action and, 16–17

Index

moral law and, 177 as motivation, 20–21 natural dialectic and, 12, 15–16 necessity and, 18n28 necessity of moral law and, 23 as normative, 122 the ought and, 98, 99, 104, 106, 185, 217 rational principles and, 19n29 respect and, 21, 71, 79– 80 synthetic use of, 6, 7n11 theoretical, 12, 15, 199–200, 206, 217 unity with theoretical reason, 15, 198–200, 217 will and, 22, 47–50, 104–105, 116–122, 128, 180–182 price, 36n5, 143–144, 161n82, 173, 210n37 principium diiudicationis/executionis, 20–21, 22, 79, 156 promising, 36n5, 133–136, 140–141, 154–155, 159–160 prudence, 37n6, 39n9, 107–108, 112–115 pure rational cognition, 15 rational beings. See also humanity as end in itself; world of understanding (intelligible world) autonomy and, 144–145, 151n71 dignity and, 162n86 end in itself and, 142–156 free will and, 182, 186 in general, 21–22, 23 imperatives and, 106 maxims and, 150, 152, 157–158, 163 moral law and, 82 necessity and, 107 perfect and imperfect, 52–54 purely (perfectly), 179, 183, 184 representation of the law and, 83n59 rationalistic tradition, 5, 15 Rawls, John, 139 realism, moral, 148, 173 realm of ends, 96, 127, 149n68, 161n83. See also categorical imperative formulas, FRE reason. See practical reason; theoretical reason regard, 35n1 respect. See also dignity; duty, acting from; esteem acting from (as motivation), 59, 60n26, 62–78, 79, 85n63, 91, 176 acting from duty and, 55–58, 67, 89, 91, 93– 94

235

autonomy and, 71, 91 as feeling, 21, 60, 69–71, 76n44, 79– 83, 213n38 good will as, 22 moral law and, 33, 56, 85, 171, 213n38 for rational nature as end in itself and, 82– 83, 142–156, 163–164, 174 reason and, 21, 71, 79– 80 as translation of Absicht, 35n1, 38n7 as unexplainable, 176 for will, 83, 105 worth (objective) and, 33, 56, 65, 69, 79– 83, 94 rules, 45–47. See also objective principles Sapiere ande (Horace), 25n41 Schiller, Friedrich, 68nn35–36 sciences, 1, 11. See also anthropology self-love, 20, 63n29, 131–133, 159, 213n38 Shaftesbury, Anthony Ashley Cooper, Earl of, 20n33 Smith, A., 20n33 spontaneity, 24, 144–145, 177, 186, 194, 196–198, 206, 216 suicide, 63, 130, 131–132, 153–154 synthetic method, 6, 7, 16–17, 30, 175n1 synthetic propositions, 18, 110n17, 127. See also categorical imperative (CI), as synthetic talents, neglect of, 136n46 talents of the mind, 36, 36–37, 39n10, 44 teleological argument, 33, 34, 47–50, 82n57 teleological theories, 41 teleology of nature (natural purposiveness), 12n19, 132, 134 temperament, qualities of, 36–40, 44, 66, 70, 83 theoretical philosophy, 16–17, 24, 100 theoretical reason. See also apriority; intelligence; universality of moral law appearances and, 186, 196–198, 209–210, 215, 216 cognition and, 144, 196–197 freedom/spontaneity and, 24, 179, 186–187, 192–193, 197–200, 206, 216 moral law and, 177, 215 necessity and, 18n28 unity with practical reason, 15, 198–200, 217 world of understanding and, 206–211, 216

236

thesis of analyticity, 178–192, 202–206, 211, 216–217 thing in itself, 194–198, 209, 210, 214–215. See also world of understanding (intelligible world) things, 144 thinking-for-oneself (Selbstdenken), 25n41 transitions. See Groundwork for the Metaphysics of Morals (Kant) universal formula. See categorical imperative formulas universality of moral law. See also categorical imperative formulas, FU and FUL; law-like character apriority and, 16, 24 common rational moral cognition and, 9, 10, 13 conceptual-analytical method and, 16–17 consequentialism and, 87– 88 experience and, 28 imperatives of prudence and, 113 laws of nature and, 17, 28n44, 30–31 maxims versus, 23n37, 76–77 morality and, 157n76 necessity as, 21–24 universalizability, 31, 90, 92. See also categorical imperative formulas, FU and FUL utilitarianism, 41–42, 86– 87 value, 36, 87– 89, 142–156. See also dignity; worth value ethics, 42 virtues ancient traditional, 36n4, 39n9, 50–51, 88n65 common rational moral cognition and, 11 doctrine of, 27, 40–41 duties of, 22 good will and, 34, 38–39, 43, 51 lack of, 43

Index

realm of ends and, 162n85 rights and duties of, 6 Schiller on, 68nn35–36 will. See also freedom; good will; imperatives; practical reason activity of, 82– 83 apriority and, 19, 77–78 authentic, 205, 212–213 CI as synthetic a priori and, 109, 182 dignity and, 88, 145 duty and, 107 hindrances to, 52–54 human beings and, 53 motivation and, 157 the ought and, 157, 180, 212–213 perfect, 105, 106n10, 181, 184 practical reason and, 22, 98–103, 119, 122 pure, 19, 22, 99, 104–105, 180–181, 184, 211–212 reason and, 22, 47–50, 104–105, 116–122, 128, 180–182 respect for, 83, 105 universal laws and, 17, 145n62 as universally legislative, 23, 83, 91– 92, 105, 145, 157–160, 207 wishing, 43, 58n25 Wolff, Christian, 14n22 world of sense, 188, 190. See also appearances; heteronomy world of understanding (intelligible world), 53, 105, 151n71, 160–161, 193. See also rational beings worlds of sense and understanding CI as synthetic a priori and, 109n16, 180–181 CI deduction and, 194–198, 202–215 freedom and, 179, 184–186 worth, 35–38, 83– 89, 92– 94, 142–149, 215. See also dignity; duty, acting from; humanity as end in itself; value Xenia (Schiller), 68n35