Groundwork for the Metaphysics of Morals 9783110204551, 9783110177077

Citation preview

Groundwork for the Metaphysics of Morals

≥

Groundwork for the Metaphysics of Morals Edited by Christoph Horn and Dieter Schönecker in cooperation with Corinna Mieth

Walter de Gruyter · Berlin · New York

" Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

ISBN-13: 978- 3-11-017707-7 ISBN-10: 3-11-017707-2 Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at "http://dnb.ddb.de#. ” Copyright 2006 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Printed in Germany Typesetting: OLD Media, Neckarsteinach Printing and binding: Druckhaus “Thomas Müntzer”, Bad Langensalza Cover design: Martin Zech, Bremen

Contents Christoph Horn and Dieter Schönecker Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii

Notes on Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Groundwork, Preface Nico Scarano (Tübingen) Necessity and Apriority in Kant’s Moral Philosophy: An Interpretation of the Groundwork’s Preface (GMS, 387–392) . . . . .

3

Groundwork I Allen Wood (Stanford) The Good Without Limitation (GMS I, 393–394). . . . . . . . . . . .

25

Christoph Horn (Bonn) Kant on Ends in Nature and in Human Agency: The Teleological Argument (GMS I, 394–396) . . . . . . . . . . . . . .

45

Marcia Baron (Indiana) Acting from Duty (GMS I, 397–401) . . . . . . . . . . . . . . . . . . . . . .

72

Harald Köhl (Berlin) The Derivation of the Moral Law (GMS I, 402–403) . . . . . . . . .

93

Groundwork II Marcus Willaschek (Frankfurt a. M.) Practical Reason. A commentary on Kant’s Groundwork of the Metaphysics of Morals (GMS II, 412–417) . . . . . . . . . . . . . .

121

Bernd Ludwig (Göttingen) Kant’s Hypothetical Imperatives (GMS II, 417–419) . . . . . . . . .

139

vi

Contents

Mark Timmons (Memphis) The Categorical Imperative and Universalizability (GMS II, 421–424). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

158

Samuel Kerstein (Maryland) Deriving the Formula of Humanity (GMS II, 427–437) . . . . . .

200

Groundwork III Klaus Steigleder (Bochum) The Analytic Relationship of Freedom and Morality (GMS III, 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

225

Corinna Mieth and Jacob Rosenthal (Bonn) “Freedom must be presupposed as a property of the will of all rational beings” (GMS III, 2). . . . . . . . . . . . . . . . . . . . . . . . . . . . .

247

Marcel Quarfood (Stockholm) The Circle and the Two Standpoints (GMS III, 3) . . . . . . . . . . .

285

Dieter Schönecker (Siegen) How is a categorical imperative possible? Kant’s deduction of the categorical imperative (GMS III,4) . . . . . . . . . . . . . . . . . . . .

301

A Bibliography on Kant’s Groundwork . . . . . . . . . . . . . . . . . . . .

325

General Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

337

Preface Kant’s Groundwork for the Metaphysics of Morals is one of the most influential texts in the entire history of moral philosophy. It offers the first and presumably the most attractive account in Kant’s writings of an ethics which can be called ‘Kantian’ in the sense in which we are accustomed to understand this classification. In earlier stages of his development, Kant sympathized with variations of a Leibniz-Wolffian perfectionism or with a Hutcheson-Humean moral sense-philosophy. In the Groundwork, however, Kant develops his own characteristic position. He now emphasizes that an adequate form of moral philosophy has to be ‘pure’, i. e. both free from all empirical elements of interest, self-love, and natural feelings as well as free from rational concepts of perfection. More generally speaking, ethics must not be grounded on anthropology, since morality is a demand, as Kant contends, which is addressed towards all rational beings as rational beings. According to Kant, ethics has to be spelled out on the basis of a ‘moral law’ that is valid for all finite rational beings. He believes that only from this point of view can moral motivation and moral obligation be formulated in an appropriate way. It is the Groundwork in which Kant develops and discusses his doctrine of the categorical imperative and where he attempts to give a ‘deduction’ for the universal validity of the moral law. Published in 1785, the Groundwork presents a type of moral philosophy that has continously inspired and provoked its readers ever since. At first sight, Kant’s Groundwork for the Metaphysics of Morals seems to be a clearly structured, well argued, and easily accessible text; not without reason it has often been used to introduce beginners to Kant’s moral philosophy. On close reading, however, the text really turns out to be most difficult and somewhat underdetermined. The Groundwork is a highly condensed treatise containing many of the topics and ideas Kant treated in his university lectures and his notes since the 1760s. Thus, interpreters face many hard questions, e. g.: Kant begins with the doctrine of the goodness of the good will; however, he never says what a good will actually is, rather, he just describes what it is not – so what is it? Why does he introduce tele-

viii

Preface

ological considerations to support his idea of a good will? Does Kant allow inclinations to somehow accompany or even support the feeling of respect, or does he not? In what precise way does he derive the categorical imperative? What kind of procedure does the categorical imperative provide to test our maxims? How are the different formulas of the categorical imperative correlated? As for the third section, it has often been described as rather dark or elusive and hardly accessible to a meaningful reconstruction. These are difficult questions and problems of interpretative exegesis. The other feature which makes the Groundwork demanding is that it raises several far-reaching claims, often insufficiently corroborated, that lead to the most fundamental questions and problems of moral philosophy such as: Can pure reason really be practical? Is Kant’s doctrine of the categorical imperative in one of its formulas convincing? Does universalization in one way or another work at all? Does Kant really provide an argument for the universal validity of a moral law? Does he really provide an argument against consequentialism? Can we make sense of his idea of autonomy? In recent years, there have been considerable achievements in researching Kant’s Groundwork and concerning Kantian moral philosophy in general. An impressive number of articles and monographs have been published, many if not most of them devoted not only to historical aspects, but also to a systematic assessment of Kant’s contribution to ethics. The articles published here are intended to summarize and to continue the work done within contemporary reseach on the Groundwork. They are the contributions delivered at a conference which took place at the University of Bonn from July 19th to 22th 2004. This conference was generously supported by the German National Research Association (DFG) and the Rheinische Friedrich-WilhelmsUniversität of Bonn. We’d like to thank our colleagues who attended the conference and contributed to lively discussions and then later to this volume. Special thanks go to Corinna Mieth for her cooperation in all questions regarding the conference and the present volume. For additional work on the manuscripts we are grateful for the support of Alexander Cotter, Nadine Dietzler, Michael Helwig, Nadine Köhne, Richard Capobianco, and Simon Weber. Bonn and Siegen, April 2006 Christoph Horn and Dieter Schönecker

Notes on Contributors Marcia Baron is Rudy Professor of Philosophy at Indiana University (Bloomington, Indiana). She is the author of Kantian Ethics Almost without Apology (1995), and with Philip Pettit and Michael Slote, Three Methods of Ethics (1997), as well as articles on a variety of topics in ethics and philosophy of law, including friendship and impartiality, patriotism, justifications and excuses, sexual consent, and manipulativeness. Christoph Horn is Professor of Philosophy at the University of Bonn. He works on various aspects of ancient philosophy (Plato, Aristotle, Plotinus, Augustine) as well as modern moral philosophy. Publications: Plotin über Sein, Zahl und Einheit (1995), Augustinus (1995), Antike Lebenskunst (1998), Politische Philosophie (2003), Grundlegende Güter (forthcoming). He edited the following volumes: Augustinus, De civitate dei (1997), (with Ch. Rapp) Wörterbuch der antiken Philosophie (2002), (with N. Scarano) Philosophie der Gerechtigkeit. Texte von der Antike bis zur Gegenwart (2002). Samuel Kerstein is Associate Professor of Philosophy at the University of Maryland, College Park. He is the author of Kant’s Search for the Supreme Principle of Morality (2002) and is working on a reconstruction and defense of Kant’s Formula of Humanity. Harald Köhl is Lecturer at the University of Saarbrücken. He was Guest Professor of Philosophy at the Humboldt University (Berlin). His most important publications are: Kants Gesinnungsethik (1990), Abschied vom Unbedingten. Über den heterogenen Charakter moralischer Forderungen (2006). Papers on practical philosophy (Kant, Schopenhauer, Williams, Rorty), theory of knowledge and semantics. Bernd Ludwig is Professor of Philosophy at the University of Göttingen. His main research areas are the philosophy of I. Kant, moral philosophy and the theory of Natural Law in the early modern period. He is editor of Kant’s Metaphysics of Morals. Most important publica-

x

Notes on Contributors

tions: Kants Rechtslehre (22005), Die Wiederentdeckung des Epikureischen Naturrechts [on Th. Hobbes] (1998). Corinna Mieth (MA 1999 and PhD 2002 at the University of Tübingen) teaches ethics and political philosophy at the University of Bonn. Her research interests are focussed on duties of assistance, on the foundation of human rights, especially social rights, on the political philosophy of John Rawls and theories of global justice. She is coauthor of a commented edition of Kant’s Groundwork with Christoph Horn and Nico Scarano. Marcel Quarfood is Senior Lecturer in Philosophy at Södertörn University College, Huddinge, Sweden. He translated Kant’s Prolegomena into Swedish, and published Transcendental Idealism and the Organism. Essays on Kant (2004), as well as articles on Kant and on the philosophy of biology. Jacob Rosenthal graduated with a Diploma in Mathematics from the University of Würzburg. Afterwards he studied philosophy at the University of Konstanz and received a PhD in 2002 with a dissertation on objective interpretations of probability (published as Wahrscheinlichkeiten als Tendenzen (2004)). Currently he teaches ethics and philosophy of science at the University of Bonn. His research interests lie in the fields of epistemology, general philosophy of science, theory of action and moral philosophy. Nico Scarano is Assistant Professor at the University of Tübingen. He works on metaethics, moral philosophy and political philosophy. His dissertation Moralische Überzeugungen. Grundlinien einer antirealistischen Theorie der Moral has been published in 2001. He edited the following books: (with Ch. Horn) Philosophie der Gerechtigkeit. Texte von der Antike bis zur Gegenwart (2002), Modelle politischer Philosophie (2003) and Ernst Tugendhats Ethik. Einwände und Erwiderungen (2006). Dieter Schönecker is Professor of Moral Philosophy at the University of Siegen. His main research areas are Kant’s ethics, metaethics, medical ethics, hermeneutics, epistemology. Publications: (with Gregor Damschen) Selbst philosophieren (2006), (in cooperation with Stefanie Buchenau and Desmond Hogan) Kants Begriff transzendentaler und praktischer Freiheit. Eine entwicklungsgeschichtliche Studie (2005),(with Allen W. Wood) Immanuel Kant: Grundlegung zur Metaphysik der Sitten. Ein einführender Kommentar (22004),

Notes on Contributors

xi

Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs (1999). Co-editor of: (with Heiner F. Klemme and Manfred Kühn) Moralische Motivation. Kant und die Alternativen (2006), Internationales Jahrbuch des Deutschen Idealismus / International Yearbook of German Idealism (2003–2005), (with Gregor Damschen) Der moralische Status menschlicher Embryonen. Argumente pro und contra Spezies-, Kontinuums-, Identitäts- und Potentialitätsargument (2003), (with Niko Strobach) Einführungen Philosophie, 15 vols. (2002 ff.), (with Thomas Zwenger) Kant verstehen/Understanding Kant. Über die Interpretation philosophischer Texte (22004) and Immanuel Kant: Grundlegung zur Metaphysik der Sitten. Neu herausgegeben und eingeleitet von Bernd Kraft und Dieter Schönecker (1999). Numerous articles and reviews in journals and dictionaries. Klaus Steigleder is Professor of Philosophy at the University of Bochum, Germany. His publications include: Kants Moralphilosophie [Kant’s moral philosophy] (2002) and Grundlegung der normativen Ethik [The Foundation of Normative Ethics] (1999). He is co-editor of: Medizinethik. Geschichte, Theorie und Ethik der Medizin [Medical Ethics. History, Theory and Ethics of Medicine] (forthcoming), Die Aktualität der Philosophie Kants [The Timelessness of Kant’s Philosophy] (2005), Bioethik. Eine Einführung [An Introduction into Bioethics] (2003). Mark Timmons is Professor of Philosophy at the University of Arizona and author of Morality without Foundations (1999) and Moral Theory: An Introduction (2002) and editor of Kant’s Metaphysics of Morals: Interpretative Essays (2002). He and Terry Horgan edited Metaethics after Moore (2006) and they are currently working on philosophical issues associated with the phenomenology of moral experience. Marcus Willaschek is Professor of Philosophy at the University of Frankfurt/M. His publications include: Praktische Vernunft. Handlungstheorie und Moralbegründung bei Kant, Stuttgart/Weimar 1992; Der mentale Zugang zur Welt. Realismus, Skeptizismus und Intentionalität, Frankfurt/M. 2003; Kant: Kritik der reinen Vernunft, Berlin 1998 (ed. with Georg Mohr); Hilary Putnam und die Tradition des Pragmatismus, Frankfurt/M. 2002 (ed. with M.-L. Raters); numerous articles in journals and essay-collections on Kant, ethics philosophy of action, metaphysics and epistemology, the philosophy of mind and language.

xii

Notes on Contributors

Allen W. Wood is Ward W. and Priscilla B Woods Professor at Stanford University. He has also taught at Cornell University and Yale University, and has held visiting appointments at the University of Michigan, University of California at San Diego, and Oxford University. He is author of nine books, most recently Kant (2005) and is general editor (with Paul Guyer) of the Cambridge Edition of the Writings of Immanuel Kant, as part of which Guyer and Wood collaborated on a new translation of the Critique of Pure Reason. Wood has also translated Kant’s Groundwork (2002).

Groundwork Preface

Nico Scarano

Necessity and Apriority in Kant’s Moral Philosophy An Interpretation of the Groundwork’s Preface (GMS, 387–392) Kant’s Preface to the Groundwork for the Metaphysics of Morals can be divided into three parts. First of all, the text begins with a topology of the philosophical disciplines, in which the project of a metaphysics of morals is delimited from other forms of theory (GMS, 387,1–388,14).1 Secondly, it contains a compactly reasoned appeal for the indispensability of such a pure moral philosophy, that is, a theory that proceeds completely a priori. Simultaneously, Kant emphasizes the uniqueness of this project (GMS, 388,15–391,14). And thirdly, the Preface offers a prospective of the task, the structure, and the method of the Groundwork, which does not itself already contain such a metaphysics of morals, but, as Kant stresses, represents a “preliminary work,” laying the foundation upon which the comprehensive theory will be developed “in the future” (GMS, 391,16–392). In the following I will go into all three parts of the Preface. The focus of my analysis, however, concerns Kant’s central argument for the indispensability of a pure moral philosophy. At decisive points, Kant clearly inserts modal considerations. For the reconstruction of the argument, it is therefore important to define precisely which form of necessity is intended. Kant uses the terms necessary and necessity in quite disparate contexts and with different meanings. He nowhere explicitly says what necessity in the domain of morality can even mean. For that reason, before I move onto the analysis of Kant’s argument, I will address more closely the logical structure of ethical principles. One must first understand the logical form that such a principle exhibits in order to be able to work out the basis of Kant’s central argument 1

This paper was translated by Aaron Looney.

4

Nico Scarano

for a pure moral philosophy. It will become evident that Kant’s argument is comprehensible if the decisive concept necessity is understood as a modal operator in the sense of modern logic. The modal status, necessity, allows moral principles to guide our counterfactual, practical reflections. This aspect of logical form can be clearly distinguished from epistemological connotations, on the one hand, and the prescriptive or imperative character of normative propositions, on the other. 1. What is a Metaphysics of Morals? (GMS, 387,1–388,14) Kant does not introduce the Preface of the Groundwork with a characterization of the work’s content; rather he attempts first of all to define the place of a metaphysics of morals within philosophy. For this task, he makes use of three criteria. Kant first differentiates philosophical theories by whether they are “formal” or “material.” Formal philosophy, according to Kant, is equated with logic. It possesses no specific object; rather it concerns itself, “without distinction among objects,” with “the universal rules of thinking in general” (GMS, 387,10 f.). In contrast, every material philosophy “has to do with determinate objects and the laws to which they are subjected” (GMS, 387,12 f.). This formulation already offers an indication of the second criterion. Kant subdivides material theories, in turn, into two classes. Kant distinguishes them by with reference to the laws to which the objects that the theories deal are subjected. He seems to assume that there are exactly two kinds of laws. And, correspondingly, he differentiates between two types of material philosophy: on the one hand, “physics,” or “doctrine of nature,” or, alternatively, “natural wisdom;” and, on the other hand, “ethics,” or “doctrine of morals,” or, alternatively, “moral wisdom.” It is a matter of the “laws of nature,” in the one case, and of the “laws of freedom,” in the other, that each theory is respectively concerned (GMS, 387,14 f.). What can Kant mean by this? The expression “laws of nature” seems to be relatively unproblematic. But what is to be understood by the expression “laws of freedom?” From Kant’s elucidation one can infer a more exact interpretation. Laws of nature are therefore laws “in accordance with which everything happens,” while the laws of freedom are those “in accordance with which everything ought to happen” (GMS, 387,25–388,1).2 Obviously, one can draw the parallel here to the modern terminologi2

See also the parallel passage in the Critique of Pure Reason, A 802/B 831.

Necessity and Apriority in Kant’s Moral Philosophy

5

cal distinction between descriptive and normative statements. At the center of a philosophical doctrine of nature, there would then stand general descriptive judgments; that means, more precisely, statements of law that describe how objects of nature act. And at the center of a doctrine of morals there would stand normative judgments that exhibit a comparable degree of generality and modal status. I will later address more thoroughly what it could mean that some normative judgments exhibit a comparable modal status to statements of law in the natural sciences. Such a parallel is far from trivial. On the contrary, it is one of the critical challenges for the interpretation of Kant’s ethical writings. 3 The first step toward a unified interpretation of the two kinds of laws consists in seeing both kinds equally as propositionally structured entities that demonstrate a clearly identifiable logical form (see 2.1 through 2.3 below).4 Next to formality versus materiality and the two types of laws, Kant introduces a third distinctive characteristic of theories. He seems to understand material theories as complex systems that can be split into individual parts. Every material philosophical theory accordingly contains one part that he characterizes as “pure,” “rational,” or as “a priori” (GMS, 388,4–14), and it may contain a second, “empirical” part. Kant presupposes that in both the doctrine of nature and the doctrine of morals the two parts can be isolated from each other. It follows, therefore, that there are two a priori types of theory: on the one hand, a “metaphysics of nature,” and, on the other, the soughtafter “metaphysics of morals” (GMS, 388,10). Thus, a metaphysics of morals would be the following type of theory: (a) It is not purely formal, but rather deals with definite objects. (b) It deals with objects in so far as these are subsumed under “laws of freedom.” (c) It contains no empirical elements. At this juncture Kant does not yet presuppose that such a theory actually exists or that it would be possible for a philosophy to work out convincingly such a theory. He has not even argued yet that it is advantageous or even necessary for philosophy to treat the a priori part separately from the empirical part. Kant pursues these lines of questions in the subsequent passages. 3

4

Thus Tugendhat (1993, 99 f.), for example, doubts whether the use of the term “necessity” in Kant’s theoretical philosophy and in his practical writings can be understood in a unified manner. For the concept of law in Kant’s ethics see also Klotz (2001). The concept of necessity as a modal quality that stands at the center of this essay is mentioned by him without, however, further explication (58).

6

Nico Scarano

2. The Indispensability of a Pure Moral Philosophy (GMS, 388,15–391,15) The Kantian appeal for a pure moral philosophy begins with a reference to the usefulness of the division of labor. Kant raises the question whether philosophy would not benefit from such a division. Ultimately, he is not concerned with whether the burden of elaborating philosophical theories should be distributed among the shoulders of a multitude of persons. Rather with this rhetorical consideration, he introduces an explicitly methodological train of thought. He asks “whether the nature of the science does not require the empirical part always to be carefully separated from the rational,” (GMS, 388,33 f.) and he confines this question to the sub-question “whether one is not of the opinion 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” (GMS, 389,6–9). The passage (GMS, 389,5–23) that is introduced by this question contains, at the same time, a positive response – the thesis that there must be a “pure moral philosophy” is “self evident.” And indeed for Kant this evidence arises “out of the common idea of duty and the moral law” (ibid.). Kant’s cursory argumentation for this thesis is difficult to grasp. Undoubtedly, it constitutes one of the key argumentative passages of the Groundwork. How can the argument be reconstructed? The point of departure for Kant’s line of argumentation lies in the “common idea of duty and the moral law.” Kant seems to assume that we (“everybody:” GMS, 389,11) may convince ourselves of the correctness of his thesis by reflecting upon what we understand by morality. What, however, is contained in our concept of morality? And with which elements of the concept is he concerned in his argumentation? In this context he makes use of an example. The moral command “you ought not to lie” refers to a law that, according to Kant, is valid “not merely for human beings” but also for “other rational beings.” That moral laws have validity for all rational beings, therefore, is the first step of argumentation towards a grounding of the thesis that there must be a pure moral philosophy. But what does Kant mean by his statement that a moral law is valid for all rational beings? And how can the argument for it be reconstructed, that is, be made as comprehensible as possible in the terms of today’s common philosophical concepts? Kant speaks in this context of an “absolute necessity” of these laws. From this necessity, it

Necessity and Apriority in Kant’s Moral Philosophy

7

follows that these laws also must be valid for other rational beings. And, in a further step, Kant concludes the apriority of moral philosophy from this necessity. The decisive term for Kant’s argumentation, necessity, especially requires interpretation. In order to obtain more exact concepts of interpretation, I would like to examine more closely in the following section the logical structure of ethical principles. 2.1. The Logical Structure of Ethical Principles At the center of every ethics are normative statements. Ethics of principle, like Kant’s, presuppose that there are ethical principles that can be formulated with the aid of normative or evaluative statements, which have a high degree of generality and therefore allow the moral judgment of a plurality of cases. Such principles have the goal of guiding our moral judgments and therewith our actions. They formulate criteria by which we can morally judge objects, for example actions or institutions. Ethical principles can have different logical strengths. They offer either a necessary or a sufficient or a both necessary and sufficient criterion for determining when an object receives a corresponding moral predicate. The categorical imperative, as defined by Kant, or, more precisely, the founding law of the categorical imperative, appears to be a quite demanding principle, one which prescribes both a necessary and a sufficient condition for morality. Kant compares this principle with a “compass in the hand” by which one “knows [his] way around very well in all the cases that come before [him], how to distinguish what is good, what is evil, what conforms to duty or is contrary to duty” (GMS, 404,1–3). Normative propositions in which principles with such a logical rigor come to expression exhibit, by means of a first approach, the following logical form: (P1) For all objects x: if, and only if, x satisfies the criterion C, does x have the moral quality M.

Such propositions contain criteria for determining when an object receives a certain moral predicate or when it is denied it. In P1 it is a matter of the conditions for the assignation of the moral predicate M. According to the all-quantified biconditional, every object that satisfies the condition C simultaneously receives the moral predicate, and vice versa. General propositions that exhibit such a form consequently

8

Nico Scarano

assist us in morally evaluating individual cases. If such principles are known to us, then we would have an extremely strong means of coming to correct valuations of our orientations in action. At the heart of Kantian ethics is the moral valuation of the objects of an entire particular class. The criterion of valuation, formulated by Kant in the categorical imperative, enables first and foremost the judgment of maxims. At the same time, the valuation of maxims leads to the valuations of actions. The normative predicate which is of ultimate concern in the valuations of actions is “morality.”5 And the criterion to which actions have to conform so that this moral quality may be attributed to them is expressed in the categorical imperative. For this criterion, I will use in the following passages the abbreviation “CI.” The fundamental principle of Kantian ethics, therefore, has approximately the following form: (P2) For all actions x: if, and only if, x satisfies the criterion CI, does x satisfy the demand of morality.

Without going into the difficult questions of interpretation here, the criterion CI can be articulated with slightly greater precision. According to Kant, the quality of the maxim from which an action results is decisive for the moral valuation of an action. For Kant, a concrete maxim for action appears to correspond to the demands of the categorical imperative only if the actor can will at the same time that her maxims become a general law. Regardless of the familiar difficulties in applying this criterion and regardless of the question about the equivalence of the three formulas depicted by Kant, the following formulation could be a preliminary rendering of the fundamental principle of his ethics: (P3) For all maxims x and all actions y: if, and only if, the underlying maxim x of the action y has the quality that the actor of y can will at the same time that x becomes a general law, does y satisfy the demand of morality.

I assume that we are dealing here with an excellent case of the “law of freedom” of which Kant speaks in the Preface. However, which as5

I interpret the categorical imperative as a test not only for the legality but also for the morality of an action. Such an interpretation is not uncontested in the contemporary literature on Kant (s., for example, Köhl, 1990, 66 f. and Wood in this volume); it does not, however, play a decisive role in my further reflections. The term “morality” in (P2) would, accordingly, simply have to be replaced by the term “legality.”

Necessity and Apriority in Kant’s Moral Philosophy

9

pect of the logical form refers to the necessity of the law is not, as yet, explicitly worked out. 2.2 The Modal Status of an Ethical Principle Ethical principles have the task of guiding our moral judgments. A more exact analysis of their logical features takes into account that with their help we judge not only factually existing objects but also objects which we merely imagine or whose existence is possible but not actual. Let us assume in a particular situation that we have two alternatives of action at our disposal. We must decide between action A and action B. If we choose A, B will never exist; if we choose B, A will likewise never exist. If we wish to judge both options of action from a moral point of view, we can do so with recourse to our ethical principles. In that case, we test whether the criteria formulated in the principles will be satisfied or not satisfied. The criteria must be applicable both to factually existing objects as well as to merely possibly existing ones in order to guide our valuation. Both the actual action A and the merely possible action B fall within the range of objects covered by the principles. This condition has consequences for their modal status. In order to express modal relations, one can make use of the possible-worlds terminology common in semantics. In case we decide for A and successfully translate this decision into action, then A is an object of the actual world, and B is solely an object of a possible world. We judge both objects with recourse to the same principles. That means, however, that these principles express something not only about the actual world but also about other, possible worlds. Since these principles indicate a strength beyond that of contingency, it appears that they exhibit the modal status of necessity. In the preliminary formulations (P1) through (P3), this logical quality has yet to be expressed. In (P1') the modal status of necessity, therefore, is explicitly taken up into the formulation: (P1') Necessarily, for all objects x: if, and only if, x satisfies the criterion C, does x have the moral quality M.

Applied to Kantian ethics, this thought results in the following formulation of Kant’s fundamental principle: (P2') Necessarily, for all actions x: if, and only if, x satisfies the criterion CI, does x satisfy the demand of morality.

10

Nico Scarano

The proposition (P3) is also to be completed accordingly: (P3') Necessarily, for all maxims x and all actions y: if, and only if, the underlying maxim x of the action y has the quality that the actor of y can will at the same time that x becomes a general law, does y satisfy the demand of morality.

The fact that ethical principles carry in themselves the modal status of necessity provides an important methodical starting point for the construction of ethical theories. Thus, the execution of the commonly performed counterfactual thought experiments in ethics is only possible under the condition that the principles of valuation support such an operation. Contingent principles are not in a position to do that. Kant makes methodical use of this logical quality of moral principles particularly in those places where he speaks of other “rational beings” (e. g. GMS, 389, 401, 408, 412, 415). The central passage in the Preface of the Groundwork in which he argues for the necessity of a pure moral philosophy also belongs to this methodical usage. It does not matter in these passages whether there actually are other such beings as, for example, the inhabitants of other planets or also God. For the Kantian argumentation it is sufficient that such beings could exist, that their existence is thinkable. The corresponding passages must be understood as methodically executed thought experiments, which make use of the particular modal status of ethical principles. Now it becomes clear why the “laws of freedom” have a comparable modal status to the “laws of nature.” Laws of nature support counterfactual arguments, too. In order to achieve this, they also must have a modal status which is higher than simple contingency. The connections formulated in them are also valid in all natural law governed, possible worlds, and in this respect, they exhibit the modal status of necessity. The difference between laws of nature and laws of freedom appears to consist primarily in the fact that the laws of nature are concerned with all-quantified, descriptive biconditionals, while the laws of freedom are concerned with all-quantified, normative biconditionals, each receiving the modal status of necessity.6 6

Actually, the type of necessity spoken of here has to be further specified. Is it a matter of “logical,” “conceptual,” “nomological,” or “metaphysical” possible worlds? In Scarano (2001, chapter 3.2), I argued that our moral principles have a comparable status to the “metaphysical necessity” introduced by Saul Kripke (1980). To Kant has to be ascribed the view that it is herewith a matter of “conceptual necessity.” I see an indication of this interpretation in the method he applies in the fi rst and second sections. He presupposes that the content or the formula of the categorical imperative can be found solely through the means of the conceptual analysis of our

Necessity and Apriority in Kant’s Moral Philosophy

11

2.3. Necessity, Normativity, and Apriority A possible but easily avoidable equivocation in the expression necessity can be cleared up at this juncture. Sometimes the expression is used in the realm of morality as a synonym of normativity or prescriptivity. Consequently, actions are necessary if they connote a “should” or if it is our duty to carry them out. This type of usage is also found in Kant. In the central “third proposition” of the first part of the Groundwork, this usage is clearly expressed: “Duty is the necessity of an action from respect for the law” (GMS, 400,18 f.). While the necessity analyzed previously refers to the moral principles, it is here a matter of the necessity of the action itself. However this aspect is terminologically classified, whether as “normative,” “prescriptive,” “evaluative,” or whether one speaks of the imperative character of moral judgments, it may be distinctly distinguished from the modal-logical concept of necessity responsible for the counterfactual variations. In the propositions (P1') through (P3') this aspect – that is, necessity in the sense of “normative,” “prescriptive,” or “evaluative” – is indeed contained. There it is connected, however, to the moral predicate, not the operator of necessity. When I speak of necessity in the following sections, I mean a modal quality of judgments and not the specificum of normativity. Kant uses the expression necessity with yet other meanings. Every interpretation depends on the clarification in each particular context of what Kant exactly intends in those corresponding places and of how each particular argument is to be reconstructed. Next to (a) the type of usage as a modal operator that makes counterfactual considerations possible and (b) the usage in the sense of an imperative character, thus in the sense of “normativity” or “prescriptivity,” there is (c) moral concepts. At the beginning of the decisive argumentation in the second section, he writes, “Regarding this problem we will first try to see whether perhaps the mere concept of a categorical imperative does not also provide us with its formula” (GMS, 420,18–20). And approximately twenty pages later, he asserts in retrospect: “Yet that the specified principle of autonomy is the sole principle of morals may well be established through the mere analysis of the concepts of morality” (GMS, 440,28–30). He, therefore, assumes that he actually was able to extract the formula of the supreme moral principle solely through a conceptual analysis. In my opinion, such a proceeding allows only one conclusion: If the moral principle can be produced solely through an analytical procedure on our concept of morality, then it would have the status of conceptual necessity. According to Kant, the founding law of the categorical imperative is valid in all conceptually possible worlds. The question of which type of necessity moral principles exhibit, however, is not essential for the ensuing reflections.

12

Nico Scarano

an often encountered usage with an epistemological meaning. If the expression is used in this sense, then it means as much as “necessary know-ability,” that is, the independence of knowledge from contingent, empirical factors. Typically, Kant uses the expression a priori for judgments that exhibit this characteristic. Since this type of usage also can be clearly distinguished from the modal one, I will speak of apriority to designate this epistemological aspect. I will use necessity solely in the first sense (a). Kant doubtlessly sees a close connection between necessity and apriority. He often moves quickly from the one concept to the other without grounding the transition. The two concepts, however, originate from varying spheres. While necessity is a matter of the modal status of judgments, apriority is an epistemological concept. In the former case, the concern is the application of predicates to objects of other possible worlds. In the later case, it is a matter of the know-ability of the relevant judgments. Between the two concepts there does not seem to be a close conceptual connection. In particular cases it must be explicitly argued for that apriority follows from necessity.7 Even if Kant sees a very close connection between the two concepts, he does not appear to assume that necessity and apriority are exchangeable concepts. In the central passage of the Preface, he formulates rather an argument for their connection. In the following section, I will sketch out a possible reconstruction of the argumentation’s structure on the basis of the conceptual differences just worked out. 2.4. The Central Argument in the Preface (GMS, 389,5–35) In the central passage of the Preface on the indispensability of a pure moral philosophy, Kant does not simply presuppose that from the “absolute necessity” (GMS, 389,13) of moral laws the apriority of moral philosophy follows. Rather he delivers a detailed, even if difficult, line of argumentation. First of all, he announces his aim of argumentation and designates his point of departure: “that it is of the utmost necessity to work out once a pure moral philosophy” or “that there must be such is self-evident from the common idea of duty and of moral laws” (GMS, 389,7–11). The aim of his argumentation, therefore, is the proof of the apriority of moral philosophy. And his point of departure for the argumentation lies in the analysis of our concept of duty or of our 7

Compare this point with Kripke (1980).

Necessity and Apriority in Kant’s Moral Philosophy

13

“idea” of moral laws. How can the individual steps of this very dense argumentation be distinguished from each other and ordered? According to Kant, the starting point, a reflection on our moral concepts, leads immediately to the first intermediate thesis, that (1) a moral law “has to carry absolute necessity with it” (GMS, 389,13). If this necessity is very narrowly interpreted, that is, in the sense of the previously worked out modal status of an ethical principle, then two peculiarities of the total argumentation will become more understandable. First of all, it will become clear that in the passage of the text an argumentative progression takes place from the given-ness of a modal quality to the proof of an epistemological quality distinguished from it. And secondly, an argument not implausible even from a contemporary viewpoint comes to light for the intermediate thesis (1). For, as shown above, it is the task of an ethical principle to guide our valuations even in the case of counterfactual considerations. If this is correct, then an analysis of our concept of morality can help bring to light the modal status that is responsible for the capacity of meeting this task. The argumentative progression from the point of departure to thesis (1), therefore, appears well-motivated. Through the analysis of our concept of morality we find that moral laws also apply to counterfactual situations, that they consequently exhibit the modal status of necessity. Kant distinguishes at this point between moral laws and moral duties. From a moral law (“the ground of an obligation”: GMS, 389,12) arises a moral duty (“obligation”) to which our actions have to conform. As an example of an obligation, Kant names “the command ‘You ought not to lie’” (GMS, 389,13 f.). It is interesting to observe that Kant ascribes the decisive “absolute necessity” not to the duty, but rather to the law that is foundational to the duty. This, too, serves as an indication that the term necessity is to be understood as a modal expression and not in the sense of “normativity” or “prescriptivity.” Otherwise Kant would speak of the necessity of an action instead of the necessity of the law foundational for obligation. The point of departure, “absolute necessity,” may be read therefore as a modal status of moral principles. How, then, are the other intermediate steps to be understood? The basis in the text is extremely narrow. One possibility of outlining the continued line of argumentation is as follows: In the next step, Kant seems to refer to a counterfactual thought experiment. Among the counterfactual situations to which duties are attributed are those in which not humans but other imaginable ration-

14

Nico Scarano

al beings have the possibility of acting. In such a thought experiment, we imagine a world in which there are other beings capable of action. Since the moral laws accepted by us first of all contain the modal status of necessity – that is, they apply to all possible worlds – and secondly, involve a proposition about all actions, it is to be concluded that even those actions of these merely imagined rational beings fall under these laws. The second intermediate thesis, therefore, states that (2) the duties arising from the moral laws are also relevant for the action of these merely imagined rational beings. In Kant’s terms: “Everyone must admit 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” (GMS, 389,11–15). At this juncture the decisive transition from necessity to apriority takes place. Since the applicability of obligation is valid for all imaginable rational beings, the cognition of these obligations or the laws that are responsible for them cannot depend on contingent features that arise from the fact that the thought experiment involves different rational beings. The laws, therefore, may solely depend on characteristics that all rational beings capable of action have in common. What all imaginable rational beings have in common, however, is solely the characteristic of being rational beings. Every other characteristic is contingent. The cognition of laws cannot depend, therefore, upon the empirical cognition of these contingent features. Kant seems to assume that a cognition contains the characteristic of apriority if it is not dependent upon empirical perceptions. Consequently, according to the next intermediate thesis (3), the cognition of moral laws must be a priori. Kant himself expresses this decisive intermediate conclusion in the following terms: Among the things to which everyone must admit is the fact that “the ground of an obligation,” that is, the moral law, “is to be sought not in the nature of the human being or the circumstances of the world in which he is placed, but a priori solely in concepts of pure reason” (GMS, 389,16–19). At the beginning of the ensuing paragraph, Kant reiterates this intermediate conclusion: “Thus […] are moral laws together with their principles essentially distinguished […] from everything else in which there is anything empirical” (GMS, 389,24–26). And he continues with his last step of argumentation, which leads him to the concluding thesis (4) that there has to be a pure moral philosophy. This step is constituted simply by the consequences drawn from the epistemological status of moral laws for the construction of philosophical theories. If moral laws can only be known a priori, then moral theory must have

Necessity and Apriority in Kant’s Moral Philosophy

15

a foundational part that proceeds purely a priori, that is, without reference to anything empirical. In Kant’s words, “all moral philosophy rests entirely on its pure part, and when applied to the human being it borrows not the least bit from knowledge about him (anthropology)” (GMS, 389,26–29). This is one of the possible approaches for a systematic reconstruction of the central argument in the Preface. Unfortunately, Kant offers very few suggestions of how he can assume that the thesis that there has to be a pure moral philosophy is self-evident. Consequently, much interpretation is required. And much of the interpretive work revolves around which systematic concept of necessity is implemented in the interpretation. The interpretation presented here bases itself upon a purely modal-logical concept of necessity, which entails neither an epistemological meaning nor an equivalence of “normativity” or “prescriptivity.” If one chooses other systematic basic concepts, the structure of argumentation also has to be interpreted in other ways. 8 2.5. The Practical Use of a Pure Moral Philosophy (GMS, 389,36–390,18) Although the aim of the argumentation appears to be complete, Kant, in the ensuing paragraph, suggests another form of reasoning for the “necessity” of a pure moral philosophy. This second argument, however, takes a completely different direction. Kant’s original question concerning the necessity of a moral philosophy is not free of ambiguity. For him, it was at first a matter of a necessity that referred to cognition (“speculation”: GMS, 389,37) as its goal. It could be summarized as follows: ‘If one wants to obtain knowledge of moral principles, then a pure moral philosophy is an indispensable means.’ In Kant’s second argument, a pure moral philosophy is “indispensably necessary” (GMS, 389,36) not only for the correct cognition but 8

A somewhat different interpretation is suggested by Schönecker/Wood (22004, 22– 29). They see in the “necessity” of moral laws primarily their categorical character (“Kategorizität”). Similarly, Timmermann (2004, 86) believes that it is decisive for the Kantian argument that moral prohibitions do not allow any exceptions. And Bittner (1989, 24–28) sees the starting point of Kant’s argument in the “inviolability” (“Unverbrüchlichkeit”) of moral laws. All of these approaches can support themselves on comments made by Kant. In their approaches, though, Kant’s presupposed connection between necessity in the theoretical sphere and necessity in the practical sphere is easily lost. In particular, with such a starting point, it remains unclear what role Kant’s counterfactual considerations about other rational beings have for the argument.

16

Nico Scarano

also for the correct action. It could be said, ‘If one wants to act correctly, then one needs a pure moral philosophy.’ Kant argues for this thesis: “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” (GMS, 390,2 f.). The implication contained in this argument seems, at first sight, quite strong, but it is not clear whether Kant actually advocates it so strongly. Do we really need a philosophical moral theory in order to act correctly? Are we compelled to rely on a kind of philosophical expertise when dealing with moral questions? Such an implication seems hardly congruent with the rest of Kant’s remarks. There remains nonetheless the possibility of a weaker interpretation. In this context, the expression “moral philosophy” does not entail only the explicit theory, whose formulation the scientific discipline of philosophy has as one of its tasks. Rather every human or rational being possesses a “moral philosophy” if she has the capacity to act from principle (“for the sake of”: GMS, 390,5). Accordingly, every being capable of action carries at least implicitly a moral philosophy in itself. Therefore, fleshed-out scientific theories can help us make explicit the implicit philosophical knowledge that is present in every person and therewith isolate the a priori part from the empirical part. Such an explicit knowledge can serve as a reliable “guiding thread” for action. According to Kant, one can speak of a “purity of morals” (GMS, 390,17) only if the normative principles with which we ground our actions are “pure” and contain no “empirical” elements. Thus, the scientific discipline of philosophy has a supporting but not a constitutive function for the morality of our conduct. A metaphysics of morals, as an explicitly fleshed-out philosophical theory, can, however, offer a supporting contribution in making the “idea of a pure practical reason,” which, according to Kant, lies in each of us, “effective in concreto” (GMS, 389,35). 2.6. A Methodological New Beginning (GMS, 390,19–391,15) Kant asserts that the process of working out such a metaphysics of morals is “an entirely new field […] to be entered on” (GMS, 390,22 f.). He therewith contrasts the type of theory he introduces with all other approaches that had been previously worked out in the history of philosophy. Kant is not humble in his claims. Not only does he claim that the philosophy before him had not found the right moral theory. It is

Necessity and Apriority in Kant’s Moral Philosophy

17

his opinion that all previous approaches were on the wrong track from their very starting points. For an interpretation of this passage in the text, the fact that moral principles, according to Kant, can serve not only as criterion for judgment but also as genuine grounds of action is significant. For Kant, actions are truly moral solely if not only the principles of valuation but also the particular grounds of action for the acting persons are free of all empirical content. Moral principles must be our grounds of action. Kant intimates therewith one of his central theses that he will more thoroughly develop in the course of the Groundwork. Kant accuses Christian Wolff and with him the entire moral philosophical tradition of not clearly recognizing this. In their moral-philosophies, they did not investigate “the idea and principles of a possible pure will” but began their investigations with “the actions and conditions of human volition in general,” which (and this is where Kant sees the methodological mistake) “are for the most part drawn from psychology” (GMS, 390,34–37). According to Kant, exclusively a metaphysics of morals, free from all empirical content and based on the analysis of a “pure will,” is an appropriate foundation for the working out of a normative theory of morality. 3. The Method and the Relation of the Groundwork to the Metaphysics of Morals (GMS, 391,16–392) At this point, Kant departs from his general reflections on a theory to be provided “someday” (GMS, 391,16) and dedicates himself to the work that follows. His remarks are extremely terse and without a total interpretation of the Groundwork almost impossible to comprehend. In particular, his comments on the method he uses are hardly understandable for a reader who is not already familiar with the work. He assigns three transitions to the three sections of the work although it is unclear whether the sections also contain the starting points and termini named in the title or whether they ultimately have only the transitions themselves as topics. Kant’s mention of an initially “analytically” and subsequently “synthetically” proceeding is also unclear and disputed among interpreters (GMS, 392,19–22). A possible interpretation consists in directly connecting the expressions “analytically” and “synthetically” to the word “method.” This would mean that Kant refers here to the two different methods of instruction (“Lehrarten”) mentioned in the Prolegomena, which appeared

18

Nico Scarano

in 1783. In the two-year older work, this process is explained: “the analytical method […] signifies […] that we start from what is sought, as if it were given, and ascend to the only conditions under which it is possible” (AA IV, 276 note). Kant expressly emphasizes that the analytical method has nothing to do with the procedure of conceptual analysis, but concerns only the method (“Lehrart”): “The analytical method, insofar as it is opposed to the synthetical, is very different from an aggregate of analytic propositions” (ibid). A second possible interpretation consists in referring the expressions “analytically” and “synthetically,” as they are used here, neither directly to the expression “method” nor, consequently, to the concept of method in the Prolegomena. Indeed, in the Preface to the Groundwork, Kant says only that he wants “to take the way” fi rst analytically and then synthetically. This leaves the possibility open to read the expressions in a conceptual-analytical or conceptual-synthetical sense. If one follows this interpretation, the outcome of the first two sections is constituted in the “search” (GMS, 392,3) for the supreme principle of morality. The goal of this “search” is an exact formulation of the categorical imperative. The procedure Kant uses for this is that of conceptual analysis. The formula of the categorical imperative can be found through the analytical process, but its validity cannot yet be proven. For this reason, Kant shifts to the conceptual-synthetical procedure in the third section and therewith to the “establishment” (GMS, 392,4) of the principle. In the first two sections, it is, accordingly, solely a matter of analytical judgments. In contrast, synthetical judgments, with which the proof of validity can be achieved, would stand at the center of the third section. To what degree Kant’s proceedings in the Groundwork actually are to be related to the Prolegomena is disputed among interpreters.9 In the end, only an exact reconstruction of the entire line of argumentation through the three sections of the work can provide insight into Kant’s method in the Groundwork. His remarks in the Preface offer little assistance in this endeavor. In the following section, therefore, I will not deal with these questions any further. Instead, I inquire briefly into a cluster of interpretation problems that the last section raises. This cluster revolves around the question of the relation of the metaphysics of morals to the Groundwork.

9

See, for example, Brandt (1988, 171–174), Bittner (1989, 29 f.), Freudiger (1993, 61– 70), Schönecker (1996, 1997) and Milz (1998).

Necessity and Apriority in Kant’s Moral Philosophy

19

(a) The first problem arises from the fact that Kant thinks two types of a “groundwork” for a pure moral philosophy are possible: on the one hand, a “critique of pure practical reason,” and on the other hand, the present work. Although he claims there is “really no other foundation” than a “critique” (GMS, 391,17 f.), he believes that one can do without such a critique and lists three reasons for this (see GMS, 391,20–392,2). It remains nevertheless unclear how the relation of the Groundwork to the intimated critique of practical of reason is to be defined. (b) Equally unclear is the precise relation of the Groundwork to the envisioned metaphysics of morals. Is the “groundwork” of a theory itself to be considered a part of this theory or is what Kant calls the “preliminary work of laying the ground” (GMS, 391,37) a part separate from the actual metaphysics? Undoubtedly, the parts of the theory developed in the Groundwork raise the claim of being independent of empirical knowledge and therefore belong to a pure moral philosophy. Thus, at least parts of the Groundwork have to be identical with parts of a metaphysics of morals. (c) Making matters more difficult in the identification of the relation between the Groundwork, a metaphysics of morals, and the mentioned “critique of pure practical reason” is the fact that Kant later published monographs with these or slightly modified titles. In any case, it may not simply be assumed that the later produced works actually deal with the types of theories mentioned in the Preface to the Groundwork. How can the relations of these types of theory to each other be more precisely defined? Is Kant’s expression “metaphysics of morals” a unified concept or are there different meanings combined here in opaque ways? An approach toward the resolution of these difficulties offers perhaps a more precise interpretation of what Kant understands by a “metaphysics of morals.” In the opening passages of the Preface, Kant had defined the metaphysics of morals with the help of three criteria:10 fi rst of all, it is a theory that is not purely formal but refers to particular objects. It refers, secondly, to objects insofar as these are under “laws of freedom.” And thirdly, it is perfectly free of empirical content. Therefore, a metaphysics of morals is a “pure moral philosophy.” One question, however, remains unanswered. What actually is a moral philosophy? Or articulated more precisely: what is the ontological status of such a theory? 10

Compare with above 1.

20

Nico Scarano

Kant’s arguments for the practical use or the necessity of such a philosophy offer a suggestion for answering this question. A moral philosophy is not only an object for the scientific discipline of “philosophy.” Rather, according to Kant, every rational being, even if at times confused, carries in itself a moral philosophy.11 That raises the question about what conception of “theory” can fulfill both of these functions simultaneously. For the clarification of this question I believe the conceptual distinction between a theory, understood as an abstract system of ordered propositions,12 and the presentation of such a theory, for example in the form of a philosophical text, can be of further assistance. A moral theory, understood as a system of propositions that stand in a particular relation of grounding to each other and whose center is made up of ethical principles, is an abstract object, which can simultaneously fulfill both functions. On the one hand, such a system can be brought to expression with the help of a philosophical text. On the other hand, it is also possible to carry such a system “in oneself,” that is, to accept the corresponding propositions. If one takes such a conception of theory as a basis, the confusing terminologies and exact relations of the types of theories mentioned by Kant could be newly formulated and made transparent in the following way: (i) A metaphysics of morals or a pure moral philosophy is an abstract object, a system of propositions complete in itself and containing all a priori judgments that are concerned with the laws of freedom. (ii) A work with the title “Metaphysics of Morals” contains a presentation of this system. The text expresses the corresponding propositions and their connections. (iii) If a person carries such a moral philosophy in herself, it means that she accepts the corresponding propositions and their connections. And if this person is rational, the propositions accepted by her will comprise the grounds of her action. (iv) A work with the title “Critique of Pure Practical Reason” could be a philosophical presentation of parts of this system of propositions (and for this reason is suitable as a kind of “groundwork”). But it contains additionally a presentation that relies on philosophical theories on other spheres of investigation. In Kant’s words, “in part I require for a critique of a pure practical reason that if it is to be completed, its unity with the speculative in a common principle must at the same time be exhibited” (GMS, 391,24–27). (iv) In contrast, a work with the 11 12

Compare with above 2.5. Compare, for example, with Nida-Rümelin (2002, 36 f.).

Necessity and Apriority in Kant’s Moral Philosophy

21

title “Groundwork for the Metaphysics of Morals” can restrict itself entirely to the presentation and mediation of the foundational part of the metaphysics of morals. This contains, as Kant expresses it, “nothing more than the search for and the establishment of the supreme principle of morality” (GMS, 392,3 f.). Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–). Kant’s Groundwork for the Metaphysics of Morals (GMS) is quoted according to the edition and translation by Allen Wood (New Haven, CT: Yale University Press, 2002).

Other works Bittner, Rüdiger (1989): “Das Unternehmen einer Grundlegung zur Metaphysik der Sitten. Zur Vorrede der Grundlegung”, in: Otfried Höffe (ed.): Grundlegung zur Metaphysik der Sitten. Ein kooperativer Kommentar, Frankfurt/M., 13–30. Brandt, Reinhard (1988): “Der Zirkel im dritten Abschnitt von Kants Grundlegung zur Metaphysik der Sitten”, in: Hariolf Oberer/Gerhard Seel (ed.): Kant. Analysen – Probleme – Kritik, Würzburg, 169–191. Freudiger, Jürg (1993): Kants Begründung der praktischen Philosophie. Systematische Stellung, Methode und Argumentationsstruktur der „Grundlegung zur Metaphysik der Sitten”, Bern/Stuttgart/Wien. Klotz, Christian (2001): “Gesetzesbegriffe in Kants Ethik”, in: Volker Gerhardt/Rolf-Peter Horstmann/Ralph Schumacher (ed.): Kant und die Berliner Aufklärung. Akten des IX. Internationalen Kant-Kongresses, Berlin/ New York, Vol. 3, 55–62. Köhl, Harald (1990): Kants Gesinnungsethik, Berlin/New York. Kripke, Saul A. (1980): Naming and Necessity, Oxford. Milz, Bernhard (1998): “Zur Analytizität und Synthetizität der Grundlegung”, in: Kant-Studien 89, 188–204. Nida-Rümelin, Julian (1994): “Begründung in der Ethik”, in: Nida-Rümelin: Ethische Essays, Frankfurt/M. 2002, 32–47. Scarano, Nico (2001): Moralische Überzeugungen. Grundlinien einer antirealistischen Theorie der Moral, Paderborn. Schönecker, Dieter (1996): “Zur Analytizität der Grundlegung”, in: KantStudien 87, 348–354. Schönecker, Dieter (1997): “Die Methode der Grundlegung und der Übergang von der gemeinen sittlichen zur philosophischen Vernunfterkenntnis”, in: Hariolf Oberer (ed.): Kant. Analysen – Probleme – Kritik. Vol. III, Würzburg, 81–98.

22

Nico Scarano

Schönecker, Dieter / Wood, Allen W. (22004): Kants “Grundlegung zur Metaphysik der Sitten”. Ein einführender Kommentar, Paderborn u. a. Timmermann, Jens (2004): “Kommentar”, in: Timmermann (ed.): Immanuel Kant: Grundlegung zur Metaphysik der Sitten, Göttingen, 84–152. Tugendhat, Ernst (1993): Vorlesungen über Ethik, Frankfurt/M.

Groundwork I

Allen Wood

The Good Without Limitation (GMS, 393–394) The opening pages of Kant’s Groundwork for the Metaphysics of Morals are some of the most discussed pages in the history of ethics. But they are also some of the easiest pages in all philosophy to misunderstand, and I think also some of the most commonly misunderstood. Kant’s aim in these pages is to win support for a revolutionary way of looking at the principle of morality by appealing to a set of judgments about the special goodness of the good will that he thinks belong to common rational moral cognition. But some of these judgments are not only much more controversial, but also much harder to understand correctly than he takes them to be. The result is that those who approach these pages with theoretical preconceptions opposed to Kant’s are not only likely to reject what he is saying, but also likely to misunderstand it in ways that rationalize that rejection. Even those who sympathize with Kant are very apt to misread him, by projecting on the text a set of preconceptions about the focus of the discussion that are at odds with Kant’s actual aims. In the Groundwork Kant’s argument from these claims about the good will to his conclusions about the foundations of ethics are also slow in developing, and it is difficult for readers to see exactly what propositions about the good will are crucial to his argument. By the time he begins to reach his fi rst conclusions late in the First Section, readers are very likely to have been distracted from the overall argument by controversies surrounding his initial claims, some of which may also involve serious misunderstandings of those claims. Also, Kant makes a number of distinct claims about the good will and its special goodness which he is not careful to distinguish from one another, and which make it harder to see exactly what he is claiming and how it is supposed to support his theoretical conclusions. I say all these things to prepare the reader for an interpretation of the text which will to a considerable extent go against the grain – it

26

Allen Wood

will be at odds not only with what people have usually thought Kant is saying but also with what I freely admit it is probably easiest to think he is saying. My argument, however, will be that the common and easy way of reading the text leads us into serious difficulties. It cannot make good sense of what Kant is talking about or of what he is saying, and it attributes to him a set of claims that are simply not there in the Groundwork – sometimes because Kant does not accept them, and sometimes because they are claims about topics that Kant does not intend to address at all in this passage. I ask only that the reader try on the reading I will suggest, and see for herself or himself whether my reading is not in the end more satisfactory than what seems at first to be the easy and obvious reading. This paper is going to focus chiefly on the claim made for the good will in the opening sentence of the First Section: “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” (GMS, 393). Further, the paper will focus less on what Kant means by “a good will” than on what he means in claiming that such a will is “good without limitation.” I want to distinguish this claim both from other claims Kant makes about the goodness of the good will in the opening pages of the Groundwork, and from some claims that he is often thought to have made but is not making. And finally, I want to try to say something about how the unlimited goodness of the good will is supposed to function in the overall argument of the Groundwork. Kant brings out the distinctiveness of the good will by comparing its goodness with that of other goods, which he divides into two main categories: “gifts of nature,” which he further subdivides into “talents of the mind” and “qualities of temperament,” and “gifts of fortune,” among which he distinguishes the three main objects of human passions and competitiveness (power, honor and wealth), and also two more general and encompassing goods: health and “that entire well-being and contentment with one’s condition, under the name of happiness.” Kant thinks the unlimited goodness of the good will makes it different from any of these other goods. And he tries to persuade us of the unlimited goodness of the good will both by comparing it with these other goods and by considering how we regard both the good will and the bad will in the context of complexes made up of good or bad volition and some of these other goods (or of the bad things opposed to them). Yet if we are to understand the claim that the good will is good without limitation, we must see clearly that it is a claim neither about

The Good Without Limitation

27

its worth in comparison to other goods nor a claim about the relative goodness of various complexes in which a good or a bad will might figure. Something might be good without limitation even if it were a rather minor good as compared with others, and something might also be good without limitation even if many of the complexes in which it figured were less good than complexes involving a bad will. Thus Ross (1954, 10) is mistaken when he says that Kant’s claim here entails that the good will “must never unite with anything else to produce a bad whole.” The claim that the good will is good without limitation is rather a claim about the constancy and unchangeability of the degree of goodness possessed by the good will itself (specifically, its immunity to having its goodness diminished) when it is combined with anything else, and the uniqueness of the good will in this respect when compared to any other good whatever. Kant asserts, and expects his readers to agree, that for every other kind of thing that is in general good, the goodness of its particular instances varies with the circumstances in which those instances are found, and the way they are combined with other things, especially with a will that is not good. Thus talents of the mind, such as understanding, wit and the power of judgment, or qualities of temperament, such as courage, resoluteness and persistence in an intention, although they “are without doubt in some respects good and to be wished for […] can also become extremely evil and harmful, if the will that is to make use of these gifts of nature, and whose peculiar constitution is therefore called character, is not good” (GMS, 393). Even happiness is good, Kant claims, only when it is enjoyed by a person having a good will, since such a will “appears to constitute the indispensable condition even of the worthiness to be happy” (GMS, 393). In contrast to all other goods, however, every instance of the good will is good, and equally good, under all conditions, in all combinations, and whatever its effects or consequences. The good will is good not through what it effects or accomplishes, not through its efficacy for attaining any intended end, but only through its willing, i. e. good in itself (GMS, 394). “Even if through the peculiar disfavor of fate, or through the meager endowment of a stepmotherly nature, this 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 that has its full worth in itself” (GMS, 394). The good will of course aims at good results, and with

28

Allen Wood

good fortune, achieves them. But they form no part of its own worth, and do not add the least bit to it. “It would be only the setting, as it were, to make it easier to handle in common traffic, or to draw the attention of those who are still not sufficiently connoisseurs, but not to recommend it to connoisseurs and determine its worth” (GMS, 394). Kant’s claim that the good will is good without limitation good thus amounts to what we might call the “nondiminishability thesis:” The goodness of the good will is not diminished by any circumstance in which it is found, by any of its effects, or by any combination with other things, however good or bad, in which it may be involved. Kant also holds a parallel “nonincreasability thesis” to the effect that the goodness of the good will cannot be increased by any of its circumstances or effects or by any combination with other goods. Kant sometimes states the claim that the good will is good without limitation by saying that it is “absolutely good” (GMS, 394; 402). I think he means this in the sense of “absolute” that he explicates in the Critique of Pure Reason when he says that we apply a predicate to something “absolutely” when we mean either that the predicate applies to the thing “in itself” or “internally” (apart from any relation the thing may have to other things) or else when it applies to the thing “in all respects” or “in every relation” (KrV A324 f./B380 f.). The good will is absolutely good, in this sense, because its goodness does not vary with its relation to any other thing, and is therefore possessed entirely in itself or apart from any relation that the good will may stand to other goods. There is another thesis that Kant asserts in the same paragraph that obviously has some close relation to these two theses. This is what we can call the “unconditionality thesis:” The good will is unconditionally good, or equally good under any and all conditions, whereas all other goods are good only conditionally. Gifts of nature are good when used by a good will but bad when used by a bad will for its ends. Gifts of fortune are bad when they lead (as through arrogance) to a bad will, or when they are enjoyed by a being that lacks the good will which is the indispensable condition for the worthiness to enjoy them. I think we may view the nondiminishability and nonincreasability theses as consequences of the unconditionality thesis, in the following way: If we consider all possible instances of some kind of thing, then that kind of thing is good without limitation only if all of these possible instances are good. If some of them are bad, then this limits the goodness of that kind of thing. Since the goodness of all good things except a good will is conditioned, those instances of other kinds of things in which the condition is not fulfilled are not

The Good Without Limitation

29

good, and they constitute a limitation on its goodness. By contrast, because the good will is unconditionally good, all possible instances of it are good, and it is therefore good without limitation. In this way, the unconditionality thesis entails the nondiminishability thesis. By a parallel argument, the unconditionality thesis implies also that since no addition of any other good to the good will can increase its goodness (but can at most increase the goodness of some combination or larger whole of which the good will and this other good are parts), the nonincreasability thesis also holds. This calls attention to an important point about all these theses about the good will. They are theses contrasting the good will as a general kind of thing with other kinds of things. They are not claims directly about particular instances of the good will or particular instances of other goods. The unlimitedness of the good will’s goodness consists in the fact that no possible instances of the good will are bad, in contrast to the limitation of the goodness of all things that are only conditionally good, since when something is only conditionally good, some possible instances of it – those in which the condition is not fulfilled – will not be good. Yet when the condition for the goodness of a conditionally good thing has been fulfilled, that instance possesses goodness without limitation – at least, without the specific sort of limitation that is to be excluded by the claim that the good will is good without limitation. A gift of nature that is well used by a good will, or a gift of fortune that is good because it is enjoyed by a person with a good will – these particular instances of these goods are wholly good, nothing about them limits their goodness. None of these theses about the good will we have been considering, therefore, enable us to infer anything directly about the comparative goodness of particular instances of different goods. We cannot infer from them that a will that is good is better than the deserved happiness of the being who possesses that good will. As far as the nondiminishability and nonincreasability theses are concerned, it might be that the deserved happiness of a person with a good will is to be regarded as a much greater good than the good will which serves as the indispensable condition of the goodness of this happiness. Or suppose we consider two agents, A and B. A has a good will, and uses every means at its command to actualize good ends, but owing to a special disfavor of fate or through the meager endowment of a stepmotherly nature, none of its good aims are realized, but the very reverse of them, an extremely bad result, comes about. B has a bad will and sets evil ends, but owing to the ironies of fortune, the very op-

30

Allen Wood

posite of what B aims at is realized, and this is a very good result, perhaps exactly the result that A, with the good will, would have willed to come about. Which complex, considered as a whole, should we value more: A’s good will, with its bad results, or B’s bad will, with its good results? Nothing in the nonincreasability thesis, or the nondiminishability thesis, implies any answer to this question. In other words, the nondiminishability and nonincreasability theses are theses are only about the goodness of the good will as a possible kind of good, and also theses only about the good will itself. There is no reason to suppose that Kant would deny that the complex of a good will plus the good results at which it aims is a better whole than the combination of the same good will with bad results that might come about despite the good will’s fruitless attempts to prevent them. Nor do the the nondiminishability and nonincreasability theses themselves make any claims about the goodness of the good will in relation to the goodness of other goods. Something whose goodness can be neither increased nor diminished might be a relatively minor good, not nearly as good as many of the things whose goodness is increased or diminished by their circumstances, or effects or by their combination with other things. As a matter of fact, however, Kant does hold that the good will is also a greater good than any of the goods that are subject to being increased or diminished: “considered for itself, without comparison, it 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” (GMS, 394). And this seems also to be something that Kant sometimes intends to say when he claims that the good will is “absolutely” good. Here, however, the claim that the good will is absolutely good does not mean that its goodness is independent of its relations to other things, but rather makes the quite distinct claim that it is a greater good, in comparison to other things, than anything else. This distinct claim, which we may call the “higher worth thesis,” does entail that when a person with a good will achieves the happiness of which she has made herself worthy, then it is the good will and not the happiness that is the greater good. The higher worth thesis may also imply – though this seems to me less certain – that in our example of A and B above, that the complex of A’s good will and its unfortunate bad results is to be preferred to the complex of B’s bad will and its fortunate but unintended good results. This remains uncertain because it might be that the goodness of the unintended good results in the case of B, or the badness of the

The Good Without Limitation

31

unintended bad results in the case of A, or especially both together, might be so great as to outweigh, all things considered, the fact that the goodness of A’s will is a higher good than any results that could have issued from it (or, as one may suppose by parallel reasoning, that the badness of B’s will is a worse thing than any harm B’s bad will could have done). Perhaps Kant intends the higher worth thesis to exclude these possibilities too: The goodness of the good will is to be so great that it is always to be valued more highly than the achievement of any good result, no matter how good, or the avoidance of any bad result, no matter how bad. Certainly if it is an unconditioned requirement of morality that we should have a good will under all circumstances, and no matter what the results, then one way to take this requirement is that a comparative judgment of this kind is implied by it. But we could also think of the unconditional requirement to have a good will simply as a requirement about how to will, and not as resting on (or as entailing) any comparative judgments about the value we should attach to the states of affairs that might result from what we will. However all this may be, the chief point to be emphasized is that the higher worth thesis is not part of, and is logically independent of, what Kant is claiming when he says that the good will is good without limitation. It remains to be considered what relationship, if any, there is between the nondiminishability thesis and the higher worth thesis within Kant’s arguments in the Groundwork. This last remark should also remind us that in the context of the Groundwork, none of Kant’s claims about the goodness of the good will are being presented merely for their own sake. Nor is the goodness of the good will being presented by Kant as a value-claim that is fundamental to his ethical system as a whole. If we wish to extract from the Groundwork the value-claims with that status, we would do far better to choose either the claim that rational nature is an end in itself or the claim that the rational will is universally legislative. Kant’s claims about the good will are rather part of his strategy for deriving certain claims about the nature of the supreme principle of morality. It is also noteworthy that this strategy is employed solely within the First Section of the Groundwork, where Kant is trying to derive a formula of the moral law solely from common rational moral cognition. The goodness of the good will is not appealed to at all in the argument of the Second Section, where Kant argues more definitively, from a philosophical conception of rational willing, and where he completes his systematic formulation of the moral law by advancing beyond the universal law formula to the formula of humanity as end in itself and

32

Allen Wood

the formula of autonomy. He mentions the worth of the good will again only at GMS, 437, after that development has been completed, and merely in order to connect up the results of the Second Section with those of the First Section. In order to say how Kant’s claims about the unique value of the good will are related to his argument, we need to say at least a little more about what Kant means by the “good will.” This is not as easy as one might think, since despite the emphasis he places on the goodness of the good will, Kant never gives us a direct and clear account of what the good will is. Instead, he immediately narrows his focus to a special case of the good will, namely, the concept of acting from duty, which, according to Kant, “contains that of a good will, although under certain subjective limitations and hindrances, which, however, far from concealing it and making it unrecognizable, rather elevate it by contrast and let it shine forth all the more brightly” (GMS, 397). It is not immediately clear how much farther Kant intends the concept of the good will to extend, beyond the concept of acting from duty. Clearly the divine will is thought of as a good will, but God utterly lacks the “subjective limitations” that would be necessary for the concept of duty to apply to him. Many readers of the Groundwork tend simply to identify the good will, at least for humans, with the will that acts from duty. But I think that is a mistake. Kant famously, if cryptically, identifies the will with practical reason (GMS, 412). Reason, however, is described in the Critique of Pure Reason as the “faculty of principles” (KrV, A299/B356). This means that the will is the faculty of practical principles – that is, the faculty determining the normative policies that guide a person’s actions. Kant distinguishes two kinds of principles: subjective principles or maxims, and objective principles or laws (GMS, 401 f.; KpV, 19). Maxims are principles adopted contingently by a particular subject to guide its actions: they are normatively valid only for the subject that adopts them, and they remain valid only as long as this subject continues to accept them. Laws are principles objectively valid for all rational beings. It is an important Kantian thesis that rational beings are capable not only of adopting maxims, valid subjectively for themselves, but are also autonomous, capable of giving laws that are valid objectively for all rational beings (GMS, 431). Will is the faculty that gives both kinds of principles, and derives actions from them (GMS, 412). Every rational will gives laws, so the distinction between a good will and bad will can have nothing to do with this objective legislation. It can have to do only with the subjective principles or maxims the will

The Good Without Limitation

33

adopts. The good will, therefore, is the will that adopts good maxims, namely, those that accord with objective laws of reason. The bad will is the will that adopts maxims contrary to these laws. We can now characterize the difference between those who think the good will, at least in the case of human beings, belongs only to subjects who act from duty, and those (such as myself) who think that the good will, even for human beings, includes some subjects who do not act from duty. The question is: Does the ground or incentive from which the subject adopts its maxims also belong to those maxims as part of their content? If the answer to this question is yes, then it would not be possible for two subjects to adopt or to act on the same maxim from different incentives, since the difference in incentives would all by itself make the maxims different. For example, it would not be possible for two merchants to adopt the same maxim: “Deal honestly with all your customers, whether experienced or inexperienced” (GMS, 397), one of them from prudence (concern for his reputation) and the other from duty. Nor would it be possible for the same merchant to act on this same maxim at one time from prudence (say, when he realizes he is being watched as he deals with an inexperienced customer) and another time from duty (when he realizes he can get away with cheating the customer, but constrains himself to deal honestly from respect for the moral law). For if the incentive itself is part of the maxim, then the merchant who deals honestly from prudence has a different maxim from the merchant who deals honestly from duty, and only the latter truly has a good will. On this view, those who act externally in the same way as the person with a good will do not really have a good will unless they act from duty. If they are honest, or beneficent, but are so only from prudence or inclination, then their maxim and their will is not good. In contrast, someone who thinks that the incentive need not be part of the maxim can say that both merchants act according to the same maxim, which, moreover, since it is in conformity with duty (i. e. is what duty would command), counts as a good maxim, the maxim making for a good will. On this view, the merchant who adopts this dutiful maxim from prudence has a good will as well as the merchant who adopts the same maxim from duty, and if the same merchant follows this dutiful maxim sometimes from prudence and sometimes from duty, he has a good will whichever incentive moves him at the moment. (Of course, a merchant who acts from prudence and whose real maxim is not to deal honestly, but only to deal honestly when he won’t get caught cheating, does not have the same maxim as the

34

Allen Wood

merchant whose maxim, adopted from prudence, is to deal honestly regardless of whether he can get away with it. The latter maxim makes for a good will, because it accords with the moral law, while the former maxim, by allowing for cheating when you can get away with it, does not accord with the moral law.) I think this latter interpretation, which allows that a person may have a good will even when they do not act from duty, is clearly preferable to the former interpretation. For it makes sense to think that Kant’s examples in the Groundwork (of the merchant, the man preserving his life, the beneficent man, the gout-sufferer who resists the temptation to indulge in unhealthy food or drink) are examples in which Kant is imagining two people (or the same person at different times) following the same dutiful maxim, one time from prudence or inclination, the other time from duty. Moreover, Kant says of beneficent or honorable actions done from inclination that they deserve praise and encouragement (GMS, 398) – something Kant should not be expected to say if such maxims exhibit a will that is not good. Further, Kant holds that we have a duty to cultivate our sympathetic feelings (MdST, 456) and even our love for others as inclination (EaD, 337 f.) on the ground that these feelings and inclinations will motivate us to do our duty. Yet surely this is inconsistent if Kant thinks any maxim adopted or followed from inclination would deprive the will of its goodness. Finally, Kant describes duty as “necessity of an action from respect for the law” (GMS, 400). From other passages, we know that Kant understands this necessity, also called “necessitation” (Nötigung) as a kind of “constraint” (Zwang), namely an inner constraint or selfconstraint through respect for the law (MdST, 379). In other words, to act from duty is to constrain oneself to do an action that accords with the moral law. However, it is obvious that such self-constraint is not always needed to achieve compliance with the law. Prudence may give the merchant a sufficient ground to deal honestly with customers, self-love may be sufficient to move us to preserve our lives or to adopt a healthy diet, and sympathy may lead to beneficence. It also makes sense to think that where self-constraint is not necessary, it is also properly speaking not even possible. For what sense would it make to say that one person who enjoys making others happy does so freely, while another does it with self-constraint? Those who wish to limit the good will to the will that acts from duty must apparently say such things. Still worse, they must say that only the person who acts with self-constraint has a good will, while the other equally beneficent

The Good Without Limitation

35

person does not have a good will. It is true that Kant has sometimes been charged with holding such absurdities as these, but they are not a necessary or even a natural way of reading what he says.1 Within the First Section itself, the argument appeals (as Kant says it will) not to claims about the good will, but to claims about acting from duty. 2 Specifically, Kant argues for the first formula of the prin1

2

Many think that for Kant it is possible to act from duty even when one has an inclination to do the act which is strong enough that no self-constraint is required. To sustain such a reading of the text, however, one is forced to do either of two things: On the one hand, one might attribute to Kant a highly mechanistic conception of action and motivation, according to which, when one has two different incentives to perform the same action, there must be a fact of the matter about which incentive caused the action, much as, if a light bulb were hooked up to two different wires, there would have to be a fact of the matter about which wire it was that delivered the current to the lighted bulb. Or, on the other hand, one would have to interpret claims in such cases about one’s ‘real motive’ as claims not about that particular action at all, but rather about one’s character or volitional dispositions, the structure of one’s volitional priorities in general – to act from duty is to give duty general priority over inclination in making one’s decisions. But neither of these options is the least bit attractive. The first requires us to ascribe to Kant, on the basis of no explicit textual support whatever, a highly implausible and unappealing theory of action. The second transforms what he evidently intends to be a claim about the moral worth of this action into a general claim about the agent, or about what the agent would do under counterfactual circumstances. (Would the agent still act in a dutiful way if the cooperating inclination were removed?) But it also does not seem that there need be any corresponding general truth about the agent’s character, or about what the agent would do, for this action to have moral worth in the case Kant is imagining -- where the agent must constrain himself through respect for the law if the dutiful action is to be performed at all. In fact, we might think that an action would have more moral worth in such a case if it were performed against the agent’s general dispositions, and in conflict with his (generally bad) character. For that would give the agent’s act of self-constraint an even more heroic quality. The more one reflects on the inconveniences and implausibilities we are required to digest in order to maintain the usual interpretation of Kant here, the more attractive becomes the idea that in this passage, all Kant means by ‘acting from duty’ is ‘acting with self-constraint through respect for law’ – something that need not, and therefore could not, occur in any case where there were a co-operating inclination that made self-constraint unnecessary (hence impossible). Another temptation to think that a will can, in the sense relevant here, act from duty even when it need not constrain itself to do so is drawn from Kant’s claim that only the action done from duty has “moral worth.” It might seem to follow from this that a dutiful action requiring no self-constraint (due to co-operating inclinations, such as sympathy) is morally worthless. Kant has sometimes been accused of holding such views, but it is easy to see how we could avoid that reading. For it is easy enough to avoid it even while understanding “acting from duty” as “acting with self-constraint out of respect for the law” once we realize that an action which lacks “moral worth” in the sense Kant means here need not be a morally worthless act, or an act of no value to morality. Obviously, an act that conforms to duty, especially when done with a good will (that is, from a maxim that conforms to duty) would have some kind of value to morality (and deserve “praise and encouragement” as Kant says it does), even if it is not done from duty. Notice that when speaking of “moral worth” in

36

Allen Wood

ciple of morality (the formula of universal law) by arguing that duty is necessity of an action only from respect for the law, and hence that the law to which this respect is directed can be limited by no condition arising from any end or inclination, so that conformity to this law must be nothing but the conformity of actions as such with universal law (GMS, 402). Later, to be sure, he also relates the way in which the law is unlimited in its content by any end or inclination to the unlimited goodness of the good will, when he says about respect for the moral law “estimation of a worth which far outweighs everything these passages, Kant often appends an adverb to “moral”: he speaks of: “true moral worth” (GMS, 398) “genuine moral worth” (GMS, 398) “authentic moral worth” (GMS, 399), and then speaks of “that so pre-eminent good we call moral” (GMS, 401). I think his talk about “moral worth” is best understood as follows: Morality cares about many things, but not all of them have true or genuine or properly moral worth. Morality approves of the happiness of a good person, for instance. But happiness, even deserved happiness, is not properly a moral good. Neither, thinks Kant, is even an action conforming to duty, when this conformity serves self-interest or inclination. It has properly moral worth – and deserves, as Kant says, not merely praise and encouragement but also esteem – only when it would not have happened but for the agent’s self-constraint out of respect for the moral law. This point is closely related to a way in which Kant’s entire discussion in these pages, and most especially his comparison of beneficent actions done from inclination and from duty, is very commonly misunderstood. Kant describes a man who fi rst (a) makes others happy because he enjoys doing so, and then, after his inclination to beneficence is eclipsed by his own sorrows, (b) tears himself out of deadly insensibility and is beneficent not from inclination but from duty (GMS, 398 f.). There are different questions one might ask in comparing the two cases (a) and (b). The question Kant means to ask is: “In which case does the action call forth from us the most properly moral esteem for the action?” This is not at all the same question as: “Which situation would we most want to be in?” or even “Which sort of beneficent action would we, as moral educators, most try to encourage?” There may be moral theories (and moral psychologies) that deliver the same answer to all these questions: that say: the act deserving the most properly moral esteem is the one we would most want to perform and most want to encourage. To people who are committed to such theories (or who, out of inattention, don’t even see the difference between the questions), Kant’s insistence that we should feel properly moral esteem for the act only in case (b), not in case (a), is often taken to imply that he would also favor case (b) in response to the other two questions – which would, of course, be perverse. But there is actually no good reason to think that Kant would not agree with us that we should prefer to do beneficent actions from inclination rather than having to constrain ourselves to do them from respect for the law, or that he would dissent from the thought that we should try to educate people to be beneficent from inclination wherever possible, rather than having to make themselves be beneficent out of respect for moral laws. His point (to which I think many readers remain simply oblivious) is that despite this, it is the man who must constrain himself to be beneficent whose action is most to be esteemed as having a worth that is properly (truly, genuinely, authentically) moral. Perhaps even when they understand the question, not all readers will agree with Kant’s answer. But if we look carefully at most of the critics who scorn what he says in this passage (and they are legion), I think we will find virtually none who has actually addressed precisely the question Kant intends here.

The Good Without Limitation

37

whose worth is commended by inclination, and that the necessity of my actions from pure respect for the practical law is what constitutes duty, before which every other motive must give way because it is the condition of a will that is good in itself, whose worth surpasses everything” (GMS, 403). However, this comparison of the ground of the moral law with the value of the good will occurs well after the universal law formula has already been derived. The analogy between the law and the good will represents a confi rming observation regarding the claims that Kant has been making, and is not directly a part of the argument for them. The same is true of Kant’s other prominent reference to the goodness of the good will in the Second Section of the Groundwork. Kant begins the paragraph immediately following the derivation and systematization of all three formulas by saying: “Now we can end at the place from which we set out at the beginning, namely with the concept of an unconditionally good will” (GMS, 437). He then argues that good will’s principle will be the same as the formula of autonomy, in which the other two formulas have been combined: “Act always in accordance with that maxim whose universality as law you can at the same time will” (GMS, 437). But the connection of this formula, or any formula of the moral law, with the good will, occurs only as a confirming afterthought, and never directly as part of their derivation. Of course, if having a good will requires acting from duty, then the arguments of the First Section that proceed from the esteem we accord to someone who acts from duty are also arguments from some comparable value property of a good will – perhaps its goodness without limitation or the unconditionality thesis or the higher worth thesis, or some combination. But recall that even those who think that a human being must act from duty in order to have a good will do not think that this is true of the divine will, which is also a good will, and hence also unconditionally good, good without limitation, and of a higher worth than any other thing. Hence arguments based on our esteem for acting from duty are not based on appeals to these value properties of the divine will, nor therefore, on the same value properties as they are found in the good human will. I think the conclusion we must draw is that, contrary to the impression so easily obtained from the famous opening sentence of the Groundwork, none of Kant’s arguments for any of the formulas of the supreme principle of morality are based on the unlimited goodness of the good will, nor on any of the other value properties that Kant regards as unique to the good will. From this point of view, we can only conclude that the many readers of Kant

38

Allen Wood

who come away from the Groundwork with the impression that his opening claim about the goodness of the good will is fundamental to his entire ethical theory are profoundly mistaken. Kant sometimes speaks (though only here in the Second Section, never in the First Section) not merely of the “good will” but of the “absolutely good will.” “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” (GMS, 437; cf. 426, 439, 444, 447). This is not merely a repetition of the claim that the good will is “absolutely good” (either in the sense that it is good unconditionally, regardless of any of its relations to other things, or in the sense that it has a higher worth than any other possible good thing). Rather, it identifies a special case of the good will, where not only does its maxim conform to the moral law, but where the principle of this will is itself conformity to moral laws. 3 This makes sense, however, only after we have identified the moral law, and can therefore specify the will whose goodness is absolute in the sense that it conforms to that law. It must not be thought that all along Kant had this “absolutely good will” in mind, and that it was only of it that he said that it is good without limitation. Rather, any will that adopts a good maxim, in whatever respect and from whatever incentive, is, to that extent, a good will, and good without limitation.4 3

4

In the Preface, Kant says: “For as to what is to be morally good, it is not enough that it conform to the moral law, but it also must happen for the sake of the law” (GMS, 390). This is usually read as saying the same as that an action has moral worth only when it is done from duty. But it does not say the same thing. For a holy will (to which the very concept of duty, and of acting from duty, could not apply) presumably also conforms itself to the moral law and acts for the sake of the law. It is better to understand Kant to be saying here that properly moral good is to be found in a will which performs dutiful actions on principle, rather than (as he puts it here) “contingently and precariously” from co-operating inclinations. In other words, we come closer to a worth that is properly moral when we are dealing with an absolutely good will, a will whose basic principle is to conform its actions to the moral law, than a good will which is merely acting on some principle or other that conforms to the law. Kant no doubt would deny that it would be possible for a finite and imperfect human will to do this always without sometimes having to act from duty (with self-constraint). But in a person whose temperament has been well cultivated to dispose it to morality, it seems that it might be possible to act quite often for the sake of the law without having to constrain oneself through respect for the law. So I do not think that what Kant is talking about here in the Preface precludes the possibility that a good will, possessing a good that is properly moral, especially if it should be an absolutely good will, might often act not only in conformity with the moral law, but also for the sake of the law, without having to act from duty. A will that acts on the principle that all its maxims should conform to the moral law is, I think, what Kant means by a will that acts “for the sake of the law.” But, as has been pointed out in the previous note, this is a will that need not act from duty

The Good Without Limitation

39

It may help here to realize that a “will” (as practical reason or the capacity to act on principles) is an abstraction. A good will is not the same as a good person. Every (finite, imperfect) person’s will is always in some respects good and in some respects bad, assuming that all of us adopts some good maxims and some bad ones. So a bad person may (sometimes, in some respects) have a good will, and a good person may (sometimes, in some respects) have a bad will. For example, the merchant who, out of prudence and with a self-serving aim, adopts the maxim of dealing honestly with his customers, still has a good will as far as that maxim is concerned. But if his maxims are generally self-serving, then on the whole he would doubtless be a bad man, and would even have (on the whole) a bad will. We must not think that because he is a bad man on the whole, his will regarding this maxim is a bad will, as long as his maxim does indeed conform to morality. Moreover, a merchant who also had the maxim (or meta-maxim) of adopting maxims in conformity with the moral law for the sake of the law (GMS, 390) would have on the whole a better will than this merchant, who adopts the dutiful maxim from prudential considerations. Even this latter merchant, however, would still not be acting from duty in cases where prudential considerations gave him sufficient grounds for following the maxim of honesty, since no self-constraint would be needed in those cases to get him to follow the maxim of honest dealing. Thus we should not confuse acting “from duty” with acting “for the sake of the law.” Nor should we think that anyone who is honest, or beneficent on principle (as distinct from being honest out of prudence or beneficent out of sympathetic inclination) is thereby acting from duty. Such thoughts, which are admittedly all too easy to entertain, misconstrue what Kant means by “acting from duty” and contribute to many common confusions about what he is saying in the opening pages of the First Section of the Groundwork. A good will must also be distinguished from a person’s good character or virtue, which is the strength of the person’s character in act(this is especially obvious in the case of a holy will, but might be true in the case of a finite human will wherever its conformity on principle to the moral law comes about without the need for self-constraint through respect for the law). If that is right, then an absolutely good will might exemplify the properly (truly, genuinely, authentically, pre-eminently) moral good, yet without acting from duty. In the First Section, however, Kant never considered the case of the absolutely good will, so he took no position on this. All the cases he considered had to do with the conformity of particular actions to particular duties or dutiful maxims. In those cases, it was only through exhibiting self-constraint through respect for the law (acting from duty) that the properly moral good could be exemplified.

40

Allen Wood

ing on good maxims (MdST, 408 f.). Good character implies a certain good will, which is why Kant speaks of the two together in Section One of the Groundwork (GMS, 393; 398); but as he makes explicit elsewhere, a good will can co-exist with lack of virtue, when a person adopts good principles, but is too fragile or weak of character to carry them out reliably (RGV, 29; GMS, 407 f.). Goodness of will might also be combined with a serious lack of wisdom or judgment, so that the actions performed may even be contrary to the good principles that make the will good. In both these ways, a person with a good will can be a person who often does what is morally wrong. When it forms part of a syndrome of moral weakness or bad moral judgment, the good will (the adoption of good principles) might be thought to lose some of its goodness, and to be itself less good than it would be when found in a person of greater virtue and moral wisdom. The more we reflect on this point, the more reason we might find to doubt that Kant is even correct in regarding the good will as good without limitation. But even if we do doubt the unlimited goodness of the will on these grounds, that doubt need not call into question the fundamental values on which Kantian ethics rests: the autonomy of reason, the universal self-legislation of the rational will, the dignity of rational nature as an end in itself. The fact that goodness of will is an abstraction – that all of us doubtless have a good will in some respects and a bad will in others – may account for the fact that in the early pages of the Groundwork Kant chooses to narrow his focus from the good will to the special case of acting from duty. For every instance of acting from duty displays a good will, even though the converse of this does not hold, and acting from duty also displays a good will under adverse or trying circumstances (“under certain subjective limitations and hindrances,” as Kant puts it) so that in acting from duty the good will shows itself with a kind of special heroism not present in many cases of the good will (the adverse circumstances “elevate it by contrast and let it shine forth all the more brightly.”) For the same reason, Kant prefers the person who acts from duty to the one who is dutiful from prudence or inclination only in the sense that the person who acts from duty is due greater esteem, not in the sense that we should prefer to be like him in the sense of being in his (adverse and unfortunate) situation. All too many readers of the Groundwork – probably because they do not distinguish the question “Who is most morally admirable?” from the question “Whom would I most want to be like?” – have badly misread what Kant is saying here, as though he thought we should prefer

The Good Without Limitation

41

having interests or inclinations adverse to the moral law over having interests and inclinations in harmony with it. There are, as I said at the outset, many different ways in which it is easy to misread this famous but treacherously brief discussion. It is no exaggeration to say that a large part of Kant’s reputation as a moral philosopher is based on such serious but widespread misunderstandings. The (perhaps surprising) conclusion that the goodness of the good will is not foundational for Kantian ethics, though in one sense true, may nevertheless be exaggerated, at least if it is stated in this unqualified form. For if we look beyond Kant’s explicit arguments I think it is nevertheless possible to see some very suggestive connections between his striking claims about the good will and some of the main theses for which he is arguing in the Groundwork. Perhaps the clearest case is the relation between Kant’s unconditionality thesis about the good will and his claim that the supreme principle of morality is a categorical imperative. An imperative is an objective principle or law of reason, or more precisely, a command of reason, that is, an objective principle or law that is directed to a being whose will does not necessarily accord with that law, but which may need to be rationally constrained to accord with it (GMS, 413 f.). A holy will, such as the divine will, necessarily acts according to laws or objective principles, and therefore no imperatives or rational commands apply to it. A categorical imperative is a command of reason whose rational bindingness on the will is independent of any end external to that imperative itself. A technical imperative that directs us to take the necessary means to an arbitrary end is rationally binding only on those who have set that end. A prudential imperative that directs our pursuit of happiness is rationally binding on us only because we are assumed to have our happiness as an end. These imperatives are therefore hypothetical rather than categorical (GMS, 414). But a moral imperative directs us to perform actions that are morally required, or at least morally meritorious, and it is rationally binding on us irrespective of any ends we may have. Hence a moral imperative is a categorical imperative. Of course, a categorical imperative might also rationally direct us to set certain ends – Kant argues that it requires us to set as ends both our own perfection and the happiness of others (MdST, 385–388). But its rational bindingness is neither conditional nor dependent on our setting those ends. A categorical imperative therefore rationally directs us to do certain things without reference to any ends we may already have, and therefore, its bindingness is not conditional on the consequences of

42

Allen Wood

our actions for any actual or possible ends given independently of the imperative itself. More precisely, however, a categorical imperative rationally directs us not so much to do certain things, as to adopt and act on certain principles (maxims), which may in turn include or entail the performance of certain actions, or the setting of certain ends. The basic effect of a categorical imperative, therefore, is to declare certain maxims of action binding on us irrespective of their relation to any independently given ends, or therefore, of their relation to any consequences (good or bad) of acting on them. The will, as we have seen above, is for Kant the rational faculty of practical principles. That is, it is the faculty through which a rational being gives itself such principles, whether subjective principles (maxims) or objective principles (laws). A will is good when its maxims accord with the laws it gives itself. For a holy will, such as the divine will, there are only objective principles, or else, what would come to the same thing, its subjective principles or maxims are necessarily identical to the objective principles it rationally legislates to itself. (This is why no rational commands or imperatives apply to such a will.) For a finite and imperfect will, however, goodness of will consists in the contingent and voluntary accord between its maxims and its laws. In other words, as we have already seen, what makes a good will good are its principles, or, in the case of a finite and imperfect will, its maxims. Now there are two possible ways in which maxims, hence wills, might be considered good. They might be good (at least in part) because of the consequences of acting on them, hence good relative to certain independently given ends. Or they might be good simply in themselves, irrespective of any such consequences or their conduciveness to any independently given ends. In the former case, the goodness of the good will itself would necessarily be increased if its consequences were good, and diminished if its consequences were bad (relative to those ends). According to Kant’s unconditionality thesis, however, and the nondiminishability and nonincreasability theses that follow from it, the goodness of the good will is unaffected by its consequences, whether these be good or bad. Therefore, the maxims constituting the goodness of a will are not good only relative to certain consequences or independently given ends, but must be regarded as good in themselves, irrespective of their consequences or the effect of these consequences for any independently given ends. That means that the principle that specifies which maxims are good must not do so by reference to any independently given ends, but only by regarding certain maxims as

The Good Without Limitation

43

good (as rationally required or rationally meritorious) in themselves, irrespective of their consequences for any independently given ends. This principle, therefore, cannot be a hypothetical imperative of any kind, but must be a categorical imperative. It follows that anyone who accepts Kant’s nondiminishability and nonincreasability theses about the goodness of the good will is committed to regarding the objective principle or law of that will, when it is addressed as a command to a finite and imperfect will, as a categorical imperative. The nondiminishability and nonincreasability theses (and therefore the claim that the good will is good without limitation) still do not settle one important issue about the bindingness of categorical imperatives. They do not, namely, settle the question whether a categorical imperative must take rational precedence over all hypothetical imperatives – in other words, whether it is rationally required that we adopt the maxims this imperative declares good in themselves even when they force us to sacrifice our happiness, or some other end given independently of this categorical imperative. This issue is settled, however, by Kant’s higher worth thesis. For this tells us that the good will is to be rationally preferred to any other good (including any independently given ends), and hence that the principle (the categorical imperative) that rationally commands us to adopt maxims that are good in themselves has rational priority over all other practical principles of reason. This means that the categorical imperative is unconditionally binding not only in the sense that its own rational bindingness is independent of any independently given ends, but also that it is binding on the will no matter what other ends the will may (contingently or even necessarily) set independently of it. In these ways, then, Kant’s various claims about the unique goodness of the good will are after all intimately related to central theses of the Groundwork. They do not play any direct role in his arguments, but their acceptance already commits us to an essentially Kantian conception of practical reason and the supreme principle of morality. Even if we doubt, on the grounds I have suggested earlier, that Kant is right that the good will is good without limitation, simply recognizing that the good will is an important good is enough to give us reason to attend to the importance of acting on moral principles and recognizing the importance for morality of our capacity to act on principles – in other words, of the will. In that sense, the famous claim with which Kant opens the First Section of the Groundwork might plausibly be read not as a sober statement of a fundamental moral thesis, but rather as a rhetorical proclamation whose success in directing our

44

Allen Wood

attention to the moral importance of willing on principle might be enough for Kant’s purposes, even if we decide on reflection that it contains more hyperbole or exaggeration than it does literal truth. Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–) (abbreviated as ‘AA’). The translation of the Groundwork will be my own, taken from Kant, Groundwork for the Metaphysics of Morals (New Haven: Yale University Press, 2002). The Critique of Pure Reason will be cited according to the A/B pagination from the first and second editions. GMS Grundlegung zur Metaphysik der Sitten, AA, IV KpV Kritik der praktischen Vernunft, AA, V KrV Kritik der reinen Vernunft, AA III, IV MdST Metaphysik der Sitten, Tugendlehre, AA,VI EaD Das Ende aller Dinge, AA, VIII RGV Die Religion innerhalb der Grenzen der blossen Vernunft, AA, VI

Other works W. D. Ross, Kant’s Ethical Theory: A commentary on the Grundlegung zur Metaphysik der Sitten (Oxford: Clarendon Press, 1954),

Christoph Horn

Kant on Ends in Nature and in Human Agency The Teleological Argument (GMS, 394–396) Many readers of the Groundwork feel some irritation about the fact that Kant directly combines, in the First Section, his concept of a good will with teleological reflections – and even with quite strange ones. Immediately after having introduced his starting thesis that good will is the only good without restriction (GMS, 393,5–394,31), Kant turns to a discussion of the teleology of human nature (GMS, 394,32–396,37). In this passage, he opposes a view of teleology which, in his eyes, is thoroughly mistaken. Moreover, he develops what he thinks to be the correct account of this topic. This is surprising since Kant, at least in his ‘critical period’, is well known as a philosopher who rejects traditional teleological claims, in particular the famous teleological argument for God’s existence. Therefore, what we find in our passage seems to be a sort of theoretical anachronism, a return to bad metaphysics. And indeed there is, at least prima facie, little to say in favor of Kant’s teleological considerations. Consequently, in the vast literature on the GMS, the passage received relatively little attention and some bad press.1 The line of thought developed in the teleological passage can be divided into four sections: (a) Kant fi rst tells us why he introduces his considerations on teleology anyway (GMS, 394,32–395,3). The reason is that he wishes to reject a serious suspicion which can be raised against his concept of good will: it could be said to be a mere fantasy or a philosophical figment. What he wants to make plausible is that the aim of nature in “assigning reason to govern our will” consists precisely in the morality of good will. 1

See e. g. Schönecker/Wood (2002, 52–4).

46

Christoph Horn

(b) After that, Kant provides a more extensive rejection of the claim that happiness (Glückseligkeit) is the true end of human nature (GMS, 395,4–27). He makes use of an argument which takes the form of a modus tollens: The natural dispositions of each animal are optimally purposive or suitable for the goals for which the animal is destined. (Let us call this the ‘principle of suitability’.) Now, if happiness were the goal for humans we would have to find, according to this principle, some indications for the alleged destination within men’s natural endowment. In truth, however, there is no sufficient ground to accept the idea that happiness is our natural goal. What we in fact find is that we are deeply determined by our possession of practical reason. If the natural goal of men consisted in the pursuit of happiness, then the possession of reason would be a relatively dysfunctional means, since it would have been more appropriate to endow humans with stronger instincts. Therefore, given the principle of suitability, the fact that we possess practical reason is not consistent with the idea of happiness as our natural end. Hence, according to Kant, the assumption that happiness is men’s natural destination should be relinquished. (c) In the next, smaller passage (395,28–396,13), Kant offers us some considerations which are intended to support the thesis that practical reason is a dysfunctional instrument to gain happiness and contentment of life (Zufriedenheit des Lebens). In Kant’s opinion, we can find an evidence for it in the empirical fact that highly cultivated persons who possess extraordinary mental capabilities tend to show a certain misology, i. e. a hatred of reason. When these people retrospectively reflect on the gains and losses of having cultivated their intellectual abilities within their biographies, they typically come to the conclusion that this development didn’t lead them to a larger amount of happiness, but to an increase of hardship. So they feel even envy for ordinary people and their non-intellectual, purely sensual way of life. Kant affirms that there is some truth in their judgment, but not in the way they put the problem: in fact, these people involuntarily validate the thesis that our practical reason is not made to bring about happiness. Hence, what Kant claims, is that their conviction is “covertly founded on the idea of another aim for their existence, possessing much greater dignity”. (d) Kant hereafter develops his own, positive view of the teleology of human nature (GMS, 396,14–37). According to him, the “true vocation” of practical reason is not “to produce volition as a means to some other aim, but rather to produce a will good in itself”. The decisive argument in Kant’s account is founded, as in section (b), on the

Kant on Ends in Nature and in Human Agency

47

principle of suitability: nature has endowed us with practical reason, and this endowment would not be quite suitable for our allegedly fundamental demand for happiness. Reason, he explains, should rather be understood as being appropriate to influence the will. The purpose of nature cannot be to fulfill our needs and interests, “since reason is not sufficiently effective in guiding the will safely in regard to its objects and the satisfaction of all our needs.” There is one further point in this passage: Kant claims that a good will is not “the single or entire good”; it is only the “highest good” in the sense of being the “condition for all the rest, even for every demand for happiness”. Apparently he thereby acknowledges that happiness is a natural end for humans, but only a subordinate one. The strife for happiness is a natural fact of human life, but it has to be put under the limiting conditions of morality. The entire passage under consideration leaves us with an impression of opacity, indeterminacy and vagueness, and it seems to be full of implicit assumptions which we would reasonably expect a defense of. Moreover, it doesn’t fit smoothly into the argument of the First Section; it rather seems to be a sort of digression. One cannot see what it may contribute to the ‘analytical’ line of thought which is intended by Kant to lead us from the concept of good will to that of duty and finally to a first formula of the categorical imperative. In its literary form, our text might rather be considered to be a polemical insertion directed against some of Kant’s adversaries unknown to us. Which point in the present text is of such an importance that Kant placed his teleological considerations in this prominent context? In the following three paragraphs, I would like to deal with the teleological passage by raising some basic problems concerning it. Firstly, I will try to evaluate the range and persuasiveness of the argument provided by Kant. The second part will be dedicated to the question of how our passage should be situated within Kant’s overall thinking about teleology. And thirdly, I will turn to the problem of which function is conceded, in the GMS, to human happiness given that it does not obtain the first place among human goals. I The line of argument advanced by Kant in the teleological passage of GMS I seems to be extremely questionable. At least, we have to reproach him for the very elliptic character and the unclear form of his argument. But I think there are many other problems to be solved.

48

Christoph Horn

As an interpreter of section (a), one has to make sense of the fact that Kant, as I already mentioned, surprisingly hopes to support his view on good will by such an unclear argument anyway. Concerning the crucial section (b), some important questions one would like to ask Kant may be these: What precise type of teleology is it that he has in mind? How can Kant hope to defend his idea of good will by a kind of doctrine which must be, in his own eyes, obscure and illegitimate? Why does he introduce such serious and far-reaching claims in such a superficial and provisionary way? How would Kant legitimize what we called the principle of suitability? Apparently, he does nothing at all here to make it plausible. Further questions are: Is it really convincing to say that the natural dispositions of each animal are indications of the goal or end towards which it is directed? What may be a typical example for such a natural goal? Who or what constituted this goal? Do there exist biological species for the sake of this end? Are men determined by nature to follow these alleged ends? Or is it rather a sort of natural inclination which humans feel with regard to such ends? Hence, are we capable of choosing to act contrary to nature, and is Kant demanding us not to do so, but to follow our ‘true nature’? If that were the case, we would have to accuse him of committing a naturalistic fallacy. Next, we have to raise the question of which natural dispositions of biological species may be regarded, from a teleological point of view, as relevant or telling. What does count here as important? Would it be persuasive to ask why men possess a cecum? It is a well-known problem of Aristotelian teleology that in describing biological species, e. g. humans, it takes into consideration no more than few selected features, called ‘proper features’ (idia) and leaves out all the others.2 Apparently, there are more or less dysfunctional features or dispositions to be found in biological species. But given this fact, couldn’t it also be true that human reason has been generated by nature without any goal, namely as being produced at random or as a feature which has been useful under former environmental conditions which nowadays no longer persist? To put the same point in a slightly different way: Why should one consider humans as a biological species having this or that set of natural characteristics? Wouldn’t it be more attractive to separate the biological and cultural features of humanity, and to put practical reason primarily on the second side? As a consequence of this, doesn’t it seem more persuasive to assume that human goals always exist in a culture-dependent context and that all are, 2

Cf. the new comprehensive monograph of Johnson (2005).

Kant on Ends in Nature and in Human Agency

49

consequently, relative to it? In order to raise the problem the other way round: Couldn’t one argue that human reason is a very helpful and perhaps even the most powerful of the natural instruments which allow to pursue someone’s happiness? At least for purposes of selfpreservation and welfare – to use the examples which Kant himself gives – practical reason seems to provide the best basis available, as is shown by the success of humanity among biological species. But with regard to our passage, there are even more problems to be solved. It remains unclear which idea of happiness Kant presupposes when he rejects to placing it on the first rank of human goals. Does he simply think of a hedonistic concept? At first glance, this seems plausible since he assumes that instincts are the best way to reach the happiness in question. But is it persuasive to say that all of what we wish or what we are striving for can be described in hedonistic terms? Furthermore, when Kant claims that practical reason does not contribute to the type of happiness presupposed by his adversaries – is that more than a caricature of any traditional philosophical concept of happiness? May it be convincing to characterize, e. g., the ancient idea of ‘flourishing life’ (eudaimonia) in such a way that reason doesn’t matter in it? And why should we take it for granted that reason is an impediment for happiness? It would be necessary to provide an argument for why practical reason cannot improve someone’s state of happiness and satisfaction, since it seems convincing to maintain that strong instincts together with a form of strategic reason do a good job in pursuing someone’s satisfaction. Why does practical reason, according to Kant, obstruct our happiness anyway? Additionally, regarding section (c), there exists a manifest problem with the distinction between intellectuals (those who are said to tend towards a sort of cynicism concerning reason) and ordinary people who are said to be directed towards sensual pleasures. As the second group of persons confirms, there is, in Kant’s eyes, the possibility to lead one’s life untroubled by the demands of reason. By this distinction, however, he seems to concede more than he prudentially should: From his point of view, Kant must not acknowledge that there are persons who entirely rely on their instincts and at the same time gain a certain degree of happiness, as long as he maintains that mankind is teleologically determined by practical reason whose impulse goes to the direction of bringing about a good will. What Kant should have claimed is that nobody can evade the impact of practical reason, even if it may seem as if certain persons succeeded in doing so. Moreover, there is a problem raised by D. Schönecker and A. W. Wood who

50

Christoph Horn

rightly indicate that Kant is in a sort of dilemma: Kant himself permits that reason can be appropriately used as an instrument to support our needs and interests, but at the same time, he suggests it to be a non-instrumental power which exercises a profound moral influence on human will. 3 Concerning section (d), there are some further questions to be added. The most important one may be: why should the influence of practical reason generate the good will which has been characterized, some pages before, as the only good without restriction? What kind of impact does Kant have in mind when he claims that practical reason is able to produce a good will? One of the difficulties which this claim raises is how a good will can be said to be a goal anyway. Is it the kind of object someone’s actions can be directed to in the same sense as it is the case with happiness? Can it serve as a goal in the way in which the majority of ancient and medieval philosophers meant happiness to be a goal, namely as a ‘final end’? Suppose that this is what Kant wants to tell us. How can, then, our natural dispositions indicate such a moral end? If it were correct to say that a naturalistic argument is not appropriate to challenge morals, why should it be suitable to support a moral point of view? A further question of some importance is why Kant takes into consideration only these two possible fi nal ends, as if such an alternative was exhaustive, and why he takes them to be, at least in this sense, mutually exclusive. We find no explicit justification for these claims in our text. The list of problems seems so overwhelming that there remains little hope to making sense of our passage. A simple strategy to acquit Kant on all these points is, now, to suppose that he did not ground his teleological considerations on his own philosophical position, but reacted on the convictions of some unnamed adversaries. By this line of interpretation, it would be superfluous to ascribe to him such a strange position. At least prima facie, it looks promising to claim that Kant simply argues on the basis of a view adopted in this context and for the sake of argument. At the beginning of our text, Kant seems to say that he is merely testing the idea of a good will in the light of teleological assumptions: “Hence we will put this idea [i. e. that of a good will] to the test from this point of view [i. e. teleological thinking]” (GMS, 395,1–3). Taken this way, what we have in our text would be an argument ad personam, not Kant’s own position. And in fact, if one compares our text with some passages to be found in his handwritten 3

Cf. Schönecker and Wood (2002, 54).

Kant on Ends in Nature and in Human Agency

51

reflections and in the transcripts of his lectures, it may seem as if Kant responded to a precisely identifiable group of 18th century philosophers, namely to naturalistic or materialist authors such as Helvétius, LaMettrie, and Mandeville who in his eyes rejected precisely what he meant by a ‘good will’ and did this by adopting teleological arguments. This view has been suggested by M. Forschner.4 But on a closer look, this line of interpretation can be excluded. If one takes into consideration other Kantian texts on teleology, one will find again all of the substantial claims of our passage. The central elements are present, e. g., in the ‘Third proposition’ of the Idea toward a Universal History with a Cosmopolitan Aim, a text published in 1784, just one year before the Groundwork. In considering the purposes of nature regarding mankind, Kant there unmistakeably maintains the validity of the principle of suitability. 5 Furthermore, he claims that humans are determined by nature to pursue their happiness independently of their instincts, namely by activating their reason. Nature equipped men only with a minimal set of means in order to force them to develop their rational capabilities. The final aim of nature is, according to Kant, to urge humans to develop their rational freedom. All achievements of mankind should be, according to nature’s purpose, the result of its own efforts. Kant says: The invention of his food, his shelter, his external security and defense (for which it [i. e. Nature] gave him neither the horns of the bull nor the claws of the lion nor the teeth of the dog, but only hands), all pleasure which can make life agreeable, no less insight and cleverness and even the goodness of his will should entirely be his own handiwork. (IGA, 19,29–34)

As the quotation shows, it is again Kant’s opinion that, according to the plan of nature, goodness of will is produced by human insight. We have no reason to think that this opinion functions as an argumentum ad personam in the Idea. What is different here compared with our text from the GMS, is that Kant does not stress the fundamental distinction between the things that man has brought about by reason: human successes in producing food and shelter, in protecting oneself and in attaining pleasure count as a rational achievements in 4

5

This and other interesting accounts of the sources and implicit presuppositions of our passage are provided by Forschner (1988 and 1989). Cf. Ak. 8.19,23–4: “Nature does nothing in vein and is not sumptuous in using the means towards its ends” (Die Natur tut nämlich nichts überflüssig und ist im Gebrauche der Mittel zu ihren Zwecken nicht verschwenderisch).

52

Christoph Horn

the same way as the generation of the moral attitude does. Leaving the differences between our two texts aside, it is remarkable that the text quoted from the Idea confirms that Kant considers nature to be a kind of quasi-divine power developing a rational plan for all biological species and providing them – even if in an extremely parsimonious way – the means necessary for their survival and their happiness. Before we come to the question of Kant’s overall view of teleology, I would like to re-examine the problem of the possible adversaries of our passage in the GMS. As I already mentioned, one crucial problem of our text consists in the fact that it doesn’t really contribute to the analytical line of argument which leads, in GMS I, from “common rational moral cognition”, centered around the concept of a good will, to the “philosophical moral cognition”, focused on the idea of a categorical imperative. Now, one of the possible explanations for Kant’s introduction of teleology could be that he thought of certain adversaries which he wanted to refute in this very context. Who may be the philosophers that Kant perhaps had in mind? What makes this question difficult to answer is that these adversaries have to defend two different claims which do not fit very well together. The first is that nature does nothing in vain, since it is a divine order, and that our natural dispositions are indicators of our natural goal. The second is that man is determined by nature to strive for happiness which seems to be understood in a hedonistic and instinct-guided way. Whereas the former seems typical for intellectualist conceptions of teleology such as that of Aristotle and the Stoics, the latter belongs to the tradition of Epicurean philosophy. But as is well known, the former is not hedonistic, and the latter is strictly a-teleological. Therefore Forschner’s solution which I already mentioned regains, at least temporarily, a certain attractivity. It may seem as if Kant had in mind the Neo-Epicurean bon sauvage-philosophers Helvétius, LaMettrie, and Mandeville. There is, however, a major obstacle for Forschner’s identification of Kant’s adversaries. When Kant characterizes the kind of happiness which they have in mind he makes use of the terms ‘preservation’ (Erhaltung) and ‘welfare’ (Wohlergehen, GMS, 395,8–9). If these authors were hedonists, Kant should have spoken of pleasure or lust. Preservation and welfare are not subjective mental states, but clearly objective goods. The absence of a hedonistic and subjective vocabulary and the occurrence of the key word ‘preservation’ makes it highly plausible that the actual Kantian adversaries are the Stoics. The ancient Stoics described self-preservation (sustasis heautou, conservatio sui) as one of the provisional ‘primary goods of nature’,

Kant on Ends in Nature and in Human Agency

53

and they put it into sharp contrast to Epicurus’ ‘cradle argument’ which tried to identify pleasure as the first natural good.6 It’s Kant’s own concept of happiness which can be characterized, at least partly, as being hedonistic and subjective, as we will see below (in part iii.). For example, shortly before our passage, he describes happiness as “contentment with one’s own condition” (GMS, 393,15–6). Two further indications for a Stoic background of Kant’s observations lie in his concept of an ‘instinct’ 7 and in the constatation that the persons feeling misology are “by no means morose or ungrateful toward the kindness of the world’s government” (GMS, 396,8–9). The expression ‘world’s government’ is apparently Stoic and does not make any sense with regard to the a-theological authors of the 18th century. If we accept that the philosopher who hates reason is, according to Kant, Rousseau8 (and perhaps authors like Helvétius, LaMettrie, and Mandeville)9, the material collected by Forschner makes a new and better sense. What we see, then, is that Kant partly defends his contemporaries against Stoic intellectualism without sharing the hedonistic naturalism of these philosophers. Kant thus opposes the kind of teleology proposed by the Stoics who regarded reason and a rational version of happiness as the ends towards which human nature is directed. Kant’s strategy hence is a double one: On the one hand, he accepts the idea that human reason, like any other natural feature of mankind, possesses a teleological significance according to what we called the principle of suitability. Insofar as this is the case, he shares the Stoic view. What he, however, tries to make plausible, against the Stoics, is that the real purpose of human reason is not happiness, but the generation of a good will. Kant here somewhat illegitimately mixes his own hedonistic and subjective concept of happiness with the Stoic one which is intellectualist and objectivist; this last one speaks of preservation and well-being as provisionary goods.10 On the other hand, he accepts the view of the a-theological Naturalists who reject 6 7

8

9 10

Cf. e. g. Chrysippus, SVF III, 178. 181. 184. Instinctus is the Latin standard translation of the Greek hormê which plays a fundamental role in Stoic theory of natural impulses. Rousseau claims in the Discours sur l’origine et les fondements de l’inégalité as well as in the Émile that a full degree of happiness has only been real for our uncivilized ancestors in a pre-rational state. For the passages in Rousseau which Kant might have had in mind see J. Ferrari (1979, 171–88). In this respect, I follow Reich (1935) and Himmelmann (2003, 130–2). According to the KpV, the Stoic view of happiness is that the wise man attains it “by the consciousness of the moral way of thinking” (im Bewusstsein der sittlichen Denkungsart: KpV, 127,14–5). Kant misinterprets Stoic eudaimonism in terms of a subjective fulfillment of desires.

54

Christoph Horn

reason insofar as it is a useless instrument to gain happiness. Kant takes them as witnesses for the falsity of the Stoic kind of teleology, but he immediately adds that there exists a more valuable kind of destination for human reason: the rise of morality. One expression which should be taken seriously as a signal is the word misology. It hints at a precisely identifiable historical source, namely Plato’s Phaedo (89d1 and d4). In Plato’s account of misology, the hatred of those contemptuous of rationality is caused by their own former love of reason and the disappointment which results from a rough and uncritical use of it. To despise rationality is thus a consequence of an intense, but disappointed enthusiasm for it. Somebody may become an enemy of arguments in the same way in which one can take the identity of a misanthrope: namely by putting too much trust in unchecked arguments, as another person may practice a naive form of philanthropy until she becomes repeatedly disappointed. Kant seems to have read the Phaedo inspired by Moses Mendelssohn, who published his own writing Phaedo or On the Immoratlity of Soul in 1767 as a selection and amplification of Plato’s dialogue. What is particularly interesting is the observation that not only does the expression misology stem from Plato’s Phaedo, but perhaps also the principle of suitability. The Phaedo seems to be the first text in Western philosophy which develops the idea of teleological explanation of nature. There, it is Socrates who reports his deep disappointment when he read as a young man the treatise of Anaxagoras whom he expected to develop a philosophy of nature according to teleological principles. As the Platonic Socrates says: And I rejoiced to think that I had found in Anaxagoras a teacher of the causes of existence such as I desired, and I imagined that he would tell me first whether the earth is flat or round; and then he would further explain the cause and the necessity of this, and would teach me the nature of the best and show that this was best; and if he said that the earth was in the centre, he would explain that this position was the best, and I should be satisfied if this were shown to me, and not want any other sort of cause. And I thought that I would then go and ask him about the sun and moon and stars, and that he would explain to me their comparative swiftness, and their returnings and various states, and how their several affections, active and passive, were all for the best. For I could not imagine that when he spoke of mind as the disposer of them, he would give any other account of their being as they are, except that this was best; and I thought when he had explained to me in detail the cause of each and the cause of all, he would go on to explain to me what was best for each and what was best for all. (Phaedo 97d5–98b3; transl. by B. Jowett).

Kant on Ends in Nature and in Human Agency

55

For the young Socrates, it is Universal Reason which structures the world according to the principle that everything is organized in the best way available. Unfortunately, Anaxagoras did not do him the favour to provide a philosophy of nature based on this fundamental principle. We are not precisely told what it could possibly mean that something is organized in such a perfect form. But we find, in Plato’s text, a clear-cut opposition between mechanistic causes which do not even count as causes and teleological explanations which alone are seen as acceptable. In our terminology, Socrates distinguishes between necessary and sufficient causes in explanation and assigns mechanistic causes to the former kind, teleological causes to the latter (Phaedo 99a4–b8; transl. by B. Jowett): There is surely a strange confusion of causes and conditions in all this. It may be said, indeed, that without bones and muscles and the other parts of the body I cannot execute my purposes. But to say that I do as I do because of them, and that this is the way in which mind acts, and not from the choice of the best, is a very careless and idle mode of speaking. I wonder that they cannot distinguish the cause from the condition, which the many, feeling about in the dark, are always mistaking and misnaming. And thus one man makes a vortex all round and steadies the earth by the heaven; another gives the air as a support to the earth, which is a sort of broad trough.

For our purpose it is interesting that the whole ‘autobiographical’ passage in the Phaedo seems to be devoted to the introduction of a new and revolutionary way of explaining the natural world: by teleological explanation.11 Apparently, there is little distance between this Socratic teleological thought and the principle of suitability which Kant formulates as follows (GMS, 395,4–7): In the natural predispositions of an organized being, i. e., a being arranged purposively for life, we assume as a principle that no instrument is to be encountered in it for any end except that which is the most suitable to and appropriate for it.

Nevertheless, we find one major difference between the principle developed in the Phaedo and the Kantian formula: Kant restricts his principle to living organisms. These are said to have instruments which are organized in the best way possible. There is no such restriction in Plato’s writings and his Socrates does not confine the principle 11

See e. g. the commentary of Gallop (1975) ad 97b8–98b6: “The present passage marks the transition from a mechanistic to a teleological conception of the natural order that was to dominate European science for the next two thousand years.”

56

Christoph Horn

to living beings. Yet, we also find a striking similarity: As in Plato, we find in Kant the claim that living organisms are purposively organized in a way that is best for a certain objective end. Kant does not mean that they are perfectly organized for survival or reproduction, i. e. for their own best; apparently, his kind of teleology is the objective one inaugurated in the Phaedo. II It may be helpful now to clarify the most relevant kinds or aspects of teleology before characterizing Kant’s observations in GMS I. One fundamental dichotomy is that between a universal teleology which is based on the thesis that the entire universe shows a coherent sort of purposiveness – directed towards one comprehensive end – and special or individual forms of teleology which claim that there exist final ends only for certain (natural) species or individuals respectively. It is typical for traditional theological approaches to defend the first version, and typical in many early modern positions to find the second form of teleological thinking. Concerning our passage in the Groundwork, we are confronted with the first form; Kant is apparently speaking of a teleological design of the entire universe. Furthermore, it makes a considerable difference if someone defends an external teleology, maintaining that, in the outer world, an entity or event exists or happens for the sake of something else, e. g., that moonshine is made for lovers, or an internal teleology which concerns the inner structure, function, or organisation of something, especially that of an organism; so we might say that the heart exists for the sake of blood circulation and thereby for the survival of animals. In § 63 of the Critique of the Power of Judgment, Kant distinguishes between an external purposiveness (äussere Zweckmäßigkeit) which he calls ‘relative’ (since it happens between two entities) and an internal one which signifies a teleological structure within an entity. To explain this innere Zweckmäßigkeit, he famously introduces the concept of an organism (or as he says, an organisirtes Product der Natur) that possesses such an internal purposiveness. An organism shows an internal structure of reciprocity and interdependence in which every part is in the same time means and end, cause and effect with regard to all other parts (cf. especially KU § 66). Nevertheless, it is obvious that Kant, in our passage, is not thinking of this intriguing phenomenon. His issue here clearly is based on external teleology.

Kant on Ends in Nature and in Human Agency

57

One should not mistake this last distinction for that which exists between extrinsic and intrinsic forms of teleology: An entity or event can be said to be a passive or insensitive part of a teleological process since it may not dispose of any inner perspective or internal dimension (regardless whether it is a means or an end of the process), or it can be described as having an inner principle of striving or aiming at a goal. Darwinism explains teleological elements of nature in the first way: e. g. a skeleton of an extinct species may help a palaeontologist to decide if the animals under consideration lived in terrestrial or aquatic habitats; the teeth may provide evidence of whether these animals were carnivores or herbivores. The Darwinist background conviction is that animals are highly suited to meet the challenges of their specific environments; but of course, adaptation is nothing but a result of arbitrary mutation and selection under competitive conditions. Aristotelianism traditionally assumes that natural entities, especially animals, follow internal impulses connected with their essence which stimulate them to go for their natural goals, whereas Darwinism even interprets inner impulses as consequences of an adaptation to the outer environment. Nearby this dichotomy lies another one of no less importance: One may discern cases of intentional teleology, when we attribute to an entity a mental, conscious, or even rational anticipation of its goal, from cases of non-intentional or functional teleology. Instead of intentional teleology, one could also speak of personal teleology. But perhaps it would make more sense to reserve this distinction for the dichotomy between a teleological structure designed by a personal, intelligent demiurge or creator (as in Plato’s Timaeus) and an impersonal cause of purposiveness in nature. Of course, this last distinction is that between theological and a-theological forms of teleology. An additional distinction of high importance for Kant is that between an objective and a subjective concept of purposiveness: Kant calls a teleological relation ‘objective’ if it has an object as its end or target. It does not play any role if the object under consideration is generated or if it is only modified or affected by the effect directed on it. Accordingly, he characterizes a teleological structure as subjective if there is no orientation towards an object but only an effect on the subject which feels the purposiveness of its organization. This is the case in aesthetic experience where the spectator of art or nature is, following Kant, confronted with the well-organized structure of his cognitive abilities. And a final Kantian dichotomy is that between formal and material (or real) purposiveness. We are confronted with a formal sort

58

Christoph Horn

of Zweckmäßigkeit, if we reflect on the mere form of something – e. g. if we appreciate something imagined independently of its real existence. Material kinds of purposiveness are those in which there exists a causal relationship between a concept and its object.12 To apply these concepts, I think we can subsume our passage in Groundwork I into the categories of universal and external teleology whereas we do not know so far if Kant’s present use of teleology is extrinsic or intrinsic, intentional or non-intentional, and personal or impersonal. What we know is that he accepts both objective and subjective forms of purposiveness. Now, what causes most of the serious irritations that we spelled out in section (i) is the fact that Kant apparently vindicates a universal and external form of teleology. In the history of philosophy, this approach is typically connected with intrinsic, intentional, personal, and objective ways of teleological thinking. But it was Kant who famously refuted such a form of speculative metaphysics which existed up to his days in the Leibnizian and Wolffian schools, particularly in the ‘physico-theological’ argument for the existence of God. The relevant passages in the Critique of Pure Reason (A620 ff./B648 ff.) and the Critique of the Power of Judgment (§ 85) are unambiguous and of no less force than Hume’s rejection of the argument in the Dialogues Concerning Natural Religion. In the fi rst Critique Kant strongly opposes an analogy between the intentional and personal production of artifacts and the creationist idea of a functionally designed world by using the words: “Reason would be unable to satisfy her own requirements, if she passed from a causality which she does know, to obscure and indemonstrable principles of explanation which she does not know” (A626/B654, transl. J. M. D. Meiklejohn). On the other hand, Kant admits, particularly in the third Critique, a remaining importance of teleological thinking even beyond the internal purposiveness of organisms. In the later text, it is obvious that Kant does not defend an intrinsic, intentional, and personal teleology. Let us suppose, then, that Kant did not hold a position in 1785 which we rejected in 1781 and 1790. What I left aside so far is the distinction between the three main objects of traditional teleological thought in philosophy: these are nature, history, and human agency. For Kant, all three topics of teleology are of major importance, and he widely contributes to them in a series 12

A useful overview on Kant’s concept of teleology and purposiveness is provided by Frank and Zanetti (1996, 1184–204). Marc-Wogau (21938, 71) differentiates between ten aspects of Kant’s concept of purposiveness.

Kant on Ends in Nature and in Human Agency

59

of texts of his critical period. With regard to the first, Kant feels it necessary to deal with natural teleology since, with regard to nature, the common spectator (not only the pre-modern one) gets the impression that everything is well-ordered and wisely coordinated: every natural species possesses its ‘ecological niche’, i. e. its separate habitat, its food and shelter, its ways to defend itself or to get itself to safety, its way of reproduction and of raising its descendants; even if one species is instrumentalizing the other for survival, no one seems to definitively destroy the other. Also humans are provided by nature with all the capacities, skills, and means which are indispensable for their lives. But even more impressive is the inner structure of an organism: as we already saw, Kant describes it as a structure of reciprocity and interrelatedness. It is a whole organized under one concept where the parts are interdependent towards each other as being both and simultaneously means and ends, causes and effects. Do all these observations provide evidence for a highly organized universe? No, for Kant, the teleology of nature isn’t something objective and real. But does that mean that it must be traced back to the ways in which we see and interpret nature? Does it reflect a merely subjective way of explanation following our way of understanding human behavior? Kant claims that teleology in nature lies beyond the dichotomy of objectivism and subjectivism. He considers organisms as purposes and thinks that man as the end for the sake of which universe has been designed without neglecting that organic life and humans are natural species among other natural species, endowed with some extraordinary abilities, but lacking of many others. It is even harder to answer the second question: Are we justified in assuming purposiveness in human history? Are most of the historical events and processes contingent ones, or do we find convincing examples for what one might call ‘meaning in history’, i. e. for comprehensive structures and ordering principles? Is there some sort of divine providence at stake in history, or at least certain lines of development and progress? The spectator may come to deeply different conclusions: She may get the skeptical impression first formulated by Thucydides that, since human nature is constant and changeless, historical disasters and successes, catastrophes and achievements will always follow the same rules. Or she will share the enlightenment idea of a moral progress or cultural development, at least in the centuries of which we have historical knowledge. Undoubtedly, for Kant, there is some evidence for a progress towards a human progress in society, economics, national and transnational politics etc. He thinks that nature is organized in a way that forces humans to develop their talents and

60

Christoph Horn

capabilities. What makes our impression so profoundly ambiguous is that we seem to have enough reason to defend both anthropological pessimism and optimism. Kant tries to solve the problem by assuming that evil conditions in nature and human history are destined to bring about civilization and, as a last consequence, even morality. It is surprising to see how far he is going in this direction: e. g. even the existence of mosquitoes and other pests is interpreted as an impulse for human development: Thus one could say, e. g., that the vermin that plague humans in their clothes, hair, or bedding are, in accordance with a wise dispensation of nature, an incentive for cleanliness, which is in itself already an important means for the preservation of health. Or the mosquitoes and other stinging insects that make the wilds of America so trying for the savages are so many goads to spur these primitive people to drain the swamps and let light into the thick, airless forests and thereby as well as by the cultivation of the soil to make their abode more salubrious (KU, § 67; 379,22–31).

Only in the third area of traditional teleology, that of human agency, does there seem to be no difficulties – at least at first glance – since we usually have no problem in presupposing the intentionality of human behavior. But even if one leaves aside the precarious questions concerning the freedom of will, there remain serious challenges connected with this topic. Practical teleology can be understood not only in the simple way of attributing intentionality to humans acts, but also in a very demanding and far-reaching sense according to which the agency of rational individuals follows an implicit order structured by general or final ends (particularly happiness or pleasure), necessary means and obligatory ends (mostly moral ones, but also instrumental ones). Teleological ethics in this sense has been the predominating form of moral philosophy in the pre-modern history of thought: most ancient and medieval philosophers defended this type of practical teleology. Especially in the Critique of Practical Reason, we fi nd passages in which Kant discusses this tradition sympathetically, even if, at first glance, he seems to criticize it. As has been shown by Barbara Herman and many others in recent years, Kant supports practical teleology with regard to some of its basic assumptions, e. g. that each agent has to value every option she is choosing as something good and every option she is avoiding as something evil.13 Kant indeed assumes that happiness is not an arbitrary end; he takes it to be the final 13

Cf. KpV, 59: nihil appetimus nisi sub ratione boni, nihil aversamur nisi sub ratione mali. See on this point Herman (1993) ch. 9.

Kant on Ends in Nature and in Human Agency

61

and comprehensive purpose of human agency.14 But on the other hand there are passages in which rational freedom or autonomy seems to be summum bonum, as Kant formulates explicitly.15 Be that as it may, if human agency is a well-structured whole of non-arbitrary fi nal ends and necessary means, it is tempting to think there may even exist a connection to the teleologies of nature and history. Now, what is the subject matter of GMS I, 394–6? Apparently, what is lacking is historical teleology. But we are confronted both with natural teleology and with considerations concerning human agency. As soon as this thematic ambiguity becomes clear, one can express the most fundamental difficulty of our passage as follows: what it seems to make so extremely weak is its connection (and, as one might think, confusion) of natural and personal teleology. Natural teleology (if legitimate) is universal and external, whereas personal teleology is individual and internal. The first one may be judged as a pre-modern, theological relic in our common consciousness without relevance for scientific explanations of nature, whereas the second is relevant, but should be confined to intentional human behavior. Therefore, the crucial question for our further interpretation of the text has to be how the two topics may, in Kant’s eyes, fit together. The situation is worsened by the insight that these two themes of teleology are completely independent from one another and therefore of no reciprocal importance even in the case that both versions are sound and acceptable. Suppose that we had good reasons to accept the principle of suitability: every natural species might be considered as optimally adapted to its objective purpose. Would that have any consequences for agents and the structure of their intentions, goals, and ends? Of course, not. Reasons to act can only be justified by a process of internal considerations, by practical deliberation, not by clarifying the structure of the external world. Whether humans are by nature (or by divine providence) destined for this or that purpose does not matter on the level of an agent’s practical reasons. If it were e. g. highly plausible that the human set of teeth is that of a carnivore species, it would nevertheless incorrect to see in this 14

15

Cf. GMS,415,28–416,1: “There is one end, however, that one can presuppose as actual for all rational beings […] and thus one aim that they can not merely can have, but of which one can safely presuppose that without exception they do have t in accordance with a natural necessity, and that is the aim at happiness. The hypothetical imperative that represents the practical necessity of the action as a means to furthering happiness is assertoric. One may expound it as necessary not merely to an uncertain, merely possible aim, but to an aim that one can presuppose safely and a priori with every human being, because it belongs to its essence.” Cf. Moralphilosophie Collins: Ak. 27/1. 344.

62

Christoph Horn

fact a reason to refuse vegetarianism. Precisely from a Kantian point of view, it would be completely excluded to accept a sort of moral naturalism in which we are told what is morally right or wrong with regard to basic facts of nature – even if they might be beyond any doubt. The even more absurd consequences would be the result if we tried to infer from our intentions, goals, and ends to the sphere of natural teleology. What we are striving for or aiming at needs not to be something that turns out to be, at the same time, fundamental for the basic structure of nature, even if it were convincing to say that these aims and ends are not only arbitrary, but essential ones. It is of crucial importance to distinguish between objective and subjective teleology, the view from outside and the view from inside, and Kant seems to commit the disastrous mistake of bringing these two aspects closely together. What should we make of this precarious situation? Can we rightly accuse Kant of such a profound confusion? Or is it preferable to interpret our text as a well-considered connection between natural and personal teleology? To understand our passage in the Groundwork, we now have to investigate the Kantian attitude towards teleology developed since his writing On the Various Races of Men (1775), continued in the Idea (1784), in Determination of the Concept of a Human Race (1785), in On the Use of Teleological Principles in Philosophy (1788), and finally in the second part of the third Critique (1790). Kant repeatedly raises the question if we are entitled to follow our impression that there exists an objective teleology in nature and in history. What he proposes is a solution which lies in between the alternative of objective and subjective teleology. Nature itself is not an intelligent being, and history is not designed by a divine ordering hand, but nevertheless, we find in it a high degree of suitability, organization, coordination, and providence. As we perceive, lots of entities are in a harmonious or cooperative relationship with each other, many things exist for the sake of others in a sort of hostile and yet balanced relationship, and many evils are suited to generate useful consequences. Despite this impression, Kant claims that it would be inadmissible to use final causes in our explanation of the natural world or history. We are not allowed to transgress the borderlines of our usual scientific explanation, namely to explain things in terms of efficient causes. The use of fi nal causes for the understanding of natural phenomena would be an illegitimate transfer from the sphere of human behavior. Natural suitability is, in contrast to the mechanistic nexus efficiens, not a constitutive or substantial principle in nature. According to his Transcendental Idealism, Kant maintains that we can legitimately make use of final causes only

Kant on Ends in Nature and in Human Agency

63

in the reductive sense of what he calls a ‘regulative principle’. On one hand, a regulative principle is not an arbitrary, but a necessary idea of our reason, i. e. a principle which plays a fundamental role in organizing our experience. On the other hand, its importance for us is not based on the real world existing independently of us. So regulative principles have the special quality that we can neither relinquish them nor interpret them as describing features of reality. Whenever they are used for descriptive purposes, a theoretical situation arises which Kant characterizes as ‘dialectic’ (i. e. illusionary) since it is based on an improper use of an idea which we find in us, but not in the outer world. In which sense do we usually believe that natural phenomena can be described teleologically? After having excluded, in the third Critique, examples of subjective purposiveness, e. g. aesthetical ones, and of a ‘merely formally objective purposiveness’ (bloß formale objektive Zweckmäßigkeit), e. g. of geometrical figures, Kant turns to the concept of organism by which he claims that the ‘objective reality’ of natural purposiveness can be demonstrated (KU, § 65). A description of nature in terms of mechanistic causality fails, according to him, to characterize a structure which, as we saw, is fundamentally teleological. What seems a little confusing is that Kant adds several other aspects which he takes as evidence for the insufficiency of mechanism, e. g. the empirical laws of nature and diverse circumstances in the world which enforce useful reactions among humans. But his decisive point is the organic structure which he takes to be the central example for an entity which exists as ‘natural purpose’ (Naturzweck). In order to exist as a natural purpose, something must be “cause and effect of itself” (KU, § 64). But man is not only a purpose-in-itself insofar as he is an organism. What counts even more is the fact that humans are, as rational beings, subject to moral law whose expressions are strictly commanding precepts. Now suppose that nature and history would leave no room for the commands of the categorical imperative: then morality would be something absurd, since its strict precepts could not be realized by humans who have to act under natural conditions. But man cannot relinquish his idea of morality as long as he takes seriously his status as a rational agent. Such an agent cannot reasonably contest the demands of morality resulting from practical teleology. So there remains only one way open: making compatible nature and human agency. In order to guarantee this compatibility between the ‘realm of nature’ and the ‘realm of morals’, one has to accept Transcendental Idealism. It allows one to accept that both natural teleology and historical teleology can

64

Christoph Horn

be vindicated on the basis of practical teleology. Already in his writing On the Use … (1788), Kant expresses this line of thought. Teleology taken in its practical meaning, i. e. morality, is forced, according to this essay, to realize its purposes. Hence it is not allowed to neglect that the necessary preconditions for this realization must be fulfilled. These preconditions contain both the final causes (Endursachen) and the supreme cause of the world (oberste Weltursache). Transcendental Idealism therefore is able to guarantee the possibility of morals in the natural world.16 As Kant says in the third Critique, man is the singular being in the world which is forced to obey the moral law in following its ends; hence, it must attribute freedom to itself and considers this freedom even as its highest purpose; and, therefore, is a purpose-initself and represents the highest end of all nature: Now we have in the world only a single sort of beings whose causality is teleological, i. e. aimed at ends and yet at the same time so constituted that the law in accordance with which they have to determine ends is represented by themselves as unconditioned and independent of natural conditions but yet as necessary in itself. The being of this sort is the human being, though considered as noumenon: the only natural being in which we can nevertheless cognize, on the basis of its own constitution, a supersensible faculty (freedom) and even the law of the causality together with the object that it can set for itself as the highest end (the highest god in the world). Now of the human being (and thus of every rational being in the world), as a moral being, it cannot be further asked why (quem in finem) it exists. His existence contains the highest end in itself, to which, as far as he is capable, he can subject the whole of nature, or against which at least he need not hold himself to be subjected by any influence from nature. – Now if things in the world, as dependent beings as far as their existence is concerned, need a supreme cause acting in accordance with ends, then the human being is the final end of creation; for without him the chain of ends subordinated to one another would not be completely grounded; and only in the human being, although in him only as a subject of morality, is uncoditional legislation with regard to ends to be found, which therefore 16

On the Use of Teleological Principles in Philosophy: Ak. 8.182,35–183,9: “But since a pure practical teleology, i. e. a morals, is determined to make real its purposes in the world, it will not be allowed to neglect its possibility within it [i. e. the world], both regarding the final causes implied in it, and regarding the appropriateness of the supreme cause of a totality of all purposes as effect, hence a natural teleology, as well as the possibility of a Nature at all, i. e. a transcendental philosophy, in order to ensure that pure practical teleology has objective reality concerning the possibility of the object in practice, namely concerning the purpose whose realization is prescribed by it [i. e. practical teleology].”

Kant on Ends in Nature and in Human Agency

65

makes him alone capable of being a final end, to which the whole of nature is teleologically subordinated. (KU, 435,15–436,2).

To resume, we find two substantial reasons given in the third Critique that Kant’s connection between natural and personal teleology is not an involuntary confusion or mistake, but a well-considered argumentative step. According to Kant, it is (a) the teleological structure of an organism and (b) the teleology of rational agency which renders plausible that man is the end-in-itself for the sake of which nature is entirely designed. Personal teleology thus provides the decisive argument for the acceptance of natural teleology which otherwise would be anachronistic, even if its status is confined to that of a regulative principle. Natural teleology must not be transferred to the level of scientific research or to speculative metaphysics. Historical teleology cannot be taken as an objective law of history. But here we should look again at to our text and raise the question if we discover here, too, any evidence for a Kantian strategy of conscious fusion between the two topics. Let us now return to the Groundwork. Note that only in the First Section of the Groundwork do natural teleology and personal teleology seem to be illegitimately mixed. There are no traces of natural teleology in the rest of the text. In GMS II and III, Kant’s use of the concept of nature is strongly reduced and only appears dependent on or in contrast to personal teleology. On page 421, he expresses the ‘formula of universal law’ in terms of a universal law of nature: “So act as if the maxim of your action were to become through your will a universal law of nature”. In this alternative version of the categorical imperative, Kant demands that one should, by a thought experiment, imagine that one’s own maxims should be valid in the universe as laws of nature. The character of a thought experiment becomes manifest by the words “as if … were” which have no parallel in the formula of universal law. The concept of a law of nature clearly implies here the impossibility to act against it. The agent, thus, is asked to imagine that her maxim might become a rule without exception. Also on page 455 f., we find the usual Kantian contrast of nature – identified with necessity and determinism – and human freedom. When Kant uses examples like that of the mosquitoes (and similar cases in § 67 of the KU), all he wants to claim is that teleology is valid for the reflecting power of judgment (reflektierende Urteilskraft), not for determining power of judgment (bestimmende Urteilskraft). The teleologies of nature and history must not be understood in a ‘constitutive’, but in a ‘regulative’ sense (KU, 379,10–16).

66

Christoph Horn

III Towards the end of our teleological passage, Kant makes this claim about the relationship of good will and happiness (GMS, 396,24–30): This will may therefore not be the single or entire good, but it must be the highest good, and the condition for all the rest, even for every demand for happiness, in which case it can be divided with the wisdom of nature, when one perceives that the culture of reason, which is required for the former, limits in many ways the attainment of the second aim, which is always conditioned, namely of happiness, at least in this life […].

We are familiar with the doctrine of the highest good developed here in nuce from the more detailed version of the Critique of Practical Reason (KpV, 108–14): The summum bonum for human beings must be conceived as a combination of morality and happiness in the sense that morality is the end for the sake of which one should act, whereas happiness is the reward that God distributes in the afterlife according to someone’s worthiness to be happy. According to this view, it would be morally inappropriate if one acted for the sake of his own happiness, but it would disappoint our natural inclinations if we couldn’t fulfill our desire for happiness. By combining morality as the end and happiness as the reward and by setting happiness under the condition of morality Kant escapes from the dilemma that either the demands of moral law or our natural desires are missed. The first mistake is, according to Kant’s account in the second Critique, that of Epicureanism, the second that of Stoicism. In our passage of the Groundwork, we encounter of the first short drafts of this doctrine. It presupposes far-reaching assumptions concerning the understanding of happiness, and it entails a harsh criticism of eudaemonism in moral philosophy. As we see from Kant’s remarks on happiness in the Groundwork, he is fully conscious of these consequences. In his pre-critical period, Kant himself defended a sort of eudaemonism which has been founded on a perfectionist account of human life. We find this account e. g. in his General History of Nature and Theory of Heaven (1755) where Kant claims that human happiness consists in the future “community with the infinite being”; at that point of his development, he believed that human nature is directed towards its perfection. Nature thus prepares man for a perfect existence with God.17 A second stage of his thought on happiness might 17

Cf. Ak. 1.322 and Himmelmann (2003, 10).

Kant on Ends in Nature and in Human Agency

67

be identified in his lectures on anthropology prior to 1777. Following Rousseau’s “sentiment of existence”, Kant there developed an optimistic view of happiness based on the principle of the perfection of pleasure via aesthetic taste; in his account, he tries to unify the Stoic and the Epicurean account of pleasure, as S. Meld Shell (2003) has made plausible. Beginning with his lectures after 1777, and especially in the ‘Doctrine of Method’ of the fi rst Critique, we find the kind of serious criticism of happiness known in more detailed versions from the Groundwork and the second Critique. Kant now interprets happiness as the comprehensive fulfilment of all the wishes and inclinations of a person: “Happiness is the satisfaction of all our desires; extensive, in regard to their multiplicity; intensive, in regard to their degree; and protensive, in regard to their duration” (KrV, A 806/B 834). In his lectures known as Praktische Philosophie Powalski [dating probably from the end of the 1770s] he defines happiness as “satisfaction of all pleasures anyway” (Ak. 25.101). Happiness is thus reduced to its subjective and hedonist components, and it is criticized precisely for those aspects. Kant now characterizes all approaches in moral philosophy which are based on the “general principle of self-love or someone’s own happiness” as “material principles” (KpV, 22,6–8) and claims that they are principally mistaken (5.39). In his late Anthropology (1798), Kant famously maintains that “all eudaemonists are practical egoists” (Anthro, 130). In all of his treatises on moral philosophy dating from the critical period, the concept of happiness plays a thoroughly negative role in Kant’s view. A very pointed statement is that of the ‘Doctrine of Virtue’ where Kant says that “if eudaemonism (the principle of happiness) is adopted as the principle instead of eleutheronomy (the principle of freedom of the inner legislation), the consequence is the euthanasia (quiet death) of all morality” (MS, 378,15–7). At a closer look, there are two different objections that Kant raises against happiness as the key concept of morals. The first depends, as we saw, on his conviction that the pursuit of happiness contradicts the principle of morality. According to Kant, the only way in which an agent can be adequately motivated is when she is acting from the viewpoint of practical reason. Solely practical rationality can both rule out egocentric motives in human agency, and guarantee a stable and permanent form of motivation. The second objection is an epistemological one. It says that all individual opinions concerning happiness are based on experience and are thus vague and transitional; they do not permit of more than a strategic approximation to “a maximum of welfare”,

68

Christoph Horn

since nobody is “capable of determining with complete certainty, in accordance with any principle, what will make him truly happy, because omniscience would be required for that” (GMS, 418,22–4). We can only act “in accordance with empirical counsels, e. g., of diet, frugality, politeness, restraint etc” (GMS, 418,25–7). In GMS I 399, Kant takes both objections together; but in our context, it is the fi rst one that he has in mind.18 But despite his principal rejection of happiness as a practical and theoretical basis of morals, Kant doesn’t contest, neither in our context nor in his general view, the fundamental significance of the human longing for happiness. Happiness is for Kant even “the most powerful and inward inclination” (GMS, 399,8). As he underlines in the second edition of the Groundwork, happiness is an “aim that they [i. e. all rational beings] not merely can have, but of which one can safely presuppose that without exception they do have it in accordance with a natural necessity” (GMS, 415,30–2). How can we describe the kind of synthesis which Kant proposes in order to connect the two antagonistic concepts of morality and happiness? Roughly spoken, there exist four strategies in moral philosophy to interpret the relationship of morality and happiness: [a] an identity thesis, [b] a reconciliation thesis, [c] a conflict thesis, and [d] an incompatibility thesis. Kant himself discusses thesis [a] in the Critique of Practical Reason and characterizes it by its assumption that happiness and virtue are ‘analytically’ connected to each other, i. e. connected by conceptual implication (KpV, 113–4). But he rejects [a] both in its Epicurean form (where moral virtue is an implication of happiness being a necessary instrument to it) and in its Stoic variant (where the attainment of full moral virtue is simply identified with the state of perfect happiness). If a philosopher, as Kant does, accepts that both morality and virtue are irreducibly founded in human nature, it clearly follows that he cannot adopt the incompatibility thesis [d]. Hence we would expect that he had to choose between [b] and [c]. But Kant’s position is a double one: His doctrine of postulates advanced in the second Critique amounts 18

Both objections don’t do justice to ancient (or medieval) Eudaemonism. As has been shown by, e. g., Irwin (1996) and Weidemann (2001), Kant firstly neglects the fact that ancient eudaemonism is based on an objective account of happiness, and is hence able to discuss the necessary and sufficient conditions in an intersubjectively comprehensible way. Secondly, Kant ignores that moral virtue doesn’t figure, at least in most of these accounts, as a means which is instrumentalized towards a selfish form of happiness. On the contrary, adopting virtue implies the transformation of a moral personality. Furthermore, it doesn’t seem adequate to characterize ancient Eudaemonism generally in terms of hedonism.

Kant on Ends in Nature and in Human Agency

69

to the reconciliation thesis [b], whereas his remarks on the principal conflict between morality and happiness clearly support the conflict thesis [c]. From the perspective of a postulated afterlife, the demands of morality and the longing for happiness can be reconciled. This is, to be sure, a quite extravagant strategy to bring about such a reconciliation. But basically, Kant defends the thesis [c]: Morality is in a deep conflict with happiness with cannot easily be resolved. In the Critique of the Power of Judgment, we find a statement that lies precisely in the line of argument which we know from our teleological passage in the Groundwork: It is easy to decide what sort of value life has for us i fit is assessed merely by what one enjoys (the natural end of the sum of all inclinations, happiness). Less than zero: for who would start life anew under the same conditions, or even according to a new and self-designed plan (but one still in accord with the course of nature), which would, however, still be aimed merely at enjoyment? It has been above what value life would have if conducted in accordance with the end that nature has set for us, which it contains in itself, and which consists in that which one does (and not merely what one enjoys), where we are, however, always merely a means to an undetermined final end. Thus nothing is left but the value that we ourselves give to our lives through that which we do not merely do but also do so purposively and independently of nature that even the existence of nature can be an end only under this condition. (KU, 434 footnote).

The value of life would be thoroughly destroyed if one would describe it in terms of happiness interpreted as the satisfaction of desires. According to Kant, no one would choose, if he could, to lead a hedonistic life for a second time. The value of life is based on what can be actively done by an agent and, as Kant continues, by the sort of action that confers purposiveness to nature.19 Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–) (abbreviated as ‘AA’). The Critique of Pure Reason will be cited according to the A/B pagination from the first and second editions. 19

I am grateful to Dieter Schönecker for his valuable comments on an earlier version of my paper and to Alexander Cotter for checking my English.

70

Christoph Horn

All quotations of the Grundlegung are taken from Immanuel Kant, Groundwork for the Metaphysics of Morals, edited and translated by Allen W. Wood, New Haven/London 2002: Yale University Press. All other textual references are to The Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press, 1992-). Anthro Anthropologie in pragmatischer Hinsicht, AA VII GMS Grundlegung zur Metaphysik der Sitten, AA, IV IGA Idee zu einer allgemeinen Geschichte in weltbürgerlicher Absicht, AA VIII KpV Kritik der praktischen Vernunft, AA, V KrV Kritik der reinen Vernunft, AA III, IV KU Kritik der Urteilskraft, AA, V MS Metaphysik der Sitten, AA, VI

Other works Ferrari, Jean (1979): Les sources françaises de la philosophie de Kant. Paris. Forschner, Maximilian (1988): “Moralität und Glückseligkeit in Kants Reflexionen”, in: Zeitschrift für philosophische Forschung 42, 351–370. Forschner, Maximilian (1989): “Guter Wille und Haß der Vernunft”, in: O. Höffe (ed.), Grundlegung zur Metaphysik der Sitten. Ein kooperativer Kommentar, Frankfurt a. M., 45–65. Frank, Manfred and Zanetti, Véronique (1996): Immanuel Kant. Schriften zur Ästhetik und Naturphilosophie, Frankfurt a. M. Gallop, David (1975): Plato, Phaedo. Translated wit Notes, Oxford. Guyer, Paul (2000): Kant on Freedom, Law, and Happiness, Cambridge/New York. Guyer, Paul (2002): “Ends of Reason and Ends of Nature: The Place of Teleology in Kant’s Ethics”, in: The Journal of Value Inquiry 36, 161–186. Herman, Barbara (1993): The Practice of Moral Judgment, Cambridge/ Mass. Himmelmann, Beatrix (2003): Kants Begriff des Glücks, Berlin/New York. Irwin, Terence H. (1996): “Kant’s Criticism of Eudaemonism”, in: S. Engstrom/J. Whiting, (edd.), Aristotle, Kant, and the Stoics. Rethinking Happiness and Duty. Cambridge, 63–101. Johnson, Monte Ransome (2005): Aristotle on Teleology, Cambridge. Korsgaard, Christine M. (1996): Creating the Kingdom of Ends, Cambridge/ New York. Langthaler, Rudolf (1991): Kants Ethik als ‘System der Zwecke’. Perspektiven einer modifizierten Idee der ‚moralischen Teleologie’ und Ethikotheologie, Berlin/New York. Marc-Wogau, Konrad (21938): Vier Studien zu Kants Kritik der Urteilskraft. Uppsala. McFarland, J. D. (1970): Kant’s Concept of Teleology, Edinburgh. McLaughlin, Peter (1990): Kant’s Critique of Teleology in Biological Explanation: Antinomy and Teleology, Lewiston. Reich, Klaus (1935): Kant und die Ethik der Griechen, Tübingen.

Kant on Ends in Nature and in Human Agency

71

Schönecker, Dieter / Wood, Allen (2002): Kants Grundlegung zur Metaphysik der Sitten. Ein einführender Kommentar, Paderborn u. a. Shell, Susan Meld (2003): “Kant’s ‘True Economy of Human Nature’”, in: B. Jacobs/P. Kain (edd.), Essays on Kant’s Anthropology, Cambridge, 194–229. Weidemann, Hermann (2001): “Kants Kritik am Eudämonismus und die Platonische Ethik”, in: Kant-Studien 92, 19–37.

Marcia Baron

Acting from Duty (GMS, 397–401) 1. Kant’s discussion of acting from duty and his remarks about moral worth are often read – and taught to students – as if they constituted an independent essay that just happened to be published where it is in the Grundlegung (or even: just happened to be published in the Grundlegung). So read, it may seem that the whole point of that discussion is to put forward a test for the moral worth of actions. This misconception then nourishes the perception that Kant’s ethics is disturbingly moralistic. “Why,” it is asked, “would we want a test for the moral worth of actions? Rightness, perhaps, but moral worth?”1 Arguably, only those eager to keep the moral score on their neighbors or (recognizing perhaps that they do not know their neighbors’ – or even their own – motives well enough to apply the test) to dream about being the gatekeepers of heaven would be interested in such a test. Not everyone puzzles over the (supposed) fact that Kant is putting forward a test for the moral worth of actions. But those who do not puzzle over why a test of moral worth is something we need are frequently troubled by the content of the apparent test. Moral worth seems to be doled out incorrectly.2 To a considerable extent these misgivings arise because GMS, 397–401, are read in isolation from the rest of Section I. To appreciate the import of GMS, 397–401, we need to attend to Kant’s avowed purpose in discussing acting from duty. He introduces the discussion by saying that to develop the concept of a good will, he will put before us the concept of duty, “which contains that of a good will, though 1

2

Thanks to Vann McGee for raising this question in conversation with me around 1977, when we were graduate students at the University of North Carolina at Chapel Hill. The thought is that actions done from admirable feeling are morally worthier, or at least as morally worthy, as actions done from duty. Among the many people who have so remarked are Annas (1984); Blum (1980); Oakley (1992); Sidgwick (1981, 223); Stocker (1976).

Acting from Duty

73

under certain subjective limitations and hindrances, which, however, far from concealing it and making it unrecognizable, rather elevate it by contrast and let it shine forth all the more brightly” (GMS, 397). The end of the discussion is also to be borne in mind: it leads to Kant’s first statement of the Categorical Imperative. Through a discussion of what it is to act from duty, he asks what sort of principle would guide a good will. The answer tells us what, if there is a supreme principle of morality, that principle is. And as he said in the Preface, “The present groundwork is […] nothing more than the search for and establishment of the supreme principle of morality” (GMS, 392). Harald Köhl is providing the chapter that addresses GMS, 402– 406, so I will not enumerate the steps that take us from acting from duty to the Categorical Imperative. But we need to remember that the discussion of acting from duty and moral worth does not stand alone. It leads us from the idea of a good will to an explication of what a supreme principle of morality, if there is one, would have to be. 2. I turn now to the details of Kant’s discussion of acting from duty. Kant says that he passes 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 have been done from duty, since they even conflict with it.” It looks as if he is saying that actions contrary to duty cannot be done from duty. That would be a strange claim: why could I not mistakenly think X was my duty, and act accordingly, when it in fact is contrary to duty? But there is no need to read Kant as ruling out this possibility, given his “already recognized” (“schon […] erkannt”). 3 Although this is not explicit, it is reasonable to suppose that by “already recognized” he means already recognized by the agent as contrary to duty (and probably also by those of us thinking about what sorts of actions might be done from duty).4 And of course if the agent recognizes the action as contrary to duty she cannot be performing such actions from duty. Kant also sets aside actions done from self-interest, i. e., actions “to which human beings have immediately no inclination, but never3

4

I am grateful to Mark Timmons for drawing my attention to Kant’s “already recognized.” In saying that it is reasonable I do not mean to imply that it is clear that “already recognized” does mean “already recognized by the agent [and probably others].” There is room for reasonable disagreement here. Indeed, Kant does seem to think it quite easy to “distinguish what is good, what is evil, what conforms to duty or is contrary to duty” (GMS, 404). Still, he does not claim that honest mistakes are impossible, and it would be uncharitable to attribute this view to him if we do not have to.

74

Marcia Baron

theless perform them because they are driven to it through another inclination.” Why set them aside? In this case, he explains, it “is easy to distinguish whether the action […] is done from duty or from a selfseeking aim.” By contrast, it is “much harder to notice this difference where the action is in conformity with duty and the subject yet has besides this an immediate inclination to it.” [“Weit schwerer ist dieser Unterschied zu bemerken, wo die Handlung pflichtmässig ist und das Subjekt noch überdem unmittelbare Neigung zu ihr hat.”]5 What exactly is Kant saying here? This much is clear: He takes the latter actions to be the cases that, for his purposes, merit attention. But a lot is not clear. The puzzles include the following: (1) Why do the latter actions – those in accordance with duty and to which the agent has an immediate inclination – merit so much attention? (2) Kant compares (a) actions done from self-interest with (b) actions to which the agent has an immediate inclination. [(a) “Handlungen […] zu denen […] Menschen unmittelbar keine Neigung haben, sie aber dennoch ausüben, weil sie durch eine andere Neigung dazu getrieben werden”; (b) “die Handlung pflichtmässig ist und das Subjekt noch überdem unmittelbare Neigung zu ihr hat.”] Why the asymmetry? Why are the latter not “done from inclination” (or why are the former not “actions for which the agent has a motive of self-interest”)? If the answer is “No reason; the asymmetry is insignificant,” contemporary commentators – myself included – who mark a distinction between acting aus Neigung and acting mit Neigung are misguided. For, thus understood, Kant would not be observing any such distinction. (3) Just what is it that is much harder to notice?

My focus in this section will be (3), though (2) will enter into the discussion because the asymmetry contributes to the challenge of figuring out what it is that is “much harder to notice.” I shall assume that Kant meant the comparison to be as he stated it, and so will preserve 5

A comment is in order concerning “Neigung.” Should this be understood narrowly, to refer only to one particular kind of desire, or more generically? At GMS, 413n, Kant writes, “The dependence of the faculty of desire on sensations is called ‘inclination.’ and this always therefore proves a need.” Strictly speaking, then, inclination is only one type of desire. But I follow Henry Allison in thinking that Kant intends “inclination” broadly in this discussion, to encompass, as Allison puts it, “momentary desires, instincts, passions, fears and disinclination” – more generally, “any stimulus to action that stems from our sensuous, as opposed to our rational nature” (Allison, 1990, 108). This reading is borne out by the variety of things in the examples at GMS, 397–399, that are loosely classified under the heading of “immediate inclination.”

Acting from Duty

75

the asymmetry.6 An answer to (1) will emerge in the course of the discussion, and evidence thought to support an alternative answer will be assessed in Sects. 3 and 4. Kant says it is “much harder to notice this difference” when the agent has an immediate inclination to the action in question than when the agent is “getrieben” by another inclination to perform the action. “This difference” refers to his assertion that it is “easy to distinguish whether the action in conformity with duty is done from duty or from a self-seeking aim.” 7 [Da “lässt sich leicht unterscheiden, ob die pflichtmässige Handlung aus Pflicht oder aus selbstsüchtiger Absicht geschehen sei.”] Let me first spell out an interpretation that I think incorrect, though it is a natural reading (particularly if the sentences are read as isolated sentences). According to this reading, Kant is saying that it is easy to discern, for any action in conformity with duty and done from selfinterest, that it indeed was done from a self-serving purpose rather than from duty; and it is harder to discern this if the action is one to which the agent has an inclination. On this reading, what is harder (or easier) to discern is the agent’s real motive. When Kant says that it is harder to notice “this difference,” he means that whereas it is easy to tell, if an action is done from self-interest, that it really is done from self-interest rather than from duty, it is much harder to determine, if an action is one to which the agent has an inclination, whether it is done from duty. This reading is implausible for at least three reasons. First, it is in some tension with Kant’s discussion at the start of Section II (GMS, 407), where he emphasizes the impossibility of knowing for certain our true motives. Second, even setting aside GMS, 407, it would be uncharitable (and implausible) to attribute to Kant so silly a view as that it is easy to detect that someone is acting from self-interest. 6

7

One reason for my assumption is that he clearly is marking the distinction between aus Neigung and mit Neigung at GMS, 400, where he writes, “Only that which is connected with my will merely as a ground, never as an effect, only what does not serve my inclination but outweighs it, or at least wholly excludes it from the reckoning in a choice, hence only the mere law for itself, can be an object of respect and hence a command” (italics added). The part I italicized allows for the possibility of an inclination (though not necessarily a supporting inclination) being present yet not influencing the decision. It is troubling that Kant seems by this to be treating all actions done from immediate inclination as actions with a self-seeking aim. I will not here enter the debate on whether Kant is a psychological hedonist with respect to all actions not done from duty. For discussion of that question, see Reath (1989), Herman (2001), and Kerstein (2002).

76

Marcia Baron

Third, it does not fit Kant’s purposes in that paragraph. Attention to the paragraph as a whole, and the context in which it occurs, makes it clear that he is not concerned there with how hard or easy it is to detect the agent’s true motive. I take Kant’s “much harder to notice this difference” to concern comparative difficulties in differentiating concepts, not in discerning motives. The claim is that acting from duty is easier to differentiate from acting from self-interest than from an action to which one has an immediate inclination.8 We do not infer from the fact that an action is done in accordance with duty that it must be done from duty – not, that is, if the action is one to which the agent has (and we see the agent to have) a strong motive of self-interest. “Thus one is honestly served; yet that is by no means sufficient for us to believe that the merchant has proceeded thus from duty and from principles of honesty; his advantage required it” (GMS, 397). By contrast, we may well draw this inference (invalid though it is) when we perceive the agent to be inclined to the action. Does Kant say what I suggested in my last sentence? No; he does not explain what contrast he intends when he begins the next paragraph with “By contrast.” I have suggested that it is this: Although we are not at all likely, if we see that the agent is “driven” by self-interest to do X, to infer from the fact that X is in conformity with duty that it is done from duty, things are different when we do not see a motive of self-interest but perceive the agent to be inclined to do X. If we do not see any motive of self-interest, we might well infer from the fact that the agent acted in accordance with duty that he acted from duty, and we might do so even if we see a strong inclination to the action in question. The claim has some intuitive plausibility. It does indeed seem to be the case that we have some tendency to conflate acting from duty and acting from inclination (assuming the action is in accordance with duty), while at the same time we view them both as sharply separate from acting from self-interest. That this is so is suggested by the clear contrast between doing something from an ulterior motive and doing it, as we say, “for its own sake.” It is noteworthy that “doing something for its own sake” covers a range of possibilities that loosely fall under 8

I will not keep repeating “immediate;” the idea is simply that one has an inclination to do the action in question, and “immediate” serves to contrast that with acting from self-interest (or as Kant has it, a mediate inclination). Note too that my clumsy wording is due, at least in part, to my efforts to preserve the asymmetry noted above. Otherwise I would write “acting from duty is easier to differentiate from acting from self-interest than from acting from immediate inclination.”

Acting from Duty

77

the headings of acting from duty and acting from inclination: doing it because it merits it or because one should do it – roughly, doing it from duty – and doing it because it is enjoyable, or because one feels like it. What is clearly ruled out by “doing it for its own sake” is doing it from some ulterior motive. It makes sense, then, to claim, as Kant does, that acting from duty is more sharply distinguished from acting from self-interest than from acting from inclination. Because acting from duty is readily conflated with acting from inclination, Kant needs to set things straight, explaining the difference between acting from duty and acting from inclination. It is important that his readers be clear on this difference, since otherwise they will not understand what an unconditionally good will is (thinking perhaps that it is something like benevolence), and are likely to get on entirely the wrong track in thinking about what sort of principle guides such a will. Without a firm understanding of the difference between acting from duty and acting from inclination, they might think that a good will could be determined by a heteronomous principle.9 My interpretation of GMS, 397, thus provides an answer to the question of why Kant devotes so much more attention to actions done from inclination than to actions done from self-interest. There is, he thinks, very little tendency to confuse the idea of acting from self-interest with that of acting from duty, whereas acting from inclination and acting from duty are less clearly distinct in his readers’ minds. A different interpretation needs to be addressed. On this alternative interpretation, Kant is again saying that we, in effect, err, but in a way that is significantly different from the way I suggested above. The error, according to this alternative interpretation, is that of failing to realize that doing X from duty is incompatible with being at the same time inclined to do X. Kant is, on this reading, at pains to emphasize the incompatibility between doing X from duty and being inclined to do X. This reading is important to address at length, because it has been the source of much antipathy to Kant’s ethics. Since actions done from duty (and only actions done from duty) have moral worth, having an inclination to do X, on this reading, prevents an action from having moral worth. And that seems troubling. Why should the fact that one has an inclination to do X preclude one’s doing X from having moral worth?

9

For yet another interpretation, see Wood (1999, Ch. 1).

78

Marcia Baron

On my interpretation, Kant’s reason for attending so closely to actions to which the agent lacks any inclination (mediate or immediate) is to bring home the difference between acting from duty and acting from inclination, a difference less apparent than that between acting from duty and acting from self-interest – and a difference vital for understanding the idea of an unconditionally good will and why there is nothing “except the universal lawfulness of the action in general” that can “serve the will as its principle” (GMS, 402). But as it is easy to suppose that Kant instead is aiming to convince us that an action cannot count as done from duty if one is also inclined so to act, I will devote the next section to assessing evidence thought to support that reading. 3. Does Kant hold that an action to which one has an immediate inclination is thereby precluded from counting as “done from duty” (and thus from having moral worth)? Before we address this, a clarification is in order. The question should not be confused with the following: Does Kant hold that an action done from inclination cannot at the same time be done from duty? I. e., does the fact that an action is done from inclination preclude its also being done from duty? I believe the answer to this question is “Yes,” but that the answer to the former question (that posed in the previous paragraph) is “No.” One can act from duty mit Neigung but cannot act aus Pflicht und Neigung.10 Some resist this distinction between (a) acting from duty, where one also has an inclination to do the action in question, and (b) acting from both duty and inclination. They claim that to have an inclination is by definition to be moved, at least a little, to act accordingly. That being the case, they argue, one cannot be inclined to do X, and then do X, without doing X at least partly from the inclination. But the “definitional” claim is not one to which Kant would, or should, accede. It presupposes a different (and less robust) view of agency than he holds. His picture is not a mechanistic one of in10

As I explain in Baron (1995, Ch. 5), and more recently in “Overdetermined Actions and Imperfect Duties” (forthcoming), I do not consider this to be an embarrassment for Kant and his defenders. I think it is correct to say that one in fact cannot do X simultaneously from duty and inclination, if by “doing X from duty” we mean doing it because it is morally required. I also note in “Overdetermined Actions” Allen Wood’s suggestion (primarily by e-mail and in conversation, but briefly indicated in Wood (1999, 44) that acting from duty should be understood more broadly, to include not only doing X because it is morally required but any action that is the result of self-constraint through moral considerations. Thus understood, acting from duty and inclination might indeed be possible. I do not believe, however, that this conception of self-constraint is Kant’s, for reasons I briefly indicate in that essay.

Acting from Duty

79

ternal forces moving us to act; so there is nothing in the nature of action or motivation, on his theory of agency, that would render nonsensical the idea that one can have an inclination to do X and do X (from duty) without actually being moved by that inclination.11 Returning now to the question with which I opened this section: Does Kant hold that an action to which one has an immediate inclination is thereby precluded from counting as “done from duty”? I will first note inadequate bases for thinking the answer has to be “Yes.” All of Kant’s illustrations in Grundlegung I of actions done from duty are actions to which the agent is not inclined (and in some cases is strongly disinclined). This has been regarded by some as strong evidence that Kant holds that an action cannot count as done from duty (and therefore, as morally worthy) if the agent is inclined so to act. But it is not strong evidence, for there is another explanation of this feature of his examples. Kant needs to provide examples of actions done from duty, and wants readers to be clear that the actions really are done from duty. That they are done from duty is most apparent if no alternative account is available. So he purposely describes conduct to which the agent is not inclined. Examples of acting from duty without any inclination so to act serve another purpose, as well. He is concerned to convince readers that actions done from duty do not need the assistance of inclination – indeed, that one can act against all felt inclination, and against one’s own interests. He therefore presents examples of acting from duty where no assistance from the inclinations is available. Now the foregoing does not purport to show that Kant in fact does hold that one can do X from duty while also being inclined to do X; it purports only to show that the passages cited as evidence against that view in fact are not evidence against it. That his examples are of actions to which the agent has no inclination (or on balance is more disinclined than inclined) is thus compatible with both views: that an ac11

Of course it would be very hard to know that one has not acted from that inclination, but that is a different matter. (And clearly the fact that I cannot tell whether I have acted from a motive I have when I do X is no reason to conclude that I must have done so.) For more on the distinction between acting mit Neigung and acting aus Neigung, see my “Freedom, Frailty, and Impurity” (1993), a discussion of some issues in Allison, Kant’s Theory of Freedom (1990). Allison replies in the same issue, and his article is reprinted in Allison (1996). See also Baron (1995, 151 f.). For a sustained argument that it makes no sense to suppose that one can act mit Neigung without acting aus Neigung (an argument that I will not be able to address here), see Latham (1994).

80

Marcia Baron

tion done from duty excludes inclination, and that it is consistent with the agent being inclined so to act. (It is also compatible with the view that an action can be done from both duty and inclination, though in fact I do not think that Kant holds that to be possible.)12 Let me expand on why Kant’s choice of examples does not show that he believes that an action from duty excludes inclination.13 Consider how he sets up his examples. He speaks first of people “so sympathetically attuned that, even without any other motive of vanity or utility to self,” they “take an inner gratification in spreading joy around them, and can take delight in the contentment of others insofar as it is their own work.” “In such a case,” he asserts, “the action […] has no true moral worth” for its maxim “lacks moral content” (GMS, 398). He then asks us to imagine “that the mind of that same friend of humanity were clouded over with his own grief, extinguishing all his sympathetic participation in the fate of others,” thus emphasizing that the motives available to him earlier are now no longer available to him. We are asked to suppose that “now, where no inclination any longer stimulates him to it, he tears himself out of this deadly insensibility and does the action without any inclination, solely from duty; only then does it for the first time have its authentic moral worth” (GMS, 398). The example brings to the fore that unless he acts from duty, his action lacks moral worth, no matter how beneficial the action and how amiable the 12

13

GMS, 400, provides strong evidence that Kant does not believe that to be possible: “An action from duty is supposed entirely to abstract from [absondern] the influence of inclination.” The evidence is strong but not quite decisive. It is a little unclear what “absondern” means here, and for that reason the evidence is not decisive, but it is difficult to imagine anything “absondern” might mean that would not entail that Kant believes an action cannot be done from both duty and inclination. Some additional evidence comes just before GMS, 402: “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 pre-eminent good which we call ‘moral.’ The relevance of this passage is apparent if we assume that “that so pre-eminent good which we call ‘moral’” is equivalent to moral worth, and bear in mind that an action has moral worth if and only if it is done from duty (GMS, 398). Additional evidence comes from recognizing what it is to act from duty for Kant, and more generally, what the relation is between incentives and action on Kant’s view of agency. Since the incentive does not operate as a mechanical force, and instead as a reason, to say that someone acts from both duty and from x is to say that duty alone is not the sole reason for which he acts. But some explanation is needed here as to why it is not the sole reason, and I can think of no explanation that avoids attributing to the agent an insufficient regard for the fact that the action in question is one’s duty. Acting from duty is incompatible with acting from desire, since it involves regarding duty as a less than decisive consideration. I explain this more fully in Baron (1995, Ch. 5) and in Baron (forthcoming). See also Allison (1990, Ch. 6). This paragraph and the following two paragraphs overlap with Baron (1995, 148 f.).

Acting from Duty

81

inclination. (It also is designed to remove any doubt that we have this incentive – the motive of duty – and that it can be effective even in the absence of any cooperating inclination.) Dissenters – those claiming that Kant holds that one cannot act from duty mit Neigung – charge that “for the first time,” in the sentence just quoted, makes it plain that Kant is saying that the man’s action lacks moral worth unless he has no inclination to it. The only way around it, they claim, is to take the liberty of inserting into the proposition “the qualification that it is the first time we can know that an action has moral worth” (Baker, 1986, 460).14 But as Barbara Herman points out, it matters that Kant is not speaking at the end of his example (at GMS, 398, 5–13) of a different person from the one described earlier in the example. He is not comparing two people, one who acts from inclination, and another who lacks inclination and acts from duty. He is saying of someone who formerly acted on others’ behalf from inclination and not from duty that only now, when the inclination to act on their behalf is not available, does his action of aiding them have moral worth. Herman explains: “Of him it is […] said: only when the inclination to help others is not available does his helping action have moral worth. For of him it was true that when he had the inclination he did not act from the motive of duty” (Herman, 1993, 18–19). Kant’s use of “first” does not signify that only when one acts without inclination does the person’s action have moral worth. He is saying of someone who, when he acts with inclination, does not act from duty, that only when he lacks the inclination to help others does his action of helping have moral worth, for only then does he act from duty. So the critics are wrong to claim that if we are to reject the “if inclinations, no moral worth” interpretation, we have to insert the qualification that only now can we know that an action has moral worth. But what about the end of GMS, 398, where Kant writes, “if nature had put little sympathy at all in the heart of this or that person, […] nevertheless would he not find a source within himself to give himself a far higher worth than that which a good-natured temperament might have”? This looks bad for those of us who hope to avoid the “if inclinations, no moral worth” interpretation. For is not Kant saying here that those who lack sympathy are able to give themselves a far higher worth than are those whom nature has endowed with plenty of sympathy? The idea, apparently, is that only the person lacking sympathy, or a heartfelt inclination to help others, can act from 14

See also Henson (1979).

82

Marcia Baron

duty. Schiller’s (in)famous jabs do not seem so clearly unwarranted after all. Kant seems to be saying that those with fellow-feeling are precluded from attaining the moral worth available to those who “by temperament” are “cold and indifferent toward the sufferings of others.” Maybe Schiller was right: one’s only choice, if one wants to be morally worthy, is to try to rid oneself of fellow-feeling.15 Looking only at the wording in that passage, the reading is plausible enough. But it would be in conflict with Kant’s theory of freedom to hold that the very having of fellow-feeling debars one from acting from duty. Kant holds that no matter what our inclinations, we are always capable of acting from duty.16 Fortunately the un-Kantian reading of the passage is not the only one possible. A much more Kantian reading is simply that in acting from duty, the man gives himself a higher worth than does the person who never acts from duty (the person with the “good-natured temperament”). The latter could act from duty, but does not. The former man acts from duty, and thus gives himself a higher worth than the latter does. Further support for understanding Grundlegung, 398–99, not to rule out the possibility of acting from duty mit Neigung comes from the many passages in which Kant emphasizes the importance of separating the pure from the empirical. These passages shed light on the motivation for choosing examples in which the agent’s inclinations are obstacles to doing his or her duty. See GMS, 388 f., and recall GMS, 397, where Kant explains that he will develop the concept of a good will in the following manner: “we will put before ourselves the concept of duty, which contains that of a good will, though under certain subjective limitations and hindrances, which, however, far from concealing it and making it unrecognizable, rather elevate it by contrast and make it shine forth all the more brightly.” The limitations 15

16

I am modifying Schiller’s suggestion only slightly. It went like this (Schiller, 1981, 221): Scruples of Conscience I like to serve my friends, but unfortunately I do it with inclination And so often I am bothered by the thought that I am not virtuous. Decision There is no other way but this! You must seek to despise them, And do with repugnance what duty bids you. I am using, with very slight alteration, Allen Wood’s translation of the passage. See Wood (1999, 28). For a discussion of Schiller’s differences with Kant on duty and inclination, see Reiner (1983). One might argue that Kant’s view is that one is always capable of doing one’s duty, not that one is always capable of acting from duty. I will not try to address that possibility here.

Acting from Duty

83

and hindrances – inclinations that conflict with duty – make the good will more evident. It should come as no surprise, then, that in each of Kant’s examples, the agent’s inclinations hinder rather than help the agent to act as morality requires. That they do should not be taken to entail that inclinations, and affect in general, are always or typically, in Kant’s view, moral hindrances.17 Kant’s discussion and commendation in Kritik der praktischen Vernunft of a method of isolation further buttress this explanation of his inclusion, in each example, of “subjective hindrances” to doing one’s duty. He likens the philosopher who arranges an “experiment” to “distinguish the moral (pure) determining ground from the empirical,” to the chemist (KpV, 92). Later in the same work he writes that “the purity of the moral principle […] can be clearly shown only by removing from the incentive of the action everything which people might count as a part of happiness” (KpV, 156, substituting “people” for “men”). 4. The arguments presented thus far undermine support for the claim that Kant holds that an action cannot be done from duty (and thus cannot have moral worth) if the agent is inclined so to act. Can we do better than that? Is there evidence that Kant in fact does hold that an action can be done from duty even when the agent is inclined so to act? I believe that GMS 397 provides very strong evidence, though not as decisive as I used to take it to be. After saying that “it is easy to distinguish whether the action in conformity with duty is done from duty or from a self-seeking aim” in the case of an action to which the agent has no immediate inclination but is “driven to it […] through another inclination,” Kant says that it is “much harder to notice this difference where the action is in conformity with duty and the subject yet has besides this an immediate inclination to it.” I reasoned that if absence of inclination were necessary for the action to qualify as an action done from duty, it would not be difficult at all to know whether an action to which the agent had an inclination was done from duty. It would be clear that it was not done from duty. Since Kant is quite explicit in saying that it is not clear, it cannot be the case that absence of inclination to X is necessary for X to qualify as an action done from duty.18 17

18

I discuss in Baron (1995, Ch. 6) a nexus of questions concerning whether there is any positive role for affect in Kant’s ethics. Lara Denis defends Kant against some of my claims in Denis (2000). See also Guyer (1993) and Sherman (1990). So I wrote in Baron (1995, 150). That Henry Allison had independently interpreted the passage in this way increased my confidence that I was reading the passage correctly. See Allison (1990, 111). When I revisited the matter while writing “Act-

84

Marcia Baron

I now think the evidence, though strong, is not decisive. It looks decisive only on the (plausible) assumption that when Kant says it is “much harder to notice this difference [...]” he means not merely that many of us find it harder to notice it but rather that it really is harder. After all, if the idea were only that many of us find it harder, that is consistent with it being the case that an absence of inclination is necessary for an action to qualify as done from duty. In other words, maybe it should not be harder, yet is (for many of us), with the result that although in fact an agent’s being inclined to do X is incompatible with her doing X from duty, we fail to realize this. Kant’s view thus might, consistent with GMS, 397, be as I described at the end of Sect. 2: although we see clearly enough that if an action is done from selfinterest it cannot be done from duty, we are not as quick to see that if one is inclined to do X (where X is in accordance with duty), one cannot do X from duty. In fact (on this alternative reading) an action cannot be done from duty if one is inclined anyway to perform the action, but since most (or many) of us do not realize that, Kant needs to persuade us. It is, on this possible reading, for that reason that he discusses at length actions from inclination, contrasting them to actions done from duty. I do not consider this reading of “harder to notice” at all plausible. In Grundlegung I, Kant is, as Thomas E. Hill, Jr. puts it, “provisionally assuming, and then analyzing and developing, the ‘knowledge’ that ordinary people have of morality insofar as they are rational” (Hill, 2002, 39). It would be surprising if in saying that it “is harder to notice this difference” Kant meant “it is harder, though it should not be.” One would expect that any contrast in Grundlegung I between what we do fi nd hard and what we should find hard would be clearly indicated, not surreptitiously implied. In sum, though it is not decisive, the evidence in support of the view that acting from duty mit Neigung is indeed possible, on Kant’s view, is very strong.19 The conflicting evidence comes only from reading GMS, 398 f., in a way that is unnecessary and moreover is in tension, for reasons explained above, with Kant’s theory of freedom.

19

ing from Duty” for Allen Wood’s edition of Groundwork, I registered some doubt (Baron, 2002, n. 4). Additional evidence can be found in the Tugendlehre. Since Kant emphasizes in that work the importance of cultivating various sentiments, it would be odd if he held that if we have sentiments that support doing what is in fact our duty – sentiments that would incline one to so act – we cannot act from duty.

Acting from Duty

85

5. Assuming that I am right to say that acting from duty is compatible with being inclined so to act, but not compatible with acting from inclination, some questions remain as to just what it is to act from duty. Could the following actions count as actions done from duty? (1) An action that is permissible but not obligatory. Let us suppose that the agent would not have performed the action had she thought it impermissible, and that she reflected on whether or not it was permissible before so acting. (2) An act of helping another or of doing something by way of developing one’s talents (practicing the violin, making arrangements to resume violin lessons, buying new strings for the violin). Let us assume that the agent opts to practice violin (or to arrange to take violin lessons) in part because she recognizes a duty to develop her talents, and that she helps another in the recognition that others’ needs make a moral claim on her (and on everyone else, as well).

It seems clear that a merely permissible action could not qualify.20 One cannot from duty alone perform a merely permissible action; one can do X from duty only if it is one’s duty, and an action that is merely permissible is not required, hence not one’s duty. Someone might retort, “What if it is the only permissible action in those circumstances?” Then it would not be merely permissible. I used “merely” to indicate that it is not, in addition to being permissible, morally required or even recommended. Another possible retort: “What if what one does is merely permissible, but there are only a few permissible actions open to one under the circumstances? Does one not act from duty in performing this action, since one has made a point of refraining from doing something impermissible, and has done so from duty?” Here it seems the correct thing to say is that what one does from duty is refrain from performing the impermissible action. One does not act from duty in opting for this particular permissible action. What, then, about (2)? It is tempting to say this: an act of helping another is not one’s duty, either. It is not a duty to help here and now, or to practice violin here and now – or, indeed, ever. I could take up a different musical instrument, or develop my talents in entirely different ways, learning a new foreign language or improving my rusty German, rereading great novels I have not read in thirty years, jogging daily, taking an astronomy course, and so on. If permissible actions 20

Unless, perhaps, one mistakenly thought it was morally required; but I will not consider that possibility here.

86

Marcia Baron

cannot be done from duty, the same would seem to be true of individual acts of helping others or of developing one’s talents. Such actions are of an interesting ilk. Each such action is of a type which, in very general terms, is required, but the action itself is not itself required.21 I am not morally required to help at every opportunity nor to develop my talents as much or as often as possible – much less to do so with respect to every talent I either have or might, with effort, come to have. The acts described in (2) are not strictly required, but are not merely permissible, either. They fall under principles of imperfect duty. Kant so categorizes them in Grundlegung, and though he does not explain the distinction between perfect and imperfect duties until later, in Metaphysik der Sitten, it is clear from the Grundlegung that then (as well as later), he viewed actions such as helping another not to be, as isolated actions, strictly required. We are required to help others, and to develop our talents, but how much (and how) cannot be specified.22 Framed differently: the maxims of helping others, and of perfecting ourselves, are obligatory, and we cannot have those maxims without adopting some more specific maxims that instantiate the very general ones; but we have considerable latitude as to which particular specific maxims we adopt. As they are not strictly required, it is hard to understand how they could possibly qualify as actions done from duty. Yet we know from GMS, 398, that Kant does think it possible to help others from duty. The best explanation is that acts of helping another count, for Kant, as acts done from duty insofar as the agent adopts from duty a maxim of helping others. The subsidiary maxims – the maxims instantiating that maxim – are also adopted from duty, and thus helping S, if one does so not for the pleasure of helping, or because one loves to be with S, but because it is morally incumbent on us to help one another, counts as an action from duty. Another way to explain it is as follows: acts of helping another are not, severally, morally required, but because they are of a type that is morally required, one can help others from duty. (The same holds for arranging to have violin lessons, practicing violin, etc.) In this way they are quite unlike merely permissible actions, and fairly similar to 21

22

I am ignoring here what I believe are exceptions (though Kant does not say that they are): aiding where the need is grave and the aid needed is easy for the agent to provide. See Baron (forthcoming). This is clear both from his remarks at GMS, 421, and in the fact that the Categorical Imperative applies differently to imperfect duties than to perfect duties. For further discussion, see Baron (1995), Baron (forthcoming), Hill (1971), and Hill (2002).

Acting from Duty

87

at least some perfect duties. I have a duty not to lie to others, but just how I fulfill this is, to some degree, up to me. Just as my duty to help others may be discharged in many ways, my duty not to lie to others can be discharged in a number of ways. Admittedly it is not exactly the same; there is less latitude in the latter than in the former; but the difference is not as huge as is sometimes supposed.23 I might discharge the duty not to lie by telling the truth (though this avenue may not be open to me if, say, it entails divulging another’s secret), by deftly changing the subject, or saying something outrageous at just the right time so as to distract or disarm my interlocutor, or more straightforwardly by indicating that I do not want to discuss that topic (though if the timing is not quite right, this may amount to giving the answer away, something I may be obliged not to do). But are they really so unlike merely permissible actions? The situation seems fairly similar to the case where there are, in a particular situation, only a few permissible actions; it is my duty to opt for one of those in order to avoid doing what is impermissible. The difference is this: the maxims of the merely permissible actions are not instantiations of a general maxim which is itself obligatory (except at a very abstract level: do not do what is impermissible). They have in common only that they are not impermissible. Unlike acts of helping others, or of developing one’s talents, they are not attached to an obligatory end. 6. Reflection on these sorts of cases leads to the thought that perhaps there is something askew in asking whether an individual action is done from duty or not. Although Kant does indeed use the word “Handlung,” which suggests individual actions, acting from duty is really a matter of acting a certain way, not performing this action from this motive. It concerns conduct over time, conduct guided by certain maxims. Likewise, moral worth seems to reside not primarily in individual actions, but in conduct – and more specifically, in the maxims, or principles, by which one conducts one’s life. That moral worth does not reside primarily in individual actions is evident in Kant’s “second proposition”: “an action from duty has its moral worth not in the aim that is supposed to be attained by it, but 23

I am highlighting the similarities in order to make sense of what I take to be Kant’s position (insofar as he had a clear position). In my “Overdetermined Actions” (forthcoming, sect. 5), I highlight the differences so as to indicate why I think one should take a different position, viz. that one can adopt a maxim of beneficence from duty but cannot help from duty in any particular instance (except where aid actually is obligatory, because it is desperately needed and easy to give).

88

Marcia Baron

rather in the maxim in accordance with which it is resolved upon,” i. e., in “the principle of the volition, in accordance with which the action is done …” (GMS, 400). I am drawing attention not to the negative part; I think Kant’s readers generally are clear that moral worth does not reside in the aim, or intended result. What seems not to be sufficiently heeded is the positive part of the assertion. The moral worth resides in the maxim. This suggests two matters that are inadequately attended to (especially by Kant’s critics, but sometimes by his defenders, as well). First, since maxims guide conduct, and are not simply attached to individual actions, moral worth is much more naturally attributed to conduct than to isolated actions. Secondly, whereas Kant’s views are often summarized as “the moral worth of an action lies in its motive,” that is not really accurate. It resides in the maxim. The mistake of thinking that moral worth lies in the motive is pernicious, because the word “motive” carries some theoretical (roughly, empiricist) baggage: it suggests a moving force that impels the agent so to act. This is, of course, quite far from Kant’s view of agency, but even those of us who are fully aware of that may still be laboring with a distorting image when we speak of “acting from duty” (or especially, “acting from the motive of duty”). If we are, then the assertion that moral worth resides in the motive is all the more misleading. Recognizing that moral worth resides in the maxim changes the picture. That Kant is far more concerned with character, and conduct over a long stretch of time, than with the moral worth or lack thereof of isolated actions is something I have argued elsewhere (Baron, 1995), focusing especially on later works (Metaphysik der Sitten and Religion). But even if we confine our attention to GMS, 397–401, it is difficult to maintain the view that Kant is offering an account of the moral worth of individual actions. First, there is the fact that he shifts from talking about the moral worth of “actions” to the moral worth of “character” (at the end of GMS, 398, and start of GMS, 399) and later (towards the end of GMS, 399) speaks of “conduct” (“Verhalten”) having moral worth. We would not expect such a shift (particularly from action to character) if he were concerned to provide an account of the moral worth of actions. Second, as noted, the fact that moral worth resides in maxims indicates that the value lies in the principle by which one conducts oneself, not in each individual action, and that evaluating individual actions for their moral worth would (quite apart from the epistemic problems) be misguided. It is nonetheless somewhat perplexing that Kant speaks of moral worth as attaching to a “Handlung” (unless “Handlung” in late 18th

Acting from Duty

89

c. German could signify conduct, rather than simply action). But it helps to make sense of this if we bear in mind the role that discussion of the moral worth of actions plays in Kant’s development of the idea of a supreme principle of morality, purified of anything empirical. To elucidate the good will and to develop therefrom the principle that guides the good will, he needs to zero in on the goodness of actions done from duty so as to bring out the following: (a) their goodness comes only from the will itself; and (b) the principle of a good will is “the universal lawfulness of the action in general,” i. e., “I ought never to conduct myself except so that I could also will that my maxim become a universal law” (GMS, 402). 7. In this essay I have focused on two interpretive questions. The first was very specific: What exactly is “much harder to notice” at GMS, 397? What contrast is Kant drawing? My answer was that Kant’s “much harder to notice this difference” concerned comparative difficulties in differentiating concepts, not in discerning motives. Acting from duty is easier to differentiate from acting from self-interest than from an action to which one has an immediate inclination. The second question was more general, and covered more of the text; and an answer to it provided a context for the first question. The second question was: Why do actions which accord with duty, and to which the agent has an immediate inclination, receive so much attention in Kant’s discussion of acting from duty? I argued that it is not because Kant is trying to show that such actions cannot be done from duty, despite appearances that he is arguing that. The idea, rather, is this. Kant is, in the interests of illuminating the concept of a good will and ultimately locating the principle that would guide such a will, explaining what it is to act from duty. To be clear on what it is to act from duty, it is necessary to isolate acting from duty from conduct that might be confused with it. There is no risk of mistaking for actions done from duty actions that we and the agent recognize as contrary to duty, so there is no need to discuss actions recognized as contrary to duty. Nor is there any danger of confusing actions from duty with actions from self-interest. But actions to which the agent has an immediate inclination – where the agent does not act from duty – might be confused with actions done from duty. Hence the need to discuss actions to which the agent has an immediate inclination. Kant then presents examples both of actions to which the agent has an immediate inclination, where the agent acts from that inclination, and actions which we can see are not done from inclination and are done from duty. So that

90

Marcia Baron

it is clear that the actions really are done from duty, Kant builds into the case that the agent is not inclined to so act. Actions done from duty differ from actions that otherwise resemble them in that the maxims of the former have moral content. Exactly what this means is not spelled out very fully in Grundlegung, and readers interested in this question would do well to read Part I of Religion.24 But it is clear that what differentiates them is not that actions from inclination spring from one sort of motive, and actions from duty spring from another;25 rather, the idea is that only actions from duty are guided by a commitment to doing whatever morality requires.26 Just as “understanding, wit, the power of judgment, and like talents of the mind,” though “good and to be wished for” can also become “evil … if the will that is to make use of [them …] is not good” (GMS, 393) likewise, inclinations such as fellowfeeling need to be guided by a commitment to doing what is right. 27 Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–). GMS Grundlegung zur Metaphysik der Sitten, AA, IV Groundwork for the Metaphysics of Morals (2002), Wood, Allen W. (ed., trans.), New Haven, Yale University Press. KpV Kritik der praktischen Vernunft, AA, V The Critique of Practical Reason (1956; 3rd ed.1993), Beck, Lewis White (ed., trans.), New York, Macmillan Publishing Co. MdST Die Metaphysik der Sitten. Metaphysische Anfangsgründe der Tugendlehre , AA,VI The Metaphysics of Morals, Part II (1996) in Gregor, Mary J. (ed., trans.): Immanuel Kant: Practical Philosophy, Cambridge, Cambridge University Press. RGV Die Religion innerhalb der Grenzen der blossen Vernunft, AA,VI Religion within the Boundaries of Mere Reason (1998) in Wood, Allen / di Giovanni, George (eds., trans.): Religion within the Boundaries of Mere Reason and Other Writings, Cambridge, Cambridge University Press. 24 25 26 27

See especially RGV 29–30 and 36. See Korsgaard (1989). For further discussion, see Baron (1995, Ch. 5). A draft of this paper was presented in Bonn, in July 2004, as part of the conference on Kant’s Grundlegung held at Rheinische Friedrich-Wilhelms-Universität. I am grateful to the conference organizers and to the participants for their helpful discussion, to Lara Denis, Dieter Schönecker, and Allen Wood for their written comments, and to Elizabeth Tropman for editorial assistance.

Acting from Duty

91

Other Works Allison, Henry E. (1990): Kant’s Theory of Freedom, New York, Cambridge University Press. Allison, Henry E. (1996): Idealism and Freedom: Essays on Kant’s Theoretical and Practical Philosophy, New York, Cambridge University Press. Allison, Henry E. (2001): “Ethics, Evil, and Anthropology in Kant: Remarks on Allen Wood’s Kant’s Ethical Thought,” Ethics 111, 594–613. Annas, Julia (1984): “Personal Love and Kantian Ethics in Effi Briest,” Philosophy and Literature 8, 15–31, reprinted in Badhwar, Neera Kapur (ed.) (1993): Friendship: A Philosophical Reader, Ithaca, Cornell University Press. Baker, Judith (1986): “Do One’s Motives Have to Be Pure?” in Grandy, Richard / Warner, Richard (eds.): Philosophical Grounds of Rationality, London, Oxford University Press. Baron, Marcia (1993): “Freedom, Frailty, and Impurity,” Inquiry 36, 431– 441. Baron, Marcia (1995): Kantian Ethics Almost without Apology, Ithaca, Cornell University Press. Baron, Marcia (2002): “Acting from Duty,” in Wood, Allen W. (ed.): Groundwork for the Metaphysics of Morals, New Haven, Yale University Press, 92– 110, published (2004) in German as “Handeln aus Pflicht,” in Ameriks, Karl / Sturma, Dieter (eds.): Kants Ethik, Paderborn, Mentis Verlag, 80–97. Baron, Marcia (forthcoming): “Overdetermined Actions, Imperfect Duties, and Moral Worth,” in Klemme, Heiner F. / Kühn, Manfred / Schönecker, Dieter (eds.): Moralische Motivation. Kant und die Alternativen, Hamburg, Felix Meiner Verlag. Blum, Lawrence (1980): Friendship, Altruism, and Morality, London, Routledge and Kegan Paul. Denis, Lara (2000): “Kant’s Cold Sage and the Sublimity of Apathy,” Kantian Review 4, 48–73. Guyer, Paul (1993): Kant and the Experience of Freedom: Essays on Aesthetics and Morality, Cambridge, Cambridge University Press. Henson, Richard G. (1979): “What Kant Might Have Said: Moral Worth and the Overdetermination of Dutiful Action,” Philosophical Review 88, 39–54. Herman, Barbara (1993): The Practice of Moral Judgment, Cambridge, Harvard University Press. Herman, Barbara (2001): “Rethinking Kant’s Hedonism,” in Byrne, Alex / Stalnaker, Robert / Wedgwood, Ralph (eds.): Fact and Value: Essays on Ethics and Metaphysics for Judith Jarvis Thomson, Cambridge, MA, MIT Press. Hill, Thomas E., Jr. (1971): “Kant on Imperfect Duty and Supererogation,” Kant-Studien 62, 55–76, reprinted in Hill (1992): Dignity and Practical Reason in Kant’s Moral Theory, Ithaca, Cornell University Press. Hill, Thomas E., Jr. (2002): Human Welfare and Moral Worth: Kantian Perspectives, Oxford, Oxford University Press. Kerstein, Samuel J. (2002): Kant’s Search for the Supreme Principle of Morality, Cambridge, Cambridge University Press.

92

Marcia Baron

Korsgaard, Christine M. (1989): “Kant’s Analysis of Obligation: The Argument of Foundations I,” The Monist 72, 311–340, reprinted in Korsgaard (1996): Creating the Kingdom of Ends, New York, Cambridge University Press. Latham, Noa (1994): “Causally Irrelevant Reasons and Action Solely from the Motive of Duty,” Journal of Philosophy 94, 599–618. Oakley, Justin (1992): Morality and the Emotions, New York, Routledge. Reath, Andrews (1989): “Hedonism, Heteronomy, and Kant’s Principle of Happiness,” Pacific Philosophical Quarterly 70, 42–72. Reiner, Hans (1983): Duty and Inclination: The Fundamentals of Morality Discussed and Redefined with Special Regard to Kant and Schiller, Santos, Mark (trans), The Hague, Martinus Nijhoff Publishers. Schiller, Friedrich (1981): “Xenien” in Trunz, Erich (ed.): Goethes Werke, vol. 1, Munich, Beck, 208–234. Sherman, Nancy (1990): “The Place of Emotion in Kantian Morality,” in Flanagan, Owen / Rorty, Amelie Oksenberg (eds.): Identity, Character, and Morality, Cambridge, MA, MIT Press, 149–170. Sidgwick, Henry (1981): The Methods of Ethics, Indianapolis, Hackett. Stocker, Michael (1976): “The Schizophrenia of Modern Ethical Theories,” Journal of Philosophy 68, 453–466. Wood, Allen W. (1999): Kant’s Ethical Thought, Cambridge, Cambridge University Press. Wood, Allen W. (2003): “The Good Will,” Philosophical Topics, 457–484.

Harald Köhl

The Derivation of the Moral Law (GMS, 402, 420f.) No new principle of morality is set forth in it [in the Groundwork] but only a new formula. … But whoever knows what a formula means to a mathematician, which determines quite precisely what is to be done to solve a problem … will not take a formula that does this with respect to all duty in general as something that is insignificant and can be dispensed with. Kant, Critique of Practical Reason, Preface

I. Posing the Question – and a Common Prejudice A supreme moral principle is an essential part of an ethical theory. This, at least, was Immanuel Kant’s conviction. Thus he had to be able, concerning his ‘moral law’, to “state its content” (GMS, 420).1 It is common knowledge that Kant made several attempts to derive the content of his moral law: in the Groundwork of the Metaphysics of Morals and the Critique of Practical Reason (GMS, 402 and 420 f.; KpV, §§ 1–7). Kant’s interpreters have not appreciated his efforts in this area. His attempts at deriving the formula of his ‘moral law’ (the ‘Categorical Imperative’) are commonly seen as difficult to understand or as being incomprehensible; or, if his derivations are believed to be understood, they are characterized as implausible or fallacious. A few examples of such critique are representative of many others: 1

What is here at issue is the basic formulation of Kant’s moral law (or his Categorical Imperative) that can be found (with small variations) in GMS (402, 421) and in KpV, § 7. One could also call it the Universal Law-Formulation of the Categorical Imperative. I will not take into account the End-in-itself-, the Autonomy-, and the Kingdom of Ends-formulations of the Imperative.

94

Harald Köhl

H. J. Paton, not famous as a Kant critic, sees a “weakness” in Kant’s argument: “If we interpret the argument [in the way Paton does] it is simply misleading.” Further, in response to Kant’s attempt at a proof in GMS I, Paton makes the general remark: “Kant has the unfortunate tendency to conflate and skim over the important transitions in his proofs” (Paton 1958, 71). – One of the most distinguished contemporary German philosophers, Ernst Tugendhat, thinks that Kant’s argument in GMS I is “of course … invalid”, the transition to the conclusion is simply “arbitrary” (Tugendhat, 1993, 130). The “argument that Kant gives [in GMS II] is surely false” (ibid. 136 f.; my insertion). “… so we can … state that his attempt to give the very idea of a categorical imperative [the Categorical Imperative] … a particular meaning has failed” (ibid. 137 f.; my insertion). – Henry E. Allison (1996, 143) fi nds a gap in the proof of GMS I and transfers this diagnosis to the argument in GMS II. Nonetheless, Allison thinks of the corresponding argument in the KpV that it doen’t contain a gap. – Allen W. Wood and Dieter Schönecker also think Kant’s proof to determine the content of the moral law in GMS I is ‘highly problematic’ because it contains a ‘gap.’ The argument in GMS II ‘is open to the same objections’ (Schönecker / Wood, 22004, 91, 93, 126). In what follows, I will try to show that, contrary to common opinion, Kant’s derivations of the basic formula of the ‘moral law’ in the GMS can be plausibly reconstructed (parts II–IV of this essay).2 Consequently, other reconstructions are targets for critique (part V). Nonetheless, the possibility of a comprehensible explication of Kant’s derivation of the Categorical Imperative does not mean that his proofs are themselves beyond criticism (part VI). II. From outside in It seems to me useful to begin with a self-made argument (albeit inspired by Kant): before reconstructing the Kantian texts. My own argument to determine the content of a moral law runs as follows: 1. If there were something like a moral law, what would it say? – This is the initial question. 2. Even if we don’t (yet) know the wording of the moral law, we know that we are searching for a law. 2

In connection with my critique of Allison, in part V of this paper, I discuss the corresponding argument in the KpV.

The Derivation of the Moral Law

95

3. From the connotations of the term ‘law’ we can conclude that a moral law is a rule that encompasses a wide range of creatures (for Kant: all rational beings). Such a law is a ‘universal law’. 3 Let us assume that as a moral law it carries with it an expectation that all addressees can be made to insightfully understand it. Hence the ‘universality’ of our law means the expectation of its being universally valid. 4. Further, we assume that we know what the formula of the moral law cannot contain. We presume that it cannot be a rule that demands certain actions because they are useful to achieve contingent aims.4 (On this point many of us have Kantian intuitions.) 5. What else could our moral law contain? In any case, its formulation must express its normative character (for beings like us): it is a universal rule. As such, it demands a particular type of action, actions in conformity with the rule. In the same sense we can speak of action that is demanded by a law and in conformity with a law (or of maxims that are in conformity with a law). In this sense, I will speak of ‘lawfulness1’ (‘Gesetzmäßigkeit1’), which is demanded by the (moral) law; if an action or maxim is in this sense lawful (‘gesetzmäßig’), it is ‘lawful1’. – It is necessary to use an index here, because the Kantian notion ‘Gesetzmäßigkeit’ has a second, crucially different meaning to which I will refer as ‘lawfulness2’ and which means something like ‘lawlikeness’; more on this later. 6. The previous point expresses the banality that a moral law, as a rule, demands action in accordance with the law. But which kind of action? What property of the law could it be that the demanded behaviour should conform with? If the law does not contain the demand to pursuit certain aims, whose realization affords the execution of certain actions (cf. above step 4.): then the only thing that remains in the law of which it can sensibly be said that actions must conform to, is the legal form, its having the character of a law, i. e. the pretension of being universally valid that belongs to the concept of a law (cf. above step 3). For the content of the moral law, this means that the demanded behaviour must be able to be prescribed in a legal formulation, i. e. in 3

4

If “practical laws” are to be laws, they must, as Kant puts it, “be valid for the will of every rational creature” (KpV, 19). “A practical law that I cognize as such must qualify for a giving of universal law: this is an identical proposition and therefore self-evident.” (KpV, 27) Even hapiness is a ‘contingent’ aim in this sense, since particular conceptions of it cannot be made universally binding.

96

Harald Köhl

the form of a law, because it is universalizable and can thus justifiably be demanded of all addressees. In other words, the actions demanded must be grounded in intentions or subjective principles (maxims) that can serve as universal laws (or: can be reformulated into such laws). 5 Here now is the place for speaking of ‘lawfulness2’ (‘Gesetzmäßigkeit2’). To act in accordance with a moral law in this sense means to constitute one’s actions in a lawful 2 manner, i. e. to act in a way that one’s action itself is ‘lawlike’ (and hence carries the expectation of being universally valid with it). The content of the moral law thus demands action that is lawful2 (gesetzmäßig2) – or maxims that are lawful2. By my lights, the foregoing argument is intelligible and conclusive. The conclusion – and so the Kantian proofs, if my considerations should substantially coincide with his – therefore can only be questionable if one or more premises are false. I will examine this possibility in the concluding part of my paper. III. Kant’s Proof in Groundwork II I will now turn to those passages in Kant’s texts where he himself attempted to derive the content of his moral law. The proof he presents in GMS II (420 f.) for the basic formula of the Categorical Imperative is, in my opinion, easier to understand than its counterpart in GMS I (402). In any case, the latter proof can be understood with the help of the deliberations developed in the preceeding part of my paper. For this reason, I take it to be methodologically necessary to begin with the latter proof. Kant surprises us here with the idea of “whether the mere concept of a categorical imperative may not also provide its formula contain5

Strictly speaking, a morally qualified maxim cannot become a moral law. Maxims express general intentions of individual persons, while laws (at least for beings like us) are normative, i. e. are rules. So, strictly speaking one must say that the moral qualification of being universalizable that a maxim may possess is a good reason to formulate a corresponding norm (law). Kant himself invites the reading just criticized when he writes: “Now I want only to know whether a maxim could also hold as a universal practical law.” (KpV, 27) But he also expressly writes: “maxims are indeed principles but not imperatives.” (KpV, 20) Charitable readers orient themselves at the best formulations that an author has found for his claims. – See also the corresponding interpretation of Bruce Aune (1979), 23 f.: “Kant characterizes practical laws in a way implying that they could not be […] maxims.” But it would, acccording to Aune, be okay to say: “… if a given maxim is universally valid, a corresponding, fully general, ‘lawlike’ statement is also ‘valid’” (my emphasis). Such a ‘statement’ is, in other words, a “candidate for the relevant law”.

The Derivation of the Moral Law

97

ing the proposition which alone can be a categorical imperative” (420; my emphasis). My italics make clear that this is one of the passages where Kant did not carefully distinguish between the plurality of categorical imperatives and the Categorical Imperative; between “all imperatives of duty” and “their principle”, the “universal imperative of duty”, out of which the former imperatives can supposedly “be derived” via its application on maxims (421; my emphasis). In any case, Kant’s strategy seems to be clear: his distinction developed in 414 ff. between categorical and hypothetical imperatives was intended to clarify the character of a categorical imperative. From his description of the character of categorical imperatives he believed he could now extract the formulation of the Categorical Imperative: (1) … when I think of a categorical imperative I know at once what it contains. (2.1) For, since the imperative contains, beyond the law, only the necessity that the maxim be in conformity with that law, (2.2) while the law contains no condition to which it would be limited, (2.3) nothing is left with which the maxim of action is to conform but the universality of a law as such; and this conformity alone is what the imperative properly represents as necessary. (3) There is, therefore, only a single categorical imperative and it is this: act only in accordance with that maxim through which you can at the same time will that it become a universal law. (420 f.; my numbering)

I will start my interpretation of this passage with sentence (3), in which Kant presents the basic formulation of his Categorical Imperative. Two meanings of ‘law’ are here at work: the “single categorical imperative” refers to the Categorical Imperative (i. e. the ‘moral law’ for morally imperfect beings). The searched-for law (singular) prescribes maxims to have a constitution so that they could ‘become a universal law’. The last formulation implies the assumption of a potential plurality of laws that could be ‘generated’ from maxims that would meet the requirements of the law. Thus, (3) implies a hierarchy of laws. Consequently I will call the law the ‘supreme’ moral law, or the ‘fundamental’ moral law.6 Kant’s formulation of that law: “act only in accordance with that maxim through which you can at the same time will that it become a universal law” (my emphasis) must be understood in the sense of my footnote 5 that the universalizability of maxims is a reason to establish a corresponding norm. In (2.1) the term ‘law’, which the Categorical Imperative “contains”, refers to the law-character which the supreme law posesses as a law: 6

See the title of KpV, § 7.

98

Harald Köhl

It is accompanied by the expectation of its universal validity. (Cf. II, 2 and II, 3.) – Kant’s formulation of the ‘necessity of the maxim’ (still in 2.1) to conform with that law, expresses the prescriptive character of the supreme law. According to Kant, every norm expresses a practical necessity (and categorical imperatives an ‘unconditional’ practical necessity); in this context that means that each maxim must conform to this law: i. e. to the law-character of the Categorical Imperative. I understand the idea of ‘conformity to the law’ (‘Gesetzmäßigkeit’) which is contained in the previous formulation in the sense of ‘lawfulness1’ (‘Gesetzmäßigkeit1’; cf. II, 5), i. e. in the sense of a demand to follow a categorically formulated rule. Note that if Kant meant here already ‘lawlikeness’ (‘Gesetzmäßigkeit2’) – the prescription that the maxims themselves should have legal form – there would have been no reason for him to continue his proof. But more importantly, in that case he would, in (2.1), have presupposed what should result only as the conclusion (2.3). In (2.2) it is established what the searched-for formulation of the Categorical Imperative (Kant: “the law”) cannot contain: “no condition to which it would be limited”. (This corresponds to II.4.) After all, the Categorical Imperative, whose content is to be ascertained, is a categorical imperative. Such imperatives, qua being ‘un-conditional’, are characterized negatively7 in the following way: they must not, in contrast to hypothetical imperatives, be followed under the ‘condition’ of the utility of the hypothetically demanded actions. – Note that this and only this point of his proof justifies the Kantian suggestion that he’s deriving the content of ‘the’ Categorical Imperative from the character of ‘a’ categorical imperative. Ad (2.3): “[N]othing is left with which the maxim of action is to conform but the universality of a law as such; and this conformity alone is what the imperative properly represents as necessary.” Now Kant has arrived at his conclusion. In II,3 I explained the “universality of a law as such” by our expectation of its universal validity. That “the maxim of action” “is to” conform to the “universality of a law as such” – a “conformity … the imperative properly represents as necessary” (represents = formulates) – I read in the sense of ‘lawfulness2’ 7

The positive descriptions that Kant gives in the GMS about the character of categorical imperatives are often overlooked. A categorically demanded action is “necessary of itself” and “good in itself” (414); good through the “form and the principle” (416) of the action itself. Its ‘form’ is the “form, which consists in universality” (436) that is demanded by the ‘principle’ of Kantian ethics. See H. Köhl (2001), part IV. See also Köhl (2004), 128, as well as Köhl (2006), Chapter 5.

The Derivation of the Moral Law

99

= ‘lawlikeness’ (= ‘Gesetzmäßigkeit2’) (my emphasis). This means the demand that a maxim, if it is to be considered as moral, can (normatively re-formulated) be a law. (See II.6.) The moral law has this demand as its content. – The formulation of the Categorical Imperative in Kant’s (3) is simply an alternative expression of what Kant has already concluded in (2.3). In the antecedent explication of Kant’s proof in GMS II, I believe the difference between ‘lawfulness1’ (= ‘Gesetzmäßigkeit1‘) and ‘lawfulness2’ = ‘lawlikeness’ (‘Gesetzmäßigkeit2’) has proofed useful. Of course, what has been shown thus far has only been an explication of the component parts out of which Kant’s reasoning consists. We are still missing an insight into the plot of his argument. This can be presented as follows: What knowledge is now available to us that we could use to determine the content of the Categorical Imperative? First, we know that we are dealing with an imperative, a rule (see (2.1) and (2.3)): “with which the maxim of action is to conform”). Second, we mean to know that this imperative should be categorical,8 not hypothetical (see (2.2) of Kant’s proof). Third, since the Categorical Imperative is meant to formulate a supreme law, it possesses a legal character: in the sense of being associated with an expectation to be valid for all X.9 So, when I’m thinking of a categorical imperative, I also know what it contains: one can say in allusion to Kant’s sentence (1) in 420. It contains a universal, supreme, categorical rule for action. Moreover it is clear that the searched-for imperative, qua rule, should prescribe something for actions or maxims: presumably a property that they should possess; an aspect in which they should ‘conform’ to the imperative. Now Kant’s question seems to me to have been the following: assuming we know nothing about the Categorical imperative at the outset except the three aforementioned aspects, which of these contains something that could be prescribed for actions or maxims of action? Not the first aspect, the prescriptive character. That would be pointless, because maxims have, in expressing intentions, a different logical form than prescriptions. Not the second aspect, since the application of the predicate ‘categorical’ to maxims is nonsensical. Thus only the third aspect remains: the legal character. So the supreme law can at 8

9

[L]aws … must be … categorical: otherwise they are not laws.” Cf. KpV, 20. “… this rule … is a law because it is a categorical imperative.” Cf. KpV, 21. With the “lawgiving” of reason the rule must be “objectively and universally valid“. Cf. KpV, 20 f.

100

Harald Köhl

most stipulate that maxims must themselves have a lawlike character, that is, be able to establish (lower-level) moral laws (plural). And is this not precisely Kant’s proof? “For, since the imperative contains, beyond the law [my third: legal character], only the necessity that the maxim be in conformity with this law [my first: prescriptive character], while the law contains no condition to which it would be limited [my second: the categorical character], nothing is left with which the maxim of action is to conform but the universality of a law as such [the result of my process of elimination]; and this conformity alone is what the imperative properly represents as necessary. There is, therefore, only a single categorical imperative and it is this: act only in accordance with that maxim through which you can at the same time will that it become a universal law.” – Now, we can say in more than simply an allusion, along with Kant ‘that, when I think of a categorical imperative, I immediately know what it contains.’ By my lights, this explication of Kant’s proof in GMS II is both intelligible and conclusive. And the moral of this story? There is only hope for a supreme moral principle, to be actually a law (that is to be universally valid), if it prescribes nothing for maxims other than this: they themselves must be universal. Stated negatively: if the fundamental moral law permitted other maxims, the law’s own claim to universality would be destroyed. For such maxims would not be compatible with its character of being a law. Of course, Kant’s proof only gives a formulation of a supreme moral principle. Justification of its content (that is of the demand expressed by it) is another matter, which Kant will address in GMS III. Even if the demand, to act according to universally valid rules of action, were justified, one could still ask if the resulting law is a constructive, or fruitful principle. IV. Kant’s Proof in Groundwork I Kant’s first attempt to prove the content of his ‘moral law’ (402) is, in my opinion, more difficult to understand than his later attempt 420 f. The difficulty comes from the perspective from which he approaches the matter. Kant’s starter for the proof is now the previously conducted search for the moral ‘motive’. For him, a morally motivated action (for morally imperfect beings) consists in acting “from duty.” Only then does an action have “moral worth” (397 ff.). Kant takes up these theses again immediately preceding his proof 402, as well as

The Derivation of the Moral Law

101

the ‘good will’ and its attributes. This is unsurprising, since the ‘Achtungs’-Thesis (400) belongs to the ‘three sentences’ about acting out of duty, through which Kant wants to further “develop” his idea of a good will (397). This thesis is Kant’s ‘third sentence’: “Duty is the necessity of an action from respect for law.” (400)10 In the following sentence where he takes up earlier presented determinations, he claims regarding an “action from duty” that in its case “there is left for the will nothing that could determine it except objectively the law and subjectively pure respect for this practical law, and so the maxim of complying with such a law” (400 f.). Because of the explicative ‘and so’ this sounds like the anticipated conclusion of Kant’s following proof: (1) But what kind of law can that be, the representation of which must determine the will, even without regard for the effect expected from it, in order for the will to be called good absolutely and without limitation? (2.1) Since I have deprived the will of every impulse that could arise for it from obeying some law, (2.2) nothing is left but the conformity of actions as such with universal law, which alone is to serve the will as its principle, (2.3) that is, I ought to act except in such a way that I could also will that my maxim should become a universal law. (3.) Here mere conformity to law as such, without having it as its basis some law determined for certain actions, is what serves the will as its principle … (402; first emphasis and inserted numbers mine.)

Sentence (1) poses the initial question (“what kind of law can that be”) and makes clear from the beginning that Kant wants to perform a proof via an analysis of the genuine moral motive. First, Kant’s talk about the “representation” of the law “which must determine the will” (my emphasis) is clearly about a motive. Second, because that “representation” is a representation of the moral law, Kant can only be referring to the moral motive to act ‘for the sake of the law’ (see 390), or in the case of human beings, ‘out of duty’. This is the motive that must be present for a will to “be called good” (401).11 Kant’s initial question is thus: what is the content of moral law itself, if the motive to act according to it is the motive to act ‘for the sake of the law’? In 10

11

Here, for the first time in GMS I, Kant mentions the (moral) ‘law’. In the ‘Preface’ to the GMS, Kant already mentions (moral) ‘laws’ (plural; in 387 f.) and soon thereafter for the first time „the moral law” (390; my emphasis). With his talk of the “representation” (of the law) that “must determine the will” in moral action, Kant takes up again his previous discussion about the “representation of the law” (401). There already he claimed that this representation “insofar as it is the determing ground of the will, can constitute the preeminent good of the will we call moral” (ibid.).

102

Harald Köhl

other words, what is implied, for the wording of the moral law, by this description of the moral motive? Ad (2.1): The moral motive ‘to act for the sake of the law’ refers ultimately to the moral law. Thus it cannot consist in the “impulse” to achieve something beyond “obeying” the law (no “regard” is present for an “effect expected” (see (1)). But then (2.2) the moral motive ‘to act for the sake of the law’ can only refer to the law-character of the law, and be the motive for generating actions with this character. Acting this way is consequently demanded by the moral law. (2.3) This last step in the proof takes place by a shift in perspectives. If – out of the perspective of the moral motive to act ‘for the sake of the law’ – this motive can only refer to the law-property φ: then – seen from the perspective of the searched-for law-formulation – a law that demands the exercise of so motivated actions accordingly can only contain the precept: Carry out actions with property φ (or act according to maxims with this property). The “conformity of actions … with universal law”, in (2.2), and the “mere conformity to law”, in (3), is that which I earlier referred to as ‘lawlikeness’ (‘Gesetzmäßigkeit2’): The moral law consists in the command to carry out actions that are capable of being laws (or whose underlying maxim can be transformed into a law). A good action (with ‘moral worth’) is one done ‘for the sake of the law’: that is (solely12) because it is capable of being a law. – I can here leave the question unanswered whether for Kant it is part of the content of the moral law that actions that occur because of their lawlikeness must be done from that motive.13 What differentiates the proofs that Kant presents in 402 and 421 f. to determine the content of the moral law? The first proof, an examination from the perspective of the moral motive, seems to me to rely on the crucial move (not explicit in Kants text) of assuming that the moral motive must ultimately refer to the law. Thus, it cannot be for the sake of profits that may result from following the law. Consequently, it can only aim at that which defines the law qua law: that is the expectation associated with it of being universally valid. In the proof 421 f., I see the crucial move as the transition from lawfulness1 (Gesetzmäßigkeit1) to lawfulness2 = lawlikeness (Gesetz12 13

I argue for this strict reading in Köhl (1990), 84 ff. If one reads into the formulation of the the Categorical Imperative the demand of genuinely morally motivated action it is impossible to explicate the difference between the ‘law-conformity’ (‘Legalität’) and ‘morality’ (‘Moralität’) of an action. I have argued for this thesis in Köhl (1990), 67, Fn. 9.

The Derivation of the Moral Law

103

mäßigkeit2). Gesetzmäßigkeit1 appears only implicitly in the proof 402.14 One easily gets the impression that here Kant‘s introduction of ‘lawlikeness’ (in (2.2) and (3)) is somewhat abrupt, thereby skipping a step in the argument – which I believe is a false impression. Kant could have carried out the proof by explicitly working with Gesetzmäßigkeit1 as follows: (1) Good actions have to be done ‘for the sake of the law’ (i. e. out of ‘the’ moral motive: ‘morality’). (2) ‘To act for the sake of the law’ implies wanting to act lawfully1 (= gesetzmäßig1; ‘law-conformity’), that is in accord with the legal precept – even if the content of said precept is for now unknown. (3) That at which the lawfully1-oriented action cannot aim is some profit resulting from obeying the law. (4) Thus it must aim at conforming to the legal character of the law, that is according to its claim to universal validity (lawlikeness = Gesetzmäßigkeit2). ‘To act for the sake of the law’ thus means to act for the sake of the lawlikeness of one’s own actions. (5) The content of the moral law thus consists in the demand for actions that are lawlike (lawful 2).

V. Three Critics of Kant, and an ally In my discussion of secondary literature I concentrate on a thread that was initially spun by Bruce Aune, and then picked up by Allan Wood and Henry Allison (among others). All three authors claim to have found a gap in Kant’s argument to determine the content of the moral law. Because Wood and Allison both lean upon Aune, I will mainly concentrate on him. It is as a consequence of the previous parts of my essay that I find Aune’s criticism of Kant erroneous. The mistakes that he makes are repeated (with variation) by Allison and Wood. Aune concentrates himself on the passage 402 in GMS I.15 Kant’s claim had been: (2.1) Since I have deprived the will of every impulse that could arise for it from obeying some law, (2.2) nothing is left but the conformity of actions as such with universal law, which alone is to serve the will as its principle … 14

15

It occurs insofar as the thought that the searched-for law contains the notion that it should be followed, that is, the prescribed actions should be in accord with the law. See for the following: Aune (1979), 28 ff.

104

Harald Köhl

Aune took up this text as follows: (S1) Kant’s idea here is that since … a good and dutiful will does not conform to a practical law because the consequences of doing so will satisfy some purpose or ‘inclination’, the only rational basis for conforming to it … must be supplied by the intrinsic character of the law itself … [i. e.] ‘the character of being a universal law.’ (28; my insertion.)

Out of this Aune develops – supposedly Kant’s – formula of the practical respectively moral law: L: Conform your actions to universal law. (29)

From this Aune distinguishes the following, Kant’s fi rst formulation of the moral law in 402 and the formulation of the Categorical Imperative 421, the latter equivalent with Aune’s C1: Act only on that maxim through which you can at the same time will that it should become … a universal law. (Ibid.)

Though Aune claims that Kant’s text allows one to believe ‘he thinks L and C1 are the same basic principle’ (ibid.), he takes this belief to be highly implausible: because it seems to him obvious – and he thinks it must have been so for Kant as well – that L and C1 differ in an important regard: (S2) C1 is useful in a way that L is not. (S3) As I have formulated it, L is a higher-order principle telling us how to conform to certain lower-order laws. (S4) But if we do not know what these lower-order laws are we shall not fi nd L a very useful principle. (S5) C1 does not seem to possess this limitation. (S6) Even if I do not know what principles are properly considered universal laws, I know they are practical principles that, being universal, are binding on all rational beings. (S7) Knowing just this much, I could no doubt decide whether I could will that the maxims of many possible actions should be (or become) such laws. … (S8) C1 would thus have practical value for me even though I had no idea how to comply with L. (30; (S1)-(S8): my numbering.)

Based on these considerations, Aune poses the question: “Which is really the fundamental practical law, L or C1 …?”, and answers it for the benefit of C1 (29). For Aune, the supposed gap in Kant’s proof consists in the absence of an argumentative bridge from L (which he presumes to have found in Kant’s text) to C1. What should we make of Aune’s interpretation? (S1) does not make it clear that Kant wanted to reach a formulation of the moral law by considering the motive for moral action. For Kant

The Derivation of the Moral Law

105

this is clear in the discussion prior to sentence (2.1). By contrast, Aune orients himself at Kant’s formulation in (2.1): “I have deprived the will of every impulse that could arise for it from obeying some law” (my emphasis). Kant is talking here about a good will or, as Aune says in (S1): “a good and dutiful will.” Kant’s thought however, as I hope to have shown in part IV, is clearer if one orients oneself by a description of a motive such as ‘acting for the sake of the law’ (or ‘out of duty’ or ‘out of respect’). These descriptions make clear that the moral motive ultimately refers to the moral law (whatever it may contain), and thus not on “any impulse that could arise for it by obeying some law.” Accordingly the searched-for law cannot prescribe aiming at such ‘impulses’ (results of action). The moral motive can only be directed at the “intrinsic character of the law itself”, i. e. its “universality”. (Instead of Aune’s “universality”, in (S1), I have spoken of the expectation of the law’s universal validity.) It follows that the searched-for law can only prescribe something that flows out of its “internal character” (S1). I am skeptical that one should, along with Aune in (S1), say that the motivational relationship to the law’s legal character is “the only rational basis for conforming to it” (my emphasis). Here not rationality is relevant, but the appropriate characterization of the moral motive, out of which an appropriate formulation of the law can be drawn. But how does Aune move from his Kant-interpretation in the sense of (S1) to his law-formulation L? And how does he gain his (S3): “As I have formulated it, L is a higher-order principle telling us how to conform to certain lower-order laws” from Kant’s text? I believe Aune’s interpretation can only come from a misreading of Kant’s sentence ((2.1) above): [s]ince I have deprived the will of every impulse that could arise for it from obeying some law …” (my emphasis). Clearly Aune reads Kant’s ‘some law’ as ‘one of a plurality of moral laws.’ But Kant’s text does not support this reading. ‘Some law’ may refer to the moral law or to ‘an arbitrary law, be it the moral law.’ In any case I am unable to find Aune’s distinction between the moral law and lower-order moral laws in Kants passage 402. Since I am unable, in this passage, to discover Aune’s distinction, I conclude that Kant in fact does not hold Aune’s L. At least not at 402. If Aune were correct that Kant holds L and bases his thought on a difference between the moral law and lower-level moral laws, then his criticism in (S4) would be correct: “But if we do not know what these lower-order laws are we shall not find L a very useful principle.” What Aune takes to be an argument against L is, from my point of view, an argument against attributing L to Kant. Since Kant, by his

106

Harald Köhl

proof at 402, first wanted to discover the formulation of the moral law (out of which concrete laws could be derived), he had, on his way to that formulation, to avoid talk of such laws. On my reading there is no mention of them. If Aune were correct in attributing L to Kant, he would also be correct in his assertion that C1 is more likely to be the searched-for fundamental moral principle than L. But since L is not held by Kant at all, Aune’s question regarding the relationship between L and C1 does not even arise. Of course, there can be no gap between two – one of which doesn’t exist. Kant’s proof in 402 may be faulty, but for this gap he cannot be held responsible. I want to move from my last point of contention with Aune to my criticisms of both Wood’s and Allison’s interpretations. Aune had formulated that “Kant’s … good and duitiful will does not conform to a practical law because …” and “the only rational basis for conforming to it …” (see (S1)), and also that “L is a higher-order principle telling us how to conform to certain lower order laws” (see (S3), my emphasis). Through this ‘conformity’-talk a misplaced notion has been imported into Kant’s Text on 402 (and took place in the history of this passage’s interpretation): what I have termed ‘lawfulness1’ (‘Gesetzmäßigkeit1). In this sense one acts lawful if one’s own action corresponds with a/the law, i. e. one does what one is commanded to do. As I have tried to show in part IV of my paper, this notion of Gesetzmäßigkeit1 is not present in 402 (in contrast to the proof in 420 f.). Rather, the “conformity of actions as such with universal law, which alone must serve the will as its principle” and the “mere conformity with the law as such” that should be present in morally motivated action ‘serving the will as its principle’ (present in 402) is captured by what I have called ‘lawlikeness’. This means the capacity of an action or a maxim to serve as the epistemic basis for a norm that can be thought to be universally valid. The presentation of proof 402 that Allen Wood gives in his book on Kant’s Ethical Thought (KET)16 is too brief to be considered a serious effort at comprehending the argument. His critique is just as brief. Wood obviously thinks the errors in Kant’s proof are so evident that it is unnecessary to explicate them. Kant’s reflections are “based on fallacious reasoning (or at least it is an argument with a large gap remaining to be filled)” (KET: 48). Wood sees the same gap in the proof that Aune saw between his L and C1. 16

See Wood (1999), 47 f.

The Derivation of the Moral Law

107

Wood’s treatment of proof 420 f. is more thorough. His criticism of Kant is beset by a misreading similar Aune’s. Thus, Wood’s reconstruction also utilizes the notion of ‘lawfulness1’ (Gesetzmäßigkeit1) understood as ‘conformity with moral laws (plural)’: The argument [420 f.] … infers that the conformity of maxims to universal laws as such follows from the mere concept of a categorical imperative. The first and most abstract version of the categorical imperative could be expressed as follows: CI: Adopt only maxims that conform to universal law as such. (KET: 78)

Out of this Wood generates the objection we have encountered in Aune (against L): The trouble with CI is that it gives us no idea at all what universal laws there are. CI does tell us nothing about the content of practical laws. The claim that our maxims should conform to universal law even seems to presuppose the existence of another universal law (or laws) whose content still remains unspecified. (KET: 78 f.)

How comes Wood to believe that one can obtain the notion that maxims should conform to moral laws (plural) from Kant’s argument in 420 f.? In Kant’s argument only the law is under consideration: (1.) … when I think of a categorical imperative I know at once what it contains. For, since the imperative contains, beyond the law, only the necessity that the maxim be in conformity with that law, (2.2) while the law contains no condition to which it would be limited, (2.3) nothing is left with which the maxim of action is to conform but the universality of a law as such; and this conformity alone is what the imperative properly represents as necessary. (My emphasis.)

Wood cannot have generated his interpretation out of Kant’s formulation of the Categorical Imperative, according to which one should only act following according to maxims that one can simultaneously desire to “become a universal law” (my emphasis). Even though the idea of a plurality of laws is implicit here it would be incoherent to speak of a maxim conforming to a law that the maxim eventually is fi rst becoming17 . (3.) There is, therefore, only a single categorical imperative and it is this: act only in accordance with that maxim through which you can at the same time will that it become a universal law. (420 f. my emphasis; original text italicizes the entire Categorical Imperative.) 17

On the problem of Kant’s claim that a ‘maxim can become a universal law’, cf. my footnote 5.

108

Harald Köhl

Wood could have discovered his interpretation only in the formulation of Kant’s sentence (2.3) that states: “nothing is left with which the maxim of action is to conform but the universality of a [singular] law as such; and this conformity alone is what the imperative properly represents as necessary” (my emphasis and insertion). Wood obviously reads “a law” as ‘one of a potential plurality of laws’ and the ‘conformity’ to law in the sense of ‘lawfulness1’ (Gesetzmäßigkeit1). But this sentence is not speaking of many potential laws, but of the character of a law as such: its universality. This is conveyed by ‘lawlikeness’: maxims should evince the same universality as a law, so that they may ‘become’ laws. That which maxims should conform to, are, on my reading of 420 f., not laws (as Wood contends) but the universality that signifies a law qua law. The difficulty that Wood had in reading Kant’s text is apparent in his vacillation between different formulations. He speaks of the “the conformity of maxims to universal laws [plural] as such”; formulates as CI: “Adopt only maxims that conform to universal law [sing.] as such”; und formulates “that our maxims should conform to universal law [sing.]” (See the Wood quotes above in my text; my emphasisses and insertions.) Wood even imports the notion of ‘conformity with moral laws’ (plural) into his understanding of the basic formulation of the Categorical Imperative 421 (as demonstrated above this idea is there not present): “… it [the basic formula] proposes a test to use on any given maxim to determine whether that maxim conforms to these universal laws“ (KET: 79; my emphasis and insertion – Wood italicizes ‘test’). However Kant’s thought that maxims eventually can ‘become’ laws is different from the notion that they should conform to laws – in the sense of ‘lawfulness1’ (Gesetzmäßigkeit1). In his explicitly critical treatment of “The fallacy in Kant’s deduction” (KET: 81 f.) Wood claims, like Aune, that the basic formula of the Categorical Imperative does not follow out of his CI. (Here again he wavers between speaking of “the conformity [of a maxim] to the universality of law in general” and saying: “… this tells us no more than that our maxims ought to conform to whatever universal laws (plural) there are” (KET: 81; emphasis and insertions mine). But since, by my lights, Kant does not hold CI, there can be no gap in moving from CI to the basic formula. Henry Allison follows Aune’s and Wood’s criticism of Kant in his paper already mentioned.18 Allison leans on Wood’s work on Hegel’s 18

See Allison (1996), footnote 6.

The Derivation of the Moral Law

109

Ethical Thought (HET),19 in which Wood previously examined Kant’s arguments about the content of the Categorical Imperative. Allison writes, like Aune: … since Kant presents this imperative [the Categorical Imperative] in a parenthetical clause intended to explicate the notion of conformity to universal law as such, it would appear that he regarded these principles as equivalent. At the very least, he must be taken as claiming that the fi rst somehow entails the second. Here, then, is the ‘crucial gap’ in Kant’s argument. (145 – my insertion.)

According to Allison, Wood wanted to show “the impossibility of deriving a viable moral principle from the mere concept of universal practical law“ (143). The way in which Allison relates the misreadings of Aune and Wood is remarkable in at least one point. To make this clear, I will introduce a distinction that Wood used in HET: Wood distinguished the “universality of applicability” and the “universality of concern” as two desirable properties of practical laws (HET: 135 f.; my emphasis; see also 165 f.). “[T]hat laws should have … ‘universality of applicability’” means “[e]verybody should be equally subject to them.” (HET: 165; my emphasis.) The “universality of concern” that we desire of good laws rests on the assumption “that everyone’s following those laws will result in a system of collective behaviour that is rational and generally beneficial” (ibid.). Allison recapitulates Wood’s distinction (which Wood uses as an objection against Kant), noting that Kant’s mistake consisted in failing to distinguish between these two meanings of ‘universality’ in his proofs to determine the content of the Categorical Imperative. Now the astonishing point I alluded to above is Allison’s claim that Wood’s “universality of applicability” (“Everybody should be equally subject to them” = the laws) is equivalent to Aune’s L, which the reader will recall is the formulation of the moral law that Kant supposedly holds: “conform your actions to universal law.” Aune elaborated this claim as follows in his (S3): ‘L is a higher-order principle telling us how to conform to certain lower-order laws.’ Rather, Wood’s ‘universality of applicability’ roughly corresponds to what I have called the legal character of a practical principle (in part II of this paper). Out of the connotations of the notion ‘law’ one can deduce, as I claimed there, that a moral law is a precept encompassing a broad circle of creatures. 19

See Wood (1990), 161 ff.

110

Harald Köhl

In my opinion, the prospects of Wood’s criticism are not encouraging if his “universality of applicability” means the same as Aune’s L. This is because, as I argued above, Kant not at all held L. Thus if Allison’s equation would be correct, Wood lacks something to set against his “universality of concern.” Regardless of whether or not Kant’s proofs to determine the content of his Categorical Imperative are correct, I find Wood’s distinction between „universality of applicability“ (as he understands it) and a “universality of concern” very helpful. But is it, as he understands it, a useful critical tool against Kant? – What Wood characterizes as ‘concern’ in his phrase ‘universality of concern’ sometimes is called by him “collective benefit” or “collective rationality.” He then uses this terminology to claim that “collective benefit […] is not identical with Kant’s FUL test [that is the basic formulation of the Categorical Imperative], which says that it must be possible to will without contradiction that the law will be universally followed“ (HET: 165 – my emphasis). Because “Kant’s conception of the law contains [only] universality of applicability, it does not contain anything like collective benefit” (ibid. – my insertion). True. But what looks to be a factual argument for the relative ‘emptiness of content’, or unfruitfulness, or the morally indeterminate character of Kant’s formulation of the law itself, Wood transforms into a criticism of Kant’s derivation of this formulation (without good reason, as I’d argue): “We cannot infer (as Kant does) that a principle with universality of applicability is eo ipso a principle with collective rationality. Such an inference is invalid.“ (HET: 166; my emphasis.) I’m unable to understand how Kant could have made this mistake in his derivation of the Categorical Imperative if he (according to Wood) should have been satisfied with a principle that merely possessed ‘universality of applicability.’ Before I return to Allison on a different point, I’d like to summarize the hitherto given critique of all three of Kant’s critics. They have allowed themselves to be mislead by the ambiguity of the term ‘Gesetzmäßigkeit’ (and related ways of speaking). ‘Gesetzmäßigkeit1’ is ‘the conformity (of an action or a maxim) to a (or the) Moral Law’, while ‘Gesetzmäßigkeit2’ means ‘the legislative form of a maxim’. Kant’s critics failed to clarify this ambiguity, consequently taking crucial places in Kant’s argument to mean ‘Gesetzmäßigkeit1’, when Kant was actually working with the notion of ‘Gesetzmäßigkeit2’. Thus they were able to import the idea of a plurality of laws into Kant’s derivation of the Categorical Imperative, a plurality to which maxims needed to conform. Kant was subsequently criticized for this idea. In fact, it is

The Derivation of the Moral Law

111

unintelligible where these laws could have originated. Such laws can only be generated by the application of the Categorical Imperative to maxims. But this Imperative can only be used after its formulation has been derived. The aforementioned laws cannot appear during the derivation. The gap in the derivation of the moral law that Kant’s critics believe they have found in GMS 402 and 420 f. cannot be found in the corresponding proof in the KpV (§§ 1–7). This, at least, is Allison’s claim. According to him, in the KpV the gap has successfully been bridged “by means of the introduction of transcendental freedom“ (144). Of course, Kant would have done himself, by introducing transcendental freedom at that place, a disservice. Who, except staunch Kantians, would still believe in the existence of a freedom which implies “independence from causal determination by everything empirical”? (151) Even Allison knows that if Kant’s so-called ‘metaphysical deduction’ (L. W. Beck) of the moral law rests on a “thick conception of transcendentally free agency”, that is “for better or worse” (154; my emphasis). If that is what grounds moral law! I take the assumption to be questionable that Kant used transcendental freedom as a premise in his proof in the KpV. I read this proof as follows: Kant goes from the thesis of § 4 to his conclusion in § 7. §§ 5/6, where transcendental freedom is introduced, is an interlude: In the ‘Theorem’ of § 4, Kant moves from the assumption “[i]f all material of a law, …, is abstracted from it, nothing remains except the mere form of a universal legislation” to the conclusion: “a rational being … must suppose that their mere form [= the form of maxims], through which they are fitted for giving universal laws, is alone that which makes them a practical law.” (KprV, 27; my insertion.) This form is prescribed for maxims in the “Fundamental law of pure practical reason” (§ 7) by the demand “One ought to act a certain way” (KpV, 30; my emphasis). The transition from § 4 to § 7 would be the transition that was, concerning the corresponding GMS-passages, criticized by Aune, Wood and Allision – an objection I challenged. So, on my view, what function does the interlude in §§ 5 and 6 have? For Kant a will “to wich only the legislative form of the maxim can serve as a law is a free will”: ‘free’ in the “transcendental sense” (KpV 29, § 5). “Granted that a will is free” [which I read as: if we can assume that …] then we can manage ‘Problem II’, which is to “find the law which alone is competent to determine it necessarily.” Because “the lawgiving form, insofar as it is contained in the maxim, is therefore the only thing that can constitute a determining ground of the [free]

112

Harald Köhl

will” (KpV 29, § 6 – my insertions). Kant here only adds to his claims in § 4 that the will in question is a free will. But it is not obvious to me, as it is to Allison, that this is an additional premise in the argument to the formulation of the law in § 7. In the ‘Remark’ to § 6 Kant seems to ask if we may presuppose the freedom of the will to arrive at a formulation of the moral law. If I am reading him correctly, his answer is ‘no’. The answer is ‘no’, because the “correct organization of our concepts” implies “that morality first reveals the concept of freedom to us” (KpV, 30; my emphasis). Kant’s “morality”, i. e. the ‘concept of morality’, I understand to be the searched-for formula of the moral law, that consequently cannot be deduced from the notion of transcendental freedom. – Looking back on § 5 this paragraph can be taken to mean: if one could begin with the notion of transcendental freedom, then one would know what the searched-for practical law was. However, as § 6 makes clear: that is an illegitimate move. I am uncertain on this point. The transition from § 4 over § 5/§ 6 to § 7 is confusing. I find the ‘Remark’ to § 6 even more confusing. If we could assume that my confusion has an objective correlate, I would diagnose Kant’s argument as follows: In the ‘Remark’ to § 6, Kant is introducing considerations that do not belong in the flow of arguments to ascertain the formulation of the moral law. Rather, they belong to Kant’s attempt to answer the question how the demand expressed by this law may be grounded. That this project of GMS III can’t be successfully carried out, is the thesis in the KpV.20 “Thus freedom and unconditional practical law reciprocally imply each other” we read in the ‘Remark’ to § 6 of the KpV. Allison is surely right in identifying this as the ‘Reciprocity Thesis’ (151 – his term) that Kant already formulated in GMS III and according to which “a free will and a will under moral laws are one and the same” (GMS, 447). Freedom and morality are “reciprocal concepts” for “one single concept” (GMS, 450). But that is precisely the point: this thesis belongs in the context of an attempt to ground the moral law. That this context is not present in the KpV was obviously of no consequent for Allison. In the GMS Kant still believed he can presuppose freedom for his argument: 20

See my reconstruction of Kant’s ‘Deduction’ of the categorial imperative’ in: Köhl, 2004, 129 ff.

The Derivation of the Moral Law

113

If therefore, the freedom of the will is presupposed, morality together with its principle follows from it by mere analysis of its concept. … Freedom must be presupposed as a property of the will of all rational beings. (GMS, 447)

Now, in the KpV, Kant no longer believes this is legitimate. My impression is that Kant was here so bent on correcting his position of the GMS that he inserted the subject of the (now believed to be impossible) deduction of the moral law in an argument that was concerned with something else: not to ground the Categorical Imperative (the demand it expresses) but to first of all discover its formulation. In his search for an understanding of Kant’s Search for the Supreme Principle of Morality (so the title of his illuminating book21), Samuel J. Kerstein unfolds what he calls “a new, criterial reading” (12) of Kant’s derivation of the moral law. His critique of the gap-theorists (Aune, Wood, Allison) is preceeded by the following reconstruction concerning both of the Kantian proofs in GMS I and II: First, Kant develops criteria that any viable candidate for the supreme principle of morality must fulfil. […] Second, Kant tries to establish that no possible rival to the Formula of Universal Law fulfils all of these criteria. Finally, Kant attempts to demonstrate that the Formula of Universal Law remains as a viable candidate for a principle that fulfils all of them. With these three steps, Kant strives to prove that if there is a supreme principle of morality, then it is this formula (12, cf. Kerstein’s ‘Summary’: 93).

The criteria developed by Kerstein in step one constitute (cf. 12) what he calls Kant’s “basic concept” of the requirements that have to be fulfilled by a supreme moral principle: “It would be practical, absolutely necessary, binding on all rational agents, and would serve as a supreme norm for the moral evaluation of action” (1). The first three of these criteria are in accordance with my own claim, concerning Kant’s proof in Groundwork II, that the moral law contains the highest categorical prescription for action, which is simultaneously universal (in part III of this paper). Concerning Kant’s derivation in GMS I, Kerstein contends (as I did above in part IV): “Kant holds his discussions of the good will, [of acting from] duty, and moral worth” to be essential for his proof, adding that he (and Kant) believe this discussion “to be necessary for the development of criteria for the supreme principle of morality” (86; my emphasis and insertion). Considering Kant’s proof in Groundwork II 21

Cambridge 2002. References in the following paragraphs are to this book.

114

Harald Köhl

Kerstein finds it “elliptical” (93). Surprisingly, he introduces an additional criterion as a constituent of the derivation: “The derivation is not complete unless ‘all imperatives of duty’ can be derived from the imperative Kant proposes as the only viable candidate for the supreme principle of morality” (87; cf. 88). Compared with this interpretation, my own reconstruction in part III has been more text-immanent. Further analysis of the differences between Kerstein’s and my analysis of Kant’s attempts to derive the correct formulation of the moral law would be beyond the scope of this paper. VI. Critical Considerations Let us assume for the sake of the argument that I have succeeded in parts II–IV of my paper to plausibly reconstruct Kant’s arguments to determine the content of the Categorical Imperative into a conclusive argument. Then why does Kant fail to persuade us? Which of his premises are problematic? 1. Surely Kant’s notions of the moral ‘law’ and moral ‘laws’ are alien to us. For his predecessors22 and contemporaries this was different. Kant’s talk of such ‘law(s)’ becomes problematic when he exploits the connotations of these expressions for certain philosophical purposes. So, for example, the argument that he presents in the “Preface” to the GMS for the benefit of a non-empirical understanding of ethics (GMS, 389) trades on his rhetoric about ‘law.’ Because only the talk of a ‘law’ makes it seem compulsory, that it “carr[ies] with it absolute necessity… does not hold only for human beings …[but also] for other rational beings” (ibid. – my insertion). By comparison, Kant’s mention of a “command” (e. g. “thou shall not lie”, ibid.) does not so clearly express an (unconditioned) necessity to act in accord with it (my emphasis). 2. Kant’s presumption of the scope of validity for a moral law is questionable as well. I would like to admit that the connotations of the expression ‘law’, and the description of the law as the ‘highest’ moral principle, imply that this precept is valid for an encompassing circle of creatures (see II.3). But whether this circle consists out of ‘all rational creatures’ is doubtful. It is even problematic that this law “does not hold for humans only” but for all humans (ibid.). At most, it could apply to all ‘rational’ or ‘competent in an xy-way’ or ‘born with capaci22

“The natural law theorists thought only a morality built around a specific concept of law and obligation would be serviceable.” Cf. Schneewind (1998, 518).

The Derivation of the Moral Law

115

ties xy’ humans. Since the use of ‘rational’ is not without alternatives, Kant’s move to ‘other rational creatures’ (and hence to all rational creatures) qua addresses of the law is also not necessary. 3. The way in which Kant introduces the definition of a ‘law’ in GMS I – “Duty is the necessity of action out of respect for law” – makes clear that he presumes that there can only be one highest moral principle. His description of the function of the law as the highest criterion of moral judgment, and his distinction between the Categorical Imperative and categorical imperatives (= laws) make clear that Kant thinks of a hierarchy of moral principles with one supreme principle. The corollary of this conception is a deductive model of moral justification (from top to bottom). This hierarchical conception of an ethical theory is optional. Some philosophers have tried to plausibly argue that we can do without a highest moral principle.23 Beyond this, there is a more plausible alternative to Kant’s conception of ethics. I am speaking of a holistic ethics, such as Ludwig Siep24 has argued for recently, and as it has been introduced to me by J. B. Schneewind. His essay “Moral Knowledge and Moral Principles”25 is an ethical counterpart to W. V. O. Quine’s epistemological holism in the concluding part of his “Two Dogmas of Empiricism”.26 Kant’s hierarchical model of moral principles and his modell of deductive reasoning in ethics raises a question about the justification of a highest principle (the law). It cannot be grounded in a higher principle. How then? It is doubtful that Kant’s justification of the moral law in GMS III was successful or that his doctrin of the ‘fact of pure reason’ in the KpV does away with the problem.27 If, furthermore, one is not impressed by attempts of ultimate justification (like those by K.O. Apel), an ‘ethical holism’ will be even more attractive. According to this position, moral laws do not necessarily need to be grounded in higher laws, and ultimately in a or the law. Because principles, including a highest principle, can also be justified ‘from the bottom’: by demonstrating that they summarize lower-level principles, moral commands and assessments in a fruitful way. Speaking more generally, moral principles, including the highest, can be justified by ap23 24 25 26 27

See Larmore (1987), Chapter VI, as well as Patzig (21983). See Siep (2004), as well as Köhl (2006), Chapter 6.4. See Schneewind (1970). See Quine (1953). For a criticism of the Doctrine of the ‘fact of pure reason’, see Köhl (2001), part VI, as well as Köhl (2006), Chapter 6.3.

116

Harald Köhl

pealing to their inferential relations to the total inventory of an ethical theory – insofar as aspects of this theory seem to be relevant in the particular case. For a pragmatist like Schneewind it is self-evident that the tools of moral justification are used relatively to concrete problem-situations. It results from these situations which of our resources are relevant in the particular case and how the ‘patchwork’ of ethical theory is to be arranged. A hierarchical order with one principle at its apex may turn out to be useful, but it is not necessarily so. Often our moral reasoning will be deductive, but it is not necessarily so. 4. Kant’s attempt to deduce the formula of the Categorical Imperative out of the character of a categorical imperative (420 f.) implicitly contains the premise that ‘categorical’ imperatives are always ‘moral’ demands. I hope to have shown in “Moral und Klugheit“ and in Abschied vom Unbedingten. Über den heterogenen Charakter moralischer Forderungen 28, that the equation moral = categorical is wrong. Rather, there are non-moral demands that are categorical in structurally the same sense as (many) moral demands. – The falsity of that premise could imply what many suspect anyway: that the Categorical Imperative is not specifically a moral principle, but simply a consistency principle whose application cannot bring about morally significant results. 5. Is a material principle for action really precluded from being a moral law? Such a principle would be a hypothetical imperative. So what? In the texts just cited I have tried to make the ‘right to existence’ of hypothetical moral precepts (next to the existence of moral categorical imperatives) plausible. So why should not also the highest moral principle (the law) be hypothetical? The moral law cannot aim at happiness: I will for the moment admit this and its justification to Kant. But what precludes a law to be moral that would demand actions for realizing virtue (or a set of virtues)? It may be that such a conception is open to grave objections. But that is not the point. It matters only whether such a law could sensibly be called a ‘moral’ (vs. non-moral) law. 29

28 29

See Köhl, 2006. I am grateful to Christoph Horn and Dieter Schönecker for their efforts to realize the translation of my paper.

The Derivation of the Moral Law

117

Literature Kant’s writings Kant’s Grundlegung zur Metaphysik der Sitten (= GMS; engl.: Groundwork of the Metaphysics of Morals) is cited according to the Akademie-Ausgabe and its pagination, volume IV (page numbers inserted into the text without abbreviation refer to the GMS). The Kritik der praktischen Vernunft (= KpV; engl.: Critique of Practical Reason) is cited according to volume V of the Akademie-Ausgabe and its pagination. English translations are taken from The Cambridge Edition of the Works of Immanuel Kant: Practical Philosophy (New York, 1996: Cambridge University Press).

Other works Allison, Henry E. (1996): “On a presumed gap in the derivation of the categorical imperative,” in: Idealism and Freedom, Cambridge, 143–154. Aune, Bruce (1979): Kant’s Theory of Morals, Princeton. Köhl, Harald (1990): Kants Gesinnungsethik, Berlin/New York. Köhl, Harald (2001): “Moral und Klugheit. Rortys Kritik an einer kantischen Unterscheidung”, in: Deutsche Zeitschrift für Philosophie 49, 19–41. Köhl, H. (2004), “Kant’s ‘Grundlegung’: neu ediert, kommentiert, interpretiert” in: Philosophische Rundschau 51. Köhl, H. (2006), Abschied vom Unbedingten. Über den heterogenen Charakter moralischer Forderungen, Freiburg i. Br. Larmore, Charles (1987): Patterns of Moral Complexity, Cambridge. Paton, Henry. J. (1971): The Categorical Imperative: A Study in Kant’s Moral Philosophy. Philadelphia: University of Pennsylvania Press. Patzig, Günther (21983): “Ein Plädoyer für utilitaristische Grundsätze in der Ethik”, in: Patzig: Ethik ohne Metaphysik, Göttingen, 127–147. Quine, Willard V. O. (1953): “Two Dogmas of Empiricism”, in: Quine, From a logical point of view, Cambridge: Harvard University Press, 20–46. Schneewind, Jerome B. (1970): “Moral Knowledge and Moral Principles”, in: G. A. Vesey (Ed.), Knowledge and Necessity, London / New York. Schneewind, Jerome B. (1998): The Invention of Autonomy, Cambridge. Schönecker, Dieter / Wood, Allen W. (22004): Kants “Grundlegung zur Metaphysik der Sitten.“ Ein einführender Kommentar, Paderborn. Siep, L. (2004), Konkrete Ethik, Frankfurt a. M. Tugendhat, Ernst (1993): Vorlesungen über Ethik, Frankfurt a. M. Wood, Allen. W. (1990): Hegel’s Ethical Thought, Cambridge. Wood, Allen W. (1999): Kant’s Ethical Thought, Cambridge.

Groundwork II

Marcus Willaschek

Practical Reason A commentary on Kant’s Groundwork of the Metaphysics of Morals (GMS, 412–417)1 1. Introduction On pages 412–417 of the GMS, Kant introduces his conception of the will as practical reason and the closely related distinction between hypothetical and categorical imperatives in a rather concise manner. In the following, I will address the much discussed definition of the will (4: 412), the concepts of both the holy will (4: 412 und 414) and the imperative (4: 413–414), as well as the differences among the different kinds of imperatives (4: 414–417), while explaining Kant’s central concepts and theses. Prior to this, however, I will present a short overview of Kant’s general line of argument. 2. Overview Practical reason is the ability to act rationally – that is, the ability of a person to rationally coordinate her goals and ends and to orient her actions according to these rationally set ends. Kant identifies this ability with the will, or with the ability of a rational being (more precisely, a being possessing reason) to be the cause of its conduct through its own “representations” (mental states). In other words, having a will means being able to act in accordance with one’s own rational representations. Within the concept of practical reason, Kant distinguishes pure from empirically qualified practical reason – a distinction that he first 1

For their valuable advice and criticism, I would like to thank Alexander Bagattini, Steffi Schadow, Dieter Schönecker as well as the participants of the preparatory conference in Bonn.

122

Marcus Willaschek

makes explicit in the KpV, and that is already anticipated (and implicitly presupposed) in the GMS (cf. 4: 389, where Kant speaks of a “practical pure reason” as well as of the question of how “pure reason can be practical,” 4: 458; cf. 461). God and the angels (if they exist) possess exclusively pure practical reason; their decisions are in no way influenced by subjective factors, and particularly not by sensible (sinnliche) inclinations, but solely by objective, and specifically moral, rational grounds. Each of their wills is a “holy will,” which necessarily and exclusively seeks the moral good. Humans, on the other hand (as the only rational beings of whose existence we have knowledge), are influenced by their naturally and socially conditioned wishes and inclinations in many ways. Their wills are, therefore, not holy, for when it accommodates their interests they can decide to act immorally. Their practical reason is therefore empirically qualified: as instrumental rationality, it serves to satisfy empirically-given inclinations in the most effective way possible. If, on the other hand, as Kant supposes, moral principles hold without exception and necessarily for all beings who possess reason, thus also for humans, then it must be rational for every rational being to follow these principles, independent of its respective inclinations or wishes. Accordingly, humans must not only have empirically qualified reason, but also pure practical reason; that is, they must be able to orient their action according to moral principles when this would not further, or would even contradict, their own interests. The human will is thus distinguished from a holy will in two ways: first, our decisions do not necessarily accord with that which would be rational to do, and second, “rational” for us does not only mean “morally good,” but also “useful for the satisfaction of subjective inclinations.” The first difference leads to the concept of an imperative, the second to the distinction between hypothetical and categorical imperatives. Imperatives are propositions in which what is good and reasonable to do is expressed in the form of a command. Such commands are not directed at a being with a holy will that necessarily does what is rational, but rather at beings such as humans, who can act rationally but who do not necessarily do so (i. e., those for whom both rational and irrational actions are possible in any given situation). Most people are thus capable of doing something because they have recognized it as rational (e. g. taking care of their health or paying their taxes), but this rational insight does not necessarily motivate them to a corresponding course of action, because their inclinations (e. g. towards indolence) or their self-interest may oppose it. The principles of rational action, therefore,

Practical Reason

123

appear for human beings as commands that express themselves in an ought: “(From a rational point of view) you ought to live healthily” or “(From a rational point of view) you ought not cheat on your taxes.” This “ought” is not an expression on the part of a power or convention (as is the case when one says that someone ought to follow the command of a superior or ought to greet one’s neighbors), but an expression of a specifically rational form of motivation, which Kant describes as rational “necessitation” (Nötigung). This necessitation serves as a motive for doing what is rational even when we find it difficult (and must therefore “force” ourselves to do it). The more rational one is in a practical respect, the more fully developed will this motive be. By prescribing an action, imperatives distinguish it as in some way “good.” Corresponding to the distinction between a pure and a practically qualified reason, there are also two kinds of imperatives: those that demand an action because it is “good” as a means to the effective satisfaction of inclinations and needs, and those that prescribe an action because it is “rational per se” or “good in itself,” without taking into account the ends of the agent. The former are hypothetical imperatives and the latter, categorical imperatives. Kant divides the hypothetical imperatives again into two classes, depending on whether they prescribe an action as a means to a possible or an actual end. Thus Kant arrives at a tripartite division of imperatives, giving several alternative descriptions of each of the three classes, depending on which of their characteristics he wants to emphasize: (1) problematically practical principles (= imperatives/rules of skill, technical imperatives), (2) assertorically practical principles (= rules/counsels of prudence, pragmatic imperatives), (3) apodictically practical principles (= imperatives/laws of morality, moral imperatives). (1) and (2) are the hypothetical, (3) the categorical imperatives. 3. The definition of the will (4: 412) 2 Kant defines the will as the “capacity to act in accordance with the representation of laws, that is, in accordance with principles,” and he identifies this capacity with practical reason. This passage has given rise to a variety of different interpretations (cf. Laberge 1989; Tim2

The following translations of passages from Kant’s works are based on the Cambridge Edition of the Works of Immanuel Kant (New York: Cambridge University Press, 1992-).

124

Marcus Willaschek

mermann 2003, § 9), and even if one takes the context into account the passage remains ambiguous. Kant begins with this observation: (1)

Everything in nature works in accordance with laws.

This claim follows from the regularity conception of causation that Kant inherited from Hume (cf. Willaschek 1992, 34–43). According to this view, a at time t0 is the cause of b at time t1, if a is of type A and b is of type B, such that from things or events of type A at t0 there always follow things or events of type B at t1. In that case, “If A at t0, then B at t1” is a universally valid law. Thus the claim holds analytically that all things in nature “work in accordance with laws,” or that their causality functions according to laws of nature. (2) Only a rational being has the capacity to act in accordance with the representation of laws, that is, in accordance with principles, or has a will.

The will is thus a capacity (i. e., an ability) of a rational being to act in a specific way, namely “in accordance with the representation of laws.” But what does it mean to act “in accordance with the representation of laws”? Which laws are meant here? It is exactly with regard to this question that interpretations diverge. In order to understand Kant correctly here, it will be helpful to review the connections among the concepts life, faculty of desire (Begehrungsvermögen), and will (cf. with regard to this and what follows Willaschek 1992, 82–90). “Life,” as Kant writes in the KpV, “is the faculty of a being to act in accordance with the laws of the faculty of desire. The faculty of desire is the faculty of this being to be, through its representations, the cause of the actuality of the objects of these representations” (5: 9; cf. already in 2: 327 as well as 4: 544). As is briefly stated in the MdS, “The capacity of a being to act in accordance with its representations is called life” (6: 211). Here, “to act” has the broad meaning of the Latin “agere”: it means the same as “to bring about” or “to cause” (cf. Gerhardt 1986). Kant understands life as the purposeful spontaneous activity (Selbsttätigkeit) of an organism (or in the case of a “living God,” cf. 3:421, the activity of a non-material entity) in accordance with the Aristotelian-Leibnizian concept of entelechia. This spontaneous activity, however, is conceived (at least in the case of physical beings) in a distinctively “modern” way as a lawful and purely causal association. A living being also “works” according to laws: it has a representation of what it desires and this representation motivates it to act in a way that leads to the realization of the

Practical Reason

125

desired object or goal. The causal connection between representation and bodily movement has the form of a natural law; more specifically, it conforms to a psychophysical lawfulness, a “law of the faculty of desire.” Because the desire comes from the living being itself, Kant can also say that it acts “from an inner principle” (4: 544). As living beings, rational beings act or work “according to laws” – and specifically according to those of the (rational) faculty of desire. At the same time, they act “according to their representations,” and thus “according to the representation of laws.” But the laws according to the representation of which rational beings act, according to the GMS, are not the “laws of the faculty of desire” according to which they “act” or “work” as living beings. As we have seen, the latter are causal laws that describe the regular connection between the mental states (“representations”) and bodily movements of a living being. Like all natural laws they hold independently of whether or not a particular living being represents them to itself. Even if Kant’s formulation might seem to suggest something else, the laws mentioned in sentence (1) are of a completely different kind than the laws of which Kant speaks in sentence (2). But what then are these laws “according to the representation of which” rational beings act? The answer does not unambiguously emerge from the aforementioned passage, but the context lends credence to the supposition that they are those “objective laws” of reason which are discussed at the end of the paragraph and in the paragraphs that follow (4: 413). What is of concern here is not only the moral law and other, more specific laws of morality, but also the “laws” of instrumental action, for, as Kant says, “All imperatives… indicate… the relation of an objective law of reason to a will that because of its subjective constitution is not necessarily determined by it” (4: 413, my emphasis). Even though Kant usually understands “practical laws” to be moral laws (cf. 5: 19), it seems that a broader reading is appropriate in this context: the laws according to whose representation only rational beings can act are the laws of rational action, whether they be laws of skill, pragmatic laws, or moral laws. Kant calls the laws themselves (but not their “representation”) “principles” or “objective principles” (cf. e. g. 4: 400 note, 4: 413). The addition “that is, according to principles” can consequently not refer to the entire preceding expression, “according to the representation of laws,” for laws or principles are something quite different from representations of laws or principles. If the explanation beginning with “that is” (“d. i.”) refers only, however, to the expression “laws,” then

126

Marcus Willaschek

the repetition of the words “according to” (“nach”) creates difficulties, for it would then be both the representation of laws and the laws (principles) themselves according to which rational beings act. This difficulty is resolved, however, if one considers what it means for Kant that a rational living being acts “according to a principle”: it has a representation of the principle in question, which motivates it to realize the represented object (the principle). Someone who acts according to the representation of a law (a principle), ipso facto acts according to the law (principle) itself. Sentence (2) should then be understood in the following way: (2') Only rational beings can orient their behavior according to the representation of the laws of rational action, and thus act according to these laws or principles themselves.

One can also say this more concisely: only rational beings have the ability to act rationally, in both the instrumental and moral sense. Kant calls this ability for rational action and effort the “will.” The human will, as one can also say with Kant, is a rational “faculty of desire in accordance with concepts” (6: 213); the “representations” through which we become cause of the actuality of represented objects are not (only) of an intuitive and sensible kind, but are (also) conceptual. Kant calls a concept that is at the same time the cause of the object represented through the concept an “end” (5: 180). While the definition of the “will” as the rational faculty of desire corresponds with the philosophical tradition, and especially with the vocabulary of the Leibnizian-Wolfian school of philosophy (cf. e. g. Baumgarten, Metaphysica § 690), in the sentence that follows Kant chooses an at least terminological path of his own: (3) Since reason is required for the derivation of actions from laws, the will is nothing other than practical reason.

The expression “practical reason” takes up the Aristotelian language of nous praktikos, but has no modern-language predecessor in the 18th century (cf. Beck 1974, 265, note 14). Hence Kant was free to define this expression according to his own needs. Nevertheless, Kant apparently does not want to give a stipulative definition, as he justifies the identification of the will and practical reason with the remark that “reason is required for the derivation of actions from laws.” One might grant Kant that “derivation” is the task of reason; at another point he expressly identifies reason with the “faculty of the determination of the particular through the universal (the derivation of prin-

Practical Reason

127

ciples)” (20: 201). However, that means only that one cannot have a will without also having reason (cf. 4: 427: “The will is thought as a capacity to determine itself to action in conformity with the representations of certain laws. And such a capacity can be found only in rational beings.”) This does not mean, however, that the will “is nothing other” than practical reason. For this, rather, two claims would have to hold, claims that Kant implicitly presupposes: first, that the will (as the capacity to act according to the representation of laws of reason) “is nothing other” than the capacity of the “derivation of actions from laws,” and second, that for this capacity, practical reason is not only necessary (Kant: “is required”), but also sufficient. The talk of a “derivation of actions” refers to Aristotle’s practical syllogism in which a general major premise (“Everything sweet must be tasted”) and a minor premise (“This is sweet”) are followed directly by the execution of the rationally commanded action (cf. EN 1147a 25 ff.). Kant also assumes that practical reason does not only consist in deriving propositions or opinions about actions but also the actions themselves: practical reason is the ability to orient one’s behavior according to the laws of reason. That is, it is not only the ability to recognize what is rational (the so-called “principium diiudicationis”), but to also do it (“principium executionis”). However, Kant’s conception differs from the Aristotelian view in one important respect: practical reason does not produce the “derived” actions directly (i. e., simply through insight into their rationality), but indirectly, via a feeling. This feeling, however, which in the case of action according to moral laws Kant terms “respect for law” (cf. 4: 400, 440) or “moral feeling” (4: 460), is itself of rational origin, for it arises through insight into the rationality of moral laws (cf. 4: 460, 5: 76). In Kant’s view practical reason delivers both objective reasons and the subjective “incentives” of rational action and is thus nothing other than the will or the ability to act rationally. This identification of practical reason and the will raises a problem, however, when Kant directly afterwards (and in many other passages) speaks of reason “determining” the will (more or less effectively). How can this be if the two are identical? The answer lies in the ambiguity of the expressions “will” and “practical reason” (cf. on this and the following Willaschek 1992, 48–53). In the GMS, as we have already seen, Kant defines “will” and “practical reason” as the ability to act rationally. What is at issue there, however, is a complex ability. It includes on the one hand the ability to bring the totality of one’s wishes and convictions into a rational order (which guarantees

128

Marcus Willaschek

consistency and completeness). Kant calls this ability reason and, insofar as it concerns primarily wishes, ends, and grounds for action, practical reason or will: “The will is therefore the faculty of desire considered not so much in relation to action (as choice (Willkür) is), but rather in relation to the determining ground of this choice of action. The will itself actually has no determining ground, but insofar as it can determine choice it is instead practical reason itself” (6: 213, my emphasis). In this passage from MdS Kant does not understand “will” and “practical reason” as he does in GMS in the broader sense, as the ability to act rationally, but rather in the narrower sense, as the ability to produce rational determining grounds for action. On the other hand, practical reason in the broad sense also includes the ability to actually act according to these rational grounds of action. This ability is what Kant calls “free choice” (freie Willkür). Choice is the capacity to act according to one’s own representations, which are connected with pleasure and displeasure (Lust and Unlust) (6: 413); choice is free insofar as pleasure and displeasure do not determine behavior but rather serve as a basis for the rational establishing of ends (cf. KrV A 802/ B 830 and, with a slight change of emphasis, 6: 413). Practical reason or will in the broader sense as the ability to act rationally includes, first, practical reason or the will in the narrower sense (that is, the ability to bring one’s wishes and ends into a consistent scheme of grounds for action) and, second, “free choice” (that is, the ability to do what is rationally desired). The will, choice, and practical reason are therefore not distinct causal instances or subject-like homunculi, but instead aspects of the complex ability to act rationally. When Kant speaks of the fact that reason determines the will or choice, he means that practical reason in the narrower sense (the capacity to set rational ends) provides the “determining grounds” for the will in the broader sense (the ability to act rationally) or free choice (the ability to act rationally). That reason determines the will thus means – depending on emphasis – either that one actually does what is rationally desired, or that what one does is actually rational. When Kant, however, equates reason and the will, then he means either (in the broader sense) the ability to act rationally or (in the narrower sense) the ability to establish ends rationally. Although Kant’s use of words is confusing and not always uniform, his conception of the will as practical reason is in fact uniform and thoroughly comprehensible. A further meaning of the term “will” in Kant has not yet been mentioned: namely, according to Kant the will is not only the ability

Practical Reason

129

to act according to laws or principles and thus rationally, but rather, as free will, also the law-giving faculty (cf. 6: 226). This means that a free rational being can not only orient its decisions and actions according to the rational principles that it is given and adherence to which serves the satisfaction of subjective inclinations, but that it also has the ability to “give” itself these principles according to which it acts and decides. Kant calls this ability “autonomy” or self-legislation (Selbstgesetzgebung) (4: 440 ff.) The fact that a rational being “gives” itself the principles according to which it acts is obviously not intended to mean that the recognition of these principles is the result of a conscious and deliberate decision that could also have turned out otherwise. It means that the rationality of a rational agent consists in the recognition of these principles; they are not given to her by a foreign authority (heteronomy), but are rather an expression of her own reason. While Kant in the GMS takes a heteronomous will to be conceivable (4: 441), in the MdS he seems to almost identify the will with the ability of selflegislation (6: 226). 4. The concept of a holy will (4: 412 and 414) Kant’s transition from the definition of the will to his conception of the imperative is constituted by the concept of the holy will, which he introduces in two steps: “If reason infallibly determines the will, the actions of such a being that are recognized as objectively necessary are also subjectively necessary, that is, the will is a capacity to choose only that which reason independently of inclination recognizes as practically necessary, that is, as good.” (4: 412). Soon afterwards, this necessarily good will turns out to be a “holy will”: “A perfectly good will would, therefore, stand just as much under objective laws (of the good), but it could not on this account be represented as necessitated to actions in conformity with law, since it of itself, according to its subjective constitution, can only be determined by the representation of the good. Hence no imperatives hold for the divine will and in general for a holy will: the ‘ought’ is out of place here, because volition is already of itself necessarily in accord with the law” (4: 414). A holy will is thus a will which “of itself, according to its subjective constitution” can do nothing other than the rational and good. This conception raises a number of questions, only two of which I want to consider more closely here: (1) Is the conception of a will that wills “only” the objectively rational, i. e., the morally good, consistent? (2) And how

130

Marcus Willaschek

does this concept of a “perfectly good will” relate to that of the “good will” that is the topic of the beginning of the fi rst section? (1) A holy will, in contrast to a human will, is a “capacity to choose only that which reason independently of inclination” recognizes as good; such a being “can be determined only through the representation of the good.” These formulations suggest that a being can only have a holy will if its will is not “affected” by inclinations, but instead acts exclusively according to the objective principles of reason. This becomes clear, too, through comparison with the human will for which, in contrast to the holy will, reason alone “does not adequately determine the will” and thus which “is also subject to subjective conditions (certain incentives)” (4: 412). Thus the holy will would be subject to no subjective conditions, such that reason alone would determine it sufficiently. Similarly, Kant writes in the KpV that the moral law has for humans “the character of an imperative, because in them, as rational beings, one can presuppose a pure will but, as beings affected by needs and sensible motives, not a holy will, that is, a will that would be incapable of maxims conflicting with the moral law” (5: 32). In this passage as well it sounds as if a being with a holy will may not be affected by “needs and sensible motives,” because it is this state of being affected that prevents humans from attaining this holy will. Such a view would be awkward, however, since a being without “inclinations” would be incapable of action. Rational principles alone are not sufficient to establish how one should act in a certain situation. With regard to the principle of instrumental action (“He who wills the end, also rationally wills the necessary means to that end”) this is obvious: it presupposes that one already has ends; what these ends are does not only depend on rational considerations, but also on that to which one has an “inclination.” Even moral principles, and above all the categorical imperative, are by themselves not sufficient to be taken to show a certain action to be rationally commanded. After all, one is supposed to test, using the categorical imperative, if the maxim by which one intends to act could be a universal law. Maxims, however, are subjective rules of action, which “reason determines according to the conditions of the subject (more often his ignorance or also his inclinations)” (4: 421; my emphasis). A being that has no inclinations (and is also not subject to any other “subjective conditions”) has also, therefore, no maxims, and thus nothing to which it could apply the categorical imperative. Thus, even a being with a

Practical Reason

131

holy will needs inclinations (or other not purely rational “subjective conditions”) in order to be able to act rationally. Now in the passage cited above, Kant does not explicitly exclude the possibility that a being with a holy will could have inclinations, but instead says only that inclinations may have no influence on its will. As we have just seen, however, even this is problematic: a “capacity to choose only that which reason independently of inclination recognizes as practically necessary, that is, as good,” is not enough to decide on concrete actions and would therefore not be a will. Kant’s characterization of the holy will as completely independent of inclinations is thus too strong. However, what is important for Kant in the concept of the holy will is not its independence from inclinations, but its necessary correspondence with the laws of reason and of the good: “The will whose maxims necessarily harmonize with the laws of autonomy is a holy, absolutely good will” (4: 439). And for this necessary correspondence, in contrast to what Kant’s formulations suggest, a complete independence from inclinations is not necessary. A being whose will is capable of being affected by inclinations, but whose inclinations necessarily (because of its intrinsic constitution) correspond to the laws of reason, would have maxims, but “would not be capable of any maxims conflicting with the moral law” (5: 32) and would thus have a holy will. In fact, many passages speak for the thesis that Kant was of the opinion that a holy will could well be affected by inclinations as long as these necessarily corresponded with the laws of reason. Kant writes in the MdS that in the case of a “holy (superhuman) being… no hindering impulses would impede the law of its will” (6: 405). And in lecture notes it is stated, “holiness is the absolute or unlimited moral perfection of the will. A holy being must not be affected by the least inclination against morality” (29: 1075). The addition “against morality” obviously only makes sense if it is at least conceivable that a holy being is at all affected by inclinations. Kant continues: “It must be impossible for him to want anything that would be contrary to the moral laws. Understood thusly, no being aside from God is holy; as every creature always has some needs, and if it wants to satisfy them, then inclinations as well, which do not in each case correspond with morality” (29: 1075). The holiness of a will thus consists in its necessary correspondence with the laws of reason and morality. This does not exclude the possibility that a being with a holy will is affected by inclinations, but only the possibility that it is affected by inclinations that could motivate it to do something

132

Marcus Willaschek

irrational and immoral. (The inclinations of a divine being – divine love, for example – would not be sensibly conditioned like ours, but would have non-sensible origins.) (2) This takes us to the second question: How does the concept of the holy will relate to that of the good will, with which Kant’s Grundlegung begins? Obviously, a holy will is also a good will, indeed even a “completely good” will (4: 414). But is the good will also always a holy will? Initially it seems plausible to assume that a good will does not necessarily have to be a holy will, but that a human will can also be good. However, upon looking more closely it is anything but clear how a good will differs from a holy will and thus whether humans can ever possess a good will. Kant characterizes the good will as the only thing that is good “without limitation” (4: 393) or “in itself” (4: 394). It has an “absolute worth” (4: 394) that is independent of whether or not what is willed is also achieved. 3 What then constitutes a good will? Wood distinguishes between two possibilities: (1) the correspondence of the maxims with the moral law and (2) action from duty. In the first case the will of a being would be a good will if this being acted “in accordance with duty”; in the second case it would be good if it acted from duty. Because Wood understands action from duty as action from moral self-constraint, he prefers the first possibility. In fact it is implausible to assume that only a being which has to force itself to perform good deeds has a good will, for then God could not have a good will. But viewing the good will, with Wood, as a will whose maxims merely correspond to the moral law, and who thus acts in accordance with duty, one aspect of the good will to which Kant attached great importance is lost – namely, the reliability and consistency of the good will. A will is not good when it accidentally chooses the right maxims (4: 426) and thus acts in accordance with duty, but instead when it does this from an inner principle that guarantees action in accordance with duty under all circumstances. This principle is the moral law: as Kant says, for a being with a good will the “objective principle” (the moral law) serves “subjectively” as its “practical principle” (cf. 4: 400 note). In other words, a being with a good will acts according to the maxim to act only in accordance with such maxims that could be universal laws (cf. Willaschek 1992, 67 ff.). Whoever acts according to this (meta)maxim, acts out of “respect for the law” and thus “from duty,” for 3

Cf. on this subject Allen Wood’s contribution to this volume.

Practical Reason

133

“duty is the necessity of an action done out of respect for the law” (4: 400). The individual actions do not then accidentally correspond with the moral law; rather, they do so necessarily. When Kant returns to the good will at the end of the second part of the GMS, it is exactly this necessity of correspondence with the moral law that he emphasizes: “We can now end where we set out from at the beginning – namely, with the concept of an unconditionally good will. That will is absolutely good which cannot be evil, hence whose maxim, if made into a universal law, can never conflict with itself. This principle is, accordingly, also its supreme law: act always on that maxim whose universality as a law you can at the same time will” (4: 437; my emphasis). A will is thus good if it cannot be evil, and therefore cannot be evil because it has made correspondence with the moral law into the “supreme law” or, as Kant says in another passage, into the “supreme condition of all maxims” (5: 31; cf. 6: 36). If this interpretation of the concept of the good will is legitimate, then the question of the distinction between the good and the holy will becomes all the more urgent, since for the holy will Kant also emphasizes the necessity of correspondence with the moral law, and he equates the holy will with the “completely good will” (4: 414) and the “absolutely good will” (4: 439). Nevertheless, there is in fact a distinction between a good human will and a holy will: in contrast to God, even humans with the best wills feel inclinations whose satisfaction is possibly incompatible with the moral law. These inclinations constitute a “temptation” to action adverse to duty, and must therefore be suppressed or overcome. This danger of moral corruption that is an irrevocable part of the human condition is what Kant in the Religionsschrift calls “radical evil” (6: 32 ff.): even the best human is radically evil (that is, evil “at the root”), since he can feel morally adverse inclinations (such as indolence, envy, selfishness, etc.), which, if they do not prevent him from following the moral law, at least make it more difficult. As we have already seen, however, in the case of a holy will all impulses are from the outset and necessarily in harmony with the demands of morality so that there is no temptation that would have to be overcome: in the case of a “holy (superhuman) being […] no hindering impulses would impede the law of its will,” and it “would thus gladly do everything in conformity with the law” (6: 405). A good will is thus one that does what is morally required not simply accidentally, but out of firm resolution and lasting principle; a good will is holy if it does not have to overcome any inner resistance

134

Marcus Willaschek

(in the form of inclinations) to do so, but rather does the good, under any conceivable circumstances, “gladly.” (To what extent holiness of the will can be demanded of humans, as Kant claims in the KpV and the MdS, is an additional problem that cannot be treated here; cf. on this Allison 1990, ch. 9). 5. The concept of the imperative and the division of the imperatives (4: 413–417) “The representation of an objective principle, insofar as it necessitates a will, is called a command (of reason), and the formula of that command is called an imperative” (4: 413). According to this definition, an imperative is the linguistic expression (“formula”) of a “command of reason,” which itself is the “representation” of an objective principle or law that “necessitates” the will. (Kant does not always hold himself to this usage, though, and often uses the terms “imperative,” “command,” “law,” and “objective principle” synonymously). “All imperatives,” as Kant continues, “are expressed by an ought”–or, more precisely, can be expressed by an “ought,” since “Act in such a way that the maxim of your action could be a universal law” is an imperative in Kant’s sense, even though it is not “expressed by an ought.” It, however, could easily be reformulated accordingly (“You ought to act in such a way, so that the maxim of your action …”). The “necessitation” of the will that is expressed in the imperative is that specific rational form of normative motivation which consists in the awareness that one ought to do what is rational to do. Kant does not try to ground the connection between reason and normativity further: he presupposes here that a being that is endowed with reason but not always rational ought to be rational (cf. Kant’s discussion of the usefulness of reason at the beginning of GMS I; 4: 395/6). He does, however, explicate the normative and evaluative dimension of reason a bit later by highlighting the connection between “ought” and “good”: imperatives say “that to do or to refrain from something would be good […]. Practical good, however [in contrast to the “pleasurable”; M. W.], is that which determines the will by means of representations of reason, hence not by subjective but objective causes, that is, from grounds that are valid for every rational being as such” (4: 413; cf. 414). According to Kant, every imperative implies a value judgment, but one that is rationally justifiable and thus claims to hold for all rational beings. (Here we are reminded of the famous thesis of

Practical Reason

135

the KpV concerning the priority of the moral law over the concept of the good; cf. 5: 57–71.) Imperatives, however, do not hold for all rational beings, because, as we have already seen, a holy will is not subject to any normative “necessitation”: “the ‘ought’ is out of place here,” for Kant, “because volition is already of itself necessarily in accord with the law” (4: 414). Kant is famous for the thesis that ought implies can: someone who ought to do A, also can do A (cf. 5: 30). In contrast, his claim here is: ought implies being able to do otherwise. Someone who ought to do A, can also not do A. (It does, for example, seem as senseless to command someone to be identical with himself, as it is to command him to not be identical with himself; he cannot stop doing the first, and he cannot do the second. In both cases the “ought” is “out of place.”) Kant’s view can be summarized by saying that imperatives are rational directives (or directives of reason). Their normative binding force is based on the fact that it is rational to follow them. Their prescriptive character is based on the fact that they are directed toward beings that can act rationally, but do not necessarily always do so. Kant continues: “All imperatives command either hypothetically or categorically. The former represent the practical necessity of a possible action as a means to achieving something else that one wants (or that it is at least possible for one to want). The categorical imperative would be one that represented an action as objectively necessary in itself, without reference to another end” (4:414). On the basis of the analytic relationship between “ought” and “good,” Kant can express the distinction between hypothetical and categorical imperatives in the following way as well: “Now, if the action would be good merely as a means to something else, the imperative is hypothetical; if the action is represented as in itself good […], then it is categorical” (4:414). The distinction between hypothetical and categorical imperatives thus has an impact on the question of why it is good and rational to act in the way commanded by the imperative. This distinction does not necessarily show in the linguistic form of the imperative: “Invest your money cautiously!” is, viewed linguistically, a categorical sentence but a hypothetical imperative, for a cautious investment is obviously good “merely as a means” and is not sought after for its own sake. “If you owe money, you ought to pay it back” is linguistically a conditional sentence but nevertheless a categorical imperative, because it represents an action “as objectively necessary in itself, without reference to

136

Marcus Willaschek

another end” (on the problem of the distinction between hypothetical and categorical imperatives cf. Brinkmann 2003, 19–51). As mentioned above, Kant divides hypothetical imperatives further into “rules of skill” and “counsels of prudence” and thus arrives at a tripartite division of the imperatives: (1) Rules of skill prescribe an action under the condition of an end concerning which it is left open whether one actually follows the end and whether it is rational to do so: “If you want A, then you ought to do B (as a necessary means thereto)!” Nothing changes in this hypothetical structure when the condition is met and one thus actually pursues the end in question: Kant does not provide for the derivation, using modus ponens, of a categorical imperative, “You ought to do B,” from the hypothetical imperative “If you want A, you ought to do B” and the empirically-contingent observation “You want A.” The imperative thus derived, “You ought to do B,” remains a hypothetical one, as its validity depends upon the condition that the person in question wants A. (2) For this very reason counsels of prudence are also merely hypothetical imperatives. They recommend an action as a means to an end that the addressee of the imperative actually has: “Because you want A, you ought to do B (as a necessary means thereto)!” Because imperatives are expressions of principles of reason and thus have universal validity, the only end in question here is one that all rational beings actually pursue (insofar as, as Kant notes, “imperatives apply to them as dependent beings,” 4: 415). The only such end is that of one’s own “happiness”: every person wants to be happy. (Kant defines happiness as a state in which “everything goes according to one’s wish and will” (5: 124)). This pursuit of happiness is a natural fact – it is part of what constitutes humans as rational living beings; thus the “purpose of happiness” is something we have by a “natural necessity” (4: 415). It can thus be presupposed to exist in every human being. Nevertheless, this purpose cannot serve as a basis for a categorical imperative, for different people find happiness in very different things. No means, therefore, can be identified that would be actually necessary for every person in order to become happy: “Because you (like every rational natural being) want to be happy, you ought to live healthily,” or, more concisely, “Live healthily!” is a rational counsel, but with some luck (Glück, fortuna) one can also be happy (glücklich, beatus) if one does not heed it. (3) Only moral (categorical) imperatives hold without exception for all (“dependent,” i. e., capable of being sensibly motivated) rational

Practical Reason

137

beings. They do not concern the “material of the action” (thus do not depend on which ends one is pursuing), but concern rather “the form and the principle from which it [the action] itself follows” (4: 416). The “form” of the action is, as has already been shown in the first section of the GMS, the “conformity to law as such,” which serves the will “as its principle” (4: 402). In this manner, Kant arrives at the categorical imperative in the singular (“never to act except in such a way that I could also will that my maxim should become a universal law,” 4: 402; cf. 421). According to Kant, out of this universal categorical imperative, as their principle, “all imperatives of duty,” i. e., all particular categorical imperatives such as the prohibition on lying, stealing, etc., “are derived” (cf. 4: 402). What unites these distinct categorical imperatives is the fact that one ought to act in the prescribed way not for an “actual or possible purpose,” but simply because the action is good “in itself.” – To prove that there actually are such actions (and thus also the categorical imperative that commands them) is the aim of the third section of the GMS. Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–) (abbreviated as ‘AA’). The Critique of Pure Reason will be cited according to the A/B pagination from the first and second editions. All textual references are to The Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press, 1992–). GMS Grundlegung zur Metaphysik der Sitten, AA, IV KdU Kritik der Urteilskraft, AA, V KpV Kritik der praktischen Vernunft, AA, V KrV Kritik der reinen Vernunft, AA III, IV MAN Metaphysische Anfangsgründe der Naturwissenschaft, AA, IV MdSR Metaphysik der Sitten, Rechtslehre, AA, VI MdST Metaphysik der Sitten, Tugendlehre, AA,VI RGV Die Religion innerhalb der Grenzen der blossen Vernunft, AA, VI TG Träume eines Geistersehers, AA, II

Other works Allison, Henry E. (1990): Kant’s Theory of Freedom, Cambridge. Beck, Lewis White (1974): Kants “Kritik der praktischen Vernunft.” Ein Kommentar, München.

138

Marcus Willaschek

Brinkmann, Walter (2003): Praktische Notwendigkeit. Eine Formalisierung von Kants Kategorischem Imperativ, Paderborn. Gerhardt, Volker (1986): “Handlung als Verhältnis von Ursache und Wirkung. Zur Entwicklung des Handlungsbegriffs bei Kant”, in G. Prauss (Hg.), Handlungstheorie und Transzendentalphilosophie, Frankfurt/M. Laberge, Pierre (1989): “La définition de la volonté comme faculté d’agir selon la représentation des lois”, in O. Höffe (Hg.), Grundlegung zur Metaphysik der Sitten. Ein kooperativer Kommentar, Frankfurt/M., 83–96. Timmermann, Jens (2003): Sittengesetz und Freiheit. Untersuchungen zu Immanuel Kants Theorie des freien Willens, Berlin. Willaschek, Marcus (1992): Praktische Vernunft. Handlungstheorie und Moralbegründung bei Kant, Stuttgart/Weimar. Wood, Allen (2006): “The Good Without Limitation”, in this volume.

Bernd Ludwig

Kant’s Hypothetical Imperatives (GMS, 417–419) I. In German language (and also in many others) there is a grammatical trap, which usually snap shuts as soon as one starts talking about Kant’s imperatives thoughtlessly. It goes like this: ‘Who wants to play the piano must practice.’ This is a true statement informing us of connections in the world. It is almost exemplary descriptive and would be wrong if for example playing the piano was like sneezing, which one does not have to practice to be able to do it. Put into second person, this statement reads ‘If you want to play the piano you must practice.’ This is another descriptive statement, particularly since it was generated by insertion into the former – a procedure which does not affect the descriptive character.1 The latter can also be expressed differently: ‘If you want to play the piano well, practice!’ It is true that the tone might change slightly (maybe from an amicable hint to a parental admonition), but put aside the educational authority (and the possibly connected sanctions) nothing has changed but the grammatical form. The grammatical indicative ‘you must practice’ has been substituted by the grammatical imperative ‘practice!’, including an expression mark at the end of the sentence. Now the mentioned trap is wide open, for expression marks also appear at the end of sentences which not only contain the grammatical form of an imperative but actually are imperatives (‘Practice!’), i. e. in Kant’s terminology: statements expressing an ‘ought’ and thus ‘an objective necessitation [Nöthigung]’ towards a certain action (see KpV, 20). This kind of statement one can address to oneself as well 1

This kind of insertion is, for instance, used when trying to take into account the peculiarities of the addressee, like: ‘If you want to play the piano well, YOU have to practice (in opposition to your advanced brother)’.

140

Bernd Ludwig

as to others. But as statements these latter imperatives do not contain any conditional clauses. Rather, these grammatical imperatives stop being imperatives, i. e. stop expressing an ‘ought’, if equipped with conditional clauses. The statement: ‘If 4 is a prime number: Practice!’ does not express an imperative, just because it does not express an ought, or a necessitation (for who could be necessitated to do what?). The one and only imperative involved here is: ‘Practice!’. Neither does ‘If you gave a promise, keep it!’ express a necessitation (the imperative ‘Keep your promise!’ rather necessitates the one who gave the promise), and ‘If you want to play the piano, practice!’ necessitates least of all. To cut a long story short: grammatical forms of imperatives are no indicator for imperatives, unless they come alone. And, on top of all that, at least this is for sure: in Kant’s writings there are no formulations of imperatives like ‘If …, then do …!’ to be found. And having said all the above, this is probably just consistent.

II. Kant’s Terminology The conception of an objective principle, in so far as it is obligatory for a will, is called a command (of reason), and the formula of the command is called an imperative. (GMS, 413)

Unlike angels and saints, human beings are not reasonable beings but only capable of reason. Their willing is not necessarily determined by reason but can also be stimulated by inclinations. This is where the necessitation, or ought, comes from on part of the reason. ‘You should do X!’ or just ‘Do X!’ is thus the natural form of the imperative which speaks to the one of whom it is expected that her or his will is not determined from subjective causes, but objectively, that is on principles which are valid for every rational being as such. (GMS, 413)

Any given imperative can be examined not only with respect to the content, but also with respect to the conditions under which it is actually commanding: If certain conditions would not apply anymore, would the necessitation, the ought, vanish and hence the imperative as such turn invalid or void? These conditions may either be changed by the addressee of the ought, or not. In the second section of the Groundwork, Kant is only interested in the former kind of conditions. Consequently, he distinguishes imperatives as follows:

Kant’s Hypothetical Imperatives

141

Now all imperatives command either hypothetically or categorically. The former represent the practical necessity of a possible action as means to something else that is willed (or at least which one might possibly will). The categorical imperative would be that which represented an action as necessary of itself without reference to another end, i. e., as objectively necessary. (GMS, 414)

As one can see from the lecture notes of Collins and Mrongovius, Kant had already given lectures in this spirit some years earlier. Here one can find formulations like this: The imperatives of skill command only hypothetically […] The imperatives of prudence command not only given a problematic condition but given an assertoric one. (AA XXVII, 246, compare 1399 f.)

There are corresponding formulations in the Reflections, too (cf. R 6805, 7207). Thus, in his writings as well as in the lectures the often called ‘hypothetical’ imperatives are explicitly introduced as ‘hypothetically-commanding’ statements, i. e. ‘hypothetical(ly)’ does expressly not refer to any kind of logical form of the imperative as a statement, but only expresses the contingency of the commanding character of the ‘formula’. In this way, an important difference to the categoricallycommanding imperatives is given at hand: the necessitation on part of the hypothetically-commanding imperatives only refers to … whatever is only necessary for the attainment of some arbitrary purpose may be considered as in itself contingent, and we can at any time be free from the precept if we give up the purpose; on the contrary, the unconditional command leaves the will no liberty to choose the opposite. (GMS, 420)

Whoever recognises himself as the addressee of the two imperatives ‘Practice playing the piano!’ and ‘Keep your promise!’ recognises at the same time that he might get rid of the necessitation expressed in the first imperative by dropping the purpose, viz. to learn to play the piano. The imperative commands hypothetically. But the given promise cannot be taken back one-sided, the cancellation of the necessitation is not within the power of the addressee. This imperative commands categorically. Accordingly, (1) for those who do not want to play the piano there simply is no imperative ‘Practice (playing the piano)’! – neither a hypothetical, nor a categorical one. On the other hand, (2) there is an imperative ‘Practice (playing the piano)!’ for everyone who wants to play the piano. In so far as statement (2) is true, the included gram-

142

Bernd Ludwig

matical imperative actually is an imperative, and, since statement (1) is true as well, the imperative in statement (2) is a hypothetical one. In this way, according to Kant’s introduction of the difference between ‘hypothetical’ and ‘categorical’ imperatives, Kant’s talk of ‘hypothetical imperatives’ turns out to be a mere short-form of ‘hypothetically-practical’ or ‘hypothetically-commanding statements’ respectively. Obviously, the same goes for the ‘categorical imperative’, for this one “commands categorically and absolutely” (AA XXVII, 247). In the Second Critique, Kant gives the equally plain statement: A ‘categorical imperative’ is ‘categorical’ only in so far as the very fact that the statement in question entails a determination of will does not depend on the particular desires and faculties of the addressee: The latter must sufficiently determine the will as will even before I ask whether I have the ability required for a desires effect or what I am to do in order to produce it, and must thus be categorical: otherwise they are not laws (KpV, 20).

Hence, applied to practical rules, the terms ‘categorical’ and ‘hypothetical’ do generally not refer to a possible (judgment-)form of an imperative, but to the mode of the determination of will only. For this reason, grammatically the two terms are no adjectives which belong to ‘imperative’, but they determine the verb ‘command’ attributively.2 There are neither ‘hypothetical imperatives’ nor ‘categorical imperatives’ sensu stricto, 3 but only ‘hypothetically-’ or ‘categoricallycommanding’ statements respectively, which, in so far as they command, are imperatives (or are called so) and which all have the same form: ‘Do X!’. ‘Hypothetical’ and ‘categorical’ are not placeholders in an imaginary ‘Kantian table of the imperatives’4, but just name the relevant relation in a judgment about the practical character of a statement given as a practical one. Strictly speaking, ‘hypothetical’ and ‘categorical’ are terms that belong to the meta-theory of imperatives. They don’t classify imperatives, but they appear in judgements 2

3

4

Also KpV, 31: “represented a priori as a categorical proposition,” KpV, 32, and MdST, 404: “commands categorically,” similar GE, 282 and 416. In German, the terms ‘kategorisch’ and ‘hypothetisch’ stand for adjectives and adverbs both. That is to say, the semantic relations fundamentally differ here from the ones in the case of ‘hypothetical judgments’: There it is not the character of the judgment which is contingent, but solely the asserting of (one of) the singular judgments. Hence, hypothetical judgments are “always” judgments, but ‘hypothetical imperatives’ sensu stricto were only imperatives if the antecedent were true. In analogy to the Table of Judgements (in KrV, B95) M. Moritz f. i. misses ‘disjunctive’ imperatives (s. 9 ff.).

Kant’s Hypothetical Imperatives

143

which state that – and why – a particular statement is an imperative at all. Basically, Kant has to take this for granted, for in his opinion ‘hypothetical’ and ‘categorical’ are judgment-predicates e definitione, and imperatives just are no truth-apt (theoretical) expressions, i. e. judgments, but commanding (practical) statements. There can be no more ‘hypothetical imperatives’ sensu stricto within the scope of Kantian concepts than, e. g. ‘hypothetical questions’ or ‘hypothetical exclamations’. This can be seen very clearly if thinking about the possible form hypothetical imperatives in the strict sense might have: if they had the prima facie obvious (albeit degenerated, 5 but probably only possible) form of hypothetical judgments with a ‘categorical imperative’ as its consequent (‘If you want to play the piano, practice!’), there was the immediate threat of the unpleasant complication of true antecedents enforcing ‘categorical imperatives’ via separation-rule. These would – appropriately to the presupposition – compete with Kant’s moral imperatives, which would involve absurd consequences.6 To take the famous example from Richard Hare: ‘If you want to inherit your uncle’s property before he has gambled it away, you must/should murder him as soon as possible.’ Well, in that case: Yes indeed, I want to be my uncle’s heir soon, hence (with elemental modus ponendo ponens): For me, it is a categorical imperative to murder my uncle as soon as possible. Hence, ‘conditional with a categorical imperative as its consequent’ cannot be the correct analysis of the so called ‘hypothetical imperatives’.7 Further: if there is an imperative to be involved at all and since the consequent cannot be a categorical imperative (as the example shows), it seems that it has to be a hypothetical imperative (tertium 5

6 7

The hypothetical judgment is not made of one judgment and one imperative but of two judgments (KrV, B98). See Patzig (1973, 209 f.). The inventor of this example, Hare (1971), shows the trouble one gets into, if one sticks to this misguided (and not Kantian) analysis. – A simple reduction ad absurdum shows that any expression of the form “E→O(M)” won’t work as a formal version of “Who wants the end, ought to want the necessary means”, with “E” standing for “Person p wants the end ”, “M” for “Person p wants the means” and “O( )” for something like “It has to be the case that …”. We only have to apply contraposition twice: let “~m→~e” stand for the proposition [underlying E→O(M)] that the means is necessary for the end in question. Since “~m→~e” is equivalent to “~~e→ ~~m”, “~M→O(~E)” as well has to be the expression of an “imperative” (by the very same reason as for which “E→O(M)” is one). But “~M→O(~E)” is equivalent to “~O(~E)→M” again and this latter is unquestionably nonsense: nobody wants the means in all cases where there is no obligation/necessity to abandon the end. Since the only non-trivial step in the “argument” above is the transition from “~m→~e” to “E→O(M)”, this must be inadequate.

144

Bernd Ludwig

non datur). However, then there only came off an infinite regress of hypothetical conditions but never a necessitation, not to mention that it remains entirely clouded what could be the content of the antecedent of each single one of the hypothetical imperatives in question. Therefore, according to the opening considerations and to Kant’s proper definitions, a statement like: ‘If you want to be happy, avoid pain!’ is not an imperative at all,8 but simply a truth-apt judgment whichcontains an imperative. At the most, it is a kind of advice which one usually uses to argue for the relevance of the imperative-formula ‘Avoid pain!’ And further, it is not even a hypothetical judgement (that is, a conditional), since it does not answer the question: ‘Under which conditions shall I avoid pain?’, but rather: ‘For what reasons shall I avoid pain?’. Only if one kept on asking after having received the answer (‘Because you want to be happy.’) and thus discovered being free of pain to be a necessary condition of human happiness, eventually a hypothetical judgment came into play: ‘If you don’t avoid pain, then you cannot be happy.’ Although at first sight Kant himself speaks a bit loosely in this context and seems to equate ‘hypothetical imperatives’ with “counsels” (GMS, 416,19; but then 24 f.), one should be aware of the fact that he gave a clear and consistent exposition of the basic concepts afore, which make it possible for the reader to specify such stenographical formulations and thus explicate them in a coherent and adequate manner. We will see that Kant isn’t very particular about the terminology in this respect when looking at the alleged analyticity of hypothetically-commanding imperatives. But first, the basic idea of the proof of the ‘possibility’ of hypothetical imperatives must be recapitulated. III. The Possibility of hypothetical Imperatives Let’s start again with some elementary considerations independent of Kant’s writings: 1. A man9 who wants to realise an end (E) which cannot be realised without any help on his part, is necessitated to do10 something definite, viz. to apply the necessary means (M). This is completely 8 9

10

G. Patzig pointed that out in 1971 already. Although the following does apply to women as well, I don’t consider this in the main text, just to avoid grammatical redundancy or discrimination by using the male pronouns pars pro toto. But not to want anything definite: if I want to step through the door, I must (rebus sic stantibus) push the door handle. But I can push the door handle without wanting to push it (e. g., if I push it accidentally).

Kant’s Hypothetical Imperatives

145

independent of his being rational in a narrower sense. If he doesn’t do M, he won’t realise E (period!). And if he doesn’t want to apply the necessary means, he e suppositione won’t (be able to) do anything leading him to his end. For this reason, he (inevitably) does M if he does E, in this sense he shall/must do M, if he wants E. 2. If he is, moreover, rational in a wider sense, i. e. knows about the necessity of M and about its viability, then he actually wants M if he wants E, i. e. he does not want E, if he does not want M, because he knows that he cannot realise E if he does not realise M. He just doesn’t know anything he could do to realise the end. He cannot give any answer to the question of how one could recognise that he wants E, for everything he does seems to show that he doesn’t want E. One would turn ones back on him, saying: ‘You don’t really want to realise E. At best you wish that E was the case.’11 Kant’s analysis of the structures which form the basis of the hypothetically-commanding imperatives is expressed in a two-stage argument: Whoever wills the end, wills also (so far as reason decides his conduct) the means in his power which are indispensably necessary thereto. This proposition is, as regards the volition, analytical; for, in willing an object as my effect, there is already thought the causality of myself as an acting cause, that is to say, the use of the means; and the imperative educes from the conception of volition of an end the conception of actions necessary to this end. (GMS, 417)

This passage claims three things:12 1. Who wants the end wants (in so far as …) the means. 2. Statement ‘(1)’ is analytic. 3. The imperative derives the concept of necessary action from the wanting the end. According to Kant, these three statements contain the answer to his question of “how we can conceive the necessitation of the will which the imperative expresses” (GMS, 417). Furthermore, Kant stresses that this question does in fact not need any exceptional discussion, that the answer goes without saying to a certain degree. Obviously, the fi rst statement is true if the second is. And the third should then result with the aid of the first. First, ad 2.: Analyticity of the ‘means-end-formula’ 11

12

Who does not want to practice the piano and knows that it cannot be done without, simply cannot do anything he himself could understand as expression of his will to become a pianist, i. e. it is not clear what is meant by ‘he wants to become a pianist’ in this case. For reasons of clarity, the three assertions are represented in an abridged way.

146

Bernd Ludwig

If the ‘means-end-formula’ is analytic, its truth can be understood by the analysis of the meaning of the words contained in the judgment, i. e. of the Kantian concepts ‘means’, ‘end’ and ‘willing’/’will’. Moreover, this insight should result without any additional idiosyncratic assumptions – according to Kant’s assurance that there is no need of any exceptional discussion in this context. In the Groundwork one can find the following about the first of these concepts: On the other hand, that which merely contains the ground of possibility of the action of which the effect is the end, this is called the means. (GMS, 427)

Using the explicit definitions of the two other concepts provided in the Metaphysics of Morals from 1797, one reaches the goal directly. An end is an object of choice (of a rational being) through the representation of which choice is determined to an action and to bring this object about. (MdST, 381)

However, this text dates from 1797. Therefore one has to be aware of the fact that it already takes into consideration the distinction between ‘Wille’ (sc. will) and ‘Willkür’ (sc. choice), which Kant doesn’t draw yet in the Groundwork and which he only uses firmly after the Religion from 1793. It is also mentioned in the introduction to the Metaphysics of Morals: The faculty of desire in accordance with concepts, insofar as the ground determining it to action lies within itself and not in its object, is called a faculty to do or to refrain from doing as one pleases. Insofar as it is joined with one’s consciousness of the ability to bring about its object by one’s action it is called choice; if it is not joined with this consciousness its act is called a wish. The faculty of desire whose inner determining ground, hence even what pleases it, lies within the subject’s reason is called the will. (MdSR, 213)

This ‘Willkür’ (choice) corresponds to ‘Wille’ in the ‘means-end-formula’ of the Groundwork. To answer the question of the analyticity of this formula one has to be aware that Kant’s concept of ‘Willkür’ (1785: of ‘Willen, GMS, 449) includes “consciousness of the ability to bring about its object” (MdSR, 213). But under this presupposition the willing of the end almost naturally comprises the willing of every “ground of possibility of the action” (GMS, 427) identified as necessary.13 For if 13

See also the quotation from GMS, 417, above: “… but if I know that it is only by this process …” Surely, one feels already necessitated if one only believes. As expected,

Kant’s Hypothetical Imperatives

147

I desired an end without wanting the means identified as necessary, I either acted ‘conscious of the inability to bring about its object’ of the end (and, hence, by definition wouldn’t really want or will), or – implicitly – contradicted myself by asserting that I could realise the end without the relevant necessary means (sc. ‘the ground of possibility of the action’). The contradiction can be seen in this: ‘I want to do something that realises E, but I only want to do such which does not realise E.’ Hence, not-willing of M is connected to knowing of missing causality for E, since the imperative-addressee knows that he does not do anything which realises E, as long as he refrains from doing anything which realises E. Thus, he wants to do M if he wants E. If he doesn’t want M, he doesn’t want E, but only wishes E. Accordingly, to Kant one can only want something, so far as one believes that it is in one’s own power. Two cases have to be considered here: Either the means is taken to be unattainable, rightly or wrongly (sc. ‘I can’t find my gun’), or they are explicitly not wanted (sc. ‘I will never kill an innocent person!’). In both cases, the wishing does not push ahead to willing. In the second case the reason is on the part of the wishing person itself (‘I wish he was dead, but I don’t want to kill him’), in the first it is not (‘I wish I had my gun at hand and killed him’).14 These conceptual relations are clearly expressed in an answer to objections against the relevant representation in the Second Critique. In the introduction to the Third Critique from 1790 it reads: An objection has been made to me […] and the definition of the faculty of desire as the faculty for being through one’s representations the cause of the reality of the objects of these representations has been criticized because mere wishes are also desires, but yet everyone would concede that he could not produce their object by their means alone. – This, however, proves nothing more than there are also desires in a human being as a result of which he stands in contradiction with himself, in that he works toward the production of the object by means of his representation alone, from which he can however expect no success, because he is aware that his mechanical powers (if I may so name those that are not psychological), which have to be determined through that representation in order to realize the object (hence mediately), are either inadequate or even aimed

14

Kant does nowhere in his argument assume that the relevant causal judgments are true. However, it is hard to judge singular cases: ‘I cannot live that way’ usually just means: ‘I do not want to live that way, for the price is too high’ – still, the bystanders sometimes consent that one cannot live that way. But then it is all about what we want to expect of each other – and not about where the natural necessities begin.

148

Bernd Ludwig

at something impossible, e. g., to make what has happened not have happened (O mihi praeteritos, etc.), or, when impatiently waiting, to make the time until the wished-for disappear. (KdU, 177; my emphasis)

Second, ad 3): Necessitation and Necessity Anyone, ‘so far as [his] reason decides his conduct,’ wants the means together with the end. In this context (i. e. the analytic meansend-formula), Kant does not mention ‘necessitation’: … but if I know that it is only by this process that the intended operation can be performed, then to say that, if I fully will the operation, I also will the action required for it, is an analytical proposition; for it is one and the same thing to conceive something as an effect which I can produce in a certain way, and to conceive myself as acting in this way. (GMS, 417)

The necessitating character of the hypothetically-commanding imperatives, which goes beyond the insight that who wants the end also wants the means, as long as the reason decisively determines his actions, shows itself in two cases: 1. If the one, whose will is not (yet) determined decisively by the reason in the relevant respect, is necessitated to find the means (without he cannot realise the end), i. e. get his reason the relevant influence. Or 2.: If the one, who already knows about the necessary means, feels necessitated to decide whether to do without the realisation of the ends or to use the means. What he surely cannot do (and hence, which he cannot be ordered to omit) is both to want the end and at the same time not want the means already known to be necessary. Hence in this second case he is not necessitated to do something, but only to make a decision: He is necessitated to “give up his end” (GMS, 420), or to use the means in question. Though, who wants end and means likewise is not necessitated at all, just like Kant’s ‘saints’: For those only laws apply which are no imperatives (GMS, 414). In this sense, a hypothetically-commanding imperative indeed necessitates, regardless of the fact that everyone who wants the end wants the means as well as long as the reason decisively determines his willing. For, who wants the end and knows what the necessary means is (sc. as long as …) is not necessitated to want the means, he just wants it. Who does not want the means is not necessitated not to want the end, he (as long as …) just does not want it. The necessitation only shows itself when incompatible wishes face each other. In that case only one of them can become a will, and if that happens, the other wish looses its force – and reason has gained decisive influence. Consequently, the ‘ought’, which is expressed by the hypothetical imperative in view of the means in question, is always a hypothetical

Kant’s Hypothetical Imperatives

149

willing at the same time. However, it is misleading to talk about ‘rational willing’ as a paraphrase of Kant’s own formula: ‘wills also (so far as reason decides his conduct)’. ‘Willing’ is exclusive to rational beings anyway, as Kant claims: “the will is nothing but practical reason” (GMS, 412,26 ff.), and, hence, a non-rational willing is an absurdity within the scope of Kant’s considerations. What Kant must have in mind here is that without influence of reason one does not develop any relation to a certain object or a certain behaviour whatsoever. If I want the radio to play music but don’t know that for making that possible I have to get up from my chair, I have no more will regarding my sitting position than regarding the position of a stone on the backside of the moon. Consequently, the parenthesis ‘so far as reason decides his conduct’ does not mean that we – metaphorically speaking – listen to reason sometimes more and sometimes less, but to what degree it speaks to us in particular cases. For after it has spoken to us, we don’t have the choice of being disobedient anymore: who knows that he cannot make the radio play without getting up, but still refuses to get up, from now on simply lacks “consciousness of the ability to bring about its object” (MdSR, 213), in short: the will for music, for he knows of all remaining actions that they won’t make the radio play. He isn’t ordered or forbidden anything but he just realises what he cannot do. If in the context given Kant inserts the clause ‘as far as reason decides [maßgeblich beeinflußt]’ he thus does not refer to the fact that reason can prevail15 against the passions (as one has to imagine it in the case of the categorical imperative) but he refers to our capacity to “deduct” (GMS, 412,28 ff.) an action from a given rule. It might happen even to Kant’s ‘saints’ that they are influenced by reason insufficiently and thus don’t develop a will concerning getting up from the chair, even if they want to listen to the radio. But in this case (as always with saints) it is not moral imperfection but mere epistemic incompleteness (or stupidity, admittedly), which is ‘maßgeblich’ at this point. Maybe the unknowing saint does not want to get up. But he must, if he wants to listen to the announcements of the Holy Father on the radio. Being a saint, he e suppositione does not struggle with irrational inclinations, but he is necessitated nevertheless to find out the adequate rule that tells him what is required to realise his end. Maybe Kant has not shown clearly enough the difference between what against the necessitation works in cases of hypothetical and cat15

This is approximately what Schönecker/Wood (22004, 119) propose as an interpretation.

150

Bernd Ludwig

egorical imperatives respectively.16 The necessitation in hypothetically-commanding imperatives is – on his conditions – no necessitation reason exerts as an antagonist of the inclinations, but always a necessitation which one inclination by the help of reason exerts against another. The addressee of the imperative does not have to do anything definite but rather to decide which of the (say, morally neutral) inclinations (or wishes) he wants to give in: to stay in the chair or to listen to the radio. The hypothetical imperative makes clear to him that he can only make his wish to stay in the chair a will at the price of letting go his wish to listen to the radio. Accordingly, Kant does not say: ‘Who wants the end (and knows about the necessary means) ought to17 want the means’ but plainly “… wants the means” (GMS, 417, lines 10 and 30). The determination of the will by a hypothetical imperative does not consist in an order to want18 only consistently. This would be futile because it would order us to refrain from doing something we cannot do anyway. Rather, the ‘determination’ consists in the imperative’s necessitating us to make our wishes consistent when it comes to actions, i. e. to the willing, by showing us what we wish, but cannot do – and in the face of that knowledge cannot want, either. In short: ‘You want, because you know that you cannot do differently!’. The necessitation is the repercussion of wanting the means upon wanting the end. It is different in the case of Kant’s categorical imperatives: you are capable, because you know that you are obliged (see KpV, 30); but the inversion does not hold that you are incapable, just because you know that your action is prohibited – human beings are no angels, whose reason is practical without any hindrances, thus determines the will, but they are always stimulated by their inclinations (GMS, 449; see also KpV, 82, and MdST, 383). In this respect the moral prescriptions which one can violate deliberately are principally different from hypothetical imperatives. Exactly because of this, according to Kant, the latter are just “Corollarien aus der Naturwissenschaft” 16

17

18

This lack of clarity fi nds an echo in Kant’s ambiguous use of the word ‘Gebot’. In the beginning he defi nes all imperatives as formulas of ‘Gebote’ (GMS, 413), but later (GMS, 416) he suggests that only categorical imperatives are such. Or ‘muss’ (must), as in Patzig (1973, 215). – At best ‘Who wishes the end, shall will the means’ would render an understandable assertion. In Kant we mostly fi nd: I ought to do something, because I want something (GMS, 441,11; 444, 4 and 11; the only exception: GMS, 419,9). Though Downie (1984, 486) consents to that, he tries to show that there is a „moral obligation of practical consistency” alike to a liability behind hypothetical imperatives.

Kant’s Hypothetical Imperatives

151

(KdU, 173; MdSR, 222; see also AA XX, 200), which might contain a determination of will. As one can easily see now, they just necessitate us, when we know about the means to an end, to decide for the means or against the end.19 Simply and solely categorical imperatives necessitate us to decide for or against something definite, hence, to do something definite. IV. Hypothetical vs. Absolute Necessitation – Achenwall Still, the perhaps surprising discovery that hypothetically–necessitating imperatives do not make particular actions necessary but rather decisions is not just a modern interpretation of Kant, but can already be found in Gottfried Achenwall’s (whom Kant thought so highly of) terminological fixing of an obviously ready-established usage: Si determinatur alter ad unicum; coactio vocatur absoluta: si ad unum ex quibusdam determinatis eligendum; dictur hypothetica. (Elementa Iuris Naturae, 1750, § 70)20

Achenwall’s remark sheds new light on the distinction of imperatives, for it questions a seemingly obvious assumption of most interpretations: even if it might seem that Kant developed the terminology of the “hypothetical imperatives” – and hence the distinction between hypothetical and categorical imperatives as such – in analogy to his doctrine of judgements, there is no positive proof for this supposition. On the contrary, there is much to be said for the thesis that Kant just adopted the distinction directly from the given moral context and only transferred it to his own terminology later. Some textual evidences illustrate that quickly: an early Reflection (R 6498; cf. R 7202) distinguishes the vis necessitans absoluta and the vis necessitans hypothetica. And one should not think only about coactio absoluta/hypothetica 21 (see the quote above), but also about the distinction between 19

20

21

Accordingly something like “O(~M→ ~E)”, that is “O(E→M)” might be closer to the point than “E→O(M)” [compare note Fn. 6 above]. But unfortunately this does not give any opportunity for fruitful formal fi nger exercises. “If someone is necessitated to do a specific act this is an absolute coercion [coactio absoluta]. If he is necessitated to choose one amongst others [eligere] this is a hypothetical coercion.” In Achenwall we find that “absolute” and “categorical” are usually closely related in the context in question: “… si voto voluntas declaratur […] hypothetica, hypotheticum, si absoluta, categoricum [dicitur.]” AA XIX, 341 (= Ius naturalis, pars posterior, § 26). For further details see Schwaiger (1999, especially 168 ff.).

152

Bernd Ludwig

absolutem und hypothetischem Naturrecht in Achenwall’s writings (cf. AA XIX, 326). Perhaps it is just coincidence, but there are at least two passages in the Groundwork, in which Kant expressly calls the categorical imperative “absolutely commanding”: … the categorical imperative, on the contrary, is not limited by any condition, and as being absolutely, although practically, necessary, may be quite properly called a command. (GMS, 416, cf. 420: “unconditional command” and, e. g., AA XXVII, 247 and 1400)

On the other hand, at least in Kant’s publications there is no talk of “unconditional” judgments as opposed to hypothetical ones. The factual differentiation between absolute/categorical necessitation on the one hand and hypothetical necessitation on the other is thus at least genetically independent from the special terminology of the Table of Judgements from 1781. And Kant’s assimilation of imperatives and judgements in the Groundwork could par conséquent be nothing but a quite powerful artifice of suggestive presentation without having a real fundamentum in re. Also, the “hypothetical” necessity was sometimes called “conditional” by Kant (e. g., R 6639). Correspondingly, one can find the distinction of “necessitatio […] categorica vel conditionalis” (R 6463). Thence, the Kantian innovation was less of the discovery of the distinction itself, than of its specific usage: maybe this is the reason why Kant can do without a systematic introduction of this distinction in the Groundwork, and – relying on its familiarity – simply introduces it with the formula: “Now, all imperatives command either hypothetically or categorically.” (See above), then interpreting this quite traditional distinction in his own and new sense: as the distinction concerning the dependency of the necessitation on a subjectively set end. And this sense, again, is not a specifically “critical” one, but was set out already in Kant’s pre-critical Untersuchung über die Deutlichkeit der Grundsätze der natürlichen Theologie und der Moral from 1764 (see UDG, 298 ff.).

V. Are hypothetical imperatives analytic? We saw above that the term “hypothetical” in the expression “hypothetical imperative” is no grammatically independent adjective, but an elliptic adverb to “commanding”. Given imperatives are hypothetically- or categorically-commanding respectively. A similar distinction should be drawn as well for the adjectives “analytic” and “synthetic.”

Kant’s Hypothetical Imperatives

153

Every now and then, the Kantian assertion that hypothetical imperatives were “analytic practical propositions” (GMS, 419) is described as: “hypothetical imperatives are analytic statements” – with the last expression standing for Kant’s “analytical judgements,” e. g. the heading title of a chapter in H. J. Paton’s classic The Categorical Imperative is ‘Imperatives of skill are analytic propositions’. Evidently, it is only a very small step from here to a question like: “Are hypothetical imperatives true in virtue of their logical form, are they even tautologies?”22 This kind of question contains a false presupposition, which I want to focus upon in the following. Kant himself does on no account say that hypothetical imperatives are analytic judgements or statements. According to him, hypothetical imperatives are analytic, at the best, in the figurative sense that one will fall back on an analytic principle if one explains their being possible: Having shown the “possibility” of imperatives of skill with the help of the means-end-formula (GMS, 417,11 and 417,23), he writes that the imperatives of prudence were “just as analytic” (as those of skill) (GMS, 417,29). At this point, this can just mean that their possibility can also be shown with the analytic formula in question only,23 for more has not been mentioned (and is neither mentioned elsewhere in the Groundwork). In a later passage, again, it is said of an imperative of prudence that it is an “analytic practical proposition” (GMS, 419,4) and that it differs from an imperative of skill only in the specific end (sc. happiness). This is an allusion to the mentioned figurative way of talking, which becomes even clearer through the attributive position of “analytic”: the reason that the statement is a practical one is a fact that can be stated by an analytical judgement about the necessitation to the application of a means, in case the will to the end is given. Hence the assumption that Kant believed hypothetical imperatives to be analytic statements in any sense or even analytic judgments (no matter how dubious an allegation like that might be anyway24) does not find any support by reference to the Groundwork (or in fact to any other of his writings), since Kant nowhere says that hypothetical imperatives themselves are “analytic”. In the three passages suggest22 23 24

Cf., e. g., Seel (1989, 160 f.). Otherwise we had to assume that some piece of text had dropped out. It has to be kept in mind that in Kant’s writings the analytic/synthetic-distinction is defi ned for categorical judgements exclusively, that is, it makes only sense for judgements of the form “S is P”. A judgement of this kind is analytical if and only if “the predicate(-term) is contained in the subject(-term)” (see e. g. KrV, B11). No matter how this might be spelled out more precisely: there is no natural way to bring any imperative (be it categorical or hypothetical) to the form: “S is P” – and we don’t have any evidence that Kant himself ever thought he could accomplish this.

154

Bernd Ludwig

ing that Kant might want to assert something like that, he is only concerned with the very familiar claim of the analyticity of the principle forming the basis of the explanation, why the use of the means is in fact imperative for her or him who wills the end. Kant never claims that the imperative itself is analytic. These negative findings get confirmed further by Kant’s talk of the “synthetic-practical” character of the categorical imperative. Even if Kant uses the formula ‘synthetical practical proposition’ sc. “synthetischer praktischer Satz” (GMS, 444, instead of ‘synthetic-practical proposition’ sc. “synthetisch-praktischer Satz”) later and, in this case, does not use ‘synthetic’ grammatically as an attribute to ‘practical’ but as a second predicate to ‘proposition’ (and of same rank as ‘practical’), this does not indicate that Kant wants to express more than what he has already written in the footnote to the formulation ‘a priori synthetic practical proposition’. Here, he directly and negatively refers to the analytic “means-end-formula” by emphasising that in case of the categorical imperative the willing of an action [is not deduced] by mere analysis from another already presupposed (for we have not such a perfect will), but connects it immediately with the conception of the will of a rational being, as something not contained in it. (GMS, 420)

Hence this is – as in the case of hypothetical and categorical – not about transferring the distinction between analytical and synthetical from categorical judgments to imperatives. Rather, it is about the synthetical character of the very categorical judgment which ties the concept of a definite “willing” to the concept of the will of a rational being. To cut a long story short: “synthetical” in “synthetical-practical proposition” refers to the syntheticity of the very judgment which expresses or explains the practical character of the imperative in question. This insight involves, firstly, that logic and grammar are in complete correspondence (like in ‘analytic-practical proposition’), and secondly, that at least Kant himself can be released from the reproach of having overlooked that imperatives (hypothetical as well as categorical ones) are not (categorical) judgments. Since nearby the formulation “synthetic practical proposition” (GMS, 444) there is no evidence that Kant wanted to say anything different here than in GMS, 420, one has to take this formulation as elliptic. The same goes for the third and last relevant passage: We cannot prove that this practical rule is an imperative, i. e., that the will of every rational being is necessarily bound to it as a condition, by a mere analysis of the conceptions which occur in it, since it is a synthetical prop-

Kant’s Hypothetical Imperatives

155

osition; we must advance beyond the cognition of the objects to a critical examination of the subject, that is, of the pure practical reason, for this synthetic proposition which commands apodeictically must be capable of being cognized wholly a priori. This matter, however, does not belong to the present section. (GMS, 440)

Here, Kant intermingles two statements: The imperative itself (sc. the ‘synthetic proposition which commands apodeictically’ sc. “der synthetische Satz der apodiktisch gebietet,” viz. “Don’t steal!”) and the judgment which says that the rule or law in question (i. e. this very imperative) in fact is an imperative. If it mattered that the imperative itself were a synthetical statement (like the latter part of Kant’s formulation might suggest at first sight), this would nevertheless be of no importance to the point in question, for it would only show that it is not possible to find out about the statement’s truth (but not about its imperative mode) by analysing its concepts. As long as “hypothetical” is supposed to be the “hypothetical” from the First Critique and Prolegomena, the statement to be analysed can only be a judgment, and in this case, it is the judgement concerning the practical character of the given imperative (just as the second half of the quote tells us, compare also GMS, 420). Quite as in the other cases the benevolent interpreter might either assume that Kant falls back on a (deeply puzzling and, unfortunately, nowhere recorded) thesis of the transmission of the well-known analytic-synthetic-differentiation from judgements to imperatives. Or he might assume that Kant was not always as careful as necessary to be taken at his word when distinguishing between the semantic levels. If the second alternative is approved, we will maintain that it is not the imperative itself which is analytic or synthetic, but the judgement (which in this case is special, affirmative categorical and necessary – according to the table of judgements) stating the practical character of statements addressed to someone who has specific ends.25 Some imperatives command hypothetically (for example: “Practise the piano!” – if she or he has the end of becoming a virtuoso) and some command categorically (“Keep your promise!”). And some seeming imperatives are in fact no imperatives at all: “Keep control of the calories now!” is just no imperative at all for her or him who does not want to reduce weight. In the first and in the last case we just have to explore what it means to have or have not a particular end. In the second case, this is 25

Thus, those proposals of correction which claim against Kant that hypothetical imperatives were not analytical become obsolete, cf. e. g. Schönecker (1999, 96).

156

Bernd Ludwig

just not enough, and the inquiry has to be postponed until the third Section of the Groundwork. Since this interpretation is compatible with all passages in the Groundwork26 in which Kant makes use of the analytic/synthetic-distinction, it does not need any demanding theses about tacit reinterpretations of the concept of analyticity, a concept which was already well-known from his other writings. VI. The End Who does not only wish but also acts, i. e. who wills anything in the Kantian sense, is necessitated to obtain adequate insights, realise them and, as a consequence of these insights, to decide between the options which are identified as being incompatible. Kant’s theory of hypothetically necessitating imperatives which are analytically-practical statements (but neither hypothetical judgments nor analytic judgments) is the subtle explication of this central but not very enigmatic fact. It is no hidden theory of practical rationality and it does not presuppose any idiosyncratic doctrines from transcendental philosophy. Hence, if Kant himself says that the possibility of “hypothetical imperatives” ‘does not need any exceptional discussion’, the aequitas hermeneutica requires to refute any interpretation as inadequate which urges us to suppose that there is anything to make out but an ingenious conceptual frame for a picture rather familiar to his contemporaries. Whether or not this conceptual frame is adequate has not been answered this way27 – but it seems that smart alternatives are quite rare up to present. Literature Kant’s writings Kant’s writings are cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–) (abbreviated as ‘AA’). The Critique of Pure Reason will be cited according to the A/B pagination from the first and second editions. All textual references are to The

26

27

Kant does not take up the distinction between “analytical” and “synthetical” imperatives later again. See further Ludwig 1999.

Kant’s Hypothetical Imperatives

157

Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press, 1992–). Ge Über den Gemeinspruch: Das …, AA, VIII GMS Grundlegung zur Metaphysik der Sitten, AA, IV KdU Kritik der Urteilskraft, AA, V KpV Kritik der praktischen Vernunft, AA, V KrV Kritik der reinen Vernunft, AA III, IV MdSR Metaphysik der Sitten, Rechtslehre, AA, VI MdST Metaphysik der Sitten, Tugendlehre, AA,VI R Reflexionen, AA, XV–IXX UDG Untersuchung über die Deutlichkeit der Grundsätze …, AA, II

Other works Achenwall, Gottfried/Pötter, Johann Stephan (1750): Anfangsgründe des Naturrechts (Elementa Iuris Naturae). Frankfurt a. M. 1995. Downie, R. S. (1984): “The Hypothetical Imperative,” in: Mind vol. 43, 481– 490. Hare, R. Melvin (1971): “Wanting: Some Pitfalls,” in: R. Blinkley et al. (eds.), Agent, Action, and Reason, Oxford, 81–97. Moritz, Michael (1960): Kants Einteilung der Imperative, Lund. Paton, Herbert James (1953): The Categorical Imperative: A Study in Kant’s Moral Philosophy, New York et.al.: Hutchinson. Patzig, Günther (1973): “Die logischen Formen praktischer Sätze in Kants Ethik,” in: G. Prauss (Hrsg.): Kant. Zur Deutung seiner Theorie von Erkennen und Handeln, Köln, 207–222. Schönecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs, Freiburg i. Br./München. Schönecker, Dieter/Wood, Allen W. (22004): Immanuel Kant “Grundlegung zur Metaphysik der Sitten”: ein einführender Kommentar, Paderborn. Schwaiger, Clemens (1999): Kategorische und andere Imperative: zur Entwicklung von Kants praktischer Philosophie bis 1785, Stuttgart-Bad Cannstadt. Ludwig, Bernd (1999): “Warum es keine ‘hypothetischen Imperative’ gibt, und warum Kants hypothetisch-gebietende Imperative keine analytischen Sätze sind,” in: H. Klemme, B. Ludwig, et al. (eds.), Aufklärung und Interpretation, Würzburg, 105–124.

Mark Timmons

The Categorical Imperative and Universalizability (GMS, 421–424) Kant’s moral theory and, in particular, his theory of right conduct, has often been judged entirely on the basis of the universal law formulation of the categorical imperative (FUL, hereafter): “Act only in accordance with that maxim through which you can at the same time will that it become a universal law” (GMS, 421). For purposes of using the idea of conformity to universal law as a principle for determining the morality of “an action in concreto” (KpV, 67), Kant re-casts FUL as: “Act as if the maxim of your action were to become by your will a universal law of nature” (GMS, 424).1 After illustrating the application of the formula of universal law of nature (FLN) in section II of the Groundwork, Kant goes on to claim that the basic requirement it expresses represents “the canon of moral appraisal of action in general” (GMS, 424).2 This canon, moreover, involves two universality tests that are supposed to be the basis for distinguishing perfect from imperfect duty. Getting such fertile results from such a seemingly meager source is at least surprising, something Kant himself notes when he writes, “The simplicity of this law in comparison with the great and various consequences that can be drawn from it must seem astonishing at first […]” (MdSR, 225). What Kant fi nds astonishing, others have found unbelievable. For instance, Hegel’s (1821, § 135) charge of “empty formalism” challenges the idea that Kant’s universality tests are themselves capable of yielding fertile results about the morality of actions. And Mill (1863, ch. 1), in similar spirit, claimed that Kant’s attempt to “deduce” ordinary moral duties from FLN fails “grotesquely.” Followers of Hegel and Mill are legion and, given the 1

2

Kant’s rationale for this re-casting is elaborated in the Critique of Practical Reason, chapter II, in a section entitled, ‘On the Typic of Pure Practical Judgment’. Kant writes: ‘I understand by canon the sum total of the a priori principles of the correct use of certain cognitive faculties in general’ (KrV, A796/B824).

The Categorical Imperative and Universalizability

159

emphasis in moral philosophy on providing a theory of right conduct, rejection of Kant’s tests has led many to reject his moral theory. 3 But the wholesale rejection of Kant’s moral theory (or even just his theory of right conduct) based on the rejection of his universality tests, is surely too hasty.4 Moreover, interpreters who are generally sympathetic to Kant’s ethics are divided over the adequacy of FLN; some are pessimistic about its adequacy, others are more optimistic, claiming that the various standard objections to Kant’s tests (including those of Hegel, Mill, and their followers) can be overcome when the tests are properly interpreted (or reinterpreted). The proper understanding and adequacy of Kant’s conception of universalizability as encapsulated in FUL and especially in its variant, FLN, continues to be a lively source of debate. Part of the aim of this commentary is to provide an overview of the major issues and questions about Kant’s universal law formulations of the categorical imperative (CI, hereafter) that have been raised by scholars over the years. These formulations (and FLN, in particular) are almost always taken as representing a moral decision procedure and, in § I, I will set out the main steps in this procedure in order to explain some of the interpretative questions that are have become the central foci of debate about it. I will then proceed in § II to articulate a set of assumptions that comprise what I will call the “strong model” of the decision procedure associated with FLN. This model involves a set of core assumptions about universalizability, but it also includes a series of strong claims about the deliberative power of Kant’s tests. Then, in § III, I explain why most recent interpreters of Kant’s tests (including those who are optimistic about these tests) reject the strong model, hoping perhaps to salvage the core, but then I go on to raise certain (by now familiar) objections against FLN that attack the core assumptions. Parts I–III, then, are mainly concerned to explain one dominant interpretation of the role of universalizability in Kant’s moral theory and indicate some recent trends in its interpretation. Many interpreters have thought that a solution to the infamous problem of relevant descriptions can protect FLN as a decision procedure from the standard objections, but in § IV I argue that this is 3

4

Even those sympathetic to Kant’s universalization tests will agree with Otfried Höffe (2002, 144) that “[Kant’s] manifold assurance that the universalization of false promising runs into contradiction stands in peculiar contrast to the absence of a precise demonstration of the point.” And the same point holds with respect to Kant’s use of the test in connection with other duties as well. A point forcefully argued by Allen Wood (1999, 97–110).

160

Mark Timmons

wrong; a solution to the problem of relevant descriptions (which can be found in Kant’s moral writings) will not save the universality tests from the objections. Assuming that this pessimistic stance is correct, we need to ask what would be lost from the Kantian project were we to reject FLN, a question I take up in § V. Here, somewhat tentatively, I will propose what I call the formal constraint interpretation of universality formulations of the CI, according to which the philosophical significance of the core idea in FUL and FLN – the idea of duty involving law-like universality – represents a formal constraint (actually a set of formal constraints) on the content of moral reasons. To suppose that this set of formal constraints on moral reasons can itself function as an adequate decision procedure is mistaken. After presenting and partially defending what I am calling the formal constraint interpretation of FUL and FLN, I turn briefly to the role of these formulations in moral deliberation. I suggest that the idea of considering whether one’s maxim can be universalized, though it is not suitable as the basis for a moral decision procedure, still may play an important role in moral deliberation. However, before going on let me make clear that here I do not plan to follow a kind of standard script for writing papers and book chapters about the CI and universalizability in which the author first devises some scheme for organizing the familiar interpretations of how Kant’s tests are supposed to generate contradictions, criticizes them, and then perhaps proposes some new and better interpretation (which is usually a matter of tweaking one of the familiar options). I do not have anything against doing this, but it has already been done quite well by others. 5 In what follows, I do make passing reference to some of these interpretations, but in looking over the past half century or so of work on Kant and universalization (and particularly the past twenty years), I am interested in sorting out some issues in ways that I have not seen in print and then proposing a reading of the significance of this notion that departs from how it is typically understood, and perhaps how Kant himself understood it.

5

Here are a few of relatively recent vintage: Korsgaard (1985), Herman (1993), Baron (1997), Wood (1999, ch. 3), and Kerstein (2002, ch. 8). In the appendix below, I have attempted to organize the various interpretations of how contradictions are supposed to be generated from an application of the CI.

The Categorical Imperative and Universalizability

161

I. Kantian Universalizability as a Decision Procedure: A Few Basics and Many Questions The standard reading of FLN has to do with its role: it is supposed to represent a moral decision procedure that makes central use of a universality test. In philosophical ethics, there are various distinct universalizability theses and principles.6 Kantian universalizability, expressed in FLN and taken as the basis of a decision procedure, has two distinctive features. First, this principle (and the universality tests it expresses) is supposed to represent a substantive moral principle in the sense that its application to actions yields definite moral verdicts about the actions being evaluated. It thus contrasts with the so-called logical (non-substantive) principle of universalizability which merely tells us that if an action, performed in some set of circumstances is right (wrong), then any relevantly similar action, performed in relevantly similar circumstances must be right (wrong). The second distinctive feature of Kantian universalizability is that the tests are formal in the sense that at their core is a consistency constraint that is supposed to do the work, so to speak, in generating deontic verdicts about actions. This feature contrasts with deliberative procedures that would, for example, make essential appeal to a theory of the good, whether by appealing to perfectionist ideals or subjective conceptions of the good based on, for example, preferences of individuals. The attraction of Kantian universalizability, then, is that it promises to yield substantive moral verdicts from a formal deliberative procedure. Let us explore this further. The content of FLN (as a moral decision procedure), then, is a general deliberative procedure that can conveniently be divided into four steps or stages, each step raising a number of interpretative questions which, as I have said, have been center stage in debates about Kant’s principle. As reflected in Kant’s famous Groundwork applications of FLN, the deliberative procedure is described from the point of view of an agent who, in one of Kant’s examples, is contemplating making a lying promise, but “still has enough conscience to ask himself: is it not forbidden and contrary to duty to help oneself out of need in such a way?” (GMS, 422). As a first step in the deliberative procedure, Kant has this person formulate the maxim upon which he would be acting were he to make the lying promise. In GMS, 422, Kant characterizes 6

For an excellent discussion of the main varieties, see Narveson (1985).

162

Mark Timmons

a maxim as “the subjective principle of acting […] [which] contains the practical rule determined by reason conformably with the conditions of the subject (often his ignorance or also his inclinations), and is therefore the principle in accordance with which the subject acts […].” Although in formulating sample maxims to be tested by FLN, Kant includes sometimes more and sometimes less information about the agent’s action and circumstances, it is clear that a maxim, when fully stated, includes a specification of: (1) circumstances of choice, (2) the action, and (3) the end, or what Kant calls the “matter” of action – the purpose or motivating reason for which one is proposing to undertake the action in question. So, the first step in the deliberative process is: 1. Formulate one’s maxim (or the maxim which one is contemplating acting upon) having the form: If/whenever _________ I will _________ in order to _________,

where the blanks are to be filled respectively with a description of one’s circumstances, action, and end. The next step involves asking the question: “how would it be if my maxim were to become universal law?” (GMS, 422). Thus, one is to: 2. Raise one’s maxim to the status of universal law by reformulating it to have the form: Of necessity, if/whenever _________ everyone will ________ in order to _________.

Next, the consistency tests are applied: 3. Consider whether, in willing this imagined law of nature to hold, one is thereby committed to a contradiction – either in conception or in volition. Finally, given the connection between the universalizability of the maxim and the deontic status of actions, we reach a moral verdict: 4. If the maxim is universalizable (it can without inconsistency be conceived and willed as a universal law of nature) then action on the maxim is morally permitted. If, however, the maxim is not universalizable, then action on the maxim is morally forbidden (wrong) and, depending on which test the maxim fails, the action mentioned in the maxim is either a violation of perfect or of imperfect duty. In relation to this step-wise sketch, let us take a brief tour of the main interpretative questions that arise at each step, and then in later sections, we shall consider some of these questions in more detail.

The Categorical Imperative and Universalizability

163

Step 1: formulating one’s maxim The main interpretative question about maxims and thus about step 1 of the deliberative procedure is how one’s circumstances, action, and end are to be described for purposes of moral testing. This is the problem of “relevant descriptions,” and its significance for FLN is nicely expressed by W. D. Ross: Any individual act is an instance of a class of acts which is a species of a wider class; we can set no limit to the degrees of specification which may intervene between the summum genus “act” and the individual act. For example, if C tells a lie to the would-be murderer, this falls (i) under the subgenus “lies told to murderous persons,” (ii) under the species “lies,” (iii) under the genus “statements.” Kant pitches, arbitrarily, on the middle one of these three classes, and since acts of this class are generally wrong, and are indeed always prima facie wrong, he says that the particular lie is wrong. But the man who tells the lie may well retort to Kant “Why should the test of universalizability be applied to my act regarded in this very abstract way, simply as a lie? … The test of universalizability applied at one level of abstractness condemns the act; applied at another level of abstractness it justifies it. And since the principle itself does not indicate at what level of abstractness it is to be applied, it does not furnish us with a criterion [in my terminology, a decision procedure, M. T.] of the correctness of maxims, and of the rightness of acts that conform to them (Ross, 1954, 32 f.).

Ross’s example might prompt the thought that the more detail the better, since in his example the description lying to a would-be murderer is both the most detailed and intuitively correct maxim for moral assessment. But in many circumstances, there will be many details (e. g., one’s age) which are not morally relevant and whose mention in one’s maxim will affect its universalizability and thus lead to intuitively mistaken moral verdicts. So, presumably, a solution to the problem of relevant descriptions would specify which features of one’s choice situation are morally relevant and thus ought to be included in a formulation of one’s maxim and, by implication, which features lack moral relevance and ought not be included. We return to this issue in § IV. From step 1 to 2: raising one’s maxim to universal law Going from step 1 to step 2 is a matter of raising one’s maxim to its universalized counterpart, but there are questions about how this is to be done. The dominant interpretation (reflected in the above scheme) is that one imagines one’s maxim raised to the status of a

164

Mark Timmons

psychological law according to which everyone who recognizes or at least believes that she is in circumstances specified in the maxim, will necessarily perform the specified action for the purpose in question. We thus are to consider the universal conformity to the maxim. In opposition to this interpretation, Thomas Pogge (1989) has defended the view that we are to raise our maxim to universal law by imagining its universal availability: we are to imagine that everyone feels “morally free” to adopt and act on the maxim in question, thus allowing that not all agents will in fact do so.7 (In step 2, this would require replacing “Necessarily everyone will” with “Everyone may”). The main philosophical importance of this interpretative issue concerns the aim of making Kant’s tests yield as wide a range of intuitively acceptable moral verdicts as possible. We will return to the issue in § III. Step 3: the consistency tests This step has generated perhaps the most discussion about Kant’s deliberative procedure, and in the secondary literature we find a variety of interpretations of exactly how contradictions can be generated in connection with non-universalizable maxims. In the Groundwork, Kant explains that the general canon of moral appraisal, encapsulated in FUL and FLN is that “we must be able to will that the maxim of our act become universal law” (GMS, 424). He goes on to describe and illustrate two apparently distinct consistency requirements implicit in the general canon: the so-called contradiction-in-conception (CC) and contradiction-in-the-will (CW) tests.8 Those fascinated by the question of how contradictions are supposed to be generated in 7

8

Pogge notes that in some places (e. g., GMS, 403) Kant poses the crucial universalizability question by asking whether one can consistently will that everyone may (as opposed to will) act on her maxim; hence, the textual basis of the universal availability interpretation. In his 1954 book, Ross argued (in effect) that the universal conformance to the false promising maxim featured in Kant’s example would not ‘destroy itself’ since “if people acted always on this maxim every promise could be relied upon to be broken, and this would in many (though not in all) cases serve as well as if it could be relied upon to be kept” (p. 30). Since he does think the universal availability interpretation “would, no doubt, lead to chaos,” he too favors this latter interpretation. Allen Wood (1999, 80) claims that Kant’s presentations of FUL and FLN in the Groundwork express the universal availability and universal conformity interpretations respectively. More recently, O’Neill (1996, chs. 2 and 6) has proposed what she calls a thin ‘modal’ interpretation of the universalizability test in terms of universal adoptability, though O’Neill is less concerned about interpreting Kant than she is about making use of a plausible understanding of this sort of test. These labels were introduced by O’Neill (Nell) in her 1975 book and have since caught on in the secondary literature.

The Categorical Imperative and Universalizability

165

connection with non universalizable maxims, have mainly focused on the CC test, and the various interpretations of it can be usefully organized and compared by considering how they stand with respect to two questions: (1) What kind of ancillary information – information in addition to the maxim and whatever is conceptually involved in its being raised to universal law – is allowed to figure in generating contradictions in conception? (2) What is the epistemic status of this additional information?

With regard to these questions, interpretations are more or less austere, where the most austere of them allow only (broadly speaking) conceptual ancillary information, knowable a priori – information about the actions being tested or about rational agency – to play a role in generating contradictions.9 More opulent views permit the inclusion of such empirical information as teleological laws or causal laws that happen to govern (or in the case of teleological laws, may be taken as if they govern) our world, including the implications of the operation of such laws in a world like ours. This issue about the kinds of information that may be used to generate contradictions affects how well the tests can generate intuitively correct results for admittedly wrong actions. Compared to austere interpretations, more opulent interpretations allow empirical information to play a role and (arguably) make the tests more useful in generating moral verdicts about a wider range of issues bearing on human conduct and thus more powerful as a decision procedure (about which more will be said in the next section). The issue concerning epistemic status affects the scope of those agents for whom the moral verdicts reached by the procedure hold. Given Kant’s repeated claim that moral laws must hold for all (finite) rational agents (including, but not limited to human agents), there is pressure to interpret the tests austerely so that minimal assumptions about, e. g., rational agency, are involved in deriving moral verdicts.10 So there would appear to be a kind of trade-off here: the more austere one’s interpretation, the greater the scope of the moral conclusions reached by the deliberative procedure, however, more opulent interpretations, though narrower in scope, allow for a greater range of implications about duties and obligations that pertain to human beings. 9 10

Galvin (1991), for example, defends an extremely austere interpretation. The most famous of these is in the preface to the Groundwork at 389, but also occur at GMS, 408 and GMS, 410n.

166

Mark Timmons

Step 4: inferring a moral verdict Most commentators have thought that the tests associated with FLN are supposed to yield moral verdicts about the deontic status (the rightness or wrongness) of actions, and we have seen how the issue of ancillary information affects the scope of such verdicts.11 But there are additional questions about this step. For instance, most interpretations seem to take for granted that Kant’s tests yield conclusions about the “objective rightness” of actions, but on some interpretations the tests only yield verdicts about the “subjective rightness” of actions.12 Moreover, some interpreters, dubious about deriving deontic verdicts from Kant’s tests, have claimed that they can be used (or ought to be understood as useful) in generating moral verdicts about the moral worth of maxims and the actions one performs on their basis.13 Before going further, let us sum up. (1) The question about the proper characterization of maxims for moral testing is obviously crucial for making sense of Kant’s tests and we will return to this issue below. (2) Although there is some dispute about how one is to raise one’s maxim to universal law, let us stick to the standard, universal conformity view according to which in raising our maxim to a universal law we consider a hypothetical situation in which necessarily anyone who satisfies the antecedent of the conditional hypothetical law, does (not just may) satisfy the consequent (perform the action in question). (3) The real heart of much dispute about Kant’s tests concerns their details and, in particular, how (on what basis) contradictions can be generated (if at all). (4) Disputes about the kinds of moral verdicts Kant’s tests are able to generate largely depend on how one comes down on the first three questions. Again, here I will stick to the standard interpretation and take for granted that 11

12

13

In MdST, 393, Kant refers to ‘inner actions’, by which I take him to be referring to such actions as consciously adopting ends and which is reflected in his discussion of maxims of ends (adopting ends) and contrasted with maxims of (overt) action. See, for instance, O’Neill (1975) for an interpretation of Kant’s tests that restricts deontic verdicts to ones about subjective rightness. See Timmons (1997) for a defense of the claim that Kant’s tests are not restricted to yielding verdicts about the subjective rightness of actions. See O’Neill (1985). In one place, Kant himself suggests that appeal to universalizability can be used for deriving both deontic verdicts and verdicts about worth. After presenting FUL in section I of the Groundwork and illustrating its use in connection with making a lying promise, Kant remarks that “with this compass in hand, [common human reason] knows very well how to distinguish in every case that comes up what is good and what is evil, what is in conformity with or contrary to duty” (GMS, 404).

The Categorical Imperative and Universalizability

167

the moral verdicts in question are primarily about the rightness and wrongness of actions.14 This concludes, then, a first pass over the main steps in the deliberative procedure associated with FLN and some of the main questions that arise at each step. We now proceed to examine in more detail the very idea of FLN as a decision procedure, which will lead us to make contact with some of the recent (and not so recent) secondary literature. II. FLN as a Decision Procedure: Core Assumptions and the Strong Model Fundamental moral principles are often cast in the dual role of providing both a decision procedure for moral deliberation and a moral criterion. As I am using the term, a moral principle represents a moral criterion of right action when it purports to specify those fundamental morally relevant (nonmoral) features (one or more), possession of which by an action, makes it morally right. According to classical hedonistic versions of act utilitarianism, for instance, the fact that a concrete action (open to an agent in some circumstance) would produce as much net happiness as would any other available alternative action, is what makes the action in question morally right; alternatives lacking this feature also lack the feature of moral rightness, and so (given a standard understanding of deontic concepts) are morally wrong. Understood in this way, the principle of utility represents a moral criterion of right action. It may also be understood as representing a moral decision procedure, according to which in moral deliberation we are to estimate the utilities associated with various alternative actions and decide accordingly. But it need not be taken as a decision procedure, and some defenders of the utilitarian theory have argued that it should not be understood in this manner.15 The point is that a moral principle need not be cast in the dual role of decision procedure and moral criterion; it may play one role without playing the other in the overall economy of some moral theory.

14

15

Doing so still leaves open questions about, for example, whether the verdicts represent claims about the objective deontic status of actions or whether they only represent verdicts about subjective rightness. See, for example, Sidgwick (1907, book IV, ch. 1, § 1), Bayles (1971), and Brink (1989, ch. 8).

168

Mark Timmons

Let me suggest, then, that we simply focus on the hypothesis that FLN represents a decision procedure, putting aside the question of whether it is also a moral criterion. I distinguish between those core assumptions of FLN as a decision procedure that are most central to its functioning in this role from what seem to be additional desiderata that may (arguably) be rejected without giving up this principle – what I call the “strong model” of FLN. In what immediately follows, I lay out the core assumptions and then proceed to describe the strong model. A. FLN: core assumptions 1. Self-sufficiency Kant claims that the CI is the sole fundamental moral principle and this seems to imply that FLN, in its role as a decision procedure, can be employed as the sole moral premise in an argument that (together with other nonmoral premises) can be used to derive deontic verdicts about actions. That is, the four step deliberation procedure set out above provides the basis for a Kantian moral argument of the form: (1) An action A is right only if (and perhaps if)16 its maxim M is universalizable; (2) Maxim M is/is not universalizable; thus, (3) A-ing (acting on maxim M) is right/wrong. Of course, this simple argument form will require that premise 2 be supported by a further argument about M’s universalizability, but the idea of FLN as self-sufficient is that it is supposed to be the only substantive moral premise in the three step argument.17 2. Coherence Kant also claims that because the CI is implicit in ordinary moral thinking –thinking which “becomes even subtle” (GMS, 404) – it can be used to derive a battery of common sense moral judgments, assumed to be correct. In short, the moral verdicts about actions that can be reached by using FLN as a decision procedure are supposed to cohere with ad16

17

Whether FLN is plausibly understood as expressing both a necessary and sufficient condition of rightness has to do with what I call below its ‘deontic strength’. The issue of self-sufficiency, notice, is separate from the question of deontic strength. Even if FLN at best expresses a necessary condition of rightness, it will be self-sufficient if it can serve as the lone moral premise in the sort of moral argument we are here concerned with. For example, Nelson Potter writes: “A successful interpretation of the categorical imperative consists of an argument having only one premise (the categorical imperative), and whatever factual and causal empirical premises are needed, from which a conclusion concerning the moral rightness or wrongness of some particular kind of action follows” (1973, 397).

The Categorical Imperative and Universalizability

169

mitted correct common sense moral verdicts about those actions. This requirement, note, will be met if, for whatever reason, the principle, together with relevant nonmoral information, yields correct moral verdicts in all actual and possible cases in which the test yields verdicts.18 3. Relevance Not only must an adequate moral principle satisfy the coherence criterion, but it must do so for the right reason in the sense that as a decision procedure the moral principle either expresses or otherwise is “internally” related to facts about actions in virtue of which they have the deontic status they have. Assume that FLN is not a moral criterion but that the humanity formulation is. Then, for FLN to satisfy the relevance condition, the fact that a maxim is not universalizable must be somehow internally connected to the fact that the action in question fails to properly respect humanity as an end in itself.19 To summarize: self-sufficiency articulates what is required for FLN to play the role of a decision procedure; coherence represents perhaps the most basic condition of adequacy of a decision procedure; and relevance serves to help explain why it is that a self-sufficient decision procedure does satisfy the coherence criterion. Giving up on some or all of these seems tantamount to giving up on FLN as an adequate decision procedure; they represent core assumptions. By contrast, the following four additional theses each represent ways in which the core may be strengthened thus making FLN a very potent decision procedure. B. FLN: the strong model 4. Deontic power There is textual evidence that Kant understood FLN as expressing both a necessary and sufficient condition of moral rightness.20 If this principle does have this kind of deontic power, then not only will it be 18

19

20

My formulation here is meant to capture a very weak coherence requirement according to which FLN never leads to deontic verdicts that conflict with common sense moral convictions. Strictly speaking, this requirement will be met even in case FLN simply fails to yield any deontic verdicts. In this case, the complaint would be that FLN is useless because completely indeterminate. Although I do not take this desideratum to be quite as central as are the other two to the success of FLN as a decision procedure, it certainly represents an assumption that one would expect it to satisfy if it indeed satisfies self-sufficiency and coherence. The importance of this desideratum is stressed by Herman (1989 and 1993). See Pogge (1989, 189 f.) on this point who cites GMS, 421, and other passages as evidence that it expresses a necessary condition and GMS, 403, as evidence that Kant took it to express a sufficient condition as well.

170

Mark Timmons

able to rule out actions and omissions that are morally wrong, but it will be able to rule in actions and omissions as obligatory. This would mean, further, that FLN would be powerful enough to generate negative duties and positive duties. 5. Act-level determinacy According to this desideratum, FLN is powerful enough to yield conclusions about the deontic status of a wide range of reasonably specific act tokens – concrete doings (or omittings) that are or might be performed by a person at a time in a particular set of circumstances. 6. Classificatory accuracy The classification at issue here is between importantly different types of duty. Kant claims that his two tests are sufficient for accurately classifying morally required actions and omissions into the categories of perfect and imperfect duty. 7. Maximum scope The scope of a moral principle is a matter of the width or generality of the class of agents to whom the principle is properly addressed. A principle has wide scope vis-à-vis humans when it holds for all accountable human agents. A principle has maximum width of scope if it holds for all finite accountable agents, human and non-human. Certainly Kant thinks that the CI has maximum scope. The more interesting and controversial thought is that derived mid-level moral generalizations of the sort featured in the Tugendlehre system of duties (what I will call moral rules) also have maximum scope. As I will proceed to explain further in the next section, these latter four desiderata are not a package deal: one might accept some of them and reject others (what I am calling a strong model comes in degrees of strength). But let us pause before moving on in order to bring together the key ideas from this and the previous section. We now have before us what I take to be the main interpretative questions about Kant’s universalizability tests (§ I) and in this section we have set out the various core assumptions implicated in taking FLN as a decision procedure as well as made note of various desiderata each of which represents one way of strengthening FLN as a decision procedure. In the next section, I will briefly indicate why many recent interpreters (some of them sympathetic to FLN as some kind of adequate decision procedure) deny one or more of the strong

The Categorical Imperative and Universalizability

171

theses. This will help us focus on how those who are sympathetic to FLN as a decision procedure understand its deliberative power. But then in §§ IV and V, I want to get down to business and argue that: (1) contrary to what some believe, the promise of FLN as a decision procedure, whether strong or weak, cannot be saved (as perhaps some think) by appeal to a theory of relevant descriptions, that therefore (2) we ought to reject FLN as an adequate decision procedure. All this is negative. But I then go on to argue, in a more positive vein, that (3) the significance of universalizability in Kant’s ethics need not rest with the CC and CW tests, rather (4) it is best understood as a metaethical constraint on moral reasons, and finally that (5) if we view Kant’s Groundwork applications of FLN to maxims in the context of the book’s overall strategy, then we should view Kant’s tests as representing ad hominem arguments addressed to morally conscientious agents who are capable of experiencing moral demands but who are tempted by considerations of self-interest to violate those demands. III. Troubles for FLN as a Decision Procedure In this section, then, I continue our investigation of FLN by first explaining the move away from strong interpretations by interpreters who remain (somewhat) optimistic about this principle as a decision procedure. However, I go on to explain why I think that a pessimistic verdict about FLN is warranted; its importance within Kant’s ethics is not primarily that of a decision procedure. A. Chipping away at the strong model Let us consider the grounds for rejecting each of the four strong theses, taking them up in reverse order. 7. Wide Scope In the preface to the Groundwork, Kant famously claims that: “Everyone must grant that a law, if it is to hold morally, that is, as a ground of obligation, must carry with it absolute necessity; that, for example, the command “thou shalt not lie” does not hold for human beings, as if other rational beings did not have to heed it, and so with all other moral laws properly so called; that, therefore the ground of obligation here must not be sought in the nature of the human being or in the circumstances of the world in which he is placed, but a priori simply in concepts of pure reason” (GMS, 389).

172

Mark Timmons

How properly to interpret this passage, especially in light of Kant’s other moral writings, is not clear. Some interpreters have read it as claiming that in deriving general moral rules from the CI, one must not appeal to any premises about the nature of human beings or the circumstances of their existence. However, in light of what Kant says in the Metaphysics of Morals (in addition to issues of philosophical plausibility), other interpreters have argued that Kant’s considered position only commits him to the claim that the fundamental moral principle – the CI – must be grounded entirely in a priori claims about rational agency. Here is one place where what I have called austere versus opulent interpretations of Kant’s tests matter.21 This particular dispute may be clarified if we distinguish three levels of moral judgment in Kant’s ethics. At the most foundational level there is the CI – the supreme principle of morality – whose basis and justification are entirely a priori and so considerations of human nature and circumstances play no role in what grounds it or how it can be justified. At the second level, we have a battery of mid-level moral generalizations of the sort featured in the Tugendlehre – duties of selfperfection whose derivation obviously depend on empirical facts about human beings, and duties of happiness to others, including duties to others as human beings.22 Finally, at the third level, we have specific concrete duties – duties to perform or refrain from performing some specific action on some specific occasion – which, as Kant makes clear, require that we take into account specific circumstances. In connection with our duties to others, Kant mentions a number of possibly relevant circumstances including “differences in rank, age, sex, health, prosperity or poverty, and so forth” (MdST, 469). The dispute over scope (which rests on the dispute over what considerations may legitimately play a role in Kant’s tests) is typically focused on the second level of moral judgment. All parties agree that a maxim will include empirical information about one’s circumstances, ends, and proposed action. However, as explained earlier, austere in21

22

Galvin (1991) and Swoyer (1983) defend the idea that only a priori ancillary assumptions are allowed in Kant’s tests, many others, including, Gram (1967), Potter (1975), Harrison (1957, 1958), O’Neill (1985), Timmons (1984), Wood (1999), interpret Kant’s tests as allowing empirical ancillary assumptions to play a role in generating contradictions. Here I am compressing two distinct levels: (1) the level where Kant appeals to very general information about human beings in deriving the basic obligations that are at the same time ends (self-perfection and happiness of others), and (2) the level of more specific duties to oneself and duties to others that is elaborated in Parts I and II of the Doctrine of Elements of Ethics in The Metaphysics of Morals.

The Categorical Imperative and Universalizability

173

terpretations only allow broadly conceptual truths to figure into the tests, while more opulent interpretations allow general empirical information – information in the form of empirical or perhaps teleological laws and their implications. I believe that the past twenty years of interpretative work on Kant’s tests shows a tendency to embrace opulent interpretations of Kant’s tests and thus reject the thesis of wide scope. One main reason for this is simply that in working out the details of Kant’s CC and CW tests, it certainly appears as if he must be appealing to empirical claims about human circumstances, motivation, and knowledge. For instance, in arguing that false promises are wrong, Kant tacitly appeals to the empirical claim that in the hypothetical situation in which one’s maxim has become a natural law, people would come to know the hypothetical law in question. In a world where everyone suffered severe memory loss as did the character Leonard in Christopher Nolan’s 2000 film, Momento, no contradiction in conception would arise. Similar remarks apply to Kant’s other applications of FLN in the Groundwork and elsewhere.23 Another, related reason in favor of opulent interpretations is greater attention to the details of Kant’s presentation of a system of duties in the Metaphysics of Morals and the seeming greater plausibility of deriving a second tier of moral generalizations on the (partial) basis of anthropological information (as Kant would call it) about human beings qua human. Thus, some interpreters argue that it is only the supreme principle of morality whose justificatory basis is (and must be) a priori, and that derived duties are (perhaps must be) justified by appeal to facts about human beings. This in turn means that Kant’s system of duties in the Tugendlehre extends only to human beings and to creatures like them in their makeup. 24 But giving up on wide scope at the level of derived moral rules, though it may conflict with some of Kant’s Groundwork declarations, does not undermine the general adequacy of Kant’s universality tests. 6. Classificatory accuracy The classificatory accuracy of the CC and CW tests is suspect. On this point, I believe that most all contemporary commentators agree that 23

24

For an overview of the empirical assumptions operative in Kant’s tests see Wood (1999, ch. 3). Among those who allow information about human beings qua human as a legitimate part of Kant’s derivation of duties disagree over the epistemic status of such information. See Swoyer (1983), Pogge (1989), and Herman (1993).

174

Mark Timmons

the “correspondence thesis” (as Wood calls it) cannot be defended.25 One would think that at the very least acts of murder would fail the CC test – seemingly they are prime candidates for violations of perfect duty. But Barbara Herman (1989), for instance, argues that instances of murder and mayhem at best only fail the CW test and thus would have to classified as violations of imperfect duty. And, indeed, no matter which particular interpretation of Kant’s CC test one accepts, it is just not plausible to suppose that all maxims upon which one might murder someone will be somehow inconceivable as a law of nature.26 This is bad news for the thesis of classificatory accuracy. 5. Act-Level Determinacy To say that a decision procedure is determinate is to say that it is powerful enough to yield deontic verdicts about a full range of particular duties – duties at the third level where questions about what a particular person ought to do, on some particular occasion (given the morally relevant details of her situation) are in focus. After all, it may be that the deliberative procedure prescribed by the FLN can be used to generate verdicts about our general duties – moral rules at the second level – but the procedure may not be of use in going from general duties to concrete cases. Perhaps, for instance, Kant’s CC and CW tests can be used to generate a set of moral rules, but applying them requires moral judgment (both in determining to which situation those rule apply and how conflicts of rules are to be adjudicated). This would limit the determinacy of FLN as a decision procedure. Again, I think that there is a trend away from thinking that Kant’s tests satisfy this desideratum. For instance, O’Neill (1996), who is sympathetic to Kant’s universality tests, restricts their use to that of generating general moral rules of justice and virtue, and Herman (1993), at least as I understand her “deliberation model,” agrees with O’Neill on this point.27 But again, rejecting this desideratum does not 25 26

27

See Wood (1999, 97–101). Christine Korsgaard’s so-called ‘practical interpretation’ of the CC test works so long as the murderer’s maxim includes some end of action (which the killing is supposed to promote or bring about) such a getting a job that would otherwise go to one’s only rival. In a world in which everyone adopts and (when relevantly situated) acts on this maxim, one would fail to achieve the getting of the job by acting on one’s maxim. But, as Korsgaard notes, her practical interpretation fails to show any contradiction in conception in cases of murder for revenge. See Korsgaard (1985). In reaching all thing considered judgments about duties in concrete circumstances, Herman claims that the general moral rules generated by Kant’s tests must involve what she calls a ‘value translation’. As I understand her view, we are to use moral judgment in considering the relative normative force of possibly competing moral

The Categorical Imperative and Universalizability

175

mean that Kant’s tests are complete failures as decision procedures. Their fertility may lie simply in generating a system of general moral rules, and that would be significant. 4. Deontic Power Even if one rejects the strong claim about act-level determinacy, one might still hold that FLN and its tests are deontically powerful if one holds that universalizability is both a necessary and sufficient condition for an act type being right. And here we come to what is by far the most common way of criticizing FLN: describing examples in which a maxim passes Kant’s tests but the action in question is nevertheless wrong and thus falsely gives a positive moral evaluation of the action (a false positive), and describing examples in which a maxim fails one or another of Kant’s tests but where the action is not wrong and thus falsely gives a negative moral evaluation of the action (a false negative). False positives challenge the sufficiency of universalizability as a test for rightness; false negatives challenge its necessity. Many recent (and not so recent) FLN optimists have been willing to grant (in light of numerous false positives) that universalizability is not sufficient for rightness. They thus reject the thesis of deontic power. Nevertheless, they maintain the core idea that it represents at least a necessary condition of rightness, perhaps to be supplemented in Kant’s theory by his other formulations of the CI. Again, we take this up in the next section. So suppose we were to agree that Kant’s tests: (1) yield moral verdicts that are limited in scope to human beings; (2) cannot be used to discriminate perfect from imperfect duties; (3) yield moral verdicts of limited determinacy; and (4) represent only negative tests of moral rightness. All of this would weaken the deliberative procedure of FLN, but it might still be useful in this role. Suppose that FLN can be used, apart from other moral assumptions, in moral arguments yielding moral verdicts which, though limited in the ways just conceded, nevertheless cohere with common sense in the verdicts it does yield, and also can be understood to yield those verdicts for the right reason. Then this principle would still be self-sufficient, coherent, and relevant – the more central requirements that a moral decision procedure must satisfy. considerations (represented by the rules) in coming to an all-things-considered moral verdict about a particular case. If this is right, the use of ‘value translation’ does not involve application of Kant’s tests in going from moral rules to concrete verdicts. (I say more about this issue in the next section.)

176

Mark Timmons

B. Attacking the core As I have said, by far the most common form of objection to FLN as a decision procedure involves describing counterexamples – false positives and false negatives – and then concluding that this principle fails the coherence desideratum. Such counterexamples abound in the secondary literature.28 In attempting to diagnose why FLN is (seemingly) susceptible to counterexamples, a number of interpreters have thought that this problem is symptomatic of a deeper problem in Kant’s ethics: the lack of principled theory of moral relevance. The idea is that in many cases, maxims that pass Kant’s tests (when they should not) are full of morally irrelevant detail about the agent and her circumstances. So, for instance, suppose I make a false promise on March 8 to Igor Cycz in order to get some money because I want to buy a metal detector. Were I to formulate my maxim with this information included, the result would be: I will make a lying promise whenever it is March 8 and I believe I can get the money this way from someone named Igor Cycz, in order to buy a metal detector. Presumably, at least on many interpretations of Kant’s tests, this maxim is consistently conceivable as a universal law: in attempting to imagine a hypothetical situation in which if everyone in these same circumstances (so described) adopts and (when appropriate) acts on this maxim, we do not end up in inconsistency. Hence, we have a false positive. Of course, information about the date and the name of my intended victim is not morally relevant in this case. And because these bits of information are responsible (in effect) for the universalizability of my maxim, the thought is that were we to have a principled way of ruling out such information as morally irrelevant, we could avoid this alleged counterexample. We might also diagnose the problem of false negatives this way. Let us go back to Ross’s example in which I am being questioned by some potential murderers whose truthful answer by me will knowingly lead to the death of their innocent intended victim. Here my maxim might be formulated this way: I will lie whenever I am asked a question if I do not want to give a truthful answer. This maxim is arguably not universalizable, but (arguably) this one is: I will lie whenever I am in a 28

Of course, this claim would have to be defended by examining the various interpretations of the CC and CW tests since what works as a counterexample to one interpretation may be avoided by one of its rival interpretations; a project I cannot undertake here. However, for recent, good overviews that substantiate my pessimism see Wood (1999, ch. 3) and Kerstein (2002, ch. 8).

The Categorical Imperative and Universalizability

177

position of being able to save an innocent person’s life in order to protect that person from being harmed. Again, were we to have in hand a principled account of morally relevant features, we would have a basis for arguing that the second maxim is to be used in Kant’s tests since it includes morally relevant information about one’s circumstances and purposes that the former lacks. IV. Will a Theory of Relevance Save FLN? I myself, from time to time, have optimistically suggested that alleged counterexamples to FLN can be dealt with in this way.29 My current skepticism is not due to the fact that Kant’s theory does not include a principled account of moral relevance – this is a quite absurd suggestion, anyway. Rather, as I will explain shortly, making use of Kant theory of moral relevance undermines FLN as self-sufficient and it cannot save Kant’s tests from all obvious counterexamples. Let us explore this issue by first bringing into focus Kant’s theory of moral relevance, and then proceed to examine how his theory might be of use in his universality tests. A. Kant’s solution to the problem of relevant descriptions So here is the hypothesis under consideration. Were Kant to have a principled account of moral relevance, then we could use it to determine which features of one’s circumstances are morally relevant and should be included in an expression of one’s maxim for moral testing, and we could also use it to determine which features are not morally relevant and should not be included in a specification of one’s maxim. Now Kant does have a theory of moral relevance that is expressed in humanity formulation of the CI and is elaborated in the Tugendlehre system of moral duties. Let me explain. 30 A moral theory (whether of right conduct or value) is in the business of specifying those fundamental features or properties of actions, persons, and their circumstances in virtue of which an action or other item of evaluation has the moral properties it has. 31 The fundamental moral principles featured in a moral theory that purport to pick out which features, most 29 30 31

Timmons (1997, 2002b, ch. 7). For more detail about Kant’s theory of moral relevance, see Timmons (1997). I am helping myself to ‘property’ talk strictly for convenience; it thus should not be read in any metaphysically heavyweight manner.

178

Mark Timmons

fundamentally, make an action right or wrong, represent (what I am calling) moral criteria. As moral criteria, such principles (or, perhaps more precisely, the features they pick out) represent a theory of moral relevance. For a classical utilitarian (and here I am repeating what was said earlier), the principle of utility (where utility is taken to be a matter of the balance of pleasure and pain that would be brought about by some action), expresses the idea that facts about the utilities of actions are what make an action right or wrong. Such facts are the most fundamental, morally relevant facts; all other considerations that might be relevant in some context derive their moral relevance through their having a bearing on utility. A moral principle, presented as an explanatory criterion, represents a theory of moral relevance. Within Kant’s ethics, it is the humanity formulation of the CI that seems most obviously suited to serve as a moral criterion. It fi nds sufficient elaboration in the Metaphysics of Morals. 32 In broad outline the idea is this. First, Kant claims that what he calls “humanity” is an end in itself – something having objective (as opposed to relative) value. (See especially GMS, 428 f.) At bottom, “humanity” refers to our rational capacities as agents who not only are able to deliberate and act on the basis of what they take to be good normative reasons, but are capable of autonomy which, as Kant says, “is the property the will has of being a law to itself (independently of every property belonging to objects of volition)” (GMS, 440). This objective value – the value of rational autonomy – is the fundamental basis upon which Kant’s theory of right conduct rests. So the central idea is that facts about the bearing of actions on rational autonomy represent the most fundamental morally relevant features of actions. 33 Now granted, the rather “thin” idea of rational autonomy is not itself terribly illuminating as a moral criterion. However, in the Tugendlehre, Kant enriches this notion by appealing to anthropological information about the specific nature of human beings – their capacities and tendencies as rational yet natural beings. Thus enriched, the idea is that the fundamental morally relevant features of actions in virtue of which they are right or wrong, so far as human beings are concerned, require maintaining 32

33

This is not the whole story, however. The principle of autonomy, for reasons that I will explain below in section V, is also deeply implicated in Kant’s theory of moral relevance. Talk of ‘bearing’ here is meant to include not only how our actions causally affect the rational autonomy of persons, but also considerations about what various actions express about rational autonomy. For more on this, see Wood (1999, 141 f.; 147 f.) and Timmons (2002a).

The Categorical Imperative and Universalizability

179

and promoting those essential capacities and tendencies that are essential to the exercise of human rational autonomy. How then do we get a theory of moral relevance from the enriched conception of humanity? There are two points to be made. First, think of the various Tugendlehre negative duties to oneself – duties to refrain from suicide, immoderate consumption of food and drink, improper use of one’s sexual organs – as indicating general types of action which, given basic anthropological facts about human nature, necessarily interfere in some way with the maintenance of humanity in ourselves. Again, think of the various positive duties to oneself – duties to develop our (specifically human) powers of body, mind, and spirit – as types of activity the principled omission of which would interfere with the maintenance and promotion of our humanity. Similar remarks apply to the various negative and positive duties to others. So the first idea here is that the various duties of Kant’s Tugendlehre system represent an anthropologically grounded specification (better: an interpretation) of the (initially) thin notion of rational autonomy. But what does all this have to do with the issue of relevant descriptions? This brings us to the second point. To consider an action as of a certain type is, in effect, to consider it under some description or other. For many such descriptions, because of their importance in moral and social life, we have single terms that serve to pick out actions under a description. To classify an action as a case of suicide is short for classifying it under a more cumbersome description. The same goes for terms such as, “gluttony,” “lying,” “avarice,” “servility,” “ingratitude,” “malice,” and “envy” that are featured in Kant’s system of duties. I am suggesting, then, that the system of Tugendlehre duties represents the outline of a principled account of moral relevance. In cases where some action of mine can be correctly classified as an instance of one or more of the various actions mentioned in the system of duties, then this fact about it is morally relevant. Any other fact about an action that is morally relevant in some context will be derivatively relevant in the sense that its relevance (in that context) can be explained by appeal to one or more of the more basic relevant features that are expressed in the system of duties. B. Why the solution does not save FLN Recall the three core assumptions associated with FLN as a decision procedure: self-sufficiency, coherence, relevance. I will now proceed to explain why Kant’s theory of moral relevance will not vindicate

180

Mark Timmons

either the self-sufficiency or the coherence assumptions. Of course, if FLN does not satisfy the coherence assumption, then it cannot possibly indicate some feature of actions that is “internally related” to what makes an action right or wrong.34 1. Self-sufficiency In connection with Kant’s theory of moral relevance and its bearing on the plausibility of the universality tests, there are two observations worth recording here. First, this proposal vindicates Ross’s complaint (see the above quote) that Kant’s universality principle itself does not specify the “level of abstractness” appropriate for maxims to then be tested by the principle. 35 And so the current proposal represents a rejection of the self-sufficiency thesis, because the moral principle of humanity is now playing a crucial role in step 1 of the FLN procedure – it provides a basis for characterizing one’s maxim so that it is suitable for moral testing. 36 Second, an immediate implication of all this is that FLN cannot be used to derive the various moral rules featured in Kant’s Tugendlehre; as we have explained, those moral rules involve terms that are presumed to pick out morally relevant features of actions. 37 So if the reliability of the tests associated with FLN depends on maxims including all and only morally relevant descriptions, and if those descriptions are taken from the moral rules featured in Kant’s ethical system, then it would seem that FLN cannot be used in arguments to derive those very rules; it rather presupposes them. But this is not the end of the story because it leaves open the possibility that FLN, though not self-sufficient as so defined, can be used in reasoning from general moral rules to specific moral verdicts in concrete circumstances. Specifically, it might serve an “adjudicative” role 34

35

36

37

However, in section V I consider what I take to the philosophical and practical relevance of Kant’s universality tests. O’Neill’s 1975 solution to the problem of relevant descriptions was simply that whatever information one’s maxim (assuming that one’s formulation of it is sincere) does contain, is relevant for use of Kant’s tests. This means, in effect, that FLN does contain a solution to the problem of relevance, contrary to Ross. See O’Neill (1975, 13). One implication of this ‘solution’ is that the tests are restricted to yielding only judgments about the subjective rightness of actions. See O’Neill (1975, 129). This, incidentally, is to reverse the view, expressed by many interpreters, that because the concept of humanity (and the associated requirement of treating humanity as an end in itself) is excessively vague, in order to judge whether some action would constitute a case of failing to treat someone as an end in themselves, one must first use FLN to determine whether one’s maxim is universalizable. See, for example, Singer (1961, 235). In the Tugendlehre, Kant only uses the universal law formulation of the CI in arguing for the duty of beneficence. See MdST, 393, 451, and 453.

The Categorical Imperative and Universalizability

181

in moral reasoning. The idea would be that in judging whether some concrete action, falling under one or more of the basic moral rules, is right or wrong, we formulate our maxim including all and only morally relevant detail and then test the resulting maxim by Kant’s tests. 38 According to this proposal, although FLN is morally loaded because its use presupposes a theory of moral relevance, it is nevertheless useful in reasoning one’s way to moral verdicts in cases where there are what Kant calls “conflicting grounds of obligation” (MdSR, 224). I believe that this proposal also comes to grief and helps reveal something important about FLN and Kant’s universality tests. Here, we need to work by example. So, consider a (fictitious) case made famous by Bernard Williams39 in which George, a young but somewhat unhealthy, unemployed chemist with a needy family, is offered gainful employment in a laboratory to pursue research into chemical and biological warfare thanks to the efforts of an older chemist who wants to help. George, being opposed to such warfare, is reluctant to take the job, but knows that if he turns it down it will go to another chemist whose dedication will likely advance the research project greatly. What is George to do? If we refer to Kant’s Tugendlehre system of duties, it is reasonably clear that the duties to others of beneficence (directed toward his family) and of gratitude (toward the older chemist) are relevant here. Malice (one of Kant’s vices) is perhaps not quite relevant in this case, since if George takes the job his main motivation will not be a hatred of others. But nevertheless a kind of lack of proper regard for the fate of others, and hence a lack of what Kant calls “active sympathy” (MdST, 457) does apply. Furthermore, the duty to oneself to avoid servility, which flows from a general duty of self-perfection to maintain and promote humanity in one’s own person, clearly applies. Arguably, this case falls under more Tugendlehre duties than the ones just mentioned – Kant claims it is a matter of judgment to determine which rules will apply to concrete cases – but for present purposes we need not make a complete inventory of the morally relevant considerations that apply to this case. So, restricting ourselves to the features just mentioned, how could we use this information in applying Kant’s universality tests? One way to proceed (taking now the perspective of George) is to consider the option of taking the job and then formulate a maxim with morally relevant detail filled in as follows: 38 39

This proposal would, of course, embrace the act-level determinacy thesis. In Smart and Williams (1973, 97 f.).

182 M

Mark Timmons

In circumstances where: I am unemployed, I have (for health reasons) dim prospects for employment, I have a family to support, and where I have been offered a job but about whose nature I have reservations (because of how the products of my labor might be used) and, moreover a job which will go to a zealot if I do not take it, I will take the job not only to provide for my family but out of gratitude to the person who has arranged for me to get it.

I submit that on either of the two most plausible interpretations of the CC test (the so-called “logical” and “practical” interpretations) this maxim will pass the test. On the logical interpretation (see the appendix), one considers whether any contradiction is involved in the very conception of one’s maxim as a universal law of nature. The problem with this interpretation is that when it comes to highly specific maxims such as M, their specificity (at least in a great many cases) ensures that they pass the test. On the practical interpretation, one considers whether the end of action (mentioned in the maxim) is achievable by the means (also mentioned in the maxim) in a hypothetical situation in which one’s maxim is universally adopted and (in the relevant circumstances) acted upon. Again, if everyone adopts M and acts on it when the specified circumstances come to pass, the resulting situation does not thereby make it impossible for one to successfully act on M in the hypothetical situation. But the problem is that the same remarks about the logical and practical interpretations apply mutatis mutandis to the contrary of M: M* In circumstances where: I am unemployed, my overall job prospects are dim, my family desperately needs my financial support, and in not taking the job I might be guilty of some degree of ingratitude, and where, furthermore if I do not take the job it will go to a zealot, I will nevertheless refuse to take the job because of the potential uses of the products of my labor and also out of respect for my own integrity.

This example (and many more can be provided) strongly suggest that the CC test cannot help us reach conclusions about concrete cases (in which we need deliberative help) when morally relevant detail is specified in the maxims being tested. What about the CW test? We have not said much about this test. But from Kant’s Groundwork examples of letting one’s (natural) talents rust and refraining from helping others in need, the blue print for its use is fairly clear. The maxims in these cases can be consistently conceived as laws of nature (they pass the CC test), but they cannot be con-

The Categorical Imperative and Universalizability

183

sistently (rationally) willed to be such laws. This is so because (1) qua rational agent, there are what we may call “rational maxims” – maxims which one ought to adopt (if one has not already done so) given one’s nature as a rational agent – and which (2) are inconsistent with the implications of the maxim being tested functioning as a universal law of nature. In Kant’s writings we find appeal to two rational maxims in his Groundwork arguments for the duties to develop one’s own talents and to help others in need. If the maxim to develop one’s own (natural) talents is a rational maxim, then arguably, so is the maxim to preserve and develop one’s perfection as a moral being. Indeed, in the Tugendlehre, the maxim to preserve and promote one’s moral powers is a fundamental positive duty to oneself. And, M (insofar as it represents a sacrifice of moral integrity) conflicts with this positive duty of moral self-perfection. But, of course (as many critics have pointed out) universalization does not play a role in Kant’s argument regarding the duties of selfperfection.40 Since we are examining the idea that Kant’s universality tests (involving raising one’s maxim to the status of a universal law of nature) might be of use in deriving concrete verdicts given a general theory of moral relevance, appealing to rational maxims concerned with self-perfection will not help us out. Let us therefore consider the rational maxim of willing to be helped by others in cases of need. Unfortunately, this seems hardly relevant to the case at hand because one is not proposing to refuse help to others in need. Granted, if George acts on M (and takes the job) his action will express a certain lack of proper regard for the welfare of others; he will not be acting with the greatest possible care. But, here we must be careful about how we interpret the duty of beneficence, otherwise this duty will be implausibly demanding.41 The CW test, just like the CC test, does not seem to be of much help in arguing for conclusions about the deontic status of concrete actions even if one appeals to Kant’s theory of moral relevance. The upshot is that the idea that a theory of moral relevance will save FLN and the associated universality tests from counterexamples is false. More precisely, I have argued that (1) given Kant’s theory of moral relevance, FLN cannot stand as a self-sufficient test for generating moral rules nor (2) can it be used in reasoning from general moral rules to conclusions about concrete cases. 40

41

But see Glasgow (2003) for an intriguing attempt to explain how considerations of temporality play a role in Kant’s universality tests and thus how it is that considerations of universalizability can be made to do work in deriving duties to oneself. For a discussion of this matter, see Rawls (1989).

184

Mark Timmons

2. Coherence As I mentioned earlier, the coherence desideratum has been central in evaluating the plausibility of Kant’s CC and CW tests. And I do think that Kant’s theory of relevant descriptions does help to explain why certain maxims – particularly those that include intuitively morally irrelevant information – are not appropriate for moral testing. But, unfortunately, Kant’s theory of moral relevance cannot save the tests from all counterexamples, even from some of the well-known ones. For instance, the maxim, I will buy toy trains but not sell them, cannot be consistently universalized, but this maxim does not seem to include morally irrelevant information about the action, nor does it seem to omit morally relevant information about the agent, her act, or her circumstances. Here we have a false negative. I think it is also clear that Kant’s tests yield false positives. With false positives the problem seems to be that fairly specific maxims pass the tests. In the above example featuring George, both M and its contrary M* pass the tests. Now, of course, we could just conclude that George’s taking the job and his turning it down are permissible actions – a case where the moral pros and cons are fairly evenly balanced. But notice that if it is the specificity of maxims that get them through Kant’s tests (as seems to be the case on both the logical and practical interpretations) then in the many cases like George’s in which a number of competing moral considerations are relevant, both the maxim and its contrary will pass Kant’s tests which implies that the action in question and its omission are both morally permissible. I do not believe it is plausible to suppose that in all such cases the correct moral verdict is that both the action and its omission are morally permissible – this amounts to massive moral indeterminacy. So, I conclude that Kant’s theory of moral relevance, used as a constraint on formulating maxims, does not save Kant’s universality tests from counterexamples and thus that the coherence desideratum – perhaps the most important of the core assumptions – is not satisfied. C. Rejecting FLN as a moral decision procedure We began with a very strong conception of a decision procedure and, following the train of a long history of interpretation, have found reason to conclude that (1) not only is FLN not a powerful decision procedure, but (2) it is not self-sufficient, (3) it does not seem able to play an adjudicative role in moral reasoning, and (4) even appealing to a principled theory of morally relevant descriptions does not save

The Categorical Imperative and Universalizability

185

the tests associated with FLN from counterexamples. Suppose, then, we agree that FLN is not a self-sufficient decision procedure, or even part of one. One option for those interested in carrying on the project of Kantian ethics is simply to jettison FLN and look to Kant’s other formulations of the categorical imperative for inspiration in developing an adequate theory of right conduct.42 But surely, the very idea of universalizability is philosophically important for Kant’s overall ethical theory, and we ought now to ask what its true philosophical significance is. V. The Formal Constraint Interpretation of the Universal Law Formulation of the Categorical Imperative Let us suppose that the universality formulations of the CI do not represent moral criteria and that they do not represent adequate decision procedures. What then? I am going to propose what I call the formal constraint interpretation of the FUL (and FLN) according to which (roughly) this formulation of the CI represents a formal constraint (or, better, a system of interconnected formal constraints) on what can count as a substantive moral reason. Thus, FUL plays an important role in Kant’s ethics, but not that of a moral criterion or of a decision procedure. So, let us make a fresh start by first focusing on Kant’s notion of universalizability. I mentioned earlier that the term “universalizability” has been used to refer to a variety of importantly different ideas and claims in ethics. Recall that what I have called “substantive” universalizability theses purport to set forth conditions which, together with relevant nonmoral information, have normative implications. According to one brand of utilitarian universalizability, an action is right just in case the following counterfactual supposition involving universalization is true: if everyone who is relevantly similarly situated were to perform the action in question, the utility thereby produced would be at least as high as the “universal conformance utility” associated with any alternative action one might perform instead. Kantian universalizability (as originally explained above) promised to represent a self-sufficient moral principle with substantive moral implications. This, as I have explained, we have reason to doubt. 42

Alan Donagan (1977), for example, develops a systematic moral theory on the basis of the humanity formulation.

186

Mark Timmons

But “universalizability” is also used as a term to refer to metaethical theses about moral concepts. The most well-known of these is the so-called “logical” principle of universalizability, which Sidgwick expressed this way: If a kind of conduct that is right (or wrong) for me is not right (or wrong) for someone else, it must be on ground of some difference between the two cases, other than the fact that I and he are different persons.43 This principle is supposed to express a necessary truth that can be known a priori, and obviously its use in moral reasoning to reach normative verdicts requires additional moral premises. There is nothing particularly Kantian about this principle, as evidenced by the fact that it plays a significant role in the utilitarian moral theories of Sidgwick (1907), R. M. Hare (1963), and M. G. Singer (1961). So, if the significance of universalizability in Kant’s ethics is entirely captured by the logical principle formulated by Sidgwick, then, contrary to what many have thought, there is nothing particularly exciting or special about Kantian universalizability. But I do not believe this is right either. My proposal starts with the suggestion that we ought to distinguish between the philosophical significance of the idea of universal law in the overall argumentative strategy of the Groundwork and its practical significance as expressed in the universalizability tests. The philosophical significance of the idea of universal law is that a proper understanding of what such law in the realm of practical reasoning is all about reveals to us an interconnected set of formal features that moral reasons, qua moral, must have and which serves to distinguish them as a class from non-moral reasons and, in particular, from prudential reasons for action. Thus, Kantian universalizability is substantive in the sense explained above – it has substantive implications about the content of moral reasons. The practical significance of Kantian universalizability is that it can be used in moral deliberation as an ad hominem argumentative device to reveal a kind of duplicity in choice and action and thus make manifest the distinctive authority of moral reasons. Before going on, let me admit that I am taking some liberties with Kant’s texts. Kant himself does not explain the significance of the universal law formulations in quite the way I am proposing. However, I believe that my interpretation of the philosophical significance of these formulations meshes nicely with dominant themes in Kant’s moral theory and that the practical significance I find (see below) in 43

Sidgwick (1907, 379; see also, 209 f.).

The Categorical Imperative and Universalizability

187

these same formulations do a great deal of justice to what the universalizability test in Kant’s writings really does help reveal. Let us now proceed to examine respectively the philosophical and practical import of the universal law formulations of the CI. A. Universality: it philosophical significance Talk of the ground(s) of obligation can be usefully viewed as involving two related issues: an issue about the content of those features of an action in virtue of which it is obligatory, and an issue about the ultimate source of the kind of normativity that is characteristic of moral obligation. The two issues are connected: to serve as a moral reason, a consideration must have a kind of normative authority, indeed, for Kant, normative supremacy.44 Let us examine the matter more closely. I interpret the content question as having to do with a specification of the kinds of considerations (their content) that represent moral reasons – considerations about actions, agents, and their circumstances, that are a basis for (ground) moral requirements. The content question, then, is: “What kinds of contentful considerations qualify as moral reasons?” More precisely, if we assume that there is a single underlying or fundamental consideration bearing on action in virtue of which an action has the deontic status it has, then we can re-state our question as: “What sort of contentful consideration qualifies as a fundamental right- and wrong-making feature?” To answer this question, we must specify some general constraints that any kind of consideration must satisfy in order to count as a moral reason, given our common sense notion of duty. My suggestion on behalf of Kant is that the universal law formulations (in effect) encapsulate a set of formal requirements, based on the common notion of moral obligation, which moral reasons, qua moral, must possess. Doing so is a central part of Kant’s attempt in the Groundwork to “search for” the supreme principle of morality – understood as a moral criterion that picks out a fundamental morally relevant reasoning providing consideration that “grounds” moral rightness. Here, then is a list of three formal constraints on what can count as a basic moral reason, implicated in the universalization formulations of the CI. 44

See T. M. Scanlon (1998, ch. 4) for discussion of these two questions about moral reasons.

188

Mark Timmons

Constraint 1: Law-like character of moral reasons In a number of places in the Groundwork, Kant says that the common sense notion of duty essentially involves the idea that moral requirements hold necessarily for all rational beings as such.45 Let me connect this idea of necessity with the idea of a moral reason. For it to be true of someone that she is under a moral obligation, it must be true that there are considerations about her circumstances in virtue of which she has the obligation in question. If we think that underlying the various and sundry morally relevant features in a particular case that bear on one’s obligations there must be some underlying right-making feature, then Kant is claiming that there must be some underlying feature of actions that makes them obligatory. We can put this in terms of moral reasons talk by saying that for Kant, underlying our obligations is some fundamental consideration that provides a (normative) reason for action. Now, according to Kant’s understanding of morality, the basic moral requirements hold necessarily for all rational agents, by which I take him to mean that necessarily all rational agents qua rational have reason to comply with those requirements. Putting these ideas together, then, we can say that fundamental moral reasons – considerations in virtue of which, most fundamentally, a rational agent is obligated to act in one way or another, must be such that necessarily all finite rational beings have such reasons. To say that necessarily all finite rational agents have such reasons for action commits one to the claim that such reasons are universal in scope. Constraint 2: Supremacy of moral reasons In the Groundwork, Kant often uses the term “supremacy” to characterize the kind of value possessed by actions done solely from duty and the kind of will that is disposed to act from such a motive. But in other places, Kant says that the moral law itself has normative supremacy or “commanding authority” (GMS, 426, 431), and, of course, he does refer to this law as the supreme principle of morality. Now it is a constant theme in Kant’s moral writings that given our ordinary notions of moral obligation, moral reasons are normatively superior to competing prudential considerations.46 So, according to this supremacy constraint, moral reasons must be such that they represent consid45 46

See esp. GMS, 389 and 408. The various contrasts between morality and prudence, including the superior normative force of the former over the latter, is particularly clear in the Critique of Practical Reason in Remark II appended to Theorem IV (KpV, 35 ff.).

The Categorical Imperative and Universalizability

189

erations that provide reasons that are superior in normative force to any competing, nonmoral reasons. Constraint 3 Proper object of respect In the Groundwork, Kant tells us tells us that respect is “a representation of a worth that [not only] infringes upon my self-love” (GMS, 401n), but “far outweighs any worth of what is recommended by inclination” (GMS, 403). He also tells us that the proper object of respect is the moral law. There are two things to note here. First, when Kant says that the law is the proper object of respect, I read him as in effect claiming that the sorts of considerations that necessarily provide reasons for action to all rational beings is a proper object of the attitude of respect. (More on this below.) Second, it is the supremacy of the law or, more precisely, moral considerations that “call forth” or demand respect47 and, in so doing our recognition of them checks our self-love and “strikes down” what Kant calls “self-conceit” – the latter a kind of smug comfort one might take in herself without having made moral considerations fundamental in one’s estimation of self worth. To sum up: moral reasons qua moral must satisfy an interconnected set of formal constraints: they must be such that (1) all rational agents have such reasons for action, (2) they must be supremely authoritative, and (3) they must be a proper object of respect. I am suggesting that these tightly interconnected ideas are (in effect) encapsulated in Kant’s universality formulations of the CI. Looking at Kantian universality in this way, we can now ask what sort of fundamental rightmaking consideration can play this role. I believe that Kant’s answer is expressed in the humanity formulation of the CI. Let me explain. B. The formal constraint argument The task, then, is to identify some consideration that has a kind of worth that necessarily counts as a supremely authoritative normative reason for all rational beings – something that is a proper or fitting object of respect. As I understand Kant’s ethical theory, it is humanity as an end in itself – having what Kant calls a “dignity” – which possesses nonrelative, absolute worth (GMS, 435 f.), and whose normative significance is expressed in the humanity formulation of the CI. In brief, the idea is that if we assume that something of value can ground reasons for action, then, of course, the dignity of humanity can ground reasons for 47

See, for instance, KpV, 80.

190

Mark Timmons

action. Moreover this dignity is something necessarily possessed by all rational agents and possesses an “absolute” nonrelative worth which, according to Kant, “outweighs” any reason grounded in inclination “or at least excludes it altogether from calculations in making a choice” (GMS, 400) and is thus a proper object of respect. That humanity as an end in itself is a fundamental right-making feature of actions is, of course, reflected in how Kant uses the humanity formulation of the CI to argue for and explain the rightness and wrongness of the various duties comprising the Tugendlehre. Assuming that the humanity formulation can provide a basis for common sense moral judgments (something Kant defends in the Tugendlehre), we can understand the humanity formulation of the CI as expressing a moral criterion of right action. This is roughly how Kant’s notion of universal law leads us from formal criteria (that constrain what can count as a fundamental moral reason bearing on action) to a substantive moral criterion. So, on my formal constraint interpretation of the CI, we are to understand Kant as (in effect) arguing as follows: The formal constraint argument 1. According to “common rational cognition,” in order for some consideration bearing on the rightness of action to count as a moral reason, it must satisfy formal constraints 1–3 listed above. 2. Humanity as an end in itself satisfies all the relevant formal constraints. 3. Humanity as an end is the only consideration that satisfies the formal constraints. Thus, 4. According to common rational cognition, humanity as an end in itself is a fundamental moral reason. Since a moral principle cast in the role of a moral criterion is in the business of specifying some one (or more) fundamental morally relevant features that provide normative reasons that purport to explain why an action has the deontic status it does, the import of this argument is that it takes us from formal constraints on moral reasons to a fundamental substantive moral principle – the humanity formulation. Of course, whether this argument and, in particular, its third premise, can be defended cannot be examined here. However, let me conclude this section with two observations about my formal constraint interpretation. First, in GMS, 436, Kant tells us that “Autonomy is […] the ground of the dignity of human nature and of every rational creature.” So, although having dignity is the fundamental value that grounds moral

The Categorical Imperative and Universalizability

191

requirements, it is the autonomy of rational beings – roughly, their capacity for “giving” law – that represents a feature of such beings that is most fundamental in Kant’s theory of right conduct and value. At the beginning of this section, I mentioned that in addition to philosophical questions about the content moral reasons, there are questions concerning the normative authority of such reasons. I also mentioned that the two issues are deeply related in Kant’s thought: in order for a consideration to count as a moral reason bearing on action, it must have a kind of normative supremacy vis-à-vis competing reasons. Kant’s account of normative authority has to do with his notion of our natures as “law giving” creatures, and so a full and complete story about moral reasons in Kant would require exploration of his notion of autonomy. This I leave for another occasion. Second, one of the desiderata that any plausible interpretation of Kant’s notion of universalizability must satisfy is that the kind of constraint expressed by this notion be stronger than the logical thesis of universalizability. This desideratum would be nicely satisfied were Kant’s universalizability procedures adequate tests of the deontic status of actions. But, as explained earlier, we have reason to doubt that FUL or FLN are adequate when cast in the role of decision procedures. But now notice that my formal constraint interpretation takes Kant’s conception of universality as a metaethical constraint on reasons being moral reasons, and as such is supposed to have enough power to collectively to pick out certain considerations, in terms of their content, as being moral reasons. So, on the one hand, it is stronger than the logical principle of universalizability. But on the other hand, understood as I propose, it constrains without serving as a selfsufficient decision procedure from which we can derive an adequate system of duties. The content that the humanity formulation adds to the universal law formulation of the categorical imperative is significant for purposes of reasoning one’s way to a system of duties. And so it is not surprising that when Kant turns his attention to deriving a rich set of duties in the Tudgendlehre, he does so on the basis of the humanity formulation.48 48

I might be accused at this point of false advertising. At the outset, I promised to provide an interpretation of universalizability, but what I have really done is offer an interpretation of the notion of universal law and what I claim are conceptually associated notions (normative supremacy and respect as a fitting attitude). This bait and switch was accomplished by vague talk of ‘universality’, so the complaint might go. Reply: if one wants to use ‘universalizability’ associated with Kant’s moral theory to refer exclusively to a consistency test on maxims involving a hypothetical scenario, I won’t complain. In the next section, I explain what I take to be the true

192

Mark Timmons

C. Universalizability as an argumentum ad homimen There is one obvious loose end that my proposal must deal with and this concerns the practical significance of Kant’s universalizability constraint. There is evidence in Kant’s texts, particularly in the Groundwork, that he thought of FLN as a self-sufficient decision procedure; and this I do not doubt. For instance, in section II, just after introducing FUL, Kant remarks that “if all imperatives of duty can be derived from this single imperative […] we shall at least be able to show what we think by it and what the concept wants to say” (GMS, 421, my emphasis). And in the sample applications we find bits of reasoning that certainly suggest that FLN is being used as a self-sufficient decision procedure; that consideration of one’s maxim as a possible universal law of nature can, together perhaps with auxiliary assumptions, generate moral verdicts. I am not denying that such passages strongly encourage the dominant perception of Kant’s universal law formulations as expressing a decision procedure. But let me suggest that, apart from what Kant perhaps wanted these formulations to represent, the tests associated with them are of use in moral deliberation. But rather than represent self-sufficient decision procedures that can be used in moral deliberation to generate verdicts about our moral duties, they serve to bring out a conflict in the deliberations of the agents featured in the examples – a conflict between considerations of self-love and moral considerations – to which the agent is antecedently committed. An agent who is considering suicide, making a lying promise, letting her talents rust, or being indifferent to others, may well be acting prudently or at least within the bounds of self-love, but Kant’s examples bring out the distinctive force of moral considerations by getting the agent to focus on the kind of impartial perspective characteristic of thinking in a morally engaged manner. To the individual who is contemplating one of these actions (or refrainings), thinking about how one’s maxim would stand as a universal law of nature serves to reveal her recognition of (if not commitment to) moral considerations. In cases where our actions (or refrainings) cannot be “justified in our own impartially rendered judgment, [this way of thinking] still shows that we really acknowledge the validity of the categorical imperative […]” (GMS, 424). To consider significance of this kind of thinking. What I have been doing, then, can be understood as presenting what I take to be the philosophical significance of the central concepts implicated in Kant’s CC and CW tests.

The Categorical Imperative and Universalizability

193

whether one’s maxim can be willed as universal law typically operates against the background of a rich set of moral assumptions. In the case of suicide, the appeal to a teleological law seems to involve an appeal to moral considerations about the (morally) proper use of a motive. In the case of lying promises, the moral acceptability of the institution of promising is taken for granted. And in the talents and beneficence examples, appeals to “rational” maxims – maxims that any non-fi nite agent qua rational would necessarily adopt (and thus maxims that any finite rational agent ought to adopt) – involve background moral assumptions. Without these assumptions, the universality arguments do not get off the ground, and with them they help reveal the kind of rationalizing duplicity that is typically involved in morally wrong action. It is in revealing this kind of duplicity that Kant’s universalizability arguments function as ad hominem arguments. So, even if the CC and CW tests do not represent self-sufficient tests of moral rightness, they are useful in helping to confirm the distinct authority of moral reasons and, in the context of moral reasoning, help reveal a kind of duplicity in an agent’s attempt to justify immoral action. VI. Concluding Remarks I have argued that Kant’s universal law of nature formulation, the typic of the more general formula of universal law, is not plausibly understood as expressing a self-sufficient moral decision procedure; rather, its chief philosophical significance lies in representing a formal metaethical constraint (really, a set of related constraints) on moral reasons. Because this constraint puts limits on what can count as a moral reason, it is stronger than what is represented by resultance, supervenience, and logical universalizability, as these “logical” constraints are currently understood. This generic formal idea of moral reasons as universally valid reasons is systematically spelled out (“unfolded” as it were) in the Groundwork through the various formulations of the categorical imperative: from the universal law formulations that express this formal constraint, to the humanity formulation that identifies the content of fundamental moral reasons, to the formula of autonomy which expresses the ultimate source of normative authority of the moral reasons.

194

Mark Timmons

Appendix: Sorting Out Interpretations of Kant’s Universalizability Tests49 There are two issues, partly (but not completely) orthogonal to one another that figure in standard ways of classifying interpretations of Kant’s CC test. 50 First, as mentioned above in section II, there is the question of what sorts of ancillary information may legitimately play a role in the CC test. Austere interpretations restrict such information to necessary truths knowable a priori, while opulent interpretations allow a posteriori claims to enter in. Second, there is the question of the precise locus of the alleged contradiction that is revealed through testing non-universalizable maxims. Here, so far as I can tell, there are two main interpretative options. One of them attempts to locate the contradiction either “within” the maxim qua universal law – attempting to make sense of Kant’s claims that certain maxims, when universalized are self-contradictory – or within a “system of nature” one of whose laws is the maxim qua universal law. Within this camp we find a variety of different views on what ancillary information may be used in generating such contradictions, and hence we find both austere and opulent versions of this species of interpretation. According to another interpretive strain of thought, the contradiction is not in the very conception of a law or a system of nature but rather involves a contradiction among one’s intentions involved in the process of raising one’s maxim to universal law. Here, then, is a brief overview of the various interpretations, including footnote references to each interpretation’s defenders. A. Inconsistency in Law Interpretations of the CC test 1. Strict conceptual interpretation On this interpretation the only auxiliary information allowed in the test is represented by analytic truths that express the meanings of those terms featured in the maxim. For example, some interpreters have claimed that given the very concept of slavery, in order for this practice to exist, there must be slaves and slave owners (who by definition cannot themselves be slaves). Thus, the very idea of a law according to which everyone holds slaves is itself logically inconceivable. In 49

50

Special thanks here to Josh Glasgow. I’ve benefited from his unpublished handout on the various interpretations of the CC test. There is far less controversy, so far as I can tell over how the CW test is supposed to work. But maybe this is owing to comparative lack of attention.

The Categorical Imperative and Universalizability

195

restricting auxiliary information to a narrow set of conceptual truths, this interpretation is necessarily austere. 51 2. Relaxed conceptual interpretation Some interpreters have proposed that in addition to the narrow set of analytic truths allowed by the strict conceptual interpretation, we may allow analytic truths about finite rational agency. Permitting this expansion presumably allows the CC test to generate a greater number of contradictions covering a wider range of alleged violations of perfect duty than does the stricter version. Again, this counts as an austere interpretation. 52 3. Teleological law interpretation Kant’s Groundwork suicide example (GMS, 421 f.) appeals to the alleged purpose of self-love and thus the idea of teleology. Some interpreters have attempted to make the appeal to teleological laws central for understanding Kant’s CC test. The idea is that certain maxims, if raised to the status of a teleological law would be inconsistent with other teleological laws. Here, we seem to get an “intra-world” contradiction, so to speak, rather than a brute inconsistency in the very conception of one’s maxim raised to universal law. Whether this interpretation counts as austere or opulent depends on how one understands the modal and epistemic status of teleological claims. 53 4. Causal law interpretation On this interpretation of the CC test one is to raise one’s maxim to the status of a casual law of nature and consider it together with other causal laws that happen to govern the actual world. The idea is that some maxims are such that in trying to conceive of them as causal laws of nature operating in a world like the actual world, a contradiction emerges: given actual causal laws (plus the “normal and predictable results” of the operation of such laws) the attempt to conjoin to those laws the universalized maxim leads to contradiction. This interpretation is clearly opulent in allowing auxiliary information knowable only a posteriori.54 51

52 53 54

Cf. Galvin (1991). Similarly, Höffe (2002, 149) argues that “experience is not required to know that a false promise harbors two conceptually internally and mutually incompatible aims”. Cf. Swoyer (1991) and perhaps Pogge (1989). Cf. Paton (1953) and Beck (1960). Cf. Harrison (1957); Dietrichson (1964); Timmons (1984); Rawls (1989); Herman (1993).

196

Mark Timmons

B. Inconsistency in Intention Interpretation of the CC test Often called the “practical contradiction” interpretation of the CC test, the idea here is that some maxims involve intended ends and actions (as means) such that were one to will that everyone act on one’s maxim, the result would be that no one (including the agent) would be able to achieve the intended end by performing the action (that is, a token of the action type) in question. Thus, in willing one’s maxim as a universal law one would in effect be willing the frustration of the end in question – at least the frustration of the ability to achieve the end by the intended means. More precisely, the contradiction emerges as follows. (1) First, take one’s maxim of obtaining or achieving end E by Aing (in C), and then (2) will, as per the test, that one’s maxim become universal law. In so willing the universality of one’s maxim, (3) one also wills the normal and predictable results, including the result that it not be the case that one obtain or achieve E by Aing (in C). Thus, (4) the intention expressed by one’s maxim (to E by Aing in C) is flatly inconsistent with what one is committed to intending in willing one’s maxim as universal law: not (to E by Aing in C). Note that this interpretation makes the CC test similar to the CW test, but does not collapse the former into the latter. The CW test requires appealing to “rational maxims” – maxims that it would be irrational to fail to have – in generating contradictions in the will. Most writers who favor this interpretation seem to be committed to an opulent set of auxiliary premises, since drawing out the implications of willing one’s maxim as universal law seems to appeal to the same considerations involved in the causal law interpretation – namely, the normal and predictable results of the operation of the hypothetical law in a world otherwise like the action world – information that can only be known empirically. Might one appeal only to a priori knowable auxiliary information in developing this interpretation? Perhaps. But then if a priori auxiliary information is all that is needed to generate contradictions, why not just opt for either the strict or relaxed version of the conceptual contradiction interpretation?55 55

Cf. Ewing (1933); M. G. Singer (1961); Gram (1967); Potter (1975); O’Neill (1975); Korsgaard (1989). – This paper was presented at the Beijing International Symposium on Kant’s Moral Philosophy in Contemporary Perspectives, May 2004, Peking University, Beijing, China. It was also presented at Friedrich-Wilhelms Universität, July, 2004, Bonn, Germany at which the papers in this volume were presented and discussed. I wish to thank audiences at both conferences for their valuable input. I would especially like to thank Christoph Horn and Dieter Schönecker for organizing the Bonn conference and editing this volume. Special thanks to Corinna Mieth for her kind help with conference arrangments. I am indebted to Josh Glasgow for his comments and suggestions on an earlier draft of this paper.

The Categorical Imperative and Universalizability

197

Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–) (abbreviated as ‘AA’). The Critique of Pure Reason will be cited according to the A/B pagination from the first and second editions. All textual references are to The Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press, 1992–). GMS Grundlegung zur Metaphysik der Sitten, AA, IV KpV Kritik der praktischen Vernunft, AA, V KrV Kritik der reinen Vernunft, AA III, IV MdSR Metaphysik der Sitten, Rechtslehre, AA, VI MdST Metaphysik der Sitten, Tugendlehre, AA,VI

Other works Aune, Bruce (1979): Kant’s Theory of Morals, Princeton, NJ: Princeton University Press. Baron, Marcia. W. / Pettit, Philip / Slote, Michael. (1997): Three Methods of Ethics: A Debate, Oxford: Blackwell. Bayles, Richard E. (1971): “Act Utilitarianism: Account of Right-Making Characteristics or Decsion-Making Procedure?” in: American Philosophical Quarterly, 8, 257–65. Beck, Lewis. W. (1960): A Commentary on Kant’s Critique of Practical Reason, Chicago: Chicago University Press. Brink, David O. (1989): Moral Realism and the Foundations of Ethics, Cambridge: Cambridge University Press. Donagan, A. (1977): The Theory of Morality, Chicago: Chicago University Press. Ewing, Alfred C. (1933): “The Paradoxes of Kant’s Ethics,” in: Philosophy, XXI, 40–56. Galvin, Richard (1991): “Ethical Formalism: The Contradiction in Conception Test,” in: History of Philosophy Quarterly, 8, 387–408. Glasgow, Joshua (2004): “Expanding the Limits of Universalization: Kant’s Duties and Kantian Moral Deliberation,” in: Canadian Journal of Philosophy, 33, 23–47. Hare, Richard M. (1963): Freedom and Reason, Oxford: Oxford University Press. Harrison, Jonathan (1957): “Kant’s Examples of the First Formulation of the CI,” in: Philosophical Quarterly, 7, 50–62. Harrison, Jonathan (1958): “The Categorical Imperative,” in: Philosophical Quarterly, 8, 360–64. Hegel, Georg F. W. (1821): Elements of the Philosophy of Right, Nisbet, H. B. (trans.), (1991), Cambridge: Cambridge University Press. Herman, Barbara (1989): “Murder and Mayhem,” in: The Monist, 92, 411–449.

198

Mark Timmons

Herman, Barbara (1993): “Moral Deliberation and the Derivation of Duties,” in: Herman, B., The Practice of Moral Judgment, Cambridge: Harvard University Press. Höffe, Otfried (2002): “The Prohibition against False Promising,” in: Höffe, O., Categorical Principles of Law, Migotti, M. (trans.), University Park, PA: Pennsylvania University Press. (Originally published as Katogorische Rechtsprinzipien, Frankfurt a. M.: Suhrkamp Verlag, 1990.) Kerstein, Samuel J. (2002): Kant’s Search for the Supreme Principle of Morality, Cambridge: Cambridge University Press. Korsgaard, Christine M. (1985): “Kant’s Formula of Universal Law,” in: Pacific Philosophical Quarterly, 66, 24–47. Reprinted in Korsgaard, C. (1996): Creating the Kingdom of Ends, Cambridge: Cambridge University Press, and in Guyer, P. (ed.), (1998): Kant’s Groundwork of the Metaphysics of Morals: Critical Essays, Lanham, MD: Rowman & Littlefield. McNair, Ted (2000): “Universal Necessity and Contradictions in Conception,” in: Kant-Studien, 91, 25–43. Mill, John S. (1863): Utilitarianism. London. Mulholland, Leslie (1978): “Kant: On Willing Maxims to Become Universal Law,” in: Dialogue, 17, 92–105. Narveson, Jan (1985): “The How and Why of Universalizability,” in: Potter, N./Timmons, M. (eds.): Morality and Universality: Essays on Ethical Universalizability, Dordrecht: Reidel. O’Neill (Nell), Onora (1975): Acting on Principle, New York: Columbia University Press. O’Neill, Onora (1985): “Consistency in Action.” in: Potter, N./Timmons, M. (eds.): Morality and Universality: Essays on Ethical Universalizability, Dordrecht: Reidel. Reprinted in: Guyer, P. (ed.), (1998): Kant’s Groundwork of the Metaphysics of Morals: Critical Essays, Lanham, MD: Rowman & Littlefield. O’Neill, Onora (1996): Towards Justice and Virtue, Cambridge: Cambridge University Press. Paton, Henry J. (1953): The Moral Law, London: Hutchinson & Co. Pogge, Thomas V. (1989): “The Categorical Imperative,” in Höffe, O, (ed.): Grundlegung der Metaphysik der Sitten: Ein Kooperativer Komentar, Frankfurt a. M.: Vittorio Klostermann. Reprinted in Guyer, P. (ed.), (1998): Kant’s Groundwork of the Metaphysics of Morals: Critical Essays, Lanham, MD: Rowman & Littlefield. Potter, Nelson (1975): “How to Apply the Categorical Imperative,” in: Philosophia, 5, 395–416. Rawls, John (1989): “Themes in Kant’s Moral Philosophy,” in E. Förster (ed.): Kant’s Transcendental Deductions, Stanford: Stanford University Press. Ross, Sir William D. (1954): Kant’s Ethical Theory, Oxford: Oxford University Press. Scanlon, Thomas M. (1998): What We Owe to Each Other, Cambridge: Harvard University Press. Sidgwick, Henry (1907): The Methods of Ethics, 7th ed, Indianapolis, IN: Hackett Publishing Co., 1981. Singer, Marcus G. (1961): Generalization in Ethics, New York: Alfred A. Knopf.

The Categorical Imperative and Universalizability

199

Smart, J. J. C. / Williams, Bernard (1973): Utilitarianism: For and Against, Cambridge: Cambridge University Press. Swoyer, C. (1983): “Kantian Derivations,” in: Canadian Journal of Philosophy, 13, 409–431 Timmons, Mark (1984): “Contradictions and the Categorical Imperative,” in: Archiv für Geschichte der Philosophie, 66, 294–312. Timmons, Mark (1997): “Decision Procedures, Moral Criteria, and the Problem of Relevant Descriptions in Kant’s Ethics,” in: Jahrbuch für Recht und Ethik (Annual Review of Law and Ethics), 5, 289–417. Timmons, Mark (2002a): “Motive and Rightness in Kant’s Ethics,” in: Timmons, M. (ed.): Kant’s Metaphysics of Morals: Interpretative Essays, Oxford: Oxford University Press. Timmons, Mark (2002b): Moral Theory: An Introduction, Lanham, MD: Rowman & Littlefield. Wood, Allen W. (1999): Kant’s Ethical Thought, Cambridge: Cambridge University Press.

Samuel J. Kerstein

Deriving the Formula of Humanity (GMS, 427–437) In Groundwork II, Kant tries to establish that if there is a supreme principle of morality, then it is (or is equivalent to) the Formula of Humanity. He offers a “derivation” of the principle: “So act that you use humanity, whether in your own person or in the person of any other, always at the same time as an end, never merely as a means” (GMS, 429, emphasis omitted). In broad outline, Kant’s derivation (GMS, 427 ff.) is not hard to discern. It takes shape against the background of his fundamental tenet that the supreme principle of morality would have to be a categorical imperative, that is, a principle binding on all of us no matter what our particular inclinations might be. First, Kant contends that if there is a supreme principle of morality (and thus a categorical imperative), then there is an objective end: something that is unconditionally valuable. A categorical imperative requires an unconditionally valuable “ground.” Second, Kant claims that this unconditionally good thing would have to be humanity. In his view, therefore, if there is a supreme principle of morality, then humanity is unconditionally good. For Kant “humanity” does not refer to the class of human beings, but rather to a set of capacities. In the Metaphysics of Morals, Kant tells us that “the capacity to set oneself an end – any end whatsoever – is what characterizes humanity (as distinguished from animality)” (MdST, 392). So at the very least, if a being has humanity, then it has the capacity to set ends. Kant, it seems, uses “humanity” interchangeably with “rational nature” (see, for example, GMS, 439). In doing so he suggests that having humanity involves having certain rational capacities. Among them are the capacity to act on maxims and hypothetical imperatives, as well the capacity to act autonomously, that is, (roughly) to conform to self-given moral imperatives purely out of respect for these im-

Deriving the Formula of Humanity

201

peratives.1 According to the third main claim in the derivation, if humanity is unconditionally good, then we must always treat it not merely as a means but also as an end. Therefore, if there is a supreme principle of morality, then we ought to do just what the Formula of Humanity says. So the supreme principle of morality, assuming there is one, must be this formula, or at least something equivalent to it. This paper focuses on the derivation’s first two steps, especially the second. Assuming that a categorical imperative would require there to be something unconditionally good, why must this unconditionally good thing be humanity, rather than something else? As Part II below points out, the argument Kant appears to invoke in response to this question, namely an argument by elimination (GMS, 428), is very disappointing. But Christine Korsgaard (1996, 106–132) and, later, Allen Wood (1999, 124–132) have claimed that this argument does not represent Kant’s best effort at a derivation. Kant, they contend, actually undertakes a “regressive” argument. I concentrate on Wood’s more recent account of the argument. According to a crucial part of Wood’s account, Kant claims that whenever an agent sets an end, he must ascribe objective goodness to it.2 But in ascribing objective goodness to his ends, Kant contends, the agent commits himself (rationally speaking) to holding his rational nature (that is, his humanity) itself to be unconditionally good. So, in sum, since a rational agent exercises his rational nature in setting ends, he must conclude that it is unconditionally valuable. Part III questions whether Kant actually makes this argument. Contrary to Wood, I do not, for example, believe that, according to Kant’s Groundwork position, whenever an agent sets an end, he is rationally compelled to ascribe objective goodness to it. I do not deny that it was open to Kant to appeal to the regressive argument or that 1

2

Here I am following Thomas Hill, Jr. (1992, 38–41). Allen Wood presents a slightly different account of what Kant means by humanity (1999, 118 ff.). As Allen Wood has pointed out to me, it is not obvious that Korsgaard attributes this claim to Kant. But I take as some indication that she does her statement that “in the argument for the Formula of Humanity, as I understand it, Kant uses the premise that when we act we take ourselves to be acting reasonably and so we suppose that our end is, in his sense, objectively good” (1996, 116). In any case, it is worth noting that Korsgaard’s account of the regressive argument incorporates a robust notion of an objectively good end. On her account, Kant tries to show that if an agent takes himself to have objectively good ends, then he is committed to the unconditional goodness of humanity. But in order for an end to be objectively good in Kant’s sense, she claims, it must be the object of rational choice, provide “reasons for action that apply to every rational being” (1996, 115) and be fully justified. All three criteria are apparent in 1996, 114–116.

202

Samuel J. Kerstein

the regressive argument is philosophically interesting. It is just not, I think, an argument that Kant himself unfurls. Part IV tries to bolster Kant’s argument by elimination with the help of material that he presents later in Groundwork II. Kant points out that rational nature is that which can possess a good will (GMS, 437). Since it is, he suggests, rational nature alone can be the unconditionally good “ground” of a categorical imperative. For all other candidates are such that if they had this status, the good will would be “subordinate” to them. But, claims Kant, “such a will cannot without contradiction be subordinated to any other object” (GMS, 437). Suitably interpreted, this reasoning strengthens Kant’s argument by elimination, I contend. The argument turns out to be stronger than an isolated reading of GMS, 427–429, makes it appear. Of course, even if this claim is correct, Kant’s argument is open to challenge. Part V focuses briefly on one objection to it. Before exploring Kant’s defense of the view that humanity must be the unconditionally good ground of any categorical imperative, we need to examine why he thinks a categorical imperative requires such a ground in the first place. Why, if there is a categorical imperative, must there be anything unconditionally good? I. In his initial step in the derivation of the Formula of Humanity, Kant claims that if there is a supreme principle of morality (and thus a categorical imperative), then there is something of absolute (i. e., unconditional) worth. In something of absolute worth alone “would lie the ground of a possible categorical imperative,” and “if all worth were conditional and therefore contingent, then no supreme practical principle for reason could be found anywhere” (GMS, 428). But why could not a principle be unconditionally binding on us if nothing was unconditionally good? According to Kant, an agent sets himself to do something, that is, determines his will, on the basis of his idea that doing this thing will enable him to secure some end. In Kant’s view, all acting has an end (KpV, 34). Kant distinguishes between subjective and objective ends. Objective ends, if there are any, would hold for all rational beings. The idea of securing them would make available to all rational beings a sufficient ground (motive) for acting. But subjective ends do not give all rational beings grounds for securing them. These ends are such

Deriving the Formula of Humanity

203

that their “mere relation to a specially constituted capacity of desire on the part of the subject gives them their worth” (GMS, 428). Suppose a particular object is a subjective end. If an agent does not value this object, either in itself or as a means to something else, then it has no worth to him. And if the object has no worth to him, intimates Kant, then he does not have a ground to secure it. For him, it is not an end. Apparently, Kant has the following view: an agent has a sufficient ground to secure an object only if he values it – or at least is rationally compelled to value it. In the latter case, the agent is presumably able, through rational reflection, to come to value the object, thereby gaining a sufficient ground to secure it. Against the background of this view, we can reconstruct the basis of Kant’s claim that if there is a categorical imperative, then there must be an objective end – something absolutely valuable. A categorical imperative would be necessarily binding on all rational agents. But a principle could not be necessarily binding on all rational agents unless each of them necessarily had a sufficient ground (motive) at his/her disposal for obeying it. Take us, human rational agents. To say that a principle is binding on us is to say that we ought to (i. e., have an obligation to) conform to it. Kant, of course, holds that if an agent ought to do something, then she must be able to do it (e. g., KpV, 125; 159). But if an agent did not have a sufficient ground available to her for conforming to a rule, then she might not be able to conform to it. Thus, if not all rational agents necessarily have a ground for obeying a principle, then it cannot be a categorical imperative. As we noted, to have a ground for doing something an agent must, according to Kant, hold (or be rationally compelled to hold) the action or its effects to be valuable. Therefore, Kant seems to conclude, if there is a categorical imperative, then there must be something that everyone holds (or must hold) to be valuable: an objective end. There must be something that everyone, in every context, is rationally committed to valuing: something that is absolutely valuable or, equivalently, unconditionally good. 3 II. In the first step of his argument, Kant tries to show that if there is a categorical imperative, then there must be some object (or objects) that all rational agents must hold to be unconditionally good. In the sec3

I am not convinced that this argument is successful. For detailed criticism of it, see Kerstein (2002, 47–54).

204

Samuel J. Kerstein

ond step, Kant claims that this something is humanity. Kant appears to pack his defense of this claim into the paragraph which begins: “Now I say that the human being and in general every rational being exists as an end in itself” (GMS, 428). For at this paragraph’s end he concludes that if there is a “supreme practical principle for reason,” then rational beings are of absolute worth (unconditionally valuable). On its face, the defense seems inadequate. It appears to amount to an argument by elimination. Kant quickly dismisses three candidates for unconditional goodness: objects of inclinations, inclinations themselves, and beings “the existence of which rests not on our will but on nature,” (GMS, 428) such as animals. Then he embraces humanity as alone suited to be an end in itself, that is, the unconditionally valuable ground of a categorical imperative. His dismissal of rival candidates for unconditional goodness seems precipitate. For example, Kant says simply that “all objects of the inclinations have only a conditional worth; for if there were not inclinations and the needs based on them, their object would be without worth” (GMS, 428). But this statement requires defense that Kant does not here seem to offer. For an opponent might reasonably object that, regarding some such objects, Kant has things backwards: it is not the case that they are valuable (just) because we desire them, but rather the case that we desire them at least partly because they are (in themselves) valuable.4 Even if the remarks Kant here makes did eliminate these candidates for absolute goodness, the question would arise as to whether he is entitled to conclude that it is humanity he is looking for. After all, might not Kant have overlooked some other candidate for absolute goodness? What about the state of affairs of all rational agents being happy? How does Kant dismiss this possibility? This candidate is not itself an inclination. It need not be considered an object of an inclination. (We can easily envisage a world in which no one desires everyone – including his enemies – to be happy.) And everyone’s happiness is not obviously something the existence of which would rest on nature rather than on our will. Kant’s argument that only humanity could be the absolutely valuable thing needed if there is to be a categorical imperative (supreme principle of morality) appears to suffer from serious shortcomings. His arguments against the other candidates he considers seem far too quick, and he does not explicitly consider at least one candidate that 4

For a defense of this sort of objection, see Gaut (1997, 176 f.).

Deriving the Formula of Humanity

205

springs quickly to mind, namely everyone’s being happy. On the face of it, Kant’s argument by elimination does not seem very promising. III. Although he acknowledges shortcomings in Kant’s argument by elimination, Wood does not despair of Kant’s establishing that humanity must be the ground of any categorical imperative (1999, 124). For, according to Wood (1999, 125), Kant offers a positive argument for this claim in the very next paragraph. A key step in Kant’s positive argument, suggests Wood (1999, 127), is an attempt to establish the following: each rational agent must (is rationally compelled to) agree that in setting himself ends, he commits himself to the view that his rational nature is unconditionally valuable. Wood suggests that Kant’s defense of this key step involves a “regress on conditions.” Inseparable from an agent’s setting ends for himself is his holding that they have a certain value. But an agent can, rationally speaking, place this value on his ends only on condition that he hold his own rational nature to be unconditionally valuable. When we look more closely, Wood suggests, we find that this key step in Kant’s argument rests on three claims. The first has to do with the kind of value the agent is committed to placing on the ends he sets. Whenever an agent sets himself an end, goes the claim, he must ascribe “objective” goodness to it. The second claim is that the agent must hold the source of the objective goodness of his ends to be his setting of them. What makes his ends good, he is rationally compelled to affirm, is that they are the objects of the exercise of a certain capacity, namely his rational choice. 5 Finally, if the source of the goodness of objectively good ends is his rational choice (rational nature), then the agent must hold his rational nature itself to be unconditionally good.6 So, in sum, since a rational agent exercises his rational nature 5

6

That Wood attributes the fi rst two claims to Kant is manifest in the following passage: “Thus Kant’s argument is based on the idea that to set an end is to attribute objective goodness to it and that we can regard this goodness as originating only in the fact that we have set those ends according to reason. The thought is . . . that rational choice of ends is the act through which objective goodness enters the world” (1999, 129). See also (1999, 127 and 129 f.). Wood makes clear that he attributes this claim to Kant when he states that “Kant does infer from the premise that rational nature is the source or ground of the objective goodness of all ends to the conclusion that rational nature itself is the underivative objective good, an end in itself” (1999, 130). See also (1999, 127 and 129).

206

Samuel J. Kerstein

in setting ends, he must conclude that it is unconditionally valuable. Even if we embrace this conclusion, we might wonder how to get from it to the further point that a rational agent must conclude that everyone’s rational nature is unconditionally good.7 We also might wonder how the exclusivity of rational nature as a possible ground for a categorical imperative gets established. That humanity is unconditionally valuable does not in itself seem to entail that nothing else is. But these are not issues for us to pursue here. Also absent from our agenda is evaluating the philosophical plausibility of the argument that Wood (and Korsgaard before him) attribute to Kant. I find the argument (in both Wood and Korsgaard’s versions) to be engaging and important, but as I have argued in detail elsewhere, I think it has serious shortcomings.8 Here I concentrate on the question of whether there are adequate textual grounds for attributing the argument to Kant in the first place. The regressive argument may be Kantian in spirit. But I doubt whether he actually employs it in the Groundwork. To begin, let us consider the first main claim that, according to Wood, Kant makes in defending the view that humanity is unconditionally valuable. For Kant, it seems, all cases of an agent’s setting an end are cases of his exercising his capacity of rational choice (humanity). Kant implies that having humanity involves having the capacity to set ends.9 The first claim is that whenever an agent sets an end, he must (is rationally compelled to) ascribe objective goodness to it. He must, says Wood, “regard it as something of value universally for all rational beings” (1999, 127). Let us assume that in Kant’s view, when an agent sets an end, he must take it to be good in some sense. Why should we agree that in the Groundwork Kant embraces the view that he must take it to be objectively good? According to Wood, at GMS, 412–414, Kant “maintains that goodness, whether moral or nonmoral, is that which reason represents as practically necessary, and hence as an object of volition for all rational beings” (1999, 127). Something is an object of volition for all rational beings just in case all such beings must regard it as valuable, Wood implies. So, he suggests, at GMS, 412–414, Kant commits himself to the view that if an agent sets himself an end, whether she views the end to be good morally or non-morally, she is rationally compelled to hold it to be objectively good. 7 8 9

Korsgaard (1996, 123) and Wood (1999, 131) each discuss this point. Cf. Kerstein (2002, 46–72). Cf. MdST, 392.

Deriving the Formula of Humanity

207

I do not see where in the specified pages Kant commits himself to this view.10 Among the central claims Kant there makes are the following: All actions prescribed by an imperative (command of reason) are in some sense practically necessary, that is, good. Actions prescribed by a hypothetical imperative are good as means to some end; actions prescribed by a categorical imperative are good in themselves.11 But it does not follow from these claims that if an agent sets herself an end, she is rationally compelled to hold it to be valuable for all rational beings. Suppose an agent sets herself an end that, she realizes, is an action commanded by a hypothetical imperative, but not a categorical imperative. It is a necessary means to some further goal she has. She would then have to agree that everyone is rationally compelled to hold the following: if someone, including the agent herself, has this further goal, then, as a means to this goal, it is good for that person to realize the end. But the agent would not be rationally compelled to hold that this end is something of value for each and every rational agent. Let me illustrate this rather abstract point with an example. Suppose Sally sets herself the end of traveling to New York. This action, she realizes, is commanded by a hypothetical imperative, but not by a categorical imperative. Traveling to New York is a necessary means for visiting the Empire State Building, which is something she has had an inclination to do for a long time. (The hypothetical imperative at issue would be something like this: “If you want to visit the Empire State Building, then you ought to travel to New York.”) If we grant the claims set out above, Sally would have to agree that everyone, including her, must hold the following: if a person has the end of visiting the Empire State Building, then it is good as a means to that end for that person to go to New York. But Sally would not be rationally compelled to agree that the end of traveling to New York is something that all rational agents must value. In GMS, 412–414, Wood suggests, Kant embraces the view that if an agent sets himself an end he is rationally compelled to hold it to be objectively good. Some central claims Kant makes there do not entail this view. If there is a passage in GMS, 412–414, that seems to license this conclusion, it is perhaps the following: “Practical good, 10

11

Thomas Hill, Jr. argues (2002, 244–274) that Kant does not embrace the view that in setting an end an agent always commits himself to the end’s objective goodness. I focus more than does Hill on the issue of whether specific texts in the Groundwork provide evidence that Kant adopts the view in question. In this context Kant is using “imperative” as a success term; he is conceiving of imperatives as valid.

208

Samuel J. Kerstein

however, is that which determines the will by means of representations of reason, hence not from subjective causes, but objectively, i. e., from grounds that are valid for every rational being as such” (GMS, 413). If we identify “practical good” with any end an agent sets, then we seem to have evidence that Kant maintained the view in question. For then Kant appears to be suggesting that what determines an agent to pursue any end of his is a ground that is “valid for every rational being as such.” And if what determines an agent to pursue any end of his is a ground valid for every rational being as such, then it seems that an agent is rationally compelled to view any end of his as valuable to every rational being. But it is questionable whether in the passage cited from GMS, 413, Kant identifies “practical good” with any (and every) end an agent sets. There is no straightforward indication that he does. He uses the term end [Zweck] not once in the paragraph from which the passage is taken. Moreover, there is another interpretation of the passage available. By “practical good,” I believe, Kant means something like “objective (practical) principles.” In the two sentences preceding the passage, Kant is describing imperatives, which are objective practical principles expressed with the help of an “ought,” since they are addressed to agents, such as human beings, who might fail to act in accordance with them. In the passage itself, Kant seems to be maintaining the following: if an agent does something at least partly on the grounds that an imperative commands it, then the agent is acting from grounds that are valid for every rational being as such. And that one maintains this does not require him to embrace the quite distinct view that in setting an end, an agent must take it to be objectively good. Suppose, for example, that Sally visits New York partly because the hypothetical imperative “If you want to visit the Empire State Building, then you ought to travel to New York,” commands this action. (I describe her as acting “partly” from the imperative because a necessary ingredient of her acting from it at all is that she have the desire to visit the Empire State Building.12) Sally is acting from grounds that are valid for every rational being as such. For assuming that the principle in quotations is really an imperative, it is the expression of an objective principle–one valid for every rational being as such. The imperative is valid for us (human beings) in that if any of us have the end of visiting the Empire State Building, then, other things being equal, we ought to travel to 12

At GMS, 428, Kant calls relative ends, such as Sally’s end to visit the Empire State Building, grounds of hypothetical imperatives.

Deriving the Formula of Humanity

209

New York. Nothing here implies that in setting herself the end of visiting New York, Sally must embrace the view that the end is valuable to all rational agents. The interpretation on the table of GMS, 413, harmonizes well with what follows. In the very next sentence Kant suggests that practical good is “distinguished from the agreeable, as that which has influence on the will only by means of sensation from merely subjective causes, those which are valid only for the senses of this or that one, and not as a principle of reason, which is valid for everyone” (GMS, 413). An objective practical principle is, of course, a “a principle of reason, which is valid for everyone.” It seems a stretch to claim that in Groundwork, 412–414, Kant endorses the view that in setting an end, an agent must hold it to be objectively good. Actually, near the beginning of the derivation of the Formula of Humanity there is evidence that he does not endorse this view. Kant says: “The ends that a rational being proposes as effects of its action at its discretion (material ends) are all only relative; for only their relation to a particular kind of capacity of desire of the subject gives them their worth, which therefore can provide no necessary principles valid universally for all rational beings and hence valid for every volition, i. e., practical laws.” (GMS, 427 f.)

Let us suppose that a particular agent, John, correctly believes these claims to be true. John has a “material end,” namely that of his attending the Super Bowl. According to Kant, it is only the relation of his end to the particular kind of capacity of desire he has that gives it its worth. In other words, it is only his setting it as an object to attend the next Super Bowl that gives his attending it any value. If he believes this, then why would he believe that everyone is rationally compelled to view his end as valuable? John would, it seems, be free to reason as follows: “Those who do not make it their object that I attend the next Super Bowl (and there are surely many) might be rationally compelled to acknowledge that my attending it is valuable to me. Yet it is not the case that they must take my attending it to be valuable to them.” Moreover, Kant says that material ends “provide no necessary principles valid universally for all rational beings.” But if material ends were, rationally speaking, objectively valuable, then they would provide at least one principle valid universally for all rational beings, namely the principle: “A material end is never to be treated as an object of no value whatsoever.” This principle would have some practical bite. It would, for example, forbid an agent from failing to

210

Samuel J. Kerstein

promote John’s end in cases in which doing so would not interfere with any of the agent’s morally required action or pursuit of his own material ends. In the passage immediately preceding his derivation of the Formula of Humanity, not only does Kant fail to embrace the view that in setting an end, an agent must hold it to be objectively good, he implies that he rejects it. The text of the Groundwork fails, I think, to imply that Kant embraces the first claim in the regressive argument for humanity as the ground of a categorical imperative. Kant does not, I contend, adopt the view that whenever an agent sets himself an end he must ascribe objective goodness to it.13 If this contention is correct, then it is unsurprising that we find little evidence that he makes the second or third claims. Where does Kant actually make the argument that an agent must hold that the source of the objective goodness of his ends is his exercising his capacity of rational choice (rational nature) in setting them (claim 2)? Where does he actually make the argument that if an agent must hold this, then he must also hold his rational nature itself to be unconditionally good (claim 3)? According to Wood, Kant makes these arguments in order to defend his contention at GMS, 429, that each of us necessarily represents his existence as that of an end in itself, that is, as that of something that is unconditionally valuable. I believe that it was open to Kant to make the arguments. But I do not find compelling textual evidence that he actually does. Indeed, Kant’s claim that each of us necessarily represents his existence as that of an end in itself, as well as the surrounding text, admits of an interpretation that does not invoke the regressive argument at all.14 In the paragraph at the end of which Kant sets out the Formula of Humanity, he says: “The ground of [the moral principle] is: Rational nature exists as end in itself: The human being necessarily represents his own existence in this way; thus to that extent it is a subjective principle of human actions. But every other rational being also represents his existence in this way consequent on just the same rational ground that also holds for me;* thus it is at the same time an objective principle from which, as a supreme practical ground, it must be possible to derive all laws of the will.” (GMS, 429) 13

14

Both Korsgaard (1996, 115) and Wood (1999, 128 f.) appeal to passages outside of the Groundwork for support of their view that, according to Kant, if an agent sets an end, he holds it to be objectively valuable. Both invoke, for example, Kant’s discussion of goodness and well-being in The Critique of Practical Reason (KpV, 58–61). For an interpretation of this discussion according to which Kant is not there committing himself to the view in question, see Hill (2002, 262 ff.). This contention challenges Korsgaard as well as Wood. See Korsgaard (1996, 122 f.).

Deriving the Formula of Humanity

211

The note indicated in the second sentence reads as follows: “Here I put forward this proposition as a postulate. The grounds for it will be found in the last Section” (GMS, 429). Kant’s claim that the human being necessarily represents his own existence as an end in itself might be a claim about human nature. Kant might be suggesting that it is natural for human beings to think of themselves as superior to non-rational beings, including other animals.15 But perhaps Kant wants to leave it open that non-human rational agents might not, as a rule, act against the background of such a view. That would explain why Kant calls doing so a “subjective” principle of acting. In any case, the plausibility of the claim that human beings do indeed represent themselves as superior to non-rational beings would not seem to depend on an argument regarding the ultimate source of value in the world. But what are we then to make of the rest of the passage, in particular of Kant’s remark that “every other rational being also represents his existence in this way consequent on just the same rational ground that also holds for me”? This remark seems to me to contain two distinct claims. The first is that we, human rational agents, have a rational ground, that is, are rationally compelled, to represent all rational agents as unconditionally valuable. The second is that non-human rational agents are also rationally compelled to view all rational agents as having such a value. But to this point in the Groundwork, Kant has proven neither of these claims. By his own lights, he has shown merely that if one assumes there to be a categorical imperative (supreme principle of morality), then one must hold humanity to be an end in itself. He has not shown (nor tried to show) that anyone who does not make this assumption is rationally compelled to hold humanity to be an end in itself. In other words, Kant has not proven the validity of the categorical imperative (Formula of Humanity). That is a task that he puts off until Groundwork III. There he tries to show that all agents (rational beings with a will), not merely human agents, must take themselves to be free and thus to be bound by the moral law.16 It is therefore not at all surprising that, in a note, Kant calls the remark in question a “postulate” and suggests that he will defend it in Groundwork III. Our reflections thus far have not left us with a good impression of Kant’s Groundwork derivation of the Formula of Humanity. He seems to offer a very weak argument by elimination in defense of his claim that the unconditionally good ground of a categorical impera15 16

See Anth, 127. See GMS, 447 f.

212

Samuel J. Kerstein

tive would have to be humanity. If I am correct, he simply does not present the more philosophically interesting regressive argument attributed to him by Wood. However, things might be brighter than they appear. Well after he unfolds his argument by elimination (but still in Groundwork II), Kant sets out considerations that supplement and reinforce it. In particular he appeals to the notion of a good will in arguing that if there is a supreme principle of morality, then humanity, but not anything else, is an end in itself. Reflection on this appeal will help us to see that Kant’s argument by elimination is stronger and more interesting than it might seem to be. IV. At GMS, 437, Kant begins a summary of a main line of argument he has developed up to this point in the Groundwork. “Now we can end,” he announces, “at the place from which we set out at the beginning, namely with the concept of an unconditionally good will.” In the next paragraph, he says the following: “Rational nature is distinguished from the rest of nature by this, that it sets itself an end. This end would be the matter of every good will [. . .] Now, this end can be nothing other than the subject of all possible ends itself, because this subject is also the subject of a possible absolutely good will; for, such a will cannot without contradiction be subordinated to any other object.” (GMS, 437)

At GMS, 428, let us recall, Kant seems to dismiss precipitately three candidates for the unconditionally good ground of a categorical imperative: inclinations themselves, the objects of inclinations, and beings “the existence of which rests not on our will but on nature.” In the passage just cited, Kant suggests an argument that might bolster his elimination of these candidates. The argument takes shape against the background of Kant’s view of the value that “common rational moral cognition” attributes to a good will. In our ordinary moral thinking, claims Kant, we hold that such a will is unconditionally good. An impartial rational spectator would, we believe, judge it to be good in every possible context in which it exists.17 Moreover, we believe that a good will is preeminently good.18 That we believe this presumably implies that, in our view, nothing that is devoid of a 17 18

See GMS, 393. See GMS, 394 and 401.

Deriving the Formula of Humanity

213

good will is as good as something that has such a will. The argument unfolds as follows. Whatever constitutes the ground of a categorical imperative must be compatible with the value we attribute to a good will (i. e., unconditional and preeminent goodness). But let us suppose that any of the rivals to humanity were the ground of a categorical imperative. We would have to acknowledge that a good will would (in some circumstances) be “subordinated” to the rival and that a good will would thereby fail to have the value we attribute to it. We would land in contradiction. Therefore, we have license to dismiss any of the rivals as possible grounds for a categorical imperative. As is perhaps already evident, the material that is to supplement Kant’s argument by elimination is material Kant develops before his presents the argument. I will not appeal to any positions Kant develops after he does so—for example his view that humanity has dignity, that is, unconditional and incomparable value. Filling in the supplemented argument’s details requires elaboration of the concept of a good will as well as of the notion of one thing’s being subordinated to another. The latter task is in a sense easier, since Kant simply does not tell us precisely what this notion amounts to. But I think it reasonable to assume that he would endorse the following claim: If x is subordinated to y, then x is less valuable than y and we thereby have sufficient grounds to use x in whatever way is necessary in order to maintain y. So for example, if plants are subordinated to rational beings, then they are less valuable than rational beings, and we have sufficient reason to harvest the former in order to preserve the latter. Regarding a good will, it suffices for our purposes to take note of two ways in which Kant seems to employ the notion as it applies to us, agents who can indulge their inclinations and thereby act contrary to the moral law. According to the first usage, a good will is a particular sort of willing or, what for him amounts to the same thing, of acting. Kant writes of “the unqualified [uneingeschränkten] worth of actions” (GMS, 411), presumably of actions done from duty, which he has previously stated to have “unconditional and moral worth” (GMS, 400). Since, according to Kant, the good will is good without qualification [ohne Einschränkung], it appears that sometimes “good will” refers to a certain kind of action, that is, action done from duty.19 According to 19

This reading of “good will” would have to be broadened to accommodate Kant’s view that perfectly rational beings such as God cannot act from duty. To them the “ought” of duty does not apply, since their willing is necessarily in accord with the law. See GMS, 414. We might attribute to Kant the view that these beings have a

214

Samuel J. Kerstein

a second usage of “good will,” it refers not to a particular kind of action an agent might perform but rather to a kind of character she might have. An agent has a good will on this usage just in case she is committed to doing what duty requires, not just in this or that particular action, but overall. Presumably if an agent has this commitment, then she will sometimes act from duty. (For example, she will invoke duty as her incentive to do what is morally required in cases in which she is tempted by her inclinations to act contrary to what morality demands.) Kant intimates that having a good will amounts to having a certain kind of character in the first paragraph of Groundwork I. Right after suggesting that a good will is good without qualification, he tells us that certain qualities of temperament, for example, courage or resolution, “are undoubtedly good and desirable for many purposes, but they can also be extremely evil and harmful if the will which is to make use of these gifts of nature, and whose distinctive constitution is therefore called character, is not good” (GMS, 393).20 Sometimes Kant employs what we might, following Karl Ameriks (1989, 54–59), call the “whole character” conception of a good will. I believe Kant to be employing the whole character conception of a good will in the passage we are discussing (GMS, 437). In any case, it is that conception of a good will that might help him to bolster his argument by elimination. Let us now return to our (GMS, 437–based) supplement to Kant’s argument by elimination. Suppose that beings “the existence of which rests not on our will but on nature,” say, species of wild animals, were the ground of a categorical imperative, namely one commanding us never to eradicate currently existing species of such animals. We would be committed to the view that such beings were not only unconditionally but also preeminently valuable. Otherwise, we might sometimes lack sufficient grounds to abide by the principle not to eradicate them. And if we might lack such grounds, we cannot take the principle to be a categorical imperative. We would presumably lack sufficient grounds to abide by the principle in question when doing so would conflict with maintaining some-

20

good will (engage in unconditionally good willing) just in case they act “for the sake of the law.” Presumably such beings are capable of doing this. And Kant does not seem averse to the idea that acting from duty is a species of acting for the sake of the law. Later Kant is discussing a man who is by temperament cold and indifferent to others, but who, from duty, acts beneficently. “It is just then,” says Kant, “that the worth of character comes out, which is moral and incomparably the highest” (GMS, 398 f.). This passage suggests that “good will” refers not merely to a particular kind of action, but to a kind of character that can be expressed in action.

Deriving the Formula of Humanity

215

thing that was also unconditionally valuable, but more valuable than a wild animal species. That something might, for example, be a person. Say that the only way to save someone’s life was to kill the last two remaining representatives of a bird species in order to make a medicine for him. If we believe correctly that persons are more valuable than species of wild animals, then, in these circumstances, we do not have sufficient reason to act in accordance with a principle commanding us never to eradicate the latter. It seems plausible to grant the possibility that more than one kind of thing is unconditionally good as well as that one unconditionally good thing is better than another. After all, what basis do we have for denying it? Yet if we grant this possibility, we find that preeminent as well as unconditional goodness is necessary to ground a categorical imperative. Getting back to our example, if species of wild animals can serve as the ground of a categorical imperative, then they must be unconditionally and preeminently good. Their being preeminently good implies that nothing that is not a species of wild animals is as good as something that is. But if species of wild animals have this value, then a good will’s value is subordinate to their value. They are worth more than it is, and so we have sufficient reason to abandon it in order to preserve them. As Kant suggests it would at GMS, 437, this subordination of a good will to species of wild animals results in a contradiction. For we have been assuming that a good will is unconditionally and preeminently good. But a good will cannot both be less valuable than species of wild animals and, as the notion of preeminence implies here, more valuable than they are. It would, I believe, be unproblematic to illustrate via an analogous chain of reasoning how Kant might eliminate other candidates for the ground of a categorical imperative, including inclinations and the objects of inclinations. Even a candidate Kant does not seem to consider in his argument by elimination, namely everyone’s being happy, is vulnerable to this reasoning. Suppose that everyone’s being happy were the ground of a categorical imperative, namely one commanding us to maximize the aggregate welfare. We would then be committed to the view that everyone’s being happy was both unconditionally and preeminently valuable. If it were merely unconditionally valuable, we might sometimes lack sufficient grounds (i. e., rational justification and thus motivation) to abide by the principle to maximize aggregate welfare. For example, we might lack sufficient grounds for abiding by this principle when doing so would prevent us from securing something more valuable

216

Samuel J. Kerstein

such as, perhaps, the existence of persons. And if we might be without adequate reason to conform to a principle, then we cannot take it to be a categorical imperative. But if everyone’s being happy has both unconditional and preeminent value, then a good will’s value is obviously subordinate to it. (For Kant, of course, not everyone who has a good will is happy and not everyone who is happy has a good will. See, for example, GMS, 442.) The object constituted by everyone’s being happy is worth more than a good will, and so we have sufficient reason to destroy the latter in order to promote the former. But this conclusion forces us into a contradiction. We are assuming along with Kant that a good will is unconditionally and preeminently good. Yet a good will cannot both be less valuable than everyone’s being happy and, as the notion of its preeminence implies here, more valuable than it is. Of course, the supplemented argument by elimination just brought to bear against a version of utilitarianism is far from invulnerable to attack. It is obviously open to philosophers to disagree with Kant’s view that, according to ordinary moral thinking, a good will is unconditionally and preeminently good. Moreover, it remains to be seen how the supplemented argument avoids dismissing Kant’s own candidate for the ground of a categorical imperative. How is it that humanity itself does not get eliminated from contention? Initially, it seems that it would. After all, if humanity is to be the ground of a categorical imperative, then we must hold it to be not only unconditionally good, but also preeminently good. Otherwise we leave open the possibility of there being other, better, unconditionally good things–things that we would, rationally speaking, have to preserve even at the cost of disobeying the principle grounded by the value of humanity. But if we hold humanity to be preeminently as well as unconditionally good, then we must, it seems, conclude that a good will is subordinate to it, contradicting our assumption that the good will is preeminently good. Fortunately, this argument ignores the special relationship that obtains between humanity and a good will. Every being who has a good will necessarily has humanity. For having a good will entails having capacities constitutive of humanity, for example, the capacity to act on categorical imperatives. An agent cannot have an overall commitment to do what duty requires according to Kant unless she has the capacity to act on principles that specify what it requires, namely moral rules. Much as it would be impossible to be an excellent classical pianist without being a pianist, so it would be impossible to have a good will without having humanity. With this point in view, we are able to see

Deriving the Formula of Humanity

217

that the supplemented argument by elimination does not, in the way the objection alleges, throw the baby (humanity) out with the bath water (rival candidates for the ground of a categorical imperative). If we hold humanity to be preeminently good, then we believe that nothing that is devoid of humanity is as good as something that has it. A good will, of course, is not devoid of humanity. So holding humanity to be preeminently good does not force us to conclude that a good will is subordinate to it, that is, has less value than it does. Here it makes sense to object that, even if this last point is correct, humanity would get eliminated as a candidate for the ground of a categorical imperative. For the idea that humanity is unconditionally and preeminently good contradicts the notion that a good will is unconditionally and preeminently good. Two different things cannot both be preeminently valuable. Given the notion of preeminence that (at least on my reading) Kant employs, it turns out that two different things can both be preeminently good. Humanity and a good will are indeed two different things; for one can have humanity without having a good will. Not every being possessed of rational nature (humanity) is committed overall to doing what moral principle requires. Some of us do not have excellent character. But there is no contradiction in maintaining that humanity is preeminently good and all the while holding that a good will is. According to the former claim, nothing that is devoid of humanity is as good as something that has it. According to the latter, nothing that is devoid of a good will is as good as something that has it. The latter claim implies that humanity without a good will is not as good as humanity with a good will. But there is no contradiction in maintaining both that nothing devoid of humanity is as good as something with it and, at the same time, that humanity devoid of a good will is not as good as humanity with one. In much the same way, it is not self-contradictory (though it may be false) to maintain both that no frying pan that is not copper is as good as any frying pan that is copper and, at the same time, to hold that every copper frying pan that is 3 mm thick is better than any copper frying pan that is less thick. V. Even when supplemented as I suggest, Kant’s argument by elimination appears to be closer to Kant’s intentions concerning the Groundwork derivation of the Formula of Humanity than the regressive argument

218

Samuel J. Kerstein

attributed to him by Korsgaard and Wood. For not only does Kant explicitly embrace the reasoning that bolsters the argument (namely at GMS, 437), but this reasoning itself relies only on concepts that Kant develops before the derivation is complete. Of course, the supplemented argument from elimination and thus the derivation as a whole is, at best, only as convincing as Kant’s view that a good will is unconditionally and preeminently valuable. This view is surely in need of defense.21 But let me conclude by considering a further challenge to the supplemented argument. If the reasoning in Part IV is sound, the argument avoids two pitfalls. It does not descend into self-contradiction; for there is nothing inconsistent in maintaining as the argument does that both humanity and a good will are unconditionally and preeminently good. Moreover, the argument does not imply that a good will is subordinated to humanity. So it avoids thereby eliminating humanity as a possible ground for a categorical imperative. However, a serious question remains: What justification does Kant have for holding that it is beings with rational nature who constitute the ground of a categorical imperative, rather than maintaining that it is merely beings with a good will who do so? What, for example, permits Kant to reject the view that it is not all of us, but rather only those of us with a good will, who never ought to be treated merely as means? Consider a principle that Kant does not discuss, namely what I call the Formula of the Good Will: “So act that you treat a good will, whether in your own person or in the person of any other, always at the same time as an end, never merely as a means.” It makes sense to stipulate that, according to this formula, having a good will necessarily involves having an overall commitment to doing what Kant’s Formula of Universal Law requires. For Kant’s derivation of the Formula of Universal Law takes place well before the argument by elimination. The supplemented argument by elimination seems to leave open the possibility that a good will serves as a ground for the Formula of the Good Will. So why must we say that it is humanity rather than a good will that remains viable as a ground for a categorical imperative?22 21

22

But the supplemented argument does not depend on a further controversial view Kant holds (GMS, 393), namely that only a good will is good without qualification. In reply, one might focus attention on the fact that Kant’s derivation of the Formula of Humanity occurs after his derivation of the Formula of Universal Law. One might then argue that Kant implicitly holds the following: whatever is to serve as the ground of a categorical imperative must serve as the ground of a principle equivalent to the Formula of Universal Law, that is, a principle that requires or permits just those actions that are required or permitted by the Formula of Universal Law.

Deriving the Formula of Humanity

219

Our response to this objection depends on our interpretation of a thorny issue concerning a good will. According to Kant, an agent can never be sure that he (or anyone else) has a good will.23 A commitment to morality over the fulfillment of inclinations is constitutive of a good will; but introspection cannot reveal the presence of any such commitment, he implies. Even if one has never violated his duty, one might nevertheless lack a good will. For one’s conformity to duty might result from a fortunate harmony in his case between the dictates of prudence and the requirements of morality, rather than from a genuine commitment to the moral law.24 But can we ever be sure that an agent lacks a good will? The answer to this question is not obvious. What if, for example, Sue has manifested, and continues to this moment to exhibit, a pattern of breaking promises, apparently just for her own financial gain? In Kant’s view the Formula of Universal Law forbids acting on maxims of false promising for financial gain (GMS, 422). Moreover, Kant assumes that it is a simple matter for all of us to determine what morality demands. 25 So can we rest assured that Sue (or anyone else) does not currently have a good will? This is not the occasion to try to answer this question. But much depends on which answer turns out to be correct. Suppose that we can be sure that an individual, say Sue, now lacks a good will. In this case, the Formula of Humanity and the Formula of the Good Will would not, for practical purposes, be equivalent. The former but not the latter would forbid treating Sue merely as a means.26 Humanity

23 24 25

26

But the good will would serve as the ground of a principle, namely the Formula of the Good Will, that is not equivalent to the Formula of Universal Law. Therefore, concludes the argument, the good will is not the ground of a categorical imperative. Perhaps Kant would embrace this argument, but I do not find it philosophically promising. Kant, of course, suggests that the Formula of Humanity is equivalent to the Formula of Universal Law (GMS, 436). But I know of no substantive interpretation of the two principles according to which they turn out to be equivalent. So I fear that this sort of argument would likely not only eliminate the good will as a possible ground of a categorical imperative, but humanity as well. See GMS, 407 f. See RGV, 36 f. See, for example, GMS, 404, and KpV, 36. This view seems to me to be unpersuasive, as I explain in Kerstein (2002, 119–129). At this point, it might be tempting to make the following argument. “Even if we are sure that Sue does not have a good will now, it is open to her to develop one in the future. On the basis of this potential she has, the Formula of the Good Will would forbid us from treating her merely as a means.” But this argument misses the mark. For the Formula of the Good Will does not forbid us from treating beings with the potential for having a good will merely as means, but rather from treating beings with a good will merely as means. And here we are assuming that we know Sue not to have a good will.

220

Samuel J. Kerstein

and a good will would constitute competing grounds for categorical imperatives, and the supplemented argument by elimination would be incomplete. Actually, if we can know that an individual lacks a good will, then the argument would be worse than incomplete. It would discount humanity as a possible ground for a categorical imperative, namely for the Formula of Humanity itself. It is easy to imagine situations in which this principle and the Formula of the Good Will would deliver incompatible moral verdicts. As we just noted, the latter would imply that treating Sue merely as a means is morally permissible while the former would entail that it is not. But in such a case an agent would find herself without sufficient grounds to abide by the Formula of Humanity. Why should she privilege the dictates of the Formula of Humanity over those of the Formula of the Good Will? It is true that the Formula of Humanity is (supposedly) grounded on something unconditionally and preeminently valuable, but the Formula of the Good Will is grounded on something that has this value and is (supposedly) even better.27 Of course, if there is a possible situation in which an agent would lack sufficient grounds to abide by a principle, then this principle is not a viable candidate for a categorical imperative. But suppose that we cannot be sure that any particular person currently lacks a good will. In order to abide by the Formula of the Good Will, it seems, we would then treat as ends in themselves all beings with humanity. For all such beings might have a good will. Humanity, says Kant, is “the subject of a possible absolutely good will” (GMS, 437). In order to insure our compliance with the Formula of the Good Will, we would have, therefore, to treat everyone with humanity as if he/she had a good will. So in reply to the objection, one might say that the Formula of the Good Will is for practical purposes equivalent to the Formula of Humanity. Since it is, a good will and humanity are not really competing grounds for a categorical imperative. If Kant’s considered view is that we can never know whether an individual lacks a good will, his supplemented argument by elimination avoids a significant obstacle.28 In any case, this argument deserves further attention, I believe. It is not only philosophically interesting, but also well-grounded in Kant’s text.

27 28

See IV above. I would like to thank the other contributors to this volume for very helpful discussion of this paper.

Deriving the Formula of Humanity

221

Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: W. deGruyter, 1902-) (abbreviated as ‘AA’). All English translations are based on The Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press, 1992–). Anth Anthropologie in pragmatischer Hinsicht, AA VII GMS Grundlegung zur Metaphysik der Sitten, AA, IV KpV Kritik der praktischen Vernunft, AA, V MdST Metaphysik der Sitten, Tugendlehre, AA,VI RGV Die Religion innerhalb der Grenzen der blossen Vernunft, AA, VI

Other works Ameriks, Karl (1989): “Kant on the Good Will,” in: Höffe, Otfried (Ed.): Grundlegung zur Metaphysik der Sitten; Ein kooperativer Kommentar, Frankfurt a. M., 45–65. Gaut, Berys (1997): “The Structure of Practical Reason,” in: Cullity, Garrett/ Gaut, Berys (Ed.): Ethics and Practical Reason, Oxford, 161–188. Hill, Thomas, Jr. (1992): Dignity and Practical Reason, Ithaca: Cornell University Press. Hill, Thomas, Jr. (2002): Human Welfare and Moral Worth, Oxford; Oxford University Press. Kerstein, Samuel J. (2002): Kant’s Search for the Supreme Principle of Morality, Cambridge: Cambridge University Press. Korsgaard, Christine M. (1996): Creating the Kingdom of Ends, Cambridge: Cambridge University Press. Wood, Allen W. (1999): Kant’s Ethical Thought, Cambridge: Cambridge University Press.

Groundwork III

Klaus Steigleder

The Analytic Relationship of Freedom and Morality (GMS III, 1)1 The first part of the Third Section of the Groundwork belongs to that genre of philosophical texts that are relatively simple to comprehend in their general orientation yet considerably more difficult to comprehend in their details. I will proceed in the following manner. As the first part introduces the third and final section of the Groundwork, I will begin 1. by situating this Third Section in the context of the entire work, clarifying its task and briefly surveying the functions of the first part. 2. I will then attempt to explicate the fi rst part of Groundwork III in greater detail. Given its brevity, I will quote the text in its entirety in the course of my reflections. 3. In conclusion, I will discuss a divergent interpretation of the text. 1. On the Position and Task of Groundwork III The task of the Third Section of the Groundwork is the justification of the supreme principle of morality. According to Kant, this is not to determine the content of the moral principle, or to ascertain how it reads. Rather, the task is to demonstrate that the moral principle is binding for us and that, correspondingly, we have a strict obligation to comply with it. In the First and particularly in the Second Section of the Groundwork, Kant has shown that we can develop the idea of an unconditional ought and that this idea leads us to a nonarbitrary and clearly defined content of a moral principle. If there are binding moral norms of action for us, then these conform to a specific moral principle. Initially, this moral principle is nothing other than an idea, i. e., a thought-entity, and it could, as Kant writes, constitute a mere “phan1

Translated from the German by Jo Ann Van Vliet, Ph. D., Tübingen, and checked by Alexander Cotter (Stonehill College).

226

Klaus Steigleder

tom” (GMS, 407). It could even be the case that no unconditional ought exists, and that actually all norms, which we generally address as moral norms, are nothing other than norms of prudence. These would then constitute not an unconditional ought but only a conditional ought. In order to better comprehend the task of Groundwork III, i. e. to demonstrate that morality is not a “phantom” but that we can and must obey the moral principle, we should recall why we are capable of formulating a nonarbitrary idea of an unconditional ought and what this idea entails. We are able to conceive the idea of an unconditional ought because we are conscious of the reality of conditional demands, of conditional norms of action. In Groundwork II Kant develops his theory of norms governing action (Kant himself does not speak of norms of action but of imperatives) as a theory of practical judgments from the perspective of an acting subject. Conditional norms of action are so structured that, for the agent, the practical necessity exists of pursuing an end or undertaking an action in order to pursue an end, if or because he or she has a certain actual end. “Because I want to lose weight, it is necessary for me to want to eat less chocolate candy” or “Because I want to avoid destitution in old age, it is necessary for me to want to put money aside for retirement” are two examples of such practical judgments.2 The intention to do or achieve something necessitates the intention to do something else which is a means of attaining the original end. In this context, “wanting” means volition, the actual resoluteness, the real intention of pursuing an end or undertaking an action in order to pursue an end.3 Since we do not automatically want to do what it is necessary for us to want to do on the basis of our actual ends (we would like to continue to enjoy chocolates with abandon, although we want to lose weight, and we would prefer to spend rather than to save money, even though we do want to avoid destitution in old age), the practical necessity of wanting to do something, arising out of our actual ends, confronts us as a demand, a necessitation, an ought. Kant differentiates two types of conditional norms of action, namely, technical norms and prudential norms. Technical norms are dependent on ends that we can have; prudential norms, in contrast, are dependent on an end that all those beings that are under the influence of feelings of pleasure and displeasure inevitably have, to wit, the end of one’s personal well-being, or, as Kant says, one’s own happiness. 2

3

Kant speaks of propositions (“Sätze”) rather than judgments. Propositions are, however, in his opinion, assertoric judgments. See Ent, 193,33; Log, § 30, note 3, 109,11–22. Kant designates such real intentions to act as maxims.

The Analytic Relationship of Freedom and Morality

227

Given the indeterminate nature of this end, we continually interpret and reinterpret it through more specific ends (e. g., living in a marriage or a long-term relationship with another human being, starting a family, engaging in a fulfilling profession, achieving affluence). Conditional norms of action have, thus, the structure that, due to an end, a practical necessity determined by this end arises for the agent, a necessity which appears as a demand, which becomes an ought.4 Ultimately, all conditional norms of action derive from the comprehensive end of “one’s own happiness,” since this underlies the choice of ends on which conditional norms of action are dependent. The comprehensive end of personal happiness arises for us, however, through an interaction between our sensibility and our reason: because we are shaped by feelings of pleasure and displeasure, we wish to achieve and sustain pleasant states of being (if possible), and to end and avoid unpleasant states of being as well (if possible). With the assistance of our reason we conceive the goal of a maximum of conveniences and a minimum of inconveniences. 5 Being knowledgeable of conditional demands, we can attain a conception of an unconditional moral ought by negating the essential characteristics of a conditional ought. An unconditional ought must be founded on an absolute practical necessity, i. e., the necessity of wanting to do something, independent of a goal given to the practical faculty of reason or resulting from its interaction with unreasoned elements or aspects. This leads to the concept of an unconditionally necessary end. This must originate in the practical faculty of reason. Since the concept of an unconditionally necessary end must be free from all empirical elements stemming from our sensibility, it must originate in a direct purposiveness of pure practical reason. Since the criteria for this purposiveness of reason cannot be external to reason; and since, correspondingly, the end, too, cannot be external to reason; and since, finally, an unconditionally necessary end cannot be something that must be realized (rather, it must be thought of as something that necessarily exists as an end) – this very end must necessarily lie in the faculty of pure practical reason. A being that possesses the faculty of pure practical reason must understand himself or herself, on the basis of this faculty, as an unconditionally necessary end and, accord4

5

For a more extensive discussion see Steigleder (2002, 23–58, Chapter 2: “Hypothetische Imperative”). “Happiness is the satisfaction of all our inclinations (extensive, with regard to their manifoldness, as well as intensive, with regard to degree, and also protensive, with regard to duration)” (KrV, B 834 /A 806).

228

Klaus Steigleder

ingly, attribute an absolute worth to himself or herself. Such a being must assume that, on the basis of his or her faculty of pure practical reason, he or she is capable of determining ends on his or her own accord which take into account the unconditionally necessary end as which he or she necessarily exists, because he or she, on the basis of his or her faculty of reason, is able to establish ends on his or her own accord. Correspondingly, there is an unconditionally practical necessity for him or her to take into account the unconditionally necessary end as which he or she, and every other being equally capable of action, exists himself or herself in all action. This is, so to speak, the quintessence of the aspects and considerations through which Kant develops the idea of an unconditional ought in the Second Section of the Groundwork. Kant does not, however, speak explicitly of “pure practical reason” in this context.6 This is in striking contrast to the beginning of the Second Section in which Kant critically valorizes the conclusions of the First Section over the philosophy of his day. Kant reserves the Third Section of the Groundwork for this quintessence, which he presents to a considerable extent, as we will see, in the first part of Groundwork III under discussion here. To be sure, it is important to observe that the requisite considerations have already been provided by Kant within the context of his explication of the idea of a categorical imperative. There, Kant preceded his analysis of the norms of action, the hypothetical and categorical imperatives, for example, with the description of his intention: “[…] we must follow and present distinctly the practical faculty of reason, from its general rules of determination to the point where the concept of duty arises from it” (GMS, 412,22–25). Earlier, Kant had emphasized, in conjunction with the conclusions of the First Section, “[…] that all moral concepts have their seat and origin completely a priori in reason” (GMS, 411,8 f.; my emphasis). And in the course of the actual explication of the idea of an unconditional ought he writes: Here, however, it is a question of objective practical laws and hence of the relation of a will to itself insofar as it determines itself only by reason; for then everything that has reference to the empirical falls away of itself, since if reason entirely by itself determines conduct (and the possibility 6

That is, from GMS, 412,26 and particularly from GMS, 420,24 on. In both explicit references to “pure practical reason” following GMS, 412,26 (namely, GMS, 440,25 f. and 445,11–15) the task of a “critique of pure practical” reason is mentioned and it is thereby referred to GMS III.

The Analytic Relationship of Freedom and Morality

229

of this is just what we want now to investigate), it must necessarily do so a priori (GMS, 427,13–18).

The further determination of this self-direction of a will exclusively determined by reason leads to the concept of a necessary end valid for all rational beings: Now, what serves the will as the objective ground of its self-determination is an end, and this, if it is given by reason alone, must hold equally for all rational beings (GMS, 427,21–24; only end (“Zweck”) is emphasized in the original). But suppose there were something the existence of which in itself has an absolute worth, something which as an end in itself could be a ground of determinate laws; then in it, and in it alone, would lie the ground of a possible categorical imperative, that is, of a practical law (GMS, 428,3–6; original emphasis).

Kant develops this even further in the idea of a will which is “giving universal law” and subject only to the law which it gives to itself, summarizing that “the autonomy of the will” is “the supreme principle of morality.” But that the above principle of autonomy is the sole principle of morals can well be shown by mere analysis of the concepts of morality (GMS, 440,28–30).

As this quotation evinces, the idea of an unconditional ought (or of morality), which we can conceive as a mere inverse of a conditional ought, leads to a moral principle having a content. Initially, such a moral principle is nothing more than something that we can and must think if we think an unconditional ought. Accordingly, this moral principle is at first only a thought-entity, a mere idea. To be sure, it is not immaterial that the content of the moral principle can be derived through “mere analysis of the concepts.” However, this does not yet establish the validity of the moral principle. It could even be the case that for us all norms of action must be reducible to norms of prudence. This would then be the case if all determining motivations for our action were to arise from our sensible structure of impulse. Our reason would then not be capable of determining original ends but only capable of administering the interests of inclinations.7 Consequently, it would be unfounded to dictate that we take into account the unconditionally necessary end in our action, 7

Cf. GMS, 406,14–25; 441,17 f.

230

Klaus Steigleder

which we must constitute for ourselves and which every other human being capable of action must constitute for us. Neither can there be a necessary end or an absolute worth for us nor could we take such an absolute worth as such into account, should it exist. Although, as Kant underscores, the content of the moral principle can be ascertained “by mere analysis of the concepts of morality,” that which is ascertained through this analysis has a different status than the hypothetical imperatives, which, according to Kant, are analytic practical judgments. Kant calls hypothetical imperatives analytic judgments (propositions), because in their case wanting to achieve an end mandates wanting to pursue another end or an action as means. Wanting to pursue another end or wanting to undertake an action is, in a certain sense, inherently wanted in wanting to achieve the original end.8 Crucial for our discussion is the fact that the volition, upon which the necessity of further volition depends, is empirically accessible. We are conscious of the fact that we have or can have ends which hypothetical imperatives are dependent on. This is why, according to Kant, it is not difficult to answer the question of the “possibility” of hypothetical imperatives, i. e., the question of the origin of the necessitating power of hypothetical imperatives. They are respectively dependent on our actual ends, the reality of which is therefore unquestionable. This is completely different in the case of the categorical imperative or the unconditional ought. Neither is this dependent on something, the reality of which is established, nor is it established from the beginning that there is such a thing as an unconditional ought. An unconditional ought, as introduced in Groundwork II, is, as indicated, a mere idea, and, correspondingly, everything ascertained through its analysis possesses the same status: it, too, is, at least for the moment, a mere idea. Thus, the question of the possibility of an unconditional ought also has another significance than the fundamental question concerning theoretical judgments in the Critique of Pure Reason “How are synthetic judgments possible?,” the reality of such synthetic judgments a priori in theoretical form being indisputable for Kant. To establish that the moral principle is not only a thought-entity but that it possesses validity for us is the task of the Third Section of the Groundwork. Kant can also define this task as the proof of the moral principle as a synthetic practical judgment a priori. What actually constitutes a synthetic practical judgment a priori? 8

For a more comprehensive discussion, see Steigleder (2002, 24–39).

The Analytic Relationship of Freedom and Morality

231

According to Kant, three kinds of practical judgments are conceivable for us as finite beings capable of action: 1. Maxims. These have the two following basic forms: a) “I want to achieve end E through my action,” and b) “I wish to carry out action A9 in order to achieve E.” The reality of maxims is indisputable. 2. Hypothetical Imperatives. These take the following basic forms10: a) “I ought to do A, because I want to achieve E.” b) “I ought (to attempt) to achieve E2 because I want to achieve E1.” That is, hypothetical imperatives demand the adoption of certain maxims because the agent has certain other maxims from which the necessity arises to adopt the maxims being demanded. Since the reality of maxims is indisputable, the reality of hypothetical imperatives is equally indisputable. 3. Categorical imperatives. Their basic form consists of the unconditional demand to do something or not to do something. Since practical judgments are practical judgments precisely because they express the volition of a subject capable of action, we can more fully comprehend the categorical imperative as a practical judgment if we realize that moral oughts articulate a practical necessity for an agent, which materializes as a demand for him or her (a necessitation). The unconditional ought of the categorical imperative thus derives from the judgment that an unconditionally practical necessity exists for the agent: in the absence of a presupposed maxim the acting subject must necessarily want to do something. The truth of a corresponding synthetic practical judgment a priori presupposes for this reason that the agent is capable of determining necessary ends and of complying with necessary ends. Thus the truth of the synthetic practical judgment a priori in question is contingent on our possession of the faculty of pure practical reason, or of freedom in the strict sense. Thus precisely this appears to be the task of the proof of the validity of the moral principle in Groundwork III: to demonstrate that we have the faculty of pure practical reason, or of freedom in the strict sense, that enables us to establish ends, or to account for necessary ends completely independent of sensible impulses while acting. The task of the “critique of pure practical reason,” which is to establish the 9

10

Here action is to be understood in the broad sense of voluntary action respectively inaction directed toward an end. Maxims can also have inaction (omission) as their subject. Hypothetical imperatives can call for the rejection of certain maxims (to not do things that I want to do) as well as the pursuit of certain goals. Here it is not my intention to outline all the forms of hypothetical imperatives but simply to indicate their basic structure or form.

232

Klaus Steigleder

truth of the moral principle as a synthetic practical judgment a priori, does not, as in the Critique of Pure Reason, lie in the determination of the scope of the use of theoretical reason independent of experience but first and foremost in the establishment of the reality of the faculty of pure practical reason. Thus, it is not unforeseen when Kant opens the Third Section of the Groundwork with a part entitled “The Concept of Freedom is the Key to the Explanation of the Autonomy of the Will,” and then in the second part attempts to defend the thesis “Freedom must be Presupposed as a Property of the Will of All Rational Beings.” Nevertheless, the prospects are not promising for Kant’s envisioned proof of the validity of the moral principle in Groundwork III. For the establishment of the reality of the faculty of pure practical reason would appear to be tantamount to the attempt to establish freedom in the strict sense. Yet one conclusion from the Critique of Pure Reason is precisely that freedom in the strict sense or transcendental freedom cannot be proven. Transcendental freedom is conceivable. As it cannot be an object of experience, and as experienceable nature as such is appearance and not a self-existent reality, it can without contradiction be conceived together with the necessity that in experienceable nature “all alterations occur in accordance with the law of the connection of cause and effect” (KrV B 232, see also B 566–586 [A 538–558]). But it is neither possible to prove nor disprove the existence of transcendental freedom. If it were necessary to prove the existence of pure practical reason in the course of the proof of the validity of the moral principle, then, on the basis of the conclusions of the Critique of Pure Reason, it is evident from the very beginning that the validity of the moral principle cannot be proven. For this reason, Kant does not make the slightest effort to prove the existence of freedom in Groundwork III. Rather, he will attempt to demonstrate that even a less ambitious undertaking is fully sufficient. It is not requisite to establish that we are free in the strict sense (that we possess the faculty of pure practical reason); rather it is requisite to establish that we must hold ourselves to be free (that we must assume that we possess the faculty of pure practical reason). The latter can be true, even if the former (that we are free in the strict sense) were false; or if we, as is the case, cannot ascertain its truth or falseness. If it is, however, true that we must assume that we have the faculty of pure practical reason, then we must further assume that we constitute an unconditionally necessary end and that we can and must take this unconditional end into account in all our action, in summary: we must

The Analytic Relationship of Freedom and Morality

233

assume that the moral principle possesses validity for us. Consequently, the moral principle is not a mere thought-entity for us but must be the authoritative guiding principle of our action. This is the line of argumentation that Kant pursues in Groundwork III. It is prefigured in the first part in Kant’s attempt to epitomize his explication of the idea of an unconditional ought in Groundwork II, his emphasis on the centrality of freedom for the evolving argumentation, and his acknowledgment of the difficult nature of the justification of the moral principle. Let us now turn to this first part in detail.

2. The Argumentation of Part 1 in Groundwork III The first part of Groundwork III can be divided into two segments. The first of these, which is comprised of the first two paragraphs, has the task of explicating the thesis in the title “The Concept of Freedom is the Key to the Explanation of the Autonomy of the Will,” and culminates in the statement that “hence a free will and a will under moral laws are one and the same.” (GMS, 447) The second segment, which the third and final paragraph comprises, addresses the task and the problem of the justification of the moral principle as a synthetic practical judgment a priori. Let us begin with the argumentation in the first segment. The first paragraph reads: Will is a kind of causality of living beings insofar as they are rational, and freedom would be that property of such causality that it can be efficient independently of alien causes determining it, just as natural necessity is the property of the causality of all nonrational beings to be determined to activity by the influence of alien causes (GMS, 446,7–12).

This sentence contains in condensed form essential elements from Kant’s theory of action. What Kant calls “will” here, he calls practical reason or free choice (freie Willkür) elsewhere. “Will” is initially to be differentiated in two directions. On the one hand, it is to be distinguished from the unfree faculty of desire (arbitrium brutum) possessed by animals, and on the other, from pure practical reason. Common to both animals and human beings, Kant asserts, is their possession of the faculty of desire (Begehrungsvermögen). Kant understands this as the faculty entailing an inner causation of the behaviour of a being through psychic respectively mental contents (representations). A billiard ball rolls as a result of an external propulsion; an animal begins

234

Klaus Steigleder

to move out of an inner impulse, because it has, for example, seen or perhaps scented potential prey. The capability to perceive or to develop representations and to govern behavior through representations is, for Kant, a characteristic of life.11 Thus, he comprehends life in a narrower sense than we do, being accustomed to consider plants, too, as “living beings.” In distinction to animals, however, the behaviour of which, Kant maintains, is completely controlled by instincts or drives, we, as rational living beings, possess the faculty of free choice or of practical reason, i. e., we have the capacity to distance ourselves from our impulses, to develop interests, and to act on the basis of reasons. We are correspondingly capable – on the basis of deliberations – to choose and pursue ends. Nonetheless, it may be the case that the motivations for the active pursuit of the ends issue from our sensible structure of impulses. Then we would not be capable of determining ends and pursuing them independent of our inclinations. Consequently, the “freedom” of our free choice would be merely a relative freedom. In the Critique of Pure Reason (B 830/A 802), Kant has called this “practical freedom” and emphasized that we can experience such practical freedom ourselves: it is accessible to our experience that we are capable of acting in relation to our sensible impulses. Insofar as free choice is an object of experience we must comprehend our power to act as also being subject to natural necessity. Ultimately we would be “determined to activity by the influence of alien causes.” This is unchanging despite the fact that these determinations include inner representations, a point that Kant will expressly emphasize again in the Critique of Practical Reason: […] it does not matter whether the causality determined in accordance with a natural law is necessary through determining grounds lying within the subject or outside him, or in the first case whether these determining grounds are instinctive or thought by reason, if […] these determining representations have the ground of their existence in time and indeed in the antecedent state, and this in turn in a preceding state, and so forth, these determinations may be internal and they may have psychological instead of mechanical causality, that is, produce actions by means of representations and not by bodily movements; they are always determining 11

“Life is the faculty of a being to act in accordance with laws of the faculty of desire. The faculty of desire is a being’s faculty to be by means of its representations the cause of the reality of the objects of these representations.” (KpV, 9,19–22; original emphasis); see also MdSR, 211,7–9: “The faculty of a being to act in accordance with its representations is called life” (Original emphasis).

The Analytic Relationship of Freedom and Morality

235

grounds of the causality of a being insofar as its existence is determinable in time and therefore under the necessitating conditions of past time, which are thus, when the subject is to act, no longer within his control and which may therefore bring with them […] natural necessity […] (KpV, 96,19–37; the phrase “causality determined in accordance with a natural law” is not italicized in the original).

Yet at the beginning of the first part of Groundwork III Kant immediately focuses on free choice, or the will, under the hypothesis that this free choice possesses the faculty of freedom in the strict sense (“and freedom would be that property of such causality that it can be efficient independently of alien causes determining it”). If we think of free choice as free in this sense, then we must not think of it as pure practical reason, i. e., we must not equate free choice with pure practical reason but must think of it as being endowed with the faculty of pure practical reason – an insight which is significant for the comprehension of the line of argumentation in Groundwork III in general. If free choice were equatable with pure practical reason, a being endowed with such a faculty would necessarily determine itself to activity through pure practical reason (would necessarily do what is morally good). This is clearly not the situation of finite “living beings,” and, in any event, not our situation insofar as we possess the faculty of pure practical reason, because otherwise the moral law would not have to confront us as a demand, an imperative. If our capacity of choice possesses the faculty of pure practical reason, then we are always capable of determining ourselves to activity “independently of alien […] determining causes.” If we thoroughly contemplate the implications of this negative determination of freedom in the strict sense, we realize that we have also been directed toward a positive content. In certain respects this repeats the establishment of an unconditional ought in Groundwork II. This originally consisted in negating the essential properties of a conditional ought. However, it proved to have given grounds for positive determinations. The preceding defi nition of freedom is negative and therefore unfruitful for insight into its essence; but there flows from it a positive concept of freedom, which is so much the richer and more fruitful (GMS, 446,13–15; original emphasis).

The positive concept of freedom is conveyed through the conception of free choice as a causality. If, namely, free choice “can be efficient independently of alien causes determining it,” then the capacity to be

236

Klaus Steigleder

the cause of such effects must be inherent without this cause itself in turn being comprehensible as the effect of a foreign, preceding, and determining cause. Since the concept of causality is not an experiential concept and, consequently, is to be comprehended as entailing a necessary relationship between cause and effect,12 we cannot conceive the relationship between an original (uncaused) cause and its effect as a lawless relationship replacing the naturalistic progression of cause and effect but rather as a lawful relationship of its own kind. A different law supplants natural law, namely, the law of pure practical reason. Since the concept of causality brings with it that of laws in accordance with which, by something that we call a cause, something else, namely an effect, must be posited, so freedom, although it is not a property of the will in accordance with natural laws, is not for that reason lawless but must instead be a causality in accordance with immutable laws but of a special kind; for otherwise a free will would be an absurdity. Natural necessity was a heteronomy of efficient causes, since every effect was possible only in accordance with the law that something else determines the efficient cause to causality; what, then can freedom of the will be other than autonomy, that is, the will’s property of being a law to itself? (GMS, 446,15–447,2)

The will has the capability to establish ends and to actively pursue ends independent of foreign determining influences. Given the fact that such independent establishment and pursuit of ends do not appear out of nowhere, but are brought about by reason, they must also be determined by the characteristics of reason, specifically, universality and necessity. Reason that establishes ends necessarily observes the law, which it is to itself, is autonomy. This thought can be developed further with various accentuations. It can be stressed that reason, which establishes ends, can constitute an unconditionally necessary end and, correspondingly, an absolute worth for itself. Kant himself sets a different accent: But the proposition, the will is in all its actions a law to itself, indicates only the principle, to act on no other maxim than that which can also have as object itself as a universal law. This, however, is precisely the formula of the categorical imperative and is the principle of morality; hence a free will and a will under moral laws are one and the same (GMS, 447,2–7). 12

We must assume (expressed in a somewhat simplified manner here) that in nature every change is the effect of a cause, which necessarily produces this effect. Which cause produces which effect is a matter which we, however, must attempt to ascertain through experience.

The Analytic Relationship of Freedom and Morality

237

Before I briefly address the relationship between the various accentuations, it is crucial to realize that, as emphasized, the will is to be conceived as free choice, which possesses the faculty of pure reason, but which is not simply pure practical reason. Such a will always has the possibility of obeying the law in determining its ends and actions, the law which it is “to itself” on the basis of its faculty of pure practical reason. But it is not simply reason that establishes its own reasonable ends. Correspondingly, the will does not necessarily observe the law, for it is not autonomy but only possesses the power of autonomy. It is a law to itself and it is “under the law,” that is, “subject to the law.” That Kant comprehends the will in Part 1 as free choice (endowed with pure practical reason) and not as pure practical reason is evident, for example, in his assumption that the will is capable of maxims, i. e., of subjective determinations. As subjective determinations of the will, maxims can deviate from precisely those laws of pure practical reason characterized by universality and necessity. Consequently, the moral law takes the form of a categorical demand only to act according to those maxims which correspond to the lawgiving of pure practical reason. This lawgiving would not be a foreign demand but a demand arising from our own faculty of reason. If, however, although we possess the faculty of pure practical reason, we were to act according to maxims contravening the moral law, we could not attribute this to natural necessity, i. e., to alien determining influences. For then we would allow ourselves to be determined by foreign influences, even though we always have the possibility of not observing these influences. We would then have to comprehend ourselves as being free in our action and as being responsible for our action, although it would not be a realization of the law of our freedom. How this occurs or would occur is totally unknown to us. Nonetheless, the distinctions and their related determinations are significant. If we, in the sense explained above, must attribute ourselves with a free will, then we must presuppose that we ourselves and every other being, which, like us, has a free will, constitute an unconditionally necessary end and possess an absolute worth which is to be taken into consideration in all action. This demand does not essentially differ from the principle emphasized by Kant “to act on no other maxim than that which can also have as object itself as a universal law.” There are, of course, differences insofar as the various formulations of the moral principle call for different procedures to examine or to determine the compatibility of a purpose or an intended action with the moral principle. In passing, I would like to note that not all of the

238

Klaus Steigleder

examination procedures are equally suited and that precisely the procedure favoured by Kant in Groundwork II is problematic.13 Here it was only important to comprehend why Kant can assert “hence a free will and a will under moral laws are one and the same.” This conclusion ends the passage which I have designated as the first segment of Part 1 in Groundwork III. Kant summarizes this once again at the beginning of the second segment, i. e., at the beginning of the third paragraph, as follows: If, therefore, freedom of the will is presupposed, morality together with its principle follows from it by mere analysis of its concept (GMS, 447,8– 10).

This thesis of the analytic relationship between the freedom of the will and the moral principle is, as shown above, not surprising in light of the conclusions of Groundwork II. These culminated in a non-arbitrary idea of the unconditional ought. As Kant’s considerations in the first section are supposed to show, this idea is grounded in the idea of a free will proper. We are still proceeding on the level of mere thought. This is why the justification of the moral principle consists, as also shown above, in establishing the moral principle as a synthetic practical judgment a priori. And we have already attempted to comprehend that this is actually tantamount to the task of establishing the reality of freedom, or, as we can now say, of a free will (in the strict sense), although this is not directly possible (in the form of a proof of freedom). Ultimately, this is exactly what Kant says in the final paragraph of Part 1. There he continues with the following remark: But the principle of morality – that an absolutely good will is that whose maxim can always contain itself regarded as a universal law - is nevertheless always a synthetic proposition; for, by analysis of the concept of an absolutely good will that property of its maxim cannot be discovered (GMS, 447,10–14).

This assertion may appear to be rather curious at first glance. Is this not exactly what Kant tried to show in Groundwork I, namely, that the 13

It is important to note in general that the investigation which has the task of developing the fundamental characteristics of the content of the moral principle is not the Groundwork, but The Metaphysics of Morals of 1797. There, Kant establishes that the moral principle justifies a number of orderly sub-principles. Correspondingly, for Kant it is not at all the task to always directly judge our actions and maxims on the basis of the moral principle. See especially Steigleder (2002, 129–274, “Teil B: Moralphilosophie als Rechts- und Tugendlehre”).

The Analytic Relationship of Freedom and Morality

239

idea of an absolutely good will leads to the idea of a moral principle? And are Groundwork II and the first two paragraphs of Groundwork III not an eloquent testimonial to the richness of the idea of the unconditional ought, and the cross-referentiality of its various aspects? The solution to this problem presumably lies in the fact that here Kant does not only focus on the absolutely good will as an idea and, correspondingly, that the determination of the capacity of the maxims of this will does not hypothetically follow. The judgment (proposition) differs on this point from the introductory sentence, which reads: “and freedom would be that property of such causality that it can be efficient independently of alien causes determining it” (GMS, 446,8–10; original emphasis), as well as from the first sentence of the paragraph under discussion, which opens with the stipulation “If, therefore, freedom of the will is presupposed […]” What cannot be derived “by analysis of the concept of an absolutely good will” is the real capacity of a subjective determination of the will to take the moral law as such into account, and therefore the reality of the faculty of pure practical reason and consequently of the moral law. Thus, it is requisite to prove the existence of the moral principle as a synthetic practical judgment a priori. While an analytic (theoretical) judgment in the simplest case of a categorical judgment makes a characteristic (“Merkmal”) explicit in the predicate term which was already conceived in the subject term, a synthetic judgment cannot remain on the internal level between subject term and predicate term.14 For here, namely in the predicate term, a new (not yet conceived) characteristic is ascribed to the subject term, and so the question arises what (which third “entity” in relation to the subject and predicate term) allows the attribution of such a new characteristic to the subject term.15 In the case of synthetic judgments a posteriori it is experience, in the case of synthetic theoretical judgments a priori these are, according to Kant, the forms of intuition, time and space. Kant draws parallels between the problem of synthetic practical judgments a priori and of synthetic theoretical judgments a priori and also inquires about the third cognition which is capable of overcoming the mere internal relationships of conceived determinations of volition and allows the attribution of reality to the determinations of unconditional volition and unconditional ought. A viable possibility for this third cognition is, on the one hand, freedom, as shown above. On the other hand, freedom is 14 15

See KrV, B 10 f./A6 f. See KrV, B 12 f./A9 f.

240

Klaus Steigleder

not a true parallel to the forms of intuition, time and space. For us, freedom is something, the reality of which is neither ascertained nor ascertainable. Such synthetic propositions are possible only in this way: that the two cognitions are bound together by their connection with a third in which they are both to be found. The positive concept of freedom provides this third cognition, which cannot be, as in the case of physical causes, the nature of the sensible world (in the concept of which the concepts of something as cause in relation to something else as effect come together) (GMS, 447,14–20).

Kant, thus, does not speak of freedom as the third cognition being sought but rather purports that the positive concept of freedom provides or creates (“schafft”) this third cognition. This third cognition, as Kant recapitulates with reference to his preceding reflections on the positive concept of freedom, is to be distinguished from the causal relationships constituting experienceable nature (and thus independent of a time structure). Of interest here is Kant’s assertion that the concept of freedom provides or creates this third cognition. This is initially quite surprising since his task seems to be the departure from the level of mere ideas. In actuality, however, as addressed above, Kant’s solution will consist in establishing that the idea of freedom cannot be a mere idea for us but that we must necessarily consider ourselves to be free in the sense of the positive concept of freedom. According to Kant, it will also be crucial to show that we must necessarily comprehend ourselves as rational beings with wills, that we cannot comprehend our sensible nature as our real nature, and that we must attribute ourselves with the faculty of pure practical reason. Consequently, Kant concludes the first section by maintaining: What this third cognition is, to which freedom points us and of which we have an idea a priori, cannot yet be shown here and now; nor can the deduction of the concept of freedom from pure practical reason, and with it the possibility of a categorical imperative as well, as yet be made comprehensible, some further preparation is required (GMS, 447,20–25).

It will eventually become evident that we must regard ourselves and every other being capable of action as an unconditionally necessary end because we cannot consider our practical evaluations and our ends as being limited to interests which issue from feelings of pleasure and displeasure. This affirms what Kant could only present as a postulate in Groundwork II, namely, that every rational being must so conceive his or her existence that he or she exists as an end in himself

The Analytic Relationship of Freedom and Morality

241

or herself.16 Kant had envisioned the proof as a task of Groundwork III. The course of the appropriate line of argumentation is established in its fi rst part. 3. A Critical Discussion of Dieter Schönecker’s Interpretation Dieter Schönecker supports an interpretation of the fi rst part of Groundwork III which differs from mine on essential points.17 He takes the decisive impetus for his interpretation from a consideration of the line of argumentation in Groundwork III. According to Schönecker, this line of argumentation would not be comprehensible if the “thesis of analyticity” (“Analytizitätsthese”) in Part 1 were to connote that, given the presupposition of the freedom of a fi nite (“sensible-rational”) being, the binding nature of the categorical imperative would follow analytically. In this case the argumentation would culminate in the third part of Groundwork III at the very latest, and not, as Kant assumes, in the fourth part. The argumentation would then, in Schönecker’s opinion, present itself as follows: following the “thesis of analyticity” (Part 1) Kant would initially show (in Part 2) that a “rational being, which has a will” must ascribe freedom to himself or herself and, consequently (in accordance with the interpretation of the “thesis of analyticity” which Schönecker rejects), must see himself or herself as a finite being obligated by the categorical imperative. That the categorical imperative is binding for us would then only require the proof that we, as potential agents, must understand ourselves as “rational beings which have a will.” The demonstration of this would be the task of Part 3, signifying that the argumentation would have to be considered completed upon its conclusion. According to Schönecker, Kant himself claims, however, that he only completes the deduction of the binding nature of the categorical imperative in Part 4. An advocate of the “thesis of analyticity” (in the interpretation Schönecker considers flawed) would have to consider this part superfluous. If one does not wish to impute “total confusion” to Kant (Schönecker, 1999, 157), then the “thesis of analyticity” must be understood differently. As Schönecker sees it, Kant does not wish to utilize this to establish an analytic relationship between freedom and the categorical imperative but rather to show that a “purely rational (perfect) being” or a 16 17

See GMS, 429,2–7, 35 f. Schönecker (1999, 147–195).

242

Klaus Steigleder

“sensible-rational being, whose will is, however, considered merely as an intelligible will” necessarily obeys the moral law (Schönecker, 1999, 160).18 Thus Kant, in Schönecker’s interpretation, wishes to emphasize that a free will and a rational or moral will respectively a “will subject to the principle of morality” are interchangeable concepts. This understanding of the “thesis of analyticity” is, in his eyes, “the only reading […] which allows the placement of the text in a coherent interpretive framework” (Schönecker, 1999, 154). This is, of course, a bold assertion. To begin with, neither the limitation of the final step of the argumentation exclusively to Part 3 nor the characterization of the aim of Part 4 (ostensibly the initial realization of deduction in Kant’s understanding) are as unequivocally clear as Schönecker purports. There may conceivably be other readings without having to impute “total confusion” to Kant.19 Moreover, Schönecker’s interpretation of the “thesis of analyticity” is not directly suggested by the text. Schönecker expends considerable effort in showing that another understanding of the introductory sentence than the obvious one is possible: “Living beings” must not necesssarily mean corporeal beings but can mean fully rational beings for Kant, as Schönecker verifies on the basis of lecture transcriptions. Additionally, Schönecker must acknowledge that Kant is “not totally without responsibility” for the (in Schönecker’s opinion) faulty interpretation of the “thesis of analyticity.”20 Kant himself speaks of the “categorical imperative” in conjunction with the “thesis of analyticity,” although it would have been better not to have done so. Finally, Schönecker must assert that Kant presupposes that for a sensible-rational being the proof of his or her possession of freedom in the strict sense would not render the proof of the necessity for action in accordance with this freedom unessential. “Even if it is successfully established that we can act freely and thus morally, it does not automatically follow that we ought to act morally” (Schönecker, 1999, 69).21 In my opinion, this thesis does not appear to be consistent with Kant’s comprehension of the structure of imperatives. This structure, as outlined above, can be summarized as follows: for a being capable of action a) in accordance with his or her faculty of reason the practical necessity exists of acting in a certain manner; this being, however, b) does not automatically conduct himself or herself as would be ap18 19 20 21

Cf. also Schönecker (1999, 157): “If a being acts freely, then it acts morally.” For an example see my own interpretation in Steigleder (2002). See Schönecker (1999, 161 f.). See also Schönecker (1999, 79, 154, and 157, for example).

The Analytic Relationship of Freedom and Morality

243

propriate to his or her faculty of reason, and therefore, c) the practical necessity to act becomes a demand, a necessitation, for the being capable of action. This structure is valid for hypothetical imperatives as well as for categorical imperatives, if they exist. In the case of categorical imperatives the practical necessity lies in the necessity of taking the unconditionally necessary end into account, as which every being, capable of acting freely in the strict sense, exists. If we must presuppose that we can act freely in the strict sense, then a necessary end for our action results from this. And insofar as we must assume that we can act freely, we must assume that we have the possibility of taking this necessary end into consideration. If we have further cause to assume that we do not automatically act as is necessary according to our faculty of reason, this necessity appears to us as an unconditional demand, becomes a categorical imperative. If, then, as Kant has established in Part 1, freedom is to be comprehended as autonomy (and not as lawlessness), then for a finite being endowed with the capacity for freedom the validity of the categorical imperative follows without question. Here no deduction (establishing the validity of a synthetic judgment a priori) is required. The intention of Kant’s proof is rather to establish that we, although we do not automatically act rationally, must necessarily see ourselves as “rational beings, which have a will.” The problem is not whether we ought to act, in the event that we can act, but whether we must necessarily (and not simply as a possibility of thought) attribute ourselves with a specific capacity, which we must necessarily see as linked to our capacity for action. Let us continue by assuming that in Groundwork III Kant is successful in presenting the appropriate proof and has thus established the categorical imperative as a valid moral principle. Nevertheless, it still remains difficult to comprehend why a finite being endowed with the faculty of pure practical reason is capable of not obeying the command of his or her faculty of reason in a responsible manner. Kant saw this difficulty as being unresolvable on grounds of principle. It is important to note that the formulation of the question itself requires a differentiation in the conception of freedom. Freedom 1 connotes the capacity of pure practical reason to be inherently purposive, freedom 2 connotes the capacity of a finite being endowed with freedom 1 to allow this freedom to be realized or not. I have attempted to express this difference in the concepts of freedom above by maintaining that even then, when we ascribe the faculty of pure practical reason to ourselves, we may by no means understand ourselves as purely rational beings. Kant later tried to capture the difference in the concepts of

244

Klaus Steigleder

freedom terminologically in the distinction between “Wille” (“will”) and “Willkür” (“free choice”). 22 Here “will”(“Wille”) signifies the faculty of pure practical reason, “free choice” (“Willkür”) the faculty of desire possibly endowed with such a will. The free faculty of desire remains, in a vaguely comprehensible manner, the center of decisionmaking, so to speak, for a finite being capable of action endowed with a will. The being capable of action is equally responsible for his or her decisions to allow or to disallow the will to act. My thesis, which also informs my interpretation of Part 1, is that Kant has already established the distinction between “will” (“Wille”) and the faculty of “desire” (“Willkür”) in the Groundwork, not terminologically but essentially. “Will” signifies in the Groundwork fundamentally and almost always “free faculty of desire” respectively “free choice” (“freie Willkür”) or “practical reason.” Of particular relevance here is the question of capacities of the “will” in the sense of “free choice.” The entire course of the study would be incomprehensible if one were to comprehend the “will” as the faculty of pure practical reason from the beginning. This is already the case, as can be shown, with Kant’s discussion of the “good will” in Groundwork I. Here, too, in Part 1 of Groundwork III being discussed, the argumentation is conclusive if we, as proposed above, understand the “will” as “desire” or “free choice.” This is clear in the opening sentence. Finite beings capable of action (“living beings”) possess, according to Kant, “free choice,” in contrast to the “arbitrium brutum” of animals.23 This is characterized as a “kind of causality.” The possibility of freedom in the strict sense is cautiously introduced: “and freedom would [!] be that property of such causality that it can [!] be efficient independently of alien causes determining it” (GMS, 446). Apparently this is a faculty of practical decisions, which is subject to “alien,” namely, sensible influences. For this faculty the “property” of freedom is being considered hypothetically and to such an extent that the sensible influences must not be determining influences, but that rather the possibility of a determination to action exists independent of sensibility. 24 While this view makes perfect sense for the will, understood as “free choice,” it 22 23 24

See MdSR, 213 f. Cf. KrV, B830 / A802. Cf. MdSR, 213 f.: “Freedom of choice is this independence from being determined by sensible impulses; this is the negative concept of freedom. The positive concept of freedom is that of the ability of pure reason to be itself practical. But this is not possible except by the subjection of the maxim of every action to the condition of its qualifying as universal law.”

The Analytic Relationship of Freedom and Morality

245

would, on the other hand, be senseless if one were to understand the will as the faculty of pure practical reason from the beginning. The presupposed distinction between “free choice”(“freie Willkür”), which (possibly) possesses the faculty of pure practical reason, and “pure practical reason” requires, as noted above, a distinction in the conception of autonomy as well. The autonomy of free choice (in the strict sense) is an endowment which would result in the impossibility of the determination of such free choice through sensibility, free choice always having the possibility yet not the necessity (“automatically”) of observing the law of pure practical reason. The autonomy of pure practical reason, in contrast, would be a necessity, an “automatism,” so to speak. Kant had considered this difference exactingly in Groundwork II when he speaks of a will that is simultaneously lawgiving and under the law, that is, subject to his or her law, 25 and distinguishes this will (which, as it must be clear by now, cannot simply be equated with pure practical reason) from the will of a purely rational being, which is lawgiving yet not simultaneously subject to this lawgiving.26 Kant apparently proceeds from this distinction when he writes in our Part 1 of Groundwork III: “hence a free will and a will under [!] moral laws are one and the same.” To be sure, Dieter Schönecker has indicated that Kant can also characterize a fully rational being as being “under” moral laws (Schönecker, 1999, 162 f.), yet the hypothetical supposition of such a being is, as I have shown, not what Kant primarily envisioned in our context.

25

26

See, e. g., GMS, 331: “Hence the will is not merely subject to the law but subject to it in such a way that it must be viewed as also giving the law to itself […].” See GMS, 433 f.: “A rational being belongs as a member to the kingdom of ends when he gives universal laws in it but is also himself subject to these laws. He belongs to it as sovereign when, as lawgiving, he is not subject to the will of any other. A rational being must always regard himself as lawgiving in a kingdom of ends possible through freedom of the will, whether as a member or as sovereign. He cannot, however, hold the position of sovereign merely by the maxims of his will but only in case he is [a] completely independent being and with unlimited resources adequate to his will.”

246

Klaus Steigleder

Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–) (abbreviated as ‘AA’). The Critique of Pure Reason will be cited according to the A/B pagination from the first and second editions. All textual references are to The Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press, 1992–). GMS Grundlegung zur Metaphysik der Sitten, AA, IV Ent Über eine Entdeckung … AA, VIII KpV Kritik der praktischen Vernunft, AA, V KrV Kritik der reinen Vernunft, AA III, IV Log Logik-Jäsche, AA IX MdSR Metaphysik der Sitten, Rechtslehre, AA, VI

Other works Schönecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs, Freiburg i. Br. Steigleder, Klaus (2002): Kants Moralphilosophie. Die Selbstbezüglichkeit reiner praktischer Vernunft, Stuttgart.

Corinna Mieth and Jacob Rosenthal

“Freedom must be presupposed as a property of the will of all rational beings” (GMS III, 2) Section III of Kant’s Groundwork is subdivided into five (or six, including the “Concluding remark”) parts. The task of this section is to show “that a categorical imperative is possible”, a task that has at least two aspects. The first concerns justification: Kant aims to show that and why the moral law is a valid principle for any rational will and in its imperative form a binding principle for our will, and no “chimerical idea” or “phantom” (GMS, 445). The second aspect concerns motivation: how our will and actions can be determined by the moral law independently of any inclinations we may happen to have, how it is possible that we take an “interest” in the moral law (GMS, 460). The second question cannot be answered: “[…] it is impossible for us to explain […] how pure reason can be practical” (GMS, 461).1 As a first step, Kant focuses on a rational will as such (parts 1 and 2). Subsequently, he confronts the difficulty surrounding the fact that our human wills are not purely rational and it is therefore not prima facie guaranteed that what is valid for a rational will as such is in the same way also valid for our wills (parts 3–5). In our commentary, we focus on the interpretation of part 2 (GMS, 447–448). To begin with, we give a short summary of part 1 in order to sketch the background of our passage. Then we carry out our task in two steps: the first is a reconstruction of the argumentation of part 2, the second is a discussion of five main problems of interpretation with regard to it. These include the ambiguity of the concept of freedom, the significance of the first-person perspective, the kind of reason to which Kant’s argument applies, the relation between freedom and indeterminism, and last but not least the extent to which Kant’s argument is successful, in other words, what it actually establishes. 1

For a thorough discussion of the meaning of the question “how a categorical imperative is possible” see Schönecker (1999, ch. 2).

248

Corinna Mieth and Jacob Rosenthal

In part 1, Kant begins with the notion of a free will. He claims that “the concept of freedom is the key to the explanation of the autonomy of the will” (GMS, 446), that an autonomous will is nothing but a will under the moral law (“reciprocity thesis”)2, and that this connection is an analytical one, i. e. morality follows analytically from the concept of freedom (“analyticity thesis”). Kant first gives a negative characterization of freedom: A free will is a will that is independent of alien determining causes, and this means for Kant that the will is not determined by natural causes according to the laws of nature. But it must be determined somehow and according to some law, since the concept of free will includes the notion of causality (because the free will is supposed to cause actions) and this notion, in turn, carries with it the concept of law. The activity of a free will, by which it causes actions, must thus follow some law. This can only be a law the will imposes on itself, for otherwise the will would be subject to an alien law and not free. So the negative characterization of freedom of the will, namely, absence of alien determining causes, leads by itself to a second, positive characterization: A free will is an autonomous will, a will that acts according to self-given laws. But, as Kant has shown in Section II, the common principle of those laws or maxims is that they can be willed as universal laws through themselves. This is what the categorical imperative commands, which therefore captures the very idea of autonomy – at least as far as its content is concerned. In its imperative form it applies only to imperfect wills, which do not by themselves act in accordance with the stated principle. The categorical imperative is also the principle of morality, as Kant attempted to show in his analysis of our common moral understanding. A free will is an autonomous will, a will under self-given laws, and this in turn is a will that acts according to the moral law. “If, therefore, freedom of the will is presupposed, morality together with its principle follows from it by mere analysis of its concept.” (GMS, 447) But can freedom of the will be presupposed? This is the subject of part 2 of Section III that we are going to discuss in this contribution. Now, since Kant has argued that a free will is a will under the moral law, the question naturally arises, if our will (or any rational will) is free. If this question can be answered positively, Kant has shown that we (or any rational being3) are subject to the moral law. Three argu2 3

Cf. Allison (1990, 2002). When we speak of “rational being” we follow Mary Gregors translation of “vernünftiges Wesen”.

“Freedom must be presupposed as a property of the will of all rational beings”

249

mentative steps can be distinguished in our passage. In a first step (“It is not enough […] endowed with a will”), Kant sets up his task: to give an a priori proof for freedom. In the second step (“I say now […] theoretical philosophy”), Kant claims that if a rational being has to view itself as free, it is really free “in a practical respect”. In the third step (“Now I assert […] every rational being”), Kant argues that indeed any rational being has to view itself as free. We will quote and interpret the passage along the three indicated steps, in order to give a reconstruction of Kant’s argument (I), and then discuss five critical points about it (II).

I. Reconstruction Step 1: It is not enough that we ascribe freedom to our will on whatever ground, if we do not have sufficient ground for attributing it also to all rational beings. For, since morality serves as a law for us only as rational beings, it must also hold for all rational beings; and since it must be derived solely from the property of freedom, freedom must also be proved as a property of all rational beings; and it is not enough to demonstrate it from certain supposed experiences of human nature (though this is also absolutely impossible and it can be demonstrated only a priori), but it must be proved as belonging to the activity of all beings whatever that are rational and endowed with a will. (GMS, 447 f.)

There is one problem with part 1 of Section III that Kant has to solve in part 2, namely he has not yet shown that we can indeed presuppose freedom as a property of the will of all rational beings (and therefore ascribe morality to ourselves as rational beings). At the end of part 1, he states that he intends a “deduction of the concept of freedom from pure practical reason” which requires “further preparation” (GMS, 447). So the idea is to show that we have to ascribe freedom to our will insofar as we are rational. Why should we look for a proof of freedom in practical reason? As a preliminary remark, Kant reminds us that we cannot take an alleged proof of our freedom from particular experiences. That might of course be tempting: Don’t we experience our freedom directly? But that would merely be an a posteriori argument for freedom, and moreover, it would at best be valid for human beings. Since the task of the Groundwork is to develop “a pure moral philosophy, completely cleansed of everything that may be only empirical

250

Corinna Mieth and Jacob Rosenthal

and that belongs to anthropology” (GMS, 389), this road is closed, and an a priori argument for freedom must be given that is valid for any rational will. Freedom has to be proved as a property of beings that are a) rational and b) have a will. So in part 1, Kant hypothetically presupposed freedom and arrived at morality analytically. Now he sets up the task of proving freedom from rationality, and so in the end he moves from rationality to morality. An outline of the completed argument would run as follows: 1. Every rational being has a free will. 2. A free will is not determined by alien causes or according to alien laws. 3. A free will is an autonomous will. 4. Autonomy is the property of the will to be a law to itself. 5. The fact that a will is a law to itself means that it acts on a maxim that can be willed as a universal law through itself, which is the principle of morality. 6. An autonomous will is a will under the moral law. 7. The will of every rational being is a will under the moral law. The argumentative program for the passage under discussion is: 1. Morality is valid for us insofar as we are rational beings. 2. Morality must be valid for all beings insofar as they are rational. 3. Morality follows from freedom as a property of the will of a rational being. 4. The reason to ascribe freedom to a will must be the same for all rational beings. 5. Freedom must be proved as a property of beings that are a) rational and b) have a will. Step 2: I say now: every being that cannot act otherwise than under the idea of freedom is just because of that really free in a practical respect, that is, all laws that are inseparably bound up with freedom hold for him just as if his will had been validly pronounced free also in itself and in theoretical philosophy. [In a footnote, Kant restates this in greater detail:] I follow this route – that of assuming freedom, sufficiently for our purpose, only as laid down by rational beings merely in idea as a ground for their actions – so that I need not be bound to prove freedom in its theoretical respect as well. For even if the latter is left unsettled, still the same laws hold for a being that cannot act otherwise than under the idea of his own freedom as

“Freedom must be presupposed as a property of the will of all rational beings”

251

would bind a being that was actually free. Thus we can escape here from the burden that weighs upon theory. (GMS, 448)

Step 2 consists in a claim expressed in the fi rst half-sentence, and its following clarification and justification. Kant does not aim at what he calls a theoretical proof for freedom: a straightforward argument to the effect that any rational will is a free will. Such an argument would be difficult or impossible to give, but that does not matter, because for Kant’s purposes it is dispensable. Instead, he claims that any being that has to view itself as free when acting (that has to act “under the idea of his own freedom”) may be considered to be free “in a practical sense”. That is, the being may be considered to be subject to the same laws as if it had been proved that it is free. What it means that any rational agent has to view himself as free will become clear in step 3, when Kant presents his argument for this claim. What is clear by now is that Kant’s argument is bound up with the perspective of the first person, because any rational agent has to presuppose his own freedom. The topic is “freedom […] as laid down by rational beings […] as a ground for their actions”. The “practical respect” mentioned here as opposed to the “theoretical respect” characterizes the kind of proof Kant is going to give for the freedom of any rational being. If any such being has to view itself as free, it also has to admit that it is subject to any law that goes along with freedom, in particular, it has to accept the moral law as a valid law for its will. Step 3: Now I assert that to every rational being having a will we must necessarily lend the idea of freedom also, under which alone he acts. For in such a being we think of a reason that is practical, that is, has causality with respect to its objects. Now, one cannot possibly think of a reason that would consciously receive direction from any other quarter with respect to its judgments, since the subject would then attribute the determination of his judgment not to his reason but to an impulse. Reason must regard itself as the author of its principles independently of alien influences; consequently, as practical reason or as the will of a rational being it must be regarded of itself as free, that is, the will of such a being cannot be a will of his own except under the idea of freedom, and such a will must in a practical respect thus be attributed to every rational being. (GMS, 448)

As in step 2, Kant starts with a claim and gives the justification afterwards. The claim reformulates the title of part 2: Freedom must be presupposed as (not: is) a property of the will of all rational beings.

252

Corinna Mieth and Jacob Rosenthal

Presupposed by whom? In the first place, freedom must be presupposed by the rational being itself who is acting, as we have already seen, i. e. freedom must be presupposed from the first-person perspective. What Kant additionally seems to claim is that “we” (from the perspective of the third person) must ascribe (“necessarily lend”) the idea of freedom to any agent who has to view himself as free (“acts under the idea of freedom”). There are two possibilities of interpretation for this, a weak and a strong one. The weak one says that this is not really an additional claim. We simply discover that the agent has to consider himself to be free and in this sense “lend him the idea of freedom”. The strong one says that we have to consider him to be free. We have to introduce a distinction here, because there is a certain ambiguity in terminology. If we speak of a rational being we might merely think (weakly) of a being that engages in practical deliberations. Any such being has to consider itself to be free, as Kant subsequently shows, but of course we don’t have to view it so. We can ascribe its practical judgments to mere impulses, while the being itself assumes at the same time that it is acting according to reason. But, if we speak of a rational being, we might also think (strongly) of a being endowed with a purely rational or perfect will, and such a being would of course have to be considered to be free by any spectator and not just by itself. The point is that we cannot know if we are rational beings in this strong sense. But what we know is that we are rational beings in the aforementioned weak sense: We engage in practical reasoning and view our practical judgments as causing our actions. In doing so, we have to view ourselves as being guided by reason and therefore as free. We do not have to think that we are rational throughout (we know that we are not), but when we deliberate what to do and come to a considered judgment, we have to ascribe this judgment to reason and so hold ourselves to be free at least in this respect. Kant states that if we think of a rational being with a will, we have to assume that it is provided with practical reason: a reason that “has causality with respect to its objects.” In this assumption, Kant identifies the will of a rational being with practical reason: We thereby think of a will as determining what is chosen or what is done in a reasonable way. The latter is important, because Kant does not think of a will here in the sense that it has a freedom to choose whatever it likes in an indeterminate way. The formulation that the will “has causality with respect to its objects” rather means that we have a reason for an action, and that the action is really due to this reason, and not due

“Freedom must be presupposed as a property of the will of all rational beings”

253

to some other external cause, e. g. an impulse. We have to remember what Kant stated in part 1 about the two kinds of causality. On the one hand, there is causality “just as natural necessity” (GMS, 446) that applies to nonrational beings as determined by alien causes. On the other hand, there is a causality independent “of alien causes determining it” (GMS, 446) which corresponds to the negative definition of freedom, namely freedom from alien causes. What Kant develops from the first person perspective amounts in the first place to showing that we have to consider ourselves to be free in this negative sense. If we engage in practical deliberation, we view ourselves as rational beings with wills, and then we cannot consistently think of ourselves as “receiv[ing] direction” from unreasonable sources: from impulse or inclination, which would mean heteronomy. Instead, we have to assume freedom in the sense of autonomy, which corresponds to the positive concept of freedom. This concept has two aspects: the aspect of spontaneity (we do not receive direction from alien causes, but our will has causality with respect to its objects), and the aspect of acting under self-given laws (as any kind of causality goes along with some law, the spontaneous will acts according to self-given laws, everything else would mean heteronomy). Now, these self-given laws are the laws of reason. For, if you reason out what to do, you must consider your judgment to be dependent on principles of reason, and not caused by an impulse, because otherwise it would be no judgment at all. You cannot say: “I judge this course of action to be best, to be the one I should adopt, but I only say so because I had no breakfast today.” The second half of the sentence simply renounces what the first one states. The first half states that you have reasons to do such-and-such, whereas the second half denies this: there is merely a natural cause operating on you. If you really think that a natural impulse drives you to judge and do such-and-such, you cannot at the same time consider this action to be a rational action, and much less the foregoing practical judgment to be a reasoned judgment, in fact, you cannot view it as your judgment at all. So, a rational being, i. e. a subject who is able to deliberate and make considered judgments, cannot “attribute the determination of his judgment not to his reason but to an impulse”. It must view its reasoning as being governed by principles of reason (which are the laws reason gives to itself). The reader will note that this line of argument applies to theoretical as well as to practical reasoning. A deliberating being, if it wants to claim validity for the resulting judgments, has to presuppose that its deliberations are guided by the laws of reason, no matter if

254

Corinna Mieth and Jacob Rosenthal

these deliberations are theoretical (what to believe) or practical (what to do). As the passage “Reason must regard itself as the author of its principles independently of alien influences; consequently, as practical reason […]” (emphasis by us), and also the foregoing sentence “Now, one cannot possibly think of a reason that would consciously receive direction from any other quarter with respect to its judgments […]” indicate, Kant is aware of this. Of course, his main interest is practical reason, consequently most of his formulations refer to the will, and so it may be disputed that he has more in mind here than practical reson, namely, reason in general. But at least the systematic fact remains that Kant’s argument, if valid at all, is as valid for theoretical deliberations and judgments as for practical ones. We will come back to this in the third comment. So, any rational (in the sense of deliberating) being has to consider itself to be autonomous, to be free in the positive sense, at least as far as its actual deliberations and judgments are concerned. In particular, it has to view itself as acting according to the laws of reason, and it has to view these laws as self-given. Therefore every deliberating being has to admit that it is subject to the principle that expresses the idea of autonomy, i. e. the moral law. In particular, we (as human beings) have to view ourselves so and thus accept the moral law as a law for our wills. It cannot be questioned that human beings are rational beings in the sense that is required for Kant’s argument to go through, for such a question, when put forward by a human being, presupposes what is questioned. Kant here gives (or at least purports to give) an argument in order to establish certain conclusions, and we try to comprehend and judge it (or, at least, we view ourselves as doing this). In doing so, we have to look upon ourselves as being subject to the laws of reason. Whenever we think about what to do or what to believe, we are bound to view ourselves as rational beings, as far as this very activity of deliberating is concerned. Let us assume that Kant’s argument is successful. Let us assume that in deliberating we have to view ourselves as free in the negative as well as the positive sense. Then the question arises why he has not thereby reached his goal. Hasn’t he shown that any being that is capable of practical deliberation is subject to the moral law, in the sense that any such being has to admit that it is subject to this law? Why does Kant not simply stop at this point, why the further parts of Section III? This is due to the fact that we are not only rational, deliberating beings, but also sensible beings. As rational beings, we are subject to the self-given laws of reason, and in particular to the moral law, but

“Freedom must be presupposed as a property of the will of all rational beings”

255

as sensible beings, we are subject to the laws of nature. So, we have a kind of antinomy here. We are subject to two kinds of laws that according to Kant pull in different directions. So, for him it is not enough to show that we have to accept the moral law as valid for us, because there is another law with the claim to determine our behaviour: the natural law that directs us by way of our inclinations and impulses. Something has to be said about the relationship of these two laws. The problem can also be posed thus: Even if we (have to) admit that we are subject to the moral law (and whatever laws of reason there might be), we know that we often act contrary to the moral law. It is clear that at least sometimes we are not determined by the laws of reason. In these cases, other laws must be guiding us. The laws of reason therefore do not have “objective necessity” for us, but at most “subjective necessity” (GMS, 449). So, it is not enough to view ourselves as rational agents. Another perspective is forced upon us, and something must be said about how these two perspectives and the laws that correspond to them relate. In addition, Kant has to argue that one perspective, that of ourselves as rational beings, can claim priority over the other, for otherwise he could not establish that we always should follow the laws of reason, no matter if we actually do so or not. If we were endowed with perfect wills, Section III of the Groundwork could have ended with part 2, but we are not. This does not mean that the first two parts of Section III are exclusively concerned with perfectly rational beings. They are not: they are concerned with all beings who deliberate what to believe and what to do, and it is shown that all those beings are in deliberating bound to look upon themselves as being guided by the laws of reason, at least as far as the ongoing deliberation is concerned. So, one could say that part 1 and 2 of Section III are concerned with the rational part of rational beings. But this is not the only aspect of those beings, if they are not perfectly rational. They know, first, that in another respect they are not guided by the laws of reason but by the laws of nature, and second, that they often act contrary to what reason requires. The argument that we are free in a practical respect may be valid, but there is also an argument that we are subject to the laws of nature, that our actions are completely determined by these laws. As Kant holds the latter considerations to be at least as weighty,4 we face a conflict of perspectives that is to be resolved. This conflict, to repeat, has two sources: First, that Kant thinks that we cannot be determined by the laws of reason and by the laws of nature 4

Cf. GMS, 456.

256

Corinna Mieth and Jacob Rosenthal

simultaneously, at least not in any straightforward sense, because the negative aspect of freedom consists in the absence of natural determination according to the laws of nature. Second, that we know very well that we are not always guided by the laws of reason. To resolve the conflict, Kant introduces an ambitious kind of compatibilism that is based upon his distinction between noumenal and phenomenal world, and two corresponding points of view. As the phenomenal world, the world of appearances, is completely determined by natural laws, freedom must be situated in the noumenal world, i. e. we have to consider ourselves to be things in themselves that spontaneously cause events in the phenomenal world. Our actions, viewed as events in the phenomenal world, are on the one hand fully determined by natural causes according to the laws of nature, and are thus mere links in a great causal chain. The very same events, viewed as our actions, are on the other hand spontaneously caused by our noumenal selves. In addition, Kant has to show that one of these conflicting perspectives, namely the homo-noumenon-stance, has priority over the other. He does this by arguing that “the world of understanding contains the ground of the world of sense” (GMS, 453) and therefore, “what belongs to mere appearance is necessarily subordinated by reason to the constitution of the thing in itself” (GMS, 461). All this constitutes one important topic of the remaining parts of Section III. (Another is the problem of motivation: how it can be that we take an interest in the moral law, independently of any inclination or impulse.) We will come back to this solution later (in the fifth comment), but only in passing because it is not our main topic. In the meantime, it is important to keep in mind that in GMS III, 2 Kant does not draw on the distinction between phenomenal and noumenal world at all, and consequently, that his argumentation in the passage under discussion does not depend on it. To assess its validity or to find it plausible, we do not have to buy Kantian dualism or his special kind of compatibilism. II. Discussion 1. There are several, and not only two, notions of freedom involved in Section III of the Groundwork, which at the beginning of the discussion should be distinguished clearly, even if either further or obvious or well-known considerations show that these notions amount to one and the same thing. What Kant actually proves in our passage is this: A reasoning being has to assume that in its reasoning it obeys the

“Freedom must be presupposed as a property of the will of all rational beings”

257

principles of reason. This, and no more, is shown. It is not proved that the being cannot view itself as standing under the laws of nature, nor, that the being has to view itself as “spontaneous”, nor, that the laws that guide the reasoning must be considered to be “self-given”, nor, that if the reasoning issues in an action, the being “could have done otherwise”. These conclusions may follow from further or earlier considerations, and Kant undoubtedly thought that at least some of them would, but the argument as such in the passage under discussion does not lead to them. So, we have to distinguish five notions of freedom, the relationship of which is to be elucidated. (i) A being is free1, if it deliberates, judges, wills and acts according to the laws of reason. This is the kind of freedom that is proved – at least in a “practical” sense – in the passage under discussion. It could be called “freedom as rationality”. (ii) A being is free2, if its deliberations and judgments, its will and actions are not determined by the laws of nature (or other “alien laws”). This is what Kant calls “negative freedom”, one could also speak of “freedom as the absence of heteronomy”. (iii) A being is free3, if it acts spontaneously, if it is the starting point of new causal chains, if its will “has causality with respect to its objects”, if it is, so to speak, an unmoved mover. This is “freedom as spontaneity”. (iv) A being is free4, if it deliberates and acts according to self-given laws. It is this kind of freedom that is most properly called “freedom as autonomy”. (v) A being is free5, if it has genuine alternative possibilities of action in the sense that its choice of an action, its will, is not pre-determined by any foregoing factor whatsoever. This might properly be called “freedom as indeterminism”. Two general remarks about these definitions: First, these notions of freedom are not necessarily meant to denote permanent properties of beings. That may be the case, but it need not, it is a further question. For example, when we take freedom as rationality, it seems that human beings sometimes deliberate, judge, will and act rationally, and sometimes they do not. When they do, they are free1, when they do not, they are not. The same holds for the other notions of freedom. There is no guarantee, nor would it be plausible, that a being that is sometimes free in a certain sense must always be free in this sense. So one could add a time index to the definitions above: “A being is freei at time t, if at time t it …”. Second, it might be objected to the definitions that it would be more plausible to refer to abilities or capacities than

258

Corinna Mieth and Jacob Rosenthal

to actual performances. That is, for example: “A being is free1 (at time t), if (at time t) it is able to deliberate, judge, will and act according to the laws of reason (no matter whether it actually does so or not)”, and analogously for the other notions of freedom. To refer to abilities in such a way may be appropriate concerning several kinds of freedom, but not concerning freedom of the will, which is at stake here. With freedom of the will, such a move would merely shift the issue one step back. If a being is able or has the capacity to do something, it depends (among other things perhaps) on its will whether it actually does so or not. But we are dealing here precisely with the question how the will of the being comes about, and what we have to assume about this coming about in order to call it free. The five notions of freedom are five different ideas what freedom of the will might consist in. It cannot consist in a certain ability of the subject, because it depends on the subject’s will whether the subject makes use of its ability or not, and therefore, as long as we talk about abilities or capacities, the will is out of focus and all problems concerning its freedom remain open. Let us add some remarks on terminology here. As we know from other important notions in Kants oeuvre, he does not always use terms in a stringent way. In the Groundwork, the notion of spontaneity plays no prominent role, it just shows up in GMS 452, when Kant talks about the “pure self-activity” of reason. But he has the same thing in mind, when he, in our passage, speaks of “a reason that […] has causality with respect to its objects” (GMS, 448). In Prolegomena, § 53, he makes the distinction between causality interpreted as natural necessity and causality as freedom, and introduces explicitly freedom as spontaneity (Prol, 344). In the Critique of Pure Reason, Kant speaks of “transcendental freedom” to describe the same thing (e. g. KrV, A 445/B 473 ff., A 532/B 560 ff.). The ability of self-determination, that in Kant’s view implies that one’s will is not determined by sensual impulses, is called “practical freedom” there (KrV, A 534/B 562). These two aspects of practical freedom are what Kant calls “positive freedom” and “negative freedom” in GMS. In GMS III, 3 reason (“Vernunft”) is described as “pure self-activity” (“reine Selbsttätigkeit”) as opposed to understanding (“Verstand”) which is restricted to the function “to bring sensual representations under rules” (GMS, 452). Reason shows a “pure spontaneity” as an ability to build ideas independently of empirical and sensual influences. 5

5

Cf. KdU, § 77.

“Freedom must be presupposed as a property of the will of all rational beings”

259

The notion of autonomy is often traced back to Rousseau’s political philosophy in which the point is made that political freedom consists in being subject to no other laws than the ones to which we have given our consent (Schneewind, 1992, 313 f.; Wimmer, 1980; Sturma, 2004). Another concept of autonomy is the “empirical” one of John Stewart Mill or Harry Frankfurt, autonomy “as action on reflectively endorsed desires, or as avoiding specific sorts of social and personal dependence.” (O’Neill, 1989, 53 f.) In Kants moral philosophy, autonomy does not mean either of these, it rather describes the fact of being subject to the moral law that we have to give to ourselves insofar as we are rational. So we are subject to a law (nomos) that we have given to ourselves (autos). In GMS III,1 Kant’s “negative freedom” is freedom2. This is the notion Kant starts from. It means willing and acting in absence of alien determining causes. We note that for Kant natural causes, and in particular our own inclinations, are alien causes and vice versa. That goes without saying for readers of the Groundwork, but is by no means unproblematic. That all and only natural causes are alien causes is in need of justification. We take the point for granted here and address it later (in the fifth comment). Now, what does it mean to act, but not from alien determining causes? How can such a situation be characterized positively? One can either think of case (iii) or of case (v), but freedom as indeterminism is ruled out by Kant as an absurdity, for it amounts to the idea that the will is under no law at all. So, the absence of alien determining causes implies freedom in the sense of (iii): A being that is free2 (“negatively”) in its willing and acting has also freedom in the sense of spontaneity. But, as any kind of causality is causality according to some law, freedom3 implies freedom4. If a being is free3, it can start causal chains, which must occur according to some law, which can only be a self-given law, for otherwise the being would be subject to an alien law and not be free2. Thus, according to Kant, (ii), (iii) and (iv) are only different characterizations of the same situation. The notion of freedom is introduced through the contrast between autonomy and heteronomy that he previously developed and explained at the end of GMS II. Freedom is not introduced through the contrast between determinism and indeterminism, which would point to freedom in the sense of (v). The autonomous will is not indetermined, but determined by self-given laws. Now, what is proved by Kant in our passage (not theoretically, but at least practically) is freedom1: freedom as rationality.6 It is the same 6

This is what Dieter Henrich (1975, 64 f.) calls “Vernunftfreiheit” (freedom of reason) or “Urteilsfreiheit” (freedom of judgment). He distinguishes it from “transzen-

260

Corinna Mieth and Jacob Rosenthal

as freedom as autonomy if and only if the laws an autonomous being gives to itself are the laws of reason. Why should this be the case? It is only the case if we identify a rational being with its reason – with its rational part. The point of Kant’s argument here is that in reasoning a rational being has to identify itself with reason, and so has to consider itself to be free in the senses (i) and (iv) simultaneously. The self-given laws are just the laws of reason. So, Kant seems to have proved successfully that in deliberating any being has to view itself as free1, free2, free3, and free4 simultaneously (whereas freedom5 is an absurdity). This preliminary judgment will be further clarified and scrutinized in the following comments. 2. Note that it is the fi rst-person perspective from which freedom as rationality must be presupposed. That is why so many commentators come to the conclusion that one of the main topics of GMS III is the self-relation of the subject (e. g. Prauss, 1993, 255; Steigleder, 2002). But then the problem arises of whether this insight about ones own practical reason is necessarily to be transferred to others, and why. You can easily say about somebody else: “He judges this course of action to be best, but his judgment is due to the fact that he had no breakfast today and is therefore in a bad mood.” But you cannot say such a thing about yourself. If you do, you contradict yourself in an outright way. Now, in making this statement about a third person, you represent this person’s will as irrational. This means that even from the third-person perspective a rational will must be considered to be subject to the laws of reason, otherwise it would simply be no rational will. But this analytic truth is not Kant’s point here. He does not say: “From the notion of a rational will, it flows analytically that it is a will under the laws of reason.” If this were his point, he would have given us a purely theoretical proof for the freedom of any rational will. Kant’s concern here is not whether a rational will is a will that follows the laws of reason, which is clear, but whether we actually have rational wills in this sense. The point is that in practical deliberation we have to represent ourselves as rational and therefore as free1. So, dentale Freiheit” (transcendental freedom) and claims that the further is relevant for theoretical contexts and rational beings in general, whereas the latter is relevant for practical contexts and for rational beings that are endowed with a will. In Henrich’s interpretation, transcendental freedom is the presupposition for the principle of morality. It is a mixture of what we call freedom 2 , freedom 3 and freedom4. Using this terminology, Kant’s argument establishes freedom of reason in a straightforward way, whereas the case for transcendental freedom is problematic and remains to be scrutinized.

“Freedom must be presupposed as a property of the will of all rational beings”

261

the title Kant has given to the passage can be misunderstood. That freedom understood as rationality is (not only: must be presupposed as) a property of a rational will, is true analytically, but whether we are entitled to view our wills as rational is not so clear. We can doubt that we have rational wills. What we cannot doubt is that we engage in practical deliberation, and what Kant shows is that in doing so we have to consider ourselves to be rational and therefore free1, at least as far as an ongoing deliberation is concerned. It is this argument that works only from the first-person perspective. I can say about your judgments that they depend on non-rational natural impulses, but I cannot say it about mine. If I say it about mine, they are no longer my judgments. Therefore, every deliberating being has to consider itself to be rational, at least with respect to its actual deliberation. It was Lewis White Beck (1960, cf. 1975), more recently followed by Christine Korsgaard (1996), who emphazised the difference between the perspective of the actor and that of the spectator. From the (first person-) perspective of the actor who has to view himself as the author of his actions and his foregoing deliberations, freedom must be presupposed. Beck and Korsgaard also speak of the “practical point of view”, refering to the doctrine of the two standpoints in GMS III, 3. Freedom is thus characterized as an idea that is a creation of practical reason – not as a mere fantasy, but as a necessary assumption. On the other hand, from the perspective of the spectator, we observe events in the world that we can describe from the “theoretical point of view”, according to the laws of nature. We do so without interpreting ourselves as involved and without interpreting the events as results of our will. While it seems possible, although very unnatural, to describe another “actor” as merely a part of nature, as if he were not free, it is of crucial importance for Kants moral philosophy that we are not only forced to interpret ourselves as free but also lend the idea of freedom to all other rational beings. If we had no reason to apply the idea of freedom to other beings, it would not be clear why we should treat them as persons, as ends in themselves, and not as mere means or things. Therefore it is no coincidence that the notion of autonomy at first appears in combination with the formula of humanity in GMS II when Kant describes “every rational being” as “the subject of all ends” and therefore as an “end in itself” (GMS, 431). “From this there follows” as Kant points out “the third practical principle of the will”, the so-called principle of autonomy, that describes the will not only as “subject to the law but subject to it in such a way that it must be viewed as also giving the law to itself and

262

Corinna Mieth and Jacob Rosenthal

just because of this as first subject to the law (of which it can regard itself as the author).” (GMS, 431) From there we are lead, according to Kant, to the idea of a kingdom of ends that is “a systematic union of rational beings through common objective laws” (GMS, 434) that can be understood as common and objective because they “abstract from the personal differences of rational beings as well as from all the content of their private ends” (GMS, 433). But this means to interpret the others as free beings as well. It seems that from the argument given in GMS III, 2 (that should give a justification for the moral point of view developed in GMS II) it follows that we have to consider ourselves to be free i. e. autonomous because we have to consider ourselves to be rational from the first-person perspective. And we know that any other deliberating being has to do so as well with respect to itself. But is this enough to show that we have to interpret others as free i. e. autonomous beings as well and therefore as ends in themselves that are an object of respect for all our maxims? This does not seem to be implied by Kant’s argument in III, 2, and is at the very least in need of additional argumentation. The problem surfaces as early as Section II, when Kant derives the formula of humanity (GMS, 429). There he writes: “The ground of this principle is: rational nature exists as an end in itself. The human being necessarily represents his own existence in this way; so far it is thus a subjective principle of human actions. But every other rational being also represents his existence in this way consequent on just the same rational ground that also holds for me; thus it is at the same time an objective principle […]” from which the categorical imperative in the form of the formula of humanity follows. But again, the question is, even if every rational being has to regard itself as an end in itself, why does he have to regard the other rational beings as ends in themselves, too? This is the same kind of problem as we face in our passage. To be sure, if every rational being has to view itself as autonomous, and if the principle of autonomy is “to act on no other maxim than that which can also have as object itself as a universal law” (GMS, 447), every rational being is bound to will according to this principle. But it is not clear what this principle implies with respect to other rational beings, whether it dictates a specific attitude towards them. In other words, if we look at the categorical imperative: It is not clear whether the formula of universal law indeed implies or is equivalent to the formula of humanity. This is claimed by Kant, of course, but to get to this result he has to make the problematic step just mentioned.

“Freedom must be presupposed as a property of the will of all rational beings”

263

3. Although Kant’s argument aims at practical deliberation and practical judgments, it is as valid for theoretical judgments. Kant’s primary interests here are the rational will and rational actions, but what he claims is in effect about judgments in general, be they practical or theoretical. The core sentences of his argument are: “Now, one cannot possibly think of a reason that would consciously receive direction from any other quarter with respect to its judgments, since the subject would then attribute the determination of his judgment not to his reason but to an impulse. Reason must regard itself as the author of its principles independently of alien influences […]” (GMS, 448). These statements are about reason in general, not just about practical reason. This is quite obvious from the continuation: “[…] consequently, as practical reason or as the will of a rational being it must be regarded of itself as free” (emphasis by us). So, Kant talks about reason in general, and only afterwards applies his considerations to the special case of practical reason. And this is very plausible. I can no more say “I judge this proposition to be true (probable, well-confirmed etc.), but this judgment is due to a non-rational impulse” than I can say “I judge this course of action to be the right one, but only due to a non-rational impulse”. So, as Kant’s formulations as well as systematic considerations show, his argument applies equally to any kind of deliberation and judgment.7 In this context it should also be kept in mind that Kant emphasizes the unity of reason in the Preface of GMS. There he says “I require that the critique of a pure practical reason, if it is to be carried through completely, be able at the same time to present the unity of practical with speculative reason in a common principle, since there can, in the end, be only one and the same reason, which must be distinguished merely in its application.” (GMS, 391) For Onora O’Neill, the mentioned “common principle” has to be the moral law, interpreted as “the supreme principle of all reason” (O’Neill, 1989, 52). She claims that Kant “argues not from reason to autonomy but from autonomy to reason. Only autonomous, self-disciplining beings can act on principles that we have grounds to call principles of reason. […] Autonomy does not presuppose but rather constitutes the principles of reason and their authority”. (O’Neill, 1989, 57) Kant has never really carried out his idea of the unity of theoretical and practical reason, and it is, of course, hardly believable that all principles of reason should somehow spring

7

Cf. Schönecker (1999, ch. 4.2.1).

264

Corinna Mieth and Jacob Rosenthal

from the bare idea of self-legislation.8 But this is only an aside. We will discuss the problematic relation between autonomy and rationality in another respect in the fifth comment. For now, we note that Kant’s argument applies to every activity of reasoning and judging. In particular, it applies to instrumental reasoning. First of all, even a perfect will would have to reason instrumentally if it aimed at the realization of an end. “Now, what serves the will as the objective ground of its self-determination is an end, and this, if it is given by reason alone, must hold equally for all rational beings. What, on the other had, contains merely the ground of the possibility of an action the effect of which is an end is called a means.” (GMS, 447) As any action pursues an end (as is emphasized repeatedly by Kant, e. g. RGV, 4; Ge, 279, note; KpV, 34), the process of practical reasoning can never do without instrumental, i. e. means-end-reasoning. In particular, this applies to a being that acts from the motive of duty. If it wants to do its duty in a certain situation, it always has to obey hypothetical imperatives, too. Second, if we consider an end that depends on a natural inclination, reason does not tell us to pursue the end. But it does tell us that if we are to pursue the end, we have to will the means. The validity of the hypothetical imperative expressed by the conditioned sentence does not depend on the mentioned inclination or any other “alien influence”. After all, reason in the form of the hypothetical imperative does neither command us to take the means, nor to pursue the end. It commands us to take the means provided that we (for whatever reason) are to pursue the end. Any rational being can follow this reasoning and recognize its validity, whether or not it actually approves of the end. This is how Kant puts it in the Groundwork: “Whoever wills the end also wills (insofar as reason has decisive influence on his actions) the indispensably necessary means to it that are within his power.” (GMS, 417) What makes this matter difficult is that insofar as the will depends on a natural inclination, it is not free in the negative sense. For example, at the end of GMS II, in “Heteronomy of the Will as the Source of all Spurious Principles of Morality” (GMS, 441), Kant says: “If the will seeks the law that is to determine it anywhere else than in the fitness of its maxims for its own giving of universal law […] heteronomy always results. The will in that case does not give itself the law; instead the object, by means of its relation to the will, gives the law to it. This relation, whether it rests upon inclination or upon representations of reason, lets only hypothetical imperatives become possible 8

Cf. Schönecker (1999, ch. 4.4.1).

“Freedom must be presupposed as a property of the will of all rational beings”

265

[…]” (GMS, 441) So, in how far is the will of a deliberating being free, when it engages in instrumental reasoning in order to fulfill a natural desire? Kant himself has changed his mind as to whether such a being has freedom in the sense of spontaneity (transcendental freedom). In the Critique of Pure Reason he opts for “yes” (KrV, A 547/B 575 f.), in the Critique of Judgment for “no”.9 Concerning freedom as rationality, the situation is as explained above: Reason does neither command us to take the means, nor to pursue the end, but it commands us to take the means provided that we are to pursue the end. So, we have to distinguish between the different notions of freedom here. Hypothetical imperatives carry with them an ought, and if one judges, according to a hypothetical imperative, that one ought to do such-and-such, one claims validity for this judgment, and consequently has to presuppose that the judgment is made according to the laws of reason. So, Kant’s argument de facto applies to these cases of reasoning, too, and establishes freedom as rationality. But we seem to have neither negative freedom, nor freedom as autonomy, nor freedom as spontaneity in these cases of instrumental reasoning, where reason serves a natural inclination. What complicates the matter further is that it is not quite clear how to describe the situation from the fi rst-person perspective. Generelly, when one engages in practical deliberation to serve a certain desire, one reflectively endorses that desire, at least rudimentarily, and thereby implicitly claims freedom as autonomy (or, at least, a certain sense of autonomy). From the third-person perspective it is easy to view somebody who is deliberating what to do as unfree in every possible sense, especially when we are Kantians and the person thinks about how to satisfy a certain natural desire. But it is hardly possible to do so with respect to one’s own desires – to view them as alien influences on the will and nevertheless continue thinking about how to satisfy them. To remember, it is the first-person perspective that is relevant for Kant’s argument here. It does not seem that definite solutions to these problems are to be found in Kant’s writings, and we will not discuss the issue further here. We merely note that the four different kinds of freedom involved here (rationality, negative freedom, autonomy, and spontaneity) do not automatically coincide. The systematic reason for the fact that the hypothetical imperatives create difficulties is that with them, on closer inspection, the different kinds of freedom tend to fall apart. Since Kant wants them to amount to one and the same thing, he himself is uncer9

Cf. the Introduction to KdU.

266

Corinna Mieth and Jacob Rosenthal

tain whether to call an action free that is due to the representation of a hypothetical imperative involving a natural desire. We will argue later that contrary to what Kant thinks, by his argument in GMS III, part 2 he is only able to prove (in a practical sense) freedom as rationality, but not the other kinds of freedom. We will expound this claim in our fifth comment. Freedom as rationality as viewed from the first-person perspective, to be sure, extends to any kind of deliberation, especially to instrumental reasoning. That Kant’s argument also applies to theoretical reasoning creates the following problem. Kant wants to give no theoretical proof for freedom, but a proof that we have to view ourselves as free. We cannot be sure that we are free, but we have to look upon us as if we were free. But what kind of standpoint is this? If you really have to consider yourself to be free, you are bound to say “I am free”. You cannot say: “Maybe I am not free, but I have to view myself as free.” The first part of this statement implies that you need not view yourself as free, and so contradicts the second. If we really follow Kant’s argument, we have to take its conclusion seriously. If I am really bound to consider myself to be free in any rational deliberation, I cannot at the same time say: “But this does not show that I am free.” The latter statement means that I entertain the possibility that I am not free, and consequently, that I do not necessarily view myself as free. The content of the judgment (“perhaps I am not free”) contradicts its presupposition (“in making judgments I have to consider myself to be free”). The position on freedom Kant describes is dialectically instable. If we honestly mean that we have to view ourselves as free, then we really have to view ourselves so and say “we are free”. The problem is independent of what “freedom” means. It would not arise if Kant’s argument applied only to practical deliberations and not to theoretical reasoning as well. In that case I could easily say: “In practical deliberations I have to consider myself to be free, but maybe I am not free.” I could say this without contradicting myself, because this judgment is a result of a theoretical deliberation, and so does not speak about its own presuppositions. If theoretical reasoning would not be affected by the question of freedom, I could entertain the possibility that I am not free in my theoretical as well as my practical judgments without rendering such a piece of reasoning automatically instable. But this is not the situation. Kant’s argument applies to theoretical reasoning as well, and so his argument, if it is valid at all, should convince us that we are free – period. Kant shows convincingly that any deliberation on what to do or what to believe, and consequently any action which is taken according to prior deliberation, takes place under the presupposition that it is made

“Freedom must be presupposed as a property of the will of all rational beings”

267

according to the laws of reason. If you seriously doubt that in your reasoning you follow the laws of reason, you no longer believe in its validity. It is not that you must be rational for this. It is simply not possible to believe in a piece of reasoning and at the same time doubt its validity, and it is simply not possible to make a certain judgment and at the same time think it is unjustified. If one wants to be very cautious, one has to construe the mentioned presupposition negatively. Of course, a rational (in the sense of deliberating) being may never have entertained the thought that in deliberating it follows the laws of reason, so it need not be his explicit opinion that it does. That it has to be its implicit opinion at least, is very plausible, but it is not fully clear what “implicit” amounts to. Perhaps our rational being never reflects on its reasoning. So, maybe it need not believe positively (not even implicitly) that in deliberating it follows the laws of reason. But at least this much is true, that it cannot make any conflicting assumption, because that would amount to the judgment that the reasoning is invalid and the judgment emerging from it unjustified. And this much is true, that if a rational being reflects on its reflecting, it has to admit that it is committed to the assumption that in reasoning it follows the laws of reason. So, we have to consider ourselves to be subject to the laws of reason, because otherwise we could not claim the validity of our reasoning and judging, which we inevitably do when we reason and judge. Kant’s statement that “every being that cannot act otherwise than under the idea of freedom” in the indicated sense “is just because of that really free in a practical respect” (GMS, 448) does mean no more than that this being has to admit the validity of the laws of freedom (or reason) for itself. The phrase does not constitute an additional step in the argument. If Kant had given an argument that we are free, this would have meant that we are free theoretically, now he has given an argument to the effect that we have to consider ourselves to be free, and this means that we are free in a practical sense, since concerning the question which laws we have to accept as valid for us, everything is as it would have been had we been given a theoretical proof for freedom. The “practical respect” does not to refer to practical deliberations and actions as opposed to theoretical deliberations and beliefs, but to the kind of argument Kant has given for our freedom, an argument that is as valid for theoretical reasoning as it is for practical deliberations. 4. We have seen that it is freedom1, freedom as rationality, that we have to ascribe to ourselves according to Kant’s argument. In deliberating and judging, we have to suppose that we deliberate and judge ac-

268

Corinna Mieth and Jacob Rosenthal

cording to the laws of reason. Any other supposition would amount to the admission that our piece of reasoning is invalid and our judgment cannot be upheld. In particular, when we reason what to do, come to a conclusion and make a corresponding decision, we have to suppose that our will is determined by the laws of reason. This kind of freedom – freedom as being subject to the laws of reason – is obviously very different from another kind of freedom we might be interested in when we are acting: freedom as a genuine, unconditioned ability to choose between various options, to will otherwise as one actually does. This would be freedom5, freedom as indeterminism. In addition, the fact that “a reason […] is practical, that is, has causality with respect to its objects” (GMS, 448), i. e. freedom as spontaneity, does not imply that the being who is endowed with this kind of freedom could will and do otherwise. Quite the contrary. If in deliberating about what to do, you have to assume that you are rational, and if you additionally assume that there is only one rational thing to do, you have to conclude that there is only one action open to you. The assumption that there is only one rational thing to do is not always accurate, of course, but according to Kant it is at least accurate in any morally relevant situation in which you have a certain obligation. So, contrary to what one might think, neither freedom as spontaneity, nor freedom as autonomy, nor freedom as rationality have anything to do with freedom as the genuine possibility to will otherwise as one actually does. This does not mean that we always act rationally, and in particular morally, or that we have to assume that we do. We do not, but as far as we don’t, we are irrational. In these cases our reason does not have “causality with respect to its objects”, our act is instead determined by “alien causes”. This happens very often, and we know this. We make mistakes in our deliberations, or we arrive at a rational judgment and nevertheless do otherwise, i. e. perform a weak-willed act because of the interference of an inclination. As Kant thinks that morality is required by reason, every immoral act is also irrational, and must therefore be a weak-willed act or due to a faulty deliberation. Provided that your practical judgment determines what you do (and if it does not, practical deliberation does not make sense), and provided that there is only one rational course of action, you have to assume that there is no genuine choice about what to do, that your will is predictable, that there is only one action genuinely open to you. You just have to find out which it is. You have to assume all this, because in practical deliberating and willing accordingly you have to view yourself as rational. You might be wrong, of course, and this may be obvious from the per-

“Freedom must be presupposed as a property of the will of all rational beings”

269

spective of the third person but not from your own perspective at that moment. If you deliberate about what to do and arrive at a practical judgment, it is only your judgment if you claim validity for it, and this means that you think you have thereby found out what the reasonable thing to do is. We all know that we often act irrationally, and can judge our own past deliberations and practical judgments thus, but this means that they are no longer our judgments. We cannot take this stance towards our present and actual judgments. Freedom as indeterminism means that you can will and do one thing, but also another thing, in an absolute or unconditioned sense. You are neither determined nor predictable in your will and your action, and afterwards it is true that you could have willed and done otherwise. This is not the kind of freedom Kant is dealing with in our passage. If you are free in the sense of rational autonomy, your will is determined by the laws of reason and you think and act under their direction and according to them. So if there is only one reasonable option, you will choose this one, you cannot will otherwise, and what you will do is entirely predictable. In particular, if there is a choice between a morally right and a morally wrong action, you will choose the former. While freedom as indeterminism is not compatible with any determination whatsoever of the will, freedom in the sense of autonomy, on the contrary, assumes that the will is determined: determined by the laws of reason. These laws are self-given laws, to be sure, but the will is nevertheless determined by them, and this excludes the performance of any action but one, at least in situations with moral obligations. Therefore, if we take Kant’s claim seriously that a lawless will would be an absurdity (GMS, 446), there is no room for freedom as indeterminism. There are two possibilities: Either you will and act according to the laws of nature (which is heteronomy), or according to the laws of reason (which is autonomy), but in either case your will is determined and your actions occur with necessity. There is no third or meta-position from which you could deliberate and choose whether to follow your natural inclinations or your reason, because, if you really were in such a meta-position, you would be able to choose against your inclinations, i. e. your will would be free in the negative sense, and then it would, according to Kant, also be free in the positive sense, which means that you would choose according to the laws of reason. The supposed third or meta-standpoint collapses.10 If you try to imagine it, you realize that you in fact imagine the situation of autonomy. There is no 10

Cf. Timmermann (2004, 146).

270

Corinna Mieth and Jacob Rosenthal

meta-choice between autonomy and heteronomy, because there is no law according to which such a meta-choice could be made. It would have to be a lawless choice and so an absurdity. It may be due to this that the concept of decision (“Entscheidung”) plays no prominent role in Kant’s practical philosophy. Either your will is determined by natural causes according to the laws of nature, or by rational grounds according to the laws of reason. All this is difficult to reconcile with ideas about the connection between moral responsibility and indeterminism, ideas that at least to some extent seem to be shared by Kant. Concerning this, there are obvious tensions in his philosophy. In Religion within the Limits of Reason Alone, for example, he says: “The human being must make or have made himself into whatever he is or should become in a moral sense, good or evil. These two [characters] must be an effect of his free power of choice, for otherwise they could not be imputed to him and, consequently, he could be neither morally good nor evil“ (RGV, 44). On the other hand, he says: “Freedom does not consist in the fact that we might have preferred the opposite, but only in the fact that our preference was not necessitated passive” (Nachlass, 4226, AA XVII, 465). Or: „But freedom of choice cannot be defined – as some have tried to define it – as the ability to make a choice for or against the law (libertas indifferentiae), even though choice as a phenomenon provides frequent examples of this in experience.“ (MdSR, 226)11 Shortly after this, he says “that freedom can never be located in a rational subject’s being able to choose in opposition to his (lawgiving) reason, even though experience proves often enough that this happens (though we still cannot comprehend how this is possible).“ (MdSR, 226)12 The last passages are particularly important, because in it Kant admits the (incomprehensible) reality of what we call freedom as indeterminism, but he refuses to call it “freedom”. Sometimes the distinction between “Wille” (will) and “Willkür”13 (faculty of arbitrary choice), that Kant explicitly introduces in the 11

12

13

“Die Freiheit der Willkür aber kann nicht durch das Vermögen der Wahl, für oder wider das Gesetz zu handeln (libertas indifferentiae), definiert werden, wie es wohl einige versucht haben, obzwar die Willkür als Phänomen davon in der Erfahrung häufige Beispiele gibt.” (MdSR, 226) “dass die Freiheit nimmermehr darin gesetzt werden kann, dass das vernünftige Subjekt auch eine wider seine (gesetzgebende) Vernunft streitende Wahl treffen kann, wenn gleich die Erfahrung oft genug beweist, dass es geschieht (wovon wir doch die Möglichkeit nicht begreifen können).” (MdSR, 226) As e. g. Beck (1960, 177, note 1) points out, it is very hard to find an adequate English translation for “Willkür”. There exist as many different proposals as translations of Kant’s works. That is why we follow Beck’s solution and keep the German expression in the following. Mary Gregor’s translation is “power of choice”. In her

“Freedom must be presupposed as a property of the will of all rational beings”

271

Metaphysics of Morals, is invoked to elucidate his reasoning in GMS III.14 As Allison puts it, it is only through the distinction between Wille and Willkür that Kant can make plausible that we are affected by our inclinations, but not determined by them. The moral „ought to“ expresses that the Willkür should bring itself under the moral law given by the will in the sense of practical reason. According to Wimmer’s interpretation, the will is identified with practical reason, whereas “Willkür” denotes our freedom to choose, in particular whether to follow the moral law or not. This choice has to occur in an indeterminate (“arbitrary”) way. There can be no law under which it stands, because it is supposed to be a genuine choice between heteronomy and autonomy. So, Willkür would be just what we call freedom as indeterminism. This does not bear on Kant’s notion of freedom, however. In the Metaphysics of Morals, Kant first of all speaks generally of a “faculty of desire in accordance with concepts” (“Begehrungsvermögen nach Begriffen”) as a “faculty to do or to refrain from doing as one pleases” (“Vermögen, nach Belieben zu tun oder zu lassen”) (MdSR, 213). In the further explanation of this “faculty of desire”, he differentiates between Wille and Willkür, binding Willkür to “bring about its object by one’s action” (“Hervorbringung des Objekts”), whereas he combines Wille with reason. Willkür is our ability to make choices, whereas the will is identified with practical reason insofar as it can determine the Willkür (MdSR, 213). But Kant refuses to talk about “freier Wille” in MdS: The will “cannot be called either free or unfree” (MdSR, 226). Only the Willkür can be called free, but Kant states clearly that this “freedom of choice cannot be defined – as some have tried to define it – as the ability to make a choice for or against the law” (MdSR, 226, cf. the quotations above). So, Kant is not prepared to apply the notion of freedom to an indeterministic act of choice. Instead, he distinguishes a negative and a positive notion of “freie Willkür” that is exactly analogous to the distinction between freedom in the negative and freedom in the positive sense given in GMS III, 1: “Freedom of choice is this independence from being determined by sensible impulses; this is the negative concept of freedom. The positive concept of freedom is that of the

14

glossary (Gregor, 1996, 650), she refers to the latin difference between voluntas (Wille) and arbitrium (Willkür). What is meant by “Willkür” is in our opinion most adequately described by “faculty of arbitrary choice”, in order to emphasize the aspect of arbitrariness. Cf. Allison (1990, 225 f.); Beck (1960, 177 ff.); Wimmer (1980); Steigleder, (2002, 94 f., 109 ff.).

272

Corinna Mieth and Jacob Rosenthal

ability of pure reason to be of itself practical.”15 (MdSR, 213 f.). So, what Kant calls “freie Willkür” in the Metaphysics of Morals plays the same role as what he calls “free will” in GMS III, 1 and 2, and has the aspects of absence of heteronomy, autonomy, and spontaneity, but not of indeterminism. Whether one prefers to talk about “Wille” or “Willkür” – in neither case in the Groundwork nor in the Metaphysics of Morals does Kant apply the notion of freedom to a choice that is arbitrarily made. As far as his arguments on freedom are concerned, freedom as indeterminism is ruled out. To be sure, at least in the Metaphysics of Morals Kant admits the reality of what we call freedom as indeterminism. But, to repeat, he is not willing to address it as a kind of freedom. In addition, he says that we cannot comprehend its possibility. And rightly so, because such a choice, whether you call it free or not, runs into the very same problems as the meta-choice mentioned above. How is the choice made, how does it come about? It is clearly not simply made in a vacuum, but made according to the foregoing deliberations of the subject who has to decide. It is, after all, the subject’s choice. If we try to imagine this situation we find that we imagine a subject that is either guided by his reason or by his inclinations, and so the choice is not lawless after all. It cannot be lawless, because it has to be guided by or be due to something. A lawless choice is indeed incomprehensible; it would be guided by nothing at all and so amount to a random event that comes about without any explanation or justification. It is hard to see how an event of this kind could be attributed to the subject in such a way that the subject is morally responsible for it. But this is only an aside: what is important here is that when Kant applies the notion of freedom in either the Groundwork or the Metaphysics of Morals, he does not have such a lawless choice in mind. 5. Kant has proved that in deliberating we have to view ourselves as free1. It is a “practical proof” that we have freedom as rationality. But is it also a proof that we have freedom in the sense of the absence of alien determining causes, freedom as spontaneity, or freedom as autonomy? These are more difficult questions. In deliberating, we may not assume anything that conflicts with the supposition that we reason according to the laws of reason. Kant believes that it would be such a conflicting assumption that the reasoning occurs according to the laws of na15

“Die Freiheit der Willkür ist jene Unabhängigkeit ihrer Bestimmung durch sinnliche Antriebe; dies ist der negative Begriff derselben. Der positive ist: das Vermögen der reinen Vernunft, für sich selbst praktisch zu sein.” (MdSR, 213 f.)

“Freedom must be presupposed as a property of the will of all rational beings”

273

ture. It cannot, according to him, be subject to the laws of reason and the laws of nature, at least not in any straightforward sense: “It would, however, be impossible to escape this contradiction if the subject who seems to himself free thought of himself in the same sense or in the very same relation when he calls himself free as when he takes himself to be subject to the law of nature with regard to the same action.” (GMS, 456) This is very problematic. After all, why couldn’t laws of reason and laws of nature simply operate on different levels? Why shouldn’t nature have built us in such a way that we deliberate according to the laws of reason (at least sometimes)? To make a somewhat crude comparison: Clearly, an electronic (or, in Kant’s time, a mechanical) calculator operates according to the laws of nature, but also according to the laws of mathematics. It is built that way. As the laws of nature and the laws of mathematics are totally different kinds of laws, there is no reason why they should automatically stand in conflict with each other, so that nothing could be subject to or follow both of them. They do not necessarily compete with each other. The same holds for the contrast between laws of nature and laws of reason. One undoubtedly has to assume that in one’s reasoning and judgments one follows the laws of reason, but this does not automatically exclude the view that these activities are also determined by the laws of nature. This would need an additional argument that is not provided by Kant. He has not proved that we have to view ourselves as free in the negative sense. Sometimes the contrast between being determined by nature and being determined by reason is put thus: Theoretical as well as practical judgments (and consequently rational actions) are made for reasons, whereas in nature there are only causes. The determinist, who thinks that everything he says and does is pre-determined by natural causes according to the laws of nature, views the formation of his judgments and actions as mere events. So, he cannot claim validity for his judgments, or rather: If he was right, he could not make any judgments at all. Thus, the determinist thesis is self-refuting. If you try to support it, you imply that all you think and say is thought and said not for reasons, but merely because certain natural causes are operating on you, and consequently you can no longer claim validity for what you say. If you do – and you must do it, if you want to engage in a discussion and put forward a certain thesis – you contradict yourself in an outright way.16 This argument rests on a misleading dichotomy between reasons and causes. When the determinist argues for his the16

See, e. g., Schönecker (1999, ch. 4.2.2).

274

Corinna Mieth and Jacob Rosenthal

sis, he of course claims to do so for reasons. And he also claims (by his very thesis) that there are natural causes that made him have these opinions, make these statements etc. But this is no contradiction. His grasping of the reasons is the cause for what he thinks and says, this grasping in turn has other causes, and so on, arbitrarily far into the past. The determinist has to presuppose that nature has built him such that he is able to understand arguments and judge them in the right way (most of the time, at least as far as the present discussion is concerned). And he has the hope that his opponent, the indeterminist, is structered alike, is structured such that when confronted with his, the determinist’s, arguments, he recognizes their validity. This expectation or hope may prove futile or not, but there is nothing in the whole szenario that would rule out the causal pre-determination of all the mental processes involved. There are obviously several preconditions that must be fulfilled before one can comprehend an argument, judge its validity etc. A certain talent is needed, a certain upbringing and education, and the like. Our understanding of arguments develops in time. It is not clear if one could in principle spell out a full causal story as to why a certain being at a certain point in time understands a certain argument, but still less is it clear why such a story should be impossible. One’s grasping of certain reasons could simply be part of the course of nature. A closer look reveals that we have to distinguish between two issues here. One is causal determinism: Are there sufficient causes for the fact that a certain being at a certain time makes a certain judgment? Are there full causal explanations for mental states? If that should be the case, the mental states (opinions, judgments, decisions etc.) of the being would be pre-determined by causes arbitrarily far in the past, they would in principle be predictable, and the being would not have freedom as spontaneity. Another, further issue is naturalism: Even if there are sufficient causes for mental states, can these causes be addressed as or somehow reduced to natural causes? If they could, the being would not have negative freedom. The point is trickier, because one might concede that one’s grasping of a certain reason is the cause for making a certain judgment, but deny that this cause can be viewed as or be reduced to a natural cause that operates according to the laws of nature. We can dimly imagine a full causal story as to why someone makes a certain judgment or comprehends a certain argument. That story would be a psychological story, by and large, but it is doubtful if it could be called naturalistic, or could in principle be reduced to, or replaced by, a naturalistic one. This is because a psy-

“Freedom must be presupposed as a property of the will of all rational beings”

275

chological story for the most part consists of ascriptions of intentional states that have, at least prima facie, normative implications, whereas a naturalistic story should be purely descriptive. In our dimly imaginable causal story as to why a certain being at a certain moment has a certain insight, the relata of the causal relationship are events or states that are not obviously naturalistically acceptable. The issue draws on the concept of nature as well as on the concept of reduction and cannot be settled without addressing these. But for Kant, these two issues are not distinguished. This is already clear from the fact that freedom in the negative sense (absence of natural determining causes) does imply for him positive freedom, i. e. spontaneity and autonomy, and consequently the absence of any pre-determining causes. Very clear on the whole issue is also the following passage from the Critique of Practical Reason: “That is to say, in the question about that freedom which must be put at the basis of all moral laws and the imputation appropriate to them, it does not matter whether the causality determined in accordance with a natural law is necessary through determining grounds lying within the subject or outside him, or in the first case whether these determining grounds are instinctive or thought by reason, if […] these determining representations have the ground of their existence in time and indeed in the antecedent state, and this in turn in a preceding state, and so forth, these determinations may be internal and they may have psychological instead of mechanical causality, that is, produce actions by means of representations and not by bodily movements; they are always determining grounds of the causality of a being in so far as its existence is determinable in time and therefore under the necessitating conditions of past time, which are thus, when the subject is to act, no longer within his control and which may therefore bring with them psychological freedom (if one wants to use this term for a merely internal chain of representations in the soul) but nevertheless natural necessity; and they therefore leave no transcendental freedom […]. Just for this reason, all necessity of events in time in accordance with the natural law of causality can be called the mechanism of nature, although it is not meant by this that the things which are subject to it must be really material machines. Here one looks only to the necessity of the connection of events in a time series as it develops in accordance with natural law, whether the subject in which this development takes place is called automaton materiale, when the machinery is driven by matter, or with Leibniz spirituale, when it is driven by representations; and if the freedom of our will were non other than the latter […] then

276

Corinna Mieth and Jacob Rosenthal

it would at bottom be nothing better than the freedom of a turnspit, which, when once it is wound up, also accomplishes its movements of itself.” (KpV, 96 f.) So, even if the “determining grounds are […] thought by reason”, as is the case when a subject is built by nature so as to comprehend certain arguments, the subject is not free in the negative sense, because his thoughts are part of a causal chain. How the links of the chain look like does not matter for Kant: When the chain is there, the subject is determined by the laws of nature. For Kant, the relation of cause to effect constitutes a natural necessity in any case, and consequently there is no place in his conception for a causal pre-determination that is not naturalistic. The causal pre-determination as such essentially reduces a being to the status of a turnspit: like the turnspit, it is a mere automaton. But Kant’s argument in our passage does only give us freedom as rationality. It is not shown, although intended by Kant, that the being cannot at the same time view itself as being subject to the laws of nature, or that the being has to look upon itself as spontaneous. There is no successful argument in GMS III as to why rationality should be incompatible with either determinism or naturalism. Kant himself concedes, in the passage quoted above, that determining grounds can drive a subject through representations thought by reason, and we have no reason to think that this is not our situation. We are given no reason to think that we are no automata spirituale. Kant’s true motive to discard this possibility is that he thinks it excludes moral responsibility. And in the Critique of Practical Reason he indeed argues from the reality of the moral law to the reality of transcendental freedom (which is the same as freedom as spontaneity). But in the Groundwork he wants to have it the other way round: He wants to prove the moral law as binding for our wills by establishing transcendental freedom, at least in a practical sense. But this kind of freedom does not come out of his argument. At the end we will discuss in how far or in what sense the argument is nevertheless successful. Kant repeatedly emphasizes that every action may on the one hand be viewed as an event determined by the laws of nature, as a link in a causal chain, and on the other hand as spontaneously caused by the agent (Prol, § 53; GMS, 455 ff.; KpV, 94 ff.; KrV, A 538 ff., B 566 ff.). Both perspectives seem indispensable to him, the former to guarantee the unity of appearances, the latter as a necessary condition for moral responsibility. He claims that these perspectives and the corresponding assertions can be reconciled: The former concerns the phenomenal, the latter the noumenal world. So he is a compatibilist in a certain sense.

“Freedom must be presupposed as a property of the will of all rational beings”

277

But his brand of compatibilism requires the distinction between phenomenal and noumenal world, and he leaves no doubt that there is no simpler or more down-to-earth solution to be had. He is an incompatibilist insofar as he is convinced that the reconciliation of the two perspectives – an agent standing under the laws of nature, on the one hand, and being free, on the other hand – requires such extreme means. They certainly are extreme: an action may be viewed as a natural event and is as such fully determined by events arbitrarily in the far past, according to the laws of nature. It could thus in principle have been predicted with complete certainty at any particular time prior to its occurrence. The very same event, however, viewed as an action, is spontaneously caused by an agent who thereby starts a new causal chain, and this is supposed to be no contradiction, because the latter is due to the agent as a thing in itself and somehow takes place outside of space and time. This is hardly credible – first of all, the action, the decision to act, and the foregoing deliberations of the agent that certainly somehow influence his decision, occupy a certain stretch in time. Second, the notion of causation presupposes that of time, so causal relations can only exist within the (spatio-)temporal world, and not between the noumenal and the phenomenal world. Third, if the action, viewed as a natural event, is in fact determined by the laws of nature, it would take place independently of what the noumenal self of the actor did. “If by ‘freedom’ we mean noumenal causation and assert that we know no noumena, then there is no justifiable way, in the study of phenomena, to decide that it is permissible in application to some but not others of them to use the concept of freedom. The uniformity of human actions is, in principle, as great as that of the solar system; there is no reason to regard statements about the freedom of the former as having any empirical consequences. If the possession of noumenal freedom makes a difference to the uniformity of nature, then there is no uniformity; if it does not, to call it ‘freedom’ is a vain pretension.” (Beck, 1960, 192) Or, as it is described by Bennett, analyzing the compatibilism Kant introduces in the Critique of Pure Reason: “Kant wants the so-called causality of freedom to be a ‘power of originating a series of events’ (B 582). That requires that freedom leave its mark on the world of events, making a difference to what occurs in that world, and that in turn implies that natural causality cannot entirely determine which events occur. We repeatedly fi nd Kant trying to have it both ways: freedom affects the world of events, and yet what happens in that world is just what would have happened if there had been only natural causality.” (Bennett, 1974, 200)

278

Corinna Mieth and Jacob Rosenthal

But inconceivable or not, the demanding kind of resolution to this complex of problems proposed by Kant is certainly not required by the kind of freedom that is proved in our passage. To repeat, we can view ourselves as being completely determined by the laws of nature and nevertheless claim validity for our reasoning and judgments and rationality for our actions, i. e. freedom in the sense of rationality. We just have to assume that nature has created us as at least partially rational beings, in such a way that when we are confronted with or consider an argument we most of the time correctly assess its validity or invalidity. This thought should not be too alien to Kant. After all, he considers and rejects the possibility that nature has endowed man with reason only to secure his happiness (GMS, 395), and states that in fact nature has given reason to man for the purpose of producing a good will: “[…] then, where nature has everywhere else gone to work purposively in distributing its capacities, the true vocation of reason must be to produce a will that is good […]” (GMS, 396).17 If nature had built us such that we follow the laws of reason when we deliberate, these laws would, in a sense, be no self-given laws. Instead, they would be imposed on us by nature, a situation which could be called heteronomy. In addition, we would not be the origin of causal chains, so we would not have freedom in the sense of spontaneity. This is the situation that Kant polemically describes as the “freedom of a turnspit” in KpV, 97. It is clear that Kant’s talk about autonomy, self-given laws, causality of practical reason with respect to its objects etc. is meant to exclude determination by natural causes according to the laws of nature. But his argument in our passage does not show that we have to view ourselves as free or autonomous in this sense. That our reasoning is subject to the laws of reason does not imply that there is not also an explanation of it in terms of laws of nature. Why should there be any necessary conflict? We have to assume that in our actual deliberation we follow the laws of reason. We don’t have to assume anything whatsoever about why we follow them, or whether there are other laws to which we are at the same time also subject. Furthermore, we do not have to presuppose that we are the origins of causal chains. The causal chain that leads to the action may begin arbitrarily far in the past and run through our practical deliberations that are embedded as links within it. So, freedom1 (rationality), which is in fact proved by Kant, does neither imply freedom2 (negative freedom) nor freedom3 (spontaneity). The latter kind of freedom is, according to 17

Cf. the teleological arguments in Toward Perpetual Peace.

“Freedom must be presupposed as a property of the will of all rational beings”

279

Kant, required to secure moral responsibility, but it is not established by his argument in GMS III, 2. Our diagnosis may be corroborated by looking at Kant’s SchulzReview he published two years before the Groundwork. Johann Heinrich Schulz had written an “Attempt at an Introduction to a Doctrine of Morals”, in which he advocates the complete causal pre-determination of all human actions. Kant calls this a “general fatalism” that “turns all human conduct into a mere puppet show” and has the consequence that “what is left us is only to await and observe what sort of decisions God will effect in us by means of natural causes, but not what we can and ought to do of ourselves, as authors.” (AA VIII, 13) This, Kant claims, is a self-refuting thesis, as: “Although he [Schulz] would not himself 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.” (AA VIII, 14) Again, this line of argument is only convincing if one already presupposes what Kant purports to establish by it. The idea that if we were causally predetermined, we could only “await and observe what sort of decisions are effected in us by natural causes” simply assumes that our (practical) deliberations that lead to our decisions are no part of the course of nature. The idea that if Schulz “is able to determine his judgment in accordance with objective grounds”, he cannot at the same time be “subject to the mechanism of merely subjectively determining causes”, simply assumes incompatibilism instead of justifying it. He who argues has to presuppose “freedom to think” in the sense of “determination of his judgment in accordance with objective grounds” – this much is clear. But Schulz need not at all deny that, he just has to say that his judgment is also determined by natural causes according to the laws of nature, because nature has created himself as a (partially, at least as far as the present argument is concerned) rational being, as a (partially) rational automaton spirituale, if one wants to put it that way. That “subjectively determining causes could subsequently change”, whereas “objective grounds are always valid”, is besides the point. We know very well that we are not always in the state of mind to judge the validity of arguments correctly, e. g. when we are tired or drunk or distracted, when we were children or when we become feeble-minded in old age. We know that the “subjectively determining causes” that make us grasp the “objective grounds that are always

280

Corinna Mieth and Jacob Rosenthal

valid” can and will subsequently change. So, again, the “freedom to think” that we have to claim for ourselves when we deliberate does only amount to freedom as rationality, but does not imply negative freedom or freedom as spontaneity. The case of freedom4 (autonomy) is most difficult to judge. In a weak sense, it is implied by Kant’s argument. In deliberating and judging, we not only have to assume that we follow whatever laws of reason there are, but also approve of our being subject to them. We could not wish to violate them while deliberating, because this would render our conclusions unjustified. In this sense the laws of reason are self-given. They need not be self-given in the sense of literally being chosen by ourselves, but we can certainly think of them as self-given. Upon reflection we realize that if we had the choice we would (have to) impose these laws on ourselves, or negatively, that we could not wish to get rid of them. But this sense of autonomy does not imply freedom2 or freedom3. We can further clarify this weak meaning of “autonomy” if we ask why the laws of nature are “alien laws”. Insofar as we are subject to them, we have, of course, not actually chosen to be so. But as we have seen, this may also be the case with the laws of reason. The crucial difference is that we are not bound to approve of our being subject to the laws of nature. In deliberating, we need not identify with nature and her laws, but we have to identify with reason and its laws. It is this point that is successfully made by Kant’s argument. One might agree with this and nevertheless object to the characterization of the laws of nature as “alien”. Insofar as we are deliberating beings, we have to identify with reason and its laws – this we can concede. But are we not also sensible beings? And insofar as we are such beings, don’t we have to view ourselves as part of nature and identify with her laws? Doesn’t Kant beg the question if he calls the laws of nature “alien” and the laws of reason “self-given”? Why are we entitled to prefer the view of ourselves as rational beings to the view of ourselves as sensible beings? These questions point to the reason why Kant has not reached his goal within the text passage under discussion, even if his argument here were to be completely successful. They are answered in the further parts of Section III by an appeal to the distinction between phenomenal and noumenal world, by attributing the laws of nature to the former and the laws of reason to the latter, and by claiming priority of the latter because “the world of understanding contains the ground of the world of sense” (GMS, 453), and therefore, “what belongs to mere appearance is necessarily subordinated by reason to the constitution of the thing in itself” (GMS, 461).

“Freedom must be presupposed as a property of the will of all rational beings”

281

Perhaps there is a more straightforward argument as to why the view of ourselves as rational beings can claim priority over the view of ourselves as sensible beings. We have to admit this order, because to deny or to question it would mean to make a consideration, and insofar we do so, we have to identify with reason and its laws. If we claimed priority for the sensible-being-view, we would argue with reason against reason, and thus contradict ourselves. This would be a kind of transcendental argument that did not draw on the distinction between noumenal and phenomenal world. It would be worth trying, and one could almost expect Kant to make such a move after his argumentation in the passage under discussion. But he does not, maybe because the success of the idea is doubtful: One would have to supply an argument why, when we engage in deliberation and consequently accept our being subject to the laws of reason, our reasoning could not lead to the result that there are other laws that could claim priority – at least most of the time, namely as soon as this very deliberation is over. Reason, it seems, might very well lead to the result that it is sometimes (or often, or nearly always) better not to be reasonable, the only necessary exception being the deliberation by which this very result is established. Maybe this reply can be successfully countered, but anyway, Kant does not take this line of reasoning. To conclude: If one starts with the notion of negative freedom (freedom 2), as Kant does, freedom as spontaneity and freedom as autonomy in a strong sense (freedom3 and freedom4) follow from this, provided that one follows Kant that the absence of alien determining causes implies the absence of any causal pre-determination whatsoever, and provided that the whole idea of spontaneous selfdetermination according to self-given laws is a conceivable one. If we further notice that in deliberating we have to view ourselves as reasonable, we realize that freedom as autonomy implies freedom as rationality. The self-given laws must be the laws of reason, and so we also have freedom1. So, if Kant had given us an argument that we have to view ourselves as free2, his line of reasoning would have been successful. But he did not do this. His argument shows that we have to view ourselves as free1, but from this the other kinds of freedom do not follow. Certainly it does not follow that we are free2 or free3. It does follow that we are free4, in a sense, but this sense of autonomy is a weak one that does neither imply that we can start causal chains by ourselves, nor that we are not completely determined by natural causes according to the laws of nature. As it does not imply freedom

282

Corinna Mieth and Jacob Rosenthal

as spontaneity, it is not enough to secure our moral responsibility according to Kant. If autonomy amounts to the simple claim that in theoretical and practical deliberations we have to accept and approve of the laws of reason and assume that we are guided by them, we could be free positively without being free negatively. This sounds paradoxical, and so confirms again that Kant had more in mind than this kind of autonomy. But all this does not mean that Kant’s argument in GMS III, 2 is a complete failure. To be sure, we have to view ourselves as being subject to the laws of reason. In deliberating, we have to presuppose that we follow them, and upon reflection, we have to approve of this situation. Insofar they are self-given. We not only have to assume that we are subject to the laws of reason, but we have to endorse them reflectively. This is a weak sense of autonomy, but it constitutes a view we necessarily take of ourselves and our relation to reason, and thus, if we accept the connection between autonomy and the moral law that is established in GMS III, 1, Kant has shown successfully that we have to view ourselves as being subject to the moral law. We have to accept the moral law as a law for our wills. But in this line of reasoning, freedom as the absence of natural determination or as spontaneity or as autonomy in a strong sense plays no role. Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–) (abbreviated as ‘AA’). The Critique of Pure Reason will be cited according to the A/B pagination from the first and second editions. English translations, except where otherwise noted, are from Groundwork of the Metaphysics of Morals, translated and edited by Mary Gregor, Cambridge: Cambridge University Press 1997, or from The Cambridge Edition of the Works of Immanuel Kant – Practical Philosophy, translated and edited by Mary Gregor, Cambridge: Cambridge University Press 1996. Ge Gemeinspruch, AA, VIII GMS Grundlegung zur Metaphysik der Sitten, AA, IV KdU Kritik der Urteilskraft, AA, V KpV Kritik der praktischen Vernunft, AA, V KrV Kritik der reinen Vernunft, AA, III, IV MdSR Metaphysik der Sitten – Rechtslehre, AA, VI Prol Prolegomena, AA, IV RGV Die Religion innerhalb der Grenzen der bloßen Vernunft, AA, VI

“Freedom must be presupposed as a property of the will of all rational beings”

283

Other works Allison, Henry E. (1990): Kant’s Theory of Freedom, Cambridge: Cambridge University Press. Allison, Henry E. (1993): “Kant’s Preparatory Argument in Grundlegung III”, in: Höffe, Otfried (ed.): Grundlegung zur Metaphysik der Sitten. Ein kooperativer Kommentar, Frankfurt a. M.: Klostermann, 314–324. Allison, Henry E. (2002): “Morality and Freedom: Kant’s Reciprocity Thesis”, in: Pasternack, Lawrence (ed.): Immanuel Kant – Groundwork of Metaphysics of Morals in Focus, London: Routledge, 182–210. Beck, Lewis White (1960): A Commentary on Kant’s Critique of Practical Reason, Chicago: University of Chicago Press. Beck, Lewis White (1975): The Actor and the Spectator, New Haven/London: Yale University Press. Bennett, Jonathan (1974): Kant’s Dialectic, Cambridge: Cambridge University Press. Eisler, Rudolf (1994): Kant-Lexikon. Nachschlagewerk zu Kants sämtlichen Schriften, Briefen und handschriftlichem Nachlaß, Hildesheim u. a.: Olms. Henrich, Dieter (1975): “Die Deduktion des Sittengesetzes. Über die Gründe der Dunkelheit des letzten Abschnittes von Kants ‘Grundlegung zur Metaphysik der Sitten’”, in: Schwan, Alexander (ed.): Denken im Schatten des Nihilismus. Festschrift für Wilhelm Weischedel zum 70. Geburtstag, Darmstadt, 55–112. Korsgaard, Christine M. (1996): “Morality as freedom”, in: Korsgaard, Christine M.: Creating the Kingdom of Ends, Cambridge: Cambridge University Press, 159–187. Korsgaard, Christine M. (1998): The Sources of Normativity, Cambridge: Cambridge University Press. O’Neill, Onora (1989): “Reason and Autonomy in Grundlegung III”, in: O’Neill, Onora: Constructions of Reason. Explorations of Kant’s Practical Philosophy, Cambridge: Cambridge University Press, 51–65. Prauss, Gerold (1993): “Für sich selber praktische Vernunft”, in: Höffe, Otfried (ed.): Grundlegung zur Metaphysik der Sitten. Ein kooperativer Kommentar, Frankfurt a. M.: Klostermann, 253–263. Schönecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs, Freiburg i. Br./München: Alber. Schönecker, Dieter / Wood, Allen (22004): Kants Grundlegung zur Metaphysik der Sitten. Ein einführender Kommentar, Paderborn: Schöningh (UTB). Schneewind, Jerome B. (1992): “Autonomy, obligation, virtue”, in: Guyer, Paul (ed.): The Cambridge Companion to Kant, Cambridge: Cambridge University Press, 309–341. Steigleder, Klaus (2002): Kants Moralphilosophie. Die Selbstbezüglichkeit reiner praktischer Vernunft, Stuttgart: Metzler. Sturma, Dieter (2004): “Kants Ethik der Autonomie”, in: Ameriks, Karl / Sturma, Dieter (ed.): Kants Ethik, Paderborn: mentis, 160–177. Timmermann, Jens (2004): “Kommentar”, in: Jens Timmermann (ed.): Immanuel Kant. Grundlegung zur Metaphysik der Sitten, Göttingen: Vandenhoeck & Ruprecht, 84–158.

284

Corinna Mieth and Jacob Rosenthal

Wimmer, Reiner (1980): “Praktische Vernunft und Freiheit. Kants doppelter Freiheitsbegriff als Voraussetzung für die Doppelung von Recht und Moral”, in: Wimmer, Reiner: Universalisierung in der Ethik. Analyse, Kritik und Rekonstruktion ethischer Rationalitätsansprüche, Frankfurt a. M.: Suhrkamp, 124–144.

Marcel Quarfood

The Circle and the Two Standpoints (GMS III,3) At the end of Section II of the Groundwork, the content and presuppositions of “the generally received concept of morality” (GMS, 445) have been brought out by means of analysis. One result of the analysis is that the concept of morality is grounded on that of autonomy of the will. Nothing has yet been done to establish that this concept of morality is not a mere “phantom” (GMS, 445). Hence the business of Section III will be to demonstrate its validity. In the fi rst subsection of GMS III, autonomy is shown to be analytically linked to freedom. The second subsection goes further, arguing that all rational beings must presuppose that their own acts of judgment are free, on pain of not being able to count them as their own, hence not to count them as judgments. Kant also states that it is enough for present purposes that the freedom in question is taken in the practical respect, so that we avoid having to prove freedom in the theoretical respect as well (GMS, 448). It thus appears that the demonstration of the validity of the concept of morality will be fairly straightforward. If judging presupposes freedom, and freedom implies autonomy, which in its turn is the basis of morality, then our mere capacity to judge should be sufficient for the validity of the moral principle to follow. This line of reasoning, which establishes morality by means of a first step exploiting the analytic connection between freedom and morality and a second step referring to the necessity for rational beings to regard themselves as free, I will call the straightforward argument. In the third subsection to Section III of the Groundwork, Kant somewhat unexpectedly raises a difficulty, declaring that “a kind of circle comes to light here from which, as it seems, there is no way to escape” (GMS, 450). The circle, which concerns the reciprocal connection between freedom and the moral principle, somehow threatens the straightforward argument sketched above. But it is not immediately clear why the reciprocity between these concepts should be considered a problem; on the contrary, reciprocity is what makes it

286

Marcel Quarfood

possible to proceed along the lines of the straightforward argument. Moreover, whatever one thinks of its cogency, Kant’s reasoning up to this point does not seem to be circular. Kant’s comments at the end of the third subsection bring some further information about the circle. He now claims to have removed “the suspicion that a hidden circle was contained in our inference from freedom to autonomy and from the latter to the moral law” (GMS, 453). So it seems that no circular reasoning has taken place after all, but that there was an appearance of circularity that had to be removed. Still, it is not clear for the reader that a circle even appears to be hidden in Kant’s argument. Other questions are raised by a closer look at the structure of the circle as it is formulated in GMS, 450, and a second time in GMS, 453. It has been pointed out that the sentences in which Kant presents what he refers to as a circle do not describe any strictly circular argument, but rather an argument with a premise that is insufficiently justified.1 If that is so, is it not misleading to call it a circle? Kant’s solution to the circle problem, developed in GMS, 451 f., consists in introducing a central tenet of transcendental idealism, the distinction between appearances and things in themselves. As applied to human beings, this distinction provides us with two standpoints. Kant attempts to show that we necessarily regard ourselves as belonging to the world of sense as well as to the world of understanding (Verstandeswelt). This raises questions as to the relation of the straightforward argument to the doctrine of transcendental idealism. Transcendental idealism appears to be required to avoid the circle (irrespective of its exact structure and of whether it is a circle actually committed or just an apparent one), but now it seems as if this might affect Kant’s claim that the straightforward argument does not need to prove freedom theoretically: is not transcendental idealism a result from theoretical philosophy? And moreover, some of Kant’s statements, for instance that “we transfer ourselves into the world of understanding as members of it” (GMS, 453), are rather bold ones to make for a philosopher who wants to limit claims of cognition to the sphere of experience. 2 Yet another problem is raised when one considers the context in which the circle is placed. The third subsection is presented as a discussion of how we can take an interest in morality. Immediately before the introduction of the circle a question is posed concerning how the 1 2

Cf. Schönecker (1999, ch. 5.2). Cf. Allison (1990, 227 f.).

The Circle and the Two Standpoints

287

moral law can obligate, and why we ought to “detach ourselves from all empirical interest” (GMS, 450). The solution to the circle problem should presumably also aid in answering this question. The present essay will attempt to cast some light on the problems of interpretation outlined above. It proceeds as follows: In part 1 the structure of the circle is studied in order to determine what kind of circle we are dealing with. Part 2 considers the circle problem in relation to Kant’s other concerns in GMS III, with the aim to see what an escape from the circle is supposed to accomplish. Whether a circle actually has been committed in Kant’s argument is also discussed here. Part 3 examines how transcendental idealism is introduced and in what way it solves the circle problem. As might be expected, not all difficulties will disappear. In particular, it will remain undecided whether the assumption of a world of the understanding is a step that abandons the practical point of view and attempts a theoretical proof, thereby perhaps even transcending the boundaries of cognition.

1. Is the circle a circle? Kant explains in the following way what the circle consists in: We take ourselves as free in the order of efficient causes in order to think ourselves under moral laws in the order of ends; and we afterwards think ourselves as subject to these laws because we have ascribed to ourselves freedom of will: for, freedom and the will’s own lawgiving are both autonomy and hence reciprocal concepts, and for this very reason one cannot be used to explain the other or to furnish a ground for it but can at most be used only for the logical purpose of reducing apparently different representations of the same object to one single concept (as different fractions of equal value are reduced to their lowest expression). (GMS, 450)

This long sentence is made up of three parts. The first part describes the circle as such. The second part (‘for, freedom and the will’s own lawgiving are both autonomy and hence reciprocal concepts’) identifies the principle on which the circular reasoning hinges, and at the same time prepares for the third and final part of the sentence, in which the circular reasoning is criticized. Thus, the circle as such is this reasoning, which I shall call the first circle passage: ‘We take ourselves as free in the order of efficient causes in order to think ourselves under moral laws in the order of ends; and we afterwards think ourselves as subject to these laws because we have ascribed to ourselves freedom

288

Marcel Quarfood

of will.’3 As we know that this is supposed to be a piece of circular reasoning, there is a strong pull to read the two parts of this passage as arguing in opposite directions, roughly along the following lines: (1) We take ourselves as free because we think ourselves under moral laws. (2) We think ourselves under moral laws because we have taken ourselves as free.

Simplifying a little bit more, it seems that (1) is a derivation of freedom from morality, whereas (2) is a derivation of morality from freedom. On this construal we certainly have a piece of circular reasoning. In view of the expectation to find a logical circle, this way of interpreting the passage is natural and it is not uncommon, but it is hardly adequate to the text.4 ‘We take ourselves as free […] in order to [um] think ourselves under moral laws’ should rather be taken to mean that freedom is adopted as a premise for deriving morality. By taking ourselves as free we can reach the conclusion that we stand under moral laws. A problem with this reading, however, is that now we do not get a genuine circle, but only a linear derivation starting from the assumption of freedom. And this should be a bit worrying, as Kant after all calls it a circle. I shall return to this issue below. The need to find a circle drives one of Kant’s earliest commentators to offer another reading of the same kind, but this time reversed. In an exposition of the Groundwork from 1788, the anonymous author (writing under the signature T-h) cites the first circle passage and then says: “thus it appears that from freedom we infer autonomy, and then from autonomy in turn we infer freedom.”5 Quite plausibly, T-h took the beginning of the passage to state that we assume freedom as a premise from which we reach the conclusion that we are autonomous.6 But then, in order to obtain a circle, T-h had to construe the second half of the passage as describing an inference from autonomy to freedom. Kant, on the other hand, states that we ‘think ourselves as 3

4

5

6

The German reads: “Wir nehmen uns in der Ordnung der wirkenden Ursachen als frei an, um uns in der Ordnung der Zwecke unter sittlichen Gesetzen zu denken, und wir denken uns nachher als diesen Gesetzen unterworfen, weil wir uns die Freiheit des Willens beigelegt haben”. For a recent example of such an interpretation, see Kim (2002, 71). Schönecker (1999, 349 f.) refers to several other instances. My treatment of this issue owes much to Schönecker (1999). T-h (1788, 271): “wir scheinen also aus der Freiheit auf die Autonomie, und dann wieder aus der Autonomie auf die Freiheit zu schliessen.” In the present context it does not matter that T-h construes the argument in terms of autonomy rather than morality.

The Circle and the Two Standpoints

289

subjects to these laws because we have ascribed to ourselves freedom,’ which can hardly be interpreted otherwise than that the inference described has freedom as premise rather than as conclusion. We cannot yet conclude that these interpretations are definitely wrong, however. Perhaps there is a wider context that can justify a reconstruction of the text in terms of a circular argument structure. But at least when looking closely at the first circle passage, we do not fi nd a circle in the strict sense. Let us now see how Kant’s second circle passage is constructed. It occurs at the end of the subsection Of the Interest Attaching to the Ideas of Morality. Just before, Kant claims to have removed “the suspicion that a hidden circle was contained in our inference from freedom to autonomy and from the latter to the moral law” (GMS, 453). Then comes the second circle passage, which specifies what kind of circle was suspected to be hidden in the straightforward argument: namely that we perhaps took as a ground the idea of freedom only for the sake of the moral law, so that we could afterwards infer the latter in turn [wiederum] from freedom. (GMS, 453)7

Perhaps clearer than in the corresponding part of the first circle passage, here it is stated that the idea of freedom is taken as a ground, that is, a premise. The last part of the passage says that once freedom is assumed, we can infer the moral law.8 It seems that here, like in the first circle passage, we have a non-circular argument starting from the assumption of freedom and having morality as its consequence. Kant goes on to explain what kind of mistake we would have made had we argued in the manner described in the circle passages. It was suspected that we were “unable to furnish any ground at all for the moral law but could put it forward only as a petitio principii [Erbittung eines Prinzips] disposed souls would gladly grant us, but never as a demonstrable proposition” (GMS, 453).9 We would thus have put 7

8

9

The German reads: “daß wir nämlich vielleicht die Idee der Freiheit nur um des sittlichen Gesetzes willen zum Grunde legten, um dieses nachher aus der Freiheit wiederum zu schließen.” The words “in turn” might be taken to indicate a reversal of argumentative order, suggesting a genuine circle after all. Perhaps Allen Wood’s translation (Kant 2002) of wiederum as “once again” gives a slightly different tone. Alternatively, the translation could be “unable to furnish any ground at all for our inference.” Compare Wood’s rendering (Kant 2002): “we could not offer any ground for the former.” In its context “the former” (jenem) might refer to “our inference” (unserem Schlusse). Here is the complete German sentence: “Nun ist der Verdacht, den wir oben rege machten, gehoben, als wäre ein geheimer Zirkel in unserem Schlusse aus der Freiheit auf die Autonomie und aus dieser aufs sittliche Gesetz

290

Marcel Quarfood

forward a petitio principii. Though the term is often used for circular arguments in general, “petitio principii” more specifically designates an assumption made without other ground than that the assumption can be used as premise for inferring a desired conclusion. A petitio understood in this way amounts to an ad hoc assumption. So, for instance, if someone asks me why I think the Big Bang theory is false, and I offer as reason that the universe is eternal, then in making this assumption I put forward a petitio principii, unless either a case can be made for the self-evidence of the assumption or independent grounds are given for it. The ground I offer (the eternity of the universe) is a petitio, and so my claim that the Big Bang theory is false can also be called a petitio. In this sense, the straightforward argument is a petitio principii. In “our inference from freedom to autonomy and from the latter to the moral law” (GMS, 453), the assumption of freedom has not been given any ground, and it seems to be made only for the sake of inferring the moral law, which therefore in its turn is groundless too. Now, it may seem somewhat implausible that Kant should reach this conclusion, considering that he has argued in GMS, 448, that a rational being must view its own judging as free. Is this not a justification for assuming freedom (at least in the practical respect), so that the premise is not simply put forward without ground? We will return to this question in the next section; before that, an explanation is needed for why Kant talks of a circle if he only means a petitio (in the sense related to ad hoc). Kant’s choice of word is not necessarily misleading. The Jäsche Logik (Log, § 92, 135) distinguishes two kinds of fallacies in the same short section, obviously seeing them as closely related. Whereas a circulus in probando builds the proposition to be proved into the premise used for proving it, a petitio principii consists in accepting a premise as immediately certain “although it still requires a proof.”10 As related to circulus in probando, petitio principii is in a loose sense “a kind of

10

enthalten, daß wir nämlich vielleicht die Idee der Freiheit nur um des sittlichen Gesetzes willen zum Grunde legten, um dieses nachher aus der Freiheit wiederum zu schließen, mithin von jenem gar keinen Grund angeben könnten, sondern es nur als Erbittung eines Prinzips, das uns gutgesinnte Seelen wohl gerne einräumen werden, welches wir aber niemals als einen erweislichen Satz aufstellen könnten” (GMS, 453; I have italicized the words under discussion). For a discussion of what jenem refers to, see Schönecker (1999, 340). Anyway, even if jenem refers to the moral law, so that the moral law is what is put forward as a petitio, this can be taken as resulting from the groundlessness of the premise for inferring the moral law, namely freedom (cf. Schönecker, 1999, 340 f.). See also Kim (2002, 70) for references to Kant’s other lectures on logic treating these two fallacies.

The Circle and the Two Standpoints

291

circle” (eine Art von Zirkel), which is the phrase with which the circle problem is introduced in GMS, 450.11 There is also another reason why the argument might be called a circle. Kant thinks that if one were to commit this fallacy, the motivation for doing so would be the desire to reach the conclusion. The train of thought goes from wanting to establish morality to finding a suitable premise from which to derive it. “We take ourselves as free […] in order to think ourselves under moral laws” (GMS, 450), and we “took as a ground the idea of freedom only for the sake of the moral law” (GMS, 453). There are also formulations in the previous discussion in the third subsection of GMS III which can be read in this way, such as this one: “in the idea of freedom [it seems that] we have actually only presupposed the moral law, namely the principle of the autonomy of the will itself” (GMS, 449). So even though the argument structure in the circle passages is that of a petitio, in an informal, psychological sense, the train of thought is circular. 12 One cause for the prevalence of interpretations misconstruing the circle passages as having the form of circulus in probando arguments is presumably that commentators have read this motivational movement of thought into the logical structure described in these passages. Another feature of Kant’s presentation of the problem that might be taken to speak in favour of an interpretation in terms of circulus in probando is the discussion in connection to the first circle passage of the reciprocity of the concepts of freedom and the will’s own lawgiving: freedom and the will’s own lawgiving are both autonomy and hence reciprocal concepts [Wechselbegriffe], and for this very reason one cannot be used to explain the other or to furnish a ground for it but can at most be 11

12

Kim (2002, 70) considers the possibility that the circle in GMS III may take the form of a petitio. But first he only discusses an argument in which morality is assumed without ground and used for deriving freedom, which is surprising since he argues that the goal of the deduction in GMS III is to establish the principle of morality. When Kim then (on p. 73) briefly treats the inverse argument, in which freedom is assumed and morality inferred, his rejection of this interpretation (as presented in Allison 1990, 221) seems to rely on taking the circle as a fallacy Kant himself is to have committed. Moreover, because he construes the fi rst circle passage (GMS, 450) as a description of a strict circle (p. 71), he looks for a circulus in probando even when he considers a petitio, so that his initial insight that the circle might be a petitio is lost. The circular train of thought could perhaps be presented quasi-formally in the following way (letting “M” represent morality and “F” freedom): (1) M (groundless assumption, psychologically based on conviction) (2) F→M (assumption on the basis of the reciprocity between these concepts) (3) F (groundless assumption, psychologically based on the wish to infer M) (4) M (conclusion from (2) and (3), but not from (1), as that would be unbearably trifling.)

292

Marcel Quarfood

used only for the logical purpose of reducing apparently different representations of the same object to one single concept (as different fractions of equal value are reduced to their lowest expression). (GMS, 450)

It might here seem as if Kant refers to inferences in both directions, from freedom to morality and from morality to freedom. If one of the members of a pair of reciprocal concepts ‘cannot be used to explain the other or to furnish a ground for it,’ this to be sure goes for both of the concepts. But, as Schönecker (1999, 335) points out, the text does in no way imply that both of the possible inferences are relevant for Kant’s discussion. For when freedom is assumed as the ground for morality, one of the reciprocal concepts is used to ‘furnish a ground’ for the other one, irrespective of whether an inference in the other direction is made or not. 2. The circle in context The previous discussion does not seem to have achieved much. We do not yet know how there might be a petitio principii involved in the straightforward argument from freedom to morality. This argument has two components: the analytical connection between freedom and the moral law, and the necessity for rational beings to regard their judgments as free. But if we now have reason to think that the circle pertains to the insufficient ground for positing freedom, we can conclude that the threat should concern the second step of the straightforward argument. Its first step merely claims that there is a conceptual link between the concepts of a free will and an autonomous will. It is in connection to the second step that we might be enticed to ascribe freedom to ourselves, so it is here that a petitio can be committed. But why should there be a problem in our ascribing freedom in a practical respect to ourselves? One possibility is that there is something wrong with the argument to the effect that “reason must regard itself as the author of its principles” and therefore must stand “under the idea of freedom” (GMS, 448), so that clinging to its conclusion would be entirely groundless. However, nothing indicates that Kant rejected the argument as such. There are two other possibilities that are more plausible. One is that we have not yet (that is, at the point where the threat of a circle occurs) done enough to justify the assumption that we are rational beings, so that the application of the second part of the straightforward argument to the case of human beings amounts to a petitio. Another possibility is that even if the argument for the

The Circle and the Two Standpoints

293

practical freedom of rational beings applies to us, this argument leads to inadmissible consequences. As will be seen, these interpretations share one common feature, as both of them stress the importance of making a distinction between the moral law (as the necessary course of action for an ideally rational being) and moral obligation in the form of the categorical imperative (as applying to beings who are both rational and sensibly affected). To begin with the first line of interpretation, when Kant in the second subsection argues that a rational being has to assume its own freedom, he points out that the proof concerns beings “that are rational and endowed with a will” (GMS, 448). The requirement that a will is needed in addition to rationality opens for the possibility that we may possess theoretical reason but yet lack practical reason (that is, will), so that the petitio one might commit here would be to merely assume without further ado that we have a will.13 Or, to go one step further in this direction, perhaps even assuming theoretical rationality in the case of human beings is to go too far, as long as independent grounds have not been given.14 If this is the right way to view the argument establishing practical freedom from the capacity to judge, the argument concerns ideally rational beings, and is therefore of no immediate consequence for non-ideal human beings. I cannot pursue the many questions that arise in the interpretation of the second subsection of GMS III, so I merely note here that this reading fits well with the circle considered as a petitio with respect to assuming freedom, as well as with Kant’s turning to transcendental idealism in order to solve the problem. For when the two standpoints doctrine of transcendental idealism is deployed, a new sphere is opened where human rationality (theoretical as well as practical) may be found, providing the straightforward argument with an independent reason for the assumption of freedom. Another significant feature of this interpretation is that it points to the crucial distinction between the moral law as such and the categorical imperative.15 This distinction is sometimes (e. g. in GMS, 420; 440; 447) marked by Kant in terms of analytic versus synthetic a priori connections. Simply put, the contrast is between a being for which morality is a necessary law of action, so that it cannot fail to act morally, and the 13 14 15

Cf. Allison (1990, 218). See the discussion in Schönecker (1999, e. g. 208, 233, 319). The radical contrast between the categorical imperative as applying to beings that are both rational and affected by sensibility on the one hand and the moral law pertaining to purely rational beings on the other is emphasized time and again in Schönecker (1999).

294

Marcel Quarfood

human being which has a rather more difficult time when it comes to acting morally. Obligation and duty are notions applicable only to the latter. The concept of duty “contains that of a good will though under certain subjective limitations and hindrances” (GMS, 397), whereas for a being whose reason “is practical without hindrance” (GMS, 449) or “holy” (GMS, 414; 439), the will and its maxims “necessarily harmonize with the laws of autonomy” (GMS, 439). From this we can see that the concept of duty analytically contains “good will” as subconcept, as well as “hindrances,” which we can identify with sensibility and the incentives it provides us with (cf. GMS, 449). According to Kant’s theory of conceptual containment, the concept of gold contains the subconcepts “metal” and “yellow”, and if we detach from the concept “metal” what we might call its limiting condition, the concept “yellow,” we can no longer infer “gold” (as we could from the combination of “metal” and “yellow,” if we assume for the sake of simplicity that these are the only subconcepts of “gold”).16 Analogously, if we detach the limiting condition of being subject to sensibility from the concept of a good will, we cannot infer ‘duty’ from the latter.17 ‘Duty,’ as a composite concept, presupposes a synthesis connecting its subconcepts, ‘good will’ and ‘sensible incentives.’ In less formal terms, what the requirement for a synthesis amounts to is that in order to explicate our subjection to duty (and thereby our subjection to the categorical imperative), we need to connect the necessary lawfulness of pure practical reason to the fact that our will is “affected by sensibility” (GMS, 449). To explicate how this synthetic connection can be justified a priori is to give a deduction of the categorical imperative. So when Kant asks (immediately before introducing the suspicion of a hidden circle in the third subsection of GMS III) “from whence the moral law obligates” (GMS, 450), he indicates that up to this point the possibility of the categorical imperative has not been addressed.18 An alternative way of explaining how the assumption of freedom could possibly be a kind of petitio – in spite of the argument in the second subsection of GMS III to the effect that a rational being must see itself as free – would involve accepting this argument as valid in principle even for the human case, already as it stands, without fur16 17

18

Cf. Prol, 267. I think “contains” in GMS, 397, is used in the exact sense it has in Kant’s theory of conceptual containment, but for present purposes it suffices to consider it as an analogy. Allen Wood’s translation, in Kant (2002).

The Circle and the Two Standpoints

295

ther premises. On this view, the problem would rather concern how to make use of the argument’s conclusion: the straightforward argument from freedom to morality appears to be irrelevant unless it can be explained why morality presents itself as duty for the human being. On this interpretation too, the third subsection of GMS III has to find a way to unite rationality and subjection to sensibility in the same being, and unless this can be accomplished it is of no avail to possess an abstract argument from freedom to morality. The conclusion of the argument is too distant from human moral life to have any cogency, despite its validity. If we are justified in considering ourselves free (at least in the practical respect), morality follows in virtue of the analytical link between freedom and morality, but we do not yet see how thus ascribing morality to ourselves can aid in understanding (or even be consistent with) the phenomenon of obligation and the fact that we are not necessarily acting morally. Unless this can be accommodated to the result of the straightforward argument, no progress can be made in the deduction of the categorical imperative. The alternative interpretation has both weaker and stronger points. Its weakest aspect is perhaps that it represents the circle in terms reminiscent of an antinomy rather than a petitio principii or even a circulus in probando. There are to be sure points of contact between the circle problem and the Antinomies of reason, but it would seem that this is not all there is to the circle. I will not go further in evaluating this interpretation, but merely point out that it converges with the previous one when it comes to what one should expect the solution to the circle problem to accomplish. Such a solution must, on this view as well as on the previous one, explain how both the validity of the moral law and the capacity to adopt immoral maxims can apply to human beings, that is, it must illuminate the possibility of obligation. An important difference between the two views as concerns what the solution to the circle problem should deliver is that on this view, there is no need to find new evidence for justifying freedom. The argument for a rational being’s having to act under the idea of freedom provides the evidence needed; we only have to find a way to accommodate this result to the nature of our moral situation. In light of the exposition up to this point, we should be able to answer the question as to who commits the circle fallacy. Is it Kant, or is it perhaps some other philosopher? It ought to be fairly clear that Kant has not committed the circle fallacy. Kant does not retract from his arguments in the beginning of GMS III, he rather points out that they are not sufficient for a deduction of moral obligation. Only a reading

296

Marcel Quarfood

stressing the “admission” in the statement that “it must be freely admitted that a kind of circle comes to light here” (GMS, 450) but neglecting other things said about the circle would have to attribute the fallacy to Kant. He continues the sentence by pointing out that there merely ‘seems’ to be no way to escape the circle. And in GMS, 453, when he claims to have solved the problem, it is ‘the suspicion that a hidden circle was contained in our inference from freedom to autonomy and from the latter to the moral law’ that is said to have been removed. So, apparently, there never was a circle, but a suspicion of a circle arose at one point. Why did it arise? According to Kant the doctrine of the two standpoints will remove the suspicion of a circle. Hence it must be the absence of this doctrine that occasions the threat. But the doctrine, central to transcendental idealism, could hardly be thought to be absent from Kant’s reflection. However, when the suspicion of a circle arises, the doctrine of the two standpoints has not yet been put to work in the Groundwork, and it could well be expected to be absent from the reflection of the reader. Rather than attributing actual circular reasoning to Kant or some other thinker, we should see the circle as a possible mistake. The one who might commit it would be a possible reader. Having followed Kant’s train of reasoning until the end of the second subsection of GMS III, such a reader, not aware of transcendental idealism, would find it hard to advance without assuming freedom by petitio (or, on the alternative interpretation adumbrated above, without getting stuck in the paradox of having proved morality and yet having to face the proof’s irrelevance to moral phenomenology). The problem encountered by a reader lacking the possibility to appeal to transcendental idealism could also be formulated in terms of the difference between on the one hand an analytic procedure following conceptual connections but lacking the resources to establish that something falls under these concepts, and on the other hand a synthetic a priori methodology capable of making explicit necessary connections not merely reducible to analytical relations. The former method would be useful for “determining the genuine principle more accurately,” but it could not explain the principle’s validity or tackle the issue of obligation (GMS, 449). Against this background, Reinhard Brandt’s suggestion that philosophers employing the analytical methodology of the Wolffian tradition would be the ones committing the fallacy of circularity gets a certain corroboration.19 From a Kan19

Brandt (1988, 186). For criticism of Brandt, see Allison (1990, 220), Schönecker (1999, 353), and Kim (2002, 74 f.).

The Circle and the Two Standpoints

297

tian point of view, a thinker who draws such analytic connections, but lacks the resources of critique by means of which synthetic a priori claims can be justified, is a dogmatist par préférence. The circle might be said to represent the route the ideal typical Wolffian dogmatist (as seen by the lights of Kant) would have to follow, even though no Wolffian actually reasoned along the lines of the circle.20 3. The two standpoints Immediately upon introducing the circle in 450 Kant presents the way out. It consists in attending to the difference between two standpoints we can adopt with respect to ourselves. On the one hand we think of ourselves “by means of freedom” as “causes efficient a priori,” on the other hand, we conceive of ourselves “in terms of our actions as effects that we see before our eyes.” The two standpoints are linked to the distinction between things in themselves and appearances. In these pages (GMS, 450 f.), Kant describes an observation possible to make even for “the commonest [gemeinste] understanding” (GMS, 450). But he states that a “reflective human being” would come to the same conclusion (GMS, 451), and the whole discussion is conducted in the technical language of transcendental philosophy, so it is endorsed also on this level of reflection. Bringing the unsophisticated person’s understanding into the discussion accords with the starting point of the Groundwork, namely that the content of morality is already implicitly assumed in everyday moral thinking. The distinction is not just imported from philosophical theory, but is corroborated in human experience (though, as we will see, it does certainly not have its source in experience). The drawing of a distinction between appearances and things in themselves is traced back to an observation [Bemerkung] made even by the commonest understanding in the form of “an obscure discrimination of judgment [Urteilskraft] which it calls feeling” (GMS, 451), to the effect that representations over which we have no control present things to us only with respect to how they affect us, that is, how they appear, but not as to how they are in themselves. The passivity of the reception of representations leads us to view them as appearances, and thereby we are committed also to assuming that there is “behind appearances something else that is not appearance” (GMS, 451).21 In 20 21

Which Brandt, incidentally, construes as a circulus in probando. Compare KrV, B xxvi, on “the absurd proposition that there is an appearance without something that appears.”

298

Marcel Quarfood

thus noticing the difference between passively received representations and those produced by our own activity, we obtain a “crude” [rohe] version of the distinction between the “world of sense” and the “world of understanding” [Verstandeswelt] (GMS, 451). The world of understanding is conceived of as the stable basis of the world of sense, which latter may vary for different “observers of the world” [Weltbeschauern] in accordance with variations in sensibility (GMS, 451). The above considerations do not concern only representations given from outside, but also what is given in inner sense. Therefore the information we can obtain about ourselves consists only of appearances, and in this case too, we “must necessarily assume something else lying at their basis,” an I in itself (GMS, 451). But with regard to “pure activity [reine Tätigkeit],” which “reaches consciousness immediately and not through affection of the senses,” one must count oneself to the “intellectual world [intellektuellen Welt]” (GMS, 451). This pure activity is identified as reason, a capacity which the human being “finds in himself” and by means of which he “distinguishes himself from all other things” (GMS, 452). Reason operates with ideas that transcend sensibility. In making the distinction between the world of sense and the world of understanding, reason marks out the limits of the understanding, which, in contrast to reason, is restricted in its activity in so far as its concepts only serve to “bring sensible representations under rules” (GMS, 452). The spontaneity of reason allows the human being to think of himself as “intelligence [Intelligenz]” (GMS, 452), that is, as belonging to the world of understanding.22 Therefore there are now two standpoints available from which to consider oneself: as belonging to the world of sense, and as belonging to the intelligible world. This result is used to remove the suspicion of there being a circle “in our inference from freedom to autonomy and from the latter to the moral law” (GMS, 453). There is no circle (i. e., no petitio), because we now have a ground for attributing freedom to ourselves, independently of our wish to take ourselves as subject to morality. If we find reason in ourselves, then we have the premise needed for concluding that we are practically free, and then we can use the analytic connection between freedom and morality to establish the validity of the latter. The threat of a circle being removed, the straightforward argument goes through. On this reading, the recourse to the two-standpoints doctrine of transcendental idealism serves the function of providing independent evi22

Compare KrV, e. g. A 682/B 710.

The Circle and the Two Standpoints

299

dence for our possession of reason, so that we can ascribe freedom to ourselves. But another, equally important reason for invoking the doctrine (which, on the alternative reading outlined in part 2 above, is the sole reason for appealing to it), is that we need to explain why morality takes the form of obligation for us. The distinction between the two standpoints offers such an explanation by means of a third standpoint, as it were. The standpoint of the intelligible world on its own does nothing to make obligation comprehensible, since a pure Intelligenz would have no duty, but only a rational nature harmonizing with morality. The standpoint of the world of sense, as heteronomy, would of course have no bearing on moral obligation either. Only from what might be called the third standpoint, in which we think of ourselves “as belonging to the world of sense and yet at the same time to the world of understanding,” is it possible to “think of ourselves as put under obligation” (GMS, 453). This is so because in thinking of myself as belonging to both worlds, the concept of the sensibly conditioned, heteronomous will can be synthetically connected with the “idea of the same will but belonging to the world of the understanding” (GMS, 454), which latter serves as the intelligible basis for the former. A sensible being which also has spontaneity recognizes itself, in virtue of this spontaneity, as a member of the intelligible world as well, and it thereby considers the intelligible aspect of itself to constitute its proper self (eigentliche Selbst) (GMS, 457). This makes the laws of the intelligible world prescriptive for it as a sensible being, which is what subjection to duty amounts to. Thus, the distinction between the two standpoints made possible by transcendental idealism opens the space required to account for the peculiar character of human morality, and so it goes as far as is possible towards an answer to the ultimately unanswerable question posed before the introduction of the circle, the question concerning why we take an interest in the moral law. Neither a purely sensible being nor a purely rational one could take such an interest. Does this way of avoiding the threat of a circle transgress the cognitive limits imposed by critical philosophy? I cannot even begin to discuss this question adequately here. The answer depends on the status accorded to our awareness of spontaneity. It is clear that Kant considers the consciousness of spontaneity as an undeniable fact. Is it possible to cash it out in terms of theoretical (metaphysical) cognition, or is it rather to be conceived of as a practical capacity, inexplicable as to its ultimate nature but appealed to as a presupposition for action? One consideration that perhaps speaks in favour of the latter view

300

Marcel Quarfood

is that both before and after introducing the distinction between the two points of view, Kant urges the reader not to look for a theoretical proof of freedom. Early on in GMS III he states that he is not “bound to prove freedom in its theoretical respect,” but only in a practical respect (GMS, 448n.). And at a later stage of his exposition, he says that the “concept of a world of understanding is […] only a standpoint that reason sees itself constrained to take outside appearances in order to think of itself as practical” (GMS, 458). Literature Kant’s writings Kant’s writings will be cited according to the pagination of Kants gesammelte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902–) (abbreviated as ‘AA’). All textual references are to The Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press, 1992–), except where otherwise noted. The German text of the Grundlegung will be quoted according to the edition by Bernd Kraft and Dieter Schönecker (Hamburg, Felix Meiner Verlag, 1999). The Critique of Pure Reason will be cited according to the A/B pagination from the first and second editions. GMS Grundlegung zur Metaphysik der Sitten, AA, IV KrV Kritik der reinen Vernunft, AA III, IV Log Logik-Jäsche, AA, IX Prol Prolegomena, AA IV Kant, Immanuel (2002): Groundwork for the Metaphysics of Morals, New Haven and London: Yale University Press.

Other works Allison, Henry E. (1990): Kant’s Theory of Freedom, New York, Cambridge University Press. Brandt, Reinhard (1988): “Der Zirkel im dritten Abschnitt von Kants Grundlegung zur Metaphysik der Sitten”, in: H. Oberer, Hariolf / G. Seel (Hrsg.), 169–191. Kim, Halla (2002): “Has Kant Committed the Fallacy of Circularity in Foundations III?”, in: Journal of Philosophical Research, 27, 65–81. Schönecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs, Freiburg i. Br./München. T-h (1788): “Vierter Versuch über die Kantsche Grundlegung zu einer Metaphysik der Sitten,” in: Deutsches Museum, 13, 264–293.

Dieter Schönecker

How is a categorical imperative possible? Kant’s deduction of the categorical imperative (GMS, III,4) Kant’s deduction of the categorical imperative is the answer to the following question: “How is a categorical imperative possible?” The answer is given in subsection 4 (Sec. 4) of chapter three of the Groundwork. It is impossible to understand this answer, and hence impossible to understand Kant’s deduction of the moral law, without taking into account the overall context of Groundwork III (GMS III). However, here I can only sketch the overall structure of GMS III, and therefore only present a sketch of what I call Kant’s thesis of analyticity.1 This thesis is developed in Sec. 1 of GMS III; however, it appears time and again in GMS III, and it deserves special attention (part 1). Part 2, then, offers a close reading and analysis of Kant’s deduction. The purpose of this paper is not to criticize or make philosophical use of Kant’s deduction. As I have argued elsewhere, one of the major obstacles of serious (i. e. historical) Kant-research is the inability (and unwillingness) to distinguish between the question of what a text means and and the question of whether what it manifests is true.2 1. Structure and task of GMS III: What does Kant want to achieve? In the preface of the GMS, Kant claims that his Groundwork is “nothing more than the search for and establishment of the supreme princi1

2

For a detailed analysis of Groundwork III, cf. Schönecker (1999); this article is based upon that book. Here I will not address the secondary literature (I did so extensively in Schönecker, 1999); as far as I can tell, little has been published since that pays close attention to the text. However, I will make some brief comments on Steigleder’s interpretation of what I call the thesis of analyticity. – Many thanks to Richard Capobianco and Alexander Cotter for checking my English. Cf. Schönecker (2001); cf. also Damschen / Schönecker (2006).

302

Dieter Schönecker

ple of morality” (GMS, 392). This ‘supreme principle’ is the categorical imperative (CI). It is generally agreed upon that the ‘search for’ the CI, i. e. its conceptual analysis, takes place in GMS I/II, whereas its ‘establishment’ is to be found in GMS III. To be more precise, this ‘establishment’ is in the answer that Kant provides to the question of ‘how is a categorical imperative possible?’ This question is first raised in GMS I (GMS, 417 ff.), and then again in GMS II, 425; its answer, however, is postponed to Sec. 4 of GMS III, the heading of which is again that question of how a categorical imperative is possible. There is no doubt that the answer to this question is what Kant himself calls a “deduction” (GMS, 447,22; 454,21; 463,21); the second paragraph of Sec. 4 begins with the formulation “And thus categorical imperatives are possible …” (GMS, 454,6, m. e.). Before we move on, it is important to realize two crucial structural elements of the text. First, Kant clearly states at the end of Sec. 1 that he cannot yet or immediately provide an answer to the question of how a categorical imperative is possible; rather, he says, this “still needs some preparation” (GMS, 447,25). Since the central question “How is a categorical imperative possible?” is indeed the heading of Sec. 4, it is only natural to assume that Sec. 2 and Sec. 3 provide that ‘preparation’, whereas the actual answer, then, is given in Sec. 4. (As we will see, this also fits very well with what really happens in those sections.) The second structural element relates to what I call Kant’s thesis of analyticity. In Sec. 1, Kant puts it as follows: a free will and a will under moral laws are the same. [paragraph] Thus if freedom of the will is presupposed, then morality follows together with its principle from mere analysis of its concept (GMS, 447,6–10).

This thesis has been widely misunderstood; as a result, Kant’s overall argument (deduction) in GMS III has been misunderstood too (often not recognized at all, as a matter of fact). His overall argument in GMS III always has been reconstructed as follows: A free will is a will under the moral law; freedom must be presupposed as a quality of the will of all rational beings; human beings are rational beings; therefore, the human will as a free will is under the moral law, which is to say the categorical imperative is valid. 3 Since premise 1 is proven in Sec. 1, premise 2 is argued for in Sec. 2 and premise 3 in Sec. 3,4 the answer 3 4

Cf. for instance Wood (1999, pp. 171–176). As a matter of fact, the claim that ‘freedom must be presupposed as a quality of the will of all rational beings’ is the heading of Sec. 2; Sec. 3 relates this to human beings in particular.

How is a categorical imperative possible?

303

to the question of “how is a categorical imperative possible?” and thus the deduction would already be given in Sec. 3. However, we already noticed that there is a fourth section the very heading of which is “how is a categorical imperative possible?”, and it is only in this section that Kant provides the deduction. If the standard interpretation of the thesis of analyticity and the deduction were correct, then the deduction would be completed by the end of Sec. 3. Clearly, Kant’s thesis of analyticity cannot mean that sensuous-rational beings (beings, like human beings, that are both sensuous and rational), who have a free will, always act morally; for that is obviously not the case. However, it also cannot mean that the free will of such a sensuous-rational being is “under” (GMS, 447,7) the moral law if this is taken to mean that sensuous-rational beings are obligated by the categorical imperative. They are indeed, but that they are obligated is something that Kant after Sec. 1 and 2 has yet to demonstrate. After all, this is why he still raises the following question after he has argued for his thesis of analyticity (Sec. 1) and the claim that freedom must be presupposed as a quality of the will of all rational beings (Sec. 2): “But why ought I to subject myself to this [moral] principle …?” (GMS, 449,11). In Sec. 3, Kant still asks “from whence the moral law obligates” (450,16), a question to which “no satisfactory answer” (GMS, 450,2) has been given yet (i. e. up to Sec. 3). This second structural observation also implies that a free will and a will under moral laws (under the CI) are not ‘the same’; that they ought to be is what the deduction has to prove, and that is why the deduction is yet to come after Sec. 2/3. What exactly this question (‘from whence does the moral law obligate?’) means, is hard to say and indeed a source of confusion for Kant himself. In any event, it is a question that Kant holds to be unanswered, and this along with the first observation that Sec. 2 and 3 are only a ‘preparation’ for the answer to the question of how a categorical imperative is possible renders the standard interpretation of the thesis of analyticity untenable. So how are we to understand Kant’s thesis of analyticity? Throughout GMS I/II, Kant repeatedly argues that for a perfectly rational and free being the moral law is not an imperative. Rather, the moral law must be understood as a rule that these beings necessarily follow. Or as Kant puts it: With regard to perfectly rational and free beings the moral law is not a synthetic, but an analytic proposition. Analyzing the very concept of a free and rational being yields the insight that the (concept of the) moral law is included in it. One could also say that the moral law describes what these beings do: By their very nature

304

Dieter Schönecker

perfectly rational beings always act morally. There are many passages that prove this point (some of which can actually be found in GMS III). A famous one in GMS II reads as follows: If reason determines the will without exception [unausbleiblich], 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 (GMS, 412,30–35).

Since human beings don’t have such a perfectly rational being, to them the moral law is a categorical imperative or, as Kant puts it, a synthetic proposition a priori. The categorical imperative 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 (GMS, 420,32–35, Fn., m.e.).

Willing the good action is not necessarily ‘contained’ in the volition of a sensuous-rational being, i. e. it cannot be ‘analytically derived’ from the volition of such a being. That is why, in contrast to a pure will, the CI does not follow by ‘mere analysis’ (GMS, 447,9) of the concept of freedom of such a sensuous-rational being. So when Kant in his thesis of analyticity states that ‘a free will and a will under moral laws are the same’ and that ‘morality follows together with its principle from mere analysis of the concept of the freedom of the will’ (GMS, 447,6–10), all he really says is this: A perfectly rational and free will always wills morally. Such a will must be understood as the will of an actually holy being; or (as we will see) it must be understood as the intelligible will of a being that is both a member of the empirical and intelligible world. The argumentative structure of GMS III strongly supports this reading. After Kant has presented his solution to the notorious circle in Sec. 3, he concludes: 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 (GMS, 453,11–15).

It is highly remarkable that Kant does not contrast the formulation ‘if we think of ourselves as free’ with ‘if we think of ourselves as not free.’ Rather, he contrasts it with ‘if we think of ourselves as obligated by

How is a categorical imperative possible?

305

duty.’ The counterpoint to freedom is not the absence or lack of freedom, but duty. Since morality without freedom is not thinkable for Kant, freedom understood as a counterpoint to duty cannot be identical with the freedom presupposed by morality for beings that are both sensuous and rational. Freedom as the counterpoint to duty must be understood as the quality of a will that is completely rational and free. However, a being that is ‘obligated by duty’ must be free, too. Therefore, to think of oneself as ‘obligated by duty’ implies to consider oneself ‘as belonging to the world of sense’, but it also means to consider oneself ‘at the same time (“zugleich”, GMS, 453,15) as belonging to the world of understanding’. 5 Hence the first part of the passage quoted above (GMS, 453,11–13) really is nothing but a reformulation of the thesis of analyticity. The second part after the semi-colon, however, makes clear why for sensuous-rational beings the moral law is an imperative and hence duty.6 The thesis of analyticity also shows up in the deduction found in Sec. 4 which we will address in great detail later: “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” (GMS, 453,25–27, m. e.). A bit later it says: “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” (GMS, 454,6–9, m. e.). Again, at the end of the Sec. 4 Kant concludes: “The moral ‘ought’ is thus his own necessary volition as a member of 5 6

On the “zugleich” cf. also GMS, 454,9 and 462,31. The passage just quoted (‘For now we see … world of understanding’, GMS, 453,11– 15) is Kant’s final answer to the problem of the notorious “circle” (GMS 450,18; 453,4) in Sec. 3. This ‘circle’ is yet another issue I cannot address here (cf. Schönecker, 1999, pp. 317–358). However, let me state that its correct understanding must begin with the insight that for Kant there is a difference between a petitio principii and a circulus in probando, and that the aforementioned ‘circle’ is an “Erbittung eines Prinzips” (GMS 453,9), not a vicious circle (‘Erbittung eines Prinzips’ is Kant’s translation of the Latin phrase ‘petitio principii, i. e. ‘begging the question’). What is still missing in Sec. 3 (just begged for) is the rationale for the human being’s belief in the idea of freedom as well as the justification for the categorical imperative. So what Kant actually does to remove the suspicion of the petitio is, first, to show how and why the human being can understand himself as a member of the intelligible world and therefore as free. Second, Kant’s discussion of the petitio – and his fi nal solution of it at the end of Sec. 3 – also draws our attention to a problem which is equally important, to wit, that by the thesis of analyticity the validity of the categorical imperative has not been demonstrated; as a matter of fact, the common misunderstanding of Kant’s overall argument in GMS III is exactly what Kant warns his readers against. – On the ‘circle’ also cf. Quarfood’s paper in this volume.

306

Dieter Schönecker

an intelligible world and is thought of by him as an ‘ought’ only insofar as he at the same time considers himself as a member of the sensible world” (GMS, 455,7–9, m. e.). And to quote yet another passage: “Under the presupposition of freedom of the will of an intelligence, its autonomy, as the formal condition under which alone it can be determined, is a necessary consequence” (GMS, 461,14–17). Kant describes an ‘intelligence’ as a rational being that considers itself solely as a member of the world of understanding. Thus, the ‘will’ in the context of the thesis of analyticity cannot simply be understood as the will of a sensuous-rational being. It must be understood either as the will of a perfectly rational being, whose free will always is a good will; or, with an eye on human beings, it must be understood as the will of a sensuous-rational being, whose will is both part of the ‘world of sense’ (“Sinnenwelt”) and of the ‘world of understanding’ (“Verstandeswelt”), whose will, however, as part of the latter world, is “the idea of reason, which would have full control over all subjective motivations” (GMS, 420,31, m. e.). It is exactly in this sense that in the context of the deduction proper (Sec. 4), Kant writes that the categorical ‘ought’ “represents a synthetic proposition a priori by the fact that my to 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” (GMS, 454,11, second emphasis mine). In this perspective the human being, too, considers his will as free, and such a free will always wills the good (‘all actions would be perfectly in accord with the principle of the autonomy of the pure will’). This is the meaning of Kant’s thesis of analyticity.7 Thus Kant writes right after stating the thesis of analyticity: Nonetheless, the latter is always a synthetic proposition: an absolutely good will is that whose maxim can always contain itself considered as universal law, for through analysis of the concept of an absolutely good will that quality of the maxim cannot be found. (GMS, 447,10–14)

With the ‘nonetheless’ (“Indessen”) Kant sets off the syntheticity of the CI from the analyticity of the “principle of morality” (GMS, 447,6) just mentioned in the sentence before. Whereas this ‘principle’ “follows […] from mere analysis” (GMS, 447,8–9) of the concept of freedom, from the analysis of the concept ‘of an absolutely good will’ it does not follow (‘cannot be found’) that ‘its maxim can always con7

I cannot discuss here Kant’s repeated claim that the categorical imperative is a synthetic proposition; in any event, I would hold that, strictly speaking, it does not make sense.

How is a categorical imperative possible?

307

tain itself considered as universal law’, i. e. that such a will is always good. And this means, first, that in the last passage quoted above the sentence after the colon (‘an absolutely good will …’) is a synthetic proposition; for after the colon the reason is given for why the ‘analysis’ cannot take place which implies that we are, indeed, dealing with a synthetcial proposition rather than an analytic one. Second, one has to behold that the ‘concept of an absolutely good will’ here does not refer to the will of a perfectly rational being; for if one analyzes such a concept, it does follow that ‘its maxim can always contain itself considered as universal law’. 8 In light of these considerations, what then does the question of how a categorical imperative is possible really mean? Lack of space does not allow to go into any analysis of the quite obvious parallel that Kant likes to draw between the famous theoretical question of how synthetic propositions a priori are possible. In any event, a closer look reveals that the parallel question “How is a categorical imperative possible?” actually includes three questions or aspects: 1. Why is the CI valid (binding, obligatory)? 2. How can freedom be understood, and why may we consider ourselves to be free? 3. How can pure practical reason bring about an interest in the moral law? The third question cannot be answered.9 The second question is answered in Sec. 2 and Sec. 3 of GMS III: In Sec. 1, Kant first argues that a perfectly rational and free will always wills (acts) morally. Sec. 2 demonstrates that because of its ability to think spontaneously (freely), a rational being must also consider itself practically free; Sec. 3 then refers to the difference between the world of understanding and 8

9

From early on this passage and especially the (formulation of the) ‘concept of an absolutely good will’ has tremendously contributed to the misunderstandings of the thesis of analyticity (cf. the report on the secondary literature in Schönecker, 1999, 168–171). For Kant does indeed avail himself of this ‘concept of an absolutely good will’ to refer to a perfectly rational (‘holy’) will (cf. GMS, 439,29–34). However, he also refers to 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 conflict with itself“ (GMS, 437,6). In this passage and context Kant clearly does not talk about perfectly rational (holy) beings. He refers to the will as ‘absolutely good’ only insasmuch its (particular) maxim can be universalized; cf. GMS, 426,10; 437,24; 437,32; 444,28. In Sec. 5 Kant provides a lengthy justification for his claim that the question of how a categorical imperative is possible can only be answered partially: “Thus the question, ‘How is a categorical imperative possible?’ can be answered to this extent: …: (GMS, 461,7, m. e.); what cannot be answered is the question of how pure practical reason indeed can be practical (cf. GMS 458,37; 459,34; 460,10; 461,25, 461,32).

308

Dieter Schönecker

the world of sense in order to argue that the human being also must understand himself as practically free. But then, it seems, Kant still thinks that the answer to that crucial question is still not answered; recall that Kant in Sec. 3 still asks “from whence the moral law obligates” (GMS, 450,16), a question that is not answered in Sec. 3 itself. For now, I will leave it open what exactly it is that question is asking for; as we will see later, Kant himself seems not to be entirely clear about its meaning. In any event, what is needed is somehow the proof that the CI is really valid, i. e. really binding on us. As long as this has not been shown, morality could very well be a “figment of the mind” (GMS, 407,17, 445,8); I will come back to this later. 2. The deduction of the categorical imperative Sec. 4, then, raises our crucial question again and finally answers it. This section is broken down into three paragraphs. In the first paragraph, Kant offers the argument proper. In the second paragraph, Kant provides the final answer to the question of how categorical imperatives are possible (“And thus categorical imperatives are possible …” GMS, 454,6). Paragraph three comes back to ‘common rational moral cognition’ (something mentioned at the end of the preface); I will not discuss it here. 2.1 Presuppositions of the deduction Let me now quote the first four sentences of Sec. 4: 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.) (GMS, 453, 17)

How is a categorical imperative possible?

309

These lines are basically a summary of what Kant has said about freedom and morality in Sec. 1–3, and so they are also a summary of his thesis of analyticity. Using Kant’s own words of paragraphs 9–12 (GMS, 452,7–453,15), the gist of it is this: ‘Now the human being actually finds in himself a faculty, and this faculty is reason as a faculty of pure spontaneity (self-activity). 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. As a rational being, hence one belonging to the intelligible world, the human being can never think of the causality of his own will otherwise than under the idea of freedom. Now, with the idea of freedom the concept of autonomy is inseparably bound up, and with the latter the universal principle of morality. 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.’ – Again, paragraph 1 of Sec. 4 is nothing but a recapitulation of these basic ideas. I will now present them in several theses (D1–4). Thesis 1 states that every human being must understand itself as a rational being: (D1) The human being finds in itself the faculty of reason, which, as an epistemic faculty, is a faculty of pure spontaneity.

It is important to see that Kant’s argument both in Sec. 3 and Sec. 4 begins with reason as a theoretical (epistemic) faculty. Only on pain of self-contradiction, he argues, one can deny that one has reason, for denying that one does presupposes that one does.10 The activity of reason is self-activity, spontaneity, freedom, and these qualities justify the human being to understand himself ‘as an intelligence.’ And as an intelligence, the human being must understand himself as a member of the world of understanding. (D2) As a rational being, a human being must understand himself as an intelligence and, in this perspective, as a member of the world of understanding.

It is important to say ‘in this perspective’ because only as an intelligence, and inasmuch a human being is an intelligence, he may con10

Cf. Schönecker (1999, pp. 196–316); here it is important to take into account Kant’s Review of Schulz’s Attempt at Introduction to a Doctrine of Morals (1783).

310

Dieter Schönecker

sider himself as a member of the world of understanding. Once the human being understands himself as a member of the world of understanding because of his theoretical faculty, he also may understand his reason as a practical faculty, i. e. he may understand his will to be a member of the world of understanding and hence to be free.11 (D3) As a rational being, hence as a being belonging to the world of understanding, the human being must understand the causality of his will under the idea of freedom.

Here it shows how important a proper understanding of the thesis of analyticity is. For what that thesis claims is exactly that the principle of morality is ‘inseparably bound up’ with freedom. (D4) Since the moral law is analytically bound up with freedom as a property of the will of a rational being that is a member of the world of understanding, the human being also may understand his autonomy and the moral law as the law of his rational volition, inasmuch as he understands himself as such a rational being that is a member of the world of understanding.

Again, it is of utmost importance to realize that with this move the validity of the moral law as a categorical imperative has not been demonstrated. All that has been shown so far is that the will of a human 11

Cf. GMS, 448,13–22.: “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” (m. e.). How are we to read ‘consequently’? The transition from the thinking I to the willing (acting) I only appears plausible if the reason, the freedom of which cannot be denied without a performative contradiction, is the very same reason that also is practical. That Kant actually has something like this in mind shows in a thought of his from the preface. There he says that “it can in the end be only one and the same reason that is distinguished merely in its application” (GMS, 391,27, m. e.). Kant calls this a “unity” (GMS, 391,25) between theoretical and practical reason “in a common principle“ (GMS, 391,26). Only if this unity is comprehensively exhibited, Kant continues, a ‘critique of pure practical reason’ can be ventured and this is something, he says, he could not do in the Groundwork. However, the truth of the matter is that in chapter three a transition to such a ‘critique of pure practical reason’ does take place, if only its “main feature” (GMS, 445,15) is exhibited. But there is no argument whatsoever for that alleged unity or identity of theoretical and practical reason. How this unity is to be understood, how Kant moves from the freedom to think to the freedom to will, and how from the concept of an intelligence to the concept of an intelligence with a will – we are left in the dark.

How is a categorical imperative possible?

311

being, considered merely as a member of the world of understanding, is analytically bound up with the moral law. This is why Kant writes in Sec. 4 of GMS III (453 f.) that all actions of a human being, if he were only a member of the world of understanding, ‘would be perfectly in accord with the principle of the autonomy of the pure will’; if a human being were ‘alone’ a member of the world of understanding, all his actions ‘would always be in accord with the autonomy of the will’. That autonomy and hence morality is a ‘consequence’, as Kant puts it several times, is only true “under the presupposition of freedom of the will of an intelligence” (GMS, 461,14, m. e.).12 For only “as intelligence” (GMS, 453,17, m. e.) does a human being understand himself as a member of the world of understanding. In Sec. 5 Kant argues that the human being must “think of him[self] as intelligence, also as thing in itself” (GMS, 459,22) and hence of himself as the “authentic self” (“eigentliches Selbst”, GMS, 457,34, m. e.). As an ‘authentic self’, a human bing gives himself the law. And so morality is that which the human being, in some sense and respect, ‘authentically wills’. That’s why the moral ought is “eigentlich ein Wollen” (‘really a volition’, GMS, 449,16, m. e.;), and that’s why it is “his own necessary volition as a member of an intelligible world” (GMS, 455,7, m. e.); I will come back to this latter thought momentarily. 2.2 The ontoethical principle Let me now quote the passage (both in German and in English) that is at the center of the deduction: Weil aber die Verstandeswelt den Grund der Sinnenwelt, mithin auch der Gesetze derselben, enthält, also in Ansehung meines Willens (der ganz zur Verstandeswelt gehört) unmittelbar gesetzgebend ist, und also auch als solche gedacht werden muß, so werde ich mich als Intelligenz, obgleich andererseits wie ein zur Sinnenwelt gehöriges Wesen, dennoch dem Gesetze der ersteren, d. i. der Vernunft, die in der Idee der Freiheit das Gesetz derselben enthält, und also der Autonomie des Willens unterworfen erkennen, folglich die Gesetze der Verstandeswelt für mich als Imperative und die diesem Prinzip gemäßen Handlungen als Pflichten ansehen müssen. (GMS, 453,31–454,5) 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 intelli12

The words “of an intelligence” are missing in Wood’s translation.

312

Dieter Schönecker

gence I will cognize myself, though 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. (GMS, 453,31–454,5)

An analysis of this sentence is quite tiresome; it is most important, however, because it contains what I call the ontoethical principle (OP). So please bear with me; also, keep in mind that we have to do this referring to the English translation. Let’s first understand the grammatical structure of this elusive sentence. The elements ‘because-therefore’ break down the sentence into two parts, the first of which obviously provides a reason (‘because’) for a conclusion drawn in the second (‘therefore’).13 Abstracting as much as we can from the actual content of what is being said, the first passage of the first part says, grammatically speaking, this: (OP1) The world of understanding contains the ground of the world of sense.

Grammatically speaking, there is no problem with this part. Reconstructing the next element of the sentence is also rather compelling (although the ‘hence’ will deserve more attention later): (OP2) Hence the world of understanding also contains the ground of the laws of the world of sense.

Now, one might think that because of the relationship between the world of understanding and the world of sense laid out in OP1 and OP2, the world of understanding is also ‘immediately legislative in regard to my will’ and ‘hence must also be thought of wholly as such’ (i. e. as ‘immediately legislative’). As we will see, that cannot be true. Rather, a correct understanding is as follows: (OP3) In regard to my will, which belongs wholly to the world of understanding, the world of understanding is immediately legislative and must also, in regard to my will, be thought of such that it (the world of understanding) contains the ground of the world of sense and its laws.

Already for grammatical reasons, the first part of this reconstruction cannot be disputed: ‘In regard to my will, which belongs wholly to 13

The translation of the German “folglich” (GMS, 454,4) with “consequently” (and the preceding colon) can be misleading. The “folglich” functions as an explanation or elucidation rather than a ‘consequence’ in any stricter sense; I’ll come back to this later.

How is a categorical imperative possible?

313

the world of understanding, the world of understanding is immediately legislative’. What is problematic here is the relation of this part of OP3 to OP1 and OP2, or to be more precise, problematic is both the function of the conjunction ‘because’ at the beginning of OP1 (‘But because …’) as well as the function of the adverb ‘hence’ at the beginning of what we have reconstructed as OP3 (‘hence is immediately legislative …’). Clearly, it would make no sense to interpret Kant as arguing that the world of understanding contains the ground of the world of sense and the ground of its laws and therefore (‘hence’) were immediately legislative for the will; it would not, because this very will already ‘belongs wholly to the world of understanding’ anyway and not to the world of sense. However, this problem of interpretation can be solved by pointing out that the ‘because’ at the beginning of the whole sentence must be put (and read) also at the beginning of OP3, such that one must read: ‘Because the world of understanding also in regard to my will, which belongs wholly to the world of understanding, is immediately legislative …’. The will is part of the world of understanding, and for this world of understanding it is true (according to OP1 and OP2), that it contains the ground of the world of sense and the ground of its laws. In what follows, Kant again avails himself (as he did with the ‘because’) of an ellipsis; what is left out is ‘in regard to my will’. The alleged fact that the will ‘belongs wholly to the world of understanding’ entails that the world of understanding ‘must also be thought of wholly as such’; that is, however, not as a world which is immediately legislative – for this is clear anyway since the will belongs to the world of understanding – but as a world for which it is ‘also’ true in regard to the will as part of it, that it contains the ground of the world of sense and its laws. Thus we get: ‘It is also true in regard to the will as part of the world of understanding that this will contains the ground of the world of sense and its laws’. As we will see later, it is indeed the crux of Kant’s deduction that the pure will as a member of the world of understanding contains the moral law as an imperative for this very will as a member of the world of sense. So we must reconstruct the first part of that sentence as follows: (OP1–3) Because the world of understanding contains the ground of the world of sense; and because hence the world of understanding also contains the ground of the laws of the world of sense; and because hence the world of understanding also in regard to my will, which belongs wholly to the world of understanding, is immediately legislative and must also, in regard to my will, be thought of such that it (the world of understanding) contains the ground of the world of sense and its laws …

314

Dieter Schönecker

The second part of the crucial sentence (GMS, 453,35–454,5) is easier to digest, yet still not a piece of cake. Let’s look at it again: ‘…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.’ It is broken down into two parts: The first part (in its English translation, of course) runs from ‘therefore as’ to ‘autonomy of the will;’ the second from ‘consequently’ to the end (‘as duties’). One might think that Kant in the first part Kant says mainly something like this: ‘therefore as intelligence I will cognize myself […] as […] 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’. Thus one might think that what is stated in the omitted part (‘although on the other side as a being belonging to the world of sense, as nevertheless …’) only becomes relevant in the later second part, because the human being as such a being (that belongs to the world of sense) regards the laws as imperatives. However, I submit that the human being in the entire second part of the ontoethical principle (sentence) must be understood as a being that does not ‘as intelligence’ cognize himself ‘subject’ to the law of the world of understanding and the autonomy of the will; rather it does so ‘as intelligence, although on the other side as a being belonging to the world of sense.’ Three reasons speak in favor of this reading: First, that’s what it says – why put that insertion at the end of the sentence? In a number of parallel passages Kant also emphasizes the ‘simultaneity’ of both perspectives with regard to the imperative character of the moral law. Secondly, inasmuch as the human being considers himself merely as intelligence, he does not cognize himself as ‘subject to’ the moral law; that’s part of the meaning of the thesis of analyticity. This concept of a being that is subject to the moral law is, strictly speaking, reserved for sensual-rational beings; inasmuch as the human being considers himself as intelligence, he considers oneself ‘merely as a member of the world of understanding’. Thirdly (and related to the last point), it is striking that Kant says that the human being cognizes himself ‘as nevertheless (‘obgleich’) subject to’ the moral law. For a being that considers itself only as intelligence, there is no need to consider itself as ‘nevertheless’ subject to the moral law – for what kind of difference could the ‘nevertheless’ indicate if not the difference between being a member of the world of understanding and the world

How is a categorical imperative possible?

315

of sense?14 Also, the ‘consequently’ (‘folglich’) only makes sense if it is related to that insertion. With regard to the first part of the ‘consequently’-sentence, I’d like to emphasize this. The ‘first’ (‘ersteren’) can only refer to the world of understanding at the beginning of the sentence; in the later part of the sentence the ‘laws’ of that ‘world of understanding’ are mentioned again. However, Kant connects the ‘first’ with the concept of ‘reason’ (‘Vernunft’) by means of an ‘i. e.’, and since the German ‘derselben’ cannot refer back to this very ‘reason’, one must read: ‘… to the law of reason, which in the idea of freedom contains the law of the understanding’s world.’ Kant then also mentions the ‘autonomy of the will’, and from all of this it follows that I cognize myself as subject to (1.) the law of the world of understanding, (2.) to reason which contains in the idea of freedom the law of the world of understanding, and (3.) to the autonomy of the will. Thus we can reconstruct the first part of the ‘therefore’-sentence as follows: (OP4) Understanding myself as a being belonging both to the world of understanding (intelligence) and to the world of sense, I cognize myself as subject to the law of the world of understanding, i. e. to reason, which in the idea of freedom contains the law of the world of understanding, and thus as subject to the autonomy of the will.

The rest of the sentence is easy. It, too, makes clear that the moral law, which describes the volition of perfectly rational beings, is an imperative for sensuous-rational beings: (OP5) I must regard the laws of the world of understanding for myself as imperatives and the actions that accord with this principle as duties.

Connecting these elements and adding the conjunctive particles, we thus get the ontoethical principle: (OP) Because the world of understanding contains the ground of the world of sense; and because hence the world of understanding also contains the ground of the laws of the world of sense; and because hence the world of understanding also in regard to my will, which belongs wholly to the world of understanding, is immediately legislative and must also, in regard to my will, be thought of such that it (the world of understanding) contains the ground of the world of sense and its laws, I cognize myself – understanding myself as a being belonging both to the world of understanding (intelligence) and to the world of sense – as subject to the law of 14

Cf. the elucidating parallel in GMS, 450,12–13 (“… consider ourselves as free in acting and thus nevertheless take ourselves to be subject to certain laws …”, m. e.).

316

Dieter Schönecker

the world of understanding, i. e. to reason, which in the idea of freedom contains the law of the world of understanding, and thus as subject to the 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.15

Thus the deduction of the categorical imperative is completed in just one sentence. This deduction is Kant’s answer to the question of how a categorical imperative is possible. Therefore, it is no surprise that right after stating OP, Kant officially gives the answer to that question as well: 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 first in accordance with reason (GMS, 454,6–15, m. e.).

Clearly, the deduction is not a deduction in any strict (deductive) sense (at least not as long as the reconstruction sticks to the original text). What then, exactly, is the key idea in Kant’s argument? The human being is aware of himself as an intelligence due to the spontaneous epistemic activities of his reason and understanding; as such an intelligence, the human being is the ‘eigentliche Selbst’ (‘authentic self’; GMS, 457,34; 458,2; 461,4). From there, Kant goes on (“folglich”; GMS, 448,18) to the ‘intelligence with a will’. Considering himself as an intelligence, the human being understands himself as a member of the world of understanding and thus as a thing in itself; its rational will, then, constitutes the ‘eigentliche Selbst’ as a practical being. This idea – again, that the will as an intelligible faculty is the ‘eigentliche Selbst’ of the human being, as opposed to the human being inasmuch as he is “only appearance of himself” (GMS, 457,35, m. e.), i. e. only a “phenomenon in the world of sense“ (GMS, 457,13) – is the core of OP. Kant argues for the validity of the categorical imperative as a moral 15

Cf. R 5086: „In der Verstandeswelt ist das substratum: intelligentz, die Handlung und Ursache: Freyheit, die Gemeinschaft: Glückseligkeit aus Freyheit, das Urwesen: eine Intelligentz durch idee; die form: moralitaet, der nexus: ein nexus der Zweke. Diese Verstandeswelt liegt schon itzt der Sinnenwelt zum Grunde und ist das wahre selbstandige“ (m. e., andere Hervorhebungen getilgt).

How is a categorical imperative possible?

317

law for sensual-rational beings with the superiority of the ontic status of the world of understanding. The human being as a thing in itself and hence the ‘eigentliche Selbst’ and its law is of higher ontic value then the human being as an appearance; and this is why the law of the world of understanding (the moral law) is binding upon the human being (as a categorical imperative) who is a member both of the world of understanding and the world of sense. That’s the basic idea behind OP. That this is, indeed, the basic idea behind OP can hardly be seen just by reading OP itself (that sentence in GMS, 453). We have to look at other passages. An external characteristic may lead the way: In that passage in GMS, 453, OP1 and OP2 are emphasized, an emphasis that at this length, as far as I know, hardly ever can be found in Kant’s writings (if at all). Exactly parallel to this, Kant again provides the answer to the question of how a categorical imperative is possible in GMS, 461. And just as in GMS, 453 f., he again avails himself of a lengthy emphasis to stress that the moral law is valid [!] for us as [!] human beings, since [!] 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 (GMS, 461,2–6).

The law of the world of appearances is the “natural law of desires and inclinations” (GMS, 453,28); these ‘desires and inclinations’, as “appearances” (GMS, 453,24, m. e.), determine human actions. However, this law just belongs ‘to the mere appearance’ and therefore ‘is necessarily subordinated by reason to the constitution of the thing in itself’. Here again it becomes clear that the whole force of Kant’s argument depends on the ontic superiority of the ‘authentic self’. And there is yet another passage that provides textual evidence for this interpretation. The human being, Kant says, as a rational being is a member of the world of understanding, and 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 (GMS, 457,33–36, m. e.)

In other words, the moral law is binding upon the human being because it stems from the pure will as the authentic self, which, as such, is of higher ontic value (not ‘only appearance of himself’). Even in the Critique of Practical Reason, in which Kant denies the possibility of a deduction of the categorical imperative, he says that it is the status

318

Dieter Schönecker

as a member of the world of understanding which “elevates a human being above himself (as a part of the sensible world)“ (KpV, 86, m. e.); hence “it is not to be wondered at that a human being, as belonging to both worlds, must regard his own nature in reference to his second and highest vocation only with reference, and its laws with the highest respect” (KpV, 87, m. e.). This, Kant says, is the “origin” (KpV, 86) of duty. This axiological position is also reflected in the last paragraph of Sec. 4 of the GMS, where Kant writes that “even the most wicked scoundrel […] transports himself in thoughts into entirely another order of things” (GMS, 454,21–31) in which he finds “a greater inner worth of his person” (GMS, 454,37): “This better person, however, he believes himself to be when he transports himself into the standpoint of a member of the world of understanding” (GMS, 454,37–455,2) Thus we can reconstruct the ontoethical principle as follows: (OP*) The world of understanding and thus the pure will as a member of this world of understanding are ontically superior to the world of sense, and therefore the law of this world and will (the moral law) is binding as a categorcial imperative for beings that are both members of the world of understanding and the world of sense.16 16

Steigleder (2006, this volume, pp. 225–246) rejects my interpretation of Kant’s thesis of analyticity and therefore also my interpretation of Sec. 4. Although it would be most interesting to have a detailed discussion, it is simply impossible to have it here for lack of space. Hence just three brief comments: First, no interpretation can be satisfying that is not comprehensive; it’s always easy, too easy indeed, to make claims about what a text means by ignoring those passages that don’t fit. (Thus, Steigleder is right to point out that there are elements in Sec. 1 that pose a problem for my interpretation; I have addressed these elements in Schönecker, 1999.) This being said, I’d like to reply, second, that Steigleder correctly points out that my interpretation is partly based on the overall structure of GMS III (that at least part of GMS II and GMS III are ‘preparation’ and that there is a Sec. 4 – to say the very least). However, it is certainly not sufficient to reply to this crucial element of my interpretation by just asserting that there “may conceivably be other readings” (Steigleder 2006, 242, m. e.) without actually providing an alternative reading that is also in a position to account for the overall structure of GMS III. As did many before him, Steigleder still argues that it is only Kant’s intention that we “must necessarily see ourselves as ‘rational beings, which have a will’” (Steigleder, 2006, p. 243). Clearly, however, this has been demonstrated no later than in Sec. 3. But what then is the purpose of Sec. 4 and that long and complicated sentence that is (or includes) what I call the ontoethical principle? Steigleder provides no answer, and it is therefore no surprise that in his book (2002, 67–96), Steigleder shows no interest whatsoever in that sentence either (he does, however, very briefly mention that the “Gesetz unserer Vernunft und das Verlangen unserer Bedürfnisse […] nicht auf gleicher Stufe stehen”, p. 89, m. e.). There may very well be other readings of it; but whether there are any, we will not figure out by ignoring it. Leaving aside that Steigleder, as I see it, pays no sufficient attention to Kant’s repeated claim that there is no categorical imperative for perfectly rational beings and does not sufficiently distinguish between the moral law as an ‘synthetic’ imperative and as an ‘analytical’

How is a categorical imperative possible?

319

2.3 Validity and Motivation I already pointed out that, according to Kant, the good is something which the human being, in some sense and respect, ‘authentically wills’. I would now like to come back to this idea. Recall that the deduction of the categorical imperative is the answer to the question of how a categorical imperative is possible. This question, we said, must be understood as the question of why the categorical imperative is valid (binding, obligatory), or more simply: Is the CI valid? Is there a good reason to abide by it? Kants himself asks ‘from whence the moral law obligates’.17 Traditionally, this question has often been formulated as the question ‘Why be moral?’, and this so-called Moral Question, I submit, can have only two possible answers: Either the answer refers to self-interest, such that the reason why one ought to act morally is that doing so serves one’s self-interest, at least in the long run; or the answer is that indeed there is a moral law, or, in a rather axiological language, that goodness does exist and makes demands on us, and that the moral law (or goodness) itself is the always overriding reason to act morally.18 Clearly, the first answer is unacceptable to Kant. It is one of Kant’s fundamental claims that the CI “does not have validity for us because it interests us” (GMS, 460,24), and given Kant’s overall understand-

17

18

law (though this distinction is, obviously, at the very heart of Kant’s question of how a categorical imperative is possible), I’d like to note, third, that Steigleder does not have an explanation for Kant’s repeatedly posed question ‘from whence the moral law obligates’. I do, by the way, agree that it is certainly worthwhile to consider the option that Kant in the Groundwork already distinguishes, as Steigleder suggests, between ‘Wille’ and ‘Willkür’; but again, this issue I cannot possibly address here. In the preface Kant already mentions the Grund der Verbindlichkeit of the CI (‘ground of an obligation’; cf. GMS, 389,12; 389,16; 391,11; 432,31; 439,31; 439,33; 448,34). He also speaks of the Realität of the CI (reality; GMS, 425,14; 449,26), of its Wirklichkeit (‘reality’; cf. vgl. 420,1; vgl. 406,15), Geltung (‘validity’; cf. GMS, 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; 461,12), Richtigkeit (‘correctness’; cf. GMS, 392,13), objektiven Notwendigkeit (‘objective necessity’; cf. GMS 442,9; 449,26; 449,30); Kant says that the CI gibt (is; cf. 419,18), that it wirklich stattfinde (‘is’; cf. GMS, 425,9) and that the human being ought to unterwerfen himself to it (‘subject himself to it’; cf. GMS, 449,12). All this concepts and formulations are probably best subsumed under the idea of the Gültigkeit dieses Imperativs (‘validity of this imperative’; cf. GMS 461,12). Also note that Kant not only speaks of the deduction of the CI, but also of its ‘Beweis’ etc. (‘proof’; cf. GMS, 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; 449,27). For our context, I assume that answers which could be classified as ‘formalistic’ (such as Karl-Otto Apel’s or Habermas’) fail from the word go; but they would classify as a third possible answer. On the Moral Question, cf. Schönecker (2006).

320

Dieter Schönecker

ing of morality developed in GMS I/II, this question, if understood as ‘Why be moral?’, obviously cannot be asking for the utility of the CI; it cannot ask for how the CI might serve one’s interests. It is important to see, though, that although Kant goes out of his way to argue that the question of how pure practical reason can be practical cannot be answered, he is not only interested in the validity of the CI, but also in its power to motivate. In this respect, he points out that pure practial reason (the pure will) is still one’s own will. Kant argues that, if this will and its law were not one’s own will, and hence the law were not one’s own law, one still were to find a nonmoral motive to actually comply with this law, i. e. a motive to want (or to do, respectively) what one would not want to do: “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” (GMS, 432,32–433,3). In the ‘metaphysics of morals’ as part of GMS II, Kant had already asked for the “possibility” (GMS, 427,17) of a will that is merely determined by reason and thus asked for the “ground of a possible categorical imperative, i. e. of a practical law“ (GMS, 428,5). This ‘ground’, he says, can only be “something whose existence in itself has an absolute worth“ (GMS, 428,3) – a rational and free being as an end in itself. Since the absolute worth of rational beings as ends in themselves is grounded in their autonomy,19 Kant then – in order not to beg the question (for the law of this autonomy is the CI the very validity of which is suspicious) – argues in GMS III with the ontoethical principle and the ontic superiority of the pure will of the ‘eigentliche Selbst’. In the context of GMS II, Kant sets forth the “postulate” (GMS, 429,35) that every rational being must understand his or her own being as such an ‘end in itself’ (a postulate that is then demonstrated in sections 2 and 3 of GMS III). He already provides arguments that this status justifies why one must, and how one can, subject oneself to the moral law as a categorical imperative: “The will is thus not solely subject to the law, but is subject in such a way that it must be regarded also as legislating to itself, and precisely for this reason as subject to the law (of which it can consider itself as the author)” (GMS, 431,21, second emphasis D. S.). It is because the human being gives himself the universal law 19

Cf. Kant’s lecture Feyerabend (1319 ff.). I cannot get into this but one should not forget that an important variant of the CI is based on the idea of the human being as an end in itself; cf. Schönecker / Wood (22004, 140–153).

How is a categorical imperative possible?

321

that this law can motivate us: “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” (GMS, 455,7, m. e.); this is why Kant so often emphasizes that “this ‘ought’ is really a volition” (GMS, 449,16, m. e.). Thus Kant not only attempts to demonstrate that there is a good reason to be moral, but at the same time – and this is partly responsible for the confusion GMS can easily cause – the deduction of the moral law provides an incentive to be moral: „Das Gute ist immer das, was ieder Mensch will, und er würde es auch immer thun, wenn es ihm nur nicht schwer würde, es auszuüben, und wenn unsere Natur so beschaffen wäre, daß wir immer nach dem Begriff des guten handelten, so wären wir recht frei“. 20 The moral insight that Kant wants to induce with the addressee of the moral law is that he, the addressee himself, in some sense already wants what he ought to do. 2.4 Some brief critical points As mentioned before, it is not the purpose of this paper to criticize (or further develop) Kant’s deduction. However, let us briefly look at some critical points which will also help us to better understand Kant’s argument. I don’t find it convincing at all, neither from an external nor from an internal point of view. Externally speaking, I would criticize Kant’s axiology as much too narrow because it allows only rational beings to have worth; but that’s a long story and, in any case, not the issue here. Internally speaking (i. e. presuming Kant’s own critical philosophy), it is quite obvious that Kant avails himself of an ontological interpretation of his own distinction between thing in itself and appearance that otherwise is merely a epistemological distinction. Let’s have a look at OP one more time. The first part of the ontoethical principle (OP1–3), we said, must be reconstructed as follows: ‘Because the world of understanding contains the ground of the world of sense; and because hence the world of understanding also contains the ground of the laws of the world of sense; and because hence the world of understanding also in regard to my will, which belongs wholly to the world of understanding, is immediately legislative and must also, in regard to my will, be thought of such that it (the world of understanding) contains the ground of the world of sense and its laws …’. As seen earlier in our detailed analysis, it seems that Kant has a general principle in 20

MM, 903.

322

Dieter Schönecker

mind according to which the world of understanding is the ground of the laws of the world of sense which is only applied to the special case of the will; hence what Kant would be saying is that the pure will, which is a member of the world of understanding, contains the ground for the laws of the will as a member of the world of sense. And as a matter of fact, right after the deduction in Sec. 4, Kant says that the good will of an evil man “constitutes by his own admission the law for his evil will as a member of the sensible world” (GMS, 455,4–6). But how are we to understand this ‘general principle’? Kant himself suggests in Sec. 4 reading it in light of his epistemological assertion that the logical subject is the ground for the laws of nature. However, that the world of understanding is the ground of the laws of the world of sense does not follow, as suggested by the ‘hence’ in OP2, from Kant’s own fundamental claim, that one must “assume behind the appearances something else that is not appearance, namely the things in themselves” (GMS, 451,12–14). On the other hand, even if one accepted that somehow the logical subject is the ground of the laws of the world of sense in terms of that epistemological assertion of Kant’s, how are we to understand the application of that ‘general principle’ on the will? At best, it seems, there is some kind of resemblance between the legislation of the logical subject and the legislation of the practical subject. And yet it is this resemblance that Kants seems to have in mind. For he closes his official answer to the question of how a categorical imperative is possible by drawing a parallel between OP1–3 and the aforementioned epistemological claim, saying that “it is approximately in this way that concepts of the understanding, which for themselves signify nothing but lawful form in general, are added to intuitions of the world of sense and through that make possible synthetic propositions a priori on which rests all cognition of a nature” (GMS, 454,15–19, m. e.). Well, Kant himself admits that this parallel is only ungefähr (approximately). All the more we have to rely in our interpretation of OP on later passages in which Kant emphasizes the ontic superiority of the world of understanding (GMS, 457; 461). However, not only is this alleged superiority in itself dubious, but it is also not in harmony with Kant’s overall understanding of the distinction between things in themselves and appearances. According to this understanding, there is, in itself, one world. This world (of things in themselves) is what it is; this very world as understood (interpreted) by us, is called the world of sense. The latter is not, however, in any sense inferior to the former unless one understands the fact that appearances, except from being appear-

How is a categorical imperative possible?

323

ances, do not exist as establishing their inferiority; if so, one would need to draw the awkward conclusion that inclinations (as appearances) are not only inferior but not even real. Yet on Kant’s epistemological distinction between things in themselves and appearances, practical reason and its law (autonomy) is just as real as inclinations and their law (heteronomy). It could very well be that, axiologically speaking, practical reason is an end in itself and thus of absolute value (as Kant postulates in GMS II). To argue, however, that, indeed, it is (as Kant does in GMS III), based on the alleged ontic superiority of things in themselves in contrast with appearances, not only makes little sense on its own, but also cannot be reconciled with Kant’s own fundamental epistemology.

Literature Kant’s writings All textual references of the Groundwork are according to Allen Wood’s translation and edition (Yale University Press, 2002). All references to the German text of the Grundlegung are to the edition by B. Kraft and D. Schönecker (Felix Meiner Verlag, 1999). All other textual references are to The Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press, 1992–). Numbers in paranthesis refer to the AA. Feyerabend Naturrecht Feyerabend, AA 27 GMS Grundlegung zur Metaphysik der Sitten, AA, IV KpV Kritik der praktischen Vernunft, AA, V MM Metaphysik Mrongovius, AA 29

Other works Damschen, Gregor / Schönecker, Dieter (2006): Selbst philosophieren, Stuttgart. Quarfood, Marcel (2006): “The Circle and the two standpoints”, in: This volume, pp. 285–300. Schönecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs, Freiburg i. Br./München. Schönecker, Dieter (2001): “Textvergessenheit in der Philosophiehistorie“, in: Schönecker, Dieter / Zwenger, Thomas (Hrsg.): Kant verstehen/Understanding Kant. Über die Interpretation philosophischer Texte, 159–181. Schönecker, Dieter (2006): “Warum moralisch sein? Eine Landkarte für Moralische Realisten”, in: Heiner F. Klemme / Manfred Kühn / Dieter Schönecker (eds.): Moralische Motivation. Kant und die Alternativen. Felix Meiner Verlag, Hamburg.

324

Dieter Schönecker

Schönecker, Dieter / Wood, Allen W. (22004): Immanuel Kant: ‘Grundlegung zur Metaphysik der Sitten’. Ein einführender Kommentar, Paderborn. Steigleder, Klaus (2002): Kants Moralphilosophie. Die Selbstbezüglichkeit reiner praktischer Vernunft, Stuttgart. Steigleder, Klaus (2006): “The Analytic Relationship of Freedom and Morality – On GMS III, 1”, in this volume, pp. 225–246. Wood, Allen W. (1999): Kant’s Ethical Thought, Cambridge University Press.

A Bibliography on Kant’s Groundwork The following bibliography is an updated version of the bibliography that was made available by Bernd Kraft and Dieter Schönecker in their edition of the Grundlegung zur Metaphysik der Sitten with Felix Meiner Verlag, Hamburg (1999). This updated bibliography takes into account literature published between 2000 und 2005. Like Kraft’s and Schönecker’s, it is committed to the idea to include only publications that somewhat specifically refer to Kant’s Groundwork rather than to Kant’s ethics in general.1 Acton, Harry B. (1970): Kant’s Moral Philosophy. New York. Allison, Henry E. (1990): Kant’s Theory of Freedom. Cambridge. Allison, Henry E. (1996): On the Presumed Gap in the Derivation of the Categorical Imperative. In: Ders.: Idealism and Freedom. Essays on Kant’s Theoretical and Practical Philosophy, Cambridge, 143–154. Alquié, Ferdinand (1974): La morale de Kant. Paris. Alquié, Ferdinand (1985): Présentation des Fondements de la Métaphysique de Mœurs. Tome II, Paris. Albrecht, Michael (1994): Kants Maximenethik und ihre Begründung. In: Kant-Studien 85, 129–146. Ameriks, Karl (22000): Kant’s Theory of Mind. An Analysis of the Paralogisms of Pure Reason. Oxford. Ameriks, Karl (2001): Zu Kants Argumentation am Anfang des Dritten Abschnitts der Grundlegung. In: Hans-Ulrich Baumgarten / Carsten Held (Hrsg.): Systematische Ethik mit Kant, Freiburg i. Br./München, 24–54. Ameriks, Karl / Sturma, Dieter (Hrsg.) (2004): Kants Ethik. Paderborn. Atwell, John E. (1969): Are Kant’s First Two Moral Principles Equivalent? In: Journal of the History of Philosophy 7, 273–284. Aubenque, Pierre (1975): La Prudence chez Kant. In: Revue de Metaphysique et Morale, 156–182. Aune, Bruce (1979): Kant’s Theory of Morals. Princeton. Baker, Judith (1988): Counting Categorical Imperatives. In: Kant-Studien 79, 389–406. Barnes, Gerald W. (1971): In Defense of Kant’s Doctrine of the Highest Good. In: Philosophical Forum 2, 446–458. Baron, Marcia (1995): Kantian Ethics almost without Apology. Ithaca. Baumanns, Peter (1982): Kants kategorischer Imperativ und das Problem der inhaltlichen Pflichtbestimmung. In: Herta Nagl-Docekal (Hrsg.): Über1

Thanks to Raoul Simon Weber (Bonn) for important help in updating the bibliography.

326

Bibliography

lieferung und Aufgabe. Festschrift für Erich Heintel zum 70. Geburtstag. 2. Teilband, Wien, 165–179. Baumanns, Peter (2000): Kants Ethik: Die Grundlehre. Würzburg. Baumgarten, Hans-Ulrich / Held, Carsten (Hrsg.) (2001): Systematische Ethik mit Kant, Freiburg, i. Br./München. Benton, Robert J. (1978): The Transcendental Argument in Kant’s Groundwork of the Metaphysic of Morals. In: The Journal of Value Inquiry (The Hague) 12, 225–237. Bittner, Rüdiger (1974): Maximen. In: Gerhard Funke (Hrsg.): Akten des 4. Internationalen Kant-Kongresses, Berlin/New York, 485–498. Bittner, Rüdiger (1980): Hypothetische Imperative. In: Zeitschrift für philosophische Forschung 34, 210–226. Bittner, Rüdiger (1983): Moralisches Gebot oder Autonomie. Freiburg i. Br./ München. Bittner, Rüdiger (1986): Handlungen und Wirkungen. In: Gerold Prauss (Hrsg.): Handlungstheorie und Transzendentalphilosophie, Frankfurt a. M., 13–26. Bittner, Rüdiger (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 Bd. 3 (Philosophie der Neuzeit / From Descartes to Kant), 209–221. Bollnow, Otto Friedrich (1966): „Als allein ein guter Wille …“. Zum Anfang der Grundlegung zur Metaphysik der Sitten. In: Wolfgang Ritzel (Hrsg.): Rationalität – Phänomenalität – Individualität. Festgabe für Hermann und Marie Glockner, Bonn, 165–174. Brandt, Reinhard (1988): Der Zirkel im dritten Abschnitt von Kants Grundlegung zur Metaphysik der Sitten. In: Hariolf Oberer / Gerhard Seel (Hrsg.): Kant. Analysen – Probleme – Kritik, Würzburg, 169–191. Brinkmann, Walter (2001): Die Goldene Regel und der Kategorische Imperativ: Rationalität und praktische Notwendigkeit. In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.): Kant und die Berliner Aufklärung, Bd. 3, Berlin/New York, 13–20. Bruton, Samuel V. (2000): Establishing Kant’s Formula of Humanity. In: Southwest Philosophy Review 16, 41–49. Bubner, Rüdiger (2001): Another Look at Maxims. In: Predrag Cicovacki (Hrsg.): Kant’s Legacy: Essays in Honour of Lewis White Beck, Rochester, 245–259. Burri, Alex / Freudiger, Jürg (1990): Zur Analytizität hypothetischer Imperative. In: Zeitschrift für philosophische Forschung 44, 98–105. Canivet, Michel (1980): Enseigner les Fondements de la métaphysique des mœurs. In: Revue Philosophique de Louvain 78, 365–384. Carnois, Bernard (1973): La Cohérence de la Doctrine Kantienne de la Liberté. Paris. Cicovacki, Predrag (2001): Zwischen gutem Willen und Kategorischem Imperativ. In: Hans-Ulrich Baumgarten / Carsten Held (Hrsg.): Systematische Ethik mit Kant, Freiburg i. Br./München, 330–354. Copp, David (1992): The ‘Possibility of a Categorical Imperative’ Kant’s Groundwork, Part III. In: Philosophical Perspectives 6, 261–284.

Bibliography

327

Craemer-Ruegenberg, Ingrid (1983): Logische und andere Eigenschaften des kategorischen Imperativs. In: Neue Hefte für Philosophie 22, 45–61. Cramer, Konrad (1972): Hypothetische Imperative? In: Manfred Riedel (Hrsg.): Rehabilitierung der praktischen Philosophie. Bd. 1, Freiburg i. Br., 159–212. Cramer, Konrad (1991): Metaphysik und Erfahrung in Kants Grundlegung der Ethik. In: Neue Hefte für Philosophie 30/31, 15–68. Croitori, Rodica (1995): Kant and Totalitarianism from the Point of View of the Groundwork of the Metaphysics of Morals. In: Hoke Robinson (Hrsg.): Proceedings of the Eighth International Kant Congress (Memphis), Milwaukee. Volume II, Part 2, 813–820. De Bolt, Darian C. (2002): Pasternack on Intrinsic Value and Moral Overridingness in Kant’s Groundwork. In: Southwest Philosophy Review: The Journal of the Southwestern Philosophical Society 18, 121–125. Delbos, Victor (1969): La philosophie practique de Kant. Paris. Delbos, Victor (1993): La moral de Kant (Nachwort zu: Immanuel Kant: Fondements de la métaphysique des mœurs. Traduction et notes par Victor Delbos). Paris, 153–202. Delfosse, Heinrich P. (2000): Kant-Index, Band 15: Stellenindex und Konkordanz zur ‘Grundlegung zur Metaphysik der Sitten’, Stuttgart/Bad Cannstatt. Denis, Lara (2001): Moral Self-Regard: Duties to Oneself in Kant’s Moral Theory. New York. Duncan, Alistair Robert Campbell (1957): Practical Reason and Morality. A Study of Immanuel Kant’s Foundations for the Metaphysics of Morals. London/New York. Dye, James W. (1978): Kant as Ethical Naturalist. In: The Journal of Value Inquiry 12, 111–125. Ebbinghaus, Julius (1986): Deutung und Mißdeutung des kategorischen Imperativs. In: Ders.: Gesammelte Schriften. Band I, Praktische Philosophie 1929–1954, herausgegeben von Hariolf Oberer und Georg Geismann, Bonn, 279–295. Ebbinghaus, Julius (1988): Die Formeln des kategorischen Imperativs und die Ableitung inhaltlich bestimmter Pflichten. In: Ders.: Gesammelte Schriften. Band II, Praktische Philosophie 1955–1972, herausgegeben von Georg Geismann und Hariolf Oberer, Bonn, 209–229. Engstrom, Stephen (1997): Kant’s Conception of Practical Wisdom. In: KantStudien 88, 16–43. Ertl, Wolfgang (2001): Das Freiheitsproblem bei Kant: Zwischen Tradition und Innovation. In: Jens Kulenkampff / Thomas Spitzley (Hrsg.): Von der Antike bis zur Gegenwart: Erlanger Streifzüge durch die Geschichte der Philosophie, Erlangen, 79–96. Fleischer, Margot (1963): Das Problem der Begründung des kategorischen Imperativs bei Kant. In: Paulus Engelhardt (Hrsg.): Sein und Ethos. Untersuchungen zur Grundlegung der Ethik, Mainz, 387–404. Fleischer, Margot (1964): Die Formeln des kategorischen Imperativs in Kants Grundlegung zur Metaphysik der Sitten. In: Archiv für Geschichte der Philosophie 46, 201–226.

328

Bibliography

Forschner, Maximilian (1974): Gesetz und Freiheit. Zum Problem der Autonomie bei I. Kant. München/Salzburg. Förster, Eckart (1992): „Was darf ich hoffen?“ Zum Problem der Vereinbarkeit von theoretischer und praktischer Vernunft bei Immanuel Kant. In: Zeitschrift für philosophische Forschung 46, 168–185. Fouillée, Alfred (1904): Kant a-t-il établi l’existence du devoir? Revue de la Métaphysique et de Morale 12, 493–523. Freudiger, Jürg (1993): Kants Begründung der praktischen Philosophie. Systematische Stellung, Methode und Argumentationsstruktur der Grundlegung zur Metaphysik der Sitten. Bern/Stuttgart/Wien. Galay, Jean-Louis (1977): Philosophie et invention textuelle. Essai sur la poétique d’un texte kantien. Paris. Gass, Michael (1994): Kant’s Causal Conception of Autonomy. In: History of Philosophy Quarterly 11, 53–70. Geismann, Georg (2002): Die Formeln des kategorischen Imperativs nach H. J. Paton, N. N., Klaus Reich und Julius Ebbinghaus. In: Kant-Studien 93, 374–384. Geismann, Georg (Hrsg.) (2002): Julius Ebbinghaus – Die NaturgesetzFormel des kategorischen Imperativs und die Ableitung inhaltlich bestimmter Pflichten. In: Kant-Studien 93, 371–373. Genova, A. C. (1978): Kant’s Transcendental Deduction of the Moral Law. In: Kant-Studien 69, 299–313. Glass, Ronald (1971): The Contradictions in Kant’s Examples. In: Philosophical Studies 22, 65–70. Green, Ronald M. (2001): What Does It Mean to Use Someone As ‘A Means Only’: Rereading Kant. In: Kennedy-Institute of Ethics Journal 11, 247– 261. Grünewald, Bernward (1988): Praktische Vernunft, Modalität und transzendentale Einheit. Das Problem einer transzendentalen Deduktion des Sittengesetzes. In: Hariolf Oberer / Gerhard Seel (Hrsg.): Kant. Analysen – Probleme – Kritik, Würzburg, 127–167. Gunkel, Andreas (1989): Spontaneität und moralische Autonomie. Kants Philosophie der Freiheit. Bern/Stuttgart. Gupta, R. K. (1971): Kant’s Groundwork of Morality. In: Studi Internazionali di Filosofia 3, 111–161. Gupta, R. K. (1973): Kant’s Problem of the Impossibility of the Categorical Imperative. In: Kant-Studien 64, 49–55. Guyer, Paul (1995): The Possibility of the Categorical Imperaive. In: The Philosophical Review 104, 353–385. Guyer, Paul (Hrsg.) (1998): Kant’s Groundwork of the Metaphysics of Morals: Critical Essays. Totowa. Guyer, Paul (2001): The Form and Matter of the Categorial Imperative. In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.): Kant und die Berliner Aufklärung, Bd. 1, Berlin/New York, 131–150. Guyer, Paul (2002): The Derivation of the Categorical Imperative: Kant’s Correction for a Fatal Flaw. In: Harvard Review of Philosophy 10, 64–80. Haardt, Alexander (1982): Die Stellung des Personalitätsprinzips in der Grundlegung zur Metaphysik der Sitten und in der Kritik der praktischen Vernunft. In: Kant-Studien 73, 157–168.

Bibliography

329

Harbison, Warren (2000): Self-Improvement, Beneficence, and the Law of Nature Formula. In: Kant-Studien 91, 17–24. Harris, N. G. E. (1988): Imperfect Duties and Conflicts of Will. In: Kant-Studien 79, 33–42. Harrison, Jonathan (1957): Kant’s Examples of the First Formulation of the Categorical Imperative. In: The Philosophical Quarterly 7, 50–62. Heidemann, Ingeborg (1956): Person und Welt. Zur Kantinterpretation von Heinz Heimsoeth. In: Kant-Studien 48, 344–360. Henrich, Dieter (1954/55): Das Prinzip der Kantischen Ethik. In: Philosophische Rundschau 2, 20–38. Henrich, Dieter (1973): Der Begriff der sittlichen Einsicht und Kants Lehre vom Faktum der Vernunft. In: Gerold Prauss (Hrsg.): Kant. Zur Deutung seiner Theorie von Erkennen und Handeln, Köln, 223–254. 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 (Hrsg.): Denken im Schatten des Nihilismus. Festschrift für Wilhelm Weischedel zum 70. Geburtstag, Darmstadt, 55–112. Henrich, Dieter (1982): Ethik der Autonomie. In: Ders.: Selbstverhältnisse. Gedanken und Auslegungen zu den Grundlagen der klassischen deutschen Philosophie, Stuttgart, 6–56. Herman, Barbara (1993): The Practice of Moral Judgement. Cambridge. Hill Jr., Thomas E. (1992): Dignity and Practical Reason in Kant’s Moral Theory. Ithaca. Hill Jr., Thomas E. / 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 und 109–177. Hochberg, Gary M. (1973): A Re-Examination of the Contradiction in Kant’s Examples. In: Philosophical Studies 24, 264–267. Hoerster, Norbert (1974): Kants kategorischer Imperativ als Test unserer sittlichen Pflichten. In: Manfred Riedel (Hrsg.): Rehabilitierung der praktischen Philosophie. Bd. 2: Rezeption, Argumentation, Diskussion, Freiburg i. Br., 455–475. Höffe, Otfried (1977): Kants kategorischer Imperativ als Kriterium des Sittlichen. In: Zeitschrift für philosophische Forschung 31, 354–384. Höffe, Otfried (1985): Introduction à la philosophie pratique de Kant. La morale, le droit et la religion. Albeuve. Höffe, Otfried (Hrsg.) (1989): Grundlegung zur Metaphysik der Sitten. Ein kooperativer Kommentar. Frankfurt a. M. Högemann, Brigitte (1980): Die Idee der Freiheit und das Subjekt. Eine Untersuchung von Kants Grundlegung zur Metaphysik der Sitten. Meisenheim am Glan – Königstein/Ts. Horn, Christoph (2004): Die Menschheit als objektiver Zweck – Kants Selbstzweckformel des Kategorischen Imperativs. In: Dieter Sturma (Hrsg.), Kants Ethik, Paderborn, 195–212. Hunter, Ian (2002): The Morals of Metaphysics: Kant’s Groundwork as Intellectual Paideia. In: Critical Inquiry 28, 908–929.

330

Bibliography

Hutchings, Patrick A. (1972): Kant on Absolute Value. A Critical Examination of Certain Key Notions in Kant’s Groundwork of the Metaphysics of Morals and of his Ontology of Personal Value. Detroit. Ilting, Karl-Heinz (1972): Der naturalistische Fehlschluß bei Kant. In: Manfred Riedel (Hrsg.): Rehabilitierung der praktischen Philosophie. Bd. 1: Geschichte, Probleme, Aufgaben, Freiburg i. Br., 113–130. Irrlitz, Gerd (2002): Kant-Handbuch: Leben – Werk – Wirkung. Stuttgart. Jacobsen, Mogens Chrom (1992): Über das Verhältnis zwischen Immanuel Kants Rechts- und Moralphilosophie. In: Danish Yearbook of Philosophy 27, 72–92. Johnson, Darrell (1995): Analytic and Synthetic Method and the Structure of Kant’s Grounding. In: Hoke Robinson (Hrsg.): Proceedings of the Eighth International Kant Congress (Memphis), Milwaukee. Volume II, Part 2, 613–620. Kaulbach, Friedrich (1988): Immanuel Kants Grundlegung zur Metaphysik der Sitten. Interpretation und Kommentar. Darmstadt. Kemp, J. (1958): Kant’s Examples of the Categorical Imperative. In: The Philosophical Quarterly 8, 63–71. Kerstein, Samuel J. (2002): Kant’s Search for the Supreme Principle of Morality. Cambridge. Kersting, Wolfgang (1983): Kann die Kritik der praktischen Vernunft populär sein? Über Kants Moralphilosophie und pragmatische Anthropologie. In: Studia Leibnitiana 15, 82–93. King, J. Charles (1968): Bradley’s ‘Duty for Duty’s Sake’ and Kant’s Ethics. In: Kant-Studien 59, 309–317. Kirchmann, Julius Hermann v. (1874): Erläuterungen zur Grundlegung der Metaphysik der Sitten. Leipzig. Klamp, Gerhard (1949): Zum Problem der Textverschiebung in Kants Grundlegung zur Metaphysik der Sitten. In: Zeitschrift für philosophische Forschung 4, 574–578. Klein, Sherwin (1988): Kant’s Methodology in Grundlegung, Section One. In: The Modern Schoolman 65, 227–244. Köhl, Harald (1990): Kants Gesinnungsethik. Berlin/New York. Konhardt, Klaus (1986): Faktum der Vernunft? Zu Kants Frage nach dem „eigentlichen Selbst“ des Menschen. In: Gerold Prauss (Hrsg.): Handlungstheorie und Transzendentalphilosophie, Frankfurt a. M., 160–184. Korsgaard, Christine M. (1996): From Duty and for the Sake of the Noble: Kant and Aristotle on Morally Good Action. In: Stephen Engstrom / Jennifer Whiting (Hrsg.): Aristotle, Kant, and the Stoics: Rethinking Happiness and Duty, Cambridge, 203–236. Korsgaard, Christine M. (1996): Creating the Kingdom of Ends. Cambridge. Korsgaard, Christine M. (1998): Einleitung zu: Immanuel Kant: Groundwork of the Metaphysics of Morals, translated and edited by Mary Gregor. Cambridge. Kraft, Bernd / Schönecker, Dieter (1999): Einleitung. In: Immanuel Kant, Grundlegung zur Metaphysik der Sitten, mit einer Einleitung hrsg. von Bernd Kraft und Dieter Schönecker, Hamburg, VII–XXXIX.

Bibliography

331

Kraft, Bernd / Schönecker, Dieter (2000): Kants Grundlegung zur Metaphysik der Sitten in der Edition der Akademieausgabe. In: Kant-Studien 91 (Sonderheft), 20–22. Krausser, P. (1968): Über eine unvermerkte Doppelrolle des kategorischen Imperativs in Kants Grundlegung zur Metaphysik der Sitten. In: Kant-Studien 59, 318–332. Kupperman, Joel J. (2002): A Messy Derivation of the Categorial Imperative. In: Philosophy 77, 484–502. Laberge, Pierre (1980): Du Passage de la Philosophie Morale Populaire à la Métaphysique des Mœurs. In: Kant-Studien 71, 418–444. Laberge, Pierre (1982): L’Espèce de cercle dont, à ce qu’il semble, il n’y pas moyens de sortir. In: Dialogue 21, 745–753. Liddell, Brendan E. A. (1970): Kant on the Foundation of Morality. A Modern Version of the Grundlegung. Translated with Commentary. Bloomington, London. Liddell, Brendan E. A. (1972): Kant’s ‘Deduction’ in the Grundlegung. In: Lewis White Beck (Hrsg.): Proceedings of the Third International Kant Congress, Dordrecht-Holland, 401–406. Löhrer, Guido (1995): Menschliche Würde. Wissenschaftliche Geltung und metaphorische Grenze der praktischen Philosophie Kants. Freiburg i. Br./ München. Maliandi, Ricardo (1974): Bedeutung und Zweideutigkeit des „als-ob“ in der Naturgesetz-Formel des kategorischen Imperativs. In: Gerhard Funke (Hrsg.): Akten des 4. Internationalen Kant-Kongresses, Berlin/New York, 540–549. Marshall, John (1982): Hypothetical Imperatives. In: American Philosophical Quarterly 19, 105–114. Marshall, John (1985): The Syntheticity of the Categorical Imperative. In: Gerhard Funke / Thomas M. Seebohm (Hrsg.): Proceedings of the Sixth International Kant Congress. Vol. II/2, 185–200. Matson, Wallace I. (1961): Foundations of the Metaphysics of Morals. In: Frank N. Magill (Hrsg.) / Ian P. McGreal (Associate Editor): World Philosophy. Essay-Reviews of 225 Major Works. Volume 3, Englewood Cliffs, N. J., 1131–1141. Mautner, Thomas (1971): Kants Grundlegung. In: The Philosophical Quarterly 22, 255–257. McCarthy, Michael H. (1976): Analytic Method and Analytic Propositions in Kant’s Groundwork. In: Dialogue 15, 565–582. McCarthy, Michael H. (1979): Kant’s Application of the Analytic/Synthetic Distinction to Imperatives. In: Dialogue (Canada) 18, 373–391. McCarthy, Michael H. (1979): Paton’s Suggestion that Kant’s Principle of Autonomy might be Analytic. In: Kant-Studien 70, 206–224. McCarthy, Michael H. (1982): Kant’s Rejection of the Argument of Groundwork III. In: Kant-Studien 73, 169–190. McCarthy, Michael H. (1984): Kant’s Groundwork Justification of Freedom. In: Dialogue (Canada) 23, 457–473. McCarthy, Michael H. (1985): The Objection of Circularity in Groundwork III. In: Kant Studien 76, 28–42.

332

Bibliography

Melches Gibert, Carlos (1994): Der Einfluß von Christian Garves Übersetzung Ciceros De Officiis auf Kants Grundlegung zur Metaphysik der Sitten. Regensburg. Melnick, Arthur (2002): Kant’s Formulations of the Categorical Imperative. In: Kant-Studien 93, 291–308. Miller, R. D. (1993): An Interpretation of Kant’s Moral Philosophy. Harrogate. Milz, Bernhard (1998): Zur Analytizität und Synthetizität der Grundlegung. In: Kant-Studien 89, 188–204. Moritz, Manfred (1972): Die Probleme der Deduktion des kategorischen Imperativs (ein Vorbericht). In: Lewis White Beck (Hrsg.): Proceedings of the Third International Kant Congress (Rochester), Dordrecht-Holland, 424–433. Mulholland, Leslie (1978): Kant: On Willing Maxims to Become Laws of Nature. In: Dialogue 17, 92–105. Oberer, Hariolf (1983): Kants praktische Philosophie. In: Karl Bärthlein (Hrsg.): Zur Geschichte der Philosophie. Bd. 2, Von Kant bis zur Gegenwart, Würzburg, 15–30. O’Neill, Onora (1989): Agency and Anthropology in Kant’s Groundwork. In: Yirmiyahu Yovel (Hrsg.): Kant’s Practical Philosophy Reconsidered. Papers presented at the Seventh Jerusalem Philosophical Encounter, December 1986, Dordrecht/Boston/London, 63–82. O’Neill, Onora (1989): Constructions of Reason: Explorations of Kant’s Practical Philosophy. Cambridge. Pasternack, Lawrence (Hrsg.) (2002): Immanuel Kant: Groundwork of the Metaphysics of Morals. London/New York. Pasternack, Lawrence (2002): Intrinsic Value and Moral Overridingness in Kant’s Groundwork. In: Southwest Philosophy Review: The Journal of The Southwestern Philosophical Society 18, 113–121. Paton, Henry James (1947): The Categorical Imperative. A Study in Kant’s Moral Philosophy, London. Paton, Henry James (1958): The Aim and Structure of Kant’s Grundlegung. In: The Philosophical Quarterly 8, 112–130. Patt, Walter (1997): The Synthetic Character of the Moral Law According to Kant. In: Hariolf Oberer (Hrsg.): Kant. Analysen – Probleme – Kritik. Bd. III, Würzburg, 21–39. Patzig, Günther (1973): Die logischen Formen praktischer Sätze in Kants Ethik. In: Gerold Prauss (Hrsg.): Kant. Zur Deutung seiner Theorie von Erkennen und Handeln, Köln, 207–222. Patzig, Günther (1980): Der kategorische Imperativ in der Ethik-Diskussion der Gegenwart. In: Ders.: Tatsachen, Normen, Sätze. Aufsätze und Vorträge, Stuttgart, 155–177. Patzig, Günther (1986): „Principium diiudicationis“ und „Principium executionis“: Über transzendentalpragmatische Begründungsansätze für Verhaltensnormen. In: Gerold Prauss (Hrsg.): Handlungstheorie und Transzendentalphilosophie, Frankfurt a. M., 204–218. Pelegrinis, Theodosius N. (1980): Kant’s Conceptions of the Categorical Imperative and the Will. London. Petrus, Klaus (1994): „Beschrieene Dunkelheit“ und „Seichtigkeit“. Historisch-systematische Voraussetzungen der Auseinandersetzung zwischen

Bibliography

333

Kant und Garve im Umfeld der Göttinger Rezension. In: Kant-Studien 85, 280–302. Philonenko, Alexis (1972): L’Œuvre de Kant. La philosophie critique. Tome II, Paris. Pistorius, Hermann Andreas (1786): Rezension der Grundlegung zur Metaphysik der Sitten. In: Rüdiger Bittner / Konrad Cramer (Hrsg.): Materialien zu Kants Kritik der praktischen Vernunft, Frankfurt a. M. 1975, 144–160. Pogge, W. (1994): Freudigers Grundlegung. In: Grazer Philosophische Studien 47, 223–239. Potter, Nelson (1974): The Argument of Kant’s Grundlegung, Chapter One. In: Canadian Journal of Philosophy. Supp. 1, 73–91. Potter, Nelson (1993): What is wrong with Kant’s four examples? In: Journal of Philosophical Research 18, 213–229. Prauss, Gerold (1983): Kant über Freiheit als Autonomie. Frankfurt a. M. Quarfood, Marcel (2001): Kant’s Practical Deduction of Moral Obligation in Groundwork III. In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.): Kant und die Berliner Aufklärung, Bd. 3, Berlin/ New York, 72–79. Rawls, John (1989): Themes in Kant’s Moral Philosophy, in: E. Förster (ed.), Kant’s Transcendental Deductions, Stanford, 81–113. Rawls, John (2000): Lectures on the History of Moral Philosophy, ed. by B. Herman, Harvard. Reath, Andrews (1989): The Categorical Imperative and Kant’s Conception of Practical Rationality. In: Monist 72, 384–410. Reath, Andrews (2003): Immanuel Kant, Groundwork of the Metaphysics of Morals (1785): Duty and Autonomy. In: Jorge J. E. Gracia (Hrsg.): The Classics of Western Philosophy: A Reader’s Guide, Malden, 346–356. Reboul, Olivier (1967): Prescription ou Proscription? Essai sur le sens du devoir chez Kant. In: Revue de Metaphysique et Morale 72, 295–320. Reich, Klaus (1935): Kant und die Ethik der Griechen. Tübingen. Reiner, Hans (1949): Eine Textverschiebung in Kants Grundlegung zur Metaphysik der Sitten. In: Zeitschrift für philosophische Forschung 4, 123–124. Reiner, Hans (1949): Stellungnahme zu den vorgelegten Bemerkungen. In: Zeitschrift für philosophische Forschung 4, 579–581. Rickless, Samuel C. (2004) From the Good Will to the Formula of Universal Law. In: Philosophy and Phenomenological Research 68, 554–577. Riedel, Manfred (1989): Kritik der moralisch urteilenden Vernunft. Kants vorkritische Ethik und die Idee einer „Grundlegung zur Metaphysik der Sitten“. In: Ders.: Urteilskraft und Vernunft. Kants ursprüngliche Fragestellung, Frankfurt a. M., 61–97. Rodriguez Lopez, Blanca (2002): Sobre el caracter hipotetico de los imperativos morales. In: Telos: Revista Iberoamerica de Estudios Utilitaristas 11, 99–119. Rollin, Bernard E. (1976): There is Only One Categorical Imperative. In: Kant-Studien 67, 60–72. Ross, Sir William David (1954): Kant’s Ethical Theory. A Commentary on the Grundlegung zur Metaphysik der Sitten. Oxford. Rossvaer, Viggo (1979): Kant’s Moral Philosophy. An Interpretation of the Categorical Imperative. Oslo.

334

Bibliography

Sandermann, Edmund (1989): Die Moral der Vernunft. Transzendentale Handlungs- und Legitimationstheorie in der Philosophie Kants. Freiburg i. Br./München. Sartiaux, F. (1917): Morale Kantienne et Morale humaine. Paris. Schaller, Walter E. (1995): From the Groundwork to the Metaphysics of Morals: What happened to Morality in Kant’s Theory of Justice? In: History of Philosophy Quarterly 12, 333–345. Schneewind, J. B. (1992): Autonomy, Obligation, and Virtue: An Overview of Kant’s Moral Philosophy. In: Paul Guyer (Hrsg.): The Cambridge Companion to Kant, Cambridge, 309–341. Schnoor, Christian (1989): Kants Kategorischer Imperativ als Kriterium der Richtigkeit des Handelns. Tübingen. Schöndorf, Harald (1985): „Denken-Können“ und „Wollen-Können“ in Kants Beispielen für den kategorischen Imperativ. In: Zeitschrift für philosophische Forschung 39, 549–573. Schönecker, Dieter (1996): Zur Analytizität der Grundlegung. In: Kant-Studien 87, 348–354. Schönecker, Dieter (1997): Die Methode der Grundlegung und der Übergang von der gemeinen sittlichen zur philosophischen Vernunfterkenntnis. In: Hariolf Oberer (Hrsg.): Kant. Analysen – Probleme – Kritik. Bd. III, Würzburg, 81–98. Schönecker, Dieter (1997): Gemeine sittliche und philosophische Vernunfterkenntnis. Zum ersten Übergang in Kants Grundlegung. In: Kant-Studien 88, 311–333. Schönecker, Dieter (1997): Die „Art von Zirkel“ im dritten Abschnitt der Grundlegung. In: Allgemeine Zeitschrift für Philosophie 22, 189–202. Schönecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs. Freiburg i. Br./München. Schönecker, Dieter (2001): What is the ‘First Proposition’ Regarding Duty in Kant’s Grundlegung? In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.): Kant und die Berliner Aufklärung, Bd. 3, 89–95. Schönecker, Dieter (2001): Kants Grundlegung zur Metaphysik der Sitten lesen. In: Information Philosophie, Heft 5, 50–56. Schönecker, Dieter / Wood, Allen W. (22004): Immanuel Kant: Grundlegung zur Metaphysik der Sitten. Ein einführender Kommentar. Paderborn. Scott, J. W. (1924): Kant on the Moral Life. An Exposition of Kant’s Grundlegung. London. Sherline, Edward (1995): Heteronomy and Spurious Principles of Morality in Kant’s Groundwork. In: Pacific Philosophical Quarterly 76, 32–46. Singer, Marcus G. (1975): Verallgemeinerung in der Ethik. Zur Logik moralischen Argumentierens. Frankfurt a. M. Singer, Marcus G. (1983): Reconstructing the Groundwork. In: Ethics 93, 566–578. Skorpen, Erling (1981): Making Sense of Kant’s Third Example. In: KantStudien 72, 415–429. Staege, Roswitha (2002): Hypothetische Imperative. In: Kant-Studien 93, 42–56.

Bibliography

335

Stattler, Benedikt (1788): Anhang zum Anti-Kant in einer Widerlegung der Kantischen Grundlegung zur Metaphysik der Sitten. München, wiederabgedruckt in: Aetas Kantiana. Brüssel. Steigleder, Klaus (2001): Hypothetische Imperative als reflexive Urteile. In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.), Kant und die Berliner Aufklärung, Bd. 3, Berlin/New York, 113–121. Steigleder, Klaus (2002): Kants Moralphilosophie. Die Selbstbezüglichkeit reiner praktischer Vernunft. Stuttgart/Weimar. Steinberger, Peter J. (1999): The Standard View of the Categorical Imperative. In: Kant-Studien 90, 91–99. Storheim, E. (1979): Kants und Fichtes Begründung der Moral. In: Klaus Hammacher / Albert Mues (Hrsg.): Erneuerung der Transzendentalphilosophie im Anschluß an Kant und Fichte. Reinhard Lauth zum 60. Geburtstag, Stuttgart-Bad Cannstatt, 411–424. Stratton-Lake, Philip (1993): Formulating Categorical Imperatives. In: KantStudien 84, 317–340. Stratton-Lake, Philip (2000): Kant, Duty and Moral Worth. London/New York. Sullivan, Roger J. (1989): Immanuel Kant’s Moral Theory. Cambridge. Sullivan, Roger J. (1994): An Introduction to Kant’s Ethics. Cambridge. Swoyer, Chris (1983): Kantian Derivations. In: Canadian Journal of Philosophy 13, 409–431. Teale, A. E. (1951): Kantian Ethics. London. Tenbruck, Friedrich (1949): Bemerkungen zu H. Reiners Textkorrektur. In: Zeitschrift für philosophische Forschung 4, 572–574. Terzis, George N. (1989): Autonomy without Indeterminism: A Reconstruction of Kant’s Groundwork, Chapter Three. In: The Modern Schoolman 67, 1–13. Timmermann, Jens (2000): Kant und die Lüge aus Pflicht: Zur Auflösung moralischer Dilemmata in einer kantischen Ethik. In: Philosophisches Jahrbuch 107, 267–283. Timmermann, Jens (2001): The Dutiful Lie: Kantian Approaches to Moral Dilemmas. In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.): Kant und die Berliner Aufklärung, Bd. 3, 345–354. Timmons, Mark (1984): Contradictions and the Categorical Imperative. In: Archiv für Geschichte der Philosophie 66, 294–312. Timmons, Mark (1989): McCarthy on Practical Necessitation in Kant. In: Kant-Studien 80, 198–207. Timmons, Mark (1992): Necessitation and Justification in Kant’s Ethics. In: Canadian Journal of Philosophy 22, 223–261. Tugendhat, Ernst (1993): Vorlesungen über Ethik. Frankfurt a. M. Vialatoux, J. (1956): La morale de Kant. Paris. Wagner, Hans (1993): Kants schwierige Bestimmung des Verhältnisses zwischen Freiheit und Moralgesetz. (Zum 3. Abschnitt der Grundlegung zur Metaphysik der Sitten. A.-A. IV 446–463). In: Reinhold Breil / Stephan Nachtsheim (Hrsg.): Vernunft und Anschauung. Philosophie – Literatur – Kunst. Festschrift für Gerd Wolandt zum 65. Geburtstag, Bonn, 3–15.

336

Bibliography

Wagner, Hans (1994): Kants Konzept von hypothetischen Imperativen. In: Kant-Studien 85, 78–84. Walsh, Dorothy (1971): The Dilemma of the Grundlegung. In: The Philosophical Quarterly 21, 149–157. Ward, Keith (1972): The Development of Kant’s View of Ethics. Oxford. Wartenberg, Thomas E. (1992): Reason and the Practice of Science. In: Paul Guyer (Hrsg.): The Cambridge Companion to Kant, Cambridge, 228– 248. Wenzel, Uwe Justus (1992): Anthroponomie. Kants Archäologie der Autonomie. Berlin. Westphal, Kenneth R. (1995): How ‘Full’ is Kant’s Categorical Imperative? Jahrbuch für Recht und Ethik 3, 465–509. Wike, Victoria S. (1987): The Role of Happiness in Kant’s Groundwork. In: The Journal of Value Inquiry 21, 73–78. Wilde, Leo Henri (1975): Hypothetische und kategorische Imperative. Eine Interpretation zu Kants Grundlegung zur Metaphysik der Sitten. Bonn. Willaschek, Marcus (1992): Praktische Vernunft: Handlungstheorie und Moralbegründung bei Kant. Stuttgart/Weimar. Williams, T. C. (1968): The Concept of the Categorical Imperative. A Study of the Place of the Categorical Imperative in Kant’s Ethical Theory. Oxford. Wimmer, Reiner (1980): Universalisierung in der Ethik. Analyse, Kritik und Rekonstruktion ethischer Rationalitätsansprüche. Frankfurt a. M. Wolff, E. (1966): Les trois Imperatifs Categoriques et les trois Postulats de la Raison Pratique chez Kant. In: Archives de philosophie 29, 37–58. Wolff, Robert P. (Hrsg.) (1969): Foundations of the Metaphysics of Morals. Translated by Lewis White Beck, with Critical Essays Edited by Robert Paul Wolff. New York. Wolff, Robert P. (1973): The Autonomy of Reason. A Commentary on Kant’s Groundwork of the Metaphysic of Morals. New York. Wood, Allen W. (1972): Kant on False Promises. In: Lewis White Beck (Hrsg.): Proceedings of the Third International Kant Congress, Dordrecht-Holland, 614–619. Wood, Allen W. (1976): Kant on the Rationality of Morals. In: P. La Berge / F. Duchesneau / B. E. Morrisey (Hrsg): Proceedings of the Ottawa Congress on Kant in the Anglo-American and Continental Traditions, Ottawa, 94–109. Wood, Allen W. (1999): Kant’s Ethical Thought. Cambridge. Wood, Allen W. (2002): Immanuel Kant, Groundwork for the Metaphysics of Morals. Edited and translated by Allen W. Wood. With Essays by J. B. Schneewind, Marcia Baron, Schelly Kagan, Allen W. Wood, New Haven.

General Index a priori, apriority 3, 5, 7, 11–16, 61, 142, 154 f., 158, 165, 171 ff., 186, 194, 196, 228–233, 238 ff., 243, 249 f., 293 f., 296 f., 304, 306 f., 316, 322 Achenwall, Gottfried 151 f., 157 Acton, Harry 325 agency 58, 60 f., 63, 65, 67, 78 ff., 88, 111, 165, 172, 195 Albrecht, Michael 325 Allison, Henry 74, 79 f., 83, 91, 94, 103, 106–111, 112 f., 117, 134, 138, 248, 271, 283, 286, 291, 293, 296, 300, 325 Alquié, Ferdinand 325 Ameriks, Karl 214, 221, 325 analytic/synthetic 17 f., 145 f., 152 f. analyticity, thesis of 241 f., 301, 303– 307 Anaxagoras 54 f. Annas, Julia 72, 91 anthropology 6, 15, 67, 250 Apel, Karl-Otto 115, 319 appearance 286, 297, 316 f., 321 f. argument by elimination 201 ff., 211 ff. Aristotelian biology 57 Aristotelian teleology 48 Aristotle 52, 127, 330 Atwell, John E. 325 Aubenque, Pierre 325 Aune, Bruce 96, 103–106, 108–111, 113, 117, 197, 325 automaton spirituale 275 f., 279 autonomy 61,129, 131, 178 f., 190 f., 229, 232 f., 236, 243, 245, 248, 250, 253 f., 257, 259–265, 268–272, 275, 278, 280 ff., 320 f. Bagattini, Alexander 121 Baker, Judith 81, 91, 325 Barnes, Gerald W. 325 Baron, Marcia 78–80, 83 f., 86, 88, 90 f., 160, 197, 325, 336 Bärthlein, Karl 332 Baumanns, Peter 325 f.

Baumgarten, Hans-Ulrich 126, 326 Bayles, Richard E. 167, 197 Beck, Lewis White 111, 126, 138, 195, 197, 261, 270 f., 277, 283, 326, 331 f., 336 Bennett, Jonathan 277, 283 Benton, Robert J. 326 Bittner, Rüdiger 15, 18, 21, 326, 333 Blum, Lawrence 72, 91 Bollnow, Otto-Friedrich 326 Brandt, Reinhard 18, 21, 296 f., 300, 326 Breil, Reinhold 335 Brink, David O. 167, 197 Brinkmann, Walter 136, 138, 326 Bruton, Samuel 326 Bubner, Rüdiger 326 Burri, Alex 326 Canivet, Michel 326 Capobianco, Richard 301 Carnois, Bernard 326 categorical imperative 7 f., 10 f., 18, 41 ff., 63, 65, 73, 86, 94–100, 104, 107– 116, 125, 130 f., 137, 141 ff., 150, 154, 158–196, 200–221, 228 f., 231, 236, 241–243, 247 f., 262, 293 ff., 301–323 categorical imperative, deduction of 108, 111, 113, 241 ff., 291, 294 f., 301– 323 categorical imperative, formula of autonomy 32, 37, 178 f., 190 f., 193 categorical imperative, formula of humanity 31, 169, 177–179, 189–191, 200–221, 261 f. categorical imperative, formula of universal law 36, 65, 93, 113, 158, 185, 193, 218 f., 262 philosophical significance of 187–191 practical significance of 192 f. as a decision procedure 161, 167–177 core assumptions of 168 f. formal constraint interpretation of 185–191

338

General Index

categorical imperative, formula of universal law of nature (FLN) 158– 185, 193 formal constraint interpretation of FLN 160 strong model of 167, 169–171 categorical imperative, possibility of 229 f., 240, 301–323 categorical imperatives (pl.) 97 f., 116 causality 58, 63 f., 124, 145, 147, 233 ff., 239, 244, 248, 251 ff., 257 ff., 268, 275, 277 f., 308 ff. character 27, 35, 39 f., 88, 214, 270 Chrysippus 53 Cicero 331 Cicovacki, Predrag 326 circle 285–299, 304 f. circulus in probando 290 f., 305 Collins 61, 141 conceptual necessitiy 10 f. consistency test 164 f. contradiction-in-conception 164, 173 f., 182, 194–198 contradiction-in-will 164, 173 f., 182 f., 196 Copp, David 326 Cotter, Alexander 69, 225, 301 Craemer-Ruegenberg, Ingrid 326 Cramer, Konrad 327, 333 Critique of practical reason 19 f., 60, 228, 231, 310 Croitori, Rodica 327 Damschen, Gregor 301, 323 Darwinism 57 De Bolt, Darian C. 327 deduction (see ‘categorical imperative, deduction of’) Delbros, Victor 327 Denis, Lara 83, 91, 327 derivation 37, 93–116, 126 f., 172 f., 200 ff., 209 ff., 217 f., 288 Descartes, René 326 determination of judgement 251, 253, 263, 279, 310 determination, natural 111, 234 f., 256, 275, 278, 282 (see also ‘pre-determination’) determination, of the will 142, 150 f., 237, 239, 244 f., 269 dignity 40, 46, 189–191, 213 Donagan, Alan 185, 197 Downie, R. S. 150, 157 Duchesneau, F. 326

Duncan, Alistair R. Campbell 327 duties of self-perfection 172, 181, 183 duty 6 ff., 11 ff., 34, 36 f., 47, 72 f., 82, 87, 97, 101, 114 f., 133, 137, 158, 160 ff., 170–174, 177, 179–183, 187 f.,192, 195, 216, 219, 228, 294 f., 299, 304 f., 318 duty, acting from 32–40, 72–90, 100– 103, 105, 113, 132 f., 213 f. Dye, James W. 327 Ebbinghaus, Julius 327 Eisler, Rudolf 283 end, necessary 227–229, 236, 240 f. end, objective 56, 200, 202 f. end, subjective 202 f. end-in-itself 31, 40, 64 f., 169, 178, 180, 189 f., 204 f., 210 ff., 229, 261 f., 320, 323 Engelhardt, Paulus 327 Engstrom, Stephen 327, 330 Epicureanism 52, 66–68 Epicurus 53 Ertl, Wolfgang 327 ethical principles 7–10, 17, 20 f. Eudaemonism 49, 53, 66 Ewing, Alfred 196 f. faculty of desire 74, 124–126, 128, 146 f., 233 f., 244, 271 Ferrari, Jean 53, 70 Fichte, Johann Gottlieb 335 Fleischer, Margot 327 Forschner, Maximilian 327 Förster, Eckart 327 Fouillée, Alfred 328 Frank, Manfred 58, 70 Frankfurt, Harry 259 free choice 233 ff., 237, 244 f. (see also ‘freie Willkür’) freedom 51, 61, 64 f., 67, 84, 232, 235 f., 239 f., 243 f., 247–282, 285, 287–288, 292, 297, 300, 304 f., 310 freedom of the will 60, 111 f., 129, 233, 236 ff., 242, 245, 247–282, 287 f., 291 f., 302 ff., 306, 309 ff. freedom, negative 235, 248, 253 f., 256– 259, 264 f., 267, 269, 271–276, 280 ff. freedom, positive 235 f., 240, 248, 253 f., 258 f., 269, 271 f., 275, 282 freedom, transcendental 111 f., 232, 258, 260, 265, 275 f. Freudiger, Jürg 21, 326 Funke, Gerhard 326, 331 Galay, Jean-Louis 328 Gallop, David 55, 70

General Index Galvin, Richard 165, 172, 195, 197 gap theorists 94, 103 ff. Garve, Christian 331, 332 Gass, Michael 328 Gaut, Berys 204, 221 Geismann, Georg 327 f. Genova, A. C. 328 Gerhardt, Volker 124, 138, 326, 328, 333–335 gifts of fortune 26, 28 gifts of nature 26–29, 214 Glasgow, Joshua 183, 194, 196 f. Glass, Ronald 328 Glockner, Hermann 326 Glockner, Marie 326 God 10, 32, 122, 124, 131–133, 213, 279 God’s existence 45, 66 good will 25–43, 45–53, 66, 72 f., 77 f., 82 f., 89 f., 101, 104 ff., 113, 129–133, 202, 212–220, 238 f., 244, 278, 294, 306 f., 322 good, absolute 28–30, 38, 133, 204 , 212, 220, 238 f., 307 (see also good will) good, highest (summum bonum) 47, 61, 63, 66 good-natured temperament 81 f. goodness, objective 52, 201, 205 ff., 210 ff. goodness, preeminent 101, 212–218 goodness,unconditional 201, 204, 206, 210 ff., 215 Gracia, Jorge J. E. 333 Green, Ronald M. 328 Gregor, Mary 248, 270 f. Grünewald, Bernward 328 Gunkel, Andreas 328 Gupta, R. K. 328 Guyer, Paul 70, 83, 91, 328 Haardt, Alexander 328 Habermas, Jürgen 319 Hammacher, Klaus 335 happiness 20, 27, 29 f., 36, 41, 43, 46–54, 60 f., 66–69, 83,116, 136 f.,144, 153, 167, 172, 204, 226 f., 278, 308 Harbison, Warren 328 Hare, Richard 143, 157, 186, 197 Harris, N. G. E. 328 Harrison, Jonathan 172, 195, 197, 329 hedonism 53, 67, 69 Hegel, Georg F. W. 108, 158 f., 197 Heidemann, Ingeborg 329 Heimsoeth, Heinz 329 Heintel, Erich 326

339

Held, Carsten 326 Helvétius 51–53 Henrich, Dieter 259, 283, 329 Henson, Richard G. 81, 91 Herman, Barbara 60, 70, 75, 81, 91, 160, 169, 173 f., 195, 197 f., 329 heteronomy 129, 253, 257, 259, 264, 269–272, 278, 308, 323 Hill, Thomas E. jr. 84, 86, 91, 201, 207, 221 Himmelmann, Beatrix 53, 66, 70 history 58–63, 65 Hochberg, Gary M. 329 Hoerster, Norbert 329 Höffe, Otfried 159, 195, 198, 329 Högemann, Brigitte 329 holism (ethical) 115 f. holy will 37–42, 121 f., 129–134, 304, 307 Horn, Christoph 116, 196, 329 Horstmann, Rolf-Peter 326, 328, 333– 335 humanity 31, 80, 178 f., 200 ff., 261 f. humanity, as biological species 49 Hume, David 58, 124 Hunter, Ian 329 Hutchins, Patrick 329 hypothetical/categorical 97 ff.,122 f., 134 ff.,141 ff., 150 ff., 207, 230 f. hypothetical imperative 41, 43, 61, 98 f., 116, 122 f.,135 f., 136, 139–156, 230 f., 243, 264 ff. counsels of prudence 123, 136 f., 141, 144, 153 rules of skill 123, 136, 141, 153 Ilting, Karl-Heinz 330 imperative 41 f., 121–123, 125, 129–131, 134–137, 139–146, 153, 155, 207 f., 226 f. imperatives, as judgements 226, 231 imperatives, structure of 136, 242 f. imperfect duties 85–87, 175 incentive 33, 38, 60, 80 f., 83, 127, 130, 214, 294, 321 inclination 30, 33 ff., 48, 66 ff., 73–85, 89, 122 f., 129 ff., 140, 149 f., 162, 189 f., 200, 204, 207, 212 ff., 219, 227, 229, 234, 247, 253, 255 f., 259, 264 f., 268 f., 271 f., 317 indeterminism 247, 252, 257, 259, 268– 272 instinct 46, 49, 51–53, 74, 234 intelligence 306, 309 ff., 314

340

General Index

interests 41, 47, 79, 122, 234, 240, 319 f. Irrlitz, Gerd 330 Irwin, Terence H. 68, 70 Jacobsen, Mogens Chrom 330 Johnson, Darrell 330 Johnson, Monte Ransome 48, 70 Jowett, B. 54 f. judgement 5, 7 ff., 11 f., 16 ff., 20, 40, 65, 115, 142 ff., 152–156, 168, 172, 174, 181, 190, 251 f., 259 ff., 263, 265 ff., 273 f., 278 f., 285, 292, 297, 310 judgement, synthetic a priori 230 ff., 239 f., 243 judgement, analytic 153, 156, 230, 238 f. judgement, practical 226–233, 252 ff., 266, 268 f., 273 judgement, table of 142, 152, 155 Kagan, Schelly 336 Kaulbach, Friedrich 330 Kemp, J. 330 Kerstein, Samuel J. 75, 91, 113 f., 160, 176, 198, 203, 206, 219, 221, 330 Kersting, Wolfgang 330 Kim, Halla 288, 290 f., 296, 300 King, J. Charles 330 kingdom of ends 262 Kirchmann, Julius Hermann v. 330 Klamp, Gerhard 330 Klein, Sherwin 330 Klotz, Christian 5, 21 Köhl, Harald 8, 21, 73, 98, 102, 112, 115–117, 330 Konhardt, Klaus 330 Korsgaard, Christine 70, 90 f., 160, 174, 196, 198, 201, 206, 210, 218, 221, 261, 283, 330 Kraft, Bernd 325, 330 Krausser, P. 331 Kripke, Saul 10, 12, 21 Kulenkampff, Jens 327 Kuppermann, Joel J. 331 Laberge, Pierre 123, 138, 331, 336 LaMettrie 51, 52, 53 Langthaler, Rudolf 70 Larmore, Charles 115, 117 Latham, Noa 79, 92 lawfulness 78, 89, 95 f., 98 f., 102 f., 106– 108, 110, 294 laws of freedom 4 f., 10, 19 f., 267 laws of nature 4, 10, 65, 124, 174, 182 ff., 195, 248, 255 ff., 261, 269, 273 f., 276– 281, 322

laws of reason 33, 125, 127, 131 f., 253– 258, 260, 265, 267–270, 272 f., 278, 280 ff. legality 8 Leibniz, Gottfried Wilhelm 275 Leonard (character of “Memento”) 173 Lidell, Brendan E. A. 331 life 124, 234 living being 124 ff., 234, 242, 244 logical structure of ethical principles 3 f., 7–11 Löhrer, Guido 331 Looney, Aaron 3 Ludwig, Bernd 156 f. Magil, Frank N. 331 Maliandi, Ricardo 331 Mandeville, Bernard 51–53 Marc-Wogau, Konrad 58, 70 Marshall, John 331 Matson, Wallace I. 331 Mautner, Thomas 331 maxim 8, 33 f., 39, 42, 65, 86–88, 96 ff., 131–134, 137, 163 f., 226, 231, 237 McCarthy, Michael H. 331, 335 McFarland, J. D. 70 McGee, Vann 72 McGreal, Ian P. 331 McLaughlin, Peter 70 McNair, Ted 198 means-end-formula 145–153, 264 mechanism 55, 275 Meiklejohn, J. M. D. 58 Melches, Gilbert Carlos 331 Melnick, Arthur 332 Mendelssohn, Moses 54 metaphysics of morals 4 f., 16 f., 19–21 metaphysics of nature 5 method 17 f. Mieth, Corinna 196, 247 Mill, John S. 158 f., 198 Miller, R. D. 332 Milz, Bernhard 18, 21, 332 misology (contempt of rationality) 54 modal status 5, 9 f., 13 moral criteria 167 f., 178 f., 190 moral law 31, 34, 37 ff., 41, 63 f., 66, 93– 116, 125, 130–135, 188, 211, 213, 219, 235, 237, 239, 242, 245, 247 f., 250, 254 f., 259, 263, 271, 276, 282, 286–299, 301–323 moral law, respect of 33–36, 127 (see also ‘duty, acting from’) moral laws (pl.) 6, 12 ff., 165

General Index moral motive (acting for the sake of the law) 38 f., 100–103, 214 (see also ‘duty, acting from’) moral principle 83, 106, 114 ff., 161, 168 f., 172, 180, 185, 190, 210, 217, 225 ff. (see also ‘supreme principle of morality’; ‘categorical imperative’) moral principle, justification of 225– 243, 301–323 moral principles 4, 10 f., 13, 15, 17, 43, 115, 122, 130, 167, 177 f. moral question 307, 319 moral worth 35 f., 72–73, 77–78, 80–81, 87–88, 100, 113 morality 3, 6–11, 16 f., 31, 47, 54, 60, 63– 69, 83 f., 102, 112, 131 f., 158, 188, 219, 225–243, 248 ff., 268, 285 f., 288 f., 291– 300, 305, 308–311 (see also ‘moral motive’) Moritz, Manfred 332 Moritz, Michael 142, 157 Morrisey, B. E. 336 motivation 35, 67, 79, 82, 123, 134, 173, 229, 234, 247, 256, 291, 306, 319–321 Mrongovius, Christoph C. 141 Mues, Albert 335 Mulholland, Leslie 198, 332 Nachtsheim, Stephan 335 Nagl-Docekal, Herta 325 Narveson, Jan 161, 198 naturalism 53, 62, 274, 276 naturalists 53 nature 5, 26 f., 28 f., 45, 47 ff., 51 f., 54 f., 57–66, 69 f., 81, 124, 158, 194, 204, 212, 214, 232, 236, 240, 248, 261, 273 f., 275, 277–280, 308, 322 nature, gifts of 20, 27 f., 214 nature, human 14, 45 f., 53, 59, 66, 68, 171 f., 178 f., 190, 211, 318 necessitation 34, 123, 134 f., 139 ff., 144 f., 148–153, 226, 231, 243 necessity 3–15, 34, 36 f., 61, 97 f., 101, 107, 114 f., 133, 135, 141, 145, 148, 152, 162, 171, 188, 226 ff., 230 f., 236 f., 242 f., 255, 269 necessity, natural 61, 65, 68, 137, 233– 237, 253, 258, 275 f. Nida-Rümelin, Julian 20, 21 Nolan, Christopher 173 nondiminishability thesis 28–30, 42 nonincreasability thesis 28–30, 42 normativity 11, 13, 134, 187 norms 225 ff.

341

O’Neill, Onora 164, 166, 172, 174, 180, 196, 198, 263, 283 Oakley, Justin 72, 92 Oberer, Hariolf 326–328, 332, 334 obligation 13 f., 165, 171 f., 181, 187 f., 203, 225, 268 f., 293–299, 303–309, 319 ontoethical principle 311–318 organism 55 f., 58 f., 63, 65, 124 f. ought 123, 130, 134–137, 139 ff., 148 ff., 208, 213, 225–229, 231, 265, 271,305 f., 311, 316, 321 ought, conditional 226–229, 235 ought, unconditional 227–231, 233, 235, 238 f. Pasternack, Lawrence 327, 332 Paton, Henry James 94, 117, 153, 157, 195, 198, 331 f. Patt, Walter 332 Patzig, Günther 115, 117, 143 f., 150, 157, 332 Pelegrinis, Theodosius N. 332 perfectionism 66 f., 161 permissible actions 85–87, 184 petitio principii 289 f., 292, 295, 305 Petrus, Klaus 332 Philonenko, Alexis 333 Pistorius, Hermann Andreas 333 Plato 54–57 pleasure 49, 51 ff., 60, 67, 86, 128, 178, 226 f., 240 Pogge, Thomas 164, 169, 173, 195, 198 postulate 68, 211, 240, 320 Potter, Nelson 168, 172, 196, 198, 333 practical deliberation 61, 252 ff., 260 ff., 265 ff., 278 ff., 282 practical principles 32, 42 f., 104, 123, 208 practical syllogism 127 Prauss, Gerold 260, 283, 326, 329 f., 332 f. pre-determination 257, 273–276, 279, 281 prescriptive 4, 11, 13, 98 ff., 135, 299 principium diiudicationis/executionis bonitatis 127 Quarfood, Marcel 285, 305, 323, 333 Quine, W. V. O. 115, 117 radical evil 133 rational being 6, 10, 14 ff., 20, 32, 42, 61, 63 f., 68, 80, 95, 104, 111, 113 f., 121 f., 124 ff., 134 f., 146, 149, 154, 170, 188 f.

342

General Index

191, 201 f., 204, 206 ff., 213, 229, 232 f., 240 ff., 247–255, 260–264, 267, 278– 281, 285, 290, 292 ff., 302–307, 315, 318 rational nature 31, 40, 74, 183, 200 ff., 205 f., 210, 212, 217 f., 240, 262, 299 rationality 54, 67, 105, 110, 122, 127, 129, 145, 148 f., 150, 156, 250, 257, 259 ff., 264–268, 272, 276, 278, 280 f., 293, 295 Rawls, John 183, 195, 198 reason, law of 41, 125, 315 reason, practical 32, 39, 43, 46 ff., 61, 67, 121 ff., 126–128, 149, 186, 233 f., 244, 249, 251–254, 260 f., 263 f., 271, 275 f., 278, 293, 304, 307, 309 f., 318 ff., 323 f. reason, pure practical 16 , 111, 122, 155, 227 f., 231 ff., 235 ff., 249, 294, 307, 320 law of 111, 236 f. reason, theoretical 232, 266 f., 293, 309 f. Reath, Andrews 75, 92, 333 Reboul, Olivier 333 regressive argument 201 f., 206, 210, 212, 217 regulative principle 63, 65 Reich, Klaus 53, 70, 333 Reiner, Hans 82, 92, 333, 335 relevant descriptions 159 f., 163, 171, 177–184 representations 121, 124 ff., 135, 147, 208, 233 f., 258, 264, 275 f., 287, 292, 297 f. respect 34 ff., 75, 86, 93, 101, 104 f., 115, 123, 126 f., 133, 140, 144, 148, 150, 159, 165, 169, 182, 189, 249 ff., 261 ff., 278, 285 f., 290, 292 f., 295, 297, 300, 311, 318 ff. Rickless, Samuel 333 Riedel, Manfred 329 f., 333 Ritzel, Wolfgang 326 Robinson, Hoke 327, 330 Rodriguez Lopez, Blanca 333 Rollin, Bernard E. 333 Rosenthal, Jacob 247 Ross, Sir William D. 27, 44, 163 f., 176, 180, 198, 333 Rossvaer, Viggo 333 Rousseau, Jean-Jacques 53, 67, 259 Sandermann, Edmund 333 Sartiaux, F. 333 Scanlon, Thomas M. 187, 198 Scarano, Nico 10, 21

Schadow, Steffi 121 Schaller, Walter 333 Schiller, Friedrich 82, 92 Schneewind, Jerome B. 114–117, 259, 283, 334, 336 Schnoor, Christian 334 Schöndorf, Harald 334 Schönecker, Dieter 15, 18, 21 f., 45, 49 f., 69, 71, 90, 94, 116 f., 121, 149, 155, 157, 196, 241 f., 245–247, 263 f., 273, 283, 286, 288, 290, 292 f., 296, 300 f., 305, 307, 309, 318–320, 323, 325, 330, 334 Schulz, Johann Heinrich 279, 309 Schumacher, Ralph 326, 328, 333–335 Schwaiger, Clemens 151, 157 Schwan, Alexander 329 Scott, J. W. 334 Seebohm, Thomas M. 331 Seel, Gerhard 153, 326, 328 self, authentical 311 self-activity see spontaneity self-determination 229, 258, 264, 281 self-interest 36, 73–78, 84, 89, 123, 171, 319 self-preservation 52 Shell, Susan Meld 67, 71 Sherline, Edward 334 Sherman, Nancy 83, 92 Sidgwick, Henry 72, 92, 167, 186, 198 Siep, Ludwig 115, 117 Singer, Marcus G. 180, 186, 196, 198, 334 Skorpen, Erling 334 Smart, J. J. C. 181, 198 Socrates 54 f. Spitzley, Thomas 327 spontaneity 253, 256–259, 265, 268, 272, 274–278, 280 ff., 298–299, 309 Staege, Roswitha 334 standpoint, problem of 293, 261, 286, 296–300, 318 Stattler, Benedikt 334 Steigleder, Klaus 227, 230, 238, 246, 260, 271, 283, 301, 318 f., 324, 334 Steinberger, Peter J. 335 Stocker, Michael 72, 92 Stoicism 52, 54, 66, 68 Stoics 52 f. Storheim, E. 335 Stratton-Lake, Philip 335 Sturma, Dieter 259, 283, 325, 329 subject, logical 322 Sullivan, Roger J. 335 superiority of the ontic status 316 f.

343

General Index supreme law 97–99, 115, 133, 210 supreme principle of morality 11, 16, 18, 21, 31, 37, 41, 43, 73, 89, 93, 100, 113 f., 172 f., 187 f., 200 ff., 204, 211 f., 225, 229, 301, 308 Swoyer, Chris 172 f., 195, 198, 335 sympathy 34 f., 80–82, 181 talents of the mind 26 f., 90 Teale, A. E. 335 teleological argument 45, 51, 278 teleology 45–69, 195 Tenbruck, Friedrich 335 Terzis, George N. 335 theology 56, 57 thing-in-itself 28, 256, 277, 297–311, 316 f., 321–323 things-in-themselves 297 though experiments 10, 13 f. Thucydides 59 Timmermann, Jens 15, 22, 123, 138, 269, 283, 335 Timmons, Mark 73, 166, 172, 177 f., 195, 198, 335 transcendental idealism 62–64, 286 f., 293, 296–299 Tropman, Elizabeth 90 Tugendhat, Ernst 5, 22, 94, 117, 335 unconditionality thesis 28 f., 37, 42 universality 37, 95, 97 f., 100, 105, 107– 110, 133, 158 ff., 187, 189, 191, 193, 196, 236 f. universalizability 97, 158–199 Vialatoux, J. 335 virtue 39 f., 68, 116, 174 Vliet, JoAnn van 225 Wagner, Hans 335 Walsh, Dorothy 335 Ward, Keith 335 Wartenberg, Thomas 335 Weber, Raoul Simon 325 Weidemann, Hermann 68, 71 Weischedel, Wilhelm 329 welfare 49, 52, 67, 183, 215 Wenzel, Uwe Justus 336 Westphal, Kenneth R. 336 Whiting, Jennifer 330

Wike, Victoria S. 336 Wilde, Leo Henri 336 will 32, 35, 38 f., 41 ff., 47, 50, 65, 75, 78, 80, 89 f., 95, 101, 103, 105 f., 121–137, 140 ff., 145 f., 148 ff., 153 f., 158, 178, 188, 202, 204, 208 ff., 228 f., 232 f., 235 ff., 241, 244, 247–284, 293, 304, 306, 308, 319 (see also ‘good will’) will as practical reason, problem of 32, 39, 121–123, 126 ff., 149, 235, 237, 244 f., 252, 263, 271, 293, 310, 320 will, autonomy of 229, 232 f., 248, 250, 259, 285, 291 f., 302, 304, 309, 311 f., 314 ff., will, intelligible 242, 299, 304, 316, 320 will, pure 17, 130, 304 f., 308, 311, 313, 316–321 will, rational 31 f., 40, 149, 247, 250 f., 260 f., 263, 316 Willaschek, Marcus 124, 128, 133, 138, 336 Williams, Bernard 181, 198 Williams, T. C. 336 Willkür 146, 244, 270 ff., 319 (see also ‘free choice’) Willkür, freie 128, 233, 244 f., 270 ff. Wimmer, Reiner 259, 271, 283, 336 wisdom 4, 40, 66 wish 27, 67, 122, 128, 136, 145–148, 150, 156, 231 Wolff, Christian 17, 297 Wolff, E. 336 Wolff, Robert Paul 336 Wood, Allen 8, 15, 45, 49 f., 77 f., 82, 90, 92, 94, 103, 106, 107–111, 113, 117, 132, 138, 149, 159, 164, 173 f., 176, 178, 198, 201, 205 f., 210, 212, 218, 221, 289, 294, 302, 311, 320, 324, 334, 336 world of sense (sensible world) 256, 280, 286, 298 f., 304–309, 312–322 world of understanding (intelligible world) 256, 280, 286, 298–300, 304 ff., 309–322 worth, absolute 136, 189 f., 202 ff., 228 ff., 236 f., 320 Yovel, Yirmiyahu

332

Zanetti, Véronique 58, 70 Zweig, Arnulf 329