ethics argument paper 1

Rethinking the

Western Tradition

The volumes in this series

seek to address the present debate

over the Western tradition

by reprinting key works of

that tradition along with essays

that evaluate each text from

di√erent perspectives.

EDITORIAL

COMMITTEE FOR

Rethinking

the

Western

Tradition

David Bromwich Yale University

Gerald Graff University of Illinois at Chicago

Geoffrey Hartman Yale University

Samuel Lipman (deceased)

The New Criterion

Gary Saul Morson Northwestern University

Jaroslav Pelikan Yale University

Marjorie Perloff Stanford University

Richard Rorty Stanford University

Alan Ryan New College, Oxford

Ian Shapiro Yale University

Frank M. Turner Yale University

Allen W. Wood Stanford University

Groundwork for the

Metaphysics of Morals

IMMANUEL KANT

Edited and translated by Allen W. Wood

with essays by

J. B. Schneewind

Marcia Baron

Shelly Kagan

Allen W. Wood

Yale University Press

New Haven and London

Published with assistance from the Ernst Cassirer Publications Fund.

Copyright ∫ 2002 by Yale University.

All rights reserved.

This book may not be reproduced, in whole or in part,

including illustrations, in any form (beyond that

copying permitted by Sections 107 and 108 of the U.S.

Copyright Law and except by reviewers for the public

press), without written permission from the publishers.

Printed in the United States of America by

Vail-Ballou Press, Binghamton, New York.

Library of Congress Cataloging-in-Publication Data

Kant, Immanuel, 1724–1804.

[Grundlegung zur Metaphysik der Sitten. English]

Groundwork for the metaphysics of morals / Immanuel Kant ;

edited and translated by Allen W. Wood ; with essays by J. B. Schneewind . . . [et al.].

p. cm.—(Rethinking the Western tradition)

Includes bibliographical references and index.

ISBN 0-300-09486-8 (cloth)—ISBN 0-300-09487-6 (paper)

1. Ethics—Early works to 1800. I. Wood, Allen W. II. Schneewind, J. B. (Jerome B.) III.

Title. IV. Series.

B2766.E6 W6613 2002

170—dc21 2002002605

A catalogue record for this book is available from the British Library.

The paper in this book meets the guidelines

for permanence and durability of the Committee on

Production Guidelines for Book Longevity of the

Council on Library Resources.

10 9 8 7 6 5 4 3 2 1

Contributors

J. B. Schneewind is professor of philosophy (emeritus) at Johns Hopkins

University.

Marcia Baron is professor of Philosophy at Indiana University.

Shelly Kagan is Henry R. Luce Professor of Social Thought and Ethics at Yale

University.

Allen W. Wood is Ward W. and Priscilla B. Woods Professor at Stanford

University.

Contents

Editor’s Preface ix

A Note on the Translation xiii

Abbreviations xvii

Text

Immanuel Kant: Groundwork for the Metaphysics of Morals (1785) 1 Preface 3 First Section: Transition from common rational moral cognition to philosophical moral cognition 9

Second Section: Transition from popular moral philosophy to the metaphysics of morals 22

Third Section: Transition from the metaphysics of morals to the critique of pure practical reason 63

Essays

1. Why Study Kant’s Ethics? 83

J. B. Schneewind

2. Acting from Duty 92

Marcia Baron

3. Kantianism for Consequentialists 111

Shelly Kagan

4. What Is Kantian Ethics? 157

Allen W. Wood

Glossary 183

Index 189

Editor’s Preface

Kant’s little book of 1785 is one of the most significant texts in the history of

ethics. It has been a standard of reference—sometimes a model to be devel-

oped and expanded on, sometimes a target of criticism—for moral philoso-

phers from the German idealist and German Romantic traditions, for Vic-

torians of the utilitarian school such as Mill and Sidgwick, for later British

idealists such as Green and Bradley, for the neo-Kantians, for twentieth-

century philosophers in both the continental and the anglophone traditions,

and for moral philosophers of all persuasions right down to the present day.

From the standpoint of the depth and originality of the ideas it contains, it

undoubtedly deserves this influence. But in the development of Kant’s own

moral thinking, it occupies a place that ought to make us question the

wisdom of treating it, the way moral philosophers customarily do, as the

definitive statement of Kant’s views on ethics.

Kant first gave notice of his intention to produce a system of moral

philosophy under the title ‘‘metaphysics of morals’’ about 1768. It took him

eighteen years to deliver even the first installment of the promised system,

which he gave a title indicative of the tentativeness and incompleteness of

what he thought he had so far accomplished: he was only laying the ground for a ‘‘metaphysics of morals’’ by seeking out and establishing its first principle.

Kant apparently began composing the Groundwork late in 1783. Letters written by Kant’s brilliant but eccentric friend J. G. Hamann report that he

began writing about moral philosophy in order to provide an ‘anticritique’

of Christian Garve’s 1783 book on Cicero’s treatise On Duties. But accord- ing to Hamann, during the spring of 1784 this critical discussion of Garve

on Cicero was transformed into something quite different, a ‘‘Prodromus

der Moral’’ (Ak 4:626–28). The title ‘‘Grundlegung zur Metaphysik der

Sitten’’ is first mentioned in a letter from Hamann in September 1784.

Hamann’s correspondence reveals him to be an avid Kant-watcher,

sometimes a helpfully critical one. But there is reason to be skeptical about

his account of the genesis of the Groundwork. Hamann’s account has in-

x Editor’s Preface

spired scholars as reputable as Klaus Reich and H. J. Paton to seek in the

Groundwork for allusions to Cicero, and even to think that they have found them. But there are no explicit references either to Cicero or to Garve’s

book about him. Kant may have been drawn to the subject of ethics in part

by reading and reflecting on Garve’s book or Cicero’s classical treatise, but

it seems unlikely that the Groundwork, as we now have it, could have grown out of a critical discussion of Garve on Cicero. The ‘‘Prodromus der

Moral’’ would seem to be a project independent of any ‘anticritique’ of

Garve that Kant could have been undertaking.

Kant was working on other topics in 1784 whose affinity with the ethical

theory presented in the Groundwork is also worth noting. For instance, he was reviewing Herder’s Ideas for the Philosophy of History of Humanity and writing two other short essays, Idea for a Universal History from a Cosmopolitan Standpoint and Answer to the Question: What Is Enlighten- ment? that reflect on human history, the social sources of the evil in human nature, the role of autonomous reason in directing our lives, and the rational

prospects for the moral progress of the human species. But perhaps no

special explanation is needed for the fact that Kant finally got around to

addressing a subject he had been promising to write on for the past sixteen

years.

Whatever the actual history of its genesis, the Groundwork went into press with Johann Hartknoch of Riga late in 1784. Throughout the winter,

and into the spring of 1785, Kant’s followers waited impatiently for its

appearance (Ak 4:628). Apparently the first copies were available on April

7. A second edition, altered in a number of passages throughout (but never

very greatly in any of them), appeared in 1786. This second edition went

through six more reprintings during Kant’s lifetime.

Kant seems always to have treated the Groundwork as a successful laying of the ground for the ethical theory presented in his later writings. But clearly

he soon came to regard it as not providing a complete or wholly clear presen-

tation even of the foundations of his system, for only three years later he

wrote a Critique of Practical Reason (1788) with the aim of clarifying those foundations, correcting misunderstandings, and answering criticisms of his

moral philosophy that had come from readers of the Groundwork. It is a matter of controversy how far what is said in the second Critique involves revisions of what Kant said in the Groundwork, but many scholars think that Kant meant to supplant the argument of the Third Section, where the

Groundwork establishes freedom of the will and relates freedom to the moral law. In the following decade Kant wrote a number of essays and treatises on

topics involving the application of his moral philosophy to politics, history,

Editor’s Preface xi

international relations, education, and religion. But it was only after he had

retired from university teaching, and as he began to realize that his mental

powers were beginning to fail him, that he finally assembled from the notes

and drafts of many years a work he called the Metaphysics of Morals, which was published as one of his very last works.

Kant’s essays and treatises of the 1790s, and especially the Metaphysics of Morals (1798), give us explicit accounts of many matters on which readers of the Groundwork customarily try to deduce the ‘‘Kantian view’’ (by triangulation, as it were) from what he says in this little foundational

treatise. Many doctrines standardly attributed to Kant on the basis of these

triangulations—on topics such as the nature of moral motivation, the rela-

tion between reason and feeling in human action, the structure of everyday

moral reasoning, and the nature of the will’s freedom—do not harmonize

very well with what Kant actually says in the Metaphysics of Morals, Religion within the Boundaries of Mere Reason, or other later works. This discrepancy strongly suggests that the Groundwork does not give us Kant’s final word on everything, and implies that where the Groundwork itself is not entirely explicit, it ought perhaps to be interpreted (often very dif-

ferently from the customary ways of interpreting it) in light of his other,

later, more explicit writings. But so influential has the Groundwork been, in comparison with his other ethical writings, that Kant will perhaps always be

burdened with what the long tradition of moral philosophers have read of

(and sometimes read into) what he said in his first foundational text on

moral philosophy.

The Groundwork is unquestionably the starting point not only for any study of Kant’s moral theory, but for any attempt to understand, develop, or

criticize any of the wide variety of ‘‘Kantian’’ ideas that have exercised

such a powerful influence on people’s thinking about morality, politics, and

religion in the centuries since this little book was first published. The trans-

lator and editor of this volume, as well as the writers of the four essays that

follow the text, hope they have presented Kant’s Groundwork in a way that will further its ongoing appropriation by everyone who thinks about the

fundamental issues raised in it.

A Note on the Translation

Kant’s Grundlegung zur Metaphysik der Sitten has had many English trans- lations. The most estimable are those by Thomas K. Abbott (1883), H. J.

Paton (1948), Lewis White Beck (1949, revised several times, most notably

in 1959 and 1990), and Mary J. Gregor (1996). Yet I have found even these

fine translations unsatisfying at certain points because, in order to provide a

smoother English reading, they are too often content to remain at a distance

from what Kant actually said, and because they sometimes commit them-

selves too much to one possible interpretation where the original text is

tantalizingly ambiguous. Also, over the years I have come to be aware that

some of their words and phrases, even some that now echo in the ears of us

who have for many years been reading the Groundwork in translation, are not the very best choices to translate precisely what Kant was saying.

In the present translation my aim has been to place the English reader, as

far as possible, in the same interpretive position as the German reader of the

original. Doing so has dictated taking pains to achieve accuracy and literal-

ness in the translation, as far as this can be made consistent with intel-

ligibility. It has also led to the attempt to preserve, as far as possible, a

consistency in terminology, not only with technical terms but even with

nontechnical ones. Where variations in meaning or context require the

same term to be translated in different ways, a numbered footnote informs

the reader of what is going on. (The unnumbered footnotes are Kant’s own.)

Kant’s paragraphing and even sentence structure have been respected, be-

cause Kant’s sentences often constitute units of argument, and modifying

them for the sake of more graceful English prose often makes the argument

harder to comprehend. Further, since my aim has been to put the English

reader in the same interpretive position as the German reader of the origi-

nal, I have not attempted to make the translation clearer or more elegant

than Kant’s German is; in fact, where Kant’s writing is obscure or awkward,

I have tried to reproduce the same murkiness and cumbersomeness in En-

glish that the German reader would encounter.

For these reasons, some will perhaps find this translation less smooth

xiv A Note on the Translation

and readable in places than the existing ones. Yet greater literalness and

transparency in a translation can often be as clarifying as confusing, as

much an invitation to ponder the meaning of the text as an obstacle to

understanding it. In such cases, the increased difficulty is, I believe, more

than compensated for by greater consistency, accuracy, and precision. I am

translating for those who want to know, insofar as they can know it from an

English translation, exactly what Kant said, so that they can have an accu-

rate basis for their own thinking, exegetical and critical, about what Kant

said. That sort of person will not be looking for easy reading.

The priorities in translating a text must obviously depend on the nature

and purpose of the text itself. Poetry should probably be translated only by

poets; philosophy certainly needs to be translated by philosophers. What

matters in a philosophical text is almost exclusively what it means. What a philosophical text means is constituted by the range of possible alternative

constructions that a reader’s philosophical imagination can justifiably put

on the words in which the text expresses its questions, doctrines, and argu-

ments. A translation succeeds, therefore, to the extent that it provides a

reliable basis for this work of imagination, neither constraining the reader

to adopt the translator’s own preferred imaginings nor suggesting possible

meanings that the original text cannot bear. Faithfulness to the precise

wording of the text is one way of achieving this; another is the use of a

consistent terminology, even if the reader must adjust slightly to an English

idiolect needed to convey the thoughts Kant expressed in German.

To a philosophical mind, the meaning of the text, taken in this sense,

matters incomparably more than the smoothness of the prose; difficult prose

is even an advantage if it provokes the kind of questioning, or even the

bewilderment, that leads to fruitful philosophical reflections that are also

really about what Kant was saying. To such a mind, in fact, there is some-

thing intellectually offensive about a translation that merely gestures in the

direction of what Kant said, leaving it to the common sense of readers (that

is, to the philosophical prejudices that a great historical text should help

them to unlearn) not merely to resolve the ambiguities, but even to deter-

mine where they are. Likewise, there is even something aesthetically repug-

nant about a translation whose smoothness of style glosses over philosophi-

cal difficulties for the sake of achieving a facility of comprehension or a

rhetorical elegance that were not in the original text.

Abbott’s translations of Kant’s foundational writings on ethics were

remarkable for their time because Abbott attempted accuracy when other

translators of philosophical works were often content with highly interpre-

tive paraphrases, or sought to interpose their own idiosyncratic readings of

A Note on the Translation xv

a text between it and the English reader. Paton, Beck and Gregor are supe-

rior to the degree that they represent increasing attentiveness to what the

text says. Such a trend seems healthy or even inevitable. The more carefully

a text is studied, the more closely and subtly it will be read, and the more

sensitive its readers will become to the need for translations that reproduce

as far as possible precisely the same interpretive situation as that confronted

by a reader of the original. Once a translation is available that achieves this

to a higher degree, philosophical readers will adjust to the inconveniences

they must incur in order to obtain the advantage. Readers who want to think

hard about what the text says will not be content with something less

accurate just because it is easier to read.

Another direct incitement to do a new translation of the Groundwork at this time was the availability of the new edition of the German text, pub-

lished by Bernd Kraft and Dieter Schönecker in Felix Meiner Verlag’s

Philosophische Bibliothek series. This text of the Grundlegung was used as the basis for the present translation. One of the special virtues of the new

Meiner Verlag edition is its attention to variations between the two earliest

versions of the text, the first published in 1785, the second a year later. The

edition usually follows the 1786 version, but notes inform the reader of the

differences. The present translation does likewise wherever textual differ-

ences make a difference in translation (which they usually do). In a few

places I have also followed the editors of the new text in making textual

emendations where the sense seems to require it. But I have done this only

reluctantly (and less often than the editors of the original text did); wher-

ever emendations are made, of course, a numbered footnote informs the

reader; in some cases, a note suggests a possible emendation, and what it

would have meant in the translation, but without actually adopting it.

This translation has benefited greatly from careful comments by, and

long discussions with, Dieter Schönecker. His care, precision, and linguis-

tic expertise and his intimate knowledge of the text of the Grundlegung saved me from many errors and led to many improvements in the transla-

tion. Schönecker and Kraft also made available to me a draft of their edi-

torial notes; I tried to reciprocate this favor by providing them with some

informational notes they did not yet have. Also helpful were textual correc-

tions and thoughtful stylistic suggestions made by Derek Parfit. In identify-

ing Kant’s references to classical philosophy and literature, I also benefited

from the expertise, erudition, and generosity of Tad Brennan, John Cooper,

and Elizabeth Tylawsky.

Abbreviations

Like this translation of the Groundwork, most writings of Kant available in English provide marginal volume: page numbers from the definitive Ger-

man edition (Ak). In the footnotes to this text of the Groundwork, the writings of Kant are cited by title in English and by Ak volume:page num-

ber. In the essays, they are cited by abbreviations listed here.

Ak Immanuel Kants Schriften. Ausgabe der königlich preussischen Akademie der Wissenschaften (Berlin: W. de Gruyter, 1902–)

Ca Cambridge Edition of the Writings of Immanuel Kant (New York: Cambridge University Press, 1992–)

G Grundlegung zur Metaphysik der Sitten (1785), Ak 4 Groundwork for the Metaphysics of Morals

KrV Kritik der reinen Vernunft (1781, 1787), cited by A/B pagination Critique of Pure Reason, Ca

KpV Kritik der praktischen Vernunft (1788), Ak 5 Critique of Practical Reason, Ca Practical Philosophy

MA Mutmasslicher Anfang der Menschengeschichte (1786), Ak 8 Conjectural Beginning of Human History, Ca Anthropology, His- tory and Education

MS Metaphysik der Sitten (1797–1798), Ak 6 Metaphysics of Morals, Ca Practical Philosophy

R Religion innerhalb der Grenzen der bloßen Vernunft (1793–1794), Ak 6

Religion within the Boundaries of Mere Reason, Ca Religion and Rational Theology

VA Anthropologie in pragmatischer Hinsicht (1798), Ak 7 Anthropology from a Pragmatic Standpoint, Ca Anthropology, History and Education Vorlesungen über Anthropologie, Ak 25

VE Vorlesungen über Ethik, Ak 27 Lectures on Ethics, Ca Lectures on Ethics

xviii Abbreviations

VL Vorlesungen über Logik, Ak 9, 24 Lectures on Logic, Ca Lectures on Logic

WA Beantwortung der Frage: Was ist Aufklärung? (1784), Ak 8 An Answer to the Question: What Is Enlightenment? Ca Practical Philosophy

Formulations of the Moral Law

Kant formulates the moral law in three principal ways. The first and third of

these have variants which are intended to bring the law closer to intuition

and make it easier to apply. These five principal formulations of the moral

law are abbreviated as follows.

First formula:

FUL The Formula of Universal Law: ‘‘Act only in accordance with that maxim through which you can at the same time will that it become

a universal law’’ (G 4:421; cf. G 4:402)

with its variant

FLN The Formula of the Law of Nature: ‘‘So act as if the maxim of your action were to become through your will a universal law of na- ture’’ (G 4:421; cf. G 4:436)

Second formula:

FH The Formula of Humanity as End in Itself: ‘‘Act so that you use humanity, as much in your own person as in the person of every other, always at the same time as end and never merely as means’’ (G 4:429; cf. G 4:436)

Third formula:

FA Formula of Autonomy: ‘‘the idea of the will of every rational being as a will giving universal law’’ (G 4:431; cf. G 4:432) or ‘‘Not to

choose otherwise than so that the maxims of one’s choice are at

the same time comprehended with it in the same volition as uni-

versal law’’ (G 4:440; cf. G 4:432, 434, 438)

with its variant,

FRE The Formula of the Realm of Ends: ‘‘Act in accordance with maxims of a universally legislative member for a merely possible

realm of ends’’ (G 4:439; cf. G 4:433, 437, 438)

Groundwork [Ak ∂:≥∫∑] for

the Metaphysics of Morals

Preface [Ak ∂:≥∫π]

Ancient Greek philosophy was divided into three sciences: physics, ethics, and logic.1 This division is perfectly suitable to the nature of the thing and one cannot improve upon it, except only by adding its principle, in order in

this way partly to secure its completeness and partly to be able to determine

correctly the necessary subdivisions.

All rational cognition is either material, and considers some object, or formal, and concerns itself merely with the form of the understanding and of reason itself and the universal rules of thinking in general, without

distinction among objects.2 Formal philosophy is called logic, but material philosophy, which has to do with determinate objects and the laws to which

they are subjected, is once again twofold. For these laws are either laws of

nature or of freedom. The science of the first is called physics, and that of the other is ethics; the former is also named ‘doctrine of nature’, the latter ‘doctrine of morals’.

Logic can have no empirical part, i.e., a part such that the universal and

necessary laws of thinking rest on grounds that are taken from experience;

for otherwise it would not be logic, i.e., a canon for the understanding or

reason which is valid for all thinking and must be demonstrated. By con-

trast, natural and moral philosophy can each have their empirical part,

because the former must determine its laws of nature as an object of experi-

ence, the latter must determine the laws for the will of the human being

insofar as he is affected by nature—the first as laws in accordance with

which everything happens, the second as those in accordance with which [Ak 4:388]

everything ought to happen, but also reckoning with the conditions under

which it often does not happen.

One can call all philosophy, insofar as it is based on grounds of experi-

1. According to Diogenes Laertius, Lives and Opinions of the Eminent Philosophers

7.39, this division was first devised by Zeno of Citium (335–265 b.c.) and was charac-

teristic of the Stoics. See, e.g., Seneca, Epistles 89.9; Cicero, On Ends 4.4.

2. Cf. Critique of Pure Reason, A50–55/B74–79.

4 Groundwork for the Metaphysics of Morals

ence, empirical, but that which puts forth its doctrines solely from princi- ples a priori, pure philosophy. The latter, when it is merely formal, is called logic; but if it is limited to determinate objects of the understanding, then3 it is called metaphysics. In such a wise there arises the idea of a twofold metaphysics, the idea of

a metaphysics of nature and of a metaphysics of morals. Physics will thus have its empirical but also a rational part; and ethics likewise; although here

the empirical part in particular could be called practical anthropology, but the rational part could properly be called morals.4

All trades, handicrafts, and arts have gained through the division of

labor, since, namely, one person does not do everything, but rather each

limits himself to a certain labor which distinguishes itself markedly from

others by its manner of treatment, in order to be able to perform it in the

greatest perfection and with more facility. Where labors are not so distin-

guished and divided, where each is a jack-of-all-trades, there the trades still

remain in the greatest barbarism. But it might be a not unworthy object of

consideration to ask whether pure philosophy in all its parts does not require

each its particular man, and whether it would not stand better with the

learned trade as a whole if those who, catering to the taste of the public, are

accustomed to sell the empirical along with the rational, mixed in all sorts

of proportions5 unknown even to themselves—calling themselves ‘inde-

pendent thinkers’,6 and those who prepare the merely rational part ‘quib-

blers’7—if they were warned not to carry on simultaneously two enter-

prises that are very different in their mode of treatment, each of which

perhaps requires a particular talent, and the combination of which in a

single person produces only bunglers: thus I here ask only whether the

nature of the science does not require the empirical part always to be

carefully separated from the rational, placing ahead of a genuine (empiri-

cal) physics a metaphysics of nature, and ahead of practical anthropology a

metaphysics of morals, which must be carefully cleansed of everything

3. 1785: ‘‘understanding, is called’’

4. Kant later includes ‘‘principles of application’’ drawn from ‘‘the particular nature of

human beings’’ within ‘‘metaphysics of morals’’ itself, leaving ‘‘practical anthropology’’

to deal ‘‘only with the subjective conditions in human nature that hinder people or help

them in fulfilling the laws of a metaphysics of morals’’ (Metaphysics of Morals, Ak

6:217).

5. Verhältnisse

6. Selbstdenker

7. Grübler

Groundwork for the Metaphysics of Morals 5

empirical, in order to know how much pure reason could achieve in both [Ak 4: 389]

cases; and from these sources pure reason itself creates its teachings a priori, whether the latter enterprise be carried on by all teachers of morals (whose name is legion) or only by some who feel they have a calling for it.

Since my aim here is properly directed to moral philosophy, I limit the

proposed question only to this: 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; for that there must be such is self-evident from the com-

mon idea of duty and of moral laws. Everyone must admit that a law, if it is

to be valid morally, i.e., as the ground of an obligation, has to carry absolute

necessity with it; 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; and likewise all the other genuinely moral laws; hence that the

ground of obligation here 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, and that every other precept grounded on

principles of mere experience, and even a precept that is universal in a

certain aspect, insofar as it is supported in the smallest part on empirical

grounds, perhaps only as to its motive, can be called a practical rule, but

never a moral law.

Thus not only are moral laws together with their principles essentially

distinguished among all practical cognition from everything else in which

there is anything empirical, but 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), but it gives him as a rational

being laws a priori, which to be sure require a power of judgment sharp- ened through experience, partly to distinguish in which cases they have

their application, and partly to obtain access for them to the will of the

human being and emphasis for their fulfillment, since he,8 as affected with

so many inclinations, is susceptible to the idea of a pure practical reason,

but is not so easily capable of making it effective in concreto in his course of life.

Thus a metaphysics of morals is indispensably necessary not merely

from a motive of speculation, in order to investigate the source of the [Ak 4:390]

practical principles lying a priori in our reason, but also because morals themselves remain subject to all sorts of corruption as long as that guiding

8. Kant’s text reads diese, which would be translated ‘‘the latter’’ and refer to ‘‘fulfill-

ment’; editors suggest amending it to dieser, which would refer to ‘the human being’.

6 Groundwork for the Metaphysics of Morals

thread and supreme norm of their correct judgment is lacking. For as to

what is to be morally good, it is not enough that it conform to the moral law, but it must also happen for the sake of this law; otherwise, that conformity is only contingent and precarious, because the unmoral ground will now and

then produce lawful actions, but more often actions contrary to the law. But

now the moral law in its purity and genuineness (which is precisely what

most matters in the practical) is to be sought nowhere else than in a pure

philosophy; hence this (metaphysics) must go first, and without it there

can be no moral philosophy at all; that which mixes those pure principles

among empirical ones does not even deserve the name of a ‘philosophy’

(for this distinguishes itself from common rational cognition precisely by

the fact that what the latter conceives only as mixed in, it expounds in

a separate science), still less of a ‘moral philosophy’, because precisely

through this mixture it violates the purity of morals and proceeds contrary

to its own end.

One should not think that what is here demanded we already have in the

propadeutic of the famous Wolff in his moral philosophy, namely in what he calls universal practical philosophy,9 and thus that here an entirely new field is not to be entered on. Precisely because it is supposed to be a

‘‘universal practical philosophy,’’ it has not drawn into consideration any

will of a particular kind, such as one determined without any empirical

motives fully from principles a priori, which one could call a ‘pure will’, but only volition in general, with all actions and conditions that pertain to it

in this universal signification; and thereby it is distinguished from a meta-

physics of morals just as general logic is from transcendental philosophy, of

which the first expounds the actions and rules of thinking in general, but the latter merely the particular actions and rules of pure thinking, i.e., those through which objects can be cognized fully a priori. For the metaphysics of morals is to investigate the idea and principles of a possible pure will, and not the actions and conditions of human volition in general, which are

for the most part drawn from psychology. It constitutes no objection to my

assertion that moral laws and duty are also discussed in universal practical[Ak 4:391]

philosophy (though contrary to all warrant). For in this too the authors of

that science remain faithful to their idea of it; they do not distinguish the

9. Christian Wolff (1679–1754), Philosophia Practica Universalis (1738–1739).

Kant uses the same title himself, however, as a subtitle to the section of the introduction to

the Metaphysics of Morals titled ‘‘Preliminary Concepts of the Metaphysics of Morals,’’

in which he discusses concepts such as freedom, duty, personhood, maxims, and laws (Ak

6:221–28).

Groundwork for the Metaphysics of Morals 7

motives that are represented as such fully a priori merely through reason, and are properly moral, from the empirical ones that understanding raises to

universal concepts through the comparison of experiences; but rather they

consider them, without respecting the distinction of their sources, only in

accordance with their greater or smaller sum (since they are all regarded as

homogeneous), and through that they make for themselves their concept of

obligation, which is to be sure not less than moral, but is so constituted as can be demanded only in a philosophy that does not judge about the origin of all practical concepts, whether they occur a priori or merely a posteriori. Now intending someday to provide a metaphysics of morals, I issue this

groundwork in advance.10 There is, to be sure, really no other foundation

for it than the critique of a pure practical reason,11 just as for metaphysics there is the already provided critique of pure speculative reason. Yet in part

the former is not of such utmost necessity as the latter, because in what is

moral human reason, even in the most common understanding, can easily

be brought to great correctness and completeness, whereas in its theoretical

but pure use it is entirely dialectical; in part I require for a critique of a pure

practical reason that if it is to be completed, its unity with the12 speculative

in a common principle must at the same time be exhibited, because it can in

the end be only one and the same reason that is distinguished merely in its

application. But I could not bring it to such a completeness here without

bringing in considerations of an entirely different kind and confusing the

reader. It is for the sake of this that instead of the term Critique of pure practical reason I have used instead Groundwork for the metaphysics of morals. But, thirdly, because a metaphysics of morals, despite its intimidating

title, is yet susceptible to a high degree of popularity and suitability to the

common understanding, I find it useful to separate from it this preliminary

work of laying the ground, in order that in the future I need not attach [Ak 4:392]

subtleties, which are unavoidable in it, to more easily grasped doctrines.

The present groundwork is, however, nothing more than the search for

10. Kant’s Metaphysics of Morals, Ak 6:205–493, was published in 1797–1798.

11. Kant published the Critique of Practical Reason, Ak 5:1–163, in 1788. But he

appears not to have intended to write a separate work with that title in 1785–1786. He

apparently planned to include a ‘‘practical’’ section in the second edition of the Critique

of Pure Reason (1787), but published the Critique of Practical Reason separately when it

grew too long for that.

12. In 1785 the definite article der is repeated; that version would be translated: ‘‘its

unity with the critique of speculative reason in a common principle.’’

8 Groundwork for the Metaphysics of Morals

and establishment of the supreme principle of morality, which already constitutes an enterprise whole in its aim and to be separated from every

other moral investigation. To be sure, my assertions about this important and

principal question, whose discussion has hitherto been far from satisfactory,

would receive much light through the application of the same principle to the

entire system, and of confirmation through the adequacy it manifests every-

where; yet I had to dispense with this advantage, which would also be

basically more a matter of my self-love than of the common utility, because

the facility of use and the apparent adequacy of a principle provide no wholly

secure proof of its correctness, but rather awaken a certain partiality not to

investigate and consider it for itself without any regard for the consequences.

The method I have taken in this work, I believe, is the one best suited if

one wants to take the way analytically from common cognition to the

determination of its supreme principle and then, in turn, synthetically from

the testing of this principle and its sources back to common cognition, in

which its use is encountered. Hence the division turns out thus:

First Section: Transition from common rational moral cognition to philosophical moral cognition.

Second Section: Transition from popular moral philosophy to the meta- physics of morals.

Third Section: Final step from the metaphysics of morals to the critique of pure practical reason.

First Section [Ak 4:393]

Transition

from common rational moral cognition

to philosophical moral cognition

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. Understanding, wit, the power of judgment,1

and like talents of the mind,2 whatever they might be called, or courage, resoluteness, persistence in an intention, as qualities of temperament, are without doubt in some respects good and to be wished for; but they 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 charac- ter,3 is not good. It is the same with gifts of fortune. Power, wealth, honor,4

even health and that entire well-being and contentment with one’s condi-

tion, under the name of happiness, make for courage and thereby often also for arrogance,5 where there is not a good will to correct their influence on

the mind,6 and thereby on the entire principle of action, and make them

universally purposive; not to mention that a rational impartial spectator can

never take satisfaction even in the sight of the uninterrupted welfare of a

being, if it is adorned with no trait of a pure and good will; and so the good

will appears to constitute the indispensable condition even of the worthi-

ness to be happy.

Some qualities are even conducive to this good will itself and can make

its work much easier, but still have despite this no inner unconditioned [Ak 4:394]

worth, yet always presuppose a good will, which limits the esteem7 that one

1. See Anthropology in a Pragmatic Respect, Ak 7:196–201.

2. Geist

3. For Kant’s distinction between ‘‘temperament’’ and ‘‘character,’’ see Anthropology

in a Pragmatic Respect, Ak 7:286–95; see also Ak 4:398–99 below.

4. Power, wealth, and honor are for Kant the three objects of the principal social

passions. See Anthropology in a Pragmatic Respect, Ak 7:271–274.

5. Mut und hierdurch öfters auch Übermut

6. Gemüt

7. 1786: Hochschätzung; 1785: Schätzung (‘‘estimation’’)

10 Groundwork for the Metaphysics of Morals

otherwise rightly has for them, and does not permit them to be held abso-

lutely good. Moderation in affects and passions,8 self-control, and sober

reflection not only are good for many aims, but seem even to constitute a

part of the inner worth of a person; yet they lack much in order to be de- clared good without limitation (however unconditionally they were praised

by the ancients).9 For without the principles of a good will they can become

extremely evil, and the cold-bloodedness of a villain makes him not only far

more dangerous but also immediately more abominable in our eyes than he

would have been held without it.

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, and 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. 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. Utility or fruitlessness can neither add to nor subtract

anything from this worth. 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.

There is, however, something so strange in this idea of the absolute worth

of the mere will, without making any allowance for utility in its estimation,

that despite all the agreement with it even of common reason, there must

nevertheless arise a suspicion that perhaps it is covertly grounded merely on

a high-flown fantasy, and that nature might have been falsely understood in

the aim it had in assigning reason to govern our will. Hence we will put this[Ak 4:395]

idea to the test from this point of view.

In the natural predispositions of an organized being, i.e., a being ar-

ranged purposively for life, we assume as a principle that no instrument is to

8. In Kant’s empirical theory of the faculty of desire, affects and passions are the two

principal obstacles to rational self-control. See Metaphysics of Morals, Ak 6:407–9;

Anthropology in a Pragmatic Respect, Ak 7:251–67.

9. Courage and self-control were, for the ancients, two of the primary moral virtues,

along with wisdom, justice, and sometimes piety. See Plato, Meno 78d–e, Republic 427e;

Aristotle, Nicomachean Ethics 3.6–12; Cicero, On Duties 1.15.

Groundwork for the Metaphysics of Morals 11

be encountered in it for any end except that which is the most suitable to and

appropriate for it.10 Now if, in a being that has reason and a will, its preser- vation, its welfare—in a word, its happiness—were the real end of nature, then nature would have hit on a very bad arrangement in appointing reason

in this creature to accomplish the aim. For all the actions it has to execute

toward this aim, and the entire rule of its conduct, would be prescribed to it

much more precisely through instinct, and that end could be obtained far

more safely through it than could ever happen through reason; and if, over

and above this, reason were imparted to the favored creature, it would have

served it only to make it consider the happy predisposition of its nature, to

admire it, to rejoice in it, and to make it grateful to the beneficent cause of it,

but not to subject its faculty of desire to that weak and deceptive guidance,

and meddle in the aim of nature; in a word, nature would have prevented

reason from breaking out into practical use and from having the presump- tion, with its weak insight, to think out for itself the project of happiness and

the means of attaining it; nature would have taken over the choice not only

of the ends but also of the means, and with wise provision would have

entrusted both solely to instinct.11

In fact we also find that the more a cultivated reason gives itself over to

the aim of enjoying life and happiness, the further the human being falls

short of true contentment; from this arises in many, and indeed in those most

practiced in the cultivated use of reason, if only they are sincere enough to

admit it, a certain degree of misology, i.e., hatred of reason;12 for after reckoning all the advantages they draw, I do not say from the invention of

all the arts of common luxury,13 but even from the sciences (which also

10. Kant’s reasons for accepting this proposition as an a priori maxim of reflective

judgment are presented in the Critique of the Power of Judgment (1790), § 66, Ak 5:376–

77.

11. Kant rejects the proposition that human happiness is an end of nature in his

writings on history and in his review of the chief work of his former student J. G. Herder

(1762–1802). See Idea toward a Universal History with a Cosmopolitan Aim (1784), Ak

8:19–20; Reviews of Herder’s Ideas for the Philosophy of History of Humanity (1785–

1786), Ak 8:64–65; Conjectural Beginning of Human History (1786), Ak 8:114–18. See

also Critique of the Power of Judgment, Ak 5:429–31. Though not an end of nature,

human happiness is an end of reason, and of morality; see Critique of Practical Reason,

Ak 5:61–62, 110–13; Metaphysics of Morals, Ak 6:387–88.

12. See Plato, Phaedo 89d–91b.

13. ‘‘Luxury (luxus) is excessive convenience in the social life of a community (so

that its convenience works against its welfare)’’; Anthropology in a Pragmatic Respect,

Ak 7:249.

12 Groundwork for the Metaphysics of Morals

seem to them in the end to be14 a luxury of the understanding), they nev-

ertheless find that they have in fact only brought more15 hardship down on[Ak 4:396]

their shoulders than they have gained in happiness, and on this account in

the end they sooner envy than despise human beings of the more common

stamp, who are closer to the guidance of mere natural instinct and do not

permit their reason much influence over their deeds and omissions. And we

must admit this much, that the judgment of those who very much moderate

the boastful high praise of the advantages that reason is supposed to supply

us in regard to happiness and contentment with life, or who even reduce it

below zero, is by no means morose or ungrateful toward the kindness of the

world’s government; but rather these judgments are covertly grounded on

the idea of another aim for their existence, possessing much greater dignity,

for which, and not for their happiness, reason has been given its wholly

authentic vocation, and to which, therefore, as a supreme condition, the

private aims of the human being must for the most part defer.

For since reason is not sufficiently effective in guiding the will safely in

regard to its objects and the satisfaction of all our needs (which it in part

itself multiplies), and an implanted natural instinct would have guided us

much more certainly to this end, yet since reason nevertheless has been

imparted to us as a practical faculty, i.e., as one that ought to have influence

on the will, its true vocation must therefore be not to produce volition as a means to some other aim, but rather to produce a will good in itself, for which reason was absolutely necessary, since everywhere else nature goes

to work purposively in distributing its predispositions. This will may there-

fore not be the single and 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 united 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, and can even diminish it to less than nothing

without nature’s proceeding unpurposively in this; for reason, which recog-

nizes its highest practical vocation in the grounding of a good will, is

capable in attaining this aim only of a contentment after its own kind,

namely from the fulfillment of an16 end that again only reason determines,

14. 1785 reads scheint instead of zu sein scheinen, which would have the effect in

translation of eliminating the words ‘‘to be’’ from this sentence.

15. 1785: ‘‘more of’’

16. 1785: ‘‘of the end’’

Groundwork for the Metaphysics of Morals 13

even if this should also be bound up with some infringement of the ends of

inclination. [Ak 4:397]

But now in order to develop the concept of a good will, to be esteemed in

itself and without any further aim, just as it dwells already17 in the naturally

healthy understanding, which does not need to be taught but rather only to

be enlightened, this concept always standing over the estimation of the

entire worth of our actions and constituting the condition for everything

else: 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 let it shine forth all the more brightly.

I pass over all actions that are already recognized as contrary to duty,

even though they might be useful for this or that aim; for with them the

question cannot arise at all whether they might be done from duty, since they even conflict with it. I also set aside the actions which are actually in

conformity with duty, for which, however, human beings have immediately

no inclination, but nevertheless perform them because they are driven to it through another inclination. For there it is easy to distinguish whether the

action in conformity with duty is done from duty or from a self-seeking aim. It is much harder to notice this difference where the action is in conformity

with duty and the subject yet has besides this an immediate inclination to it. E.g., it is indeed in conformity with duty that the merchant should not

overcharge his inexperienced customers, and where there is much commer-

cial traffic, the prudent merchant also does not do this, but rather holds a

firm general price for everyone, so that a child buys just as cheaply from

him as anyone else. 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; but here it is not

to be assumed that besides this, he was also supposed to have an immediate

inclination toward the customers, so that out of love, as it were, he gave no

one an advantage over another in his prices. Thus the action was done

neither from duty nor from immediate inclination, but merely from a self-

serving aim.

By contrast, to preserve one’s life is a duty, and besides this everyone has

an immediate inclination to it. But the often anxious care that the greatest

part of humankind takes for its sake still has no inner worth, and its maxim

has no moral content. They protect their life, to be sure, in conformity with [Ak 4:398] duty, but not from duty. If, by contrast, adversities and hopeless grief have

17. This word added in 1786

14 Groundwork for the Metaphysics of Morals

entirely taken away the taste for life, if the unhappy one, strong of soul,

more indignant than pusillanimous or dejected over his fate, wishes for

death and yet preserves his life without loving it, not from inclination or

fear, but from duty: then his maxim has a moral content.

To be beneficent where one can is a duty, and besides this there are some

souls so sympathetically attuned18 that, even without any other motive of

vanity or utility to self, 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. But I assert that in such a case the action, however it may

conform to duty and however amiable it is, nevertheless has no true moral

worth, but is on the same footing as other inclinations, e.g., the inclination

to honor, which, when it fortunately encounters something that in fact

serves the common good and is in conformity with duty, and is thus worthy

of honor, deserves praise and encouragement, but not esteem; for the

maxim lacks moral content, namely of doing such actions not from inclina-

tion but from duty. Thus suppose the mind of that same friend of humanity were clouded over with his own grief, extinguishing all his sympathetic

participation19 in the fate of others; he still has the resources to be beneficent

to those suffering distress, but the distress of others does not touch him

because he is20 sufficiently busy with his own; and now, where no inclina-

tion 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. Even

more: if nature had put little sympathy at all in the heart of this or that

person, if he (an honest man, to be sure) were by temperament cold and

indifferent toward the sufferings of others, perhaps because he himself is

provided with particular gifts of patience and strength to endure his own,

and also presupposes or even demands the same of others; if nature has not

really formed21 such a man into a friend of humanity (although he would not

in truth be its worst product), 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? By all means! Just here begins the worth of

character, which is moral and the highest without any comparison, namely[Ak 4:399]

that he is beneficent not from inclination but from duty.

To secure one’s own happiness is a duty (at least indirectly), for the lack

18. teilnehmend gestimmte Seelen

19. Teilnehmung

20. 1785: wäre

21. gebildet

Groundwork for the Metaphysics of Morals 15

of contentment with one’s condition, in a crowd of many sorrows and amid

unsatisfied needs, can easily become a great temptation to the violation of duties. But even without looking at duty, all human beings always have of themselves the most powerful and inward inclination to happiness, because

precisely in this idea all inclinations are united in a sum. Yet the precept of

happiness is for the most part so constituted that it greatly infringes on some

inclinations and yet the human being cannot make any determinate and

secure concept of the sum of satisfaction of them all, under the name of

‘happiness’; hence it is not to be wondered at that a single inclination,

which is determinate in regard to what it promises and the time in which its

satisfaction can be obtained, can outweigh a wavering idea; and the human

being, e.g., a person with gout, could choose to enjoy what tastes good and

to suffer what he must, because in accordance with his reckoning, here at

least he has not sacrificed the enjoyment of the present moment through

expectations, perhaps groundless, of a happiness that is supposed to lie in

health. But also in this case, if the general inclination to happiness does not

determine his will, if for him, at least, health does not count as so necessary

in his reckoning, then here, as in all other cases, there still remains a law,

namely to promote his happiness not from inclination but from duty, and

then his conduct has for the first time its authentic moral worth.

It is in this way, without doubt, that those passages in scripture are to be

understood in which it is commanded to love our neighbor and even our

enemy. For love as inclination cannot be commanded; but beneficence

solely from duty, even when no inclination at all drives us to it, or even

when natural and invincible disinclination resists, is practical and not path- ological love, which lies in the will and not in the propensity of feeling, in the principles of action and not in melting sympathy;22 but the former alone

can be commanded.

The second proposition23 is: an action from duty has its moral worth not in the aim that is supposed to be attained by it, but rather in the maxim in accordance with which it is resolved upon; thus24 that worth depends not on [Ak 4:400]

the actuality of the object of the action, but merely on the principle of the volition, in accordance with which the action is done, without regard to any object of the faculty of desire. It is clear from the preceding that the aims we

may have in actions, and their effects, as ends and incentives of the will, can

22. schmelzender Teilnehmung

23. Kant does not say explicitly what the ‘‘first proposition’’ was, but presumably it is

that an action has moral worth only if it is done from duty.

24. This word added in 1786

16 Groundwork for the Metaphysics of Morals

impart to the actions no unconditioned and moral worth. In what, then, can

this worth lie, if it is not supposed to exist in the will, in the relation of the

actions to the effect hoped for? It can lie nowhere else than in the principle of the will, without regard to the ends that can be effected through such action; for the will is at a crossroads, as it were, between its principle a priori, which is formal, and its incentive a posteriori, which is material, and since it must somehow be determined by something, it must be determined

through the formal principle in general of the volition if it does an action

from duty, since every material principle has been withdrawn from it.

The third proposition, as a consequence of the first two, I would express

thus: Duty is the necessity of an action from respect for the law. For the object, as an effect of my proposed action, I can of course have an inclina- tion, but never respect, just because it25 is merely an effect and not the activity of a will.26 Just as little can I have respect for inclination in general,

whether my own or another’s; I can at most approve it in the first case, in the

second I can sometimes even love it, i.e., regard it as favorable to my own

advantage. 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. Now

an action from duty is supposed entirely to abstract from27 the influence of

inclination, and with it every object of the will, so nothing is left over for the

will that can determine it except the law as what is objective and subjec- tively pure respect for this practical law, hence the maxim* of complying with such a law, even when it infringes all my inclinations.[Ak 4:401]

The moral worth of the action thus lies not in the effect to be expected

from it; thus also not in any principle of action which needs to get its motive

from this expected effect. For all these effects (agreeableness of one’s con-

dition, indeed even the furthering of the happiness of others) could be

brought about through other causes, and for them the will of a rational being

*A maxim is the subjective principle of the volition; the objective princi- ple (i.e., that which would serve all rational beings also subjectively as a

practical principle if reason had full control over the faculty of desire) is the

practical law. 25. Kant’s pronoun here is in the feminine, which could refer to ‘‘effect’’ but not to

‘‘object,’’ which seems to be the intended referent. Editors therefore often emend the

pronoun to the neuter.

26. 1785: ‘‘an effect of my will’’

27. absondern

Groundwork for the Metaphysics of Morals 17

is therefore not needed; but in it alone the highest and unconditioned good

can nevertheless be encountered. Nothing other than the representation of the law in itself, which obviously occurs only in the rational being insofar as it, and not the hoped-for effect, is the determining ground of the will,

therefore28 constitutes that so pre-eminent good which we call ‘moral’,

which is already present in the person himself who acts in accordance with

it, but must not first of all be expected from the effect.** [Ak 4:402]

But what kind of law can it be, whose representation, without even

**One could accuse me of merely taking refuge behind the word respect [Ak4:401] in an obscure feeling instead of giving a distinct reply to the question

through a concept of reason. Yet even if respect is a feeling, it is not one

received through influence but a feeling self-effected through a concept of reason and hence specifically distinguished from all feelings of the first

kind, which may be reduced to inclination or fear. What I immediately

recognize as a law for me, I recognize with respect, which signifies merely

the consciousness of the subjection of my will to a law without any media- tion of other influences on my sense. The immediate determination of the

will through the law and the consciousness of it is called respect, so that the latter is to be regarded as the effect of the law on the subject and not as its cause. Authentically, respect is the representation of a worth that infringes on my self-love. Thus it is something that is considered as an object neither

of inclination nor of fear, even though it has something analogical to both at

the same time. The object of respect is thus solely the law, and specifically that law that we lay upon ourselves and yet also as in itself necessary. As a law we are subject to it without asking permission of self-love; as laid upon

us by ourselves, it is a consequence of our will, and has from the first point

of view an analogy with fear, and from the second with inclination. All

respect for a person is properly only respect for the law (of uprightness,

etc.) of which the person gives us the example. Because we regard the

expansion of our talents also as a duty, we represent to ourselves a person

with talents also as an example of a law, as it were (to become similar to the person in this) and that constitutes our respect. All so-called moral interest consists solely in respect for the law. [The parenthetical material in the penultimate sentence was added in 1786. Cf. Critique of Practical Reason, Ak 5:71–89. In the Metaphysics of Morals, Kant lists four feelings that are produced directly by reason and can serve as moral motivation. These are

‘‘moral feeling,’’ ‘‘conscience,’’ ‘‘love of human beings,’’ and ‘‘respect’’

(Metaphysics of Morals, Ak 6:399–403).] 28. 1785: ‘‘thus’’

18 Groundwork for the Metaphysics of Morals

taking account of the effect expected from it, must determine the will, so

that it can be called good absolutely and without limitation? Since I have

robbed the will of every impulse that could have arisen from the obedience

to any law, there is nothing left over except the universal lawfulness of the

action in general which alone is to serve the will as its principle, i.e., I ought

never to conduct myself except so that I could also will that my maxim become a universal law. Here it is mere lawfulness in general (without grounding it on any law determining certain actions) that serves the will as

its principle, and also must so serve it, if duty is not to be everywhere an

empty delusion and a chimerical concept; common human reason,29 indeed,

agrees perfectly with this in its practical judgment, and has the principle just

cited always before its eyes.

Let the question be, e.g.: When I am in a tight spot, may I not make a

promise with the intention of not keeping it? Here I easily make a distinc-

tion in the signification the question can have, whether it is prudent, or

whether it is in conformity with duty, to make a false promise. The first can

without doubt often occur. I do see very well that it is not sufficient to get

myself out of a present embarrassment by means of this subterfuge, but

rather it must be reflected upon whether from this lie there could later arise

much greater inconvenience than that from which I am now freeing myself,

and, since the consequences of my supposed cunning are not so easy to foresee, and a trust once lost to me might become much more disadvan-

tageous than any ill I think I am avoiding, whether it might not be more

prudent to conduct myself in accordance with a universal maxim and make it into a habit not to promise anything except with the intention of keeping

it. Yet it soon occurs to me here that such a maxim has as its ground only the

worrisome consequences. Now to be truthful from duty is something en-

tirely different from being truthful out of worry over disadvantageous con-

sequences; in the first case, the concept of the action in itself already con-

tains a law for me, whereas in the second I must look around elsewhere to

see which effects might be bound up with it for me. For if I deviate from the

principle of duty, then this is quite certainly evil; but if I desert my maxim[Ak 4:403]

of prudence, then that can sometimes be very advantageous to me, even

though it is safer to remain with it. Meanwhile, to inform myself in the

shortest and least deceptive way in regard to my answer to this problem,

whether a lying promise is in conformity with duty, I ask myself: Would I

be content with it if my maxim (of getting myself out of embarrassment

through an untruthful promise) should be valid as a universal law (for

29. 1785: ‘‘but common human reason’’

Groundwork for the Metaphysics of Morals 19

myself as well as for others), and would I be able to say to myself that

anyone may make an untruthful promise when he finds himself in embar-

rassment which he cannot get out of in any other way? Then I soon become

aware that I can will the lie but not at all a universal law to lie; for in

accordance with such a law there would properly be no promises, because it

would be pointless to avow my will in regard to my future actions to those

who would not believe this avowal, or, if they rashly did so, who would pay

me back in the same coin; hence my maxim, as soon as it were made into a

universal law, would destroy itself.

Thus I need no well-informed shrewdness to know what I have to do in

order to make my volition morally good. Inexperienced in regard to the

course of the world, incapable of being prepared for all the occurrences that

might eventuate in it, I ask myself only: Can you will also that your maxim

should become a universal law? If not, then it is reprehensible, and this not

for the sake of any disadvantage impending for you or someone else, but

because it cannot fit as a principle into a possible universal legislation; but

for this legislation reason extorts immediate respect from me, from which,

to be sure, I still do not have insight into that on which it is grounded (which the philosopher may investigate), but I at least understand this much, that it

is an estimation of a worth which far outweighs everything 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. Thus in the moral cognition of common human reason we have attained

to its principle, which it obviously does not think abstractly in such a

universal form, but actually has always before its eyes and uses as its

standard of judgment. It would be easy here to show how, with this compass [Ak 4:404]

in its hand, it knows its way around very well in all the cases that come

before it, how to distinguish what is good, what is evil, what conforms to

duty or is contrary to duty, if, without teaching it the least new thing, one

only makes it aware of its own principle, as Socrates did;30 and thus that it

needs no science and philosophy to know what one has to do in order to be

honest and good, or indeed, even wise and virtuous. It might even have

been conjectured in advance that the acquaintance with what every human

being is obliged to do, hence to know, would also be the affair of everyone,

30. This would appear to be Kant’s interpretation of Socrates’ ‘‘human wisdom’’

(Plato, Apology 20c–24b). Compare Metaphysics of Morals, Ak 6:411.

20 Groundwork for the Metaphysics of Morals

even of the most common human being. Here31 one cannot regard without

admiration the way the practical faculty of judgment is so far ahead of the

theoretical in the common human understanding. In the latter, if common

reason ventures to depart from the laws of experience and perceptions of

sense, then it falls into sheer inconceivabilities and self-contradictions, or at

least into a chaos of uncertainty, obscurity, and inconstancy. But in the

practical, the power of judgment first begins to show itself to advantage

when the common understanding excludes from practical laws all sensuous

incentives. It then even becomes subtle, caviling with its conscience, or

with other claims in reference to what is to be called right, or even in

wanting sincerely to determine the worth of actions for its own instruc-

tion,32 and, what is most striking, it can in the latter case do so with just as

good a hope of getting things right as any philosopher might promise to do;

indeed, it is almost more secure in this even than the latter, because the

philosopher has33 no other principle than the common understanding, but

the philosopher’s judgment is easily confused by a multiplicity of consider-

ations that are alien and do not belong to the matter and can make it deviate

from the straight direction. Would it not accordingly be more advisable in

moral things to stay with the judgment of common reason, and bring in

philosophy at most only in order to exhibit the system of morals all the more

completely and comprehensibly, and its rules in a way that is more conve-

nient for their use (still more for disputation), but not in order to remove the

common human understanding in a practical respect out of its happy sim-

plicity, and through philosophy to set it on a new route of investigation and

instruction?

There is something splendid about innocence, but it is in turn very bad[Ak 4:405]

that it cannot be protected very well and is easily seduced. On this account

even wisdom—which consists more in deeds and omissions than in knowl-

edge—also needs science, not in order to learn from it but in order to

provide entry and durability for its precepts. The human being feels in

himself a powerful counterweight against all commands of duty, which

reason represents to him as so worthy of esteem, in his needs and inclina-

tions, whose satisfaction he summarizes under the name of ‘happiness’.

Now reason commands its precepts unremittingly, without promising any-

thing to inclinations, thus snubbing and disrespecting, as it were, those

impetuous claims, which at the same time seem so reasonable (and will not

31. 1785: ‘‘Nevertheless’’

32. 1785: Belohnung (‘‘reward’’); 1786: Belehrung (‘‘instruction’’)

33. 1785: ‘‘can have’’

Groundwork for the Metaphysics of Morals 21

be done away with by any command). From this, however, arises a natural dialectic, that is, a propensity to ratiocinate against those strict laws of duty and to bring into doubt their validity, or at least their purity and strictness,

and,34 where possible, to make them better suited to our wishes and inclina-

tions, i.e., at ground to corrupt them and deprive them of their entire dignity,

which not even common practical reason can in the end call good.

Thus common human reason is impelled, not through any need of spec- ulation (which never assaults it as long as it is satisfied with being mere

healthy reason), but rather from practical grounds themselves, to go outside

its sphere and to take a step into the field of practical philosophy, in order to receive information and distinct directions about the source of its principle

and its correct determination in opposition to the maxims based on need and

inclination, so that it may escape from its embarrassment concerning the

claims of both sides and not run the risk of being deprived, through the

ambiguity into which it easily falls, of all genuine ethical principles. Thus

even in common practical reason, when it is cultivated, there ensues un-

noticed a dialectic, which necessitates it to seek help in philosophy, just as befalls it in its theoretical use; and therefore the first will find no more

tranquillity than the other anywhere except in a complete critique of our

reason.

34. 1785: ‘‘at least’’

Second Section[Ak4:406]

Transition from popular moral philosophy

to

the metaphysics of morals

If we have thus far drawn our concept of duty from the common use of our

practical reason, it is by no means to be inferred from this that we have

treated it as a concept of experience. Rather, if we attend to the experience

of the deeds and omissions of human beings, we encounter frequent and, as

we ourselves concede, just complaints that one could cite no safe examples

of the disposition to act from pure duty; that, even if some of what is done

may accord with what duty commands, nevertheless it always1 remains doubtful whether2 it is really done from duty and thus has a moral worth. Hence3 in all ages there have been philosophers who have absolutely denied

the actuality of this disposition in human actions, and have ascribed every-

thing to a more or less refined self-love, yet without bringing the correct-

ness of the concept of morality into doubt; rather, they have mentioned4

with inward regret the fragility and impurity of human nature,5 which is, to

be sure, noble enough to make an idea so worthy of respect into its precept,

but at the same time is too weak to follow it, and uses reason, which ought

to serve it for legislation, only in order to take care of the interest of

inclinations, whether singly or at most in their greatest compatibility with

one another.[Ak 4:407]

In fact it is absolutely impossible to settle with complete certainty

through experience whether there is even a single case in which the maxim

1. 1785: ‘‘thus’’

2. 1785: ‘‘that’’

3. 1785 omits this word and treats the following sentence as a clause subordinate to

the previous sentence.

4. 1786 adds this verb construction Erwähnung taten

5. In Religion within the Boundaries of Mere Reason, Kant lists ‘‘fragility’’ (the

inability to hold to good maxims, once they are adopted) and ‘‘impurity’’ (the need for

nonmoral incentives to do one’s duty) as the two lesser degrees of the radical evil in

human nature, along with the highest degree, ‘‘depravity’’ (the propensity to place incen-

tives of inclination ahead of those of duty) (Ak 6:29–30).

Groundwork for the Metaphysics of Morals 23

of an otherwise dutiful action has rested solely on moral grounds and on the

representation of one’s duty. For it is sometimes the case that with the most

acute self-examination we encounter nothing that could have been power-

ful enough apart from the moral ground of duty to move us to this or that

good action and to so great a sacrifice; but from this it cannot be safely

inferred that it was not actually some covert impulse of self-love, under the

mere false pretense of that idea, that was the real determining cause of the

will; so we would gladly flatter ourselves with a false presumption of a

nobler motive, while in fact even through the most strenuous testing, we

can never fully get behind the covert incentives, because when we are

talking about moral worth, it does not depend on the actions, which one

sees, but on the inner principles, which one does not see.6

One cannot better serve the wishes of those who ridicule all morality, as

a mere figment of the mind overreaching itself though self-conceit, than to

concede to them that the concepts of duty must be drawn solely from

experience (as one is gladly persuaded, for the sake of convenience, in the

case of all other concepts); for in this way one prepares for them a certain

triumph. From love of humanity I will concede that most of our actions are

in conformity with duty; but if one looks more closely at ‘‘the imagination

of the thoughts of their hearts,’’7 then everywhere one runs into the dear

self, which is always thrusting itself forward;8 it is upon this that the aim is

based, and not on the strict command of duty, which would often demand

self-renunciation. One does not need to be an enemy of virtue, but only a

cold-blooded observer, who does not take the liveliest wish for the good

straightway as its reality, in order (especially with advancing years, and a

power of judgment grown shrewder through experience and more acute for

observation) to become doubtful at certain moments whether any true vir-

tue is ever really to be encountered in the world. And here nothing can

protect us from falling away entirely from our ideas of duty and preserve in

our soul a well-grounded respect toward its law, except the clear conviction

that even if there have never been actions that have arisen from such pure

6. Cf. 2 Corinthians 4:18: ‘‘While we look not at the things which are seen, but at the

things which are not seen: for the things that are seen are temporal; but the things which

are not seen are eternal.’’

7. ihr Dichten und Trachten; this is an allusion to the phrase Tichten und Trachten in

the Lutheran translation of Genesis 6:5, which reads (in the King James version): ‘‘And

God saw that the wickedness of man was great in the earth, and that every imagination of

the thoughts of his heart was only evil continually.’’

8. See Anthropology in a Pragmatic Respect, § 2, Ak 7:128–30.

24 Groundwork for the Metaphysics of Morals

sources, yet nevertheless we are not talking here about whether this or that[Ak 4:408]

happens, but rather reason commands, for itself and independently of all

appearances, what ought to happen; hence actions, of which perhaps the

world has up to now given no example and about which one might, ground-

ing everything on experience, very much doubt even their feasibility, are

nevertheless commanded unremittingly by reason; and that, e.g., pure hon-

esty in friendship can no less be demanded of every human being, even if up

to now there may not have been a single honest friend,9 because this duty, as

duty in general, lies prior to all experience in the idea of a reason determin-

ing the will through a priori grounds. If one adds that unless one wants to dispute whether the concept of

morality has any truth and relation to any possible object, one could not

deny that its law is of such an extensive significance that it would have to be

valid not merely for human beings but for all rational beings in general, and not merely under contingent conditions and with exceptions, but with abso- lute necessity, then it is clear that no experience could give occasion for inferring even the possibility of such apodictic laws.10 For with what right

could we bring into unlimited respect, as a universal precept for every

rational nature, that which is perhaps valid only under the contingent condi-

tions of humanity, and how should laws for the determination of our will be taken as laws for the determination of the will of a rational being in general,

and only as such also for our will, if they were merely empirical and did not

take their origin fully a priori from pure but practical reason? Nor could one give worse advice to morality than by trying to get it from

examples. For every example of morality that is to be represented to me as

such must itself be previously judged in accordance with principles of

9. ‘‘Friendship thought of as attainable in its purity or completeness (between Orestes

and Pylades, Thesesus and Pirithous) is the hobbyhorse of writers of romances. On the

other hand, Aristotle says: ‘My dear friends, there are no friends!’ ’’ (Metaphysics of

Morals, Ak 6:470). The statement attributed to Aristotle is based on Diogenes Laertius,

Lives and Opinions of Eminent Philosophers 5.1.21.

10. The original meaning of ‘apodictic’ is ‘self-evident’ (from the Greek ≤apó +

deíknumi). But Kant more typically uses it in the sense of ‘necessary’ (this is its apparent

meaning in the Table of Judgments, Critique of Pure Reason A70/B95); yet an epistemic

element of certainty is often intended as well. For example: ‘‘Geometrical propositions

are all apodictic, i.e., combined with consciousness of their necessity’’ (Critique of Pure

Reason B 41; cf. A160/B199); ‘‘[Mathematical cognition] carries with it thoroughly

apodictic certainty (i.e., absolute necessity), hence rests on no grounds of experience’’

(Prolegomena to Any Future Metaphysics, § 6, Ak 4:280).

Groundwork for the Metaphysics of Morals 25

morality as to whether it is worthy to serve as an original11 example, i.e., as

a model; but it can by no means by itself12 supply the concept of morality.

Even the holy one of the Gospel must first be compared with our ideal of

moral perfection before one can recognize him as holy; he says this about

himself too: Why do you call me (whom you see) good? No one is good (the

archetype of the good) except the one God (whom you do not see).13 But

where do we get the concept of God as the highest good? Solely from the [Ak 4:409]

idea that reason projects a priori of moral perfection and connects insepara- bly with the concept of a free will. In morality there is no imitation, and

examples serve only for encouragement, i.e., they place beyond doubt the

feasibility of what the law commands, they make intuitive what the practi-

cal rule expresses universally; but they can never justify setting aside their

true original,14 which lies in reason, and in directing ourselves in accor-

dance with examples.

If, then, there is no genuine supreme principle of morality which does

not have to rest on pure reason independent of all experience, then I believe

it is not necessary even to ask whether it is good to expound these concepts

in general (in abstracto), as they, together with the principles belonging to them, are fixed a priori, provided that this cognition is distinguished from common cognition and is to be called ‘philosophical’. But in our age this

might well be necessary. For if one were to collect votes on which is to be

preferred, a pure rational cognition abstracted from everything empirical,

hence a metaphysics of morals, or popular practical philosophy, then one

would soon guess on which side the preponderance15 will fall.16

11. 1785: ‘‘genuine’’

12. zu oberst

13. ‘‘ ‘Why do you call me good?’ Jesus answered. ‘No one is good except God

alone’ ’’ (Luke 18:19; cf. Matthew 19:17, Mark 10:18). As in note 6 above, compare also

2 Corinthians 4:18.

14. Original

15. 1785: ‘‘the truth’’

16. Kant’s references to ‘‘popular philosophy’’ are primarily allusions to a movement

of German Enlightenment philosophers, centered chiefly in Berlin, whose best-known

representatives were Christian Garve (1742–1798), Moses Mendelssohn (1729–1786),

Christoph Meiners (1747–1810), and Christoph Friedrich Nicolai (1733–1811). Other

critical references to this movement can be found throughout Kant’s writings (Critique of

Pure Reason A x, A855/B883; Prolegomena, Ak 4:261–62, 371–83; What Does It Mean

To Orient Oneself in Thinking? Ak 8:133–46; On the Common Saying ‘‘That May Be

Correct in Theory, but Does Not Work in Practice’’ Ak 8:278–89; Metaphysics of Morals,

26 Groundwork for the Metaphysics of Morals

This condescension to popular concepts17 is to be sure very laudable

when the elevation to principles of pure reason has already been achieved to

full satisfaction, and that would mean first grounding the doctrine of morals on metaphysics, but procuring entry for it by means of popularity, once it stands firm. But it is quite absurd to want to humor popularity in the first

investigation, upon which depends the correctness of principles. Not only

can this procedure never lay claim to the extremely rare merit of a true

philosophical popularity, since there is no art in being commonly under- standable if one relinquishes all well-grounded insight; this produces only a

disgusting mish-mash of patched-together observations and half-reasoned

principles, in which superficial minds revel, because there is always some-

thing serviceable for everyday chitchat, but which insightful people dis-

regard, feeling confused and dissatisfied without being able to help them-

selves; yet philosophers, who can very well see through the illusion,18 find

little hearing when for certain occasions they decry this supposed popu-[Ak 4:410]

larity, in order, through acquiring determinate insight, finally to gain the

right to be popular .

One need only look at the essays on morality adapted to this favored

taste; then one will sometimes encounter the particular vocation of human

nature (but occasionally also the idea of a rational nature in general), some-

times perfection, sometimes happiness, here moral feeling, there fear of

God, some of this and some of that, all in a wondrous mixture, without its

occurring to anyone to ask whether the principles of morality are to be

sought anywhere in the knowledge of human nature (which we can obtain

only through experience); and if not, if these principles are to be encoun-

tered in pure concepts of reason, fully a priori, free from everything empiri- cal, and nowhere else even in the smallest part, then one may seize the

Ak 6:206; On Turning Out Books, Ak 8: 433–37; Logic, Ak 9:19–20, 148). Despite this,

Kant was on terms of friendship and mutual admiration with at least two members of the

movement, namely Mendelssohn and Garve. Some scholars have maintained the thesis

that Garve’s translation, with notes, of Cicero’s On Duties greatly influenced the Ground-

work itself, including its account of the good will and its three formulations of the moral

law. See Klaus Reich, ‘‘Kant and Greek Ethics,’’ Mind 47 (1939), and A. R. C. Duncan,

Practical Reason and Morality (London: Nelson, 1957), chap. 11. For a convincing

refutation of this thesis, see Reiner Wimmer, Universalisierung in der Ethik (Frankfurt:

Suhrkamp, 1980), pp. 183–84; and Dieter Schönecker, Kant: Grundlegung III. Die De-

duktion des kategorischen Imperativs (Freiburg: Alber Verlag, 1999), pp. 61–67.

17. Volksbegriffen

18. Blendwerk

Groundwork for the Metaphysics of Morals 27

initiative by entirely separating this investigation as pure practical philoso-

phy, or (if one may use such a disreputable term) as metaphysics* of morals,

bringing it for itself alone to its entire completeness, and deferring the

expectations of the public, which demands popularity, until the completion

of this undertaking.

But such a fully isolated metaphysics of morals, mixed with no anthro-

pology, with no theology, with no physics or hyperphysics, still less with

occult qualities (which one might call ‘hypophysical’), is not only an indis-

pensable substrate of all theoretical cognition of duties which is securely

determined, but it is at the same time also a desideratum of the highest

importance for the actual fulfillment of its precepts. For the pure representa-

tion of duty and the moral law in general, mixed with no alien addition from

empirical stimuli, has, by way of reason alone (which thereby for the first

time becomes aware that it can for itself be practical), an influence on the

human heart so much more powerful than all other incentives** that might [Ak 4:411]

*One can, if one wants, distinguish the ‘pure’ philosophy of morals

(metaphysics) from the ‘applied’ (namely to human nature) ( just as ‘pure’

mathematics and ‘pure’ logic are distinguished from ‘applied’). By this

terminology one is directly reminded that moral principles are not grounded

on the peculiarities of human nature, but must be subsistent a priori for themselves; but from them human practical rules must be derivable, as for

every rational nature.

**I have a letter from the late excellent Sulzer, in which he asks me what the cause might be that the doctrines of virtue, however convincing they

may be to reason, yet accomplish so little. My answer, through being pre-

pared so as to be complete, came too late. Yet it is nothing except that the

teachers have not brought their concepts to purity, and because they were

trying to do too much by scaring up motivations to be morally good from

everywhere, in trying to strengthen their medicine they ruin it. For the most

common observation shows that when one represents an upright action as it

is carried out with a steadfast soul even under the greatest temptations of

distress or of enticement, separate from every intention for any advantage

in this or in another world, it leaves far behind and eclipses every similar

action which is affected even in the slightest with an alien incentive; it

elevates the soul and inspires the wish to be able also to act that way. Even

moderately young children feel this impression, and one should never rep-

resent duty to them otherwise than this. [Johann Georg Sulzer (1720–

1779), director of the philosophical division of the Prussian Academy of

Sciences (1777–1779). The letter in question is usually thought to be the

28 Groundwork for the Metaphysics of Morals

be summoned from the empirical field, that reason, in the consciousness of

its dignity, despises the latter, and can gradually become their master; in

place of this, a mixed doctrine of morals, composed from incentives of

feelings and inclinations and simultaneously from concepts of reason, must

make the mind waver between motivations that cannot be brought under

any principle, and can lead us only very contingently to the good, but often

also to the evil.

From what we have adduced it is clear that all moral concepts have their

seat and origin fully a priori in reason, and this as much in the most common human reason as in that reason which is in highest measure spec-

ulative; that these concepts cannot be abstracted from any empirical, and

therefore mere contingent, cognition; that their dignity lies precisely in this

purity of their origin, so that19 they serve us as supreme practical principles;

that whatever one adds to them of the empirical, one withdraws that much

from their genuine influence and from the unlimited worth of actions; that it

is not only of the greatest necessity for theoretical aims, when it is merely a

matter of speculation, but it is also of the greatest practical importance, to

demand that their concepts and laws should be taken from pure reason, to

expound them pure and unmixed, indeed, to determine the range of this

entire practical or pure rational cognition, i.e., the entire faculty of pure

practical reason; but not as speculative philosophy permits, or indeed at

times finds necessary, making the principles dependent on the particular[Ak 4:412]

nature of human reason, but rather, since moral laws are to be valid for

every rational being in general, to derive them from the universal concept of

a rational being in general; and in such a way all morality, which needs

one dated December 8, 1770 (see Ak 13:51), which, however, does not

directly raise the question Kant says it does. What Sulzer does say is this: ‘‘I

really wished to hear from you whether we may soon hope to see your work

on the metaphysics of morals. This work is of the highest importance, given

the present unsteady state of moral philosophy. I have tried to do something

of this sort myself in attempting to resolve the question, ‘What actually is

the physical or psychological difference between a soul that we call vir-

tuous and one which is vicious?’ I have sought to discover the true disposi-

tions of virtue and vice in the first manifestations of representations and

sensations, and I now regard my undertaking of this investigation as less

futile, since it has led me to concepts that are simple and easy to grasp, and

which one can effortlessly apply to the teaching and raising of children. But

this work, too, is impossible for me to complete at present’’ (Ak 10:112).]

19. This word added in 1786

Groundwork for the Metaphysics of Morals 29

anthropology for its application to human beings, must first be expounded completely, independently of anthropology, as pure philosophy, i.e., as

metaphysics (which it is possible to do in this species of entirely separate

cognitions); but we must also be conscious that without being in possession

of this, it would be futile, I will not say to determine precisely for specula-

tive judgment what is moral about duty in everything that conforms to duty,

but that it would even be impossible in a common and practical use, chiefly

in moral instruction, to ground morality on its genuine principles and

thereby to effect pure moral dispositions and implant them in people’s

minds for the highest good of the world.20

But now in order to progress by natural steps in this work not merely

from the common moral judgment (which is here worthy of great respect)

to the philosophical, as has already been done, but also from a popular

philosophy, which goes no further than it can get through groping by means

of examples, up to metaphysics (which is not any longer held back by

anything empirical and, since it must cover the entire sum total of rational

cognition of this kind, goes as far as ideas, where even examples desert us),

we must follow and distinctly exhibit the practical faculty of reason from its

universal rules of determination up to where the concept of duty arises from

it.

Every thing in nature works in accordance with laws. Only a rational

being has the faculty to act in accordance with the representation of laws, i.e., in accordance with principles, or a will. Since for the derivation of actions from laws reason is required, the will is nothing other than practical reason. If reason determines the will without exception, then the actions of

such a being, which are recognized as objectively necessary, are also sub-

jectively necessary, i.e., the will is a faculty of choosing only that which reason, independently of inclination, recognizes as practically necessary,

i.e., as good. But if reason for itself alone does not sufficiently determine the

will, if the will is still subject to subjective conditions (to certain incentives)

which do not always agree with the objective conditions, in a word, if the [Ak 4:413]

will is not in itself fully in accord with reason (as it actually is with human beings), then the actions which are objectively recognized as necessary are

subjectively contingent, and the determination of such a will, in accord with

objective laws, is necessitation, i.e., the relation of objective laws to a will which is not thoroughly good is represented as the determination of the will

of a rational being through grounds of reason to which, however, this will in

accordance with its nature is not necessarily obedient.

20. Vom höchsten Weltbesten

30 Groundwork for the Metaphysics of Morals

The representation of an objective principle, insofar as it is necessitating

for a will, is called a ‘command’ (of reason), and the formula of the com-

mand is called an imperative. All imperatives are expressed through an ought and thereby indicate the

relation of an objective law of reason to a will which in its subjective

constitution is not necessarily determined by that law (a necessitation).

They say that it would be good to do or refrain from something, but they say

it to a will that does not always do something just because it is represented

to it as good to do. Practical good, 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. It 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.*[Ak 4:414]

A perfectly good will would thus stand just as much under objective

laws (of the good), but it would not be possible to represent it as necessi- tated by them to lawful actions, because of itself, in accordance with its subjective constitution, it can be determined only through the representa-

tion of the good. Hence for the divine will, and in general for a holy will, no imperatives are valid; the ought is out of place21 here, because the volition is of itself already necessarily in harmony with the law. Hence imperatives are

*The dependence of the faculty of desire on sensations is called ‘inclina-[Ak4:413]

tion’, and this always therefore proves a need. But the dependence of a contingently determinable will on principles of reason is called an interest. This occurs, therefore, only with a dependent will, which does not always

of itself accord with reason; with the divine will one cannot think of any

interest. But the human will, too, can take an interest without therefore acting from interest. The former signifies the practical interest in the action, the second the pathological interest in the object of the action. The first indicates only the dependence of the will on principles of reason in itself,

the second on those principles of reason on behalf of inclination, where,

namely, reason furnishes only the practical rule as to how the need of

inclination is to be supplied. In the first case the action interests me, in the

second the object of the action (insofar as it is agreeable to me). In the First[Ak4:414]

Section we have seen that with an action from duty it is not the interest in an

object that has to be looked to, but merely the action itself and its principle

in reason (the law).

21. am unrechten Orte

Groundwork for the Metaphysics of Morals 31

only formulas expressing the relation of objective laws of volition in gen-

eral to the subjective imperfection of the will of this or that rational being,

e.g., to the human being.

Now all imperatives command either hypothetically or categorically. The former represent the practical necessity of a possible action as a means

to attain something else which one wills (or which it is possible that one

might will). The categorical imperative would be that one which repre-

sented an action as objectively necessary for itself, without any reference to

another end.

Because every practical law represents a possible action as good, and

therefore as necessary for a subject practically determinable by reason, all

imperatives are formulas of the determination of action, which is necessary

in accordance with the principle of a will which is good in some way.22 Now

if the action were good merely as a means to something else, then the imperative is hypothetical; if it is represented as good in itself, hence neces- sary, as the principle of the will, in a will that in itself accords with reason,

then it is categorical. The imperative thus says which action possible through me would be

good, and represents the practical rule in relation to a will23 that does not

directly do an action because it is good, in part because the subject does not

always know that it is good, in part because if it did know this, its maxims

could still be contrary to the objective principles of a practical reason.

The hypothetical imperative thus says only that the action is good for

some possible or actual aim. In the first case it is a problematically,24 in the [Ak 4:415] second an assertorically practical principle. The categorical imperative, which declares the action for itself as objectively necessary without refer-

ence to any aim, i.e., also without any other end, is valid as an apodictically practical principle.

One can think of that which is possible only through the powers of some

rational being also as a possible aim of any will, and hence the principles of

the action, insofar as it is represented as necessary in order to achieve any

aim to be effected through it, are infinitely many. All sciences have some

22. 1785: ‘‘for some aim’’

23. 1785: ‘‘the will’’

24. In his (unpublished) First Introduction to the Critique of the Power of Judgment

(Ak 20:200 note), Kant retracts the term ‘problematical’ for this kind of imperative,

replacing it with the term ‘technical’, which he also uses already in the Groundwork (Ak

4:416).

32 Groundwork for the Metaphysics of Morals

practical part, consisting of the problems whether25 any end is possible for us

and of imperatives about how it can be attained. These can therefore in

general be called imperatives of skill. Whether the end is rational and good is not the question here, but only what one has to do in order to achieve them.

The precepts for the physician, how to make his patient healthy in a well-

grounded way, and for the poisoner, how to kill him with certainty,26 are to

this extent of equal worth, since each serves to effect its aim perfectly. Be-

cause in early youth one does not know what ends he will run up against in

life, parents seek chiefly to have their children learn many things, and they concern themselves about skill in the use of means toward all kinds of discre- tionary ends, about none of which they can determine whether it will per- haps actually become an aim of his pupil in the future, but about any of

which, however, it is possible that he might someday have it, and this con- cern is so great that they commonly neglect to educate and correct their

judgment over the worth of the things that they may perhaps make their ends.

There is one end, however, that one can presuppose as actual for all rational beings (insofar as imperatives apply to them, namely as dependent

beings) and thus one aim that they not merely can have, but of which one can safely presuppose that without exception27 they do have it in accor- dance with a natural necessity, and that is the aim at happiness. The hypo- thetical imperative that represents the practical necessity of the action as a

means to furthering happiness is assertoric. One may expound it as neces- sary not merely to an uncertain, merely possible aim, but to an aim that one

can presuppose safely and a priori 28 with every human being, because it[Ak 4:416] belongs to his essence.29 Now one can call skill in the choice of means to his

own greatest well-being prudence* in the narrowest sense. Thus the imper-

*The word ‘prudence’ is taken in a twofold sense; in the first it can bear

the name of ‘worldly prudence’ and in the second that of ‘private prudence.’

The first is the skill of a human being to have influence on others, in order to

use them for his aims. The second is the insight to unite all these aims to his

own enduring advantage. The latter is really that to which the worth of the

first is reduced, and about someone who is prudent in the first way but not in

the second way one can better say that he is clever and sly, but on the whole

imprudent.

25. Aufgaben, daß, a construction somewhat opaque in meaning and almost as awk-

ward in German as ‘‘problems that’’ would be in English.

26. sicher, which could also be translated ‘‘safely’’

27. insgesamt

28. ‘‘and a priori ’’ added in 1786

29. 1785: ‘‘to his nature’’

Groundwork for the Metaphysics of Morals 33

ative that refers to the choice of means to one’s own happiness, i.e., the

precept of prudence, is always hypothetical; the action is commanded not absolutely but only as a means to another aim.

Finally, there is one imperative that, without being grounded on any

other aim to be achieved through a certain course of conduct as its condi-

tion, commands this conduct immediately. This imperative is categorical. It has to do not with the matter of the action and what is to result from it, but

with the form and the principle from which it results; and what is essentially

good about it consists in the disposition, whatever the result may be. This

imperative may be called that of morality. The volition in accordance with these three kinds of principles is also

clearly distinguished by a difference30 in the necessitation of the will. Now in order to make this noticeable too, I believe that the most suitable termi-

nology to use in ordering them is to say that they are either rules of skill, or counsels of prudence or commands (laws) of morality. For only law carries with it the concept of an unconditional and objective, hence universally valid necessity, and commands are laws that must be obeyed, i.e., followed even against inclination. The giving of counsel contains necessity, to be sure, but can be valid merely under a subjective, pleasing31 condition,

whether this or that human being counts this or that toward his happiness;

the categorical imperative, by contrast, is not limited by any condition, and

as absolutely, though practically necessary, can be called quite authentically

a command. One could also call the first imperative technical (belonging to art), the second pragmatic* (to welfare), the third moral (belonging to free [Ak 4:417] conduct in general, i.e., to morals).

Now the question arises: How are all these imperatives possible? This

*It seems to me that the authentic signification of the word ‘pragmatic’

could be determined most precisely in this way. For those sanctions are called ‘pragmatic’ which really flow not from the rights of states, as neces-

sary laws, but from provision for the general welfare. A history is written ‘pragmatically’ when it makes us prudent, i.e., teaches how the world could take care of its advantage better than, or at any rate at least as well as, the

world of antiquity has done.

30. Ungleichheit, which might also be translated ‘‘inequality.’’ Kant may be suggest-

ing, that is, not only that the three imperatives are different in kind, but also that the three

kinds of necessitation have unequal rational weight: moral necessitation is unconditional,

hence prior to the other two, overriding them in cases of conflict; pragmatic necessitation

by imperatives of prudence, in turn, overrides technical necessitation by imperatives of

skill that merely tell us how to achieve some optional end we have contingently chosen.

31. gefälliger; editors often correct this to zufälliger, ‘‘contingent.’’

34 Groundwork for the Metaphysics of Morals

question does not demand the knowledge how to think the execution of the

action that the imperative commands, but rather merely how to think the

necessitation of the will that the imperative expresses in the problem. How

an imperative of skill is to be possible probably needs no particular discus-

sion. Whoever wills the end, also wills (insofar as reason has decisive

influence on his actions) the means that are indispensably necessary to it

that are in his control. As far as volition is concerned, this proposition is

analytic; for in the volition of an object, as my effect, is already thought my

causality as an acting cause, i.e., the use of means; and the imperative

extracts the concept of actions necessary for this end out of the concept of a

volition of this end (to be sure, synthetic propositions belong to determining

the means themselves to a proposed aim, but they have nothing to do with

the ground, with making the act32 of the will actual, but rather with how to

make the object actual). That in order to divide a line into two equal parts in

accordance with a secure principle I must draw two arcs from its endpoints

—this mathematics obviously teaches only through synthetic propositions;

but that if I know that the specified effect can occur only through such an

action, then if I completely will the effect, I would also will the action that is

required for it—that is an analytic proposition; for to represent something

as an effect possible through me in a certain way and to represent myself, in

regard to it, acting in this same way—those are entirely the same.

Imperatives of prudence would be equally analytic, and entirely coin-

cide with those of skill, if only it were so easy to provide a determinate

concept of happiness. For here, as there, it would be said: whoever wills the

end, also wills (necessarily in accord with reason) the sole means to it in[Ak 4:418]

his control. Yet it is a misfortune that the concept of happiness is such an

indeterminate concept that although every human being wishes to attain it,

he can never say, determinately and in a way that is harmonious with

himself, what he really wishes and wills. The cause of this is that all the

elements that belong to the concept of happiness are altogether empirical,

i.e., have to be gotten from experience, while for the idea of happiness an

absolute whole, a maximum of welfare, is required, in my present and in

every future condition. Now it is impossible for the most insightful, and at

the same time most resourceful, yet finite being to make a determinate

concept of what he really wills here. If he wills wealth, how much worry,

envy, and harassment33 will he not bring down on his shoulders?34 If he

32. Aktus

33. Nachstellung

34. Kant ends this sentence, which seems halfway between an assertion and a rhetori-

cal question, with a period instead of a question mark.

Groundwork for the Metaphysics of Morals 35

wills much cognition and insight, perhaps that could only give him a more

acute eye, to show him all the more terribly those ills that are now hidden

from him and yet cannot be avoided, or to burden his desires, which already

give him quite enough to do, with still more needs. If he wills a long life,

who will guarantee him that it would not be a long misery? If he wills at

least health, how often have bodily discomforts not deterred him from

excesses into which unlimited health would have allowed him to fall, etc.?

In short, he is not capable of determining with complete certainty, in accor-

dance with any principle, what will make him truly happy, because omni-

science would be required for that. Thus one cannot act in accordance with

determinate principles in order to be happy, but only in accordance with

empirical counsels, e.g., of diet, frugality, politeness, restraint, etc., of

which experience teaches that they most promote welfare on the average. It

follows from this that the imperatives of prudence, to speak precisely, can-

not command at all, i.e., cannot exhibit actions objectively as practically

necessary; that they are sooner to be taken as advisings (consilia) than as commands (praecepta) of reason; that the problem of determining, cer- tainly35 and universally, what action will promote the happiness of a rational

being, is fully insoluble, hence no imperative in regard to it is possible,

which would command us, in the strict sense, to do what would make us

happy, because happiness is an ideal not of reason but of imagination,

resting merely on empirical grounds, of which it would be futile to expect [Ak 4:419]

that they should determine an action through which to attain the totality of a

series of consequences which are in fact infinite. This imperative of pru-

dence, meanwhile, would be an analytically practical proposition if one

assumes that the means to happiness could be specified with certainty;36 for

it is distinguished from the imperative of skill only in this, that with the

latter the end is merely possible, but with the former it is given: since,

however, both merely command the means to that which it is presupposed

that one wills as an end, then the imperative that commands the volition of

the means for him who wills the end is in both cases analytic. Thus there is

also no difficulty in regard to the possibility of such an imperative.

By contrast, how the imperative of morality is possible is without doubt the sole question in need of a solution, since it is not at all hypothetical, and

thus the necessity, represented as objective, cannot be based on any presup-

position, as with the hypothetical imperatives. Yet in this connection it must

not be left out of account that whether there is any such imperative any-

where cannot be settled by any example, hence not empirically; but the

35. sicher

36. sicher

36 Groundwork for the Metaphysics of Morals

worry is rather that all those that seem categorical might be, in some hidden

wise, hypothetical. E.g., if it is said: ‘‘You ought not to make a deceiving

promise,’’ and one assumes that the necessity of this omission is not mere

advice for the avoidance of some ill or other, so that it might really mean:

‘‘You should not make a lying promise, so that if it were revealed then you

would lose your credit’’; if an action of this kind37 must be considered as

evil for itself, then the imperative forbidding it would be categorical; then

one still cannot with certainty give an example in which the will is deter-

mined merely by the law, without any other incentive, although it might38

appear so; for it is always possible that fear of disgrace, or perhaps also an

obscure worry about other dangers, might secretly have had an influence on

the will. Who39 can prove through experience the nonexistence of a cause,

since experience teaches us nothing beyond the fact that we do not perceive

one? But in such a case the so-called moral imperative, which appears as

such to be categorical and unconditioned, would in fact be only a pragmatic

precept, which alerts us to our own advantage and merely teaches us to pay

attention to it.

Thus we will have to investigate the possibility of a categorical impera- tive entirely a priori, since here we cannot have the advantage that its[Ak 4:420] reality is given in experience, so that its possibility would be necessary not

for its establishment but only for its explanation.40 Meanwhile, we can

provisionally41 have insight into this much: that the categorical imperative

alone can be stated as a practical law, while the others collectively are, to be sure, principles of the will, but cannot be called ‘laws’; for what it is necessary to do for the attainment of a discretionary aim can be considered

in itself to be contingent, and we can always be rid of the precept if we give

up the aim; whereas the unconditioned command leaves the will no free

discretion in regard to the opposite, hence it alone carries with it that

necessity which we demand for a law.

Secondly, with this categorical imperative, or law of morality, the

ground of difficulty (of having insight into its possibility) is very great

37. 1785: ‘‘but rather if one asserts that an action of this kind’’

38. 1785: ‘‘even if it might appear so’’

39. 1785: ‘‘For who’’

40. Erklärung, which could also be translated ‘‘definition.’’ Kant holds that for a well-

formed (real) definition of a thing, we require a demonstration of its (real) possibility. See

Critique of Pure Reason A727–30/B755–59.

41. 1785: ‘‘But we can provisionally’’

Groundwork for the Metaphysics of Morals 37

indeed. It is a synthetically practical proposition* a priori, and since there is so much difficulty in gaining insight into the possibility of propositions of

this kind in theoretical cognition, it is easy to gather that there will be no

less in the practical.

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 for-

mula, containing the proposition which alone can be a categorical impera-

tive; for how such an absolute command is possible, even if we know how it

is stated, will still demand particular and difficult effort, which, however,

we will postpone until the last section.

If I think of a hypothetical imperative in general, then I do not know beforehand what it will contain until the condition is given to me. But if I

think of a categorical imperative, then I know directly what it contains. For since besides the law, the imperative contains only the necessity of the

maxim,** that it should accord with this law, but the law contains no [Ak 4:421]

condition to which it is limited, there remains nothing left over with which

the maxim of the action is to be in accord, and this accordance alone is what

the imperative really represents necessarily.

The categorical imperative is thus only a single one, and specifically

this: Act only in accordance with that maxim through which you can at the same time will that it become a universal law. Now if from this one imperative all imperatives of duty can be derived

as from their principle, then although we leave unsettled whether in general

*I connect the deed a priori with the will, without a presupposed condi- [Ak4:420] tion from any inclination, hence necessarily (though only objectively, i.e.,

under the idea of reason, which would have full control over all subjective

motivations). This is therefore a practical proposition that does not derive

the volition of an action analytically from any other volition already presup-

posed (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.

**A maxim is the subjective principle for action, and must be distin- guished from the objective principle, namely the practical law. The former [Ak4:421] contains the practical rule that reason determines in accord with the condi-

tions of the subject (often its ignorance or also its inclinations), and is thus

the principle in accordance with which the subject acts; but the law is the objective principle, valid for every rational being, and the principle in

accordance with which it ought to act, i.e., an imperative.

38 Groundwork for the Metaphysics of Morals

what one calls ‘duty’ is an empty concept, we can at least indicate what we

are thinking in the concept of duty and what this concept means.42

Because the universality of the law in accordance with which effects

happen constitutes that which is really called nature in the most general sense (in accordance with its form), i.e., the existence of things insofar as it

is determined in accordance with universal laws, thus the universal impera-

tive of duty can also be stated as follows: So act as if the maxim of your action were to become through your will a universal law of nature. Now we will enumerate43 some duties, in accordance with their usual

division into duties toward ourselves and toward other human beings, and

into perfect and imperfect duties:*

(1) One person, through a series of evils that have accumulated to the

point of hopelessness, feels weary of life but is still so far in possession of[Ak 4:422]

his reason that he can ask himself whether it might be contrary to the duty to

himself to take his own life. Now he tries out whether the maxim of his

action could become a universal law of nature. But his maxim is: ‘From

self-love, I make it my principle to shorten my life when by longer term it

threatens more ill than it promises agreeableness’. The question is whether

this principle of self-love could become a universal law of nature. But then

one soon sees that a nature whose law it was to destroy life through the same

feeling44 whose vocation it is to impel the furtherance of life would contra-

*Here one must note well that I reserve the division of duties entirely for[Ak4:421]

a future metaphysics of morals; the division here therefore stands only as a discretionary one (to order my examples). For the rest, I understand by a

perfect duty that which permits no exception to the advantage of inclina-

tion, and I do have perfect duties that are not merely external but also internal, which runs contrary to the use of words common in the schools;

but I do not mean to defend that here, because for my aim it is all the same

whether or not one concedes it to me. [Cf. Metaphysics of Morals, Ak 6:240, 391–98, 413, and the detailed taxonomy of duties of virtue, Ak

6:417–68. The ‘‘use of words common in the schools,’’ according to which

perfect duties are externally enforceable actions, is based on Samuel Pufen-

dorf (1632–1694), De Jure Naturale (1672), 1.1.19–20. But Pufendorf’s distinction was anticipated by Hugo Grotius (1583–1645) and had been

taken up also by, among others, Christian Thomasius (1655–1728) and J. G.

Sulzer.]

42. sagen wolle

43. herzählen, which could also be translated ‘‘reckon’’ or ‘‘calculate’’

44. Empfindung

Groundwork for the Metaphysics of Morals 39

dict itself, and thus could not subsist as nature; hence that maxim could not

possibly obtain as a universal law of nature, and consequently it entirely

contradicts the supreme principle of all duty.

(2) Another sees himself pressured by distress into borrowing money.

He knows very well that he will not be able to pay, but he also sees that

nothing will be lent him if he does not firmly promise to pay at a determi-

nate time. He wants to make such a promise; yet he has conscience enough

to ask himself: ‘‘Is it not impermissible and contrary to duty to get out of

distress in such a way?’’ Supposing he nevertheless resolved on it, his

maxim would be stated as follows: ‘If I believe myself to be in pecuniary

distress, then I will borrow money and promise to pay it back, although I

know this will never happen’. Now this principle of self-love, or of what is

expedient for oneself, might perhaps be united with my entire future wel-

fare, yet the question now is: ‘‘Is it right?’’ I thus transform this claim45 of

self-love into a universal law and set up the question thus: ‘‘How would it

stand if my maxim became a universal law?’’ Yet I see right away that it

could never be valid as a universal law of nature and still agree with itself,

but rather it would necessarily contradict itself. For the universality of a law

that everyone who believes himself to be in distress could promise what-

ever occurred to him with the intention of not keeping it would make

impossible the promise and the end one might have in making it, since no

one would believe that anything has been promised him, but rather would

laugh about every such utterance as vain pretense.

(3) A third finds in himself a talent, which could, by means of some [Ak 4:423]

cultivation, make him into a human being who is useful for all sorts of aims.

But he sees himself as in comfortable circumstances and sooner prefers

to indulge46 in gratification than to trouble himself with the expansion

and improvement of his fortunate natural predispositions. Yet he still asks

whether, apart from the agreement of his maxim of neglecting his gifts of

nature with his propensity to amusement, it also agrees with what one calls

‘duty’. Then he sees that, although a nature could still subsist in accordance

with such a universal law, though then the human being (like the South Sea

Islanders) would think only of letting his talents rust and applying his life

merely to idleness, amusement, procreation, in a word, to enjoyment; yet it

is impossible for him to will that this should become a universal law of nature, or that it should be implanted in us as such by natural instinct. For as

a rational being he necessarily wills that all the faculties in him should be

45. Zumutung

46. 1785: ‘‘and he prefers it that he indulge’’

40 Groundwork for the Metaphysics of Morals

developed, because they are serviceable and given47 to him for all kinds of

possible aims.

(4)48 Yet a fourth—for whom it is going well, while he sees that others have to struggle with great hardships (with which he could well help them)

—thinks: ‘‘What has it to do with me? Let each be as happy as heaven wills,

or as he can make himself, I will not take anything from him or even envy

him; only I do not want to contribute to his welfare or to his assistance in

distress!’’ Now to be sure, if such a way of thinking were to become a

universal law of nature, then the human race could well subsist, and without

doubt still better than when everyone chatters about sympathetic participa-

tion49 and benevolence, and even on occasion exerts himself to practice

them, but, on the contrary also deceives wherever he can,50 sells out, or

otherwise infringes on the right of human beings. But although it is possible

that a universal law of nature could well subsist in accordance with that

maxim, yet it is impossible to will that such a principle should be valid without exception51 as a natural law. For a will that resolved on this would

conflict with itself, since the case could sometimes arise in which he needs

the love and sympathetic participation of others, and where, through such a

natural law arising from his own will, he would rob himself of all the hope

of assistance that he wishes for himself.

Now these are some of the many actual duties, or at least of what we take

to be duties, whose partitioning52 from the single principle just adduced[Ak 4:424]

47. ‘‘and given’’ added in 1786

48. Kant’s text, although it emphasizes the word ‘‘fourth,’’ omits the (4) required by

the parallel with his three other examples.

49. Teilnehmung

50. 1785: ‘‘wherever one can’’

51. allenthalben

52. Abteilung; some editors correct this to Ableitung, ‘‘derivation.’’ In favor of the

emendation is that if Kant meant ‘classification,’ one would expect him to use Einteilung

(‘‘division,’’ as he did above, Ak 4:421); Abteilung refers more properly to one of the

parts or subcategories marked out by a division or classification than it does to the act of

dividing or classifying or to the entire system of classification; where it does refer to an

act of dividing, abteilen means the partitioning off of one space from another, and not the

creation of a system of classification. The construction Abteilung aus dem einigen Prinzip

is also awkward, in the same way that this English translation of it is; and no such

construction is found anywhere else in Kant’s writings. Further, Kant did speak earlier of

being able to ‘‘derive’’ (ableiten) all imperatives of duty from a single categorical impera-

tive (Ak 4:421). But despite all these reasons, the emendation to Ableitung (‘‘derivation’’)

Groundwork for the Metaphysics of Morals 41

clearly meets the eye. One must be able to will that a maxim of our action should become a universal law: this is the canon of the moral judgment of

this action in general. Some actions are so constituted that their maxim

cannot even be thought without contradiction as a universal law of nature, much less could one will that it ought to become one. With others, that internal impossibility is not to be encountered, but it is impossible to will

remains doubtful. Against it is the following: Kant goes on in the present paragraph to

discuss the relation of his principle only to the classification of duties, not to their

derivation. Further, it is not at all clear that when Kant spoke of deriving duties from a

single categorical imperative, he meant to restrict the formula of that imperative to the

two formulations that have been presented so far. He may well have meant that a deriva-

tion of duties would require the entire system of formulas, first introduced later at Ak

4:436. His practice in the Metaphysics of Morals strongly suggests the latter position.

There Kant does propose to derive an entire system of ethical duties; but only the duty of

beneficence (which pertains only to the fourth example here) is related to anything in the

present formula of the moral law (Ak 6:453). This is possible only because the maxim of

pursuing one’s own happiness (and the consequent volition of others’ voluntary assis-

tance, as required to achieve this end) can be ascribed to all rational beings, so that the

principle of morality can require them to adopt it in a universalizable form. Apart from

this unique case, universalizability enables us only to disqualify certain specific maxims,

and cannot yield anything like a positive duty (e.g., to refrain from suicide, keep prom-

ises, or develop talents). All fifteen of the other ethical duties explicitly enumerated there

(including three of the four that are exemplified here) are derived by appeal to the second

formula, that of humanity as end in itself (first stated in the Groundwork at Ak 4:429).

Kant says that suicide is a ‘‘debasing of humanity in one’s person’’ (Ak 6:422–423); the

duty to develop one’s natural perfection is ‘‘bound up with the end of humanity in our own

person’’ (Ak 6:391–92; cf. 6:444–46). In the Metaphysics of Morals, Kant treats promis-

ing under the heading of externally enforceable right rather than of ethics (as his use here

of the term ‘right’ might also imply). There Kant even denies that it is either possible or

necessary to demonstrate that promises ought to be kept (Ak 6:273). The principle of right

is distinct from the supreme principle of morality (Ak 6:230); but the fundamental right

(the innate human right to freedom) is said ‘‘to belong to every human being by virtue of

his humanity’’ (Ak 6:237). But he does discuss the ethical duty not to lie under the

heading of strict duties to oneself, where it is said to be a violation of ‘‘the humanity in his

own person’’ because it uses his capacity to communicate as a mere means (Ak 6:429).

Kant’s definitive presentation of the duties enumerated here thus has far less affinity with

the present discussion of them (based on the formula of the law of nature) than it does

with his discussion of them below in connection with the second formula, that of human-

ity as end in itself (Ak 4:429–30).

42 Groundwork for the Metaphysics of Morals

that their maxims should be elevated to the universality of a natural law,

because such a will would contradict itself. One easily sees that the first

conflict with strict or narrow (unremitting) duty, the second only with wide

(meritorious) duty, and thus all duties regarding the kind of obligation (not

the object of their action) have been completely set forth53 through these

examples in their dependence on the one principle.

Now if we attend to ourselves in every transgression of a duty, then we

find that we do not actually will that our maxim should become a universal

law, for that is impossible for us, but rather will that its opposite should

remain a law generally; yet we take the liberty of making an exception for ourselves, or (even only for this once) for the advantage of our inclination.

Consequently, if we weighed everything from one and the same point of

view, namely that of reason, then we would encounter a contradiction in our

own will, namely that objectively a certain principle should be necessary as

a universal law and yet subjectively that it should not be universally valid,

but rather that it should admit of exceptions. But since we consider our

action at one time from a point of view that accords entirely with reason,

and then, however, also the same action from the point of view of a will

affected by inclination, there is actually no contradiction here, but only

a resistance of inclination against the precept of reason (antagonismus), through which the universality of the principle (universalitas) is trans- formed into a mere general validity (generalitas), so that the practical principle of reason is supposed to meet the maxim halfway. Now although

this cannot be justified in our own impartially rendered judgment, it proves

that we actually recognize the validity of the categorical imperative and

(with every respect for it) allow ourselves only a few exceptions, which are,

as it seems to us, insignificant and forced upon us.[Ak 4:425]

Thus we have established at least this much: that if duty is a concept that

is to contain significance and actual legislation for our actions, then this

duty could be expressed only in categorical imperatives, but by no means in

hypothetical ones; likewise, which is already quite a bit, we have exhibited

distinctly and for every use the content of the categorical imperative which

would have to contain the principle of all duty (if there is such a thing at all).

But we are still not ready to prove a priori that there actually is such an imperative, that there is a practical law which commands for itself abso-

lutely and without any incentives, and that it is a duty to follow this law.

With the aim of attaining that, it is of the utmost importance to let this

serve as a warning that one must not let it enter his mind to try to derive the

53. 1785: ‘‘are completely set forth’’

Groundwork for the Metaphysics of Morals 43

reality of this principle from the particular quality of human nature. For duty ought to be the practically unconditioned necessity of action; thus it

must be valid for all rational beings (for only to them can an imperative

apply at all), and must only for this reason be a law for every human will. That which, by contrast, is derived only from what is proper to the particu-

lar natural predisposition of humanity, or from certain feelings and propen-

sities, or indeed, if possible, from a particular direction of human reason,

and would not have to be valid necessarily for the will of every rational

being—that can, to be sure, be a maxim for us, but cannot yield any law; it

can yield a subjective principle, in accordance with which we may have a

propensity and inclination, but not an objective one, in accordance with

which we would be assigned to act, even if it were to go directly contrary to all our propensities, inclinations, and natural adaptations; it even proves all

the more the sublimity and inner dignity of the command in a duty, the less

subjective causes are for it and the more they are against it, without on this

account the least weakening the necessitation through the law or taking

anything away from its validity.

Now here we see philosophy placed in fact at a perilous standpoint,

which is to be made firm, regardless of anything either in heaven or on earth

from which it may depend or by which it may be supported. Here it should

prove its purity54 as self-sustainer of its own laws, not as a herald of those

that an implanted sense or who knows what tutelary nature whispers to it,

which, taken collectively, although they may be better than nothing at all, [Ak 4:426]

yet they can never yield the principles that reason dictates and that must

have their source fully a priori and therewith at the same time their com- manding authority: expecting nothing of the inclination of the human be-

ing, but everything from the supremacy of the law and the respect owed to

it; or else, if that fails, condemning the human being to self-contempt and

inner abhorrence.

Thus everything that is empirical is, as a contribution toward the princi-

ple of morality, not only entirely unfit for it, but even highly disadvan-

tageous to the purity55 of morals themselves, in which precisely consists the

sublime56 worth of a will absolutely good in itself and elevated above all

price,57 that the principle of the actions is free of all influences of contingent

grounds that only experience can provide. One cannot be given too many or

54. Lauterkeit

55. Lauterkeit

56. Cf. below, Ak 4:439–40.

57. Cf. below, Ak 4:434.

44 Groundwork for the Metaphysics of Morals

too frequent warnings against this negligent or even base way of thinking,

which seeks out the principle among empirical motivations and laws, since

human reason in its weariness gladly reposes on this pillow and, in the

dream of sweet illusions58 (which lets it embrace a cloud instead of Juno),59

supplants the place of morality with a bastard patched together from limbs

of quite diverse ancestry, which looks similar to whatever anyone wants to

see, but not to virtue, for him who has once beheld it in its true shape.*

The question is therefore this: Is it a necessary law for all rational beings to judge their actions always in accordance with those maxims of which

they themselves can will that they should serve as universal laws? If it is,

then it must be bound up (fully a priori) with the concept of the will of a rational being in general. But in order to discover this connection, one must,

however much one may resist it, take one step beyond, namely to meta-

physics, though into a domain of metaphysics that is distinguished from

that of speculative philosophy, namely into the metaphysics of morals. In a[Ak 4:427]

practical philosophy, where what are to be established are not grounds for

what happens, but laws for what ought to happen, even if it never does happen, i.e., objectively practical laws, there we do not find it necessary to

institute an investigation into the grounds why something pleases or dis-

pleases, how the gratification of mere sensation is to be distinguished from

taste, and whether the latter is distinct from a universal satisfaction of

*To behold virtue in its authentic shape is nothing other than to exhibit[Ak4:426]

morality denuded of all admixture of the sensible and all ungenuine adorn-

ment of reward or self-love. How completely it eclipses everything else that

appears charming to inclinations, everyone can easily be aware of by means

of the least attempt of his reason, if it is not entirely corrupted for abstraction.

58. Vorspiegelungen

59. In Greek mythology, Ixion (a legendary king of Thessaly) schemed to win the love

of Hera, queen of the gods (Latin name: Juno). Her husband, Zeus, discovered his inten-

tion and formed a cloud, Nephelē, that resembled Hera. By the cloud Ixion conceived

Centaurus (for which the scholiast gives the false etymology ‘‘what penetrates the air’’).

Centaurus was the ancestor of the centaurs, a race of beings half human and half equine

(perhaps Kant’s ‘‘bastard patched together from limbs of quite diverse ancestry’’ is a

reference to them). Zeus punished Ixion for his presumptuousness by having him bound

on a wheel in Hades that turns forever. The myth is told by Pindar, Pythian Ode 2.21–50.

Since Kant’s knowledge of Latin poetry was better, he is more likely to have known the

Ixion story from Ovid (Metamorphoses 4.461, 9.124, 10.42, 12.503–5) or Virgil (Geor-

gics 3.38, 4.484; Aeneid 6.601), although these later versions emphasize Ixion’s under-

world punishment rather than the story of Juno and the cloud.

Groundwork for the Metaphysics of Morals 45

reason; on what the feelings of pleasure and displeasure rest, and how from

them arise desires and inclinations, and from these, again, through the

cooperation of reason, maxims arise; for all that belongs to an empirical

doctrine of the soul, which constitutes the second part of the doctrine of

nature, if one considers it as philosophy of nature insofar as it is grounded on empirical laws. Here, however, we are talking about objectively practi- cal laws, hence about the relation of a will to itself insofar as it determines

itself merely through reason, such that everything that has reference to the

empirical falls away of itself; because if reason for itself alone determines conduct (the possibility of which we will investigate right now), it must

necessarily do this a priori. The will is thought as a faculty of determining itself to action in accord

with the representation of certain laws. And such a faculty can be there to be encountered only in rational beings. Now that which serves the will as

the objective ground of its self-determination is the end, and this, if it is given through mere reason, must be equally valid for all rational beings. By

contrast, what contains merely the ground of the possibility of the action

whose effect is the end is called the means. The subjective ground of desire is the incentive, the objective ground of volition is the motive; hence the distinction between subjective ends, which rest on incentives, and objective

ones, which depend on motives that are valid for every rational being.

Practical principles are formal when they abstract from all subjective ends; but they are material when they are grounded on these, hence on certain incentives. 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 faculty of desire of the subject gives them their worth,

which therefore can provide no necessary principles valid universally for all [Ak 4:428]

rational beings and hence valid for every volition, i.e., practical laws. Hence

all these relative ends are only the ground of hypothetical imperatives.

But suppose there were something whose existence in itself had an abso- lute worth, something that, as end in itself, could be a ground of determinate laws; then in it and only in it alone would lie the ground of a possible

categorical imperative, i.e., of a practical law.

Now I say that the human being, and in general every rational being,

exists as end in itself, not merely as means to the discretionary use of this or that will, but in all its actions, those directed toward itself as well as those

directed toward other rational beings, it must always at the same time be considered as an end. All objects of inclinations have only a conditioned worth; for if the inclinations and the needs grounded on them did not exist,

then their object would be without worth. The inclinations themselves,

46 Groundwork for the Metaphysics of Morals

however, as sources of needs, are so little of absolute worth, to be wished for

in themselves, that rather to be entirely free of them must be the universal

wish of every rational being.60 Thus the worth of all objects to be acquired through our action is always conditioned. The beings whose existence rests

not on our will but on nature nevertheless have, if they are beings without

reason, only a relative worth as means, and are called things; rational beings, by contrast, are called persons, because their nature already marks them out as ends in themselves, i.e., as something that may not be used merely as

means, hence to that extent limits all arbitrary choice61 (and is an object of

respect). These are not merely subjective ends whose existence as effect of

our action has a worth for us; but rather objective ends, i.e., things whose existence in itself is an end, and specifically an end such that no other end can

be set in place of it, to which it should do service merely as means, because without this nothing at all of absolute worth would be encountered any- where; but if all worth were conditioned, hence contingent, then for reason

no supreme practical principle could anywhere be encountered.

If, then, there is supposed to be a supreme practical principle, and in

regard to the human will a categorical imperative, then it must be such from

the representation of that which, being necessarily an end for everyone,

because it is an end in itself, constitutes an objective principle of the will,[Ak 4:429] hence can serve as a universal practical law. The ground of this principle is:

Rational nature exists as end in itself. The human being necessarily repre- sents his own existence in this way;62 thus to that extent it is a subjective principle of human actions. But every other rational being also represents

his existence in this way as consequent on the same rational ground as is

valid for me;* thus it is at the same time an objective principle, from which, as a supreme practical ground, all laws of the will must be able to be

derived. The practical imperative will thus be the following: Act so that you

*This proposition I here set forth as a postulate. In the last section one

will find the grounds for it.

60. ‘‘Considered in themselves, natural inclinations are good, i.e., not reprehensible,

and to want to extirpate them would be not only futile, but harmful and blameworthy as

well; we must rather only curb them, so that they will not wear each other out but will

instead be harmonized into a whole called ‘happiness’ ’’ (Religion within the Boundaries

of Mere Reason, Ak 6:58).

61. Willkür

62. See Conjectural Beginning of Human History, Ak 8:114; Anthropology in a

Pragmatic Respect, Ak 7:127, 130.

Groundwork for the Metaphysics of Morals 47

use humanity,63 as much in your own person as in the person of every other, always at the same time as end and never merely as means. We will see whether this can be accomplished.

In order to remain with the previous examples,

First, in accordance with the concept of the necessary duty toward one- self, the one who has suicide in mind will ask himself whether his action

could subsist together with the idea of humanity as an end in itself. If he destroys himself in order to flee from a burdensome condition, then he

makes use of a person merely as a means, for the preservation of a bearable condition up to the end of life. The human being, however, is not a thing,

hence not something that can be used merely as a means, but must in all his actions always be considered as an end in itself. Thus I cannot dispose of

the human being in my own person, so as to maim, corrupt, or kill him.64

(The nearer determination of this principle, so as to avoid all misunder-

standing, e.g., the amputation of limbs in order to preserve myself, or the

risk at which I put my life in order to preserve my life, etc., I must here pass

over; they belong to morals proper.)65

Second, as to the necessary or owed duty toward others, the one who has it in mind to make a lying promise to another will see66 right away that he

wills to make use of another human being merely as means, without the end also being contained in this other. For the one I want to use for my aims

through such a promise cannot possibly be in harmony with my way of

63. Menschlichkeit; this term refers to one of our three fundamental predispositions:

(1) animality (through which we have instincts for survival, procreation, and sociability);

(2) humanity, through which we have the rational capacities to set ends, use means to

them, and organize them into a whole (happiness); and (3) personality, through which we

have the capacity to give ourselves moral laws and are accountable for following them

(see Religion within the Boundaries of Mere Reason, Ak 6:26–28; Anthropology in a

Pragmatic Respect, Ak 7:322–25). ‘Humanity’ thus means the same as ‘rational nature’,

and Kant’s use of it involves no retraction of the claim that moral commands must be

valid for all rational beings, not only for members of the human species.

64. In the Metaphysics of Morals, Kant discusses the duty not to maim oneself in

connection with the duty forbidding suicide (Ak 6:422–23). Verderben (‘‘corrupt’’) there-

fore probably carries with it the broad sense of ruining or destroying (sc. one’s body or

parts of it) rather than the narrower sense of moral corruption. Duties to oneself as a moral

being, which Kant classifies as duties against lying, avarice, false humility (or servility),

and duties as moral judge of oneself, are dealt with separately, 6:428–42.

65. zur eigentlichen Moral

66. einsehen

48 Groundwork for the Metaphysics of Morals

conducting myself toward him and thus contain in himself the end of this[Ak 4:430]

action.67 Even more distinctly does this conflict with the principle of other

human beings meet the eye if one approaches it through examples of attacks

on the freedom and property of others. For then it is clearly evident that the

one who transgresses the rights of human beings is disposed to make use of

the person of others merely as a means, without taking into consideration

that as rational beings, these persons ought always to be esteemed at the

same time as ends, i.e., only as beings who have to be able to contain in

themselves the end of precisely the same action.*

Third, in regard to the contingent (meritorious) duty toward oneself, it is not enough that the action does not conflict with humanity in our person as

end in itself; it must also harmonize with it. Now in humanity there are predispositions to greater perfection, which belong to ends of nature in

regard to the humanity in our subject; to neglect these would at most be able

to subsist with the preservation of humanity as end in itself, but not with the furthering of this end.

Fourth, as to the meritorious duty toward others, the natural end that all human beings have is their own happiness. Now humanity would be able to

subsist if no one contributed to the happiness of others yet did not inten-

tionally remove anything from it; only this is only a negative and not a

positive agreement with humanity as end in itself, if everyone does not aspire, as much as he can, to further the ends of others. For regarding the

subject which is an end in itself: if that representation is to have its total effect on me, then its ends must as far as possible also be my ends. This principle of humanity and of every rational nature in general as end

*Let one not think that the trivial quod tibi non vis fieri, etc. [What you do not want to be done to yourself do not do to another] could serve here as

a standard or principle. For it is only derived from that principle, though

with various limitations; it cannot be a universal law, for it does not contain

the ground of duties toward oneself, nor that of the duties of love toward

others (for many would gladly acquiesce that others should not be benefi-

cent to him, if only he might be relieved from showing beneficence to

them), or finally of owed duties to one another, for the criminal would argue

on this ground against the judge who punishes him, etc. [Here Kant is

distinguishing his principle from the so-called Golden Rule of the Gospels:

‘‘Therefore, all things whatsoever ye would that men should do to you, do

ye even so to them’’ (Matthew 7:12; cf. Luke 6:31).]

67. It is essential to Kant’s conception of a promise that it involves a ‘‘united will’’ of

the promisor and the promisee (Metaphysics of Morals, Ak 6: 272).

Groundwork for the Metaphysics of Morals 49

in itself (which is the supreme limiting condition of the freedom of the [Ak 4:431] actions of every human being) is not gotten from experience, first, on

account of its universality, since it applies to all rational beings in general,

and no experience is sufficient to determine anything about that; second,

because in it humanity is represented not as an end of human beings68

(subjectively), i.e., as an object that one actually from oneself makes into an

end, but as an objective end which, whatever ends we may have, is to

constitute as a law the supreme limiting condition of all subjective ends,

hence must arise from pure reason. The ground of all practical legislation,

namely, lies objectively in the rule and the form of universality, which makes it capable of being a law (at least a law of nature) (in accordance with

the first principle), but subjectively it lies in the end; but the subject of all ends is every rational being as end in itself (in accordance with the second

principle): from this now follows the third practical principle of the will, as

the supreme condition of its harmony with universal practical reason, the

idea of the will of every rational being as a will giving universal law. All maxims are repudiated in accordance with this principle which can-

not subsist together with the will’s own universal legislation. The will is

thus not solely subject to the law, but is subject in such a way that69 it must

be regarded also as legislating to itself,70 and precisely for this reason as subject to the law (of which it can consider itself as the author).71

Imperatives represented in the above way, namely of the lawfulness of

actions generally similar to an order of nature, or of the universal prefer- ence of the end of rational beings themselves, just by being represented as categorical, excluded from their commanding authority all admixture of

any interest as an incentive; but they were only assumed as categorical, because one had to assume such a thing if one wanted to explain the concept

68. 1785: ‘‘of the human being’’

69. 1785: ‘‘not subject to the law except in such a way that’’

70. 1785: ‘‘as a self-legislating [being]’’

71. On the distinction between the ‘‘legislator’’ of a law (who promulgates and

attaches sanctions to it) and the ‘‘author’’ of a law (whose will actually imposes the

obligation), see Metaphysics of Morals, Ak 6:227. Although Kant frequently speaks here

of the rational being as ‘‘legislator’’ of the moral law, his position (more precisely ex-

pressed, in this terminology) is that only the rational being who is obligated can be the

author of the law; Kant allows that we can speak of God (or the ‘‘supreme head of the

realm of ends’’) as the legislator of the moral law (see below, 4:433–34; Moral Philoso-

phy Collins, Ak 27:282–83; and Religion within the Boundaries of Mere Reason, Ak

6:99–100.

50 Groundwork for the Metaphysics of Morals

of duty. But that there are practical propositions which command cate-

gorically cannot be proven for itself here, just as little as this can still

happen72 anywhere in this section; yet one thing could have happened,

namely that the withdrawal of all interest in the case of volition from duty,

in the imperative itself, through any determination that it could contain, is[Ak 4:432]

indicated as the specific sign distinguishing the categorical from the hypo-

thetical imperative, and this happens in the third formula of the principle,

namely the idea of the will of every rational being as a universally legisla- tive will. For if we think of such a will, then although a will that stands under laws

may be bound by means of an interest in this law, nevertheless it is impossi-

ble for a will that is itself supremely legislative to depend on any interest;

for such a dependent will would need yet another law, which limited the

interest of its self-love to the condition of a validity for the universal law.

Thus the principle of every human will as a will legislating universally through all its maxims,* if otherwise everything were correct about it, would be quite well suited for the categorical imperative by the fact that precisely for the sake of the idea of universal legislation, it grounds itself on no interest and hence it alone among all73 possible imperatives can be unconditioned; or still better, by converting the proposition, if there is a categorical imperative

(i.e., a law for every will of a rational being), then it can command only that

everything be done from the maxim of its will as a will that could at the same

time have as its object itself as universally legislative; for only then is the

practical principle and the imperative it obeys unconditioned, because it

cannot have any interest at all as its ground.

Now it is no wonder, when we look back on all the previous efforts that

have ever been undertaken to bring to light the principle of morality, why

they all had to fail. One saw the human being bound through his duty to

laws, but it did not occur to one that he was subject only to his own and yet universal legislation, and that he was obligated only to act in accord with his own will, which, however, in accordance with its natural end, is a

universally legislative will. For if one thought of him only as subject to a

law (whatever it might be), then this would have to bring with it some[Ak 4:433]

interest as a stimulus or coercion, because as a law it did not arise from his

*I can be exempted here from providing examples to elucidate this[Ak4:432]

principle, since those that first elucidated the categorical imperative and its

formula can all serve here for precisely that end.

72. 1785: ‘‘just as little as this still cannot happen’’

73. This word added in 1786

Groundwork for the Metaphysics of Morals 51

will, but rather this will was necessitated by something else to act in a certain way in conformity with the law. Through this entirely necessary

consequence, however, all the labor of finding a supreme ground of duty

was irretrievably lost. For from it one never got duty, but only necessity of

action from a certain interest. Now this might be one’s own interest or

someone else’s. But then the imperative always had to come out as condi-

tioned, and could never work at all as a moral command. Thus I will call

this principle the principle of74 the autonomy of the will, in contrast to every75 other, which on this account I count as heteronomy. The concept of every rational being that must consider itself as giving

universal law through all the maxims of its will in order to judge itself and

its actions from this point of view, leads to a very fruitful concept depending

on it, namely that of a realm of ends.76

By a realm, however, I understand the systematic combination of vari- ous rational beings through communal laws. Now because laws determine

ends in accordance with their universal validity, there comes to be, if one

abstracts from the personal differences between rational beings, as likewise

from every content of their private ends, a whole of all ends—(of rational

beings as ends in themselves, as well as of their own ends, which each may

set for himself ) in systematic connection, i.e., a realm of ends—can be

thought, which is possible in accordance with the above principles.

For rational beings all stand under the law that every one of them ought to77 treat itself and all others never merely as means, but always at the same time as end in itself. From this, however, arises a systematic combination of rational beings through communal objective laws, i.e., a realm that, because

these laws have as their aim the reference of these beings to one another78 as

ends and means, can be called a ‘realm of ends’ (obviously only an ideal).

But a rational being belongs as a member to the realm of ends if in this

74. 1785: ‘‘Thus I will call this the principle of’’

75. The editors suggest jenem, which would translate: ‘‘in contrast to that other,

which.’’

76. The obvious source for Kant’s conception of a ‘‘realm of ends’’ is Leibniz’s

conception of the ‘‘city of God’’ as the ‘‘realm of minds,’’ and the relationship of the

‘‘realm of nature’’ to this ‘‘realm of grace.’’ Gottfried Wilhelm Leibniz (1646–1716),

Discourse on Metaphysics (1686), § 36; Principles of Nature and Grace Based on Reason

(1714), § 15; Monadology (1714), §§ 85–90.

77. 1785: ‘‘may’’

78. 1785: ‘‘as their aim their relation to one another’’

52 Groundwork for the Metaphysics of Morals

realm it gives universal law but is also itself subject to these laws. It belongs

to it as supreme head, if as giving law it is subject to no will of another.79[Ak 4:434] The rational being must always consider itself as giving law in a realm

of ends possible through freedom of the will, whether as member or as

supreme head. It can assert the place of the latter, however, not merely

through the maxim of its will, but only when it is a fully independent being,

without need and without limitation of faculties that are adequate to that

will.

Morality thus consists in the reference of all action to that legislation

through which alone a realm of ends is possible. But the legislation must be

encountered in every rational being itself, and be able to arise from its will,

whose principle therefore is: ‘Do no action in accordance with any other

maxim, except one that could subsist with its being a universal law, and

hence only so that the will could through its maxim at the same time consider itself as universally legislative’. Now if the maxims are not through their nature already necessarily in harmony with this objective principle of the

rational beings, as universally legislative, then the necessity of the action in

accordance with that principle is called ‘practical necessitation’, i.e., duty. Duty does not apply to the supreme head in the realm of ends, but it does to

every member, and specifically, to all in equal measure.

The practical necessity of acting in accordance with this principle, i.e.,

duty, does not rest at all on feelings, impulses, or inclinations, but merely on

the relation of rational beings to one another, in which the will of one

rational being must always at the same time be considered as universally legislative, because otherwise the rational being could not think of the other rational beings as ends in themselves. Reason thus refers every maxim of the will as universally legislative to every other will and also to every action

toward itself, and this not for the sake of any other practical motive or future

advantage, but from the idea of the dignity of a rational being that obeys no law except that which at the same time it gives itself.

In the realm of ends everything has either a price or a dignity.80 What has a price is such that something else can also be put in its place as its

equivalent; by contrast, that which is elevated above all price, and admits of no equivalent, has a dignity.

That which refers to universal human inclinations and needs has a mar-

79. See note 59 above (Ak 4:431).

80. The apparent source for this distinction is Seneca, Epistles 71.33. But it is an

atypical passage in the Stoic literature, since typically both pretium and dignitas refer to

the value of (preferred) indifferents rather than to virtue.

Groundwork for the Metaphysics of Morals 53

ket price; that which, even without presupposing any need, is in accord with a certain taste, i.e., a satisfaction in the mere purposeless play of the powers [Ak 4:435]

of our mind, an affective price; but that which constitutes the condition under which alone something can be an end in itself does not have merely a

relative worth, i.e., a price, but rather an inner worth, i.e., dignity. Now morality is the condition under which alone a rational being can be

an end in itself, because only through morality is it possible to be a legisla-

tive member in the realm of ends. Thus morality and humanity, insofar as it

is capable of morality, is that alone which has dignity. Skill and industry in

labor have a market price; wit, lively imagination, and moods have an

affective price; by contrast, fidelity in promising, benevolence from princi-

ple (not from instinct) have an inner worth. Lacking these principles, nei-

ther nature nor art contain anything that they could put in the place of them;

for the worth of these principles does not consist in effects that arise from

them, in the advantage and utility that they obtain, but rather in the disposi-

tions, i.e., the maxims of the will, which in this way are ready to reveal

themselves in actions, even if they are not favored with success. These

actions also need no recommendation from any subjective disposition81 or

taste, regarding them with immediate favor and satisfaction, and no imme-

diate propensity or feeling for it:82 they exhibit the will that carries them out

as an object of an immediate respect, for which nothing but reason is

required in order to impose them on the will, not to cajole them from it by flattery, which latter would, in any event, be a contradiction in the case of duties. This estimation thus makes the worth of such a way of thinking to be

recognized as dignity, and sets it infinitely far above all price, with which it

cannot at all be brought into computation or comparison without, as it were,

mistaking and assailing83 its holiness.

And now, what is it that justifies the morally good disposition or virtue in

making such high claims? It is nothing less than the share that it procures for the rational being in the universal legislation, thereby making it suitable as a member in a possible realm of ends, for which it by its own nature was

already destined, as end in itself and precisely for this reason as legislative

in the realm of ends, as free in regard to all natural laws, obeying only those

that it gives itself and in accordance with which its maxims can belong to a

universal legislation (to which it at the same time subjects itself ). For [Ak 4:436]

81. Disposition

82. dieselbe, which would appear to refer to ‘‘morality’’; some editors substitute the

plural, so that this pronoun refers instead to ‘‘these actions.’’

83. ‘‘mistaking and assailing’’ = vergreifen

54 Groundwork for the Metaphysics of Morals

nothing has a worth except that which the law determines84 for it. The

legislation itself, however, which determines all worth, must precisely for

this reason have a dignity, i.e., an unconditioned, incomparable worth; the

word respect alone yields a becoming expression for the estimation that a rational being must assign to it. Autonomy is thus the ground of the dignity of the human and of every rational nature.

The three ways mentioned of representing the principle of morality are,

however, fundamentally only so many formulas of precisely the same law,

one of which unites the other two in itself. Nonetheless, there is a variety

among them, which is to be sure more subjectively than objectively practi-

cal,85 namely that of bringing an idea of reason nearer to intuition (in

accordance with a certain analogy) and, through this, nearer to feeling. All

maxims have, namely,

(1) a form, which consists in universality, and then the formula of the moral imperative is expressed thus: ‘That the maxims must be chosen as if

they are supposed to be valid as universal laws of nature’;

(2) a matter,86 namely an end, and then the formula says: ‘That the rational being, as an end in accordance with its nature, hence as an end in

itself, must serve for every maxim as a limiting condition of all merely

relative and arbitrary ends’;

(3) a complete determination87 of all maxims through that formula, namely ‘That all maxims ought to harmonize from88 one’s own legislation

into a possible realm of ends as a realm of nature’.* A progression happens

here, as through the categories of the unity of the form of the will (its universality), the plurality of the matter (the objects, i.e., the ends), and the allness or totality of the system of them.89 But one does better in moral

*Teleology considers nature as a realm of ends, morality a possible

realm of ends as a realm of nature. In the former, the realm of ends is a

theoretical idea for the explanation of what exists. In the latter, it is a

practical idea to bring about that which does not exist but what can become

actual through our deeds and omissions and what we are to bring about in

accord with precisely this idea.

84. 1785: ‘‘except the one the law determines’’

85. This could also be translated ‘‘more subjective than objectively practical.’’

86. Kant’s text reads Maxime; but editors universally correct this to Materie, as seems

absolutely required by the second sentence of (3) below.

87. See Critique of Pure Reason, A571–83/B599–611.

88. 1785: ‘‘as’’

89. See Critique of Pure Reason, A70–71/B95–96, A80/B106.

Groundwork for the Metaphysics of Morals 55

judging always to proceed in accordance with the strict method and take as ground the universal formula of the categorical imperative: Act in accor- dance with that maxim which can at the same time make itself into a universal law. But if one wants at the same time to obtain access for the [Ak 4:437] moral law, then it is very useful to take one and the same action through the

three named concepts and thus, as far as may be done, to bring the action

nearer to intuition.

Now we can end at the place from which we set out 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 it is made into a universal law, can never conflict with itself. This principle is therefore

also its supreme law: ‘Act always in accordance with that maxim whose

universality as law you can at the same time will’; this is the single condi-

tion under which a will can never be in conflict with itself, and such an

imperative is categorical. Because the validity of the will as a universal law

for possible actions has an analogy with the universal connection of the

existence of things in accordance with universal laws, which is what is

formal in nature in general, the categorical imperative can also be expressed

thus: Act in accordance with maxims that can at the same time have them- selves as universal laws of nature for their object. This, therefore, is the way the formula of an absolutely good will is constituted.

Rational nature discriminates itself from the rest in that it sets itself an

end. This would be the matter of every good will. But since, in the idea of a

will that is absolutely good without a limiting condition (of the attainment

of this or that end), every end to be effected has to be thoroughly abstracted from (as it would make every will only relatively good), the end here has to

be thought of not as an end to be effected but as a self-sufficient end, hence only negatively, i.e., never to be acted against, which therefore has to be

estimated in every volition never merely as means but always at the same

time as end. Now this cannot be other than the very subject of all possible

ends, because this is at the same time the subject of a possible absolutely

good will; for this will cannot without contradiction be set after any other

object. The principle:90 Accordingly, ‘Act in reference to every rational

being (to yourself and others) so that in your maxim it is always valid at the

same time as an end in itself’ is, fundamentally, the same as the principle

‘Act in accordance with a maxim that at the same time contains its own

universal validity for every rational being’. For that I ought to limit my [Ak 4:438]

maxim in the use of means to every end to the condition of its universality

90. 1785: ‘‘But the principle: . . .’’

56 Groundwork for the Metaphysics of Morals

as a law for every subject, says just as much as that the subject of ends, i.e.,

the rational being itself, must be made the ground of all maxims of actions

never merely as means, but as the supreme limiting condition in the use of

all means, i.e., always at the same time as end.

Now it incontestably follows from this that every rational being, as an

end in itself, would have to be able to regard itself at the same time as

universally legislative in regard to all laws to which it may be subject,

because precisely this suitableness of its maxims for the universal legisla-

tion designates it as an end in itself, just as the fact that this dignity (pre-

rogative) before all mere beings of nature brings with it to have to take its

maxims always from its own point of view but also at the same time from

that of every other rational being as a universally legislative being (which is

why they are also called ‘persons’). Now in such a way a world of rational

beings (mundus intelligibilis) 91 is possible as a realm of ends, and specifi- cally for all persons through their own legislation as members. Accord-

ingly,92 every rational being must act as if it were through its maxims

always a legislative member in a universal realm of ends. The formal

principle of these maxims is: ‘Act as though your maxim should serve at the

same time as a universal law (for all rational beings)’. A realm of ends is

thus possible only in accordance with the analogy with a realm of nature,

but only in accordance with maxims, i.e., with self-imposed rules, whereas

the latter is possible only in accordance with laws of externally93 necessi-

tated efficient causes. Regardless of this, even though nature as a whole is

regarded as a machine, nevertheless one also gives to it, insofar as it has

reference to rational beings as its ends, on that ground, the name ‘realm of

nature’. Such a realm of ends would actually be brought about through

maxims, the rule of which is prescribed by the categorical imperatives of all

rational beings, if they were universally followed. Yet although the rational being might punctiliously follow these maxims himself, he cannot for that

reason count on everyone else’s being faithful to them, nor on the realm of

nature and its purposive order’s harmonizing with him, as a suitable mem-

ber for a realm of ends that is possible through him, i.e., on its favoring his

expectation of happiness; thus the law ‘Act in accordance with maxims of a[Ak 4:439]

universally legislative member for a merely possible realm of ends’ still

remains in full force, because it commands categorically. And precisely in

this lies the paradox that merely the dignity of humanity as rational nature,

91. intelligible world

92. 1785: ‘‘Nevertheless’’

93. 1785: ‘‘laws also of externally’’

Groundwork for the Metaphysics of Morals 57

without any other end or advantage to be attained through it, hence the

respect for a mere idea, ought nevertheless to serve as an unremitting

precept of the will, and that the sublimity of the maxim consists in just its

independence of all incentives, and the dignity of every rational subject

consists in being a legislative member in the realm of ends; for otherwise it

would have to be represented as subject only to the natural law of its needs.

Although the natural realm, too, as well as the realm of ends, is thought of

as united under a supreme head, and the latter thereby would no longer

remain a mere idea but obtain true reality, so that through this the maxim

would receive the accretion of a strong incentive; yet no increase of its inner

worth would thereby come about; for irrespective of that, this sole un-

limited legislator must always be so represented as judging the worth of the

rational beings only in accordance with their selfless conduct as prescribed

by itself merely through that idea. The essence of things does not alter

through their external relations, and it is in accordance with that which

alone constitutes the absolute worth of the human being, without thinking

about such relations, that he must be judged by whoever it may be, even by

the highest being. Morality is thus the relation of actions to the autonomy of the will, that is, to the possible universal legislation through its maxims.

That action which can subsist with the autonomy of the will is permitted; that which does not agree with it is impermissible. The will whose maxims necessarily harmonize with the laws of autonomy is a holy, absolutely good will. The dependence of a will which is not absolutely good on the principle

of autonomy (moral necessitation) is obligation. Thus the latter cannot be referred to a holy being. The objective necessity of an action from obliga-

tion is called duty. From what has just been said one can now easily explain how it is that

although under the concept of duty we think a subjection to the law, we at

the same time represent to ourselves a certain sublimity94 and dignity in a [Ak 4:440] person who fulfills all his duties. For to be sure, to the extent that the person

is subject to the moral law, there is no sublimity in him, but there is to the extent that he is at the same time legislative in regard to this law, and is only for that reason subject to them. Also we have shown above how neither fear

nor inclination, but solely respect for the law, is the incentive that can give

the action its moral worth. Our own will, insofar as it would act only under

the condition of a possible universal legislation through its maxims, this

will possible to us in the idea, is the authentic object of respect, and the

94. On the sublime, see Critique of the Power of Judgment, §§ 23–29, Ak 5:248–78;

Anthropology in a Pragmatic Respect, § 68, Ak 7:243.

58 Groundwork for the Metaphysics of Morals

dignity of humanity consists precisely in this capacity for universal legisla-

tion, although with the proviso95 that it is at the same time itself subject to

this legislation.

Autonomy of the will as the supreme principle of morality

Autonomy of the will is the property of the will through which it is a law

to itself (independently of all properties of the objects of volition). The

principle of autonomy is thus: ‘Not to choose otherwise than so that the

maxims of one’s choice are at the same time comprehended with it in the

same volition as universal law’. That this practical rule is an imperative, i.e.,

the will of every rational being is necessarily bound to it as a condition,

cannot be proven through the mere analysis of the concepts occurring in it,

because it is a synthetic proposition; one would have to advance beyond the

cognition of objects and to a critique of the subject, i.e., of pure practical

reason, since this synthetic proposition, which commands apodictically,

must be able to be cognized fully a priori; but this enterprise does not belong in the present section. 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. For thereby it is found that its principle

must be a categorical imperative, but this commands neither more nor less

than just this autonomy.[Ak 4:441]

Heteronomy of the will as the source of all ungenuine principles of morality

If the will seeks that which should determine it anywhere else than in the suitability of its maxims for its own universal legislation, hence if it, insofar

as it advances beyond itself, seeks the law in96 the constitution of any of its

objects, then heteronomy always comes out of this. Then the will does not give itself the law but the object through its relation to the will gives the law

to it. Through this relation, whether it rests now on inclination or on repre-

sentations of reason, only hypothetical imperatives are possible: ‘I ought to

do something because I will something else’. By contrast, the moral, hence categorical, imperative says: ‘I ought to act thus-and-so even if I did not will

anything else’. E.g., the former one says: ‘I ought not to lie, if I want to retain

my honorable reputation’;97 but the latter says: ‘I ought not to lie, even if I did

95. Beding

96. 1785: ‘‘hence if it advances . . . , and seeks the law in’’

97. bei Ehren bleiben

Groundwork for the Metaphysics of Morals 59

not incur the least disgrace’.98 The last must therefore abstract from every

object to the extent that it has no influence on the will, hence practical reason (will) does not merely administer some other interest, but merely proves its

own commanding authority as supreme legislation. Thus, e.g., I should seek

to promote someone else’s happiness, not as if its existence mattered to

me (whether through immediate inclination or any satisfaction indirectly

through reason) but merely because the maxim that excludes it cannot be

comprehended in one and the same volition as a universal law.

Division of all possible principles of morality

from the

assumed fundamental concept of heteronomy

Here as elsewhere, human reason in its pure use, as long as it has gone

without critique, has previously tried all possible incorrect routes before it

succeeds in getting on the only true one.99

All principles that one may take from this point of view are either

empirical or rational. The first, from the principle of happiness, are built on [Ak 4:442] physical100 or moral feeling;101 the second, from the principle of perfection, are built either on the rational concept of it as a possible effect102 or on the

concept of a self-sufficient perfection (the will of God) as determining

cause of our will.103

Empirical principles are everywhere unsuited to having moral laws grounded on them. For the universality, with which they are to be valid for

all rational beings without distinction, the unconditioned practical neces-

sity, which is imposed on these beings through them, drops out if the ground

of these principles is taken from the particular adaptation of human nature or from the contingent circumstances in which it is placed. Yet the principle

of one’s own happiness is most reprehensible, not merely because it is false

98. Schande

99. Cf. Critique of Pure Reason, A xii.

100. Kant associates this with the hedonism of Epicurus (341–270 b.c.) (Critique of

Practical Reason, Ak 5:40).

101. Kant associates this with the moral sense theory of Francis Hutcheson (1694–

1727) (Critique of Practical Reason, Ak 5:40).

102. Kant associates this with the position of Christian Wolff and the Stoics (Critique

of Practical Reason, Ak 5:40).

103. Kant associates this with the divine command morality of Christian August

Crusius (1715–1775) (Critique of Practical Reason, Ak 5:40).

60 Groundwork for the Metaphysics of Morals

and experience contradicts the pretense that one’s own welfare always

accords with conducting oneself well; also not merely because it contrib-

utes nothing to the grounding of morality, since making a happy human

being is something other than making a good one, and making him prudent

and sharp-witted for his own advantage is something other than making him

virtuous; but rather because it attributes incentives to morality that would

sooner undermine it and annihilate its entire sublimity, since they put the

motivations for virtue in the same class as those for vice and only teach us to

draw better calculations, but utterly extinguish the specific difference be-

tween them; by contrast, moral feeling, this allegedly special sense* (how-

ever shallow the appeal to it may be, since those who cannot think believe they can help themselves out by feeling when it comes to universal laws, even though feelings, which by nature are infinitely distinguished from one

another in degree, cannot yield an equal standard of good and evil, nor can

one validly judge for others at all through his feeling) nevertheless remains

closer to morality and its dignity by showing virtue the honor of ascribing to

it immediately the satisfaction and esteem we have for it, and not saying directly to its face, as it were, that it is not its beauty, but only our advantage,[Ak 4:443]

that attaches us to it.

Among rational grounds104 of morality, the ontological concept of per- fection (however empty, indeterminate, hence unusable it may be for find- ing in the immeasurable field of possible reality the greatest suitable sum

for us, and however much it has an unavoidable propensity to turn in a

circle in order to distinguish the reality talked about here specifically from

every other, and cannot avoid covertly presupposing the morality it ought to

explain) is nevertheless better than the theological concept, of deriving

morality from a divine, all-perfect will, not merely because we do not intuit

his perfection, but can derive it solely from our concepts, of which morality

is the foremost one, but because if we do not do this (which, if we did,

would be a crude circle in explanation), the concept of his will that is left

over to us, the attributes105 of the desire for glory and domination, bound up

*I count the principle of moral feeling to that of happiness, because[Ak4:442]

every empirical interest promises a contribution to welfare through the

agreeableness something affords, whether this happens immediately and

without any aim to advantage or in regard to the latter. Likewise one must,

with Hutcheson, count the principle of sympathetic participation in an- other’s happiness under the same moral sense assumed by him.

104. rationalen oder Vernunftgründe

105. Eigenschaften

Groundwork for the Metaphysics of Morals 61

with frightful representations of power and vengeance, would have to make

a foundation for a system of morals that is directly opposed to morality.

But if I had to choose between the concept of moral sense and that of

perfection in general (both of which at least do not infringe morality, even if

they are not at all suitable for supporting it as a foundation), then I would

determine myself for the latter, because, since it at least transfers the deci-

sion of the question from sensibility to the court of pure reason, even if here

it decides nothing, nevertheless it preserves unfalsified the indeterminate

idea (of a will good in itself ) for closer determination.

Besides, I believe I can dispense with an extensive refutation of all these

doctrines.106 It is so easy, and even those whose office it is to declare

themselves for one of these theories presumably have such good insight

into it (because their hearers would not tolerate a postponement of judg-

ment) that it would be only superfluous labor. What interests us more here is

to know that these principles everywhere set up nothing but heteronomy of

the will as the first ground of morality and just for this reason must neces-

sarily miscarry regarding their end. [Ak 4:444]

Wherever an object of the will has to be taken as the ground in order to

prescribe the rule determining that will, there the rule is nothing but heter-

onomy; the imperative is conditioned, namely: if or because one wills this object, one ought to act thus or so; hence it can never command morally,

i.e., categorically. Now the object may determine the will by means of

inclination, as with the principle of one’s own happiness, or by means of a

reason directed to objects of our possible volition in general, in the principle

of perfection; then the will never determines itself immediately through the representation of the action, but only through the incentive, which the

foreseen effect of the action has on the will; I ought to do something be- cause I will something other than that, and another law in my subject must therefore be taken as ground, in accordance with which I necessarily will

this other thing, which law once again needs an imperative that limits this

106. Kant’s fuller taxonomy (Critique of Practical Reason, Ak 5:40) divides theories

of heteronomy into four kinds. ‘‘Objective’’ theories are either (1) ‘‘internal’’ (the theory

of perfection) or (2) ‘‘external’’ (divine command theory). (3) ‘‘Subjective internal’’

theories include both the theory of physical and the theory of moral feeling. This taxon-

omy makes a place for yet another classification not discussed in the Groundwork,

namely (4) ‘‘subjective external’’ theories. These include the theory that morality is

grounded on education, which Kant associates with Michel Montaigne (1533–1592), and

the theory that morality is grounded on the civil constitution, which Kant associates with

Bernard Mandeville (1670–1733).

62 Groundwork for the Metaphysics of Morals

maxim. For because the impulse that the representation of an object possi-

ble through our powers is supposed to exercise on the subject’s will in

accordance with its natural constitution, whether it be of sensibility (of

inclination and taste) or of understanding and reason—which, in accor-

dance with the particular adaptation of its nature, that faculty exercises with

satisfaction in an object107—it is really nature that would give the law,

which as such would have to be not only cognized and proven through

experience, and hence is in itself contingent and thereby becomes unsuit-

able for an apodictic rule such as the moral rule has to be; but it is always only heteronomy of the will, the will does not give the law to itself, but rather an alien impulse gives it by means of the subject’s nature, which is

attuned to the receptiveness of the will.

The absolutely good will, whose principle must be a categorical impera-

tive, will therefore, undetermined in regard to all objects, contain merely

the form of volition in general, and indeed as autonomy, i.e., the suitability of the maxim of every good will to make itself into a universal law is itself

the sole law that the will of every rational being imposes on itself, without

grounding it on any incentive or interest in it.

How such a synthetic practical proposition a priori is possible, and why it is necessary, is a problem whose solution no longer lies within the bound-

aries of the metaphysics of morals, neither have we here asserted its truth,[Ak 4:445]

much less pretended to have a proof of it in our control. We showed only,

through the development of the generally accepted concept of morality,108

that it is unavoidably attached to, or rather is grounded on, an autonomy of

the will. Thus whoever takes morality to be something, and not a chimerical

idea without truth, must at the same time concede the stated principle of it.

This section, therefore, like the first one, was merely analytical. Now that

morality is no figment of the mind, which follows if the categorical impera-

tive, and with it autonomy of the will, is true and absolutely necessary as a

principle a priori—this requires a possible synthetic use of pure practical reason, upon which, however, we may not venture without preceding it with a critique of this very faculty of reason, which we have to exhibit in the last section as the main feature of this critique in a way sufficient for our aim.

107. Instead of ‘‘reason—which . . . object,’’ 1785 reads: ‘‘reason takes in perfection

in general, (whose existence either from itself or only depending on the highest self-

sufficient perfection),’’

108. allgemein im Schwange gehenden Begriffs der Sittlichkeit

Third Section

Transition

from the metaphysics of morals

to the critique of pure practical reason

The concept of freedom

is the key to the definition1 of autonomy of the will.

The will is a species of causality of living beings, insofar as they are rational, and freedom would be that quality of this causality by which it can be effective independently of alien causes determining it; just as natural necessity is the quality of the causality of all beings lacking reason, of being determined to activity through the influence of alien causes.

The proposed definition2 of freedom is negative, and hence unfruitful in affording insight into its essence; yet from it flows a positive concept of freedom, which is all the more rich in content and more fruitful. Since the

concept of a causality carries with it that of laws in accordance with which must be posited, through that which we call a cause, something else, namely

its result; therefore freedom, even though it is not a quality of the will in

accordance with natural laws, is not for this reason lawless, but rather it has

to be a causality in accordance with unchangeable laws, but of a particular

kind; for otherwise a free will would be an impossibility.3 Natural necessity

was a heteronomy of efficient causes; for every effect was possible only in

accordance with the law that something else determined the efficient cause

to causality; what else, then, could the freedom of the will be, except [Ak 4:447]

autonomy, i.e., the quality of the will of being a law to itself? But the

proposition ‘The will is in all actions a law to itself’ designates only the

principle of acting in accordance with no other maxim than that which can

also have itself as a universal law as its object. But this is just the formula of

the categorical imperative and the principle of morality: thus a free will and

a will under moral laws are the same.

Thus if freedom of the will is presupposed, then morality follows to-

gether with its principle from mere analysis of its concept. Nonetheless, the

1. Erklärung

2. Erklärung

3. Unding

64 Groundwork for the Metaphysics of Morals

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. Such synthetic propositions, however, are possible

only when both cognitions are combined with one another through the

connection with a third in which they are both to be encountered.4 The

positive concept of freedom makes for5 this third, which cannot be, as with physical causes, the nature of the world of sense (in whose concept the

concepts of something as cause comes together in relation to something else as effect). What this third thing must be, to which freedom points and of

which we have an idea a priori, still cannot be directly indicated here, and to make comprehensible the deduction of the concept of freedom from pure

practical reason, with it also the possibility of a categorical imperative,

instead still needs some preparation.

Freedom must be presupposed

as a quality of the will of all rational beings.

It is not enough that we ascribe freedom to our will, on whatever grounds,

if we do not also have sufficient grounds to attribute the same quality also to

all rational beings. For since morality serves as a law for us merely as for

rational beings, it must also be valid for all rational beings, and since it must be derived solely from the quality of freedom, therefore freedom must also

be proved as a quality of the will of all rational beings, and it is not enough to

establish it from certain alleged experiences of human nature (although this[Ak 4:448]

is absolutely impossible, and it can be established solely a priori); but rather one must prove it of the activity of rational beings in general, who are

endowed with a will. Now I say: Every being that cannot act otherwise than

under the idea of freedom is precisely for this reason actually free in a practical respect, i.e., all laws inseparably combined with freedom are valid

for it, just as if its will had also been declared6 free in itself and in a way that is

valid in theoretical philosophy.* Now I assert that we must necessarily lend

*I take this route, of assuming freedom as sufficient for our aim only as

rational beings ground it on the idea in their actions, so that I may not be obligated to prove freedom also in its theoretical intent. For even if this

latter is left unsettled, these same laws that would obligate a being that is

actually free are still valid for a being that cannot act otherwise than as

4. See Critique of Pure Reason, A7–10/B11–14.

5. schafft

6. erklärt

Groundwork for the Metaphysics of Morals 65

to every rational being that has a will also the idea of freedom, under which

alone it would act. For in such a being we think a reason that is practical, i.e.,

has causality in regard to its objects. 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; conse-

quently 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.

Of the interest attaching to the ideas of morality

We have ultimately traced the determined concept of morality to the idea

of freedom; but we cannot prove this freedom as something actual, not even

in ourselves, nor in human nature; we saw only that we have to presuppose [Ak 4:449]

it if we would think of a being as rational and as endowed with conscious-

ness of its causality in regard to actions, i.e., with a will; thus we find that

from precisely the same ground we have to attribute to every being en-

dowed with reason and will this quality, to determine itself to action under

the idea of its freedom.

From the presupposition of these ideas,7 however, there also flowed the

consciousness of a law of acting: that the subjective principles of actions,

i.e., maxims, have always to be taken so that they can also be valid objec-

tively, i.e., universally as principles, hence serve for our own universal

legislation. But why ought I to subject myself to this principle, and specifi-

cally as a rational being in general, hence through this also all other beings

endowed with reason? I will concede that no interest drives me to it, for that would yield no categorical imperative; but I must necessarily take an inter- est in it, and gain insight into how that happens to be; for this ‘ought’ is

really a volition that would be valid for every rational being, under the

condition that reason were practical in him without any hindrances; for

beings, such as we are, who are also affected through sensibility, as with

incentives of another kind, with whom what reason for itself alone would

under the idea of its own freedom. Thus we can free ourselves of the burden

that pressures theory.

7. diese Ideen, which seem to refer to the ‘‘ideas of morality’’ mentioned in the

subheading; because this yields a doubtful meaning for the sentence, some editors amend

the text to read diese Idee (sc. the idea of freedom).

66 Groundwork for the Metaphysics of Morals

always do does not always happen, that necessity of action is called only an8

‘ought’, and the subjective necessity is different from the objective.

It therefore appears as if in the idea of freedom we really only presup-

posed the moral law, namely the principle of the autonomy of the will itself,

and could not prove its reality and objective necessity for itself; and then we

would still have gained something quite considerable, more than would

have happened otherwise, by at least determining the genuine principle

more precisely, but in regard to its validity and the practical necessity of

subjecting ourselves to it, we would have come no further; for to someone

who asked us why the universal validity of our maxim as a law has to be the

limiting condition of our actions, and on what we ground the worth that we

attribute to this way of acting—a worth that is to be so great that there can

nowhere be any higher interest—and how it happens to be that the human

being believes he feels his personal worth through it alone, and that over[Ak 4:450]

against it an agreeable or disagreeable condition is held to be nothing—to

him we can give no satisfactory answer.

We indeed find that we can take an interest in a constitution of person-

ality9 that carries with it no interest at all in the condition, if only the former

makes us susceptible to partaking in the latter just in case reason should

effect the distribution of it, i.e., that the mere worthiness of being happy,

even without the motive to partake in this happiness, could interest for

itself: but this judgment is in fact only the effect of moral laws whose

importance has already been presupposed (if we separate ourselves from all

empirical interest through the idea of freedom), but that we ought to sepa-

rate ourselves from this, i.e., consider ourselves as free in acting and thus

nevertheless take ourselves to be subject to certain laws in order to find a

worth merely in our person, and that this could compensate us for the loss of

all the worth procured for our condition; and how this is possible, thus from whence the moral law obligates—in such a way we still gain no insight into this.

One must freely admit it10 that a kind of circle shows itself here, from

which, it seems, there is no way out. In the order of efficient causes we

assume ourselves to be free in order to think of ourselves as under moral

laws in the order of ends, and then afterward we think of ourselves as

subject to these laws because we have attributed freedom of the will to

ourselves, for freedom and the will giving its own laws are both autonomy,

8. 1785: ‘‘in the’’

9. persönliche Beschaffenheit

10. This word added in 1786

Groundwork for the Metaphysics of Morals 67

hence reciprocal concepts, of which, however, just for this reason, one

cannot be used to define11 the other and provide the ground for it, but at

most only with a logical intent to bring various apparent representations of

the same object to a single concept (as different fractions with the same

value are brought to the lowest common denominator).

But one way out remains for us, namely to seek whether, when we think

of ourselves through freedom as a priori efficient causes, we do not take a different standpoint from when we represent ourselves in accordance with

our actions as effects that we see before our eyes.

No subtle reflection is required to make the following remark, but rather

one can assume that the commonest understanding might make it, even if in

its own way, through an obscure distinction of the power of judgment that it [Ak 4:451]

calls feeling: that all representations that come to us without our choice

(like those of sense) give us objects to cognize only as they affect us, so that

what they might be in themselves remains unknown to us; hence that as

regards this species of representations, even with the most strenuous atten-

tion and distinctness that the understanding might add to them, we can

attain merely to the cognition of appearances, never to things in them- selves.12 As soon as this distinction is made (perhaps merely through the variation noted between the representations that are given to us from some-

where else, in which we are passive, and those which we produce solely

from ourselves, and thus prove our activity), then it follows of itself that one

must concede and assume behind the appearances something else that is not

appearance, namely the things in themselves, even if of ourselves we are

satisfied that since they never can become known to us except as they affect

us, we can never come any nearer to them and can never know what they are

in themselves. This13 must yield a distinction, though a crude one, of a

world of sense from a world of understanding,14 of which the first, in accordance with the variations in sensibility of many ways of contemplat-

ing the world, can also be extremely varied, whereas the second, on which it

is grounded, always remains the same. Even about himself and in accor-

dance with the acquaintance that the human being has of himself through

inner sensation, he may not presume to cognize how he is in himself. For

since he does not, as it were, make himself and gets his concept not a priori

11. erklären

12. Cf. Critique of Pure Reason, B xv–xxii, A26–30/B42–46, A32–49/B49–73,

A490–97/B518–25.

13. 1785: ‘‘They’’

14. Cf. Critique of Pure Reason, A235–60/B294–315.

68 Groundwork for the Metaphysics of Morals

but empirically, it is natural that he can take in information even about

himself through inner sense and consequently only through the appearance

of his nature and the way his consciousness is affected, whereas he neces-

sarily assumes about this constitution of his own subject, which is com-

posed of sheer appearances, that it is grounded on something else, namely

his I, however that may be constituted in itself, and must therefore count

himself in regard to mere perception and the receptivity of sensations as in

the world of sense, but in regard to whatever in him may be pure activity (what attains to consciousness not through the affection of the senses but

immediately), he must count himself as in the intellectual world, of which, however, he has no further acquaintance.

A reflective human being must draw such a conclusion about all things[Ak 4:452]

that might come before him; presumably it is also to be encountered in the

commonest understanding, which, as is well known, is very much inclined

to expect behind the objects of sense always something invisible and for

itself active, but is corrupted by the fact that it wants to make this invisible

once again into something sensible, i.e., into an object of intuition, and

thereby does not become by any degree the wiser.15

Now the human being actually finds in himself a faculty through which

he distinguishes himself from all other things, and even from himself inso-

far as he is affected by objects, and this is reason. This as pure self-activity is elevated even above the understanding in the respect that although the latter is also self-activity and does not, like sense, contain mere representa-

tions that arise only when one is affected by things (hence passive), it can

produce no other concepts from its activity except those that merely serve

to bring sensible representations under rules and thereby to unite them in one consciousness, without which use of sensibility it would think nothing

at all,16 while by contrast, reason, under the name of the ideas, shows such a

pure spontaneity that it thereby goes far beyond everything that sensibility

can provide it, and proves its most excellent occupation by distinguishing

the world of the senses and the world of the understanding from one an-

other, thereby, however, delineating the limits of the understanding itself.17

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; hence it has two standpoints, from

which it can consider itself and cognize the laws for the use of its powers,

15. klüger

16. See Critique of Pure Reason, A84–130/B116–69.

17. See Critique of Pure Reason, A293–309/B349–66.

Groundwork for the Metaphysics of Morals 69

consequently all its actions: first, insofar as it belongs to the world of sense, under natural laws (heteronomy), and second, as belonging to the intelligi- ble world, under laws which are independent of nature, not empirical, but

rather grounded merely in reason.

As a rational being, hence one belonging to the intelligible world, the

human being can never think of the causality of its own will otherwise than

under the idea of freedom; for independence of determinate18 causes of the

world of sense (such as reason must always attribute to itself ) is freedom.

Now with the idea of freedom the concept of autonomy is inseparably bound up, but with the latter the universal principle of morality, which in the

idea grounds all actions of rational beings just as the natural law grounds all appearances. [Ak 4:453]

Now the suspicion has been removed that we aroused above, that there

was a hidden circle contained in our inference from freedom to autonomy

and from the latter to the moral law, namely that we perhaps took freedom

as a ground only for the sake of the moral law in order afterward to infer the

latter once again from freedom, hence that we could not offer any ground

for the former, but rather only as begging a question,19 which well-disposed

souls might concede to us, but which we could never set up as a provable

proposition. For now we see that if we think of ourselves as free, then we

transport ourselves as members into the world of understanding and cog-

nize the autonomy of the will, together with its20 consequence, morality; but

if we think of ourselves as obligated by duty,21 then we consider ourselves

as belonging to the world of sense and yet at the same time to the world of

understanding.

How is a categorical imperative possible?

The rational being counts himself as intelligence in the world of under-

standing, 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 ap-

18. bestimmten; following Kant’s formulation later at Ak 4:455, editors often emend

this to bestimmenden (‘‘determining’’).

19. sondern nur als Erbittung eines Prinzips. If there is a verb here, it is not obvious

what it is; the clause might refer by parallel construction to aufstellen könnten (‘‘could set

up’’).

20. 1785: seiner, the most natural grammatical referent of which would have to be

‘‘will.’’ 1786 changes this to ihrer, whose natural referent is ‘‘autonomy.’’

21. verpflichtet

70 Groundwork for the Metaphysics of Morals

pearances 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 in-

clinations 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 na-

ture. (The former would rest on the supreme principle of morality, the

second on that of happiness.) But because22 the world of understanding contains the ground of the world of sense, hence also of its laws, hence is immediately legislative in regard to my will (which belongs wholly to the

world of understanding), and hence must also be thought of wholly as such,

therefore as intelligence I will cognize myself, though on the other side as a[Ak 4:454]

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 impera-

tives and the actions that accord with this principle as duties.

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; 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.23

The practical use of common human reason confirms the correctness of

this deduction. There is no one, even the most wicked scoundrel, if only he

is otherwise accustomed to use his reason, who does not wish, if one lays

22. 1785: But because emphasized along with what follows

23. Cf. Critique of Pure Reason, B 162–65.

Groundwork for the Metaphysics of Morals 71

before him examples of honesty in aims, steadfastness in following good

maxims, sympathetic participation, and general benevolence (and in addi-

tion combined with great sacrifices of advantage and convenience) that he

might also be so disposed. But he cannot bring it about on account of his

inclinations and impulses, while at the same time he wishes to be free of

such burdensome inclinations. Thus through this he proves that with a will

free of the impulses of sensibility, he transports himself in thoughts into

entirely another order of things than that of his desires in the field of

sensibility, since from that wish he can expect no gratification of desires,

hence no condition that would satisfy any of his actual or even thinkable

inclinations (for then the very idea that entices him to the wish would forfeit

its superiority), but he can expect only a greater inner worth of his person.

This better person, however, he believes himself to be when he transports [Ak 4:455]

himself into the standpoint of a member of the world of understanding, to

which the idea of freedom, i.e., independence of determining causes24 of the sensible world, involuntarily necessitates him, and in which he is conscious

of a good will, which constitutes by his own admission the law for his evil

will as a member of the sensible world, the law with whose authority he

becomes acquainted when he transgresses it. The moral ‘ought’ is thus his

own necessary volition as a member of 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.

Of the uttermost boundary of all practical philosophy

All human beings think of themselves, regarding the will, as free. Hence

all judgments about actions come as if they ought to have happened even if they have not happened. Yet this freedom is no experiential concept, and also cannot be one, because freedom always remains even though experi-

ence shows the opposite of those requirements that are represented as nec-

essary under the presupposition of freedom. On the other side it is just as

necessary that everything that happens should remain unexceptionably de-

termined in accordance with natural laws, and this natural necessity is also

not an experiential concept, precisely because it carries with it the concept

of necessity, hence of a cognition a priori. But this concept of a nature is confirmed through experience and must unavoidably be presupposed if

experience, i.e., cognition of objects of sense connected in accordance with

universal laws, is to be possible. Hence freedom is only an idea of reason, whose objective reality is doubtful in itself, but nature is a concept of

24. 1785: ‘‘to which the idea of freedom from determining causes’’

72 Groundwork for the Metaphysics of Morals

understanding that proves its reality from examples in experience and nec- essarily must prove it.

Now although from this arises a dialectic of reason, since in regard to the

will the freedom attributed to it appears to stand in contradiction with

natural necessity;25 and at this fork in the road, with a speculative intent, reason finds the route of natural necessity much more traveled and useful

than that of freedom: yet with a practical intent the footpath of freedom is the only one on which it is possible to make use of one’s reason for deeds

and omissions; hence it is just as impossible for the subtlest philosophy as[Ak 4:456]

for the commonest human reason to ratiocinate freedom away. Thus the

latter26 must presuppose that no true contradiction is encountered between

freedom and the natural necessity of precisely the same human actions, for

it can give up the concept of nature just as little as it can that of freedom.

Nevertheless this seeming contradiction must be done away with at least

in a convincing way, even if one could never conceive how freedom is

possible. For if even the thought of freedom contradicts itself or the thought

of nature, which is just as necessary, then it, as opposed to natural necessity,

had to be27 completely given up.

But it is impossible to escape this contradiction if the subject, which

supposes itself free, were to think itself in the same sense or in precisely the same relations when it calls itself free as when it assumes it is subject to the natural law in regard to that very action. Hence it is an unremitting problem

of speculative philosophy to show at least that its deception of a contradic-

tion rests on the fact that we think of the human being in another sense and

in other relations when we call him free than when we take him, as a piece

of nature, to be subject to its law, and that both not only can very well stand side by side, but also that they have to be thought as necessarily united in the same subject, since otherwise no ground can be supplied why we should

burden reason with an idea that, even if it can be united without contradic- tion with another that is satisfactorily confirmed, yet nevertheless involves us in an enterprise in which reason in its theoretical use is put in a very tight

spot. This duty, however, lies merely on speculative philosophy, so that it

can free the way for practical philosophy. Thus it is not put at the discretion

of the philosopher whether he will remove this seeming conflict or leave it

25. Cf. Critique of Pure Reason, A532–58/B560–86.

26. Diese, which apparently refers to ‘‘the commonest human reason,’’ but in the

context of the entire paragraph could conceivably be taken to refer to ‘‘reason’’ earlier.

27. mußte; some editors read müßte, ‘‘would have to be.’’

Groundwork for the Metaphysics of Morals 73

untouched; for in the latter case the theory about it is bonum vacans,28 and the fatalist can with grounds enter into possession of it and expel all morals

from its supposed property as taken possession of without title.

Yet one can still not say that the boundary of practical philosophy begins

here. For the settlement of that contest does not at all belong to it, but rather

it only demands of speculative reason that it should bring to an end the

disunity in which these theoretical questions involve it, so that practical [Ak 4:457]

reason can have tranquillity and security against external attacks that might

contest the terrain on which it wants to build.

But the legal claim,29 even of common human reason, on freedom of the

will is grounded on the consciousness and the admitted presupposition of

the independence of reason from all merely subjectively determined causes,

which together constitute that which belongs merely to sensation, hence

under the general term ‘sensibility’. The human being who in such a wise

considers himself as an intelligence sets himself thereby in another order of

things, and in a relation to determinate grounds of an entirely different kind,

when he thinks of himself as an intelligence with a will, consequently as

endowed with causality, than when he perceives himself as a phenomenon

in the world of sense (which he actually is too), and subjects his causality,

regarding external determination, to natural laws. Now he soon becomes

aware that both can take place at the same time, indeed even that they must.

For that a thing in its appearance (belonging to the world of sense) is subject to certain laws, of which the very same thing as thing or being in itself is independent, contains not the least contradiction; but that he must represent and think of himself in this twofold way rests, regarding the first,

on the consciousness of himself as an object affected through sense, and as

far as the second goes, on the consciousness of himself as intelligence, i.e.,

as independent in his use of reason of sensible impressions (hence as be-

longing to the world of understanding).

Hence it comes about that the human being presumes to claim a will that

lets nothing be put to its account that belongs merely to its desires and

inclinations, and on the contrary thinks of actions through itself as possible,

or indeed even as necessary, that can happen only by disregarding all de-

sires and sensible stimuli. The causality of these actions lies in him as

intelligence and in the laws of the effects and actions in accordance with

principles of an intelligible world, of which he perhaps knows nothing

further except that there it is solely reason, and indeed a reason that is pure

28. ‘‘vacant good,’’ i.e., a piece of property belonging to no one

29. Rechtsanspruch

74 Groundwork for the Metaphysics of Morals

and independent of sensibility, that gives the law, and likewise, since in that

world he himself only as intelligence is the authentic self (as human being,

by contrast, only appearance of himself ), those laws apply to him imme-

diately and categorically, so that whatever inclinations and impulses (hence

the entire nature of the world of sense) stimulates him to, they cannot

infringe the laws of his volition as intelligence, even that he is not responsi-[Ak 4:458]

ble to the first and does not ascribe it to his authentic self, i.e., his will,

although he does ascribe to it the indulgence that it would like to bear

toward them, if, to the disadvantage of the rational laws of the will, he were

to concede them influence on its maxims.30

Through the fact that practical reason thinks itself into a world of under- standing, it does not overstep its boundaries, but it would if it tried to intuit or sense itself into it.31 The former is only a negative thought, in regard to the world of sense, which gives no laws to reason in determination of the

will, and is only in this single point positive that that freedom, as a negative

determination, is at the same time combined with a (positive) faculty and

even with a causality of reason, which we call a ‘will’, so to act that the

principle of the actions is in accord with the essential constitution of a

rational cause, i.e., the condition of the universal validity of the maxim as a

law. If, however, it were to fetch an object of the will, i.e., a motivation, from the world of understanding, then it would overstep its boundaries and

presume to be acquainted with something of which it knows nothing. The

concept of a world of the understanding is therefore only a standpoint, apart from appearances, which reason sees itself necessitated to take in order to think of itself as practical, which, if the influences of sensibility were determining for the human being, would not be possible, but which is

necessary insofar as his consciousness of himself as intelligence, hence as a

cause that is rational and active through reason, i.e., freely efficient, is not to

be renounced. This thought obviously carries with it the idea of another

order and legislation than that of the natural mechanism that pertains to the

world of sense, and makes necessary the concept of an intelligible world

(i.e., the whole of rational beings as things in themselves), but without the

30. In this sentence it is unclear whether some of the pronouns refer to ‘‘the human

being’’ or to ‘‘his will.’’ The present translation assumes the latter option. If one were to

take the former, however, the end of the sentence should read: ‘‘i.e., his will, though he

does ascribe to himself the indulgence he would like to bear toward them if, to the

disadvantage of the rational laws of the will, he were to concede them influence on his

maxims.’’

31. wenn sie sich hineinschauen, hineinempfinden wollte

Groundwork for the Metaphysics of Morals 75

least presumption here to think of them further than merely as regards their

formal condition, i.e., the universality of the maxim of the will, as law, hence the will’s32 autonomy, which alone can subsist with freedom; where-

as on the contrary, all laws that are determined to an object give heteron-

omy, which is encountered only in natural laws and can also pertain only to

the world of sense.

But then reason would overstep all its boundaries if it undertook to

explain how pure reason could be practical, which would be fully the same [Ak 4:459] as the problem of explaining how freedom is possible. For we can explain nothing unless we can trace it back to laws the object

of which can be given in some possible experience. But freedom is a mere

idea, whose objective reality can in no wise be established in accordance

with natural laws, hence also not in any possible experience; for the same

reason, because no example may ever be attributed to freedom itself in

accordance with any analogy, freedom can never be comprehended, nor

even can insight into it be gained. It is valid only as a necessary presupposi-

tion of reason in a being that believes itself to be conscious of a will, i.e., of a

faculty varying from a mere faculty of desire (namely, of determining itself

to action as an intelligence, hence in accordance with laws of reason, inde-

pendently of natural instincts). But where the determination in accordance

with natural laws ceases, there too ceases all explanation, and there is nothing left over except defense, i.e., aborting the objections of those who pretend to have looked deeper into the essence of things and therefore

brazenly declare33 freedom to be impossible. One can only show them that

the presumed contradiction they have found in it lies elsewhere, since in

order to make the natural law valid in regard to human actions, they neces-

sarily have to consider the human being as appearance, and now when one

demands of them that they should also think of him as intelligence, also34 as

thing in itself, they are still considering him as appearance, to which ob-

viously the separation of his causality (i.e., of his will) from all natural laws

of the world of sense in one and the same subject would stand in contradic-

tion, but that contradiction goes away if they would keep in mind, and even

32. als Gesetze, mithin der Autonomie des letzteren; the referent of the last term

would have to be ‘‘will’’ (singular), but the apparent antecedent is plural. Some editors

therefore amend Gesetze to Gesetz.

33. erklären; so the clause could also be translated: ‘‘brazenly explain freedom as

impossible.’’

34. 1785: ‘‘as intelligence, but’’

76 Groundwork for the Metaphysics of Morals

admit, as is only fair,35 that behind appearances things in themselves (though

hidden) must ground them, and one cannot demand of their effective laws

that they should be the same as those under which their appearances stand.

The subjective impossibility of explaining36 the freedom of the will is the same as the impossibility of bringing to light and making comprehensi-

ble an interest * that the human being could take in moral laws; and never-[Ak 4:460] theless he actually does take an interest in them, the foundation of which in

us we call ‘moral feeling’, which is falsely given out by some as the stan-

dard of our moral judgment, since it has to be regarded rather as the subjec- tive effect that the law exercises on the will, for which reason alone provides the objective grounds.

In order for a sensibly affected rational being to will that which reason

alone prescribes the ‘ought’, there obviously must belong to it a faculty of

reason to instill a feeling of pleasure or satisfaction in the fulfillment of duty, hence a causality of reason to determine sensibility in accordance with its

principles. It is entirely impossible, however, to gain insight, i.e., to make

comprehensible a priori, how a mere thought that contains nothing sensible in it would produce a sensation of pleasure or displeasure; for that is a

particular kind of causality, of which, as of all causality, we can determine

nothing at all a priori, but rather we have to ask experience alone about it. But since experience can provide no relation of cause to effect except that

between two objects of experience, but here pure reason, through mere ideas

(which yield no object at all for experience), ought to be the cause of an effect

*‘Interest’ is that through which reason becomes practical, i.e., becomes[Ak4:459]

a cause determining the will. Hence one says only of a rational being that it

takes an interest in something; creatures without reason only feel sensible

impulses. Reason takes an immediate interest in an action only when the[Ak4:460]

universal validity of its maxim is a sufficient determining ground of the

will. Such an interest is alone pure. But if it can determine the will only by

means of another object of desire or under the presupposition of a particular

feeling of the subject, then reason takes only a mediated interest in the

action, and since reason for itself alone without experience can bring to

light neither objects of the will nor a feeling grounding it, the latter interest

would be only empirical and not a pure rational interest. The logical interest

of reason (to promote its insights) is never immediate but presupposes aims

of its use.

35. billig

36. erklären, which could also be translated here as ‘‘defining’’

Groundwork for the Metaphysics of Morals 77

which obviously lies in experience, it is entirely impossible for us human

beings to have an explanation how and why the universality of the maxim as a law, hence morality, should interest us. Only this much is certain: that it does not have validity for us because it interests us (for that is heteronomy and dependency of practical reason on sensibility, namely a feeling ground-

ing it, which could never be morally legislative), but rather that it inter- [Ak 4:461]

ests us because it 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. Thus the question ‘How is a categorical imperative possible?’ can be

answered to this extent: one can state the sole presupposition under which

alone it is possible, namely the idea of freedom, and to the extent that one can

have insight into the necessity of this presupposition, which is sufficient for

the practical use of reason, i.e., for the conviction of the validity of this imperative, hence also of the moral law; but how this presupposition itself is possible, no insight into that can be gained through any human reason. Under

the presupposition of freedom of the will, its autonomy, as the formal condition under which alone it can be determined, is a necessary conse-

quence. To presuppose this freedom of the will is also not only (as specula-

tive philosophy can show) entirely possible (without falling into contradic- tion to the principle of natural necessity in the connection of appearances in

the world of sense), but it is also without any further condition necessary to impute37 to it practically all its voluntary38 actions, i.e., necessary as condi-

tion in the idea, to a rational being, who is conscious of its causality through

reason, hence of its will (which is distinguished from desires). But now how pure reason can for itself be practical, without any other incentive that might

be taken from anywhere else, i.e., how the mere principle of the universal validity of all its maxims as laws (which obviously would be the form of a pure practical reason), without any material (object) of the will in which one

might previously take any interest, should for itself yield an incentive and

effect an interest that would be called purely moral—or in other words, how pure reason could be practical—all human reason is entirely incapable of explaining that, and all the effort and labor spent in seeking an explanation

are lost.

It is precisely the same as if I sought to get to the ground of how freedom

37. unterzulegen

38. willkürlich

78 Groundwork for the Metaphysics of Morals

itself, as the causality of a will, is possible. For there I forsake the philosophi-

cal ground of explanation and have no other. Now of course I could enthuse[Ak 4:462]

about39 in the intelligible world that is left over to me, in the world of

intelligences; but although I have an idea of it, which has its own good ground, I still have not the least acquaintance with it and also can never reach one through every striving of my natural faculty of reason. It signifies

only a ‘something’ that is left over if I have excluded everything from the

determining grounds of my will that belongs to the world of sense, merely in

order to limit the principle of motivation from the field of sensibility, by

setting boundaries to it and showing that it does not embrace all in all, but

that outside that principle I am still more; but I am not any further acquainted

with this ‘more’. Of pure reason, which thinks this ideal, there is left over to

me to be thought, after the separation of all matter, i.e., the cognition of

objects, only the form, namely the practical law of the universal validity of

maxims, and in accord with this, reason in reference to a pure world of

understanding as possible efficient cause, i.e., as determining the will; here

the incentive has to be entirely lacking; it would have to be this idea of an

intelligible world itself that is the incentive, or that in which reason orig-

inally would take an interest; but to make this comprehensible is precisely

the problem that we cannot solve.

Now here is the supreme boundary of all moral inquiry; to determine it,

however, is already of great importance, so that, on the one side, reason, in a

way harmful to morality, does not look around in the world of sense for the

supreme motivation and for a comprehensible but empirical interest, but on

the other side, so that, in what for it is the empty space of transcendent

concepts, under the name of the intelligible world, it does not beat its wings

powerlessly, without moving from the spot and losing itself among figments

of the mind. Besides, the idea of a pure world of the understanding, as a

whole of all intelligences, to which we belong as rational beings (though on

the other side at the same time members of the world of sense) is always a

usable and permissible idea on behalf of a rational faith,40 even if at its

boundary all knowledge has an end, in order to effect a lively interest in the

moral law in us through the splendid ideal of a universal realm of ends in themselves (rational beings), to which we can belong as members only when we carefully conduct ourselves in accordance with maxims of freedom as[Ak 4:463]

though they were laws of nature.

39. herumschwärmen

40. Glaubens; see Critique of Pure Reason, A820–31/B848–59.

Groundwork for the Metaphysics of Morals 79

Concluding remark

The speculative use of reason, in regard to nature, leads to the absolute necessity of some supreme cause of the world; the practical use of reason, in regard to freedom, also leads to absolute necessity, but only of the laws of actions of a rational being as such. Now it is an essential principle of every use of our reason to drive its cognition to the consciousness of its necessity (for without this it would not be cognition of reason). But it is also just as

essential a limitation of precisely the same reason, that it can gain no insight either into the necessity of what exists or what happens, or into that which

ought to happen, unless grounded on a condition under which it exists, or happens, or ought to happen. In this wise, however, the satisfaction of

reason is always deferred through the constant questioning after the condi-

tion.41 Hence it seeks restlessly the unconditionally necessary and sees

itself necessitated to assume it, without any means of making it comprehen-

sible; it is fortunate enough if it can only discover the concept that is

compatible with this presupposition. Thus it is no fault of our deduction of

the supreme principle of morality, but only an accusation that one would

have to make against human reason in general, that it cannot make compre-

hensible an unconditioned practical law (such as the categorical imperative

must be) as regards its absolute necessity; for we cannot hold it against

reason that it does not will to do this through a condition, namely by means

of any interest that grounds it, because otherwise it would not be a moral,

i.e., a supreme, law of freedom. And thus we indeed do not comprehend the

practical unconditioned necessity of the moral imperative, but we do com-

prehend its incomprehensibility, which is all that can be fairly required of a philosophy that strives in principles up to the boundary of human reason.

41. Cf. Critique of Pure Reason, A310–40/B366–98, A408–20/B435–48, A497–

515/B525–43.

Rethinking Groundwork for the

Metaphysics of Morals

Why Study Kant’s Ethics?

J. B. SCHNEEWIND

Kant’s Groundwork for the Metaphysics of Morals is a very hard book to understand. Those of us who have put this volume together plainly think it’s

worth trying to do so. Why? There are three reasons.

First, Kant created a dramatically new way of thinking about morality

and about ourselves as moral beings. He held that all previous attempts to

spell out the principles of ethics had been mistaken. In the Groundwork he presented the fundamentals of a different vision of morality. And in later

writings he showed how to work out the details of morality using his new

formulation of its basis. To understand Kant’s ethics historically is to come

to see the emergence of a major new option in Western thought.

Second, Kant’s ethical thought has been profoundly influential. It is one

of the two or three most important contributions that modern moral philoso-

phers have made to our culture. The Groundwork has always been the main text used to learn about Kant’s ethics. Anyone who wants to understand the

history of nineteenth- and twentieth-century moral philosophy and its im-

portance for society has to understand this book.

Third, the positions Kant took in the Groundwork are very much alive in moral philosophy today. A renewal of scholarship, commentary, and philo-

sophical discussion concerning the book began around the middle of the

last century.∞ Many misunderstandings have been cleared away, and Kant’s

other writings on ethics have been brought in to illuminate this one. New

philosophical insights from recent work are being used to show the depth

and importance of what Kant said. Kantian views of morality are a central

topic of contemporary moral philosophy. In developing Kant’s positions to

bring out their pertinence today, advocates of Kantian views depart more

or less from what he himself actually said. But an understanding of the

Groundwork is indispensable for anyone who wants to take part in current discussions of ethics.

These are strong claims about the importance of a short book. It would

84 Jerome B. Schneewind

take another book, and a longer one, to support them all. Here I will try first

to sketch Kant’s epoch-making break with the past and then to indicate

some of the developments underlying his importance for contemporary

moral philosophy.

I . Kant’s Historical Revolution

Two quotations will get us going. The first comes from St. Thomas Aqui-

nas, the great thirteenth-century synthesizer of Roman Catholic doctrine.

The second is from Kant.

Law directs the actions of those that are subject to the government of

someone. Hence, properly speaking, none imposes a law on his own

actions. (Summa Theologiae IaIIae 93.5)

The rational being must always consider itself as giving law in a realm

of ends . . . Morality thus consists in the reference of all action to that

legislation through which alone a realm of ends is possible. But the

legislation must be encountered in every rational being itself, and be

able to arise from its will. (G 4:434)

Aquinas and Kant agree that morality centrally involves law and obe-

dience to law. Both could agree with an important passage in St. Paul’s

Epistle to the Romans (2:14–15): ‘‘when the Gentiles, which have not the

law, do by nature the things contained in the law, these, which have not the

law, are a law unto themselves: Which shew the works of the law written in

their hearts, their conscience also bearing witness.’’ The gentiles do not

have the written Jewish law, but they find an unwritten law in their hearts or

consciences. For Aquinas, the law is put there by God. He finds it unthink-

able that human beings might legislate the moral law that we are all to obey.

Kant thinks that our own reason gives us the law. Morality can be under-

stood only if we see that each of us is equally a lawgiving member of the

group of those who must also obey the moral law. He holds that each of us is

both to legislate the law and to obey it.

Kant’s remark comes from a part of the Groundwork in which he intro- duces the term ‘autonomy’ to indicate what is distinctive about his own

view. The word had long been current in political discussions. An autono-

mous state was one that ruled itself. It could make its own laws without

asking permission from rulers of other states. Kant took the term from

political discourse and changed its meaning. He applied it to individuals

Why Study Kant’s Ethics 85

and to the morality that ought to govern the relations of persons to them-

selves and to one another, regardless of the political laws under which they

lived. He said that morality is a human creation. It is the legislation that

comes from our own rational will.

We can see what a radical innovation this was by looking briefly at the

history of moral philosophy. Ancient thought about ethics from Socrates to

the time of St. Augustine centered on the question of human flourishing (a

translation of the Greek eudaimonia). What is the best life for a human? Classical philosophers all thought that having good relations with others

constituted a major part of the kind of life anyone would want to live.

Consequently they held that in pursuing a good life for ourselves we would

not only have to control our own passions and desires; we would also have

to act thoughtfully and justly toward others. On this view virtue and happi-

ness are inseparable.

Christianity gave a new twist to the search for a good life. Our ultimate

good, theologians held, is to be found in a loving union with God. God

made us so that we all seek such a union. We may not realize that that is

what we want. But we are always dissatisfied with earthly goods. And this

dissatisfaction shows that the pagans were wrong to think that we might

find happiness in the present life. Moreover we are deeply flawed and sinful

beings. We ought to live in loving friendship with other people. But we are

dominated by selfish desires. Morality teaches us what we ought to do, but

we find in ourselves a stubborn resistance to doing it. Instead we seek what

we misguidedly think is our own individual good. We must be made to obey

God’s laws by threats of punishment. Morality thus becomes something

external to our own nature—at least to our fallen nature. It does not come

from within us, emerging as our own concern in the course of our natural

development. It has to be imposed on us.

What is God’s relation to the laws he imposes on all human beings—the

laws of nature, as they came to be called? Aquinas held that God’s intellect

is the source of these laws. God commands us to obey them because he

knows that they contain the core of justice and virtue. Two other medieval

thinkers, Duns Scotus and William of Ockham, proposed an alternative.

God is inscrutable and beyond human understanding, and he imposes on us

whatever laws he chooses. The laws of nature contain the core of justice and

virtue simply because God wills that we obey them. Later historians labeled

these positions ‘‘intellectualism’’ and ‘‘voluntarism.’’ Both positions were

very much a part of arguments about morality with which Kant was famil-

iar. Martin Luther’s teaching derived from the thought of the medieval

voluntarists, and Kant was raised a Lutheran. The German philosopher

86 Jerome B. Schneewind

Leibniz and his follower Christian Wolff were strongly opposed to volun-

tarism, and Kant learned Wolff’s views from his first philosophy teachers.

Luther puts his view bluntly and forcefully: ‘‘God is he for whose will

no cause or ground may be laid down as its rule and standard; for nothing is

on a level with it or above it . . . What God wills is not right because he

ought or was bound so to will; on the contrary, what takes place must be

right, because he so wills.’’≤ This side of Lutheran teaching has a prominent

place in the thought of Samuel Pufendorf, whose work on natural law was

studied throughout Europe for over a century after its publication in 1672.

His central point is simple. Christianity, as St. Paul shows, teaches that

morality is obedience to law. Law is the command of a superior, and only

God’s commands can establish a morality for all humans. But then there

cannot be moral requirements binding God, because—obviously—God

can have no superior. It follows that there can be no morality common to

God and humans. Pufendorf ridicules the very idea. ‘‘For who,’’ he asks,

‘‘dare reason thus? Pay your debts, because God pays his. Be grateful,

because God is kind to them that serve him . . . Honor your parents, because

God honors his. Are not these reasonings manifestly absurd?’’≥

Pufendorf expresses what was then a widespread Christian feeling: we

should humbly adore and obey God even if we do not understand him.

Morality is simply obedience. There were many other Christians who found

this position deeply disturbing. It puts God in the position of a tyrant and

makes us into servile subjects. But Christ taught that the essence of the law

is love: we are to love God above all and our neighbor as ourselves. And a

Pufendorfian God, who lays down arbitrary laws and gives no rationale for

them, is not, the critics said, a God who can be loved. If we are to love God

we must understand his moral commands as expressions of his love for us.

It must be possible for us to understand morality as common to God and

ourselves, however mysterious other aspects of God’s activity may be. If

we govern ourselves by following laws that we see are just and right, we

will be acting as nearly in God’s way as we can. This is what it means to say

that we are made in God’s image. It seemed impossible for the voluntarists

to explain how this could be so.

In Kant’s time there were very few atheists. Most people not only be-

lieved in God but also agreed that he was somehow indispensable for

morality. The Scots philosopher David Hume was an exception. He devel-

oped a view of morality in which God played no role. He saw human

feelings as the source of morality. Purely natural explanations could be

given of those feelings. And there is nothing about us that takes us out of the

Why Study Kant’s Ethics 87

realm of nature. We must see ourselves and our morality as causally deter-

mined parts of a causally determined nature.

Hume’s view avoided the problems of the relation of God to morality,

but it seemed to many people to deprive humans of any special dignity or

worth. For Hume, we are only a kind of animal—cleverer than the rest, but

otherwise not very different from them. Religious believers could not see

morality in this way. They could not accept the idea that the natural world is

the only world there is. We must belong to a supernatural spiritual world,

they held, in which God is supreme. They therefore had to face the problem

of how to preserve human dignity in a universe governed by the kind ofGod

they could not give up.

The intellectualists thought that the voluntarists could not solve the

problem but that they themselves could. If we can know the eternal truth

about morality and are able to bring ourselves to follow its directives just

from concern for righteousness, then we are self-governed. We have a

special relation to God. We are the only part of his creation that can obey his

laws just because we know what they are. It is because of our knowledge—

especially our moral knowledge—that we are entitled to think that we are

made in God’s image.

Some intellectualists held that there are eternal self-evident principles

governing morality. They compared such principles to the axioms of geom-

etry. God knows the moral axioms and so do we. God necessarily governs

himself by them. When we have an intuitive grasp of the moral axioms and

work out how they apply to the case at hand, we are self-governed when we

act as they direct. Other intellectualists made no appeal to such intuitively

evident axioms. They held that God always acts for the best. He created the

most perfect possible universe, and we are to be like him. We are to bring

about the most perfect results we can. To do so we need to know which of

the options before us will bring about the greatest increase in perfection. We

are self-governed when we know what is best and decide to bring it into

existence. Both sorts of intellectualists rejected purely naturalistic theories

of the world and used their moral philosophies as part of their defense of a

religious outlook.

Kant rejected Hume’s naturalism and insisted that we are not merely

natural beings, like the animals. Morality itself shows us that we have free

will. Kant shared the intellectualists’ aim of defending human dignity and

argued that free will gives it to us. But he feared that the intellectualists

were all too likely to think that only people with superior minds could be

morally good. The moral feelings that Hume saw as central could be shared

88 Jerome B. Schneewind

equally by everyone, but plainly some people are much smarter than others.

Does it follow that moral knowledge is not equally available to everyone

alike? It is hard to understand geometry; it may be equally hard for many to

understand morality. It would also be difficult for most people to calculate

which among the choices before them will bring about the greatest increase

in perfection. In either case not everyone can be self-governed. Most people

will have to obey the few who knew the truth about morality. This was a

view that Christian Wolff held. Kant found it abhorrent. Hume’s reliance on

sentiment was part of a naturalistic view that Kant rejected. A purely intel-

lectualist morality apparently led to an elitism that he also found unaccept-

able. He thus seemed to be forced into a voluntarist view. But Kant rejected

the servile attitude that seemed to go with that theory. His new idea of the

autonomy of the will gave him a way to resolve this complex problem. He

accepted the voluntarist claim that morality stems from will, but he trans-

formed the conception of will by making it into a special form of rationality

—practical rationality. He could then say that because the will is itself

rational it contains a law within it, governing all its activities. We do not

need to grasp eternal truths or to calculate complex consequences. A simple

formula governs our legislative activity. It enables us to test our plans for

action and to reject some and accept others. And just as we can think out for

ourselves what we ought to do, so we can motivate ourselves to do it. We do

not need rewards or threats from others to make us act morally. God’s will

and ours are alike in these respects. What God necessarily wills is what we

ought to will. We and God are fellow legislators of a single moral commu-

nity. We are equal to God, not merely his servile subjects, because of our

moral autonomy.

II . Kant’s Current Importance

The conception of morality as autonomy was Kant’s fundamental innova-

tion in moral philosophy. In working out his vision of humans as autono-

mous agents Kant developed new ideas about freedom and the nature of

action that are still being discussed. His view also had important implica-

tions for longstanding positions about the structure and content of morality.

One of these implications accounts for much of the significance of Kantian-

ism in current discussions. It is that no principle of human happiness can be

the foundation of morality or show its point. Utilitarianism is the label most

commonly applied to views that make happiness central to morality. Kant-

ianism is opposed to all such views.

Why Study Kant’s Ethics 89

Kant never denied the importance of happiness. He criticized the Stoics

for thinking that the pleasure of self-approval arising from awareness of

living a virtuous life was happiness enough for human beings. Happiness is

the satisfaction of desire, he held, and he insisted that finite beings such as

we are need to have our desires satisfied. He held it to be an important duty

for each of us to help others achieve happiness as they understood it. But he

denied that morality is simply the set of virtues or directives that lead to

happiness, either our own or that of everyone affected. He held that moral-

ity has a different role in our lives. Morality’s function is to set the limits

within which it is permissible for us to seek our own happiness and to help

others pursue theirs.

Kant had several reasons for rejecting a morality of happiness. One of

them is this. We have little if any control over what desires we have. To say

that what we ought to do is determined by what people want is to subordi-

nate ourselves to our causally determined nature. It is, in other words, to

abandon our autonomy. But the moral law forbids us to do so. More gener-

ally, Kant thinks that we cannot accept any morality holding that the goods

relevant to deciding what to do are made good, and can be known to be

good, without any appeal to what is morally right or obligatory. Kant holds,

against this, that only pleasures and pains that are allowed by the moral law

are morally relevant. In order for a pleasure to count as relevant in deciding

what to do, it must be one that can be obtained by a morally permissible act.

So there must be a way of determining what acts are permissible or imper-

missible prior to knowing what goods in the situation are morally relevant.

And Kant thought, of course, that there is such a way. The categorical

imperative tells us whether or not we may act on any plan of action, and

from this we can learn what acts we may or may not do.

Kantianism is thus an alternative to utilitarianism. It is not the only one.

Various forms of intuitionism hold that we can grasp a number of self-

evident moral truths by which to guide our actions. One of these truths tells

us to be benevolent, or to help others attain happiness, but that is not the

only principle. We are also to tell the truth, keep promises, and be just.

Principles like these may come into conflict with the principle that tells us

to increase the happiness of others. And it is far from self-evident that in

cases of such conflict benevolence ought always to win out.

Intuitionism often seems to be the best account available of the com-

monsense morality that most of us share. We don’t come to philosophy with

some single universal principle that we use to get answers to all our moral

questions. And we do seem to think that it’s just obvious that we should

keep promises, tell the truth, help others in need, be just, and so on. Yet

90 Jerome B. Schneewind

intuitionism has defects. It doesn’t suggest any mode of reasoning for

settling controversies with other people about what to do. And it doesn’t

give us any way of criticizing our pretheoretical moral convictions. Yet we

know that people in the past have thought it ‘‘just obvious’’ that women

should obey men, that people of color were inferior to white people, and

that gay and lesbian practices were abhorrent and unnatural. It’s hard to

escape the thought that some of our ‘‘obvious’’ beliefs might be as be-

nighted as these. But how can we tell?

One appeal of utilitarianism and of Kantianism is that each promises a

way of arguing about moral disagreements and of criticizing socially ac-

cepted moral beliefs. During much of the nineteenth century British and

American philosophers took utilitarianism and intuitionism to be the major

alternatives in moral theory. Kant played little if any role in English-lan-

guage ethics. And during much of the twentieth century English-language

moral philosophers did not discuss the substantive issues of morality. They

were concerned rather with whether moral beliefs were susceptible of ra-

tional proof. Was morality, as Hume held, simply a matter of sentiment? If it

was rational, in what way? Was moral language used to express thought that

could be true or false, or was it, rather, used just to express emotions?

These debates took place against the background of a widely shared

assumption that utilitarianism captured the content or substance of morality,

whether morality rested on reason or on feeling. Many philosophers were

convinced, however, that utilitarianism led to morally unacceptable conclu-

sions. But to oppose it they had little but intuitions or strong convictions.

The first major effort to go beyond intuitionism and provide a systematic

alternative to utilitarianism was John Rawls’s A Theory of Justice (1971). Rawls presented a way of arguing in support of principles of justice that did

not derive them from the good consequences that would follow from obey-

ing them. Instead his principles could be used to determine the relevance of

alleged goods and harms to moral decisions. He asserted what is now called

‘‘the priority of the right to the good,’’ and he linked this strongly with

Kant’s moral views.

Rawls’s work led to a great deal of constructive philosophical interest in

Kant’s ethics. Building on the scholarly work that had been published since

the 1940s, American and British philosophers produced a substantial body

of work influenced strongly by Kant’s moral thought. New interpretations

of almost every aspect of Kant’s ethics were soon followed by important

works applying Kant’s insights to contemporary problems, such as world

hunger. The English-language moral philosophy of the past decades has

revolved to a large extent around controversies concerning Kantian moral

Why Study Kant’s Ethics 91

philosophy. Kant’s views continue to vitalize discussions. And their source

is largely in the text of this volume.

notes

1. H. J. Paton’s Tbe Categorical Imperative (London: Hutchinson, 1947) is a landmark in the modern revitalization of Kant scholarship. It was

preceded by a few German contributions. John Rawls’s A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971) showed how much

some Kantian ideas could be used to help with current issues. Much of the

current discussion of Kantian ethics reflects developments of Rawls’s ideas.

2. Martin Luther: Selections from His Writings, ed. John Dillenberger (Garden City, N.Y.: Doubleday, 1962), pp. 195–96.

3. Quoted in J. B. Schneewind, The Invention of Autonomy (New York: Cambridge University Press, 1998), p. 140.

Acting from Duty

MARCIA BARON

Readers of the Groundwork are often taken aback by Kant’s discussion of acting from duty and moral worth. This reaction is not new. In 1796 Fried-

rich Schiller wryly expressed his distaste for what he took to be Kant’s

position:

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.∞

Schiller apparently read Kant as holding that if one does something nice

for someone, and does it with inclination, one is not virtuous. This is an

extension of a more understandable misreading of the First Section of the

Groundwork. Kant definitely does hold that actions have moral worth only if they are done from duty, and if one thinks that he also holds that an action

does not count as ‘‘done from duty’’ if the agent had (at the time of the

action) an inclination so to act, one will conclude that he holds that an

action can have moral worth only if the agent acted against her inclinations.

Because Kant discusses only instances of acting from duty that are

contrary to inclination (or at least unsupported by inclination), it is easy to

think that only acts that are contrary to inclination can be done from duty.

But the reason he discusses only these cases of acting from duty is that he

needs examples in which it is as evident as it can be that the person is acting from duty. It is evident if the action lacks any other incentive. It is not in the

person’s interest, and he lacks any inclination so to act, so he must be acting

from duty. In the Critique of Practical Reason Kant likens the philosopher who arranges an experiment to ‘‘distinguish the moral (pure) determining

ground from the empirical’’ to the chemist (KpV 5:92). ‘‘The purity of the

moral principle,’’ he writes, ‘‘can be clearly shown only by removing from

Acting from Duty 93

the incentive of the action everything which men might count as a part of

happiness’’ (KpV 5:156). This is precisely what he has done in his examples

at G 4:397–99. He has ‘‘arranged an experiment’’ to distinguish the moral

determining ground from the empirical.

Intuitively this makes sense. Think about when it is clearest that some-

one’s motivation is purely moral: it is when the person is under pressure to

act in a way that she sees to be immoral. Insofar as we believe that she has

no incentive to do x other than the belief that it would be wrong not to do x, we regard her as acting morally (in the strong sense, meaning not only that

she is not acting wrongly, but also that her motivation is purely or quint-

essentially moral).

In thinking about Kant’s examples at G 4:397–99, we need to keep in

mind the context of his discussion of acting from duty. Kant embarks on that

discussion in order to develop the concept of a good will.≤ The concept of

duty ‘‘contains that of a good will, though under certain subjective limita-

tions and hindrances.’’≥ To explicate this concept he puts before us exam-

ples in which it is absolutely clear that the agent is acting from duty, and to

this end he constructs the examples in such a way that other explanations of

the agent’s conduct—self-interest, a desire to please others, or just a warm-

hearted desire to lessen another’s unhappiness—are not available. In a

more typical case it would be less clear whether the agent is acting from

duty or from some inclination, but his aim is not to present typical cases.

Once we understand his aim, the impression that he was saying that acting

from duty precludes having an inclination to the action begins to fade.

Further evidence that acting from duty does not preclude having an

inclination to act accordingly (and doesn’t require that, on balance, one’s

inclinations favor acting contrary to duty) comes from Kant’s remarks at

G 4:397 on the difficulty of discerning whether an action was done from

duty if the subject has an inclination so to act.∂ If having an inclination to do

x were incompatible with doing x from duty, how could there be any diffi- culty? It would be clear from the fact that the agent had this inclination that

the action was not done from duty. That it is not clear, according to Kant,

indicates that acting with such an inclination is compatible with acting from

duty.∑

I

Even if they acknowledge that Kant is not saying that only acts contrary

to inclination can be done from duty, many readers will nonetheless find

94 Marcia Baron

Kant’s position on acting from duty troubling. For what he clearly does hold is that acting from inclination, no matter how laudable the inclination, is in

some important way not as good—not as morally worthy—as acting from

duty. This seems strange, and it is disturbing to many people, even though

they fully understand that one can act from duty without gritting her teeth,

hating what she is doing, resenting having to do it, and so on. It is disturbing

for two reasons.

First, surely it is better to enjoy helping the needy than not to. Yet the

reader of the Groundwork is left with the impression that it does not add anything at all, that it does not make it better. As long as one acts from duty,

it doesn’t seem to matter to Kant whether one does it with gritted teeth or

does it with pleasure. This is disturbing. Surely helping from duty with

pleasure is better than helping from duty with gritted teeth. Yet it seems that

on Kant’s view—at least as it comes across in the Groundwork—it makes no difference.

The second reason why it is disturbing is this. Compare two people. One

helps because she wants to; the other helps from duty and also wants to. To

many readers it seems that there is nothing superior about the conduct of the

second person. Yet Kant undeniably holds that the conduct of the second

person has moral worth, while that of the first person does not. Why? What

is so great about acting from duty?

II

Let us consider these two concerns in turn. To make the first concern

clearer, let me offer an example. I recently helped an acquaintance by

taking care of her child for a few afternoons when she was undergoing

chemotherapy. When she thanked me, I didn’t say, ‘‘Think nothing of it. We

all have to do our duty,’’ but something more along the lines of ‘‘I was glad I

could help.’’ Why does the first seem inappropriate, and what, if anything,

does the fact that it seems inappropriate tell us?

One reason it would be inappropriate to say ‘‘We all have to do our duty’’

in response to an expression of thanks has to do with how we usually

understand the word ‘duty’. We think of duty as something imposed on us,

something that we may well not endorse. We associate the word ‘duty’ with

legal requirements, requirements of one’s role (e.g., as a parent) or one’s

office (e.g., treasurer), requirements of one’s job (to hold office hours),

requirements of those who are members of a club (to pay dues), and expec-

Acting from Duty 95

tations, which may not quite rise to the level of requirements, of members

of a club or an academic department, or residents of a neighborhood. If I

live in a town in which it is required by law that we remove the snow from

the sidewalk in front of our homes, it is my duty to shovel my walk (or

arrange for someone else to do so); if I live in a neighborhood in which we

all do this and expect one another to do so, here, too, I may regard it as my

duty. In each case I need not endorse it. I may think the law silly, or wish that

I lived in a neighborhood in which a more laissez-faire attitude toward

snow removal prevailed. To say ‘‘We all have to do our duty’’ in reply to an

expression of gratitude suggests that this is something I did not exactly

choose to do but felt pressured to do.

It is important to understand that ‘duty’ (Pflicht) in Kant’s ethics is tied not to social expectations or laws, but to rationality. ‘Duty’ for Kant means,

roughly, what one would do if one were fully rational. We are rational

beings, but we are to be contrasted with those beings—if we can imagine

such creatures—with ‘‘holy wills.’’ Beings with holy wills cannot do other

than what reason prescribes.∏ We, by contrast, feel the tug of inclinations,

which not infrequently are at odds with reason. Hence we—rational beings

whose wills are not and cannot be holy—experience the sense of duty as a

constraint, because we are pulled toward what reason demands and toward

what our inclinations urge. To do our duty is not to give in to social expecta-

tions but to do what reason requires (which may or may not be the same

thing).

One reason, then, why it is inappropriate to say ‘‘We all have to do our

duty’’ in response to an expression of gratitude has to do with connotations

of ‘duty’ in ordinary speech, connotations that have no bearing on Kant’s

conception of duty.

A related reason why it is inappropriate is that in general when one does

a favor for another and is thanked for it, it is inappropriate to draw attention

to the effort one expended on behalf of the other, or to dwell on it if the other

draws attention to it. If someone thanks me for a present and says, ‘‘That’s

too generous!’’ it wouldn’t do to say, ‘‘It was a lot of money and a lot of

effort and I was horribly busy then so didn’t really have the time, but after

all you are a good friend.’’ Or if a friend comes to visit and I throw a party

for her and she thanks me and says, ‘‘I really appreciate it. I can see it must

have taken a lot of work,’’ I should not say in reply (unless perhaps in pure

and clear jest), ‘‘It sure did, and I’m utterly exhausted, so I’m glad you

appreciate it.’’ We can see from reflecting on these exchanges that the

inappropriateness of saying, ‘‘We all have to do our duty’’ tells us nothing at

all that indicates that acting from duty is in some way objectionable. It only

96 Marcia Baron

reflects the fact that a certain graciousness, even modesty (in the form of not

drawing attention to or dwelling on the effort to which one went for one’s

friend), is in order in such situations.

But this only partly answers the worry. The issue is not only what the

benefactor says to the recipient, but also the benefactor’s attitude (even if not conveyed to the recipient). It is not just that it is undesirable that one

disclose to the recipient that she found it to be quite a burden. That she found it to be such a burden is itself not good. And so the point remains: it is

preferable that the person who does us a favor does it with pleasure rather

than simply because she sees it to be her duty. This need not mean unalloyed

pleasure or pleasure free of ambivalence. We might prefer that, too—that a

friend do us a favor with unalloyed pleasure—but that is not a preference

that moral theory should strive to accommodate. (Why shouldn’t my friend

be ambivalent about spending two hours driving me to the airport, or giving

up a delicious stretch of free time to visit me in the hospital? I would have to

be extremely self-centered to think that she is wanting as a friend—or, even

more implausibly, as a person—if she is ambivalent.) But a preference that

people aid us with pleasure rather than without does not seem petty or self-

centered or in any other way an unreasonable or unworthy preference.

There is no getting around it: the character of someone who helps with

pleasure—and, more generally, who acts from duty with pleasure—seems

clearly to be better, ceteris paribus, than the character of one who acts from duty without pleasure. Can Kant recognize this, or is his view of moral

motivation and character at odds with it?

To determine this we need to look at Kant’s other works. There is noth-

ing in the Groundwork that suggests that he does not or cannot recognize this (that is, nothing that suggests that his view of moral motivation and

character is at odds with it), but there is also nothing to suggest that he does.

The work to look at is the Metaphysics of Morals, for which the Ground- work for the Metaphysics of Morals was, as the title indicates, intended to be the foundation.

Kant indicates at a number of points in the Metaphysics of Morals that a person who acts from duty would ordinarily act with pleasure. Certainly a

virtuous person would.

The rules for practicing virtue . . . aim at a frame of mind that is

valiant and cheerful in fulfilling its duties. . . . What is not done with pleasure but merely as compulsory service has no inner worth for one

who attends to his duty in this way and such service is not loved by him;

Acting from Duty 97

instead, he shirks as much as possible occasions for practicing virtue.

(MS 6:484)

The context in which the passage occurs bears mention. The quotation is

from a section of the Metaphysics of Morals titled ‘‘Ethical Ascetics,’’ in which Kant argues that the picture of the cultivation of virtue that takes as

its motto the Stoic saying ‘‘Accustom yourself to put up with the misfor- tunes of life that may happen and to do without its superfluous pleasures’’ is defective. Kant explains: ‘‘This is a kind of regimen . . . for keeping a man healthy. But health is only a negative kind of well-being. . . . Something must be added to it . . . This is the ever-cheerful heart, according to the idea

of the virtuous Epicurus’’ (MS 6:484–85).π It is part of being virtuous that one takes pleasure in life and in living as one does, i.e., takes pleasure in

helping others, developing one’s talents, and so on.

When we turn to Kant’s discussion of specific virtues and vices, there,

too, it is evident that he does not regard the sentiments with which we act as

a matter of indifference. We have a duty, he says, to cultivate our compas-

sionate impulses (susceptibility to which ‘‘Nature has already implanted’’

in us [MS 6:456]).∫ We also have duties of gratitude (MS 6:455)Ω and duties

not to be envious (MS 6:459) and not to gloat over others’ misfortunes (MS

6:460). ‘‘To rejoice immediately in the existence of such enormities de- stroying what is best in the world as a whole, and so also to wish for them to

happen, is secretly to hate human beings; and this is the direct opposite of

love for our neighbor, which is incumbent on us as a duty’’ (MS 6:460). It is

also a duty to be forgiving (but not to tolerate wrongs meekly or to renounce

rigorous means for preventing the recurrence of wrongs by others [MS

6:461]).

Reading only the Groundwork, it is easy to get the impression that sentiments do not matter (or, worse, that they matter only negatively, as

obstacles to acting as we should). We need to bear in mind that the aim of

Groundwork for the Metaphysics of Morals is simply to lay the groundwork for the later work. My point is not only that the Metaphysics of Morals presents the final form of Kant’s ethics, the fully developed theory—al-

though I think it does.∞≠ Given the particular task of the Groundwork, it is no wonder that we find little there about sentiments, and that what we do

find underscores their irrelevance, or their lack of ‘‘standing.’’ The main

task of the Groundwork is to seek out and establish the supreme principle of morality and to show that it can and must have a purely nonempirical

foundation. Hence the insignificance of emotions and sentiments to moral-

ity is much more apparent in that work than is their significance. They do

98 Marcia Baron

not play any role in grounding morality,∞∞ and so we hear little about them in

that work—except what they do not do.

Before turning to the second objection to Kant’s discussion of acting

from duty, I want to mention a common misunderstanding, related to the

one noted above, that may lie behind the view that Kant is unable to recog-

nize that acting from duty with pleasure is better than acting from duty with

gritted teeth. It is sometimes supposed that as Kant sees it, someone who

acts from duty has as his purpose simply ‘‘duty’’ (whatever that would

mean). On this mistaken view, whereas the person who helps another just

because he wants to cares about the other’s happiness, the person who helps

from duty is indifferent to it, caring only about doing his duty. The first is

seen as having as his purpose to help the other; the second is seen as having

as his purpose only to do his duty. The temptation to draw a sharp division

between caring about doing one’s duty and caring about the content of that

duty can be explained only by an error noted earlier, that of thinking that

duty (Pflicht) in Kantian ethics is something imposed on us from without. That error may lead us to think that to care about doing one’s duty is to care

about pleasing an authority figure, getting ‘‘brownie points,’’ or maintain-

ing social respectability. But since duty is something that the agent ra-

tionally endorses, it would hardly make sense for someone to care about

doing his duty but to be indifferent to the content of the particular duty.

A careful reading of the text makes it plain that someone who helps from

duty does not care only about doing his duty. Kant emphasizes that what

distinguishes an action done from duty from an action done from inclina-

tion is not its purpose, but the maxim by which it is determined—that is,

why the person chose so to act (G 4:399). Both the agent who acts from duty

and the agent who acts from inclination want to help the person they are

helping, but only the former sees the person’s needs as making a moral

claim on him.∞≤

Still, one might think that because one person is drawn to the action by

emotion, and the other is not, only the former will help another with plea-

sure. But the fact that what led the agent to the action was not emotion or

feeling does not mean that one will lack emotion and feeling when she acts.

As Christine Korsgaard explains,

Once you have adopted a purpose and become settled in its pursuit,

certain emotions and feelings will naturally result. In particular, in ordi-

nary circumstances the advancement or achievement of the purpose will

make you happy, regardless of whether you adopted it originally from

Acting from Duty 99

natural inclination or from duty. So a dutiful person, who after all really

does value the happiness of others, will therefore take pleasure in mak- ing others happy.∞≥

Once again, the Metaphysics of Morals brings out more sharply than the Groundwork does that conducting oneself as one morally ought usually brings with it certain sentiments. The happiness of others is an obligatory

end. We have a duty, that is, to make others’ ends our own.∞∂ Someone

lacking in sympathy for others, someone unmoved by the plight of human

suffering, would seem—without some special explanation—not to have

adopted as an end the happiness of others. At the very least, the person

seems not to have integrated that end into her life in a way that develops the

kinds of feelings that Kant thinks are part of having, and seeking to pro-

mote, the end.∞∑

Now in unusual circumstances one’s ability to enjoy helping may be

marred. This is the situation of the man whose mind is so ‘‘clouded over with

his own grief’’ that ‘‘the distress of others does not touch him’’ (G 4:398).

But as noted above, this is not typical, and Kant constructed the example as

he did in order to isolate the moral incentive and to make it absolutely clear

that the man was acting from duty. Even in this case, however, there is

reason to hope that having torn himself ‘‘out of this deadly insensibility’’ (G

4:398), the man now does take pleasure in helping others. If he doesn’t yet,

he very likely soon will (if he doesn’t relapse into his insensibility and cease

to help others). Kant says of beneficence that if ‘‘someone practices it often

and succeeds in realizing his beneficent intention, he eventually comes actu-

ally to love the person he has helped’’ (MS 6:402).

III

I turn now to the second reason why Kant’s account of acting from duty and

moral worth is troubling even to those who understand that acting from

duty does not require that one lack an inclination so to act or that one not

take pleasure in what one does from duty. To recap, the first was that surely

it is better to enjoy helping the needy than not to; yet Kant shows no signs in

the Groundwork of recognizing this. We have seen, though, that when he discusses virtues and vices and duties to cultivate certain qualities of char-

acter, it becomes clear that he does not take the sentiments and attitudes

with which one acts to be a matter of indifference, and that he regards taking

pleasure in doing what is morally required to be a part of virtue. The second

100 Marcia Baron

objection is that as long as one acts with the right sentiments, does it really

add anything if the person acts from duty? If one helps others and does it not

for gain but from the kindness of his heart, why think that there is some-

thing wanting if he does not act from duty?

One reason to think that acting with the right sentiments—compassion,

sympathy, eagerness to help—is not full virtue is suggested by the follow-

ing quotation from Lord Shaftesbury, a philosopher writing before Kant’s

time (and not someone usually thought of as displaying Kantian colors):

If a Creature be generous, kind, constant, compassionate; yet if he can-

not reflect on what he himself does, or sees others do, so as to take notice

of what is worthy or honest; and make that Notice or Conception of Worth and Honesty to be an Object of his Affection; he has not the

Character of being virtuous.∞∏

This strikes a chord with Kantians, calling to mind the opening paragraphs

of the First Section of the Groundwork. The good qualities of temperament alone are only conditionally good; for ‘‘without the principles of a good will

they can become extremely evil’’ (G 4:394). Only the good will is uncondi-

tionally good. Without a good will one may be generous without any sense

of when or to whom one should be generous, or of the importance of

generosity. One may be self-controlled without a good sense of the impor-

tance of self-control—and to what ends one should control or redirect one’s

emotions. Generosity in the form of helping others to carry out their rob-

bery is not virtuous; nor is self-control in a situation in which, in order not to

jeopardize one’s social standing, one controls one’s impulse to speak out

against injustice.

Shaftesbury’s point is similar.∞π A person who enjoys—even loves—

helping others (and who does help, even at great cost to her well-being)

does not exhibit virtue if she does not reflect on her conduct—and the

conduct of others—and care about its ethical character. For her conduct to

count as virtuous, it has to matter to her that she acts virtuously.

Part of virtue, Lord Shaftesbury suggests, is having a sense of virtue.

The virtuous person has a sense of the moral (or, as many virtue ethicists

would prefer to say, ethical) dimensions of her conduct, and cares not only

about the welfare of others but also about virtue itself. Is Lord Shaftesbury

saying, then, that the virtuous person must have (and act from) a sense of

duty? Not quite. Let’s distinguish three conceptions of virtuous persons.

First, imagine someone who has all the right sentiments but no sense of

duty—no sense of being under a constraint to act as she does.

Acting from Duty 101

(1) The virtuous agent just does what is virtuous without any sense of

constraint about it. She does what is virtuous effortlessly, naturally, with-

out having to give any thought to the matter. This is simply her way of

doing things. If asked, ‘‘Why do you do that?’’ she would be at a loss.

‘‘That is just what I do,’’ she would say.

Both Lord Shaftesbury and Kant would reject this conception of a virtuous

person.∞∫ This would not constitute virtue.

Here is another picture of the virtuous person, one that contrasts with (1)

and does justice to Lord Shaftesbury’s point.

(2) The virtuous agent acts virtuously with an understanding that what

she is doing is virtuous, and she does so out of a recognition of the

importance of so acting. She sees it as making a claim on her; it is not all

the same to her whether she acts virtuously or not.

But there is still another way of thinking of the virtuous person, which

arguably also does justice to Lord Shaftesbury’s point.∞Ω

(3) The virtuous agent wants to be virtuous. She sees virtue to make a

claim on her but only because she wants to be virtuous, or cares about

virtue. Its making a claim on her is conditional on her caring about it.

Both (2) and (3) are consistent with Shaftesbury’s point, but only (2) is a

Kantian position. Why? Duty involves constraint. It is not just that one wants to do what is right, wants to do her duty; one feels and recognizes herself to be under a constraint. In (2) we see that constraint: she recognizes

virtue to make a claim on her, and to do so unconditionally. Its making a

claim on her is not conditional on her wanting to be virtuous or caring about

virtue. In (3), there is no such constraint; for virtue makes a claim on her

only conditionally on her caring about virtue.

Duty is not a matter of ‘‘affection’’ but of commitment. That is why I said

that it is not quite right to understand Lord Shaftesbury as saying that the

virtuous agent must have and act from a sense of duty. One acts from duty

not by virtue of doing what is right because one wants to do what is right;

rather, one does what is right because rightness makes a claim on one. One

recognizes that one should.

We are now in a position to see what is missing—part of what is missing,

anyway—in someone who acts with good sentiments but not from duty: a

commitment to doing what morality requires. One’s attitudes, emotions,

and desires are shaped and moderated by one’s commitment to doing what

102 Marcia Baron

morality requires. They—especially the emotions and desires—are not

always in harmony with it, even in the most virtuous person; and because of

this, even the most virtuous person needs such a commitment. The virtuous

person, like everyone else, needs a sense of duty.

It is instructive to see what, on Kant’s view, distinguishes good people

from others. This is elaborated in Religion within the Boundaries of Mere Reason. The difference between the good man and one who is evil cannot, Kant says, lie in the incentives that they incorporate into their maxims.

Rather, it ‘‘must lie in their subordination,’’ i.e., which of the two incentives he makes the condition of the other (R 6:36). The good and the evil alike

adopt into their maxims the moral law as well as other incentives. The

difference between them is that the good person subordinates the other

incentives to the moral law, while the evil person does just the opposite. The

good man makes pursuing what he fancies conditional on its being morally

permissible to do so, whereas the evil man makes abiding by morality’s

requirements conditional on this squaring with what he fancies.

We may want to take issue with Kant’s characterization of the evil

person. It seems too generous and benign a portrayal. My reason for draw-

ing attention to this passage, however, is simply to bring out that what is

crucial to a person’s moral goodness is that he or she puts morality first,

giving it priority over any competing considerations. This is the heart of

acting from duty.

Since Kant introduced the notion of duty to explicate that of a good will,

this should not surprise us. The goodness of good qualities of temperament,

talents of the mind, and gifts of fortune is conditional on the possessor’s

having a good will. As noted earlier, without a good will, such qualities as

courage and resoluteness can become ‘‘extremely evil and harmful’’ (G

4:393). Similarly, one’s conduct needs to be undergirded and shaped by a

commitment to doing what is right. Kindhearted acts, generous acts, acts of

loyalty and fidelity are good only insofar as they are guided and supported

by a sense of duty.

This brings us to the second component that is missing in someone who

acts with good sentiments but not from duty: recognition that the fact that

an act is kindhearted, or generous, or loyal does not ensure that it is right. In

addition to lacking a commitment to put morality first, the person who acts

with good sentiments but not from duty is not attuned to the possibility that

the benevolent act might happen to be impermissible.

Here is another way of getting at the same point. To suppose that some-

one could be fully virtuous simply by having whatever affections, desires,

Acting from Duty 103

and affective responses are ideal (including an affection for virtue), but

without any sense of duty, overlooks the fact that acts that issue from ideal

desires may nonetheless be morally imperfect. Compassion may lead us to

help when help is unfair to another, or when what we are helping the other

to do is morally wrong for some other reason.

To see the importance of both components, contrast the Kantian view to

a view that acting from duty is of value only as a backup, as a second best.

Ideally (according to the backup model), one would act from a desire,

unmediated by any sense of duty, to do the act in question or to bring about

the good result. It is only when someone is deficient in sympathy or fellow

feeling that the motive of duty is of value. It needs to ‘‘kick in,’’ on this

view, when the real thing, the feeling we would like everyone to act from, is

missing (or eclipsed by conflicting feelings, such as resentment or an-

noyance). If I weren’t so easily annoyed, if I didn’t nurse old grudges and

thus feel unwarranted resentment, if I weren’t lazy and self-centered and

thus inclined just to put myself first, I would have no need for the motive of

duty. I would act as I should without any sense of constraint, any sense that

whatever I might prefer to do, this is what I must do, for morality requires it.

This view could take either of two slightly different forms. In one form,

it holds that only those lacking in virtue need to act from duty. Virtuous

people do not. Their emotions, feelings, and desires—more generally, their

affective natures—are exactly what they should be. Always. So simply by

acting from their desires, virtuous people act as they should. In its other

(more modest and more plausible) form, the position is that even the most

virtuous sometimes are in need, motivationally, of a sense of duty. Even the

most virtuous became weary or afraid, and do not always act as they should

without the help of the thought ‘‘I really should do this.’’

The Kantian view is closer to the second, more modest position, but still

is in fundamental disagreement with it. It takes issue with the idea that it is

only because of a deficiency—only because even the best of us are some-

times weary or afraid—that purely moral motivation is of value. The per-

son who just wants to do x (which is morally required) and does not recog- nize that she ought to do x, fails to acknowledge the moral importance of x, and specifically, that x makes a moral claim on her. In addition, there is an implicit assumption in the view that acting from

duty is of value only as a backup that needs to be challenged. The assump-

tion is that a perfect affective nature would guarantee perfect conduct. An

example will illustrate why this is not the case. Imagine someone tempted

to favor someone close to him, but who refrains from doing so because

104 Marcia Baron

under those circumstances it would be unfair to others. The circumstances

might be that one’s child is upset about something and wants a great deal of

attention over several days, and asks not to have to go to day camp in order

to spend more time with her parents. The parents, however, have obliga-

tions to students and colleagues (and let us imagine that what is owed to the

students—attendance at dissertation orals, preparation for the orals, con-

sultation with the students—cannot be postponed without detriment). In

some situations it might be possible to grant the child her wish and still do

right by the students and colleagues, but let us imagine that this is not the

case. My claim is that it may well be the case that a perfect affective nature

would pull one in a direction different from rightness: one’s emotional

leaning is toward one’s child (depending on her age, and perhaps on why

she is upset), but one acts against that leaning because on reflection it seems

clearly to be wrong to request that the dissertation orals be postponed, or

simply not to prepare for the orals (and to take the child along to them).

The same point emerges if we picture the situation a bit differently: it

might be the case that one thinks that the child should not rely on the parent

so much for comfort, and is better off attending the day camp; here, too,

affect, in one with an ideal affective nature, might be at odds with one’s better

judgment. Affect might incline one toward granting the child’s request,

while reason says, ‘‘No, the child is using this as a crutch.’’ Arguably it is part

of love—at least love for one’s children—to want to protect and to soothe;

someone who never felt this in a situation in which she or he saw that

protecting or soothing was not possible or was contraindicated would proba-

bly be a rather cold individual. Yet in some situations in which everyone

except the cold and unfeeling will want to soothe and to shelter, it will be

right to refrain from sheltering (and to take care not to overdo the soothing).

The point here is that there is no reason to assume that affect in one with an

ideal affective nature will never be at odds with what is (and what one sees to

be) right. One need not be affectively less than perfect in order for one’s

emotions and feelings not to line up perfectly with what is, and what one sees

to be, right. For the same reason, a perfect affective nature does not guarantee

perfect conduct. Hence the need for a sense of duty—for a concern to do

what is right, even if one’s desires are impeccable.

IV

I have argued that acting from duty is compatible with having an inclination

so to act and with acting with pleasure. Furthermore, although Kant doesn’t

Acting from Duty 105

say so in the Groundwork, it is clear in the Metaphysics of Morals that he holds that acting with pleasure is not only compatible with acting from duty,

but is a part of being virtuous. But although it is part of being virtuous, there

is still something for Schiller to find objectionable. Virtue in Kant’s theory involves overcoming obstacles. One need not despise one’s friends or de- spise what one is doing in order for one’s conduct to count as an action done

from duty. And one needn’t despise one’s friends in order for one’s benefi-

cent actions toward them to be virtuous. Nonetheless, virtue involves con-

straint and the recognition of constraint. Only those beings capable of

acting contrary to reason (i.e., all human agents) are subject to moral con-

straints. ‘Oughts’ apply, indeed make sense, only for such beings. ‘‘[F]or a

holy will, no imperatives are valid; the ought is out of place here, because the volition is of itself already necessarily in harmony with the law. Hence imperatives are only formulas expressing the relation of objective laws of

volition in general to the subjective imperfection of the will of this or that

rational being, e.g., to the human being’’ (G 4:414). Virtue, Kant writes in

the Metaphysics of Morals, ‘‘is the strength of a human being’s maxims in fulfilling his duty.—Strength of any kind can be recognized only by the

obstacles it can overcome, and in the case of virtue these obstacles are

natural inclinations, which can come into conflict with a human being’s

moral resolution’’ (MS 4:394).

Should we be troubled by this? Should we prefer Aristotle’s conception

of the virtuous person as free of inner conflict, free of contrary inclina-

tions?≤≠ There is something appealing about the Aristotelian conception.

We admire the person who would never be tempted to act wrongly; we

regard such a person as a saint, as possessing some remarkable inner good-

ness.≤∞ But there is greatness in those who had to overcome obstacles—

obstacles coming from within the personality of the individual—to become

the exemplary people they became. (Think of Eleanor Roosevelt. Would we

admire her more if she had not had a fear of public speaking and other

obstacles from within that she had to overcome in order to carry out her

extraordinary work against social injustice?) Of course, whether we admire

people more or less for inner struggles that they had to overcame—and

in some instances never fully overcome—depends in part on what those

obstacles are. Someone seems less, not more, virtuous, if to be virtuous he

has to struggle against a desire to torment others. (And the point is even

clearer if the desire he had to overcome was a desire to rape, mutilate,

murder for the thrill of it, and so on.) We are very glad that he overcame it,

and we may admire him for having overcome it or, if the desire remains, for

never giving in to it (although it seems a bit odd to speak of admiration

106 Marcia Baron

here), but surely it would be a very rare person who saw him as more

virtuous for having had those obstacles. This complication can be addressed

by stipulating that although virtue is exhibited only when the agent finds it

hard to do the virtuous act in question, it matters what makes it hard. Not

every obstacle qualifies.

Setting aside that complication, suppose we do hold that virtues are

exhibited only when the person finds it very hard to do the act in question.

We would hardly hold—nor does logic require us to hold—that people who

do not find it hard should try to bring it about that they do.≤≤ If it is the case,

in other words, that virtue is shown only by overcoming obstacles, it does

not follow that we should try to put obstacles in our path in order to be able

to qualify as virtuous.≤≥ This brings into relief another error in Schiller’s

reasoning. If it is true that I do not act virtuously if I have no contrary

impulses to overcome when I serve my friends, it does not follow that I

should try to despise them in order to qualify as virtuous. It might seem to

be the right thing to do if my goal were to earn as many moral brownie

points as possible, i.e., to qualify as acting virtuously as often as possible,

but that would be self-defeating. By making that my goal, I would be

making myself less virtuous. In any event, Kant does not take the position

that we are to perform as many morally worthy acts as possible. Nor is

virtue additive. It is not a function of how many morally worthy acts we

perform (or of the proportion of our actions that are morally worthy). We are

to strive to perfect ourselves morally—to make the thought of duty an all-

sufficient motive, so that we never think to ourselves, ‘‘It is wrong, but

should I do it anyway?’’ or ‘‘I see it is what I morally ought to do, but even

so, what about the tempting alternative?’’ Perfecting ourselves morally also

involves setting ends for ourselves, ends that are subsidiary to the two

obligatory ends (the happiness of others and our own perfection) as well as

cultivating our characters so that we lack, as far as possible, the various

vices Kant lists and have, as far as possible, the virtues.

I see no reason to be troubled by Kant’s conception of virtue. Others may

prefer the Aristotelian paradigm, but this ‘‘no inner conflict’’ model seems

to me to require us to swallow a lot harder. Virtue, on the Aristotelian

model, would have far less to do with personal effort and accomplishment,

and thus seems to reflect very little on the agent. It reflects, instead, either

luck or an upbringing in which one has been conditioned to take pleasure in

exactly what one should take pleasure in, and not in anything else. In

addition, the model of virtue is an otherworldly one: to strive to be virtuous

is to strive to be something a human cannot be, a being that either has no

affective nature at all, or whose affects are invariably in accord with reason

Acting from Duty 107

(with what duty requires). On Kant’s view, we are to aim to be virtuous

humans, not to be divine. This is as it should be.≤∂

notes

1. Friedrich Schiller, Xenien ‘‘The Philosophers,’’ in Goethe, Werke, vol. 1, ed. Erich Trunz (Munich: Beck, 1982). I am using, with very slight

modification, Allen Wood’s translation of the passage in his Kant’s Ethical Thought (Cambridge: Cambridge University Press, 1999), p. 28. 2. And to understand the structure of the entire First Section, keep in

mind that through explicating the concept of a good will, ‘‘to be esteemed in

itself and without any further aim,’’ Kant introduces the idea of a categori-

cal imperative (although he does not introduce that terminology until the

Second Section). ‘‘Since I have robbed the will of every impulse that could

have arisen from the obedience to any law, there is nothing left over except

the universal lawfulness of the action in general which alone is to serve the

will as its principle, i.e., I ought never to conduct myself except so that I could also will that my maxim become a universal law’’ (G 4:402). 3. I should clarify that ‘‘subjective limitations and hindrances’’ qualify

the concept of the will, not the way this concept is contained in the concept

of duty. Thanks to Dieter Schönecker for pointing out the need to clarify this

point. To understand why the concept of duty contains that of a good will,

see G 4:413–14, bearing in mind that a duty is what one ought to do or

(equivalently for Kant) what one would do if one were fully rational.

4. The passage I am citing is obscure, however, and no interpretation I

have thought of or read seems quite satisfactory. Because of its obscurity,

this piece of textual evidence seems to me less decisive than it did when I

wrote Kantian Ethics Almost without Apology (Ithaca: Cornell University Press, 1995), p. 150. For a different reading of the passage, see Wood,

Kant’s Ethical Thought, chap. 1. 5. In Kantian Ethics Almost without Apology I argue that on Kant’s

view, acting from duty is not compatible with acting from inclination, but is compatible with the agent’s having such an inclination, provided that the inclination is not part of the ground of the agent’s action. Presence of

inclination is consistent with the action’s being done from duty, and thus

having moral worth, but an action cannot qualify as an action done from

duty if it is done from inclination. This view is also taken by Henry Allison

in Kant’s Theory of Freedom (New York: Cambridge University Press, 1990).

6. See G 4:414.

108 Marcia Baron

7. For other discussions of Stoicism and Epicureanism, see The Critique of Practical Reason (KpV 5:116) and Religion within the Boundaries of Mere Reason (R 6:57–59). 8. It is ‘‘not in itself a duty to share the sufferings (as well the joys) of

others,’’ but ‘‘it is a duty to sympathize actively in their fate; and to this end

it is therefore an indirect duty to cultivate the compassionate . . . feelings in

us, and to make use of them as so many means to sympathy based on moral

principles and the feeling appropriate to them’’ (MS 6:457).

9. Of gratitude Kant writes that it requires that one not regard ‘‘a kind-

ness received as a burden one would gladly be rid of’’ but instead take ‘‘the

occasion for gratitude as a moral kindness, that is, as an opportunity given

one to unite the virtue of gratitude with love of man, to combine the cor- diality of a benevolent disposition with sensitivity to benevolence (atten- tiveness to the smallest degree of this disposition in one’s thought of duty),

and so to cultivate one’s love of man’’ (MS 6:456).

10. See Allen Wood, ‘‘The Final Form of Kant’s Practical Philosophy,’’

Southern Journal of Philosophy 36, suppl. (1997), 1–20. 11. However, as Kant explains in the Metaphysics of Morals, in one

important way some of them do ‘‘lie at the basis of morality’’: ‘‘moral feeling, conscience, love of one’s neighbor, and respect for oneself . . . are subjective conditions of receptiveness to the concept of duty.’’ They are not ‘‘objective conditions of morality,’’ but ‘‘every man has them, and it is by

virtue of them that he can be put under obligation’’ (MS 6:399).

12. See Christine Korsgaard, Creating the Kingdom of Ends (New York: Cambridge University Press, 1996), p. 58.

13. Ibid. p. 59. I have altered the text to correct a typographic error.

Although the original publication in the Monist 72 (July 1989) has ‘‘ad- vancement or achievement,’’ instead of ‘‘or’’ we find ‘‘of’’ in the more

recent publication.

14. See MS 6:385–88.

15. I am grateful to Lara Denis for encouraging me to bring out this

point.

16. Anthony Ashley Cooper, Earl of Shaftesbury, ‘‘An Inquiry concern-

ing Virtue or Merit’’ (1699; reprinted 1732), in British Moralists, ed. L. A. Selby-Bigge (1897; reprint, Indianapolis: Bobbs-Merrill, 1964).

17. But there are also dissimilarities. In addition to the one noted just

below in the text, there is another, depending on how much rests on Shaftes-

bury’s choice of the word ‘reflect’. If Shaftesbury has in mind philosophical

reflection of a sort that only the well-read, well-educated, or intelligent

could engage in, Kant would disagree with the quoted passage. As Allen

Acting from Duty 109

Wood reminds me, Kant is very reluctant to think of well-educated, philo-

sophical people as having any moral advantage over those who lack the

occasions and tools for moral reflection. Recall G 4:404: ‘‘it needs no

science and philosophy to know what one has to do in order to be honest

and good, or indeed, even wise and virtuous.’’

18. So would Aristotle. See Nicomachean Ethics II.6 as well as Eude- mian Ethics VII.15. Jennifer Whiting emphasizes the similarity between Aristotle and Kant on this and related points in her ‘‘Self-Love and Authori-

tative Virtue: Prolegomenon to a Kantian Reading of Eudemian Ethics viii 3,’’ in Aristotle, Kant and the Stoics, ed. Whiting and Stephen Engstrom (New York: Cambridge University Press, 1996).

19. I say ‘‘arguably’’ because I might be putting too fine a point on

Shaftesbury’s choice of the word ‘Affection’. Since my concern is with

Kant, not with Shaftesbury, I will not pursue that possibility here.

20. Aristotle, Nicomachean Ethics VII. 21. This may exhibit a misapprehension, on our part, of saintliness. See

R. M. Adams, ‘‘Saints,’’ Journal of Philosophy 81 (1984): 392–401. Adams takes issue with the characterization of moral excellence in Susan Wolf’,

‘‘Moral Saints,’’ Journal of Philosophy 79 (1982): 419–39. He suggests that there is considerably more depth and complexity in saints than Wolf

supposes, and, most relevant for my purposes, more inner struggle.

22. As Allen Wood observes in Kant’s Ethical Thought, ‘‘Needless to say, Kant does not advocate that we cultivate hatred of our friends or other

contra-moral desires in order to give ourselves opportunities to resist them.

(That would make about as much sense as putting your loved ones in

needless danger so as to give yourself the opportunity to display your

courage by rescuing them)’’ (p. 29). We should not, however, be all that

surprised if some readers think that Kant holds this; indeed, some might not

even regard it as an absurd view. After all, some people try to justify the

existence of evil—i.e., explain why an omnipotent, benevolent deity would

have created a world with so much human suffering and other evil, or

allowed such evil and suffering to arise—by saying that were it not for the

existence of evil, there would be no opportunity for us to exhibit (heroic)

virtue in battling it; likewise, were there no human suffering, there would be

no opportunity to show one’s moral mettle by seeking to ameliorate it.

23. I once met an erstwhile rescue worker who told me there was noth-

ing he had loved so much as the excitement of emergency calls. He spoke

rapturously about the adrenalin rush he had felt when hurrying to the scene,

not knowing what it would hold. It would be silly for him to rid himself

(even if he could) of this thrill in going to accident scenes, even though it

110 Marcia Baron

probably is true that it is more admirable—inspires more esteem—if some-

one takes part in rescue efforts despite finding them distressing, and takes

pleasure only in actually helping.

24. I am very grateful for helpful comments from Lara Denis, Dieter

Schönecker, Allen Wood, and Santiago Zorzo.

Kantianism for Consequentialists

SHELLY KAGAN

Kant’s moral philosophy represents one of the most significant approaches

to the foundations of ethics. For obvious reasons—including the simple

fact that Kant offered no distinctive name for his general approach to ethics

—views of this same, basic sort are typically known as Kantian. But this

common practice, natural as it is, carries with it an obvious danger as well:

there is a temptation to assume that Kant himself is the last word on Kant-

ianism, rather than merely being an important advocate of this sort of view. This can lull us into overlooking the possibility that in various places Kant

may have been mistaken about the implications of Kantianism; and it can

also make us feel needless pressure to reconstruct Kantianism in precisely

the terms in which Kant himself presented it. As a result, we may narrowly

focus on the details of Kant’s particular views, at the expense of appreciat-

ing the fuller significance and general interest of Kantianism. (In contrast,

we are quite used to thinking of Bentham, Mill, and Sidgwick as merely

being leading representatives of the general utilitarian approach, without thinking that any one of them has the last word on utilitarianism itself.)

In this paper I want to discuss one significant strand of Kantianism in

ethics. I focus, in particular, on certain ideas put forward in the Groundwork for the Metaphysics of Morals. But I must emphasize the point that the ideas I will be discussing are primarily put forward here as being Kantian, rather

than Kant’s. The position I will be discussing is certainly inspired (at a

minimum) by Kant’s own discussion in the Groundwork, and I will periodi- cally turn to the text of the Groundwork itself for guidance and comparison. But this essay is not intended as a piece of Kant scholarship. Rather, it is

intended as a contribution to understanding Kantianism. Indeed, because of

this, I will hereafter refer to kantianism (rather than Kantianism) where the lower case ‘‘k’’ is intended to mark the idea that I am primarily interested in

the type of approach that Kant represents, rather than Kant exegesis per se. What I want to do, then, is to sketch the basic elements of a possible kantian

112 Shelly Kagan

approach, and indicate why I think the view has abiding significance for

moral philosophy.

My primary goal is expository. I hope to say enough to make it clear why

kantianism is worth taking seriously—even by those who may, at the end of

the day, choose not to accept it. It is not my intention to offer anything like a

full presentation of kantianism (we will only be considering a few of the main ideas discussed in the Groundwork), nor is it my intention to offer anything like a full defense of kantianism. While I hope to say enough to show why one might find kantianism attractive and plausible, the argu-

ments I offer are only rough sketches, and many important objections will

go unanswered (or unmentioned).

I have a secondary goal as well, reflected in my choice of title. Kant

himself believed that kantian foundations supported a deontological rather

than a consequentialist normative theory.∞ Since most philosophers have

assumed that he was right about this, those sympathetic to consequentialism

have typically had little interest in understanding kantianism. But in fact it

is far from clear whether Kant was right about this.≤ So I hope to offer an account of kantianism that consequentialists may find congenial. In any

event, if I am right in thinking that kantian foundations are themselves

fairly plausible, then it behooves those who want to reject those foundations

to identify exactly where they think those foundations go wrong. (Of

course, given the obscurity of much of Kant’s writing, it may not be surpris-

ing that few consequentialists have actually attempted to do this.) Accord-

ingly, I want to offer a guide to kantianism that may be of particular use to

consequentialists.

But this further goal is indeed only secondary. My primary purpose is to

sketch the main lines of a potentially attractive version of kantianism.

Questions about the particular normative implications of kantian founda-

tions can be put aside until we have a better handle on the kantian founda-

tions themselves.

I . Autonomy and the Formula of Universal Law

Where then should we begin? Kantianism begins with freedom. More par-

ticularly, it begins with the fact that we are free, and with an account of that

freedom. So we must begin with that account. (It is worth noting, however,

that the Groundwork itself does not begin with the idea of freedom, but rather works backward toward it, arguing in the first two sections that if

there is to be such a thing as morality, then we must be free—that freedom

Kantianism for Consequentialists 113

is the basis of morality. Unfortunately, in the Groundwork itself Kant says rather little explicitly about how exactly we are supposed to be able to move

from the assumption of freedom back ‘‘up’’ to morality (see G 4:446–47).

Thus we must depart from Kant exegesis almost immediately.)

Kantianism begins with freedom. But I think we will better understand

the relevant notion of freedom if we begin instead with rationality. What,

exactly, is it to be rational?

Suppose we start with theoretical rationality. As a theoretically rational

being, I am capable of examining my various beliefs and seeing whether it

makes sense for me to hold them. Thus, in the first place, I have standards for evaluating beliefs, in the light of which I can ask whether or not I am

justified in holding a given belief. I might, for example, appeal to various

principles of logic, discovering that some of my beliefs commit me to

accepting still other beliefs; or I might appeal to various rules of scientific

methodology, finding that, given the available evidence, I am unjustified in

accepting some further belief. But rationality in the theoretical domain goes

beyond the mere evaluation of my beliefs: I can change my beliefs in light of my judgments concerning the extent to which they meet (or fail to meet)

the relevant standards. Normally, that is, when I see that the evidence better

supports one claim rather than another, my beliefs change accordingly.

Roughly, then, theoretical rationality consists in my ability to evaluate my

beliefs in light of the standards relevant for evaluating beliefs, and to alter

my beliefs in the light of those evaluations.

Practical rationality is similar. As a practically rational being I am capa-

ble of examining my various desires, goals, intentions, actions, and the like,

so as to see which of these make sense in the circumstances. Here too, then,

I have standards in terms of which my plans can be evaluated, goals as-

sessed, actions endorsed or criticized. Nor are these various practical ele-

ments merely subject to evaluation; I can change my goals, my intentions, and the like, in light of my judgments concerning the extent to which these

meet (or fail to meet) the relevant standards. Thus practical rationality

consists in my ability to evaluate actions, intentions, and so forth, in light of

the standards relevant for these, and to alter these elements in light of those

evaluations.

Generalizing, then, we can say that rationality—whether practical or

theoretical—consists in the ability to evaluate beliefs and acts (and so

forth) with an eye to whether they meet the relevant standards, and to alter

our beliefs and acts in light of those evaluations.

In this way rationality goes beyond mere intelligence. Nonhuman ani-

mals, I presume, also have beliefs and desires, and act in a way that is often

114 Shelly Kagan

appropriate to their circumstances. Some animals may well be extremely

adept at achieving goals and forming appropriate beliefs about their en-

vironment. Thus they display varying (and perhaps considerable) degrees

of intelligence. We could say that intelligence consists in the ability to produce beliefs and actions that in point of fact are appropriate (that is,

conform to the relevant standards); animals are often intelligent in this

sense. But only rational creatures are capable of articulating the standards against which beliefs and actions are to be evaluated, and only rational

creatures are capable of consciously comparing beliefs and actions (real or imagined) against those same standards (G 4:412, 427).

It is worth emphasizing as well the point that as rational beings we are

capable of rejecting the beliefs and actions (and the like) that don’t meet what we take to be the relevant standards. We modify our behavior and our

beliefs in light of what we think appropriate. For example, we are not

normally forced to act on desires that we happen to have, when we conclude that such desires don’t make sense, or that acting on them in present circum-

stances would be inappropriate (by whatever standards we here take to be

relevant). In this way, too, rational beings are different from merely intel-

ligent animals. For it seems plausible to view animals as mere ‘‘playthings’’

or ‘‘puppets’’ of their desires—incapable of evaluating them, and thus

incapable of rejecting them. In contrast, rational beings are in an important

sense free: if we conclude that a given desire makes no sense (perhaps we recognize that it was based on what we now see to be a mistaken belief ) or

that a given intention is inappropriate, we are free to step back from that

desire or intention, and to refuse to act on it.

Of course, the simple fact of the matter is that humans are not perfectly rational. At times we misapply our own standards and fail to see that a

belief cannot be justified (given the relevant standards). Or we may find

ourselves incapable of abandoning certain beliefs, even though we can see that these beliefs are not in fact justified. Similarly, at times we may find

ourselves giving in to desires, even though we see full well that acting on

this desire, in this situation, doesn’t actually make sense, or is otherwise

inappropriate. Thus we are, at best, only imperfectly rational. Still, it would

be implausible to suggest that we are not rational at all (in this sense), for we

clearly are capable of articulating standards for evaluating beliefs and ac-

tions, and we are typically capable of evaluating our beliefs and actions in

the light of those standards; and often, at least, we are capable of modifying

our behavior and beliefs in the light of those evaluations. Humans may not

be perfectly rational, but we are rational nonetheless, even if only imper-

fectly so.

Kantianism for Consequentialists 115

The account of rationality that I have been sketching is, indeed, only a

sketch. But even so, it remains significantly incomplete, in that I have not

yet drawn attention to an important further fact: not only are we capable of

articulating relevant standards, and evaluating and modifying beliefs and

actions in light of those standards; the standards themselves are things that we can evaluate and modify. That is, for any given standard that I might use

to evaluate a belief or an act (or an intention, and so forth), I can ask of the

standard itself whether it makes sense, whether it is indeed an appropriate standard to be used in this way in these circumstances. In effect, I can ask

whether the given standard itself meets the standards (whatever they are)

relevant for evaluating standards. And armed with these evaluations, I can in turn reject any given standard, modify it, or replace it. Thus, as a rational

being I am free not only to reject, modify, or endorse my various beliefs and

actions—I am also free to reject, modify, or endorse the standards I appeal

to in evaluating beliefs and actions. I am not forced to accept and appeal to

standards that do not make sense to me or that seem unjustified or inap-

propriate. I am free to alter the standards as I see fit.

And the same is true, of course, with regard to the ‘‘second order’’

standards that I may use to evaluate the ‘‘first order’’ standards. These

higher order standards can themselves be subject to critical evaluation: I

can ask whether the standards I use for evaluating standards are themselves

appropriate, whether they themselves meet the relevant (‘‘third order’’)

standards (whatever I may take these to be) for evaluating such (second

order) standards. And I can modify these higher order standards as seems

appropriate in light of these further evaluations. And so on, and so forth, all

the way up (or all the way down): no standard is itself forced upon me, no

standard is immune to potential criticism or evaluation. I am free, in princi-

ple, to evaluate any standard whatsoever, to ask whether it makes sense to

me, whether it is indeed an appropriate standard to use. The principles or

standards by which I evaluate beliefs and actions are themselves subject to

rational assessment and open to modification or rejection. Put another way,

the rules of rationality are not forced upon me (against my will, as it were): I

need only appeal to standards that make sense to me, that seem appropriate

in light of whatever principles, rules, or standards I endorse.

Our examination of the nature of rationality has thus led us to an impor-

tant insight. The rules or standards to which I appeal in rationally assessing

beliefs and actions are themselves subject to rational assessment, and at no

point need I simply accept a relevant rule or standard as simply given—

from ‘‘out there,’’ as it were, forced upon me despite its making no sense.

On the contrary, the relevant rules or standards need only be accepted if

116 Shelly Kagan

they, too, make sense in light of whatever rules and standards I reasonably

accept. We could put the point this way: the laws of rationality are not

forced upon reason from the outside. Rather, reason is free to reject those

standards (at whatever level) that do not make sense to itself. Reason is its

own last court of appeal. It chooses what standards to obey. In short: reason

is autonomous (G 4:440). The fact that reason is autonomous in this way is certainly not altogether

obvious. Indeed, Kant believed that previous moral philosophers had failed

to recognize the autonomy of reason, and certainly had failed to appreciate

the implications of reason’s autonomy for ethics (G 4:432–33). Most moral

philosophies have been founded in heteronomous conceptions of reason,

where some ultimate principle of reasoning is simply taken as ‘‘given’’

(from outside reason’s control) and beyond question (G 4:441–44). But

kantians believe that since we are autonomous (insofar as we are rational),

all such approaches to ethics must fail. If there is to be any hope for a sound

foundation for ethics, it must take account of our autonomy.

In the account I have been sketching, the ideas of reason, freedom, and

autonomy are tightly connected. Clearly, much more needs to be said, both

in defense of the general kantian picture I have been presenting, and by way

of further clarification of the three related concepts. But I am going to

restrict myself here to two quick remarks.

First, our analysis of rationality has led us to a picture of rational beings

as free. So eventually the kantian must confront the question of whether the

freedom that we take ourselves to have (as rational beings) is genuine or a

mere illusion. Kant himself postpones the discussion of this issue until the

third section of the Groundwork, and even there the discussion is cursory. In this essay I shall make no attempt whatsoever to pursue this question.≥ I

believe it plausible to hold that we are free, in the relevant sense, but I won’t

attempt to defend this claim here. And so, along with Kant in the first two

sections of the Groundwork, we can view the rest of our discussion as taking the form of a conditional: if we are free, what follows? Second, I want to say a word more about the concept of autonomy. Kant

typically expresses the thought that reason is autonomous by saying that

reason is the author or source of the rules and standards used by reason (e.g., G 4:431). But it is not clear that our concept of rationality can take us

quite this far. Suppose we grant the kantian that the freedom involved in

rationality means that there are no sound or valid standards for rational

assessment that cannot themselves withstand the scrutiny of rational assess-

ment. This would mean that there are no valid rules of reasoning that reason

doesn’t itself ‘‘accept,’’ or ‘‘will,’’ or ‘‘approve.’’ We might capture this idea

Kantianism for Consequentialists 117

by saying that reason must itself ‘‘sign off’’ on any purported rules of

rationality that are themselves to be binding upon reason. (There are no

rules binding upon reason that reason wishes itself free of, no rules that it

considers unreasonable rules.) But is it also true that we must think of

reason as the author of these rules (the ground of their validity)? Kant apparently thinks so, though it is not clear why. Perhaps (and this is sheer

speculation) he believes that it is inexplicable how reason could have this kind of veto power over rules of reasoning (so that no rule it disapproves of

is valid) unless reason is itself the source of the validity of the (valid) rules of reasoning. This claim is not completely unattractive, and so I shall follow

Kant here in speaking of reason as the source or author of its own rules. I

believe, however, that this further claim is not strictly needed by the kant-

ian. So long as it is conceded that reason’s autonomy means that reason

must ‘‘sign off’’ on any principles of rationality if they are indeed to be

sound—that no standard for rational assessment is valid unless reason itself

can approve of it—the kantian has, I believe, all that he needs.

Now kantians believe that given the autonomy of reason, certain im- plications fall out concerning the rules or standards that reason can give to

itself. In particular, they believe that once we recognize the autonomy of

reason, we are committed to accepting a certain fundamental rule—the

universal law formulation of the categorical imperative (FUL). Here is a possible reconstruction of the main line of thought.

Whenever I act, my acting presupposes that there is reason to do what-

ever it is that I am doing, that my act makes sense in the given circum-

stances. In effect, each action presupposes some rule or principle (though

not necessarily the same rule from act to act) that endorses the act, a rule in

the light of which the act can be seen as reasonable. Typically, of course,

these underlying principles or rules will only be implicit, but were we to

make them explicit, they might say something along the following lines:

under such and such circumstances, given such and such desires or such and

such goals, there is reason to act in such and such a way. As I say, we rarely

make such rules explicit (and even less frequently attempt to state them

fully and with care), but whenever I act, I presuppose some such rule—a

rule which, if sound, would validate my action, by showing why it is that I

have reason to do whatever it is that I am doing. (In many cases, of course,

one acts spontaneously, or simply ‘‘goes with the flow.’’ But in such cases,

presumably, the principle implicit in one’s act is precisely one that endorses

acting spontaneously in circumstances of this sort.)

So when I act, I presuppose a rule or principle that claims that I have

reason to do what I do (given the circumstances, and so forth). But which

118 Shelly Kagan

rules should I act on? This much seems clear: I should only act on rules that

are themselves valid. (The precise term of commendation used here isn’t important for our purposes. We could equally well talk of those rules that

are sound, or legitimate, or good, or reasonable.) I should only do what it

truly makes sense for me to do; so I should only act on those rules that are

themselves correct in their claims about what it is that I have reason to do. I

should only act on those principles that are valid.

But given that I am autonomous, the rules are up to me. Valid rules are

valid by virtue of my signing off on them, by virtue of my approving of

them as a rational being.

So this means: I should only act on rules that I can sign off on. I should

only act on rules that I can rationally choose to be rules. Put in slightly

different terms: I should act only upon rules that I can (rationally) will to be rules.

But rules are laws. They tell everyone what to do (or believe, or intend, and so on) in relevant circumstances. They say, for example, that in such

and such circumstances, given such and such desires, one has reason to

perform an act of such and such a type. But this means (if the given rule is

valid) that everyone has such a reason—provided that they have the rele- vant desires and find themselves in the relevant circumstances. Of course

not everyone will necessarily find themselves in the relevant circumstances,

or with the relevant desires—but it is true of everyone that if they were in the relevant circumstances (and so forth) then they would have reason to perform an act of the relevant sort. Rules are universal, providing the same reasons (under the relevant circumstances) to everyone.

So we can restate our earlier conclusion. Instead of saying that I should

act only upon rules that I can (rationally) will to be rules, we can say: I

should act only on those rules that I can (rationally) will to be universal

laws.

This is Kant’s formula of universal law, though his own favored state-

ment of it makes use of a piece of jargon. Kant typically talks about maxims, which for our purposes we can take to be first person statements of inten-

tions (‘‘I will perform such and such an act in such and such circumstances,

given such and such goals’’). Each such maxim corresponds to an implicit

principle (‘‘if one is in such and such circumstances, with such and such

goals, then one has reason to perform such and such an act’’), and so we

could restate the formula at which we have arrived as follows: I should act

only on those maxims, where I can (rationally) will that the corresponding

principle be universal law. Simplifying a bit further still, we can say: act

only on those maxims that I can will to be universal laws. And this is

Kantianism for Consequentialists 119

exactly what Kant tells us. Here is his own statement of the formula of

universal law:

FUL: ‘‘Act only in accordance with that maxim through which you at the

same time can will that it become a universal law’’ (G 4:421).

Kant’s decision to state FUL in terms of maxims rather than the corre-

sponding principles carries certain risks, for one can normally state one’s

intentions in a way that only gives a partial indication of what one takes

oneself to be doing, and why it seems to make sense. Thus, for example, if

my intention is to close the door to keep out the person attacking me, so as

to save my life, it will normally be correct to say, as well, that I intend to

close the door. But if we then focus on ‘‘I will close the door’’ as a statement

of my maxim, we will have no idea (or at best a poor idea) of why I think it

makes sense to do this in the present circumstances, and thus no idea (or at

best a poor idea) of just what the corresponding principle is supposed to be

that I am to examine so as to see whether I can indeed will it to be universal

law. These problems could have been avoided had Kant stated FUL directly

in terms of examining complete statements of the underlying principles.

But so long as we bear in mind that the real question is always whether a

purported reason-giving principle is indeed one that we can rationally will

to be universal law, we should be able to make use of Kant’s own formula-

tion without too much confusion.

Now the argument I have just sketched moves from our autonomy to

FUL, a requirement to act only on certain types of maxims (in Kant’s

formulation). But if this argument is sound, then the resulting requirement

should apply equally to everyone, that is, to every rational being. For if reason is autonomous, and autonomy yields FUL, then FUL is binding upon

all rational beings. That is to say: all rational beings should obey FUL; they

must do it if they are to act rationally. We can express this point in kantian jargon by saying that FUL is a categorical imperative (one binding upon all rational beings; see, e.g., G 4:432). Of course this does not mean that all

rational beings will obey FUL. As we have already noted, humans, at least, are only imperfectly rational, and thus may often fail to conform to FUL,

sometimes knowingly. But everyone should obey FUL: they have reason to do so, based on the mere fact that they are rational. If the argument is sound,

then FUL is a categorical imperative.

Kant says there is exactly one categorical imperative, though it has

several equivalent formulations (G 4:420–21, 436). FUL is supposedly

only one of the different ways of stating this single imperative. Another of

the formulations, the formula of autonomy (FA), goes like this:

120 Shelly Kagan

FA: ‘‘the idea of the will of every rational being as a will giving univer-

sal law’’ (G 4:431).

Note that Kant doesn’t even bother to state this version in the form of an

imperative at all! Presumably, however, what he is most concerned to im-

press upon us here is the idea that it is autonomy (the fact that reason is the

source of its own laws) that provides the basis for FUL: given the former,

we are led to the latter. The argument I have been sketching tries to make

good on this thought. (To get full equivalence, of course, we would also

need to go on to argue as well that given FUL we can derive an imperative

along the lines of ‘‘Act autonomously!’’ or ‘‘Act in keeping with your

autonomy!’’ I won’t attempt to argue that here.)

But is the argument sound? Can we actually derive FUL from the mere

assumption of reason’s autonomy? I am not sure. Doubtless several steps of

the argument could be questioned, but the most important issue, I believe, is

this. Is it really true that the only rules or standards that I could autono-

mously will are universal? Must the reason-giving principles I endorse be principles that would equally give everyone a reason? Putting the same point in a slightly different way, is it really true that the only rules that I

could freely give to myself are rules that make similar prescriptions for

everyone? Unless something like this is true, then all that autonomy will

demand is that I act on maxims that I can (autonomously) will. We won’t

have a requirement that I act only on maxims that I can will to be universal

law. And so we won’t have made it all the way to FUL. So we need to ask: is

it really true that the only principles I can autonomously give myself are

universal?

Now it might seem that the answer to this question is obvious. For it

seems obvious that I can (and should!) endorse principles that recognize

that what I have reason to do normally differs from what you have reason to do. For example, I may have reason to eat right now, while you do not.

In thinking about this question, however, it is important to bear in mind

the point, already noted, that the requirement that the reason-giving princi-

ples be (ones that I can will to be) universal laws only amounts to a require-

ment that people in the same circumstances have the same reasons. Thus, universality here only amounts to the requirement that if someone else were in the same circumstances (that is, whatever the principle takes to be the

relevant circumstances) then they too would have reason to perform the

same kind of act. (And it should be noted that, depending on the given

principle, the relevant circumstances may well include a specification of the

person’s desires or goals as well as more ‘‘external’’ circumstances.) So

Kantianism for Consequentialists 121

even if the principles I give myself are universal, this doesn’t mean that

everyone has reason to do the same specific types of acts, for people will

still find themselves in differing circumstances.

In typical cases, at least, when we find ourselves thinking that one

person has reason to do something that another person does not, this will be

because we think there is some relevant difference in their circumstances—

and a full specification of the relevant reason-giving principles will take

note of these circumstances. Thus, for example, I may believe that I should

eat, while you should not, but this may be because I believe that only

hungry people should eat, and I recognize that you are not hungry. (Or

perhaps I believe that people on diets shouldn’t eat between meals, and you

are on a diet and it is between meals; or that you need to get to a class, and I

do not, and so on.) Despite initial appearances to the contrary, then, the

underlying principle will actually be universal: anyone who is similarly

situated (with regard to hunger, dietary needs, availability of food, more

pressing demands, and so forth) will have similar reason to eat (or not). If

this is right, then at the very least most of the principles I can actually sign

off on will indeed be universal laws in the relevant sense.

But is it truly impossible for me to autonomously will principles that are not in this way universal? Can’t I simply endorse a rule that says that I (but not others) should do such and such an act in this case (but not in other cases that are otherwise similar)?

Here I can only reply that when I honestly contemplate such irreducibly

person specific or irreducibly case specific principles I find them virtually

unintelligible. I cannot fathom the idea that I might have reason to do

something in a certain kind of case, while you do not—even though there is

not a single relevant difference between us. This is not to say that I can’t

imagine someone ‘‘stating’’ such a principle, nor do I mean to claim that I

wouldn’t understand what someone affirming such a principle would be

attempting to do. Rather, I simply find that I cannot take seriously the

possibility that such a principle would be one that merits endorsement. If in

the circumstances someone has reason to act in a given way, then it seems to me that anyone at all who genuinely found themselves in relevantly

identical circumstances would have reason to act in the same way. Which is to say, when I ask myself what sorts of reason-giving principles I can truly

imagine autonomously giving to myself—fully accepting upon complete

rational reflection—the only such principles are ones that are universal.

In my own case, then, if I am indeed to restrict myself to maxims that I

can autonomously will, then I must restrict myself to maxims that I can will

to be universal laws. Perhaps others differ from me in this regard. The idea

122 Shelly Kagan

seems just barely possible, though, again, I find that I can’t take the thought

seriously. As far as I can see, any rational being would find that the only maxims he could autonomously will would be maxims that he could will to

be universal laws. And if this is indeed correct, then given the autonomy of

reason something like FUL may well follow for all rational beings what-

soever. In short, FUL may indeed be a categorical imperative.

To be sure, other questions about this step of the argument could be

pressed, and other stages of the argument could be challenged as well. So I

would not want to claim that the validity of the derivation of FUL (from the

assumption of autonomy) has now been established. But I hope that I have

said enough at this point to make it clear why the kantian’s appeal to FUL is

a position worth taking seriously. The claim that reason is autonomous is, I

think, a plausible one, and the further claim that autonomy yields FUL is

not, I believe, one that can be easily dismissed. If nothing more, these

claims are sufficiently plausible (even if one ultimately rejects one or the

other of the pair) that what I have said should make it clear why many

people have found FUL so compelling.

II . Understanding the Formula of Universal Law

Suppose, then, that we grant the kantians the validity of FUL (if only for the

sake of argument). Even if we do this, it is hardly obvious how FUL is to be

applied, how it is to be put to work. Nor is it the least bit obvious whether—

as kantians believe—FUL has sufficient ‘‘bite’’ that it can be used to gener-

ate concrete moral guidance. So let us put aside further questions about the

derivation of FUL, and turn instead to the question of what follows from it.

Granted that I must only act on maxims that I can will to be universal law,

how exactly am I to decide what to do?

The first thing to notice is that FUL itself doesn’t actually provide us

with maxims; it only serves to rule some of them out. We bring candidate

maxims to FUL, to see whether they are acceptable. The point here is easy enough to grasp if we recall that maxims are, in effect, statements of what

one intends to do in a given situation. What we should imagine then is that faced with the given situation, I have come up with some tentative plan of

action, something that I propose to do (perhaps to serve some desire or goal

I have). Armed with this tentative plan, then, I turn to FUL to see if it is

legitimate to act on it. FUL is, in effect, a test of maxims: it tells me to act only on maxims that have a certain feature.

For the moment, let’s leave the details of that test aside, and focus on the

Kantianism for Consequentialists 123

negative form of the imperative. FUL tells me to act only on maxims that

pass a certain test. Thus, if some maxim fails the test, FUL commands me not to act on it. Notice, however, that although FUL tells me to act only on maxims that pass the test, it does not require me to act on all the maxims that

do pass the test. Apparently, then, if a maxim passes the relevant test you may act on it (FUL, at least, won’t rule this out); but absent any further argument, it seems, there won’t be any requirement to act on the maxim. We must restrict ourselves to acting on maxims that pass the test, but among

the maxims that do pass, which we choose to act upon is up to us.∂

Suppose, then, that some maxim fails the FUL test (whatever, exactly, it

turns out to be). What can we conclude? If FUL is indeed a categorical

imperative, binding upon all rational beings, then we must conclude that it

is forbidden to act on that maxim. But what follows in the alternative case,

where the given maxim passes the test? Here we have to be more cautious.

Obviously enough, if a given maxim passes FUL, then as far as FUL itself is

concerned there is nothing objectionable about acting on the maxim. But

we cannot yet safely conclude that it is indeed permissible to act on the maxim in question, because, for all that we have said so far, there might be

some other imperative—beyond FUL—that must be taken into account as

well. After all, even if kantians are right in thinking that reason’s autonomy

supports FUL, it doesn’t yet follow that this is the only fundamental princi-

ple supported by our autonomy. Perhaps there are additional tests that must be passed as well. If so, then passing FUL will be necessary for permissibil-

ity but not sufficient.

Presumably Kant means to put this possibility aside with his insistence

that FUL is the only categorical imperative. (Because of this belief, he typically refers to it simply as ‘‘the’’ categorical imperative, though as we

have noted Kant also believes that this imperative can be stated in several

different, though equivalent, ways.) But even if Kant could prove that FUL (in its various formulations) is indeed the only categorical imperative,∑ that

wouldn’t necessarily put the worry to rest. For what if there were addi-

tional, basic principles (that is, principles not derived from FUL) that,

although not categorical, nonetheless validly applied in particular cases?

Even if FUL is the only categorical imperative, nothing yet rules out the possibility that a maxim might pass FUL but nonetheless fail to pass these

further (noncategorical) principles.

What the kantian needs to claim then (regardless of whether FUL is the only categorical imperative) is that even if there are any further valid princi-

ples (not themselves derived from FUL), it is not actually possible for a

maxim to pass FUL but to violate these further principles. Happily, this may

124 Shelly Kagan

not be an implausible claim for the kantian to make. Imagine that a given

maxim violates some such principle, P. Now given the autonomy of reason,

any valid principle of reasoning, including P, must be one that I rationally

favor. But if I truly continue to endorse P (even in light of its ruling out the

maxim in question) then I cannot rationally favor any principle incompat-

ible with P—including, in particular, the underlying principle correspond-

ing to the maxim. Thus, given my acceptance of P, I cannot in fact rationally

will the maxim to be universal law. That is, if the maxim violates P, it fails

FUL as well.

What this means, then, is that even if there are additional principles (not themselves derived from FUL), so long as a given maxim does pass FUL it

will pass those additional principles (if any) as well.∏ Thus, provided that a

maxim passes FUL, it is indeed permissible to act upon it.

I think, therefore, that we can put aside the potential complications that

threatened to arise from the existence of additional tests beyond that pro-

vided by FUL. We can say, straightforwardly, that if a maxim passes FUL

then it is permissible to act on it. And we can combine this result with a

point already made, that if a maxim fails FUL it is forbidden to act on it. Summing all of this up then we can conclude, quite simply, that it is permis-

sible to act on a maxim if and only if it passes FUL.

It would, however, be easy to become confused about what we have

shown so far. Suppose that in some situation I consider a maxim, M, that

would permit me to perform an act, A, in those circumstances. And let us

suppose, as well, that this maxim fails FUL. It would be natural to think that

what this shows me is that it is forbidden to do A (at least, in these circum-

stances). But in point of fact this doesn’t actually follow at all. From the

mere fact that M fails FUL, all that immediately follows is that one should

not act on M. That is, one should not do A for the particular reasons given by M. The maxim M, after all, corresponds to a particular reason-giving

principle, and that principle picks out certain features of the situation, and

tells me that by virtue of these features I have reason to do A. The fact that

M fails FUL shows me that this particular claim about what I have reason to

do (and why) is mistaken. Thus, if I do have reason to do A it is not for those (purported) reasons. But all of this is still compatible with the possibility

that there may be other (genuine) reasons to do A—even in this very

situation. For there may still be some other reason-giving principle which is sound—a principle that focuses on different features of the very same

situation, and tells me that by virtue of those features I have reason to do A. In short, even though M fails FUL, some other maxim that would permit me

to do A may still pass.

Kantianism for Consequentialists 125

Thus, even though M fails FUL, we cannot yet conclude that it is forbid-

den to do A (in this situation). To reach that conclusion we would need to

examine various other maxims as well, that is, the various other maxims

that would also instruct me to do A. It is only if all such ‘‘permission giving’’ maxims fail FUL as well that we can conclude that doing A is

forbidden. (Since I may only act on maxims that do pass FUL, if all such permission giving maxims fail, then I am indeed forbidden to do A.)

This is not to say, of course, that before concluding that a given type of

act is forbidden in a certain situation one must literally examine a huge

(perhaps infinite) number of maxims. It is possible that when a particular

maxim fails FUL we will be able to see precisely why it fails, and general-

ize to other, relevant maxims. In effect, we may be able to test large classes

of maxims at (more or less) the same time. But logically speaking the point

remains, that the failure of a single maxim does not suffice to establish that a

given act is forbidden; that requires, rather, the failure of all maxims that

would permit the act (in those circumstances). (Similarly, of course, to

establish that a given type of act was forbidden under all circumstances, we would need to show that all such permission giving maxims would fail,

regardless of what circumstances the maxims specify as relevant.)

In the last few paragraphs I have been freely talking about actions as

permissible or forbidden. What kind of permissibility is this? So far, the

answer is rational permissibility. FUL provides a test for reason-giving principles, allowing us to conclude, in certain cases, that an action is ra-

tionally forbidden (say) because no genuinely adequate reason supports

doing it. But the kantian believes that FUL captures a central moral idea as well. It serves to sort the morally permissible from the morally forbidden. If

this is right, then rationality meets morality here: if the autonomy of reason

requires you to conform to FUL, and acts forbidden by FUL are morally

forbidden, then reason requires you to obey morality.

To understand why the kantian thinks FUL can plausibly be taken not

only as a requirement of rationality but also as the basic principle of moral-

ity, it may be helpful to turn to a concrete example. Kant asks us to consider

a case where I attempt to borrow some money, promising to pay it back,

even though I know full well that I will be unable to keep such a promise.

(The same basic example is discussed at two different places—G 4:402–3

and 422—though only the second discussion makes explicit that the case

involves money.) Kant supposes that my maxim here tells me that ‘‘when I

am in a tight spot’’ I will ‘‘make a promise with the intention of not keeping

it’’ (G 4:402). And here is part of Kant’s discussion of whether this maxim

passes FUL:

126 Shelly Kagan

I ask myself: would I be content with it if my maxim (of getting myself

out of embarrassment through an untruthful promise) should be valid as

a universal law (for myself as well as for others), and would I be able to

say to myself that anyone may make an untruthful promise when he

finds himself in embarrassment which he cannot get out of in any other

way? Then I soon become aware that I can will the lie but not at all a

universal law to lie. (G 4:403)

Several details of this argument will require more careful discussion

later. Here I only want to draw attention to the plausibility of the idea that

FUL is indeed concerned with fundamental moral aspects of the situation.

In effect, Kant is telling us that immorality is a matter of cheating—making

an exception of oneself (cf. G 4:424). When I tell a lie, or make a promise I

don’t intend to keep (or butt in line, or kill someone for personal gain, and

so forth), I am playing by rules that I don’t favor others acting on as well.

After all, it is not as though someone who is immoral wants others to act in

the same way! On the contrary, what I want when I act immorally is that

everyone else should play by one set of rules (the moral rules) while I alone

get to act on a different set of rules. Here I am, then, proposing to act in a certain way, in a certain situation, but it is perfectly clear that I cannot

rationally will that everyone act in the same way in similar situations. There

is a (purported) reason-giving principle that I propose to act on, but I can’t

reasonably favor that others act on it as well. This is the telltale sign of

immorality, says the kantian. I want to treat myself differently than every-

one else gets treated; I want one set of rules for myself, and another set of

rules for everyone else. When I violate FUL, acting on a principle that I

cannot will to be universal law, I try to make an exception of myself, even

though I see full well that there is nothing at all that I consider a relevant

difference between myself and others; and that is the mark of immorality.

That is why FUL is a requirement, not only of rationality, but of morality

as well. And so we can conclude: if an act is forbidden by FUL, it is morally

forbidden. But can we similarly conclude that if an act is permitted by FUL,

it is morally permissible? As before, however, this conclusion assumes that

FUL is not only one test among many, but is indeed the only fundamental

principle—now, the only fundamental moral principle. This, too, is a claim that Kant appears to make (though it is not clearly distinguished from the

earlier claim that FUL is the only fundamental rational principle), and for

the sake of argument, at least, let us grant it as well (we’ll consider its

plausibility later). Then we can say that an act is morally permissible if and

only if it is permitted by FUL.

Kantianism for Consequentialists 127

Once again, it is important to avoid misunderstanding. We have just

concluded that an act will be morally permissible if and only if it is permit-

ted by FUL. And as we have already discussed, an action will be permitted

by FUL provided that there is some maxim that passes FUL that permits the

action in the circumstances. Note, however, that nothing that we have said

requires that this maxim be the one that the person is actually acting upon.

Provided that there is some permission giving maxim that passes FUL, it will be morally permissible to perform the act in question, even if the

person is acting on some other maxim, and that maxim fails FUL! Of course, if the person is acting on another maxim, and that maxim fails

FUL, there will be plenty that is amiss. The person will be acting on a

maxim that is unsound, both rationally and morally. That is to say, she will

be performing the action for the wrong reasons—for ‘‘reasons’’ that are not actually adequate reasons for action at all. What’s more, she will be per-

forming the action for reasons that are not morally legitimate. As such, the

person may well be open to moral condemnation of one sort or another. But

this is not to say that what she is doing is morally forbidden. Rather, we will have a case of someone who is doing an action that is perfectly permissible

morally, but is doing so for the wrong reason. In kantian jargon we can say

that such a person is conforming to the moral law, but not acting for the sake

of the moral law (G 4:390).

The distinction being drawn here is a perfectly familiar one. We all have

the idea of someone doing the morally right thing, but for the wrong rea-

sons. For example, Kant discusses a shopkeeper who gives correct change

to his customers, but does so only out of fear of being caught and having

business suffer (G 4:397). Presumably, we will all agree that giving correct

change is a morally permissible (indeed morally obligatory) thing to do.

And so we would agree that when the shopkeeper does this his action is

morally permissible; he is conforming to the moral law. This is true even

though he acts out of fear—acts for the morally wrong reasons. Thus,

despite the fact that the maxim he acts on is unsound, that it would (as we

may suppose) fail FUL, it remains true that the action he performs is mor-

ally permissible. And what makes it morally permissible is the very fact that some other maxim that enjoins giving correct change would pass FUL. In short, an act is morally permissible if and only if some permission giving

maxim passes FUL, whether or not the person in question is actually acting

on that maxim.

Let us now return to the question, earlier set aside, of how exactly we are

to determine whether or not a given maxim does pass FUL. The basic idea,

of course, is clear: a maxim passes FUL just in case I can will it to be a

128 Shelly Kagan

universal law. But how, exactly, can I tell whether or not I can ‘‘universal-

ize’’ a maxim in this way? What, exactly, do I do when I try to determine

whether a maxim can be universalized?

On what I take to be the standard proposal here, I should begin by trying

to imagine a world where everyone does in fact conform to the reason- giving principle corresponding to the maxim being tested. If the maxim

enjoins me to perform an act of type A, in such and such circumstances,

then I am to imagine a world in which everyone performs acts of type A when in circumstances of that sort. I attempt to imagine a full compliance world, as we might call it, and then I ask myself two questions about this world. First, is such a world truly possible, or does something go wrong in

trying to imagine it? Second, assuming that such a world is indeed possible

(that nothing goes wrong in the relevant sense), can I rationally will it? The

first question, in effect, is supposed to tell me whether the principle cor-

responding to the maxim could actually be a universal law; the second, whether I can will it to be such. To pass FUL, I must be able to answer both questions in the affirmative.

According to this interpretation, then, there are two distinct ways in

which a maxim could fail to pass FUL, corresponding to the two questions

I’ve just distinguished. In effect, there are two distinct subtests. This seems

to be Kant’s own view of the matter, in any event: he says that some maxims

‘‘cannot even be thought without contradiction as a universal law,’’ while other maxims that also fail FUL generate no such ‘‘internal impossibility’’;

for these other maxims, rather, the will would ‘‘contradict itself’’ if it at- tempted to will the maxims to be universal laws (G 4:424).

More significantly, Kant seems to think this distinction picks out some-

thing important, generating different types of moral requirements. These

are obscure matters, and Kant says little about them in the Groundwork, but roughly the picture seems to be this: when maxims fail at the first step, this

is supposed to generate ‘‘perfect’’ duties, while ‘‘imperfect’’ duties (which

are, despite the name, perfectly genuine duties) are generated by maxims

failing at the second step (cf. G 4:424). But it is far from obvious why the

two subtests should be invested with anything like this kind of significance.

FUL says that one should not act on maxims that cannot be willed to be

universal law. It does not say that it matters why a given maxim cannot be so willed. So it is far from clear that the kantian should follow Kant in holding

it significant at which step a given maxim fails.

For that matter, it must be admitted as well that it is far from clear what

precisely we are supposed to be concerned with as we consider the two

subtests. In a moment we will turn to an examination of some of Kant’s own

Kantianism for Consequentialists 129

examples. At the very least this should help us get clearer about what Kant

thought could lead to a maxim’s failing FUL. Whether, at the end of the day,

we agree with Kant that it makes a difference at which step a maxim fails

(or whether, indeed, maxims can fail in only two basic ways) is a matter of

less importance.

Kant discusses four main examples in the Groundwork. I am going to discuss only two of these, but I am going to do so in some detail. (In

thinking about these examples, it is also worth bearing in mind the point

that Kant is only human. In certain cases he may simply be wrong about

what FUL entails. Kantians can embrace FUL while still rejecting one or

more of Kant’s own views concerning which particular moral requirements

emerge from it.)

A. The False Promise

The first example I want to examine is Kant’s second, a return to the

false promise case that we have already had a look at. Recall that Kant

claims that ‘‘I can will the lie but not at all a universal law to lie.’’ Here is

Kant’s initial argument for this claim—that I cannot will the maxim to be a

universal law:

for in accordance with such a law there would properly be no promises,

because it would be pointless to avow my will in regard to my future

actions to those who would not believe this avowal, or, if they rashly did

so, who would pay me back in the same coin; hence my maxim, as soon

as it were made into a universal law, would destroy itself. (G 4:403)

And here is the argument the second time around, when Kant returns to

the case as one of his four examples:

Yet I see right away that it [my maxim] could never be valid as a

universal law of natureπ and still agree with itself, but rather it would

necessarily contradict itself. For the universality of a law that everyone

who believes himself to be in distress could promise whatever occurred

to him with the intention of not keeping it would make impossible the

promise and the end one might have in making it, since no one would

believe that anything has been promised him, but rather would laugh

about every such utterance as vain pretense. (G 4:422)

Now the basic line of argument here is clear enough. If we try to imagine

a world in which everyone lies, or makes insincere promises, so as to

130 Shelly Kagan

achieve personal goals by deceiving others, we find that something goes

wrong. No one would believe you when you tried to make such a promise.

But what, exactly, is it that goes wrong here? Some think that what we

find is that it is literally impossible for there to be a world in which everyone lies or makes insincere promises. Perhaps in a world where promises are so

routinely broken, the very institution of promising would disappear (or,

alternatively, would never have come into being). So there cannot be a

world in which everyone makes promises they do not intend to keep. The

maxim of lying to get out of a tight spot could not be a universal law,

because there literally could not be a world in which everyone complies

with this maxim. This interpretation sits nicely with Kant’s saying that in

such a world ‘‘there would properly be no promises,’’ that the universality

of the law would make such promises ‘‘impossible’’—that (as he later puts

it) the maxim ‘‘cannot even be thought without contradiction’’ to be a universal law (G 4:424). If there literally cannot be a world in which every- one acts on the maxim, I cannot will it to be universal law, and the maxim

fails FUL.

Others interpret the argument somewhat differently. Taking their cue

instead from Kant’s remarks that making the promise in such a world would

be ‘‘pointless,’’ that it would be impossible to achieve ‘‘the end one might

have’’ in making the promise, they conclude that what actually goes wrong

is this: in a world in which promises are routinely broken, it is much more

difficult, and perhaps even impossible, to achieve the goal specified in the maxim (getting out of a tight spot by deceiving others) by performing the

action specified in the maxim (making an insincere promise). Insincere promising works more effectively (and perhaps only at all) against a general

background in which people keep their promises. Thus, making the maxim

be a universal law—one that everyone has reason to act on—undercuts the

effectiveness of the maxim itself. And this involves a kind of practical

contradiction: if I will my maxim to be universal law, I make it harder to

achieve the very goal specified by the maxim by acting on that maxim.

From the point of view of someone willing the maxim, then, it is irrational

of me to will it to be universal law. So I cannot will the maxim to be

universal law, and it fails FUL.∫

This second interpretation, it should be noted, assumes that it is not

rational for someone who accepts the maxim to will that everyone act on the maxim, since this makes it harder to achieve the goal specified in the maxim

(in the specified manner). The argument thus presupposes some principle to

the effect that it is not rational to favor things that make it harder to achieve

one’s goals. This is not an objection to the argument, of course, for presum-

Kantianism for Consequentialists 131

ably we would indeed want to endorse some such principle of instrumental reasoning (although the details of the principle might be a matter of debate).

Kant himself, for example, earlier in the Groundwork (G 4:417), defends the claim that ‘‘Whoever wills the end, also wills (insofar as reason has

decisive influence on his actions) the means,’’ and it looks as though, on the

second interpretation, this principle, or some near relative of it, is assumed.

Again, this observation is not intended as an objection to the second

argument. If we are to sometimes reject maxims on the grounds that we

cannot autonomously will them to be universal laws, then presumably one

reason this may happen is because we find that willing the maxim to be

universal law would be an act that would fall short in terms of one or

another standard that we rationally endorse. Thus, if we do rationally en-

dorse some principle of instrumental reasoning, it is not problematic for the

kantian to appeal to that principle when arguing that one cannot will a

particular maxim to be universal law.

Regardless of which interpretation we accept, it is worth drawing atten-

tion to the fact that the argument makes use of various contingent, empirical

facts. The argument assumes, for example, that people have memories, and

will recognize the fact that promise breaking has become widespread, and

that this will result in either a breakdown of promising (on the first inter-

pretation) or a disinclination to trust the promises of others (on the second

interpretation). I note this point only to put to rest the widely held belief that

kantians think that morality is entirely a priori, something that can be established without appeal to empirical facts.Ω At best, FUL itself has this

kind of status. As we can see, however, more specific moral conclusions—

such as a prohibition against lying or making insincere promises—are

derived from FUL through the use of empirical truths.∞≠

Suppose we grant, if only for the sake of argument, that Kant has suc-

cessfully shown that the maxim in question cannot pass FUL. For reasons

we have already discussed, however, it won’t yet follow that it is morally

forbidden to make an insincere promise in this case. To reach that conclu-

sion, after all, we must argue that not only this maxim, but any other maxim

that would permit lying here, would fail as well. Kant doesn’t try to general-

ize his argument, to cover the other relevant maxims, but it is easy to see

how the attempt might go. The features of the maxim that seem relevant to

its failure are ones that would appear in any permission granting maxim

relevant to the case at hand. Thus if one maxim that would permit lying here

would fail, others should as well. (We’ll consider an objection to this claim

below.)

Can the argument be generalized even further? Can we derive not only a

132 Shelly Kagan

prohibition against lying in this particular case, but a general prohibition

against all lying whatsoever? Kant thought so, and notoriously claimed that

it is never morally permissible to tell a lie. But this is a point at which many

kantians part company from Kant himself. To take the standard case, many

kantians believe it permissible to lie to a would-be murderer so as to protect

his innocent victim hiding in your basement. And it seems at least possible

that a maxim that would permit this act could pass FUL. After all, the

existence of at least some insincere promising is compatible with the con-

tinued existence and effectiveness of promising (for there are, let us admit,

insincere promises made in the real world, yet promising has not been

rendered impossible or ineffective). Thus it seems possible that a maxim

that enjoined promise breaking or lying in sufficiently rare or special cir-

cumstances (for example) might yet pass FUL. Perhaps a maxim that per-

mitted lying to the would-be murderer is one such.∞∞ Kant may have thought

that FUL supported an absolute prohibition against lying, but the kantian

need not follow him in this regard.

B. The Maxim of Nonaid to Others

Kant’s fourth example involves a person who has a chance to aid another

in need, but is tempted to pass him by without offering assistance. Kant

imagines the person’s maxim to be one of complete refusal to provide aid

(‘‘I will not take anything from him or even envy him; only I do not want to

contribute to his welfare or to his assistance in distress’’), and he says of this

case:

But although it is possible that a universal law of nature could well

subsist in accordance with that maxim, yet it is impossible to will that such a principle should be valid without exception as a natural law. For a

will that resolved on this would conflict with itself, since the case could

sometimes arise in which he needs the love and sympathetic participa-

tion of others, and where, through such a natural law arising from his

own will, he would rob himself of all the hope of assistance that he

wishes for himself. (G 4:423)

Once again, the basic line of argument is fairly straightforward. Kant

says that although there could be a world in which everyone acts on the

maxim in question—a world where no one helps others—you cannot will

this maxim to be universal law. You cannot rationally will that indifference

to the needs of others be universal law, for you might find yourself in a

Kantianism for Consequentialists 133

situation where you need the help of others. The maxim thus fails at the second subtest: it cannot be willed to be universal law.

The first thing to notice about this argument is that it, too, appeals to

empirical facts, here the fact that each of us has needs that we cannot always

meet on our own. The second thing to notice is that it, too, makes use of

something like the principle of instrumental reasoning. The thought is that

each of us has goals of some sort, goals that we will want to achieve. But

this makes it irrational to favor things that would make it more difficult to

achieve those goals. Yet this is precisely what we will have done, in at least

certain logically possible scenarios, if we will that it be universal law that no one help another. Once I recognize that I, too, can be in need of the aid of

others, I cannot rationally favor a principle that would mean that I not get

the help I need.

Notice, as well, that the relevant question is not particularly what I

would will, were I in the situation where I needed help. That is no more relevant than the question of what I would will in the case where I don’t

need help. Rather, the question is what I am rationally prepared to will here and now as a principle to govern the case where I need help. Presumably, it is not rational of me (here and now) to be indifferent to my own need in that

possible case. So I cannot (here and now) favor a rule that would mean that

that need would go unmet (were it to arise). That is why I cannot will the

maxim to be universal law. Thus it fails FUL.

Being clear about this point helps us to understand why it is irrelevant

for someone to object that in the actual world they simply do not need

anyone’s help. Even if that were the case, it would remain a live possibility

that the situation could be different: for anyone other than a deity, one could find oneself in need. And the thought, then, is that it cannot be rational to

will, with regard to such a situation, that one not get the aid one would need.

Sometimes it is thought that whatever the force of this argument, it fails

against an imagined ‘‘rugged individualist’’ who truly favors getting by

completely on his own. Such a person, it is suggested, can will that the

maxim of nonaid be universal law—for when he contemplates the pos-

sibility that he would himself be in need of the aid of others, he insists that

even in such a case he (here and now) prefers that he die (in the given case)

rather than be helped by others. (Of course, were he actually in a position of

extreme need, he might lose his resolve and desire help. But as we have

seen, that is strictly irrelevant. What matters is that here and now he wills

that he not be aided, even in that case.)

I believe, however, that the kantian may have an answer to this objection

134 Shelly Kagan

available to him. For even the rugged individualist wants help of a particu-

lar kind—namely, to be left alone. This is easily seen if we imagine some-

one else bent on ‘‘aiding’’ him, despite his protests. The individualist wants

the cooperation of others, just as the rest of us do; it is just that aid and

cooperation take an unusual form in his case: leaving him to do things

completely by himself. If this is right, then not even the individualist can

favor a principle that would enjoin everyone to refuse to provide each with

the particular aid that they need, for that would strip the individualist of

what he most needs—to be left alone. If this is right, then none of us—not

even the rugged individualist—can will a maxim of nonaid to be universal

law.

Of course, as always, even if this is right it doesn’t yet show us that it is

morally forbidden to refuse to aid others. Doing that would require showing

that not only this particular maxim but other, similar maxims would fail

FUL as well. Once again, Kant doesn’t attempt to generalize the argument,

but here too it is not difficult to see how that more general argument might

go: all humans (at least) are finite in ability, capable, in principle, of needing

help (of some sort) from others; thus for any maxim at all that would simply

permit disregarding the needs of others, no one can rationally will the

maxim to be universal law.

But there remains a further worry. It might be objected that no principle

at all could avoid the objection being raised against a principle of nonaid.

For if, as the argument claims, it is irrational for me to will a principle (such

as a principle of nonaid) that might leave me unable, or less able, to achieve

my goals, then won’t it be similarly irrational for me to will a principle that

requires providing aid to others? After all, acting on a requirement to provide aid can itself leave me unable, or less able, to achieve one or

another of my goals. Thus, won’t the very same principle of instrumental

reasoning that supposedly makes it irrational to favor a principle of nonaid

also make it irrational to favor a principle requiring aid to others? How, then, can any principle at all—whether requiring aid or not—pass FUL?

Presumably the kantian must claim that an adequate answer to this

worry involves balancing the various needs and aims I might have that

might go unmet under the differing principles. I am looking for a principle

that I can will to be universal law. And since, logically speaking, I might

find myself in either one of the relevant roles (aid provider or aid recipient),

I have to ask myself which costs I would rather endure. But in at least some

cases—for example, when the gain to the needy when aid is provided is

significantly greater than the loss to the person who actually provides the

aid—the answer to this question is clear. Presumably, then, the principle of

Kantianism for Consequentialists 135

instrumental reasoning can lead me (here and now) to favor principles that

do require providing aid in cases of this sort. But if this is right, then FUL

will indeed support some sort of requirement to provide aid after all.

Doubtless, further questions could be raised about both of these exam-

ples (and as I have already noted, Kant discusses two other examples in the

Groundwork as well). But I hope I have said enough to give at least some sense of how FUL is supposed to be used as a test for maxims and for

deriving moral obligations.

Our discussion should also put to rest one common objection to FUL,

namely, that it has no ‘‘bite,’’ that any maxim at all can pass. For as we have

now seen, it’s not implausible to think that certain maxims do indeed fail

FUL. Thus, whatever its other shortcomings may be, at least FUL isn’t

altogether devoid of content.

There are, however, other general objections to FUL that merit further

discussion. Let me quickly mention four. All of them concern the adequacy

of FUL from the moral point of view. First, it is sometimes objected that FUL is raising a morally irrelevant concern when it asks us to consider a

world where everyone acts on the maxim in question. After all (the objec-

tion notes), in the real world typically it simply isn’t going to happen that everyone acts on a given maxim. From the moral point of view, then, why should we concern ourselves with such an unrealistic possibility?

Recall, however, that the kantian’s position is that if a maxim passes

FUL, then it is morally permissible to act on it, indeed, morally permissible

for anyone at all to act on it. It hardly seems irrelevant, then, to consider a

world in which everyone does act on the given maxim. This would simply be a world in which—in the relevant way at least—everyone is acting in a

manner that is supposedly morally permissible. Surely it makes sense to

insist that it must at least be possible for everyone to act in a morally permissible manner, and indeed, to insist further that it must be reasonable

to favor a world in which everyone acts in a morally permissible manner. (It cannot be preferable, from the moral point of view, that some act in a

morally forbidden manner.) A world in which everyone acts morally must

be both possible and attractive. Thus, in directing our attention to a full

compliance world, FUL is not at all directing our attention to a morally

irrelevant possibility.

But this immediately suggests a second objection: even if the full com-

pliance world is indeed a world worth considering when testing maxims

from the moral point of view, it is quite another matter to suggest that this is

the only world worth considering, or even the most important. After all, in the real world not everyone is going to act morally, and so it is important to

136 Shelly Kagan

know how one is permitted (or required) to act in the face of immoral

behavior by others. It would seem that the relevant question with regard to

such cases of partial compliance is what I can will with regard to a world in which not everyone is acting on the maxim in question. But FUL apparently never asks us to consider such worlds: it restricts our attention to asking whether I can will a given maxim in a world in which everyone is acting on

the maxim. Thus FUL inappropriately disregards the very real possibility of

immoral behavior (partial compliance). Worse still, because of this neglect,

it can generate morally implausible guidance, since acts that might be

perfectly attractive were everyone to be acting morally (ones that I can will

for the full compliance world) might be catastrophic when done in the face

of immoral behavior.

Presumably, this difficulty about how to properly evaluate maxims for

dealing with partial compliance might not be particularly worrisome if

there were further tests, beyond FUL, that needed to be passed as well

before it was permissible to act on a given maxim. If there were such further

tests, then they might do a better job of evaluating whether a maxim can

properly handle cases of merely partial compliance. We could appeal to

these further tests to rule out maxims passed by FUL that were inadequate

in this regard. But as we have already noted, Kant believes that FUL (and

its equivalent, alternative formulations) is the only fundamental principle

needed, and kantians have typically followed him in this. So it is worth

asking whether FUL has the ability to handle the problem of imperfect

compliance on its own.

I believe that it does. The problem, I think, lies not with FUL itself, but

with what I earlier called the ‘‘standard proposal’’ for interpreting FUL.

According to this interpretation, recall, to see whether a maxim passes FUL

I need only ask whether I can will that the principle corresponding to the

maxim be one that everyone acts upon. That is, I need only consider the full

compliance world—whether it is possible, and whether I can rationally

favor it. But why should we take FUL to be so easily satisfied? According to

FUL, after all, I should only act on maxims that I can will to be universal law. In particular, then, I have to ask whether the appropriate principle is

one that I can rationally will for all cases to which it applies. Now one such case, to be sure, may well be the case of full compliance. But often enough

the principle in question will apply to other cases as well, cases of imperfect

compliance; and so I must ask whether I can rationally will that the princi-

ple govern those cases as well. Thus, contrary to the claim put forward by the objection (and reinforced by the standard interpretation), FUL does not

actually disregard consideration of partial compliance worlds, worlds

Kantianism for Consequentialists 137

where not everyone is acting morally. On the contrary, it demands that we

consider such worlds as well, before signing off on a principle. Only if we

can will the principle for cases of imperfect compliance as well (assuming

that it applies to such cases) is it really true that we can will the maxim to be

universal law.

It may also be worth recalling, in this regard, that the principles we favor

need not prescribe the same type of action regardless of circumstances. In

particular, then, we might favor principles that tell us to act in one way

when others are acting similarly, and in quite another way when they are

not. Thus the principles that pass FUL may enjoin one kind of behavior

when others are acting morally, and quite another in the face of immoral

behavior. In short, there is no good reason to believe that FUL will be

unable to generate appropriate moral guidance for dealing with cases of

noncompliance.

A third objection complains that in point of fact no maxims (or perhaps only very few maxims) can actually pass FUL. In particular, perfectly

harmless maxims—maxims that intuitively it ought to be permissible to act

upon—fail. If this is correct, of course, then we have some reason to reject

FUL: if it fails maxims that ought to pass, then it isn’t a very good test of the

validity of a maxim. Here is an example of the sort of problem that people

have in mind when they raise this worry. Suppose that I form the intention

of going to the local pizza house, and ask whether my maxim (‘‘I will go to

Naples for lunch’’) can pass FUL. I must ask whether I can will this maxim

to be universal law; and apparently this involves trying to imagine a world

in which everyone—at a minimum, all five billion humans—goes to Na- ples for lunch! But as soon as I do this I see that either this is literally not

possible (not everyone could fit) or it would involve a practical contradic-

tion (it would make it much more difficult to get lunch). Thus my maxim

fails FUL. But this—the objection concludes—is absurd. Surely going to

the local pizza house is morally permissible (special circumstances aside),

and if FUL condemns my maxim, so much the worse for FUL.

In answering this objection, the first thing to remember is that even if

this maxim does fail FUL, that doesn’t entail that it is morally impermissi-

ble to have lunch at Naples. So long as another maxim that permits having

lunch at Naples passes FUL, then it will be perfectly permissible to have

lunch there. At worst, all that would follow is that the short maxim we are

here testing—‘‘I will go to Naples for lunch’’—does not provide a com-

pletely accurate account of what I have reason to do. And this is not, in fact,

an implausible claim. For as a moment’s reflection makes clear, whether it

makes sense for me to go to Naples depends on any number of factors not

138 Shelly Kagan

mentioned in the maxim as stated, for example, whether or not I am hungry,

whether or not I want pizza, whether or not the restaurant is crowded,

whether or not it is nearby, and so forth. Presumably I do not have reason to go to Naples regardless of how crowded it is, how inconvenient it is to get to

it, and so on. Thus the simple maxim ‘‘I will go to Naples for lunch’’ cannot

in fact be plausibly taken to be a complete account of what I have reason to

do and why. That requires a much fuller statement, one that, for obvious

reasons, I rarely have occasion to try to articulate fully. Normally, the

relevant extra conditions are left implicit, and so the short maxim is perhaps

best understood as a kind of shorthand for that fuller statement.

Once we keep this point in mind, and try to universalize an appropriately

full statement of the maxim (or universalize the short maxim, understood to

implicitly contain the various necessary qualifications), we find that the

maxim can indeed pass FUL. I can certainly will that everyone go to Naples

if it is convenient, if it isn’t too crowded, if they want pizza, and so forth.

After all, obviously enough, one or another of these conditions won’t be

met for almost any person we might consider (most, for example, are much

too far away for it to be convenient). And so, when we imagine a world in

which everyone acts on this maxim, we won’t imagine a world with billions

trying to crowd into the local restaurant. Rather, we imagine a world in

which those who want pizza and are nearby (and so forth) go. And this is a

world, it appears, that we can readily will.

In short, if we take the simple maxim to be a complete statement, it does

fail FUL, but appropriately so, while the fuller maxim passes. And if we

take the simple maxim to be shorthand for that fuller maxim, then of course

it passes as well. Either way, there will indeed be a maxim that passes FUL

that permits me to go to Naples (special circumstances aside), and so,

contrary to the claim of the objection, FUL won’t forbid this morally in-

nocuous act.

The third objection claimed (albeit incorrectly) that too little passes FUL.

The final objection that I want to consider, our fourth, makes the opposite

complaint, that too much passes. For as we have just seen, a complete

specification of one’s maxim might include any number of clauses and

conditions. (FUL does not restrict us to testing ‘‘simple’’ maxims: any

maxim can be put forward for testing.) The worry, then, is that if one is

sufficiently clever in formulating one’s maxim, one can always arrive at a

version that will pass FUL, no matter how morally unacceptable the act in

question. For example, suppose I want to murder you. Even if (as we might

suppose) the straightforward maxim ‘‘I will murder those I want dead’’

would fail FUL, I need only propose, instead, a maxim that includes, say, my

Kantianism for Consequentialists 139

proper name. Suppose, then, that I try the maxim ‘‘If I am named Shelly

Kagan then I will murder those I want dead.’’ If this maxim can indeed pass

FUL, then I am permitted to murder you (whether or not this is in fact my

maxim). But this would clearly be unacceptable. So if the rigged maxim does

indeed pass FUL, we will simply have to reject FUL.

The objection then continues by insisting that this maxim does, in fact,

pass FUL. After all, there is presumably no impossibility about having a

world in which everyone named Shelly Kagan kills at will (indeed I may well be the only person named Shelly Kagan in the world), and it certainly

seems that I (Shelly Kagan!) can be in favor of a principle that gives me this

extra freedom. So it looks as though I can will the maxim to be universal

law, and FUL unacceptably permits me to kill at will. (Similar results could

presumably be achieved by replacing my name with a definite description

that uniquely picks me out, for example, ‘‘If I am a professor of philosophy

at a midsize university, with three children, and a wife who works as a

midwife, etc., etc., . . . then . . .’’ For simplicity, however, I’ll stick to

introducing the proper name.)

In fact, however, I think it far from obvious that I can rationally will the

maxim in question to be universal law. After all, although I believe that I am

one of at best a handful of people named Shelly Kagan—perhaps, indeed,

the only one—I could presumably be mistaken about this. Perhaps there is a

vast extended clan, currently living peacefully in the jungle, all of whose

members are named Shelly Kagan. I can hardly rationally favor a principle

that would permit this vast group to kill at will. And even if (as certainly

seems likely) this possibility is unrealized in the actual world, there could be such a world, and it simply isn’t true that I (here and now) am prepared to

will with regard to such a world that all the Shelly Kagans in that world be

permitted to kill at will. Thus it isn’t really true that I can rationally will that

the maxim ‘‘If I am named Shelly Kagan then I will murder those I want

dead’’ be a universal law. Accordingly, the fourth objection fails as well.

Generalizing from the failure of this particular example, it seems we can

say the following. Although nothing in FUL, in and of itself, places restric-

tions on the content of the maxims that we bring for testing—we can add

whatever silly clauses and conditions we’d like—proper application of

FUL does have the result of ruling out maxims that introduce irrelevant

conditions. If a maxim is couched in terms of conditions that are in point of

fact rationally and morally irrelevant, we will discover that we are not

genuinely prepared to will that the maxim be a universal law.

But the discussion of the third objection has already suggested a comple-

mentary point as well, namely, that proper application of FUL will also

140 Shelly Kagan

have the result of ruling out maxims that lack relevant conditions. If a maxim is overly simplistic, we will find that we are not genuinely prepared

to will that either. Taking these points together, then, the kantian claims that

FUL provides a sufficiently subtle and sophisticated test to guide us toward

plausible moral principles, ones that are sensitive to the relevant features of

acts and their circumstances while disregarding the irrelevant features.

III . Kantianism and Consequential ism

What would those moral principles look like? That is, given FUL, what

kind of normative moral theory emerges? Putting the question like this

appropriately emphasizes the fact that what we have been primarily dis-

cussing up to this point is the kantian account of the foundations of ethics. (In this regard we have been following the lead of Kant himself in the

Groundwork, the very title of which, after all, reveals that its primary concern lies with foundational issues.) We have not yet much concerned

ourselves with describing the particular normative principles (roughly, the more directly action guiding principles, such as those requiring promise

keeping or aiding others) that would emerge from that account, except as a

means of illustrating FUL at work.∞≤ I have, of course, tried to portray that

kantian account of the foundations of ethics as attractive and worth taking

seriously—and if I have succeeded in this endeavor, then my primary

purpose in this essay is accomplished. Still, it is natural to wonder about the

normative level as well. Given FUL, what kinds of normative principles are

we led to?

We have, of course, already taken a quick look at two particular exam-

ples. FUL, we have seen, rules out moral principles that would permit me to

be indifferent to the needs of others, or to lie (or make a false promise)

simply because this would be personally convenient. Obviously enough,

given the time, we could apply FUL to a variety of other cases as well, and

doing this over a sufficiently wide range of cases would doubtless enhance

our understanding of FUL’s plausibility and adequacy. But instead of con-

tinuing to focus on particular cases, I want to step back and ask, in a general

way, whether we can say anything helpful about the overall structure of the

moral theory that would emerge from FUL.

I raise this question, of course, because most kantians have thought it

fairly clear that FUL supports a deontological moral theory. Kant himself

certainly believed this. Indeed, even those who reject deontology—conse-

quentialists being the most prominent among this group—have typically

Kantianism for Consequentialists 141

accepted this claim as well, and thus concluded that avoiding deontology

requires rejecting the kantian account of the foundations of ethics. Now it is

certainly true that nothing that I have said in this essay constitutes a full

defense of kantianism. One might reject the account of autonomy that I

sketched at the outset, for example, or deny that autonomy leads to FUL. If

one does this, of course, then even if it is true that FUL does support

deontology, given a rejection of FUL this won’t threaten, say, one’s accep-

tance of consequentialism. On the other hand, some will find the kantian

account of autonomy and its implications sufficiently attractive, and FUL

sufficiently plausible in its own right, that they are prepared to accept the

moral principles supported by FUL, even if this requires revising some of

their previously held moral opinions. If FUL does indeed support deontol-

ogy rather than consequentialism, this may then provide a powerful argu-

ment in favor of deontology.

But there is a third possibility as well, of course, which is that Kant and

most kantians are wrong when they claim that FUL supports deontology. If

it should turn out that FUL actually supports consequentialism instead, then

to the extent that one finds the kantian account of the foundations of ethics

attractive, this will actually provide an argument in favor of consequential-

ism, rather than deontology.

Of course, one point is certainly true. If the kantian account of the

foundations of ethics is correct, then the basis of ethics looks rather unlike the accounts typically offered by consequentialists. For historically speak-

ing, at least, most consequentialists (though certainly not all) have grounded

their consequentialism in what we might call foundational consequentialism —the claim that the ultimate basis of the (valid) normative moral principles

lies in an appeal to the significance of the overall good. In contrast, the

kantian account that we have been sketching gives no particularly important

role at the foundational level to the concept of the good at all. The ultimate

basis of morality, for the kantian, is not the good, but rather freedom. For

this reason, it is appropriate to say that the kantian account of the foundations

of morality is foundationally deontological, rather than foundationally

consequentialist.

But it is one thing to insist that the kantian account of the foundations of

ethics is usefully classified as deontological; it is quite another to insist that

the particular normative principles that emerge from that account are them-

selves deontological. For absent further argument, there is no particular

reason to assume that deontological foundations must yield deontological

moral principles.∞≥ When I claim, then, that FUL may well support conse-

quentialism rather than deontology, I have in mind a claim not about the

142 Shelly Kagan

foundational level, but rather one about the normative level, the level that

concerns the various action guiding principles themselves. FUL may itself

be grounded in a nonconsequentialist account (this much certainly seems to

be true), but what emerges from FUL may well be a consequentialist rather than a deontological normative theory.

Evaluating this claim, of course, requires at least a working account of the

distinction between deontological and consequentialist theories (at the nor-

mative level). Simplifying somewhat, the following should do for our pur-

poses. Consequentialism holds that an act is morally permissible if and only

if it has the best overall consequences (of those acts available to the agent).

Deontology rejects this simple account of right and wrong, insisting that cer-

tain acts are morally forbidden, even when they would lead to better results

overall. Deontologists thus embrace constraints—prohibitions against per- forming the offensive types of acts, even when doing so would lead to better

results. Typical examples of constraints include prohibitions against lying,

harming the innocent, failure to keep one’s promises, and so on.∞∂

Deontologists normally also reject consequentialism on the further

ground that it is too demanding, always requiring the agent to perform the

act that would lead to the best results overall, no matter how great the

sacrifice involved to the agent himself. Deontologists thus typically em-

brace options as well—permissions to avoid promoting the overall good when the cost to the agent would be too great. For example, deontologists

typically don’t believe we are required to sacrifice huge portions of our

income to famine relief, even though if we did so a great many lives might

be saved. Such sacrifice is doubtless praiseworthy (they say), but it is

strictly optional: we are permitted, instead, to pursue our own individual

projects—as well as going to concerts, eating at expensive restaurants, and

so forth—even though our time and money could do much more good were

it spent in other ways. Most deontologists do insist, of course, that some-

times sacrifices for others are morally required (for example, when I can

rescue someone at minimal cost to myself ); but consequentialism goes too

far (they say) in putting no limits on the obligation to promote the overall

good.

While most deontologists accept both constraints and options (thus holding that consequentialism sometimes permits what is actually forbid-

den, and sometimes requires what is actually optional), I think it fair to say

that so long as a theory contains constraints, it would normally be consid-

ered deontological, whether or not it contained options as well. In contrast, the presence of options alone (that is, without constraints as well) would not

suffice to render a theory deontological. For our purposes, then, in asking

Kantianism for Consequentialists 143

whether FUL supports deontology, the key question facing us is whether or

not FUL supports constraints.

Nonetheless, it may be helpful if we begin with the question of whether

FUL supports options. For even though deontologists need not accept op- tions, all consequentialists reject them. Thus, if FUL is to generate a conse-

quentialist normative theory, it must reject options as well. Let us therefore

postpone, for the moment, the question of whether FUL supports con-

straints. Even if there are constraints, I might still be morally required to do as much good as I can by permissible means (that is, those means not

forbidden by constraints). So in asking whether FUL supports options or

not, we are asking whether FUL supports a requirement to do as much good

as one can—within the limits of constraints (if any). Now we already know, from the discussion of the aid example, that FUL

generates a requirement to aid others; FUL does not allow us to be indif-

ferent to the good that we can do. But many kantians have thought it plain

that FUL does not require us to do as much good as we can (within con- straints—a qualification that I will hereafter leave implicit). While FUL

sometimes requires us to promote the good (such as helping to meet the

needs of others), it does not require us to do all that we can in this regard. The claim of these kantians, then, is that FUL generates a requirement to

aid, but a limited one; when the cost of providing aid to others is too great, I am not required to do it.

But it is far from obvious that FUL will actually support this kind of

limitation on the requirement to provide aid. It is certainly true, of course,

that a maximally demanding requirement to promote the good will poten-

tially impose considerable costs upon me. Indeed, in the real world I might

find myself required to make huge sacrifices, while benefiting little, or not

at all, from the fact that others are similarly required to promote the overall

good. But in evaluating alternative principles concerning aid I must bear in

mind the fact that I am looking for a principle that I can rationally favor for

all worlds to which it applies. I cannot restrict my attention to the costs and benefits that I actually expect; I must consider all possible costs and all

possible benefits. And since I have no more reason to be concerned with the

costs that I might have to pay (as benefactor) than with the benefits I might

receive (as recipient), it seems reasonable to favor a principle that provides

the best overall balance of costs and benefits. But this is precisely what is

done by a requirement to promote the overall good: it requires sacrifices only in those cases in which an even greater amount of good overall is

thereby achieved. Thus, when I ask myself what sort of requirement to

provide aid I can rationally favor to be universal law, it may well be that I

144 Shelly Kagan

must favor a requirement to bring about the best possible results overall.

Anything less demanding will be inadequate.

Indeed, this implication of the aid example may have been staring us in

the face all along, even if we did not previously draw it. For any require-

ment to provide aid at all will impose costs on those who have to provide

the aid. If, nonetheless, I cannot rationally favor a maxim that would allow

me to remain indifferent to the needs of others—and this, after all, is what

Kant and kantians have always claimed—this must be because when I bear

in mind the logical possibility that I might be either benefactor or recipient,

I am led to balance the potential costs and benefits, and thus come to favor a

principle that at a minimum requires aid when the benefits to those in need are significantly greater than the costs to those providing the aid. This is

what we argued when discussing the original aid example. But this line of

thought, if it is sound at all, has no obvious stopping point short of a general

requirement to promote the overall good. The same balancing that leads me

to favor a principle requiring aid when the benefits are ‘‘significantly’’

greater than the costs will, it seems, similarly lead me to favor a principle

requiring aid whenever the benefits are greater than the costs, period. Thus, if FUL supports any requirement to provide aid at all, it should support a

requirement to promote the overall good. As always, there are a variety of objections that might be raised against

the argument I have just been sketching. But once again, my purpose is not

to offer a full defense of the claim that FUL rejects options. I merely wanted

to indicate one main line of thought that might lead one to hold that FUL

supports a general requirement to promote the overall good—despite what

many kantians seem to believe.∞∑

So let us suppose, if only for the sake of argument, that FUL does rule

out options. As we have already noted, this is still compatible with FUL

generating a deontological system. For we have not yet considered the

question of whether FUL supports constraints. If it does, of course, then despite the general requirement to do as much good as possible within the

limits of those constraints, it will still be true that certain kinds of action will be forbidden even when performing acts of the given kinds would lead to

better results overall. Thus, so long as FUL supports constraints—even if it

does reject options—it will in fact generate deontology rather than conse-

quentialism. Accordingly, our next question must be whether FUL supports

constraints.

Now it might seem obvious, in light of our earlier discussion, that FUL

does indeed support constraints. For our very first illustration of FUL at

work seemed to show that it rules out making insincere promises, or, more

Kantianism for Consequentialists 145

generally, lying. But if FUL does support a moral prohibition against lying,

doesn’t it follow trivially that it supports constraints, and thus that it sup-

ports deontology?

In point of fact, however, this conclusion does not follow so readily, for

consequentialists themselves will be among those who support a moral

prohibition against lying. Normally, after all, lying leads to worse results

overall (counting everyone’s interests equally) and so lying will typically

be forbidden—even by consequentialists. In particular, in a typical case of

false promising the overall results would be better if one refrained from

making the insincere promise. Consequentialists will thus join deontolo-

gists in forbidding me to make insincere promises on the mere grounds that

I need the money, or am in a tight spot, and so forth. And this means, of

course, that from the mere fact that FUL prohibits making the insincere

promise in such a case, we cannot yet determine whether FUL supports a

constraint against lying and making insincere promises—even when (un- like the normal case) lying would have better results overall. Thus we are

not yet in a position to tell whether FUL supports deontology or consequen-

tialism. (Similarly, of course, for normal cases of promise breaking, harm-

ing the innocent, and so forth.)

What is needed, rather, if we are to settle the matter, is a case where it is

stipulated that lying would lead to better results overall. If FUL would forbid lying even in a case of this kind, then indeed it would be clear that

FUL generates a deontological normative theory—since it would support a

moral principle that forbids lying even when lying is necessary to achieve

the best results overall. But we have not yet investigated whether FUL

prohibits lying even in cases of this sort; and I don’t think it obvious that it does.

Of course, as we have already noted, Kant himself believed that FUL (or

its equivalent) rules out all cases of lying, no matter what the circumstances. Were he right about this, obviously enough, FUL would be the basis of a

particularly strict form of deontology. But as we have also noted, many

kantians refuse to follow Kant on this matter, holding that under the right circumstances FUL can indeed pass a maxim that would permit lying (for

example, lying to a would-be murderer). So at a minimum, we shouldn’t

take it as obvious that FUL will pass no maxims that permit lying when this is necessary to promote the overall good.

Presumably, we might attempt to settle the matter by considering a

particular case where lying is stipulated to lead to better results overall, and

then testing various maxims that would permit lying in such a case—so as

to see whether any of these lie permitting maxims could pass FUL. In

146 Shelly Kagan

principle an investigation of this sort could tell us whether FUL forbids

lying even when such an act leads to better results overall. If it does, this

would show that FUL supports a constraint against lying, and thus supports

deontology rather than consequentialism.

But such an investigation would have a variety of drawbacks. First of all,

suppose we took some such maxim—say, a maxim of the form ‘‘I will lie

under such and such circumstances’’—and found that it could not pass

FUL. As we know, this would show that one should not act on that maxim.

But it would not actually show that FUL forbids lying in such cases. It would only show that one should not act on that maxim, that if lying is permitted, the reasons why it is permitted are not adequately captured in the

particular maxim being tested. It would still be possible that some other

maxim would pass FUL, a maxim that would permit lying in the case at

hand.

On the other hand, suppose we found a maxim that permitted lying in the

particular case imagined. That would of course show that it was permissible

to act on that maxim, and thus permissible to tell a lie in that case, and

thus—by hypothesis—permissible to tell a lie in at least one case where

doing so leads to better results. But this still wouldn’t necessarily constitute

a defense of consequentialism, for the fact that lying here leads to better

results might be irrelevant (or inadequate, by itself ) to explaining why the maxim passed FUL. There might well be other cases where lying would

also lead to better results, yet where telling a lie would not be permitted by the particular maxim that permitted it in the original case. In short, even if

lying is permitted in some cases where this happens to have the best results, we couldn’t necessarily conclude that it was permitted in all cases where this had the best results. So even if we did find a maxim that permitted lying

in our original case, we wouldn’t necessarily have shown that FUL supports

consequentialism. To do that, we will need to show that FUL permits lying

in all cases where this has the best results overall. But of course even this wouldn’t suffice, for it might be that FUL permits

lying in all cases where this leads to the best results but nonetheless rules out other types of actions, regardless of the results. Perhaps, for example, lying is permitted when this promotes the overall good, but there is, nonetheless, a

constraint against bodily harm to the innocent, even when this is necessary to

bring about the best results overall. If something like this were the case, then,

of course, it would still be true that FUL supports deontology. So long as

there is any constraint at all—any prohibition against performing an act with

good results overall—FUL supports deontology rather than consequential-

ism. In short, focusing on maxims concerned with lying alone will be too

Kantianism for Consequentialists 147

narrow a method of investigation to settle the question of whether FUL

supports deontology or consequentialism.

What we want to know, of course, is whether there are any actions at all,

of any sort whatsoever, that are forbidden even when performing actions of

that sort is necessary to bring about the greatest amount of good overall. If

any act, of any kind, is forbidden even when the results would be better,

then FUL supports deontology. What the consequentialist must insist,

therefore, is that any act is permissible, so long as it leads to the best results overall. But since the permissibility of an act follows so long as there is a

single maxim that passes FUL that permits the given act, what the conse-

quentialist must claim is that for each act that has the best results, there is

some maxim or the other that would permit the act, that passes FUL. Now in principle, I suppose, it could be a different maxim in each case.

But this hardly seems likely. For as we have seen, maxims that pass FUL are

supposed to do so by virtue of referring to the various features of the

situation that actually provide the agent with adequate reason for acting in

the specified manner. A valid, fully specified maxim would pick out all and

only those features of the situation that make it reasonable for the agent to

act in the given way. According to consequentialism, however, what ul- timately justifies an agent’s performing a given act is always the very same reason, namely, that the act would lead to the best results overall. Thus the

consequentialist believes that in any given case, the act that leads to the best

results is the appropriate act to perform, and the ultimate reason why it is the right act to perform is the very fact that it leads to the best results. Thus

we should expect the consequentialist to hold that the principle ‘‘act in the

way that has the best results overall’’ is universally valid, and that the quite

general maxim ‘‘I will act in the way that has the best results overall’’ will

pass FUL (no matter what the particular case at hand).

So let us consider that maxim. If it passes FUL, then, of course, it is

permissible to act on it, which means that it will always be permissible to

perform the act that has the best results overall—whatever type of act that may be, and whatever the circumstances. In short, if the consequentialist maxim (as we might call it) passes FUL, then it is never forbidden to perform the act with the best results, FUL does not support constraints, and

FUL does not support deontology.

Does the consequentialist maxim pass FUL? I believe it does. At the

very least it must be admitted that if it fails FUL it is not obvious how and

why it does so. Consider the sorts of difficulties that have plagued maxims

in our previous examples. On at least one interpretation of the false promis-

ing example we literally cannot imagine a world in which everyone makes

148 Shelly Kagan

false promises. Is there any comparable impossibility with regard to a world

where everyone acts in such a way as to produce the best results overall?

Obviously not. A world where everyone promotes the overall good is, sad

to say, highly unrealistic, but there is no conceptual impossibility involved

in trying to imagine it.

On the alternative interpretation of the false promising example, the

existence of a world where everyone makes false promises makes it more

difficult to achieve the end specified in the maxim itself through the means

specified by the maxim (that is, getting out of a tight spot by making a false

promise is less likely in a world where everyone tries to do this). When one

imagined the maxim as universal law, the maxim’s course of action became

a less effective means to the maxim’s own end. This was a kind of practical

contradiction. Is there any comparable practical contradiction involved in

imagining a world where everyone promotes the overall good? Again, the

answer is obviously not. A world where everyone promotes the overall

good is not a world that makes it more difficult to bring about the best

results overall. On the contrary, it is likely to be a world that makes it easier

to bring about the best results overall. Thus, whatever our interpretation of

the first step of the FUL test, there seems to be no reason to think that

universalizing the consequentialist maxim leads to a ‘‘contradiction in

thought.’’

Nor, so far as I can see, is there any reason to believe that universalizing

the consequentialist maxim leads to problems at the second step, generating

a ‘‘contradiction in will.’’ When we imagined a world where no one aided

others in need, this was indeed a coherent possibility, but we found we

could not will the relevant maxim to be a universal law. Given that we

ourselves could have needs (that we were unable to meet without aid from

others), it violated a principle of instrumental reasoning to favor a maxim

that if made a universal law would necessarily leave those needs unmet. But

is there any comparable violation of instrumental reason involved with

willing it to be a universal law that everyone is to bring about as much good

as possible? It is far from clear that there is.

To be sure, if it is a universal law that everyone is to bring about as much

good as possible, then there may arise cases in which I may have to make

significant sacrifices for others. From the point of view of instrumental

reasoning this is undesirable, and gives me some reason to oppose such a

requirement. But we have, of course, already considered this point. Since I

am asking what I can will to be universal law, I must also consider the

possibility that I might be the recipient of the aid. In effect, I must weigh all the potential costs against all the potential benefits, and when I do this—or

Kantianism for Consequentialists 149

so I have argued—instrumental reasoning will lead me to favor a principle

in which sacrifices are required precisely when the benefits are greater than

the costs. That is to say, instrumental reasoning will lead me to favor a

principle requiring each of us to act in the way that has the best results

overall. Are there other reasons to think that I cannot rationally favor its being a

universal law that everyone is to act in such a way as to maximize the

overall good? At the very least it is not obvious what they might be.

Of course, one might object to such a law on the very ground that it

would permit violating constraints! Intuitively, after all, certain acts are simply morally forbidden, despite their results. But promoting the overall

good might sometimes require performing acts of these intuitively unac-

ceptable kinds. Isn’t this adequate grounds for refusing to will the maxim to

be a universal law?

In point of fact, however, it is not at all clear that such intuitions are even

relevant in thinking about which maxims pass FUL. FUL, after all, was supposed to be the basis of morality, the source of the valid moral principles

(whatever they turn out to be). It can’t play this role if we are going to

presuppose various moral principles (whether directly, or by relying on moral intuitions) in determining what can, and what cannot, pass FUL. Put

another way, given the kantian account of the foundations of ethics, appeals

to moral intuitions are logically beside the point, until we have confirmed

their accuracy independently, through appeal to FUL (cf. G 4:408–10). Thus we cannot appeal to the intuitive plausibility of constraints, and use

this as a reason for claiming that principles that violate such constraints

must fail FUL. Rather, we must first decide what passes FUL—and we

must do this on independent grounds. And what this means, of course, is

that despite the intuitive appeal of constraints, we don’t yet have reason to

think that FUL generates constraints. For all that, of course, there might well be further arguments available to

those who want to claim that I cannot rationally will it to be a universal law

that everyone do the act with the best results overall. If such further consid-

erations were offered, and found to be compelling, then it would indeed turn

out that the consequentialist maxim cannot pass FUL. I certainly haven’t

attempted to discuss all possible arguments along these lines. But it must be

admitted, I think, that it isn’t obvious what these further arguments might

look like. And so I think we should conclude—even if only tentatively—

that our maxim can indeed pass FUL. Or, at a minimum, we should at least

admit that this possibility is not one that can be readily dismissed.

But if the maxim passes FUL then it is always permissible to act on it. It

150 Shelly Kagan

is always permissible to do the act that will have the best results overall.

Thus, if the consequentialist maxim passes, there are no constraints. FUL

simply doesn’t generate them.

Putting together the results of these various arguments, we can say, at a

minimum, that it should not be taken to be obvious that FUL supports a

deontological normative theory. On the contrary, there is at least some

reason to believe that FUL yields no constraints at all, despite what Kant

and most kantians have assumed. Indeed, there is some reason to believe

that FUL supports a normative theory with neither constraints nor options.

On such a theory, each of us is simply required to do as much good as

possible. But this, of course, is consequentialism.

Here is a slightly different way to see how consequentialism is supported

by FUL (assuming that the arguments we have been considering are sound).

We have just argued that despite what kantians have typically thought, it

may well be the case that kantian foundations support the claim that it is

always permissible to do the act that will have the best results overall. By itself, of course, this result (even if correct) wouldn’t yet show that we are

required to do the act with the best results. But this further conclusion would indeed follow given the earlier claim that we are required to do as

much good as possible within the limits of whatever constraints there may

be. For if we are always permitted to do the act with the best results, there are no constraints. Thus the requirement to do as much good as possible within the limits of constraints reduces to the simple requirement to do as much good as possible. Each of us is required to do the act with the best

results overall. But this, again, is precisely the claim of consequentialism.

These same basic ideas (if they are accepted) can be rearranged once

more, into an even more straightforward ‘‘proof’’ that FUL supports conse-

quentialism. To begin with, since the consequentialist maxim passes FUL,

agents are always permitted to perform the act with the best results overall. But in point of fact, contrary to what most people have thought, no other maxim will pass FUL as well∞∏ (since any maxim that permitted doing less,

or required doing something different, would run afoul of the principle of

instrumental reason, and thus could not be willed to be universal law). Thus

agents are actually required to do the act that would best promote the overall good. In short, given FUL—and assuming, of course, that the argu-

ments we have been considering are correct—everyone is required to do

the act with the best results overall, just as consequentialism claims.

Once again, it is worth emphasizing that I do not take these remarks to

constitute a full defense of the claim that kantian foundations support a

consequentialist normative theory rather than a deontological one.∞π But I

Kantianism for Consequentialists 151

hope I have said enough to show that this possibility is one that must be

taken very seriously indeed, despite the fact that Kant and almost all kant-

ians after him have rejected it (as have indeed almost all those who have

studied kantianism, whether sympathetic to it or not).

If one were to attempt to complete the project of grounding consequen-

tialism on a kantian basis, much would still need to be done. Beyond the

obvious point, that the various arguments sketched here would need to be

developed more fully (and a host of objections would need to be considered

in greater detail), the most important remaining task would be this. A conse-

quentialist theory is incomplete until combined with a theory of the good.

Knowing that we are required to do as much good as possible does not yet

generate determinate guidance until we know what makes one outcome

better or worse than another. What we need, then, is an account of the

intrinsic goods for the sake of which we should act. If the kantian account of

the foundations of morality is correct, of course, then the intrinsic goods

must be ones that we can autonomously set for ourselves as ends. Kantians

believe there are such goods, however, and so the possibility of erecting a

complete consequentialist theory on kantian foundations remains, I believe,

both appealing and important. But I won’t attempt to sketch here what an

adequate kantian theory of the good might look like. That must be left for

another occasion.

Let me return, finally, to a point noted much earlier, when we first

introduced FUL. Kant, it will be recalled, claims that FUL is itself only one

way of stating the same basic imperative. That is, he held that there are

other ways of formulating the very same categorical imperative in quite

different language. For example, at one point in the Groundwork Kant claims that the categorical imperative can also be stated like this (the for-

mula of humanity):

FH: ‘‘Act so that you use humanity, as much in your own person as in the

person of every other, always at the same time as end and never merely

as means’’ (G 4:429).

And at a different point he claims that the categorical imperative can also be

stated like this (the formula of the realm of ends):

FRE: ‘‘That all maxims ought to harmonize from one’s own legislation

into a possible realm of ends as a realm of nature’’ (G 4:436).

It certainly must be admitted that it is far from obvious that these dif-

ferent formulas are truly equivalent, generating the very same guides to

action. Indeed, not all kantians agree with Kant about the supposed equiva-

152 Shelly Kagan

lence. Of course, for that matter, it is also far from obvious how these

alternative formulas are best understood, and how they are to be applied.

Unfortunately, pursuing these related issues would involve considerable

further discussion, and so we cannot consider them here.∞∫

I do, however, want to address one final question that might naturally

arise at this point. If there are different formulations of the categorical

imperative, is there any particular justification for focusing, as I have, on

FUL, as opposed to some of the alternative formulations? Perhaps not.

After all, if they are genuinely equivalent, then they must all support the

same moral principles. And if I am right in thinking that FUL may lead to

consequentialism, then if they are equivalent the other formulas should lead

to consequentialism as well. I find that a plausible claim as well, but I won’t

attempt to defend it here.∞Ω

But Kant himself, in surveying the alternative formulations of the cate-

gorical imperative, makes an interesting remark. With regard to some of the

other formulas, Kant suggests, it might well be the case that they are more

intuitive and accessible. But if we want a strict accounting of what to do, he

says, then we should turn to FUL (G 4:436–37). I have followed Kant’s lead

in this regard, and focused on FUL itself. Of course, I have also argued that

Kant may well be mistaken about where, precisely, FUL takes us. Kantian-

ism, I have argued, represents a significant account of the foundations of

ethics. But contrary to the claims of most kantians, and Kant himself, these

foundations may well lead us to consequentialism.

notes

1. I mean here to distinguish between foundational theories and more

‘‘normative’’ theories—theories involving basic moral requirements such

as those concerning harm doing, promise keeping, and so forth. For the

distinction between these two levels of theory, see Shelly Kagan, ‘‘The

Structure of Normative Ethics,’’ Philosophical Perspectives 6 (1992): 223– 242, or Normative Ethics (Boulder: Westview Press, 1998). I will have more to say about the distinction between deontological and consequential-

ist normative theories below.

2. Two important precedents for challenging this widely held view are

David Cummiskey, Kantian Consequentialism (Oxford: Oxford University Press, 1996); and Richard Hare, ‘‘Could Kant Have Been a Utilitarian?’’ in

his Sorting Out Ethics (Oxford: Oxford University Press, 1997). The former is particularly sensitive to the details of Kant’s own position. But insofar as

Kantian Consequentialism is primarily concerned with the formula of hu-

Kantianism for Consequentialists 153

manity rather than with the universal law formulation of the categorical

imperative, the present essay is perhaps best viewed as being complemen-

tary to that work rather than simply duplicating it. I should perhaps note

explicitly that while I have qualms about various details of Cummiskey’s

arguments, I am, of course, in broad agreement with his main conclusions.

The same is largely true for Hare’s discussion as well, though I am uncon-

vinced that it is precisely utilitarianism—rather than some other conse- quentialist theory with a more complicated theory of the good—that

emerges from Kant’s account.

3. I will note, however, that Kant’s own discussion of freedom is made

complicated by his unargued assumption of incompatibilism—the claim

that freedom is incompatible with determinism (see, e.g., G 4:446–47 or

455–56)—and that this is a view that the kantian need not accept.

4. Might the line of thought that leads from autonomy to FUL support an

even stronger conclusion? If autonomy requires that I restrict myself to

acting on reason-giving principles that I can autonomously will to be uni-

versal law, does it also require that I act on all those principles that I can so will? This is an important question, but I won’t pursue it here (except to

note that the distinction between what I can will, and what I do will, will be

relevant). For simplicity, let’s continue to follow Kant’s lead and consider

FUL only in its familiar, ‘‘negative’’ formulation.

5. His reasons for claiming this are not altogether clear or persuasive. At

G 4:402 and 420–21 he seems to have in mind something like the following

disjunctive argument: (1) the validity of imperatives must be based either

on their content or on their form. But (2) considerations of content yield no

categorical imperatives, and (3) the only categorical imperative based on

form is FUL. So (4) the only categorical imperative is FUL. Now one worry

about this argument is that it is difficult to see how to reconcile (2) with the

later search (at G 4:428–29) for a formulation of the categorical imperative

based on its inevitable content, a search that supposedly successfully results

in the derivation of the formula of humanity. But since Kant holds that the

formula of humanity is itself simply another way of formulating the same

imperative as FUL, perhaps (2) could be replaced with (2%): the only cate-

gorical imperative derivable from considerations of content is equivalent to

FUL. He could then still conclude with (4%)—that the only categorical

imperative is FUL or its equivalent. The more serious difficulty with the

argument, however, is that even if we grant (1) (and it is not clear that we

should) neither (2) (or (2%)) nor (3) seems adequately defended or obviously

correct.

6. Are there any such additional principles—valid, but not derived from

154 Shelly Kagan

FUL? I don’t see why the kantian should deny their existence. Indeed, as we

will note later, many applications of FUL seem to make use of some sort of

principle of instrumental reasoning. Kant defends his own favored version

of this principle, but it is noteworthy that this defense doesn’t make refer-

ence to FUL at all (see G 4:417). So there may be at least one such further

principle, and I don’t see why there shouldn’t be others.

7. Kant speaks here of a universal law ‘‘of nature,’’ since his discussion

of the four examples actually proceeds in terms of the formula of the law of

nature (FLN)—a variant of FUL which he introduces at G 4:421. For our

purposes, however, the differences between FUL and FLN are unimportant.

8. See Christine Korsgaard, ‘‘Kant’s Formula of Universal Law,’’ re-

printed in her Creating the Kingdom of Ends (New York: Cambridge Uni- versity Press, 1996), for a fuller discussion of these and other interpreta-

tions, including a defense of the second.

9. Unfortunately, Kant seems to be confused on this point, sometimes

apparently holding the view just shown to be mistaken—that the familiar

moral rules are themselves a priori as well. (See, e.g., G 4:389, 408, or 410–12.) In any event, the claim that FUL is a priori is less clearly mis- taken, and Kant certainly believed it too (see, e.g., G 4:419–21), though

whether it is correct depends on, among other things, whether the autonomy

of reason is something that can be established a priori. 10. This has an interesting implication, which I will mention in passing.

People often take the familiar moral rules (to keep your promises, to tell the

truth, and so forth) as themselves being categorical imperatives, binding

upon everyone. But in light of what we have just noted, we must reject this

view. (We would need to reject it in any event, if we insisted on taking

seriously Kant’s claim that FUL and its alternative formulations represent

the only categorical imperative.) If the derivation of particular moral rules makes essential use of contingent empirical facts, then those rules will

themselves only be binding given the facts in question. This means that moral rules will not be binding upon all rational beings, regardless of what else is true. Thus the familiar moral rules are not categorical—since cate-

gorical imperatives must be binding upon all rational beings without condi-

tion (see G 4:416). What is true, of course, is that they are binding, nonethe- less, for those rational beings for whom the relevant empirical facts do

obtain, and in a world like ours that may well mean for all human beings

whatsoever. In particular, then, while the familiar moral rules are not cate-

gorical, they are not conditional upon the particular desires and goals of the

people involved. (It must be admitted, however, that Kant himself seems

Kantianism for Consequentialists 155

confused on this point as well, suggesting at various places that the familiar

moral rules are indeed categorical. See, e.g., G 4:389, 408, or 410–12.)

11. For one example of an argument to this effect, see Christine Kors-

gaard, ‘‘The Right to Lie: Kant on Dealing with Evil,’’ in Creating the Kingdom of Ends. 12. Again, for the distinction between the two levels of moral theory see

either ‘‘The Structure of Normative Ethics’’ or Normative Ethics. 13. I’ve argued elsewhere that, in general, deontological foundational

theories need not support deontological normative theories (and, similarly,

that consequentialist foundations needn’t support consequentialist norma-

tive theories). See Part 2 of Normative Ethics. 14. Some deontologists are absolutists with regard to these constraints,

holding the relevant types of acts to be forbidden no matter how much good

would be done—or harm avoided—by performing them. Other deontolo-

gists are moderates about constraints, believing it permissible to infringe

the constraint when enough good is at stake. For our purposes, however, the

distinction won’t be important.

15. Kant’s own views on this subject are less clear. But it is striking that

when discussing the aid example in terms of the formula of humanity Kant

concludes that each person must ‘‘aspire, as much as he can, to further the

ends of others’’ (G 4:430). This certainly looks like a denial of options.

Perhaps, then, Kant would have agreed that FUL rejects options as well,

since he believed that the formula of humanity is equivalent to FUL.

16. At least, not if we are taking the maxims to be fully specified. The consequentialist can readily admit, of course, that many other (‘‘abbrevi-

ated’’ or ‘‘shorthand’’) maxims will also pass as well—when tested against

the implicit background assumption that the act in question has good results

overall.

17. Let me quickly mention another argument that is sometimes used to

defend the claim that kantian foundations support deontology. It turns on

the distinction between perfect and imperfect duties. A perfect duty, Kant

says, ‘‘permits no exception to the advantage of inclination’’ (G 4:421 note)

and is ‘‘unremitting’’ (G 4:424); in contrast, then, an imperfect duty pre-

sumably leaves one with some latitude as to how and when it is to be

satisfied. Arguably this entails that one must never violate or otherwise

infringe a perfect duty for the mere sake of fulfilling an imperfect one. If we

then add the further assumption that the familiar duties such as the require-

ment not to lie, to keep one’s promises, not to harm the innocent, and so

forth, are perfect duties (because, supposedly, they all fall out of FUL at the

156 Shelly Kagan

first step), while the duty to aid others is merely an imperfect duty (because

it is generated only at the second step), we seem to have the desired deon-

tological conclusion that one must promote the good, but not when this

requires telling lies, harming the innocent, and so on. There are, however, a

great many problems with this argument, not the least of which is the point,

previously noted, that nothing in the account of FUL itself warrants invest-

ing the question of the stage at which a duty is generated by FUL with

anything like this kind of significance. (For further discussion of the at-

tempt to use the perfect/imperfect distinction as an argument for deontol-

ogy, see Cummiskey, Kantian Consequentialism, chapter 6.) 18. Though I will note the obvious point that if one does accept more

than one of these formulas as expressing a genuine categorical imperative,

while denying their equivalence, one must deny Kant’s claim, also pre-

viously noted, that there is exactly one categorical imperative. (Of course, it

could still be the case that one of these categorical imperatives was the most

basic, and the others could be derived from it.)

19. Though, again, see Cummiskey, Kantian Consequentialism, for a defense of the claim that it is actually consequentialism rather than deontol-

ogy that is supported by the formula of humanity.

What Is Kantian Ethics?

ALLEN W. WOOD

Kant is the most influential moral theorist of modern times. Many philoso-

phers, of whom I am one, think that of all figures in the history of ethics,

Kant did the best job of identifying what lies at the heart of moral values

and principles, providing a philosophical defense of our core moral convic-

tions, and constructing a moral theory on the basis of them. At the same

time, nobody today actually subscribes to every aspect of Kant’s thought

about morality, especially to Kant’s moral opinions on certain subjects. To

enlightened people in our day, some of Kant’s views on lying, or suicide, or

the death penalty, or the duty to obey political authority, or human sexuality

—taken literally—are unacceptable. Even some of what Kant took to be

fundamental tenets of his ethical thought, such as his theory of noumenal

free will, seem dubiously metaphysical to most level-headed people.

Some take these stumbling blocks to indicate the theory’s basic spirit. I

think that is wrong, but we Kantians have to concede this much to them, that

it is a bit anachronistic for us to insist on talking about ‘acting from duty’ and

‘categorical imperatives’, perhaps even about ‘autonomy’, ‘ends in them-

selves’, and the ‘realm of ends’. We do so because we want to link ourselves

to the spirit of the Enlightenment, of which Kant is the most influential

representative. And we are entirely right to want to link ourselves to that spirit. At the same time, we have to admit that the eighteenth-century En-

lightenment is far from where we are now, partly because of the very success

of the Enlightenment in changing people’s minds, but largely because of the

many victories of countermovements to the progressive Enlightenment tra-

dition, many of them in the second half of the twentieth century.

As a result, even in our fierce anachronism, we Kantians relate posi-

tively to Kant’s writings only because we are silently transposing their

content, hearing them in a different key, so to speak, from that in which they

were originally written. But that metaphor is still too timid. For it is neces-

sarily the entire style of our moral convictions that has changed. Kant can

be to us only as Haydn or Mozart or Boccherini is to Shostakovich or

158 Allen W. Wood

Bartók or Ellington. For this reason we owe it not only to the critics but to

ourselves to do more than Kantians usually do by way of saying what

exactly we think Kant got right, and how it should be expressed now.

I have been talking about ‘‘us Kantians’’ in the plural, but I can’t go on

pretending that I am speaking for anyone but myself. However, I do hope

that many of those favorably disposed to Kantian ethics will share the views

I put forward. No doubt others will think my interpretation of Kant is

wrong, or (what can be, but is not necessarily, the same thing) they will

disagree with what I identify as the rational kernel of Kantian ethics and

rush to the defense of what I have decided to discard as its inessential husk.

But the resulting controversies can only help clarify both our actual ethical

thinking and our view of how Kant’s ethical thought is to be appropriated.

I . ‘‘Moralism’’

Kant theorized about morality over two hundred years ago, and he wrote for

a very different audience. Even now many people associate morality very

closely with religious beliefs and commitments, but in Kant’s time this

connection was still closer, and in many quarters it was considered im-

proper even to question it, or to attempt to discuss morality at all apart from

religion. Because morality was often conceived as obedience to God’s will,

it was wholly natural for a philosopher to treat moral principles as a species

of law or imperative, even if the philosopher is not thinking of the divine

will as the author of such commands. It was a decisive step away from the

spirit of moral authoritarianism when Kant considered our own rational will

to be the author of morality’s commands. Another important point is that in

the eighteenth century moral duty was treated almost universally with what

we can only regard as a kind of innocent earnestness, whose rhetorical

expression we can receive only with condescending irony or cynical skepti-

cism.∞ Kant writes about the authority of moral commands with a kind of

assurance that no one could today.≤

A large part of the common resistance to Kantian moral theory is simply

a reaction to the rhetoric of stern ‘‘moralism’’ in which Kant presents his

theory. Let’s face it: we all know people who are ‘‘too moral by half’’ in

their own lives—and also too intrusively interested in the morality of oth-

ers. They are always ready with cloying platitudes, which they are ready to

illustrate with tediously familiar little stories that grossly oversimplify the

problems real people face in a complicated world, and at the same time they

exhibit unrelentingly a viciously narrowminded, ‘‘moralistic’’ view of the

What Is Kantian Ethics? 159

world.≥ Such people are offensive enough in our private dealings with them,

but when they turn their zeal toward public political ends, as they have in

our country during the last generation, then the results are loathsome to

contemplate. Many thinking people reject Kant’s moral philosophy because

they react to the rhetoric of his writings in the same way they do to the

offensive spirit of such private or public moralists. The instincts of such

people are healthy, but when they direct their rage at Kant, they seriously

misunderstand him.

Once we get past his eighteenth-century rhetoric, I do not think Kant’s

ethical theory really is ‘‘moralistic’’ at all in this negative sense. Morality is a complex social and psychological phenomenon, but one of its chief ele-

ments—especially when it is preached by moralists—is a peremptory

claim by certain cultural forces (often highly traditional and very regressive

ones) to authority over the lives of individuals. Above all what makes it

hard to take ‘‘moralizing’’ seriously is our understandable skepticism about

the thought that letting ‘‘morality’’ (in this sense) have its way with me

could be compatible with the attitude I should take toward my life. For as a

mature adult human being, I rightly regard that life as mine to direct accord-

ing to my own lights. I cannot permit myself to be ordered about sub-

missively by pompous moralizers enforcing outworn codes of conduct or

by religious fanatics claiming to represent God’s authority over me.

It is one of Kant’s chief merits that he articulated just those reasons for

resisting ‘‘moralism’’—but he did so in the name of ‘‘morality’’ itself. For

Kant’s basic idea is that our primary commitment should be to directing our

own lives according to our own best rational judgment, and he accordingly

reconceived the principle of morality itself as a principle of rational auton- omy. Our resistance to Kant is often a result of the understandable difficulty we have in taking seriously the revolutionary concept of morality he is

proposing. We hear only his moralizing talk, and interpret his theory of

morality as though it were just another shabby subterfuge some moralist is

using to manipulate us. In this light, I think it would help us to understand

Kant better if instead we heard his moralistic rhetoric partly as a desperate

attempt to persuade the conservative audience of his day (including the

morally conservative part of himself ) that his fundamental idea of rational

autonomy is compatible with traditional moral standards and values. In

other words, Kant has good reason to fear that ‘‘moralists’’ (in the sense we

rightly find objectionable) will regard his theory as dangerously libertine.

We should perceive his moral sternness as an attempt (perhaps not wholly

successful) to disarm such fears. One of the things about Kantian ethics that

should most appeal to us, therefore, is that it helps us explain and justify our

160 Allen W. Wood

resistance to the very rhetoric in Kant’s own writings that we often find

most offensive. We rightly resist any morality based on tradition or custom

(perhaps sanctified by religious humbug) as an affront to our worth as self-

directing adult human beings, who have the capacity to think for ourselves

and the right to regard our own free thinking as the final arbiter of whether

morality (or anything else) deserves our allegiance.

Of course ‘‘moralism’’ is out of fashion for bad reasons as well as good

ones. People also reject it on the basis of various shabby rationalizations and

self-serving illusions. Kant was aware of many of them in their eighteenth-

century forms. Many people succumb to the sentimental notion that in

directing our lives our ‘‘natural feelings’’ can always be trusted. Or they hold

the shallow belief that there really are no reasons except selfish reasons, or

the vicious philosophy that things go best for all if we forget about other

people and just ‘‘look out for number one.’’ Or they say that whether this is

‘‘best’’ or not, it’s the way things are and one has to be a ‘‘realist’’ about life.

Or perhaps they subscribe to the absurd Romantic (and Nietzschean) idea

that some people are ‘‘special’’ (‘‘higher’’ or more ‘‘creative’’—that is, that

some people are like themselves as they imagine themselves to be), and

special people do not need to respect other people’s rights or concern

themselves with their needs.

Since Kant’s day, the self-deceptive appeal of these ideas to our selfish

complacency has been fueled by misunderstandings of Darwin’s theory of

evolution and by the perceived failure of all sorts of attempts, whether

religious or secular in inspiration, to make the world a better place. I con-

fess that Kantian ethics also appeals to me because it is ‘‘moralistic,’’ in the sense that it gives no quarter to any of these widespread and pernicious

notions. Kant had a conception of human nature that was in many ways

dark and cynical, but that only made him the more insistent that we must not

give in to the temptation to treat what is done as the proper standard for

what should be done. Kant’s moralistic rhetoric, I submit, is not offensive

(but is even appealing) if we perceive it as a reflection of his recognition

that human affairs are very far from being what they should be, and an

attack on the self-complacency with which too many people rationalize

their shortsighted and selfish ways of life.

II . Moral Deliberation

As Kant’s theory presents it, morality is fundamentally a way of thinking

about what to do. It is a mode of deliberation that could be described as

What Is Kantian Ethics? 161

‘free’ or ‘ultimate’. It is free in that it is unrestricted regarding the kind of reason one offers for a course of action, and ultimate in being the last instance of decision about how we should live our lives. Moral deliberation is, first, not restricted to reasons arising from instrumental reasoning about particular ends we have contingently set. It may consider the best way to

attain an end, but it is fundamentally about which ends should be ultimate

and primary in our lives. Second, it is not restricted to considering self-

interested ends, such as my own welfare, or the welfare of some particular

group to which I belong or which I happen to care about. Moral deliberation

provides grounds for being concerned about my own welfare and happi-

ness, but it is not limited to prudential reasoning about how my—or any- one’s, or everyone’s—happiness is best to be promoted. Moral deliberation

is concerned also with knowing why, and deciding under what conditions,

anyone’s happiness is a worthy object of pursuit.

Kant treats moral deliberation as a species (even a paradigm) of rational

deliberation. The Kantian conception of moral deliberation thus goes

against a familiar dogma that rational deliberation is solely about means

and can never be about ends (whose adoption is not supposed to be subject

to rational considerations at all). The basic error behind this dogma is the

assumption that practical reasoning is always grounded on desire. It is one

of Kant’s chief insights that there are some desires we have only because we

have reasons to have them, and these reasons are independent of any desire. Some desires (such as hunger) we just find in ourselves. Others we have

for reasons (even hunger is a rational desire to the extent that we recognize it as expressing our need for nutrition). Some rational desires are nonmoral

and some are moral in content. Morality, in fact, is about the objects of all

these desires and the reasons grounding them. Moral deliberation creates

rational desires by making these reasons evident to us. It focuses on a

certain type of ultimate value and ultimate rational concern about how we

should act. This value, as Kantian theory conceives it, is the worth of the rational beings who are moral agents, who have the value Kant describes by

saying they have ‘dignity’ and by calling them ‘ends in themselves’. Moral

deliberation is usually concerned with someone’s welfare or happiness, but

not only with the agent’s own welfare, or with the welfare of those agents

she may happen to like or with whom she may sympathize. Moral delibera-

tion regards the value of happiness or welfare as grounded on something

deeper, namely on the objective worth of the beings themselves whose

happiness is in question. Because other rational beings have the same worth

that I do, moral deliberation therefore necessarily takes the welfare of

others into account as well as my own.

162 Allen W. Wood

This feature of moral deliberation has led some philosophers to think of

morality as a kind of expanded or collective prudence, based ultimately on a

single end, the common good or greatest happiness of all. John Rawls

famously criticized this type of moral theory for overlooking (or under-

estimating the importance of ) ‘‘the distinction between persons.’’∂ The ar-

gument associated with this last slogan is often taken to be the most basic

and authentically Kantian reason for resisting the ‘‘greatest happiness’’

morality. But I think we badly misunderstand Rawls’s Kantian objection to

utilitarianism if we depict the disagreement between Kantian and utilitarian

views as fundamentally one in which the utilitarians favor human commu-

nity or the unity of human ends while the Kantians ignore these values or

give them second place to the abstract rights of isolated individuals. This is

a basic misunderstanding. Kant’s own basic principle is that we must see

ourselves as part of an ideal community, a ‘‘realm of ends’’ in which all

human ends are united and mutually supporting.∑ The most basic object of

value in Kantian ethics is therefore the rational being not as an isolated

individual but as the object of a set of common laws uniting rational beings into a community. The right way to put Rawls’s Kantian criticism is rather to say that it

charges utilitarianism with misunderstanding the true nature of that human

community which is (for both Kantianism and utilitarianism) the primary

value. Kant holds that people should be united by ends they freely and

rationally pursue in common. The fundamental aim of moral theory is to

determine the principles according to which to achieve that agreement,

hence the right kind of community. The point of Rawls’s criticism is that when utilitarianism conceives of the right end as simply the collective

maximization of personal good, it thinks of human community as if it were

best guided by a single prudential reasoner (in effect, a benevolent despot)

who dictates actions to an aggregation of will-lessly obedient (or perhaps

craftily manipulated) subjects. Respect for individual human beings, and

for their role as independent judges, equal participants in the process of

deciding for the community, must then be taken account of subsequently, as

something that the utilitarian dictator integrates (of necessity, paternalisti-

cally) into the collectively prudential calculations as one factor among

many in what is most beneficial to them. For the Kantian, however, a

community of rational beings must be conceived from the ground up as the

rational agreement of a plurality of distinct and equal persons who freely

choose to unite their ends on terms that respect each one’s autonomy. The

crucial thing, therefore, is not to determine the means to a single given

What Is Kantian Ethics? 163

collective end. It is something more basic than that: It is to determine the

principles of association through which any rational system of collective

ends is to be set.

This point has far-reaching implications for the way we should reason

about collective goods. We cannot simply integrate ‘‘process values,’’ asso-

ciated with the way ends are set and actions chosen, into the ends people

pursue, as if these were simply one kind of good to be weighed against

others in determining the content of the ends and the most efficient means to

them. It gets the fundamental role of these values wrong to ascribe ‘‘utility’’

to the fact that an action keeps a promise or an institutional arrangement

respects the rights of those who participate in it, and then to treat these as

‘‘utilities’’ to be weighed against other ‘‘utilities’’ in determining the best

‘‘overall consequences’’ of actions or social arrangements.∏ To flatten out

moral deliberation into a system of collective prudence, with all goods

treated as products of action, is to ignore the fact that the value of individual

agents and of certain kinds of community among them has the first claim on us in structuring the way we should deliberate both as individuals and as

communities.

Both the ground and the nature of these objections to consequentialist

theories are distorted when they are represented as a mindless devotion to

rule-following under the traditional title of ‘‘deontology’’—as though the

Kantian position is that the rules themselves (although they represent noth-

ing of value) have some overriding authority to which we should be pre-

pared to sacrifice everything that is of value. Kant’s theory is based rather

on the overriding value of rational beings themselves, whose rational

choice, on terms of mutual respect, is to be the final authority over what

might have enough value to make it worth pursuing as an end of action.

III . Three Features of Morali ty

‘‘Morality,’’ as a kind of rational deliberation, has three crucial features

from which Kant’s theory proceeds. The first is that it considers on their merits all reasons for acting, whatever their character. Thus although Kant holds that every action is for the sake of some end that is to be brought

about, he also recognizes that the most fundamental reasons are those that

do not presuppose an end but ground the setting of the highest ends we should bring about. The Kantian way of putting this point is that the impera-

tive of moral deliberation is ‘‘categorical’’ rather than ‘‘hypothetical.’’ This

164 Allen W. Wood

leads to Kant’s first formulation of the principle of morality, the formula of universal law (FUL) and its variant, the formula of the law of nature (FLN) (G 4:421).

The second feature, in one way still formal but in another way the most substantive value thesis on which Kantian ethics rests, is that it is concerned

with the reasons we have for taking account of the standpoint of others, hence for caring about their welfare as well as our own, and choosing on

principle to unite that standpoint with ours under common laws for com-

mon ends. This presupposes that as a rational agent I have good reasons for

respecting other rational agents, and treating their existence and their wel-

fare as having objective value in principle equal to my own. From these

considerations arises Kant’s second formula of the principle of morality, the

formula of humanity as an end in itself (FH) (G 4:429) The third feature of morality, easier to overlook but just as important as

the other two, is that it motivates us by appealing to our conception of who we are and to incentives of self-worth associated with this conception. This leads to Kant’s third (and most definitive) formula of the moral law, the

formula of autonomy (FA), and its variant, the associated formula of the realm of ends (FRE) (G 4:431, 433). It is essential to moral deliberation that the fundamental reason I have

for following it is that in my own eyes I will have less worth if I do not. A being that always did the morally right thing, and always desired to do what

it does, would nevertheless not be a moral agent at all if it did not think (or

had no tendency to be moved by the thought) that it had greater self-worth

for so acting and less self-worth if it had not so acted.π This is why Kant

begins the Groundwork by appealing to the esteem we feel for the good will, and regards the moral law as adequately grounded only when it has

been shown how obedience to it is associated with our self-respect (G 4:393–99, 431–38). It is also why Kant regards autonomy of the will and

the dignity of self-legislation as the only possible grounds for a categorical

obligation.

IV. Morali ty as Social Custom

One common criticism of moral theories (both Kantian and non-Kantian) is

based on the thesis that morality is not fundamentally a kind of rational deliberation at all. Morality, it is said, has its origin not in individual ratioci-

nation but in social customs and the shared attitudes they involve. A society

has certain rules of conduct whose observance involves the subordination

What Is Kantian Ethics? 165

of individual preferences, and even individual well-being, to what is re-

garded as the good of the whole. Failure to comply with those rules subjects

the offender to disapproval and contempt in the opinion of others. Some

theories of the origin of morality locate the truly original expectations and

opinions more precisely, in the threat of paternal disapproval of the child, or

even in the mother’s refusal of the breast to the infant, which subsequently

become ‘‘internalized’’ or take the form of the ‘‘superego’’ or ‘‘ego-ideal’’

standing over against the real person and her interests. Others fasten on the

variety and contingency of social customs, using this as an argument that

we should not expect moral principles to be universal or grounded in any-

thing we should dignify with the name of ‘reason’. Some of these ap-

proaches to morality insist that moral philosophy, instead of adopting the

intellectualistic and individualistic form of practical deliberation, should

orient itself toward this more basic ‘‘ethical life’’ (Sittlichkeit), or lived social practices, and the set of traits, feelings, and ‘‘thick’’ self-conceptions

corresponding to them.∫

It may come as a surprise to some who argue this way that Kant largely

agrees with them as far as the historical origins and social roots of morality

are concerned. Yet he holds that this does not support their objections to his

kind of moral philosophy, but in fact discredits them. In his essay Conjec- tural Beginning of Human History, Kant locates the primal origin of moral- ity in the act of sexual refusal through the concealment of desired genital organs. This device excites sexual desire by transforming its object into something imaginary, and simultaneously awakens respect for the person by closing off access to what is desired (MA 8:113). The effect is to bring

into being a kind of social power over individual agents located within

these agents themselves. By means of this power their behavior can be

controlled through their sense of their own worth and the worth of others

and through an experience of conflict between that kind of worth and their

natural desires. Kant calls this new state of consciousness the sense of

‘‘decency’’ (Sittsamkeit) or ‘‘propriety’’ (Anständigkeit). By conforming to accepted customs (Sitten), or acting in a manner that is deemed fitting (anstehend), people arouse the esteem of others, which they value as a condition of their sense of self-worth. Kant regards such behavior as closely

tied both to people’s competitive impulses (their ‘‘unsociable sociability’’)

and to their propensity to deceive both others and themselves, by conceal-

ing what would lead others to think less of them and by pretending to

supposed merits they do not possess.

Kant thereby recognizes that the psychological and historical origins of

morality are something toward which thinking people are right to adopt a

166 Allen W. Wood

critical and even a highly ambivalent attitude. Kant thus recognizes some of

the good reasons why reflective people are rightly skeptical of ‘‘morality’’

and ‘‘moralism.’’ He acknowledges the continuing dominance of these pat-

terns in social life when he treats ‘‘semblance’’ (Schein), ‘‘illusion’’ (Täu- schung), and ‘‘deception’’ (Betrug) as characteristic of human sociability, and even as increasingly characteristic of it as it becomes more fully devel-

oped or ‘‘civilized’’ (VA 7:149–50).Ω

No doubt some of the most conspicuous examples of bad conduct involve

the defiance of what is socially approved, whether the infringement of

people’s recognized rights or the disregard of their accepted claims to

respect and concern. A view that roots morality in social practices and

customary attitudes has no trouble giving an account of what is wrong with

such conduct. It has a harder time, though, in accounting for what is objec-

tionable in the evils that spring precisely from this Sittlichkeit itself, from the conformity to accepted customs and the display of traits recognized as

‘‘virtues’’ and from the harmonious relation to all those feelings and ‘‘thick’’

conceptions that express the life of an existing community. If the basic

principles of existing communities really embodied a decent approximation

of the way people should think about themselves and should treat one

another, then such approaches to morality might get things at least approx-

imately right. But as long as the truth remains closer to the reverse of this—

as long as our societies remain scenes of massive oppression, injustice, and

corruption—moral philosophy does better by encouraging us to alienate

ourselves from the ethical life of our existing communities and instead to

take as fundamental a purely rational deliberative standpoint. It is therefore

one of the chief advantages of Kantian ethics that it is abstractly rationalistic and self-alienating in precisely this way.

Kantian ethics is ‘‘individualistic’’ in the sense that it takes the capacity

for rational deliberation in individuals to be the ground of the moral life, or

even to be the ground of all value whatever. It is also individualistic in its

assertion of the importance of the rights of human individuals and in its

rejection of notions of collective moral responsibility or guilt. But it is not

individualistic at all in several other important respects in which it is too

often thought to be. First, as we have just seen, Kantian ethics recognizes that

the psychological origins of morality are social. Kant agrees with Rousseau

in regarding moral evil (what Kant calls the ‘radical propensity to evil in

human nature’) as a product of our social condition (R 6:27, 95–96). Second,

and just for this reason, Kant also holds that the struggle against evil cannot

be effective unless it is social. He does not think that individuals struggling

on their own against their evil propensity will make much progress toward

What Is Kantian Ethics? 167

good (R 6:96–102). Third, the final end of morality is social rather than

individual: it is a realm of ends, in which the ends of all rational beings would

be systematically united and mutually supporting (G 4:433–35).

For Kant the clearest model of a realm of ends in ordinary human life is

friendship, in which (he says) friends unite their ends into a collective end

in which their individual happinesses are swallowed up (MS 6:469–72, VE

27:422–23, 675–77). Genuine friendship also involves a commitment to

the ideal of perfect friendship, in which human ends are united in this

way. Because it involves this commitment, friendship, for Kant, makes us

worthy of happiness (whether or not it actually makes us happy); and we

have an ethical duty to enter into relations of friendship with others (MS

6:471). These fundamentally anti-individualistic strains in Kant’s ethics

have often been neglected by his sympathizers and his critics alike.

V. Morali ty as a Doctrine of Duties

Kant’s aim in the Metaphysics of Morals was to present a ‘‘doctrine of duties’’—a system of moral obligations that results from applying this

principle to human nature (G 4:388, MS 6:216–17). No doubt we find it

harder than Kant or his audience did to think about our lives as something to

be structured by our ‘‘duties’’—as though we should carry around a little

handbook of moral duties for our edification and consult it hourly to decide

what we should be doing. But down to this day it is also sadly fashionable in

moral philosophy, both Kantian and non-Kantian, to think that moral theory

must consist in some kind of rational decision procedure (such as some

version of the principle of utility or Kant’s FUL or FLN), through which we

could reckon up what we ought to do in every situation that presents itself to

us. An ethical theory then becomes a kind of meat grinder into which we

feed empirical facts, turn the crank, and out comes the series of acts we

ought to perform, one after the next, like neat little sausages on a string. The

position known as ‘‘Kantian constructivism,’’ which emphasizes the FUL

(or FLN) and regards moral goodness or rightness properties constructed

through the application of a ‘‘CI-procedure,’’ is a sad example of this re-

pellent picture, in which an abstract, hidebound, and often counterintuitive

process of formulating maxims and testing them for universalizability is

supposed to take over our lives, tell us at every juncture what to do, and

leave us no room to direct our conduct in the wide variety of ways that

intelligent people actually do direct it.∞≠

Kant’s famous four examples in the Groundwork (G 4:421–25) are

168 Allen W. Wood

often misunderstood as his chief contribution to the theory of ordinary

moral deliberation. But there is no reason to regard them as intended in any

such way. They occur in the course of Kant’s development of a system of

formulations of the moral law, and they are intended to illustrate the first

and most abstract formulation (FUL or FLN), so that the reader may see

how that formulation, and the reasoning involved in it, might relate to some

relatively familiar cases of moral duties and temptations to violate them. It

is absurdly hasty to jump to the conclusion that a philosopher would think

that illustrations suitable to that purpose are also a universal model for all

moral reasoning.

Kant gave a very different account when he actually came to work out

his system of duties in the Doctrine of Virtue. There he redescribed our

ethical duties as ends (MS 6:382–86, 394–95). He saw the moral life not as a matter of applying a decision procedure, but of devoting ourselves to

certain ends, working out our own personal priorities among them (with a

good deal of latitude, or Spielraum, about this) and making specific judg- ments about how best to pursue them under particular circumstances. He

also conceived of these ends, regarded as objects of morality, not as general

categories of good to be maximized (as utilitarians want to maximize the

sum or average of pleasure over pain) but rather as particular ends set under

contingent conditions—for instance, as the development of a certain talent

or capacity in ourselves or the promotion of the happiness of a certain

individual or group (MS 6:386–88). This is why Kantian ethics regards

both the choice of the specific ends and the order of priorities among ends as

necessarily involving latitude—as fundamentally a matter of individual

choice (MS 6:388–94).

Following Kant’s theory, the choices between ends and maxims are

constrained only by the enforceable rights of other people and by specific

relationships in which we stand to specific individuals (through relation-

ships of family or friendship or through professional obligations) (MS

6:468–69). In his personal moral opinions, Kant is infamous for his defense

of certain inflexible moral rules (against lying and suicide, for example).

But his theory clearly allows for exceptions to moral rules—treating excep- tivae as one of the twelve fundamental categories of moral thinking (KpV 5:66). The ‘‘casuistical questions’’ discussed in the Metaphysics of Morals are mostly questions about when we should or may make exceptions to

common moral rules (conspicuously among them, rules about lying and

suicide). On many of the issues raised in these questions Kant does not

come to any firm or general conclusion, regarding the proper reflection on

What Is Kantian Ethics? 169

these questions as ‘‘not so much a doctrine about how to find something as rather a practice in how to seek truth’’ (MS 6:411). The spirit of Kant’s theory is open-textured and flexible regarding the scope and demanding-

ness of moral duties.∞∞

VI. The Supreme Principle of Morali ty

Kant’s aim in the Groundwork was to ‘‘seek out and establish the supreme principle of morality’’ (G 4:392). It may seem anachronistic to think of our

moral ends as grounded in a single ‘‘fundamental principle of morality.’’ We

may think of moral values as an unordered plurality, and insist that different

people might subscribe to different moral principles, which would seem to

be incompatible with regarding all moral deliberation as falling under a

single fundamental principle. But in recognizing a fundamental principle of

morality, Kantian ethics really has no such implications. For Kant’s theory

itself directs us to pursue an indeterminate plurality of ends, and to apply

the fundamental moral principle through a plurality of moral laws, princi-

ples, and precepts. Ideally, all could be arrived at by applying the supreme

principle to empirical facts about human nature and the human condition,

many of them only imperfectly binding, and nearly all of them admitting of

exceptions. Such a derivation, however, is far removed from everyday life,

although as an ideal it can sometimes play an important role in our deeper

critical reflection on what principles we should follow and what ends we

ought to pursue.

The point of looking for a supreme principle can be seen when we reflect

on the fact that morality arises out of social custom or propriety when

people begin to think for themselves (MA 8:144–46). Thinking for oneself is a prerogative, even a duty, of rational adulthood; Kant thinks of it as a

critical historical development not only in the life of an individual but, even

more significantly, in the collective historical life of human civilization.

Kant’s name for this historical crisis is ‘enlightenment’ (WA 8:35). He

characterizes the thinking of enlightenment by providing three rules, or

‘‘maxims,’’ for its successful practice: (1) think for yourself; (2) think from

the standpoint of everyone else, and (3) think consistently (KpV 5:294–95;

VA 7:200, 228–29, VL 9:57). To ‘‘reason’’ is to think for oneself, to draw

the ultimate source of one’s acts and judgments from oneself rather than

taking them from the opinions of others or even from one’s own feelings

and desires, regarded uncritically as something simply given. The first

170 Allen W. Wood

maxim of rational thinking also explains how we are to understand Kant’s

insistence that the principle of morality is a priori, a thesis that often proves a stumbling block to the acceptance of Kantian ethics because of empiricist

confusions and prejudices. All cognition for Kant is the result of our fac-

ulties’ operating on what is given to us in sensible intuition. A proposition is

known a priori for Kant if its content is determined not by what is given to us from outside but through the exercise of our own faculties (KrV A1/B1).

To say that the fundamental principle of morality is a priori is therefore to say that its source lies in my own critical reflection on the reasons I have for

acting, rather than in commands coming from outside me, or on desires I

happen to find in myself prior to any rational reflection, rather than produce

in myself through considering the reasons why their objects are valuable

and worth pursuing. The claim that there is an a priori practical principle means that as reasonable beings we are always capable of reflecting crit-

ically on what is given, and accepting or rejecting it for reasons we are

capable of recognizing and validating through our own thinking.

Thinking for oneself is genuine thinking only if it subjects itself to the

authority of reasons, which means grounds that are equally valid for all

thinkers. For finite and fallible beings such as ourselves, such grounds can

be found and critically certified only if one tests one’s thoughts by adopting

the standpoint of others—ideally, of all others, since genuine reasons must

be equally valid for all. In the second maxim of thinking, Kant is asserting

that we can do this only through coming to understand the viewpoint of

others by communicating with them, and in a manner in which all are free to

express their viewpoints freely. It is for this reason that he declares freedom

of communication to be a necessary condition for the very existence of

reason itself (KrV A738/B766). The universal validity of reason also de-

pends on uniting the thinking of all standpoints into a set of principles that

can claim to be grounded for all of them. This is what Kant means by the

third maxim, that of thinking ‘‘consistently’’—which he declares to be the

most difficult of the three maxims to apply successfully (KpV 5:295).

This free, unconditional, unified, and critical thinking from a universal

standpoint is a never-ending process. Anything concrete that we identify as

a ‘‘principle’’ for such thinking enjoys this status only provisionally; our

thinking should continue to test and correct it, or at any rate to refine,

reinterpret, and rearticulate it in response to new circumstances or to view-

points we still have not adequately considered. Philosophical formulations

of such a principle are really placeholders for something that philosophers

and reflective individuals will forever seek after, collectively as well as

individually.

What Is Kantian Ethics? 171

VII. Formulations of the Moral Principle

In the Groundwork Kant formulates the moral law in three different ways. They constitute a developing progression, and correspond closely (as I have

already indicated) to the three features of moral deliberation distinguished

above. Kant begins with the form of the law, as a categorical imperative,

which leads to FUL and FLN: the requirement that every maxim be conso-

nant with the form of universal law (or law of nature). Then he takes up the

matter of the law, or the end that could motivate rational obedience to a

categorical imperative, which is found in FH, based on the dignity of hu-

manity or rational nature as an existing end in itself. Putting together the

concept of universal law with that of the dignity of rational will, Kant then

derives the final and most comprehensive formula, the idea of every ra-

tional will as universally legislative (FA), and the moral law as grounded on

autonomy. Morality is conceived as a system of laws that, if universally

followed, would constitute the community of rational beings as a ‘‘realm of

ends,’’ a mutually supporting purposive system in which the ends of all

rational beings form a harmonious community (FRE).

The fact that Kant develops the moral law in this systematic way entails that the later formulas of the law are in general both more adequate and

more definitive expressions of the principle than the first formula—the

well-known FUL or FLN and its much-discussed universalizability tests for

maxims. Thus when Kant derives his system of ethical duties in the Meta- physics of Morals, it is the worth of humanity as an end in itself, not the testing of maxims for universalizability, to which he appeals.∞≤ When Kant

presents a ‘‘universal formula’’ of the moral law in the Groundwork, the Critique of Practical Reason, and the Metaphysics of Morals, the formula he uses is that of autonomy (FA), or the adoption of maxims that not only

might be willed as universal laws, but include in themselves the volition that they can actually hold as part of a system of universal laws.∞≥ In the final

section of the Groundwork Kant provides a deduction of the moral law through its equivalence with the freedom of the rational will that must be

presupposed in all our actions, making use once again of FA.∞∂ Kant never

undertakes any unconditional validation of FUL or FLN, apart from its

development into and subsumption under the principle of autonomy (FA).∞∑

The most important idea in Kantian ethics is that the authority of moral

deliberation is based on the autonomy of the agent’s rational will. This idea

rests fundamentally on the assertion of an objective value: the equal dignity

of all rational beings as ends in themselves. The best picture of what moral

deliberation seeks is that of the realm of ends, or a rational community in

172 Allen W. Wood

which all the freely set ends of human beings ultimately converge. The

most useful formulation of the fundamental principle for generating moral

rules is the principle that the dignity of every rational being is entitled to

equal and unlimited respect. By contrast, for Kantian theory the practice of

testing maxims for universalizability has a relatively limited role to play in

moral deliberation.

The overemphasis on FUL, both by Kant’s critics and by his sympathiz-

ers, has had a mischievous effect on the interpretation, and therefore the

reception, of Kantian ethics. On the one hand, it has misled many of Kant’s

sympathizers into thinking that what is distinctive and valuable about his

ethical theory consists in some distinctive ‘‘decision procedure’’ rather than

in its substantive conception of rational nature and autonomy as the real

grounds of ethical value. Kant’s own quite limited use of the universalizabil-

ity tests (in his complex exposition of the moral principle in the Ground- work) is then ignored, and the tests are treated as if they were some sort of universal moral decision procedure or algorithm (a sort of Kantian response

to the equally bogus act-utilitarian project of precisely calculating and

comparing all the felicific tendencies of all the practical options open to us in

any situation). On the other hand, it has misled Kant’s critics into thinking

that when they find defects in Kant’s first formula when it is regarded as a

universal and self-sufficient moral decision procedure, they have discovered

a good reason simply to dismiss Kantian ethics as a whole without further

ado.

If the main issue about Kantian ethics were whether the universalizabil-

ity test for maxims is a satisfactory universal algorithm for all moral delib-

eration, then the right assessment of Kant’s ethical theory would be that it is

pretty worthless. There are several reasons why the universalizability tests

are not up to playing any such role. First, since the tests are suited only to

testing individual maxims, one by one, for permissibility, they can never

yield anything like a positive moral rule or duty. The test never enables you

to conclude that suicide or making false promises is wrong, but only that it

is wrong to perform these actions on the specific maxims tested in the

Groundwork; for all the universalizability tests might ever show, there might be other, wholly universalizable maxims on which acts of suicide or

lying promises could be made. Second, the universalizability tests are noto-

riously subject to false negatives—entirely innocent maxims that neverthe-

less fail the tests. Take the maxim that whenever I arrive at a doorway at the

same time as another, I will always yield right of way to the other and will

never go through first. There is nothing wrong with that maxim, but it could

never be adopted by all agents as a universal law. The problem here is

What Is Kantian Ethics? 173

simply that there are many maxims that cannot themselves be made (or

willed as) universal laws (or laws of nature) but also do not violate univer-

sal moral laws (on any plausible conception of what these might be). Third,

the universalizability tests are also threatened with false positives—max-

ims that pass the tests but are impermissible on any reasonable moral view.

The most basic reason for this problem is that any action is intentional (or

follows ‘maxims’) on many different descriptions and at many levels of

generality or specificity. The universalizability tests by themselves provide

no criterion for determining how, or at what level of specificity, the maxim

is to be formulated, or (therefore) which intentional features of the action

are relevant to its moral evaluation. As Hegel correctly put it, the univer-

salizability test would be fine if we already had determinate moral laws, but

the test itself can never generate any such principles.∞∏

None of these problems is fatal to Kantian ethics when FUL, FLN, and

the associated universalizability tests are used in the limited way they are

actually employed by Kant in the Second Section of the Groundwork. There he says explicitly that when we find ourselves tempted by a maxim

that violates the moral law, we are aware of ‘‘willing the opposite of our

maxim to hold as a universal law’’ (G 4:425). In other words, Kant intends

us to presuppose determinate moral laws and categories of duty deriving

from the autonomous will. The aim of the tests, as Kant tells us, is to

illustrate his first (preliminary) formulation of the moral principle. Through

his illustrations, we recognize the contradiction in our will that would result

if we tried to integrate these specific unlawful maxims into what we will as

universal laws. In these examples, Kant is not attempting to use FUL (or FLN) as an algorithm for generating all moral principles or making all

moral decisions. The examples are only a device for enabling us to see how

certain maxims that we already recognize as contrary to duty can be ex-

hibited as unlawful in the light of the (first and most abstract) formula of the

law that he has just derived.

The same four examples are considered again a bit later in light of the

formula of humanity as end in itself (FH). This later treatment is in every

case a more transparent and cogent way of motivating the general principle

of duty involved in the examples. Kant grounds the prohibition of suicide,

for instance, on respect for humanity in our own person, and the duty

against deceiving others on respect for the other and the requirement that

we adopt only ends that others can rationally share. Whether or not these

considerations ground the particular duties cited (I think they equally

ground large categories of exceptions to the prohibition on suicide, for

example), they also involve an appeal to substantive values on which deter-

174 Allen W. Wood

minate positive duties could be based—something that cannot be said

about the universalizability tests.

Why don’t readers of the Groundwork see this, and direct more attention to Kant’s more authentic derivations of duties? The reason, I think, is that it

is all too evident that arguments from the formula of humanity do not

provide even the appearance of an ethical sausage machine—a universal

algorithm for grinding out what any agent whatever should do under any

conceivable set of circumstances. Although it may sometimes be self-

evident what the dignity of humanity requires of us, sometimes it is not.

Sometimes this is a matter of dispute, and clearly there is no simple, general

way of resolving such disputes by anything like a mathematical calculation

or universal decision procedure. Hence readers who are looking for the

universal moral algorithm (in spite of Kant’s lack of an intention ever to

provide one) are not tempted to think they have found it in FH, but they can

integrate a certain interpretation of FUL (and FLN) into their delusions

about what moral philosophy is supposed to be. This misguided expecta-

tion, however, only sets up Kant’s theory for quick dismissal when it is soon

recognized that these formulas fail to satisfy the unreasonable demands

being made on them.

If we rest our theory of duties on FH, as Kant actually does in the

Metaphysics of Morals, then we have to admit that the fundamental princi- ple of morality yields no universal decision procedure for all cases. Moral

deliberation generates moral rules applicable in particular cases only in a

loose fashion, leaving (as Kant himself says) considerable ‘‘play-room’’ for

individual discretion and judgment (MS 6:390). It will also depend heavily

on what Kant called ‘‘practical anthropology,’’ that is, on our fallible, con-

stantly changing, and always deeply problematic knowledge of what hu-

man beings and the human predicament are like (both in general and under

specific social and historical conditions).

A major source of error here is a common misconception about what

moral theory is for, and especially about the function of the fundamental

principle in such a theory. The function of a fundamental principle can

never be directly to settle difficult moral issues; it can serve only to provide

the right general framework in which moral rules and controversial issues

should be raised and discussed. Even then, any formulation of it must be

regarded as provisional—an object of constant critical reflection and con-

tinual reinterpretation and rearticulation. The rethinking of traditional mis-

interpretations of Kant’s own formulations of the moral law and their rela-

tion to one another is intended here as only the beginning of such a process,

which of course should go beyond getting Kant right and begin to say how

What Is Kantian Ethics? 175

even his concepts of ‘universal law’, ‘rational nature’, and ‘autonomy’

might be reinterpreted or revised.∞π

VIII. Supernatural Freedom

I will conclude by taking up a vexed topic I barely touched on at the begin-

ning, which becomes a major theme in the Third Section of the Groundwork: Kant’s thesis that morality depends on transcendental or noumenal freedom.

I think it was an important insight on Kant’s part to see free agency as a

special kind of causality, namely a causality that acts under normative

principles, hence a capacity to choose between alternatives according to

one’s judgment about which alternative is permitted or required by a norm. I

also think he was correct to argue, on the basis of an acute analysis of what

rational normativity involves, that if we are free in this sense, then we are

bound by the moral law whose content he developed through FUL, FLN,

FH, FA, and FRE. But a great many of us today regard all our rational

capacities, including those required to give ourselves moral laws that are

categorically binding, as belonging to what we must ascribe to ourselves as

beings of nature. We are animals who have evolved through a process of

natural selection, and all our rational capacities must in principle be wholly

explainable through causal laws of nature. To the extent that we believe this,

we do not need to resolve the metaphysical problem of freedom as Kant did,

by transporting ourselves into an unknowable noumenal world. If our posi-

tion is as I have described it, then morality should be grounded on freedom,

as Kant thought, but on the sort of freedom we recognize ourselves as having

as part of our equipment as the rather remarkable but contingently evolved

natural animals that we are, and not on the sort of unknowable supernatural

freedom Kant thought we must postulate if we are to think of ourselves as

free at all.

Of course I do not mean to deny that the metaphysical problem of free

will is a genuine problem. I especially want to make it clear that I am not

representing myself as having some final solution to it. We still need to

investigate empirically the nature and biological underpinnings of our ra-

tional capacities. It will be no trivial task to reconcile our scientific theories

about ourselves with the freedom and reason we necessarily presuppose in

ourselves even in the act of undertaking such investigations. I am not

convinced that we have yet reached that theoretical goal, and I think that

there are philosophical as well as empirical obstacles still standing in the

way of our human self-understanding. Like Kant, I am dissatisfied with the

176 Allen W. Wood

shallower compatibilist solutions to the free will problem that have been

offered to it (and are still offered today).∞∫

If, in order to account for our capacities as rational beings and agents,

the best we could ever do were to postulate that we are supernaturally free

in a noumenal world, then that is what we would be stuck with. Kant would

be sadly right about the free will problem. But I doubt that Kant’s extrava-

gant metaphysics is the best we will ever be able to do with this problem.

The basic point, however, is that Kantian ethics is no more hostage to the

free will problem than any other ethical theory would be that regards us as

reasonable and self-governing beings. Any defensible ethical theory has to

deal with us as beings who ask ethical questions about what we should do,

can answer them according to reasons, and can act in accordance with the

answers they come up with. That’s what we have to start with in formulat-

ing any ethical theory, whatever psychology, biology, or metaphysics we

think lies behind it.

Any ethical theory that tried to deny that we are free, or to reduce our

natural capacities to empiricist mechanisms of sentiment and desire, would

get us obviously wrong in ways in which Kantian ethical theory gets us

right. Kant’s own doctrines perhaps play into the hands of such theories, by

representing our practical freedom and capacity for rational self-govern-

ment as something metaphysically controversial—something that requires

a mind-boggling leap of faith into the supersensible. Kant did this because

he actually found it appealing to associate our human dignity with the

thought that we are supernatural beings, or have a divine spark and destiny

that transcends anything our empirical science can ever hope to grasp. It

seemed fitting to him that those who are committed to morality would think

of themselves as belonging to a supernatural realm (a corpus mysticum or Kingdom of Grace), while those who think of themselves only as natural

beings are condemned to be enslaved by their lower desires, failing to live

up to their rational capacities; we should expect such earthbound material-

ists to treat themselves and other rational beings with contempt (like the

mere things they think we all are).

After two centuries, however, neither the philosophy nor the sociology

of this picture any longer carries any conviction for an enlightened person.

The picture itself is a holdover from the very religious prejudices Kant so

admirably opposed in his own time.

If we are capable of rationally recognizing that people are free and equal

self-governing beings and ends in themselves, then we do not have to bow

down to anything supernatural in order to sustain our commitment to treat-

ing them as their dignity requires.

What Is Kantian Ethics? 177

All thinking people reject the antiscientific attitudes of religious funda-

mentalists who reject the theory of evolution on the ground, as they some-

times put it, that ‘‘if we think of ourselves merely as animals, then we will

act like mere animals.’’ It is one measure of philosophical progress since the

eighteenth century that then such thoughts were shared by a great many of

the most enlightened people (including Kant), whereas now they are char-

acteristic only of people who are conspicuously unenlightened. The thing a

Kantian should now say is that of course we are animals—namely, animals

that have somehow evolved capacities to direct their lives according to

principles of reason and to recognize the objective value of their own

rational nature as an end in itself. The regrettable thing is only that we

animals with rational capacities too often fail to exercise them.

Kantian ethics is about having a rational conception of ourselves which

commits us to autonomy, human equality, and cosmopolitan community. It

is an ethical theory that is still as gripping in its inspiration and as radical in

its implications as it was when it emerged from the Enlightenment tradition

a little over two centuries ago. That is why Kant still is, and still should be,

the most influential ethical theorist in the philosophical tradition to which

our civilization belongs.

notes

1. A lack of innocent earnestness in one’s attitude toward morality was,

of course, not unknown in Kant’s time either; nor was it unknown even to

Kant himself, when he wrote that people ‘‘put on a show of affection,

respect for one another, modesty and impartiality, but without deceiving

anyone, because it is generally understood that they are not sincere’’ (VA

7:151). But Kant appears to regard this universal deception as inevitable,

and he even approves of it to the extent that he thinks people may actually

make themselves better through the attempt to live up to their false profes-

sions of moral goodness. In our time, even the pretense of moral earnestness

is not so universal, and people’s reactions to it are much more varied, and

often openly skeptical.

2. We can no longer even take for granted, as Hume did, that those who

profess doubt about all moral distinctions and obligations must be insin-

cerely disputatious, and that no one will take these doubts seriously. Hume,

Enquiry concerning the Principles of Morals, in Hume: Enquiries, ed. P. Nidditch (Oxford: Clarendon Press, 1975), sec. I, pp. 169–70.

3. If you don’t know what I mean here, then just start reading William

Bennett, ed., The Book of Virtues (New York: Simon and Schuster, 1993),

178 Allen W. Wood

and you should find out soon enough. Even the term ‘moralist’—in the

eighteenth century a perfectly acceptable word for someone who thinks

about morality or ethics—now embarrasses us; but since a term for the

same thing is still indispensable, people have had to invent the term ‘ethi-

cist’—arguably the ugliest neologism of recent times that can’t be blamed

either on the computer or on French literary theory.

4. John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard Univer- sity Press, 1971), p. 27.

5. Before making this criticism of classical utilitarianism, Rawls had

already laid down as the ‘‘main idea’’ of his theory the concept of an

‘‘original agreement’’ through which free and equal persons might enter

into a community; A Theory of Justice, p. 11. 6. For one attempt to do this see Amartya Sen, ‘‘Consequential Evalua-

tion and Practical Reason,’’ Journal of Philosophy 97 (September 2000): 477–502.

7. But we will mangle this point too if we are still stuck on the false

thought that desire has to be the starting point for deliberation. For then we

will express it by saying that the moral motive is a desire for self-approval. This is, of course, a desire we all have. But it is a commonly observed fact

of human life that it can usually be satisfied most easily not by doing the

right thing but by deceiving oneself into thinking one has done it, or by

adjusting one’s perception of moral demands so that they are satisfied by

what one wanted to do anyway. Hence a moral theory that starts out this

way, if it is consistent, can only be morally bankrupt, and that to a spectacu-

lar degree. The point is not whether we do or do not approve of ourselves,

but rather whether we act in such a way as to be worthy of our own

approval. So the issue for genuine moral self-concern is not how to satisfy

the desire for self-approval, but how to satisfy the conditions under which

we ought to approve of ourselves.

8. It is beyond the scope of this essay to say how far I think Hegel

subscribes to that view, and how large a gulf separates him from some

present-day ‘‘communitarians,’’ who sometimes take his name in vain

when advocating their popular but pernicious views.

9. Kant even argues that we must tolerate (or even welcome) the sem-

blance of genuine moral virtue most people assume, because it, and the

desire for honor that motivates it, are among the most readily available and

least dispensable means we have for winning others, and even ourselves,

over to genuine morality (VA 7:151–53). I will leave unexplored the ques-

tion how far we ought to follow him in this dubious direction, although I do

What Is Kantian Ethics? 179

want to point out that this line of thinking should move some of his critics, if

they are consistent, to qualify their negative opinion of him.

10. The term ‘‘CI-procedure’’ and the notion that Kantian ethics in-

volves a ‘‘constructive’’ (hence antirealist) metaethics have grown popular

through the influence of Rawls’s essay ‘‘Kantian Constructivism in Moral

Theory: The Dewey Lectures 1980,’’ Journal of Philosophy 77 (September 1980): 515–75. See also his interpretation of Kant in Rawls, Lectures on the History of Moral Philosophy, ed. Barbara Herman (Cambridge, Mass.: Harvard University Press, 2000), pp. 143–328. There is much in Rawls’s

presentation of Kant, especially in these lectures, that is true, illuminating,

and important. But regarding the theme of ‘‘constructivism,’’ I find Rawls’s

presentation of Kantianism in section 40 of A Theory of Justice, pp. 251– 257, much more accurate and sympathetic. Further, I think Kantian ethics is

both distorted and harmed by the attempt to associate it with metaethical

antirealism and by the way the idea of a ‘‘CI-procedure’’ perpetuates over-

emphasis on the formula of universal law. Kant does not directly address

issues in twentieth-century metaethics; any interpretation on these issues

has to depend on the way one construes hints in Kant’s texts or the overall

shape of Kantian theory based on the interpreter’s own philosophical pre-

dilections. My own predilections are in more a realist than an antirealist

direction, and I think Kant’s theory is read more sympathetically if it is read

more realistically. But I also think that even textually the hints on the realist

side are stronger than those on the antirealist side, so at least it should not

automatically be taken for granted that Kantianism is not a metaethically

realist view, as has been suggested recently by Philip Stratton-Lake, ‘‘Kant

and Contemporary Ethics,’’ Kantian Studies 2 (1998): 1–14. Far more harmful than the ‘‘constructivist’’ interpretation, however, has been its per-

petuation of the traditional association of Kantian ethics with the univer-

salizability test, and the common assumption that testing maxims for uni-

versalizability is Kant’s chief (or even his only) significant contribution to

ethical reasoning. For those who want to read Kant sympathetically, it is

worth noting that Rawls himself clearly admits the ‘‘CI-procedure’’ is inca-

pable of providing the content that (on this interpretation) it is supposed to

provide. (See Lectures on the History of Moral Philosophy, p. 163.) Read- ing Kant in Rawls’s way, therefore, seems best designed to yield an inter-

pretation of Kant’s ethical theory as an unsuccessful rough draft of the

theory Rawls was later to provide. It is not, then, on the whole, a very

sympathetic interpretation of Kant. To his credit, however, Rawls rejects

the idea of the ‘‘CI-procedure’’ as providing something like an algorithm

180 Allen W. Wood

for moral decision making: ‘‘It is a serious misconception to think of the CI-

procedure as an algorithm intended to yield, more or less mechanically, a

correct judgment. There is no such algorithm, and Kant knows this. It is

equally a misconception to think of this procedure as a set of debating rules

that can trap liars and cheats, scoundrels and cynics, into exposing their

hand. There are no such rules’’ (Lectures on the History of Moral Philoso- phy, p. 166). 11. Kant famously denies that there can be any ‘‘conflicts of duty’’ (MS

6:224). But what he is denying in this very brief discussion of a very

complex topic is only that there could be no case in which one obligation

could ‘‘cancel’’ another. It is not clear that Kant ever truly raised the issue

whether there are ‘‘moral dilemmas’’ (in the form in which that issue has

been discussed recently by moral philosophers). He does allow that dif-

ferent obligating reasons (rationes obligandi) can conflict, requiring us to choose the stronger over the weaker.

12. In only one of the sixteen ethical duties enumerated in the Doctrine

of Virtue is there an appeal to anything like universalizability. This is the

duty of beneficence, which involves universalizability in the unique case of

the one maxim that everyone necessarily adopts (namely, of willing one’s

own well-being and hence, on the basis of Kant’s empirical thesis that the

vital ends of all human beings are tightly dependent on the voluntary assis-

tance of others, necessarily willing that others promote it too) (MS 6:452).

Eleven of the fifteen remaining duties are based explicitly on appeals to the

worth of humanity, and the other four are based on it by implication, be-

cause they are classified as falling under duties with that explicit basis. For

documentation of these claims, see Kant’s Ethical Thought (New York: Cambridge University Press, 1999), pp.139–41.

13. See ibid., pp. 182–90.

14. See ibid., pp. 171–82.

15. Kant’s famous use of the formula of the universal law of nature in

discussing the four examples at G 4:421–25 depends on and presupposes a

system of duties, projected (but not supplied) in the Groundwork, and depending on the formulas of humanity as end in itself and of autonomy,

which are still not available to him in this discussion. The fact that the

universalizability tests for maxims have the intended background accounts,

I believe, for the shortcomings critics have often noted in the formula of the

law of nature (and to which Kant’s defenders have misguidedly and unsuc-

cessfully tried to reply) when this formula is interpreted (as I think, misin-

terpreted) as a universal and self-sufficient moral decision procedure.

16. Hegel, Elements of the Philosophy of Right, § 135A.

What Is Kantian Ethics? 181

17. I view Rawls’s conception of principles chosen in an initial situation

and Habermas’ and Apel’s conceptions of an ethics of domination-free

communication as recent positive examples of the sort of thing I have in

mind here.

18. Nor am I convinced that there is anything but snake oil and pseudo-

science in most of the flashy attempts to explain our moral and social

capacities on the basis of such a reduction.

Glossary

Abbruch tun: infringe ableiten: derive Absicht: intention, aim absondern: separate; abstract* Achtung: respect Affekt: affect All: (the) all allgemein: universal; general Allgemeingültigkeit: universal

validity

Allheit: totality anerkennen: recognize angenehm: agreeable Anlage: predisposition Anlockung: enticement Anmaßung: presumption Anschauung: intuition Ansehen: authority an sich: in itself Anspruch: claim Antrieb: impulse Arbeit: labor Art: way, kind, species auferlegen: impose Aufgabe: problem aufheben: abolish auflösen: resolve, solve; analyze Aufmunterung: encouragement aufsuchen: search

*When the word is so translated, a footnote will inform the reader.

ausfindig machen: bring to light ausmachen: constitute; settle

Bedeutung: significance, signification Bedingung: condition Bedürfnis: need befördern: further, promote Befugnis: warrant, authorization begabt: endowed Begehrungsvermögen: faculty of

desire

begreifen: comprehend Begriff: concept beharrlich: persisting beilegen: apply (to), attribute (to) Belieben: discretion Bemühung: toil, effort Berharrlichkeit: persistence berichtigen’: correct Beschaffenheit: constitution;

property

Beschäftigung: enterprise, concern, business

besonder: particular besorgen: take care Besorgnis: concern beständig: permanent, constant bestimmen: determine

184 Glossary

Bestimmung: determination, vocation (in ethical

contexts)

betrachten: consider Betrachtung: inquiry; consideration Beurteilung: judgment bewahren: confirm Bewegungsgrund: motive Bewegursache: motivation beweisen: prove Bewußtsein: consciousness Beziehung: reference; relation Bild: image billig: equitable, fair billigen: approve Boden: terrain Böse: evil Bösewicht: scoundrel

darlegen: establish Darstellung: exhibition, presentation,

display

dartun: establish Dauer: duration Deutlichkeit: distinctness, clarity Ding: thing

echt: genuine Ehre: honor ehrlich: honest, honorable Eigendünkel: self-conceit Eigenschaft: quality eigentlich: real, authentic Eigenliebe: self-love einbilden: imagine Einbildungskraft: imagination Einfalt: simplicity Einfluß: influence einräumen: concede Einrichtung: adaptation

Einschränkung: limitation einsehen: have (gain) insight into einstimmen: harmonize Empfänglichkeit: receptivity Empfindung: sensation, feeling* entlehnen: get Erfahrung: experience erfordern: require Erhabenheit: sublimity Erhaltung: preservation erheben: elevate erkennen: cognize, know* Erkenntnis: cognition Erklärung: explanation; definition;

declaration

erlaubt: permissible Erläuterung: elucidation, illustration Erscheinung: appearance erteilen: impart

fähig: susceptible; capable Fähigkeit: capacity festsetzen: establish Freiheit: freedom, liberty Freundschaft: friendship Frist: term

Gabe: gift Gebiet: domain Gebot: command Gebrauch: use, employment Gebrechlichkeit: fragility gefallen: like, please gefällig: pleasing Gefühl: feeling Gegengewicht: counterweight Gegenstand: object Geist: spirit, mind Geldnot: pecuniary distress gemäß: in accord

Glossary 185

gemein: common Gemeinschaft: community

(communio);

interaction

(commercio)

Gemüt: mind, heart* genugtun: satisfy Genuss: enjoyment Geschäft: business, concern,

enterprise

gescheit: clever Geschicklichkeit: skill Gesetz: law Gesetzgeber: legislator Gesetzgebung: legislation, giving law Gesetzmässigkeit: lawfulness Gesichtspunkt: point of view Gesinnung: disposition Gewalt: control Gewerbe: trade Glaube: belief, faith Glied: member Glück: (good) fortune; luck glücklich: happy, fortunate Glückseligkeit: happiness Grad: degree Grenze: bound(ary) Grund: ground gründlich: well-grounded Grundsatz: principle Gültigkeit: validity Gunst: favor

Handlung: act(ion) Hang: propensity heilig: holy herrlich: splendid hervorbringen: produce, bring forth hinreichend: sufficient Hirngespinst: figment of the mind

Hochschätzung: esteem

Idee: idea Imperativ: imperative Inbegriff: sum total

Kennen: acquaintance kennen: be acquainted with, know Kenner: connoisseur Kenntnis: acquaintance klar: clear Klugheit: prudence Kraft: power, force (physical

contexts)

Kritik: critique, criticism Kunst: art

Laster: vice Laune: mood lauten: be stated lauter: pure* Lehrbegriff: doctrine Lehre: doctrine Leidenschaft: passion liebenswürdig: amiable Lob: praise Lüge: lie Lust: pleasure

Macht: might, power Materie: matter meinen: hold or express opinion(s) Mensch: human being menschlich: human (adj.) Menschlichkeit: humanity Mittel: means Moral: morals Moralität: morality Muster: model Mut: courage

186 Glossary

Nachteil: disadvantage Naturgesetz: natural law Neigung: inclination Not: distress Nötigung: necessitation Notwendigkeit: necessity Nützlichkeit: utility

Oberhaupt: supreme head oberst: supreme Object: object

Person: person Pflicht: duty pflichtmäßig: in conformity with

duty

Preis: price Prinzip: principle Prüfung: examination

Quelle: source

Ratgebung: advice Ratschlag: counsel Raum: space Recht: right (n.) Rechtschaffenheit: uprightness Redlichkeit: honesty Regel: rule Reich: realm rein: pure Ruhe: tranquillity

Sache: thing Satz: proposition Schande: disgrace schätzen: estimate; esteem Schätzung: estimation Schein: illusion; semblance scheinen: seem

schicklich: suitable Schlauigkeit: cunning Schranke: limit(ation) schwärmen: enthuse Schwärmerei: enthusiasm Seele: soul Selbstbeherrschung: self-control Selbstbestimmung: self-

determination

Selbstliebe: self-love Selbstmord: suicide Selbstverleugnung: self-renunciation Sinn: sense, meaning Sinnenwelt: world of sense sinnlich: sensible, sensuous Sinnlichkeit: sensibility Sitten: morals, morality Sittlichkeit: morality Sollen: ‘ought’ Standpunkt: standpoint Stoff: material

Tat: deed Tätigkeit: activity Tauglichkeit: suitability Teilnahme: sympathy; sympathetic

participation

Treue: fidelity Triebfeder: incentive Tugend: virtue Tun und Lassen: deeds and

omissions

Übel: ill Übereinstimmung: agreement Überlegung: reflection Übermut: arrogance übersteigen: surpass Übertretung: transgression Überzeugung: conviction

Glossary 187

Umfang: range unbegreiflich: incomprehensible Unlauterkeit: impurity unnachlaßlich: unremitting Unschuld: innocence unterordnen: subject Untersuchung: investigation Unvermögen: incapacity Urbild: archetype Urheber: author Ursache: cause Urteilskraft: power of judgment Urwesen: original being

verabscheuen: abhor verachten: despise Verachtung: contempt Veränderung: alteration Verbindlichkeit: obligation Verbindung: combination

(conjunctio)

verdienstlich: meritorious Vereinigung: unification Vergnügen: gratification Verhalten: conduct Verhältnis: relation Verkehr: traffic; commercial traffic Verknüpfung: connection (nexus) Verlegenheit: embarrassment Vermögen: faculty vernuenfteln: ratiocinate Vernunft: reason verschaffen: obtain Versprechen: promise Verstand: understanding Verstandeswelt: world of the

understanding

Versuchung: temptation verwerflich: reprehensible Vollendung: completion

vollkommen: perfect (v.) Vollkommenheit: perfection vollständig: complete Vorgeben: pretense Vorschrift: precept Vorsorge: provision vorstellen: represent Vorstellung: representation

(repraesentatio)

Vorteil: advantage

wählen: choose Wahn: delusion Wahrnehmung: perception Weg: route Weltweisheit: philosophy Wert: worth, value Wesen: being, entity (ens); essence

(essentia)

Widerspruch: contradiction Widerstreit: conflict Wille: will Willkür: (power of ) choice willkürlich: voluntary; arbitrary wirklich: actual, real Wirkung: effect Wissen: knowledge Wissenschaft: science Witz: wit Wohl: well-being Wohlergehen: welfare Wohlgefallen: satisfaction wohltätig: beneficent Wohltun: beneficence Wohlwollen: benevolence Wollen: volition wollen: will Würde: dignity

Zergliederung: analysis

188 Glossary

zufällig: contingent Zufriedenheit: contentment Zusammenhang: connection; nexus Zusammensetzung: composition;

synthesis

zusammenstimmen: harmonize Zustand: condition, state Zwang: coercion Zweck: end Zweckmässigkeit: purposiveness

Index

Abbott, T. K., xi–xv

Absolute worth, 46, 54, 57, 66

Acting from duty, 13–17, 22–23,

92–110

Adams, R. M., 109

Affects, 10

Agreeableness, 30

Allison, H., 107

Animality, predisposition to, 47

Antagonism, 42

Anthropology, practical, 4–5, 28–

29

Apel, K.-O., 181

Apodictic, 24, 62

Appearances, 67–69

Aristotle, 10, 24, 105–106, 109

Arrogance, 9

Augustine, Saint, 85

Author of legislation, 49

Autonomy, 50–52, 54, 58, 63–65,

66–69, 84, 112–122

Avarice, 47

Beck, L. W., xiii, xv

Beneficence, 14, 102, 132–140, 180;

duty of, 40, 48–49

Bennett, W., 177–178

Bentham, J.,111

Bonum vacans, 73

Bradley, F.H., ix

Brennan, T., xv

Casuistical questions, 168

Categorical imperative. See Impera-

tive: categorical

Character, 9, 14

Cheerfulness and virtue, 97

Christian ethics, 85–88

Cicero, ix–x, 3, 26

CI-procedure, 167, 179

Circle, 66–69

Civilization, 166

Command, 24, 35

Common rational cognition, 6, 8–9,

20–21, 25

Communitarians, 178

Community, moral, 162

Conscience, 17

Consequentialism, 140–152

Constraints, 142

Constructivism, 179

Cooper, J., xv

Courage, 9, 10

Crusius, C. A., 59

Cummiskey, D., 152–153, 156

Decency, 165

Denis, L., 108, 110

Deontology, 140–151

Depravity, 22

Dialectic, 21, 72

190 Index

Dignity, 21, 28, 43, 52–53, 56, 57

Diogenes Laertius, 3, 24

Division of labor, 4

Duncan, A. R. C., 26

Duties, 167–171; division of, 38–

41, 47–49; to oneself, 38–39, 47–

48; to others, 39–40, 48–4; per-

fect and imperfect, 38, 41, 47–48,

128, 155

Duty, 18, 22, 38, 95; conflicts of,

180; transgression of, 42

Duty, acting from. See Acting from

duty

End, 45–47; discretionary, 46; ob-

jective, 46–47

End in itself, 45, 52. See also FH

Ends, self–sufficient and to be ef-

fected, 55

Enlightenment, 25, 157

Epicurus, 59, 97, 108

Esteem, 9, 14, 60

Ethical life, 165

Ethicist, 178

Ethics, 3

Examples, moral, 25

FA, xviii, 49–51, 52, 55–56, 58, 77,

119, 164, 171

Faith, rational, 78

Feeling, 17, 45, 76

FH, xviii, 41, 45–49, 54, 151, 164,

171, 173–175

Flattery, 53

FLN, xviii, 38–40, 54, 164, 167,

171–173. See also FUL

Formula of Autonomy. See FA

Formula of Humanity as End in it-

self. See FH

Formula of the Law of Nature. See

FLN

Formula of Realm of Ends. See FRE

Formula of Universal Law. See FUL

Fragility, 22

FRE, xviii, 51–52, 54–55, 151, 164,

171

Freedom, 63–69, 71–74, 79, 114–

117, 175–177; external, 48

Friendship, 24, 92, 167

FUL, xviii, 18–19, 37, 153, 164,

167, 171–173; derivation, 112–

122; interpretation, 122–140

Garve, C. ix–x, 25, 26

Gifts of nature, 9

God, 25, 30, 49, 52, 59–61, 85–88;

fear of, 26

Golden Rule, 48

Good will. See Will

Good, practical, 29–30

Gospel, 25, 48

Gratitude, duties of, 97

Green, T. H., ix

Gregor, M., xiii, xv

Grotius, H., 38

Habermas, J., 181

Hamann, J. A., ix–x

Happiness, 9, 11–12, 20, 26, 31–35,

59, 60, 85, 89, 93; indirect duty to

secure one’s own, 14–15

Hare, R. M., 152

Health, 3, 15

Hegel, G. W. F., 178, 180

Herder, J. G., x, 11

Heteronomy, 25, 50–51, 58, 69

Highest good, 25, 29

History, pragmatic, 33

Index 191

Holy one of the Gospel, 25

Honor, 9, 58–59

Humanity, predisposition to, 45–47

Hume, D., 86–87, 177

Hutcheson, F., 59–60

Imperative, 30–37; assertoric, 31–

33; categorical, 33–37, 42, 58–

61, 163–164; hypothetical, 31–

33, 37, 42, 45, 58–61, 163–164;

problematic, 31–32; universal

formula of, 55. See also FUL;

FLN; FH; FA; FRE

Imperfect duty, 38, 41, 47–48

Impurity, 22

Incentive, 22–23, 27–28, 29, 42

Inclination(s), 16, 17, 20, 22, 30, 42,

46–47, 62, 71, 73, 93; acting

from, 13–15

Individualism, 166–167

Inner judgment, 43

Innocence, 20

Instinct, 11–12, 39–40

Intellectualism, moral, 87

Intelligible world, 56, 67–73

Interest, 30, 65–69, 76, 77, 78

Intuitionism, 89–90

Jesus, 25

Judgment: moral, 19–20, 41; pre-

supposes freedom, 65

Juno and the cloud, 44

Justice, 10

Kant, I., his place in the history of

ethics, ix, 83–91. Works: Anthro-

pology in a Pragmatic Respect

(1798), 9, 11, 23, 46–47, 57, 166,

169, 178; Conjectural Beginning

of Human History (1786), 11, 46–

47, 165, 169; Critique of the

Power of Judgment (1789), 11,

31, 57, 169; Critique of Practical

Reason (1788), 3, 7, 11, 17, 24,

25, 59, 61, 92, 108, 168, 169–170;

Critique of Pure Reason

(1781/1787), 7, 24, 25, 36, 54, 59,

64, 67, 68, 70, 72, 78, 79, 92–93,

170; Groundwork of the Meta-

physics of Morals (1785), ix–xi,

3–79, 92–110, 111–156, 163–

164, 167–177; Idea for a Univer-

sal History with a Cosmopolitan

Aim (1784), x, 11, 165; Lectures

on Ethics Collins (1784–1785),

49; Lectures on Logic Jäsche

(1799), 26, 169; Metaphysics of

Morals (1797–1798), xi, 4, 6, 7,

10, 11, 17, 19, 24, 25–26, 38, 41,

47, 48, 49, 96–97, 99, 105, 108,

167, 168, 174, 180; On the Com-

mon Saying: That May Be Correct

in Theory But Does Not Work In

Practice (1793), 25; On Turning

Out Books (1798), 26; Pro-

legomena to Any Future Meta-

physics (1783), 24, 25; Religion

Within the Boundaries of Mere

Reason (1793–1794), xi, 22, 46–

47, 49, 102, 108; What Does It

Mean To Orient Oneself in Think-

ing? (1786), 25; What is Enlight-

enment? (1784), x, 169

Kantianism and consequentialism,

140–152

Kantianism and kantianism, 111

Korsgaard, C., 98, 108, 154, 155

Kraft, B., xv

192 Index

Law, moral, 5, 17–19, 36–37, 84,

118; acting for the sake of, 6; holi-

ness of, 53

Laws, empirical, 45

Legislator, 49

Leibniz, G. W., 51, 86–87

Logic, 3, 27

Love, 15, 17, 40

Luther, M., 85–87, 91

Luxury, 11

Lying, 5, 18, 36, 39, 48, 58–59,129–

132

Mandeville, B., 61

Mathematics, 24

Maxim(s), 15–16, 18–19, 37, 52,

54–55; complete determination

of, 54-55; form of, 54; matter of,

54; universalizability tests of, 37–

42. See also FUL, FLN

Maxims of thought and communica-

tion, 169

Means, 45

Meiners, C., 25

Mendelssohn, M., 25, 26

Merchant, honest, 13

Meritorious duty, 42, 48; see also

Duties: perfect and imperfect

Metaphysics, 3–5, 6, 8, 29, 44

Metaphysics of morals, 4–8, 26–27,

29, 43–44, 63

Mill, J. S., ix, 111

Misology, 11

Moderation, 10

Montaigne, M., 61

Moral deliberation, 160–163

Moral dilemmas, 180

Moral feeling, 17, 26, 60

Moral philosophy, popular. See Pop-

ular Moral Philosophy

Moral worth, 13–17, 28

Moralism, 158–160, 177–178

Morality, 163–164; and custom,

164–176

Morals (Moral), 4, 47

Motive, 23, 46

Natural necessity, 72

Nature, law of, 38. See also FLN

Nature, tutelary, 43

Necessitation, practical, 16, 29, 33,

43, 52, 57

Nicolai, C. F., 25

Nietzsche, F., 160

Obligation, 7, 57

Options, 142

Parfit, D., xv

Passions, 10

Paton, H. J., x, 91

Paul, Saint, 84, 86

Pecuniary distress, 39

Perfection, 26, 59–60; duty of, 39–

49, 48

Permissible, 57

Person, 46, 71; worth of, 10

Personality, 66; predisposition to,

47

Philanthropy, 13–14, 15, 17–18, 23.

See also Love

Philosophical moral cognition, 8–9

Philosophy, divisions of, 3

Philosophy, moral, 6, 20–21, 25. See

also Morals

Philosophy of nature, 45

Philosophy, popular. See Popular

moral philosophy

Physics, 3

Plato, 10, 11, 19

Index 193

Popular moral philosophy, 8, 22–29

Postulate, 46

Power, 9

Pragmatic, 33

Praise, moral, 14

Precept, 35

Predispositions of human nature, 47

Preserving one’s life, 13–14

Price, 43, 52–53

Principle of morality, 28; a priori, 5–

6; formulas of, 8; supreme, 8,

169; synthetic, 62. See also Law,

moral; FUL; FLN; FH; FA; FRE

Principle(s), 15–16, 17, 26; a priori,

22–24; formal and material, 45.

See also Law; Maxim

Promising, 18–19, 36, 29, 48, 126,

129–132

Property, 48

Propriety, 165

Prudence, 18, 31–33, 60; counsels

of, 33–35

Psychology, 6

Pufendorf, S., 38, 86

Purposiveness, natural, 10–11

Quantity, categories of, 54

Rational nature, 24, 26, 43, 46. See

also Humanity, predisposition to

Rawls, J., 90–91, 162, 178, 179–

180, 181

Realm (Reich), 51–52

Realm of ends, 51–52; 53, 56–57,

78, 167, 176; member of, 51, 53,

56; supreme head of, 52

Realm of nature, 54

Reason: instrumental, 31–32; practi-

cal, 28, 29–37, 77, 113–117; pru-

dential, 32–35; pure practical, 7–

8, 45; theoretical and practical,

113–116

Refusal, sexual, 165

Reich, K., x, 26

Respect, 16–17, 57, 165

Rights, human, 48

Schiller, F., 92,107

Schönecker, D., xv, 26, 107, 110

Science, pure and empirical parts of,

3–5

Scripture, holy, 15, 23, 25, 48. See

also Gospel

Self-contempt, 43

Self-control, 10

Self-legislation, 49–50

Self-love, 22–23,38, 39

Sen, A., 178

Seneca, 3, 52

Servility, 47

Sexual desire, 165

Shaftesbury, third earl of, 100–102,

108–109

Sidgwick, H., ix, 111

Skill, imperatives of, 32

Socrates, 19, 85

South Sea Islanders, 39

Stoics, 3, 59, 89, 108

Stratton-Lake, P., 179

Strict duty, 42, 43, 57

Suicide, 13–14, 38–39, 47

Sulzer, J. G., 27–28, 38–39

Superego, 165

Sympathy, 14, 40, 102–103

Talents: duty to develop, 39–40, 48;

of the mind, 9

Taste, 44, 62

Teleology, 54. See also Purposive-

ness, natural

194 Index

Temperament, 9

Thing, 46

Things in themselves, 67–69. See

also Intelligible world

Thomas Aquinas, Saint, 84

Thomasius, C., 38

Transcendental idealism, 67–73, 77

Tylawsky, E., xv

Universal practical philosophy, 6

Unsociable sociability, 165

Utilitarianism, 88–90, 111, 162–

163, 168

Vice, 28

Virtue, 10, 23, 27–28, 44, 92–95,

105

Voluntarism, moral, 85–86

Wealth, 9

Welfare, 9, 11. See also Happiness

Will, 15–16, 29–37, 62–66, 69–70,

74; free, 25; good, 9–13, 19, 55,

62; holy, 25, 29–30, 37; pure, 6.

See also Freedom

Whiting, J., 109

Wimmer, R., 26

Wisdom, 10, 19–20

Wolf, S., 109

Wolff, C., 6, 59, 86, 88

Wood, A., 107, 108, 109, 110

World of sense, 67–77

World of understanding, 67–77

Worthiness to be happy, 9

Zeno of Citium, 3

Zorzo, S., 110

Rethinking the Western Tradition A l s o ava i l a b l e i n t h e s e r i e s :

T h e P r i n c e

b y N i c c o l ò M a c h i a v e l l i

Tr a n s l a t e d b y A n g e l o C o d e v i l l a

T h e I d e a o f a U n i v e r s i t y

b y J o h n H e n r y N ewm a n

E d i t e d b y F r a n k M . Tu r n e r

C u l t u r e a n d A n a r c h y

b y M a t t h ew A r n o l d

E d i t e d b y S a m u e l L i p m a n

D i s c o u r s e o n t h e M e t h o d

a n d M e d i t a t i o n s o n F i r s t P h i l o s o p h y

b y R e n é D e s c a r t e s

E d i t e d b y D a v i d We i s s m a n

T h e S o c i a l C o n t r a c t a n d

T h e F i r s t a n d S e c o n d D i s c o u r s e s

b y J e a n - J a c q u e s R o u s s e a u

E d i t e d a n d w i t h a n I n t r o d u c t i o n b y S u s a n D u n n