. 1
( 8)



>>

Kant™s Search for the Supreme Principle of Morality


At the core of Kant™s ethics lies the claim that if there is a supreme prin-
ciple of morality, then it is not a utilitarian or Aristotelian perfectionist
principle, or even a principle resembling the Ten Commandments.
The only viable candidate for the supreme principle of morality is
the Categorical Imperative.
This book is the most detailed investigation of this claim. It con-
structs a new, criterial reading of Kant™s derivation of one version of
the Categorical Imperative: the Formula of Universal Law. This read-
ing shows this derivation to be far more compelling than contempo-
rary philosophers tend to believe. It also reveals a novel approach to
deriving another version of the Categorical Imperative, the Formula
of Humanity, a principle widely considered to be the most attractive
Kantian candidate for the supreme principle of morality.
Lucidly written and dealing with a foundational topic in the history
of ethics, this book will be important not just for Kant scholars but
for a broad swath of students of philosophy.

Samuel J. Kerstein is Associate Professor of Philosophy at the Univer-
sity of Maryland, College Park.
Kant™s Search for the Supreme
Principle of Morality


SAMUEL J. KERSTEIN
University of Maryland, College Park
published by the press syndicate of the university of cambridge
The Pitt Building, Trumpington Street, Cambridge, United Kingdom

cambridge university press
The Edinburgh Building, Cambridge cb2 2ru, uk
40 West 20th Street, New York, ny 10011-4211, usa
477 Williamstown Road, Port Melbourne, vic 3207, Australia
Ruiz de Alarc´ n 13, 28014 Madrid, Spain
o
Dock House, The Waterfront, Cape Town 8001, South Africa
http://www.cambridge.org

Samuel J. Kerstein 2002
C


This book is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.

First published 2002

Printed in the United Kingdom at the University Press, Cambridge

Typeface itc New Baskerville 10/12 pt. System LTEX 2µ [tb]
A


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

Library of Congress Cataloging in Publication data
Kerstein, Samuel J., 1965“
Kant™s search for the supreme principle of morality / Samuel J. Kerstein.
p. cm.
Includes bibliographical references and index.
isbn 0-521-81089-2
1. Kant, Immanuel, 1724“1804 “ Ethics. I. Title.
b2799.e8 k45 2002
170“dc21 2001043918

isbn 0 521 81089 2 hardback
Contents




page xi
Acknowledgments
xiii
Key to Abbreviations and Translations

Introduction: Derivation, Deduction, and the Supreme
1
Principle of Morality
i.1 No Modest Claim 1
i.2 The Basic Concept of the Supreme Principle of Morality 1
i.3 Derivation and Deduction of the Categorical Imperative 4
i.4 The (Alleged) Gap in the Derivation of the Formula
7
of Universal Law
i.5 Terminological and Thematic Clari¬cations 10
i.6 Outline of the Book 11
1 16
Fundamental Concepts in Kant™s Theory of Agency
1.1 Aims and Limits of the Discussion 16
1.2 Maxims: A Basic Account 16
1.3 Maxims and Other Rules of the Same Form 19
1.4 The Will 20
1.5 Determining Grounds of the Will 21
1.6 Acting from Inclination: Three Interpretations
22
and Their Importance
1.7 Acting from Inclination in the Groundwork and in the
24
Metaphysics of Morals
1.8 Material Practical Principles: Acting from Inclination
29
in the Critique of Practical Reason
2 Transcendental Freedom and the Derivation of the Formula
33
of Universal Law
2.1 Derivation in the Critique of Practical Reason: Allison™s
33
Reconstruction


vii
Contents
viii

2.2 34
A Thick Account of Kantian Rational Agency
2.3 36
Desire and Justi¬cation of Action
2.4 39
Practical Law and Justi¬cation of Action
2.5 42
Practical Law and the Formula of Universal Law
3 46
The Derivation of the Formula of Humanity
3.1 46
Outline of the Derivation
3.2 The Supreme Principle of Morality and Unconditional
47
Value
3.3 The Unconditional Value of Humanity:
54
Kant™s Argument
3.4 55
Korsgaard™s Reconstruction: Preliminaries
3.5 56
The Supreme Principle of Morality and Good Ends
3.6 From Good Ends to the Unconditional Value
59
of Humanity: The Regressive Argument
3.7 65
The Failure of the Regressive Argument
3.8 Shortcomings in the Derivation of the Formula
71
of Humanity
4 The Derivation of the Formula of Universal Law:
73
A Criterial Reading
4.1 73
Main Steps of the Derivation on the Criterial Reading
4.2 74
Korsgaard™s Reading of the Derivation
4.3 77
The Structure of Groundwork I
4.4 The Failure of One Version of the Traditional Reading
77
of the Derivation
4.5 The Challenge Posed by Aune™s Version of the
78
Traditional Reading
4.6 From Duty and Moral Worth to Two Criteria for the
80
Supreme Principle of Morality
4.7 Law as Motive: A Third Criterion for the Supreme
82
Principle of Morality
4.8 86
The Criterial Reading and Groundwork II
4.9 Coherence with Ordinary Moral Reason:
87
A Fourth Criterion
4.10 The Apriority of the Supreme Principle of Morality 89
4.11 Rejecting the Traditional Interpretation of the
91
Groundwork II Derivation
4.12 Summary 93
5 95
Criteria for the Supreme Principle of Morality
5.1 95
Plan of Discussion: Focus on First Criterion
5.2 96
Moral Worth and Actions Contrary to Duty
5.3 98
Two Conditions on Acting from Duty
5.4 104
All Actions from Duty Have Moral Worth
Contents ix

5.5 106
Only Actions from Duty Have Moral Worth
5.6 109
The Second Criterion and Its Grounds
5.7 110
The Third Criterion and Its Grounds
5.8 112
Relations between the Criteria
6 114
Duty and Moral Worth
6.1 114
Aims of the Discussion
6.2 116
Moral Worth and Helping a Friend from Duty
6.3 118
One Thought Too Many?
6.4 119
The Moral Worth of Actions Contrary to Duty
6.5 A Disturbing Asymmetry in Kant™s View
119
of Moral Worth
6.6 121
Failure of Will or Unfortunate Event?
6.7 Moral Permissibility and Moral Worth in the
124
Metaphysics of Morals
6.8 127
The (Alleged) Transparency of Moral Requirements
6.9 129
Odious Actions and Moral Worth
6.10 Sympathy and Moral Worth 132
6.11 Summary 138
7 139
Eliminating Rivals to the Categorical Imperative
7.1 139
Aims of the Discussion
7.2 140
A Sweeping Argument against All Rivals
7.3 145
The Structure of Act Utilitarianism
7.4 146
Against Act Utilitarianism
7.5 148
Against Expectabilist Utilitarianism
7.6 152
Against Perfectionism
7.7 153
Kantian Consequentialism?
7.8 155
Against a Principle Akin to the Ten Commandments
7.9 158
Further Nonconsequentialist Rivals
7.10 Summary 159
8 Conclusion: Kant™s Candidates for the Supreme Principle
160
of Morality
8.1 Kant™s Candidates and Criteria for the Supreme
160
Principle of Morality
8.2 Two Formulas and the Basic Concept of the Supreme
162
Principle of Morality
8.3 165
Two Formulas and Further Criteria
8.4 167
Two Formulas and Ordinary Moral Consciousness
8.5 Formula of Universal Law: Practical Contradiction
168
Interpretation
8.6 Formula of Universal Law: Universal Availability
171
Interpretation
8.7 174
Fundamentals of the Formula of Humanity
Acknowledgments




This book would not have been completed without help and support from
a variety of sources.
I would like to thank Terence Moore and Brian R. MacDonald of
Cambridge University Press for their patience and expertise in guiding me
through the publication process.
Material from four of my papers has been reworked into the book.
Chapter 1 incorporates “Kant™s (Not So Radical) Hedonism,” in Kant und
die Berliner Aufkl¨ rung. Akten des IX. Internationalen Kant-Kongresses, vol. 3,
a
ed. V. Gerhardt, R.-P. Horstmann, and R. Schumacher (Berlin: Walter de
Gruyter, 2001), pp. 245“253. Part of Chapter 3 stems from “Korsgaard™s
Kantian Arguments for the Value of Humanity,” Canadian Journal of
Philosophy 31 (March 2001): 23“52. Sections of Chapters 4 and 7 have been
adapted from a paper I coauthored with Berys Gaut: “The Derivation with-
out the Gap: Rethinking Groundwork I,” Kantian Review 3 (1999): 18“40.
Finally, parts of Chapters 5 and 6 were published in “The Kantian Moral
Worth of Actions Contrary to Duty,” Zeitschrift f¨ r Philosophische Forschung 53
u
(1999): 530“551. I acknowledge with appreciation the permission of the
publishers to use material from these papers.
Most of the book was written during the academic year 1999“2000, which
I spent as a Fellow at the National Humanities Center in Triangle Park, North
Carolina. I would like to thank the National Endowment for the Humanities
for supporting my stay there. The administrators and staff at the National
Humanities Center could not have been more encouraging and helpful. In
particular I would like to thank Karen Carroll, who edited an early version
of my manuscript. (I would also like to thank Jane Strong for editing a later
version.) Preliminary work on the manuscript was made possible by support
from the University of Maryland, College Park, in the form of a General
Research Board grant that relieved me from my teaching duties during the
fall of 1996. I would like to thank the University of Maryland for this support,
as well as for granting me leave to work at the National Humanities Center.
xi
Acknowledgments
xii

For their comments and criticisms of portions of this book, I would like
to thank audiences at the British Kant Society Annual Meeting, the Central
Division Meeting of the American Philosophical Association, the Midwest
Study Group of the North American Kant Society, Duke University, the
University of St. Andrews, and the University of North Carolina, Chapel
Hill.
From early on I have been fortunate to have had outstanding teachers. I
would like to thank No¨ l Carroll and Victor Gourevitch for their guidance,
e
both philosophical and personal. I am grateful to Bonnie Kent who took
the time to teach me not only how to work in the history of philosophy but
to appreciate the importance of doing so.
I have learned a great deal about Kantian ethics from discussion and/or
correspondence with many philosophers, including Paul Cohen, Mich` le e
Crampe-Casnabet, Garrett Cullity, David Cummiskey, Raymond Geuss,
St´ phane Haber, Thomas Hill Jr., Dieter Sch¨ necker, Ralf St¨ cker, and
e o o
Allen Wood. I owe a special debt of gratitude to Berys Gaut. Some cen-
tral ideas in the book stem from our collaborative work, and Berys has been
generous in encouraging me to develop them at greater length. Readers for
Cambridge University Press, as well as two others, offered comments that
have, I think, enabled me to strengthen several of my arguments. During my
stay at the National Humanities Center, I pro¬ted from (often ambulatory)
dialogue with many colleagues, including Ruth Grant, Michelle Mass´ , e
Louise McReynolds, Bernard Reginster, Daniel Sherman, Eleonore Stump,
Timothy Taylor, and Marjorie Woods. I was especially fortunate to have been
able to discuss philosophy with Thomas Christiano, who not only provided
intellectual inspiration, but patiently helped me to work out some key points
in the book. My friends and colleagues at the University of Maryland, espe-
cially Judith Lichtenberg and Corey Washington, have aided me at several
points, both intellectually and personally, in carrying out this project.
I am deeply grateful for the help and support I have received from
R¨ diger Bittner, Thomas Pogge, and Michael Slote. From the beginning,
u
these philosophers have played essential roles in the book™s development.
Each gave me valuable advice on my project as it unfolded, and offered tren-
chant and productive comments on the manuscript as a whole. My approach
to Kantian ethics owes a great deal to each of them.
Finally, I would like to thank my in-laws John and Jane Strong, my
parents Howard and JoAnn Kerstein, and especially my wife Lisa Strong,
for their constant encouragement during the writing of this book.
Key to Abbreviations and Translations




Except for references to the Critique of Pure Reason, all references to Kant
are to the Preussische Akademie der Wissenschaften edition of his works
(Berlin: Walter de Gruyter [and predecessors], 1902). References to the
Critique of Pure Reason are to the standard A and B pagination of the ¬rst and
second editions. I list here the German title, academy edition (Ak.) volume
number, and abbreviation for each of the works I cite. Under each entry,
I specify the English edition I have consulted. The translations I employ
sometimes vary from those of these English editions.
Anthropologie in pragmatischer Hinsicht (Ak. 7)
Anth
Anthropology from a Pragmatic Point of View, tr. Victor L. Dowdell.
Carbondale: Southern Illinois University Press, 1978.
Grundlegung zur Metaphysik der Sitten (Ak. 4)
GMS
Groundwork of the Metaphysics of Morals, tr. Mary J. Gregor.
In Immanuel Kant: Practical Philosophy, 42“108. Cambridge:
Cambridge University Press, 1996.
Kritik der praktischen Vernunft (Ak. 5)
KpV
Critique of Practical Reason, tr. Mary J. Gregor. In Immanuel Kant:
Practical Philosophy, 138“271. Cambridge: Cambridge University
Press, 1996.
Kritik der reinen Vernunft (1st ed. (A) 1781; 2nd ed. (B) 1787;
KrV
Ak. 3“4)
Critique of Pure Reason, tr. N. Kemp Smith. New York: St. Martin™s
Press, 1965.
Kritik der Urteilskraft (Ak. 5)
KU
Critique of Judgment, tr. Werner S. Pluhar. Hackett: Indianapolis,
1987.
Erste Einleitung in der Kritik der Urteilskraft (Ak. 20)
KUE
In Critique of Judgment, tr. Werner S. Pluhar. Hackett: Indianapolis,
1987.

xiii
Key to Abbreviations and Translations
xiv

LE Vorlesungen uber Moralphilosophie, “Moralphilosophie Collins”
¨
(Ak. 27)
Lectures on Ethics, “Moral Philosophy: Collins™s Lecture Notes,”
tr. Peter Heath, 37“222. Cambridge: Cambridge University Press,
1997.
Die Metaphysik der Sitten (Ak. 6)
MS
The Metaphysics of Morals, tr. Mary J. Gregor. In Immanuel Kant:
Practical Philosophy, 363“603. Cambridge: Cambridge University
Press, 1996.
Die Religion innerhalb der Grenzen der blossen Vernunft (Ak. 6)
Rel
Religion within the Limits of Reason Alone, tr. T. M. Greene and
H. H. Hudson. New York: Harper & Row, 1960.
All of the English editions incorporate academy edition page numbering in
their margins, except for the KrV and Rel. When I cite the Rel, I give the
academy edition page number followed by that of the English edition.
Introduction: Derivation, Deduction, and the Supreme
Principle of Morality




i.1 No Modest Claim
If there is a supreme principle of morality, then it is the Categorical Imper-
ative. This claim, which lies at the core of Kant™s ethics, is nothing if not
ambitious. Establishing it would amount to proving that absolutely no prin-
ciple other than the Categorical Imperative “ no utilitarian principle, no
perfectionist principle, no principle along the lines of the Ten Command-
ments “ is a viable candidate for the supreme principle of morality. How
does Kant (or might he) try to prove this? Does he (or might he) succeed?
Questions of this sort are what this book is about. To answer them, we must
understand what Kant means by claiming that if there is a supreme principle
of morality, then it is the Categorical Imperative.


i.2 The Basic Concept of the Supreme Principle of Morality
To begin we need to know how Kant conceives of the supreme principle of
morality. According to (what I call) his basic concept, this principle would
possess four characteristics. It would be practical, absolutely necessary, bind-
ing on all rational agents, and would serve as the supreme norm for the
moral evaluation of action. I call this concept of the supreme principle of
morality basic because it emerges immediately in Kant™s critical writings in
ethics.1 Already in the Preface to the Groundwork of the Metaphysics of Morals
it is manifest that, in Kant™s view, the supreme principle must have these
features.
It belongs to Kant™s basic concept of the supreme principle of morality
that it constitute the supreme norm for the moral assessment of action. This
means several things. The principle would distinguish between morally per-
missible actions, that is, ones that conform with the principle, and morally
impermissible actions, that is, ones that con¬‚ict with the principle (see
GMS 390). It would also specify which actions are morally required. As
1
Search for the Supreme Principle of Morality
2

Kant suggests in the Groundwork Preface, the supreme principle of morality
would not only be the basis for appraising an action™s moral requiredness,
permissibility, or impermissibility, but also its moral goodness (GMS 390).
Whether an action is morally good depends on how it relates to this princi-
ple. In particular, to be morally good an action must both conform with and
be done “for the sake of ” the principle. Finally, as the supreme norm for
the moral assessment of action, the supreme principle of morality would be
such that all genuine duties would ultimately be derived from it (see GMS
421).2 The supreme principle would justify these duties™ status as such.
Kant says that the supreme principle of morality “must hold not only for
human beings but for all rational beings as such” (GMS 408; see also GMS
389, 425, 442; KpV 32, 36).3 The supreme principle of morality would have
an extremely wide scope: one that extended not only to all rational human
beings but to any other rational beings who might exist “ for example, God,
angels, and intelligent extraterrestrials. In Kant™s view, the supreme principle
of morality would have to possess what I call “wide universal validity.” It would
have to be binding on all rational agents, at all times and in all places. This
is the second feature that, according to Kant™s basic concept, the supreme
principle of morality would have to possess.
To say that the supreme principle of morality is binding on us (human
agents) is to imply that we have an obligation to act in accordance with it.
We ought to but, as a result of privileging inclinations over duty, might not
follow its dictates. The same could also be said for any nonhuman rational
agents who had characteristics, for example, natural cravings, on the basis
of which they might act contrary to the supreme principle. The supreme
principle™s being binding on these agents would imply that they had an
obligation to act in accordance with it. For all agents “affected by needs
and sensible motives,” the supreme principle of morality would count as
an “imperative” (KpV 32). It would set out a command that we genuinely
ought to obey, although we might not obey it (GMS 414). We can conceive
of beings, however, on whom the supreme principle would be binding but
regarding whom it would be incorrect to say that they had an obligation to
obey it. According to Kant, one can be obligated to do something only if
there is a possibility that he will fail to do it.4 Yet some beings, for example,
God, might be such that they cannot fail to obey the supreme principle of
morality. It would thus make no sense to say that they had an obligation to
obey it. For them, the supreme principle of morality would be a law but not
an imperative (GMS 414, 439; KpV 32).
A third feature the supreme principle of morality would have to possess
is that of being absolutely necessary (GMS 389). Kant™s description of this
feature answers the question of what it would mean for the supreme princi-
ple of morality to be binding on an agent. On every agent within its scope,
for Kant every rational agent, the principle would hold without exception
(GMS 408). For example, a human agent would always be obligated to act
Introduction 3

in accordance with the supreme principle, no matter what he wants to do.
For us, the supreme principle of morality would be an unconditional com-
mand. That we were obligated to perform the action it speci¬ed would not
be conditional on our having any particular set of desires.
Finally, it is worth making explicit that for Kant the supreme principle of
morality must be practical “ it must be a rule on account of which agents can
act. Kant implies this in the Groundwork Preface by specifying that morally
good actions involve an agent™s acting for the sake of the moral law, that
is, the supreme principle of morality (GMS 390). In the Critique of Practical
Reason, he de¬nes practical principles, of which the supreme principle of
morality would be one, as propositions that “contain a general determina-
tion of the will,” thereby suggesting that this principle would be something
on the basis of which an agent can set himself to do something (KpV 19“20).5
One might conceive of the supreme principle of morality as a purely theoret-
ical tool. For example, one might take it to be a rule that could be employed
(perhaps by a team of experts) to categorize something an agent has done in
terms of its rightness or wrongness, but which (perhaps due to its enormous
complexity) could not be used by the agent himself in deciding what to do.
This would be a very un-Kantian conception of the supreme principle of
morality. For Kant the supreme principle must be able to ¬gure directly in
an agent™s practical deliberations.
From the very outset of his ¬rst great work in ethics, Kant operates with a
certain basic concept of the supreme principle of morality. It is evident from
the Preface of the Groundwork that he thinks of this principle as practical,
absolutely necessary, binding on all rational agents, and the supreme norm
for the moral evaluation of action.
Three remarks are in order regarding Kant™s basic concept of the supreme
principle of morality. First, as we will see, there is more to Kant™s concept
of the supreme principle of morality than is captured in this basic concept.
There are more features that, in Kant™s view, the supreme principle would
have to possess. It would, for example, have to be such that a proponent
of its being the supreme principle of morality could coherently claim that
obeying it “from duty” would have moral worth. The second point concerns
the provenance of the four features that belong to (what I call) Kant™s basic
concept. Kant, I think, would claim that if we “ that is, beings who possess
“common rational moral cognition” “ re¬‚ect a bit on what the supreme
principle of morality would be like, we ¬nd that it would have to possess these
four features.6 Kant makes it clear that, according to him, commonsense
morality is committed to the view that absolute necessity and wide universal
validity must be features of the supreme principle of morality. Implicit in
“the common idea of duty and of moral laws,” says Kant, is that “a law, if it
is to hold morally, that is, as a ground of an obligation, must carry with it
absolute necessity; that, for example, the command ˜thou shalt not lie™ does
not hold only for human beings, as if other rational beings did not have to
Search for the Supreme Principle of Morality
4

heed it, and so with all other moral laws properly so called” (GMS 389).7
The third remark regarding Kant™s basic concept of the supreme principle
of morality concerns its role in this book. We will be probing arguments for
the claim that if there is a supreme principle of morality, corresponding to
Kant™s basic concept of such a principle, then it is the Categorical Imperative.
For purposes of this book, Kant™s basic concept of the supreme principle of
morality is assumed. As readers will quickly see, assuming this concept does
not at all render it trivial or easy to establish that the Categorical Imperative
is the only viable candidate for the supreme principle of morality.


i.3 Derivation and Deduction of the Categorical Imperative
To re¬ne further our understanding of what Kant means by claiming that
if there is a supreme principle of morality, then it is the Categorical Imper-
ative, we need to place the claim into the context of the work in which it
initially arises, the Groundwork of the Metaphysics of Morals. Kant divides the
Groundwork into a Preface and three sections. In the Preface, he says: “[T]he
present Groundwork is . . . nothing more than the search for and establish-
ment of the supreme principle of morality” (GMS 392). In Groundwork I and II,
Kant searches for the supreme principle of morality in the sense that he
tries to discover what this principle would be, assuming there is such a prin-
ciple. Kant presents the Categorical Imperative by name for the ¬rst time in
Groundwork II: “[A]ct only on that maxim through which you can at the same
time will that it become a universal law” (GMS 421, Kant™s emphasis omit-
ted). Right after he presents this principle, he says: “Now, if all imperatives
of duty can be derived from this single imperative as from their principle,
then, even though we leave it undecided whether what is called duty is not as such
an empty concept, we shall at least be able to show what we think by it and
what the concept wants to say” (GMS 421, emphasis added). Throughout
Groundwork II, Kant reminds us that he is there offering no proof that the
Categorical Imperative is absolutely necessary and universally binding, and
thus no proof that genuine moral duties derive from it (see GMS 425, 431).
At the end of Groundwork II, Kant tells us what, in his view, he has demon-
strated to that point: “[W]hoever holds morality to be something and not a
chimerical idea without any truth must also admit the principle of morality
brought forward” (GMS 445). The “principle of morality brought forward”
is, of course, the Categorical Imperative. So by the end of Groundwork II,
Kant takes himself to have completed his search for the supreme principle
of morality by showing that if there is a supreme principle of morality, then
it is the Categorical Imperative.
Let us call an argument aimed at proving that if there is a supreme prin-
ciple of morality, then it is some particular principle, a “derivation” of this
principle.8 As we will see, Kant carries out a derivation of the Categorical
Imperative not only in the Groundwork but in the Critique of Practical Reason
Introduction 5

as well. He offers several arguments for the conclusion: if there is a supreme
principle of morality, then it is the Categorical Imperative.
A successful derivation would prove this conditional conclusion. It would
complete Kant™s search for the supreme principle of morality (or, more pre-
cisely, his search for what would be this principle, if anything is). But, as we
have seen, in the Preface Kant says that the Groundwork does more: it estab-
lishes the supreme principle of morality (GMS 392). In Groundwork III, Kant
tries to close a possibility left open by Groundwork I“II: the possibility that
duty is an empty concept, that is, that we actually have no (moral) duties.
He aspires to prove that the Categorical Imperative is valid: absolutely nec-
essary and binding on all rational agents (GMS 461).9 Kant suggests in the
Groundwork as well as later in the Critique of Practical Reason that proving this
would amount to giving a “deduction” of the supreme principle of morality
(see GMS 454, 463; KpV 47, 48). Kant™s usage of the term “deduction” in the
Critique of Pure Reason signals that to carry out a deduction of the Categorical
Imperative would be to show that we have a right, that is, suf¬cient justi¬ca-
tion, for considering it to be valid (KrVA 84“85/B 116“117). By the end of
Groundwork II, Kant takes himself to have shown that those of us who believe
there to be a supreme principle must embrace the Categorical Imperative
as this principle. Yet that we who believe that there is such a principle must
embrace the Categorical Imperative does not entail that it is actually binding
on us “ that we actually have the duties this imperative speci¬es. Our belief
in morality might be mistaken. A successful derivation of the Categorical
Imperative would not eliminate the possibility that morality is a “chimerical
idea.”
The aim of producing an effective derivation of the Categorical Imper-
ative seems less aspiring than that of giving a deduction of it. A derivation
that worked would show us what the supreme principle of morality would
be, if there was one, but, unlike a deduction, it would not show us that any
given principle was actually binding on us. By giving a deduction of the Cat-
egorical Imperative, Kant would answer two different opponents. First, he
would answer a moral skeptic, someone who holds that we are not obligated
to do anything at all. For he would establish that we are obligated to act only
on maxims that we can, at the same time, will to be universal laws. Second, if
Kant provided a deduction of the Categorical Imperative, he would answer
a “moral particularist,” namely someone who believes in the reality of moral
distinctions “ for example, that there are right actions and wrong ones “ but
who denies that there are any moral principles binding on all rational agents
or even all human agents.10 For Kant would demonstrate that the Categori-
cal Imperative is just such a principle. By giving a successful derivation of the
Categorical Imperative, Kant would refute neither the moral skeptic nor the
moral particularist. Both opponents would remain free to agree with Kant
that if there were a supreme principle of morality, then it would have to be
the Categorical Imperative, yet to deny that there is any such principle.11
Search for the Supreme Principle of Morality
6

It would be remiss not to mention that by the end of Groundwork II Kant
takes himself to accomplish more than a derivation of the Categorical Im-
perative. In addition to demonstrating that if there is a supreme principle of
morality, then it is the Categorical Imperative, he also thinks he proves a
stronger claim: if morality tout court is not an illusion, then it has a supreme
principle, namely the Categorical Imperative: “[W]hoever holds morality to
be something and not a chimerical idea without any truth must also admit
the principle of morality brought forward” (GMS 445, emphasis added). So,
in effect, Kant implies that by the end of Section II, we have a response to
moral particularism. Moral particularism entails moral skepticism, suggests
Kant; morality not based on principle would be no morality at all.
I do not discuss this suggestion. Nor do I focus on Kant™s deduction of the
Categorical Imperative. Instead, I concentrate on Kant™s derivation. The aim
of generating a successful derivation of the supreme principle of morality is,
I think, suf¬ciently ambitious to warrant our full attention. If Kant attains it,
then he shows that as far as candidates for the supreme principle of morality
are concerned, the Categorical Imperative is (and will be) the only game in
town.
Even though our focus is on Kant™s derivation, and not his deduction,
of the Categorical Imperative, it is worth noting that Kant eventually seems
to abandon the project of providing a deduction. In the Critique of Practical
Reason, published three years after the Groundwork, he asserts:

[T]he moral law is given, as it were, as a fact of pure reason of which we are a priori
conscious and which is apodictically certain, though it be granted that no example of
exact observance of it can be found in experience. Hence the objective reality of the
moral law cannot be proved by any deduction, by any efforts of theoretical reason,
speculative or empirically supported, so that, even if one were willing to renounce
its apodictic certainty, it could not be con¬rmed by experience and thus proved a
posteriori; and it is nevertheless ¬rmly established of itself. (KpV 47; see also KpV
48 and 93)

This passage raises many complex issues, but for our purposes a brief treat-
ment suf¬ces. In Groundwork III, Kant implies that he is undertaking a de-
duction of the Categorical Imperative (GMS 461, 463). Yet in this second
Critique passage, Kant suggests that the “objective reality” (i.e., validity) of
the moral law is “¬rmly established of itself ”; it does not need to be proved
through philosophical argument. In stating that the moral law is given as a
fact of pure reason of which we are a priori conscious and which is apodic-
tically certain, Kant is apparently suggesting that the moral law necessarily
presents itself to each rational agent as a valid practical requirement. To
use R¨ diger Bittner™s description, Kant seems to be implying that “one is
u
cognizant of [the moral law] in such a way that in all practical considera-
tions one knows of its validity and has to take this validity into account.”12
Since we are cognizant of the moral law in this way, Kant appears to hold,
Introduction 7

there is no need for arguments to show us that we are genuinely bound by
it. The project of deduction he undertakes in Groundwork III is, Kant now
thinks, an unnecessary one. That it is unnecessary to prove the validity of
the Categorical Imperative does not entail that it is impossible to do so. Yet
Kant even goes so far as to make the further claim that this project cannot
succeed: “[T]he objective reality of the moral law cannot be proved by any
deduction.”13 Kant™s grounds for this further claim need not concern us.
However, that he makes it strengthens the impression that he eschews the
Groundwork III attempt to prove the validity of the Categorical Imperative.
If, as it appears, Kant abandons this attempt, it does not, of course, follow
that we ought to do so. Kant might have failed to appreciate the strength of
his own arguments. But I do not try to make the case that he did.14


i.4 The (Alleged) Gap in the Derivation of the Formula
of Universal Law
Readers familiar with Kant™s derivation of the Categorical Imperative might
wonder why it merits a book length treatment. After all, according to the
received view, it falls conspicuously short. Kant sketches his derivation of
this principle in both Groundwork I and II. Here are central (and famously
dif¬cult) passages in each:

But what kind of law can that be, the representation of which must determine the
will, even without regard for the effect expected from it, in order for the will to
be called good absolutely and without limitation? Since I have deprived the will of
every impulse that could arise for it from obeying some law, nothing is left but the
conformity of actions to universal law as such, which alone is to serve the will as its
principle, that is, I ought never to act except in such a way that I could also will that my
maxim should become a universal law. Here mere conformity to law as such, without
having as its basis some law determined for certain actions, is what serves the will as
its principle, and must so serve it, if duty is not to be everywhere an empty delusion
and a chimerical concept. (GMS 402)

When I think of a hypothetical imperative in general I do not know beforehand what
it will contain; I do not know this until I am given the condition. But when I think
of a categorical imperative I know at once what it contains. For since the imperative
contains, beyond the law, only the necessity that the maxim be in conformity with
this law, while the law contains no condition to which it would be limited, nothing is
left with which the maxim of action is to conform but the universality of a law as such;
and this conformity alone is what the imperative properly represents as necessary.
There is, therefore, only a single categorical imperative and it is this: act only on
that maxim through which you can at the same time will that it become a universal law.
Now, if all imperatives of duty can be derived from this single imperative as their
principle, then, even though we leave it undecided whether what is called duty is
not as such an empty concept, we shall at least be able to show what we think by it
and what the concept wants to say. (GMS 420“421)
Search for the Supreme Principle of Morality
8

In both passages, Kant argues for a conditional claim. If duty is not an
“empty” or “chimerical” concept, that is, if there are genuine moral obliga-
tions, then the Categorical Imperative is the principle of these obligations,
the supreme principle of morality. In both passages, Kant is offering a deriva-
tion, or part of a derivation, of the Categorical Imperative.
If we are to believe the received view, both the Groundwork I and the
Groundwork II derivation fail. They fail because they contain a crucial gap.
In each, Kant embraces a principle that is, for practical purposes, virtually
uninformative. Without argument, he then jumps to the Categorical Imper-
ative as the only viable candidate for the supreme principle of morality.
Bruce Aune offers an in¬‚uential expression of the received view. Aune
argues that both versions of the derivation fail, but let us follow him in fo-
cusing on Groundwork I.15 In the very sentence in which Kant sets out for
the ¬rst time the principle we refer to as the Categorical Imperative, he says
that “nothing is left but the conformity of actions to universal law as such,
which alone is to serve the will as its principle” (GMS 402). According to
Aune, Kant™s saying this amounts to his embracing the principle L : “Con-
form your actions to universal law.”16 L, suggests Aune, “is a higher-order
principle telling us to conform to certain lower-order laws.”17 L “formulates
the basic moral requirement”; it commands that we conform our actions
to these lower-order laws: principles that are necessarily binding on all of
us.18 But L does not tell us what these laws are. It fails to indicate, for exam-
ple, that among them we would ¬nd “Do not commit suicide,” rather than,
say, “Minimize your suffering.” Kant, Aune says, jumps directly from L to
the Categorical Imperative, which Aune calls C1: “Act only on that maxim
through which you can at the same time will that it should become a univer-
sal law.”19 In Groundwork I, Kant assumes that “we conform to universal law
(and so satisfy L) just when we obey C1 and act only on maxims that we can
will to be universal laws.”20
Yet, notes Aune, this assumption is far from obvious, as it is easy to il-
lustrate. Kant holds that in acting on a maxim of nonbene¬cence “ for
example, “To maximize my happiness, I will refrain from helping others in
need” “ I would be disobeying C1 (GMS 423). Suppose Kant is right about
this. According to the assumption in question, then, in acting on this maxim,
I would not be conforming to universal law: to a principle that is necessarily
binding on all of us. But it is unclear why I would not be. For all Kant has
shown thus far, it could be that a principle necessarily binding on all of us is:
“Always do what you believe will maximize your own happiness.” In acting on
my maxim of nonbene¬cence, I could be conforming to this universal law.
Kant, Aune suggests, embraces L as the basic requirement of moral action.
Kant af¬rms that if there is such a thing as moral action, then it is action
conforming to universal law. But then, without argument, Kant jumps to the
conclusion that the only way for an action to conform to universal law is for
it to conform to C1. The gap Aune ¬nds in Kant™s Groundwork I derivation is
Introduction 9

between the (for practical purposes) uninformative principle L and C1, the
Categorical Imperative.21
Aune is far from alone. Several other philosophers, even ones sympathetic
to a Kantian approach in ethics, have claimed to ¬nd a gap of this sort.22 In
their view, in neither Groundwork I nor II does Kant succeed in defending a
move he makes from a practically uninformative principle to the Categorical
Imperative.
Allen Wood, for example, has recently interpreted the Groundwork I and
II derivations in essentially the same way as Aune. According to Wood, in
both derivations Kant tries to establish that “our maxims ought to conform to
whatever universal laws there are.”23 But then Kant jumps without argument
from this rather empty principle to the Formula of Universal Law. Kant
illegitimately takes for granted that the only way to conform to whatever
universal laws there are is to conform to the Formula of Universal Law.
Henry Allison discusses another characterization of the practically un-
informative principle from which Kant (supposedly) moves directly to the
Categorical Imperative. On this characterization, the principle is (what I
call) the “principle of rightness universalism”:

RU: If a maxim or action is judged permissible for a rational agent in
given circumstances, it must also be judged permissible for any other
rational agent in relevantly similar circumstances.24

RU is rather vague: for one, it is not clear what are to count as “relevantly
similar circumstances.” However, this version of the traditional reading fo-
cuses on (what it sees as) Kant™s move directly from RU to the Categorical
Imperative. According to this version, Kant presents the Categorical Imper-
ative in a parenthetical clause aimed at explicating the prescription that
the will conform its actions to universal law as such, namely RU. Kant then
implicitly identi¬es RU with the Categorical Imperative or, at the very least,
claims that the former entails the latter.25
Obviously the two principles are not equivalent. Suppose someone acts
on Kant™s famous maxim of false promising: “When I believe myself in need
of money, I shall borrow money and promise to repay it, even though I
know that this will never happen” (GMS 422). If she acts on this maxim,
then, for well-known reasons I need not here restate, she violates the Cate-
gorical Imperative.26 But she does not necessarily violate RU. If she holds
her acting on the false-promising maxim to be morally permissible, nothing
need prevent her from judging that in circumstances relevantly similar to
her own, someone else™s acting on it would be morally permissible as well.
And the notion that RU entails the Categorical Imperative has little, if any,
more plausibility than the notion that the two principles are equivalent.
Kant gives us no reason to think that someone who embraced RU would
be rationally compelled also to endorse the Categorical Imperative. Once
Search for the Supreme Principle of Morality
10

again, it turns out that Kant™s argument suffers from a glaring gap. Whether
the practically uninformative principle is RU or L, Kant cannot legitimately
move directly from it to the Categorical Imperative.


i.5 Terminological and Thematic Clari¬cations
This book explores responses to the common view, just elaborated, that Kant
fails miserably at defending a foundational claim in this ethics, namely the
claim that if there is a supreme principle of morality, then it is the Categorical
Imperative.
Before sketching the book™s structure, I need to make a few clari¬cations,
some terminological, some thematic. I have used the term “the Categorical
Imperative” to refer to the principle Kant states at Groundwork 421 (cited
in i.4) and variant expressions of this principle, such as the one he gives at
Groundwork 402 (also cited in i.4). Kant himself refers to this principle as the
“categorical imperative,” without capitalization (GMS 421). I have adopted
the capitalization in order to emphasize that the term “categorical impera-
tive” need not be used to refer to the particular principle Kant sets forth at
Groundwork 421. In another, broader, Kantian usage, the term “categorical
imperative” refers to any principle that is absolutely necessary and binding
on all rational agents.27 A categorical imperative in this sense is a “practical
law” (GMS 420, 425, 428, 432; KpV 41). A burden of Kant™s discussion in
Groundwork I“II is to show that if there is a categorical imperative (that is
also the supreme, practical norm for the moral assessment of action), then it
is the Categorical Imperative. For the sake of clarity, I sometimes substitute
the term “Formula of Universal Law” for the “Categorical Imperative.”
In Groundwork II, Kant tells us that he has represented the supreme prin-
ciple of morality in “three ways” (GMS 436). He has represented it in the
Formula of Universal Law, as well as in two other formulas. These other two
are often referred to in the Kant literature as the Formula of Humanity and
the Formula of the Kingdom of Ends. The Formula of Humanity is this: “So
act that you treat humanity, whether in your own person or in the person
of any other, always at the same time as an end, never merely as a means”
(GMS 429, emphasis omitted). The Formula of the Kingdom of Ends seems
to run as follows: “[A]ll maxims from one™s own lawgiving are to harmo-
nize with a possible kingdom of ends as with a kingdom of nature” (GMS
436).28 According to Kant, these “three ways of representing the principle
of morality are at bottom only so many formulas of the very same law, and
any one of them of itself unites the other two in it” (GMS 436). So it seems
that for Kant these three formulas are, in a practical sense, equivalent “ for
example, any action that is morally impermissible according to one is also
morally impermissible according to each of the others.
In this book I discuss only the Formula of Universal Law and the Formula
of Humanity, leaving aside the Formula of the Kingdom of Ends.29 I focus
Introduction 11

on the ¬rst two formulas because they are the most familiar and, I think,
the most forceful Kantian candidates for the supreme principle of moral-
ity. Kant™s claim that all three are formulas of the “very same law” appears
to imply that the Formula of Universal Law and the Formula of Humanity
generate the same results regarding the moral status of actions.30 I do not
believe that they do, but an account of why will have to wait until Chapter 8.
Since I hold that the Formula of Universal Law (the Categorical Impera-
tive) and the Formula of Humanity differ in their implications regarding
the moral status of actions, I view them ultimately as competitors (albeit from
the same stable) for status as the only viable candidate for the supreme prin-
ciple of morality. This book considers derivations of two different Kantian
candidates for the supreme principle of morality: the Formula of Universal
Law and the Formula of Humanity.


i.6 Outline of the Book
Let me now explain brie¬‚y how the book unfolds and what it aims to show.
According to a traditional and widely accepted reading, there is a conspic-
uous gap in Kant™s Groundwork derivation of the Formula of Universal Law.
The book is composed of two main parts. In the ¬rst, I criticize contem-
porary responses to the traditional interpretation; in the second, I con-
struct a response of my own “ a response that leads to a new approach to
Kant™s derivations of both the Formula of Universal Law and the Formula
of Humanity.
If one accepts the traditional view that Kant™s Groundwork derivation of
the Formula of Universal Law plainly fails, it makes sense to look outside the
Groundwork for a derivation of this principle. Henry Allison does just this.
Appealing to the Critique of Practical Reason, Allison constructs an argument
(available to Kant if not explicitly made by him) that, in Allison™s view, estab-
lishes that if there is a supreme principle of morality, then it is the Formula
of Universal Law. According to Allison, this argument succeeds whereas that
of the Groundwork fails, since, unlike the latter, it relies on the assumption
that rational agents have what Kant calls “transcendental freedom” “ that
is, “independence from everything empirical and so from nature generally”
(KpV 97). I maintain in Chapter 2 that even if we accept Allison™s use of
the controversial notion of transcendental freedom, this derivation fails.
In short, Allison claims that as transcendentally free, rational agents, we
require a nonsensuously based justi¬cation of our maxims. Moreover, this
justi¬cation must be the maxims™ conformity to some practical law. But, con-
cludes Allison, this law could only be the Formula of Universal Law. I argue
that Allison does not successfully eliminate the possibility that conformity
to some different law justi¬es our maxims.
Of course, the Formula of Universal Law is not the only principle Kant
advocates. Among the others we ¬nd the Formula of Humanity, a principle
Search for the Supreme Principle of Morality
12

that many consider to be the most attractive Kantian candidate for the
supreme principle of morality. Does Kant establish that if there is such a
principle, then it is the Formula of Humanity? Chapter 3 focuses on this
question. There are two key steps in this derivation, which Kant undertakes
in Groundwork II. First, Kant claims that if there is a supreme principle of
morality (and thus a categorical imperative), then there is an objective end:
something that is unconditionally good. Second, he claims that this uncon-
ditionally good thing must be humanity. (If Kant proves these claims, he
shows that if there is a supreme principle of morality, then humanity is un-
conditionally good. But if humanity is unconditionally good, Kant can go
on to argue, then we are rationally compelled to do what the Formula of
Humanity commands, that is, always to treat it as an end in itself.) Recently
Christine Korsgaard has offered an in¬‚uential reconstruction of Kant™s de-
fense of these two key steps, especially the second. I contend that despite
Korsgaard™s efforts, the defense of neither step is adequate. Kant falls far
short of establishing that if there is a supreme principle of morality, then it
is the Formula of Humanity.
Given the inadequacy of both Kant™s Groundwork derivation of the For-
mula of Humanity and his second Critique derivation of the Formula of
Universal Law (as reconstructed by Allison), the prospects for establishing
that only a Kantian principle could be the supreme principle of morality
seem very grim indeed. The second part of the book aims to show that we
can make more progress toward establishing this than one might think.
Chapter 4 challenges the traditional reading of Kant™s Groundwork deriva-
tion of the Formula of Universal Law, the reading according to which the
derivation contains an unwarranted jump from a practically empty princi-
ple to this formula. The chapter introduces a new, criterial reading of the
derivation, according to which it has three main steps. First, Kant develops
criteria that any viable candidate for the supreme principle of morality must
ful¬ll. These criteria include, but are not limited to, those that belong to
his basic concept of this principle. Second, Kant tries to establish that no
possible rival to the Formula of Universal Law ful¬lls all of these criteria.
Finally, Kant attempts to demonstrate that the Formula of Universal Law
remains as a viable candidate for a principle that ful¬lls all of them. With
these three steps, Kant strives to prove that if there is a supreme principle
of morality, then it is this formula. Defending a rejection of the traditional
interpretation of this derivation in favor of the criterial reading obviously
requires considerable textual analysis. Much of Chapter 4 focuses on dif-
¬cult passages in the Groundwork, including the ones cited in i.4. I aim to
show that the text of Kant™s derivation (in both Groundwork I and II) permits
the criterial reading. At the end of Chapter 4, I offer a preliminary list of
criteria, in addition to the ones contained in his basic concept, that Kant
develops for the supreme principle of morality.
Chapter 5 focuses on this list of four criteria. How are we to interpret
the criteria, and how does Kant defend them? The criterion that demands
Introduction 13

most of our attention can be stated thus: the supreme principle of morality
must be such that all and only actions conforming to this principle because
the principle requires it “ that is, all and only actions done from duty “
have moral worth. An advocate of a particular principle as the only viable
candidate for the supreme principle of morality must, according to Kant,
be able (rationally speaking) to maintain that an agent™s action has moral
worth if and only if she does it from duty, that is, because this principle
requires it. Chapter 5 probes both the meaning of this criterion and Kant™s
arguments for it.
It is one thing to understand this criterion and Kant™s defense of it; it
is quite another to embrace the criterion. Chapter 6 poses the question of
whether we should do so. I argue that we should accept one part of the crite-
rion (modi¬ed slightly) but reject another part. We should accept the idea
that the supreme principle of morality must be such that all instances of
willing to conform to it because the principle requires it have moral worth;
but we should reject the notion that the supreme principle must be such
that only instances of willing to conform to it because the principle requires
it have moral worth. An advocate of a certain candidate for the supreme
principle of morality, say the Formula of Universal Law, must acknowledge
that an agent™s action can have moral worth even if she does not do it be-
cause this principle requires it. Indeed, I argue that Kantian considerations
rationally compel the advocate to acknowledge that actions forbidden by the
Formula of Universal Law can have moral worth.
By the end of Chapter 6 we will have a complete list of Kant™s criteria
for the supreme principle of morality. In addition to the four that belong
to Kant™s basic concept of this principle, there are four others, modi¬ed in
accord with the argument of the chapter. According to these, the supreme
principle of morality must be such that: (v) every case of willing to conform
to it because the principle requires it has moral worth; (vi) the moral worth
of willing to conform to the principle because the principle requires it stems
from its motive, not from its effects; (vii) an agent™s representing the princi-
ple as a law, that is, as a universally and unconditionally binding principle,
provides him with suf¬cient incentive to conform to it; and, ¬nally, (viii) a
plausible set of duties (relative to ordinary rational moral cognition) can be
derived from the principle.
The ¬rst step of Kant™s derivation is to establish criteria for the supreme
principle of morality; the second is to show that no possible rival to the For-
mula of Universal Law ful¬lls all of them. Chapter 7 focuses on this second
step. In the ¬rst instance, the criterial reading I defend is a reading of Kant™s
derivation of the Formula of Universal Law. It is, however, open to Kant to
employ the same steps in deriving the Formula of Humanity. In any case, the
chapter tries to show that with the help of some of these criteria “ ones the
plausibility of which I defend “ Kant can eliminate key competitors to both
of these principles. For example, relying on criteria v and vi, Kant is able
to construct a kind of argument, which I call a “valuational argument,” that
Search for the Supreme Principle of Morality
14

succeeds in eliminating many consequentialist candidates for the supreme
principle of morality, including a utilitarian principle such as: “Always per-
form a right action: one that yields just as great a sum total of well-being
as would any alternative action available to you.” However, the valuational
type of argument does not apply to nonconsequentialist principles, such
as this detheologized imperative based on the Ten Commandments: “You
ought to honor your father and mother; you ought not to kill; you ought
not to commit adultery; you ought not to steal; you ought not to bear false
witness; you ought not to covet anything that is your neighbor™s.” But, as
Chapter 7 also tries to show, Kant is not without effective recourse against
such principles.
To complete the second step of his derivation of the Formula of Universal
Law, Kant must demonstrate that no possible rival to this principle ful¬lls
all of the criteria he develops. He must eliminate not just a few familiar
rivals but all possible principles other than the Formula of Universal Law
as contenders for the supreme principle of morality. Yet, from the outset,
it is hard to see how Kant could eliminate all possible contenders, if only
because it is unclear how he could prove that he had even taken all of them
into account. In my view, Kant does not prove this. I do not claim that Kant
successfully dismisses all rivals to the Formula of Universal Law (or that
he could successfully dismiss all rivals to the Formula of Humanity). I do,
however, defend the view that he presents compelling arguments against
some main rivals, including many consequentialist principles.
On the criterial reading, the third step of Kant™s derivation of the Formula
of Universal Law is to show that, unlike its rivals, this principle remains as
a viable candidate for one that ful¬lls the whole set of criteria Kant has
developed for the supreme principle of morality. (Showing that this formula
actually does ful¬ll the whole set of criteria would involve giving a deduction
of it. One of the criteria, one that belongs to Kant™s basic concept, is that
the supreme principle of morality be binding on all rational agents. No
derivation could show even that the Formula of Universal Law is binding
on all human rational agents, that is, that all of us are genuinely obligated
to conform to it. A deduction, not a derivation, of the Formula of Universal
Law would be needed for this.) In Chapter 8 I argue that the Formula of
Universal Law stands as a viable candidate for ful¬lling Kant™s basic concept
of the supreme principle, if we are willing to modify this concept slightly
to accommodate my criticisms in Chapter 6 of how Kant views the relations
between the supreme principle of morality and moral worth. The Formula
of Universal Law is also not disquali¬ed by three of Kant™s further criteria.
However, a serious problem arises regarding the fourth additional criterion,
namely the one according to which the supreme principle of morality must
be such that a plausible set of duties (relative to ordinary rational knowledge
of morals) can be derived from the principle. The Formula of Universal Law
is dif¬cult to interpret; there is much debate about how, precisely, to apply it
Introduction 15

in determining whether acting on a particular maxim is morally permissible.
So the question remains: which duties stem from it? I do not offer anything
approaching a thorough discussion of this question. But I try to show that on
some leading interpretations of the Formula of Universal Law, this principle
fails to generate moral prescriptions that square with common sense.
As I mentioned earlier, the criterial reading applies in the ¬rst instance
to Kant™s Groundwork derivation of the Formula of Universal Law. Yet there
seems to be no reason why Kant could not take the same steps in a deriva-
tion of the Formula of Humanity that, according to this reading, he goes
through in his derivation of the Formula of Universal Law. (If, as I hold, the
two formulas are not equivalent, then a successful derivation of the latter
would actually preclude a successful derivation of the former.) I argue in
Chapter 8 that, like the Formula of Universal Law, the Formula of Human-
ity remains as a viable candidate for a principle that satis¬es Kant™s basic
concept of the supreme principle of morality (if we modify this concept
slightly), as well as three of the four further criteria Kant develops. But does
the Formula of Humanity generate a plausible set of moral prescriptions?
This question is dif¬cult, since the Formula of Humanity itself poses inter-
pretive challenges. Without pretending to give a full treatment of the issue,
I argue that the Formula of Humanity holds more promise on this score
than does the Formula of Universal Law, although it too has some troubling
aspects.
This is where the book ends. It begins in Chapter 1 with a brief examina-
tion (too brief, I am afraid, to be entirely satisfactory) of some basic concepts
in Kant™s theory of agency. We have already invoked the notions of a maxim,
the will, acting from inclination, and so forth. We need to clarify them in
order to proceed without confusion.
As is already apparent, the book focuses mainly on arguments Kant makes
in the Groundwork and the second Critique, since these are the works in
which Kant is concerned with deriving the supreme principle of morality.
Of course, I invoke discussions in Kant™s other works in ethics, for example,
the Metaphysics of Morals. However, the book does not in any way aim to give
a comprehensive account of Kant™s ethical doctrine.
In sum, the book sets out a new reading of Kant™s Groundwork derivation
of the Formula of Universal Law. It tries to show that this argument is philo-
sophically far richer than the traditional interpretation suggests. No, Kant
does not succeed in proving his strikingly ambitious claim that if there is a
supreme principle of morality, then it is the Formula of Universal Law. But
he does offer some strong reasons for rejecting rivals to this principle. What
is more, Kant™s derivation of the Formula of Universal Law opens the door
to a heretofore unexplored way of defending the Formula of Humanity, a
principle that many of us ¬nd especially attractive as a candidate for the
supreme principle of morality.
1

Fundamental Concepts in Kant™s Theory of Agency




1.1 Aims and Limits of the Discussion
Kant peppers each of his major works in practical philosophy with comments
pertaining to what it means for us, rational agents, to act. Philosophers
disagree on how best to interpret these comments, which are often dif¬cult
and sometimes obscure.1 I offer some readings here that, I believe, cohere
with Kant™s texts, but they are surely not the only defensible readings. My
aim in this chapter is to set out a plausible interpretation of (part of ) Kant™s
theory of agency, an interpretation that will be useful as a reference point
in discussions to come. Important issues regarding Kant™s theory of agency,
such as whether Kant does or should conceive of acting on a maxim on the
model of Aristotle™s practical syllogism, are not addressed here. A thorough
reading of Kant™s theory of agency, let alone a defense of it, would require
a book in itself.
The chapter is divided into two main parts. The ¬rst focuses on a few
key concepts in Kant™s theory. In 1.2“3, I offer an account of Kant™s notion
of a maxim; then I turn very brie¬‚y to Kant™s conceptions of the will (1.4)
and of the will™s “determining grounds” (1.5). The second main part of the
chapter concerns Kant™s account of actions not done from duty, that is, ones
done on “material practical principles” (1.6“8). Understanding this account
requires some painstaking textual analysis. I explain in section 1.6 why, in
light of the main aims of this book, it is important to grasp Kant™s account
of actions not done from duty.


1.2 Maxims: A Basic Account
Let us begin, then, with the concept of a maxim. Kant tells us that a maxim
is a subjective principle of acting (GMS 421, note).2 By following R¨ diger
u
Bittner and considering the sense in which a maxim is a subjective principle
and that in which it is a principle of acting, we can develop a basic account
16
Kant™s Theory of Agency 17

of maxims.3 Having an example of a maxim at hand helps us to do so.
Suppose that Mary has adopted the maxim M: From self-love, I will shorten
my life when its longer duration threatens more troubles than it promises
agreeableness.4 A maxim is subjective in three respects. First, if there is a
maxim, then there is a subject “ that is, an agent “ who holds it. A maxim
is always some agent™s rule. If neither Mary nor anyone else held M, then
it would not be a maxim.5 Second, an agent chooses his own maxims. Kant
calls maxims “rules imposed upon oneself” (GMS 438). At any time he is
free to discard the maxims he presently holds and to adopt new ones. Mary
may have held M for the past thirty years, but it is up to her whether she
will hold it even for the next thirty seconds. Third, an agent™s maxim is
a subjective principle in that it applies only to her own action (KpV 19).
Mary™s maxim expresses what she requires herself to do if continuing to live
threatens more evil than satisfaction for her. It does not tell anyone else
what he is required to do in these circumstances.
Maxims are not just subjective principles; they are subjective principles
of acting. Agents act on (nach) maxims. This means that maxims play a role
in the generation of their actions. An agent does not merely apply a maxim
in hindsight to his action after it has occurred. If Mary has acted on M by
taking poison, then M, or, more likely, a less precise representation of it,
has contributed to the generation of her action. Of course, that someone
has adopted a maxim “ that is, given herself the requirement of acting in
a certain way under certain circumstances “ does not entail that she will
act on it. The occasion for acting on it may simply never arise. Mary may
never come to believe that her life™s continuing threatens more troubles
than agreeableness. Even if the occasion for acting on a maxim does arise, an
agent is free not to act on it. She may just choose not to abide by the principle
of action that she has given herself. Although faced with the prospect of a
miserable old age, Mary might obey the Categorical Imperative and refrain
from acting on M, that is, refrain from killing herself.6
Philosophers typically hold that for Kant, all acting is acting on a maxim.7
It is not hard to defend this interpretation. According to Kant, all of an
agent™s actions are either morally permissible or morally impermissible.8
The Categorical Imperative “ “Act only on that maxim through which you
can at the same time will that it become a universal law” (GMS 421, emphasis
omitted) “ gives us a procedure for determining whether an action per-
formed on a maxim is morally permissible. A person™s action is morally per-
missible only if she can will the maxim on which she performs it to become a
universal law. If she cannot do so, then the action is morally impermissi-
ble. The principle does not give us a procedure for determining whether
an action performed on no maxim is morally permissible. Kant, of course,
takes the Categorical Imperative to be the supreme principle of morality. He
suggests that it is the canon of the moral estimation of our action as a whole
(GMS 424). If there were questions of moral permissibility to which the test
Search for the Supreme Principle of Morality
18

embodied in the Categorical Imperative could give no answer, then Kant™s
claim that this imperative is the supreme principle of morality would be hol-
low. With these considerations in mind, it is easy to show that, for Kant, all
acting is acting on a maxim. Suppose that agents could perform actions with-
out doing so on any maxim. The Categorical Imperative procedure would
then yield no answer to the question of their moral permissibility, and the
Categorical Imperative would thus not be the supreme principle of morality.
Since Kant af¬rms it to be the supreme principle of morality, he must hold
that agents perform each and every one of their actions on a maxim.
Kant™s own examples of maxims illustrate that, at a minimum, they are
rules that specify a type of action to be performed in a type of situation,
for example, “When I believe myself to be in need of money, I shall borrow
money and promise to repay it, even though I know that this will never
happen” (GMS 422). When fully speci¬ed, however, it seems that a maxim
also includes a description of the agent™s end in doing what she does. In the
Groundwork and in the Metaphysics of Morals, Kant suggests that all maxims
contain a (description of ) an end (GMS 436, MS 395).9 The end implied in
the maxim of false promising is presumably that of getting money. Moreover,
some of the maxims Kant discusses contain descriptions of an incentive,
for example, the maxim on which Mary™s maxim is based: “From self-love, I
make it my principle to shorten my life when its longer duration threatens
more troubles than it promises agreeableness” (GMS 422, emphasis added).
Here the agent™s end, that is, the state of affairs he would aim to realize if he
acted on the maxim, remains implicit, although it is obviously something like
that of being free from that suffering which is not outweighed by happiness.
The agent™s incentive “ that which would motivate him to act if he acted on
the maxim “ is explicit; it is “self-love.”10
The notion that when fully spelled out, maxims contain descriptions of
an agent™s incentive for acting gains support from Kant™s well-known claim
in the Religion within the Limits of Reason Alone that the “freedom of the will
[Willk¨ r] is of a wholly unique nature in that an incentive can determine the
u
will to an action only so far as the individual has incorporated it into his maxim”
(Rel 23“24, English ed. 19). Later, in connection with Henry Allison™s at-
tempt to ¬ll the (apparent) gap in the Groundwork derivation, we discuss
this claim in detail. For now, note that, in Kant™s view, we have freedom of
the will. Moreover, if our will is determined to an action, some incentive
constitutes a basis for this determination.11 All of our actions are such that
we have some incentive for performing them. (The typical sneeze or slip on
a banana peel does not count as an action in the relevant sense.) Therefore,
Kant™s claim in the Religion implies that whenever we act, we do so on some
maxim that, if fully speci¬ed, would include a description of our incentive
for acting.12 A fully expressed maxim would include not only a description
of a kind of action to be performed in a kind of situation, but also a spec-
i¬cation of the agent™s end and of his incentive in performing it. A fully
Kant™s Theory of Agency 19

expressed maxim would take the form of a rule that includes each of these
elements. Of course, when we act, we might not have each of these elements
in mind.13


1.3 Maxims and Other Rules of the Same Form
Before ending our discussion of maxims, we need to address one more issue,
namely that of how to distinguish them from other rules of the same form.
This issue is important. Suppose that someone in taking a karate lesson acts
on the rule: “From self-love, every Monday at 3 p.m. I take live karate lessons
in order to improve my endurance and ¬‚exibility.” It seems reasonable to
assume that, at the same time, she might also be acting on a different, more
general rule: “From self-love, during my free time I exercise in order to stay
in shape.” If we took both rules to be maxims on which the agent acted,
then Kant would face a serious problem. At least on one common reading,
acting on the ¬rst rule would violate the Formula of Universal Law, whereas
acting on the second would not. I take it to be obvious that acting on the rule
of exercising during one™s free time is in accordance with this formula. But
consider the rule of taking karate lessons with a live instructor on Monday at
3:00 p.m. Not every agent could take live karate lessons Monday at 3:00 p.m.
An agent cannot take a live lesson without a live instructor. But if all agents
were taking live karate lessons Monday at 3:00, then there would be no
instructor available to give lessons at this time. Given that not every agent
could take live karate lessons every Monday at 3:00 p.m., it is not possible (as
a rational being) to will that it become a universal law that every agent does
so.14 If both rules count as maxims, then it seems that our agent™s action of
taking a karate lesson is morally impermissible. For she is acting on a maxim
such that she cannot, at the same time, will that it become a universal law.
To avoid the dif¬culty suggested by this example, we must have a means of
deciding which of the rules an agent acts on counts as the maxim of his
action.
Unfortunately, Kant does not explain how to do this. The best way in
my view is to specify that the maxim of an agent™s action is the fundamental
rule, of the form required of a maxim, on which he acts.15 (Recall that, at
least implicitly, a maxim must have the form of a subjective rule according
to which, from a speci¬ed incentive, an action is to be taken in designated
circumstances in order to realize some end.) More speci¬cally, a practical
rule Q of the requisite form has status as the fundamental rule of this form
on which an agent performs an action when it ful¬lls either one of the
following two conditions: Q is the only such practical rule on which he
performs the action; or Q is not the only such rule on which the agent
performs the action but is rather the most general rule of this form on
which he does so. If Q ful¬lls this second condition, it governs the agent™s
selection of a more speci¬c rule of the same form, that is, a rule ancillary to
Search for the Supreme Principle of Morality
20

Q, through which rule he implements Q ( performs an action). The practical
rule “From self-love, every Monday at 3:00 p.m. I take live karate lessons in
order to improve my endurance and ¬‚exibility” is an example of one that
might be ancillary to the maxim “From self-love, whenever I have free time,
I exercise in order to stay in shape.” An agent who adopted the latter might
take up the former as a rule for implementing it. She would presumably do
so because, as it happens, she has Monday afternoons free, wants to improve
her endurance and ¬‚exibility, and judges that training in a martial art would
be a good way of doing so. Given her circumstances, she would choose to
act on her maxim by acting on this more speci¬c rule. Of course, another
agent who had adopted this maxim might choose a different rule through
which to act on it.
In sum, a maxim is a subjective principle of acting. It is a subjective prin-
ciple in that it is held by some agent, it can be freely adopted or discarded by
her, and it applies only to her own actions. An agent™s maxims are principles
of acting in that they play a role in the generation of her actions. When
fully expressed, a maxim includes a description of a kind of action to be
performed in a kind of situation, as well as a speci¬cation of the agent™s end
and incentive in performing it. Not all rules of this form count as maxims,
however. An agent™s maxim is the fundamental rule of this form on which
she acts. This reading of Kant™s views regarding maxims is by no means thor-
ough (or thoroughly defended), but it will, I hope, serve to ¬x ideas for
discussions to come.


1.4 The Will
Another key concept in Kant™s theory of agency is that of the will. Unfortu-
nately, Kant™s account of the will is a terminological mire. In the Groundwork
and the second Critique, he typically uses Wille to refer to an agent™s capacity
to act on rules, for example, maxims or imperatives (see, e.g., KpV 32).16 But
he also uses Wille to refer in addition to an agent™s capacity to give herself
the rules on which she has the capacity to act, for example, to legislate for
herself maxims or imperatives (e.g., GMS 431 and KpV 33). Later, in the
Metaphysics of Morals, Kant typically employs Wille to refer only to the latter
capacity (e.g., MS 213). We might call an agent™s capacity to act on rules the
“executive Wille” and his capacity to give himself these rules the “legislative
Wille.”17 In the Metaphysics of Morals (and elsewhere) Kant employs another
term, Willk¨ r, that is sometimes translated as “will.”18 For our purposes, it
u
will be safe to consider Willk¨ r as the same capacity as executive Wille, that
u
is, the capacity to act on rules.19
Fortunately, we need to focus only on Kant™s notion of the executive Wille,
to which I refer here simply as the will. According to Kant, to exercise the
capacity of will “ that is, to will “ is to act. That is why Kant de¬nes the
(executive) Wille as the capacity to act on principles (GMS 412). Willing is
Kant™s Theory of Agency 21

more than wishing or even deciding to do something. Someone might wish
or decide to realize some object (e.g., to get away for a weekend at a bed and
breakfast) yet change his mind and never actually make any effort to realize
this object (e.g., never do any planning for the getaway). Willing involves
making some effort to realize what one wills. In this sense, it is a kind of
acting. In what follows, I alternate between speaking in terms of willing and
in terms of acting. For our purposes, the two amount to the same thing:
trying (on the basis of some rule[s]) to secure some objective.


1.5 Determining Grounds of the Will
The will is a capacity to act on rules. But what is a “determining ground”
of the will? As determining grounds of the will, Kant mentions (at least)
ends, inclinations, the expectation of pleasure, the principle of one™s own
happiness, and the moral law (see respectively MS 381; KpV 81, 22, 35, 72).
I assume that each of the determining grounds (Bestimmungsgr¨ nde) of the
u
will he mentions counts as such by standing in some particular relation
to willing. But, to my knowledge, Kant never says explicitly just what this
relation is. It seems to me plausible to interpret determining grounds of the
will as motivating reasons or, more simply, motives for willing. They are what
bring about willing. In Kant™s view, however, each item on the list actually
brings about an agent™s willing only if she has taken account of it in her
maxim, that is, made it part of a rule on which she acts. In other words, each
of these items on its own might count as an incentive for an agent™s acting,
but the items actually motivate her to act only if she has incorporated them
into some self-given rule.20 For example, an agent might have an inclination
to eat ice cream. But, according to Kant, this inclination determines her will
(i.e., actually motivates her) only if she has taken account of it in some
maxim “ for example, one of allowing herself small pleasures to promote
her happiness.21
One might wonder whether determining grounds of the will count not
only as motivating but also as “justifying” reasons for acting. That depends
on the sense of justifying reason one employs. Let us consider one particular
kind of determining ground of the will, namely inclinations. Obviously, that
someone has a particular inclination as a motive does not entail that, from
an impartial perspective, her acting on this motive is justi¬ed. (Acting from
the inclination to be the richest person in the county, a businessperson
might hire someone to kill her competitor.) Determining grounds of the
will are not justifying reasons in the sense of reasons that, from an impartial
perspective, always do in fact justify an agent™s action. Moreover, that an
agent has a particular inclination as a motive does not even entail that,
from her own perspective, her acting on this motive actually justi¬es her
action. If a particular inclination serves as an agent™s motive in acting, then
she has incorporated this motive into one of his maxims. But she might
Search for the Supreme Principle of Morality
22

hold that her acting on this maxim is itself ultimately unjusti¬ed because
it is morally unjusti¬ed. For example, if the agent has Kantian leanings, she
might believe that her indulging her inclination to be the richest person
in the county by acting on a maxim of ordering a hit on her competitors is
contrary to Kantian duty and therefore ultimately unjusti¬ed.
However, Kantians have recently emphasized that, as a rational being,
an agent must believe that acting on her maxim is in some sense good or
rationally justi¬able.22 If she does not meet this “justi¬cation requirement”
by holding that acting on the maxim is good morally, she must meet it by
holding that acting on it is good prudentially. She would, for example, meet
the requirement by virtue of believing that, given her end (e.g., to be the
richest person), taking the means to it speci¬ed in the maxim (e.g., killing
her competitor) is good in that it will likely be effective. In short, although
a given determining ground of the will need not constitute a reason that
actually justi¬es what an agent does, either from an impartial or from even
her own perspective, she must hold that it is good, in some sense, for her to
act on the maxim in which this determining ground has been incorporated.


1.6 Acting from Inclination: Three Interpretations
and Their Importance
This brief examination of maxims, the will, and determining grounds of the
will puts us in position to do some ¬nal stage setting for the main arguments
of this book. In sections 1.6“8, we focus on Kant™s account of actions that
are not done from duty.23
Since these sections involve painstaking textual analysis, it is helpful be-
fore proceeding to have some idea of how they further the main aims of this
book. In Chapter 4, I begin to defend a criterial reading of Kant™s derivation
of the Categorical Imperative. According to this reading, Kant develops cri-
teria for the supreme principle of morality. He then tries to show that no rival
to the Categorical Imperative for status as this principle can ful¬ll the full set
of criteria. Finally, Kant suggests that the Categorical Imperative remains as
a viable candidate for ful¬lling the full set. So Kant™s criteria for the supreme
principle of morality are obviously crucial to my reading of his derivation.
One criterion he develops is the following: the supreme principle of morality
must be such that all and only actions done because the principle requires
it “ that is, all and only actions done from duty “ have moral worth. It is
not possible to comprehend this criterion, let alone to gauge its plausibility,
without grasping what, according to Kant, it means to act from duty. But
in order to grasp this we need to understand Kant™s account of actions not
done from duty. For example, only by understanding this account can we
see that for Kant all actions done from duty are done from duty alone. For
Kant there simply are no “overdetermined” actions, ones done (at the same
time) from both duty and inclination (section 5.3). Since Kant™s criterion
Kant™s Theory of Agency 23

does not allow that an action can be done from both duty and inclination,
it implies the view that absolutely no actions have moral worth other than
those done exclusively from the incentive of conforming to moral principle.
(In Chapter 6 I argue that this view is implausible. Kant should drop it from
his criterion and maintain instead merely that all actions from duty have
moral worth.)
In the spirit of Kant™s practical philosophy, though not in its idiom, we
might call actions not done from duty “nonmoral” actions. For Kant, of
course, not all nonmoral actions are immoral. A nonmoral action can be
morally permissible: even though it is not done from duty, it can be in accor-
dance with it “ for example, the action of a shopkeeper not overcharging
an inexperienced customer (GMS 397). According to Kant, all nonmoral
actions “ that is, all actions not done from duty “ are done from inclination
(GMS 413, note).
Many philosophers believe that Kant defends a radically hedonistic ac-
count of non moral action. According to the traditional interpretation, Kant
holds that whenever an agent acts nonmorally, she is motivated solely by the
desire for pleasure.24 Pointed criticisms of Kant have arisen from the no-
tion that he embraces this account, with one philosopher going so far as to
charge that Kant™s account is not only false, but “utterly repugnant, deroga-
tory, and degrading.”25 The most obvious objection to the account is that it
fails to square with the phenomena. Agents seem to be motivated by more
than a desire for pleasure, even when they are not acting from duty. Con-
sider a serious pianist who in practicing a sonata is acting solely from her
inclination to master the piece. Depending on the circumstances, many of
us would ¬nd plausible her opinion that her motivation for practicing in-
cludes a desire to play the piece beautifully: a desire that she does not aim to
satisfy for the sake of the pleasure its satisfaction promises. If the traditional
reading is correct, then Kant defends a suspect account of nonmoral action.
Recently Andrews Reath has offered an innovative and in¬‚uential argu-
ment against the traditional construal of Kant™s account.26 Philosophers
have misinterpreted the relations Kant believes to hold between pleasure
and inclinations, says Reath. Contrary to the traditional reading, Kant does
not claim that in trying to satisfy an inclination, an agent is always motivated
by the prospect of gaining pleasure for herself. He claims rather that plea-
sure plays a role in the development of inclinations.27 An agent would not
develop an inclination for an object, say, mastering a piano sonata, unless
she expected that she would gain pleasure from realizing it. Once an agent
has an inclination for an object, however, in pursuing it she need have no
hedonic motivation at all. Once she has an inclination to master a sonata,
the agent™s motives in practicing it need not include her own pleasure.
I trust that the appeal of Reath™s interpretation is evident. Unfortunately,
the interpretation fails to cohere with Kant™s doctrine, or so I contend. Exam-
ination of Kant™s de¬nitions of inclination, as well as some of his remarks on
Search for the Supreme Principle of Morality
24

material practical principles, suggests that he did indeed hold each action
from inclination to have hedonic motivation. Nevertheless, for philosophers
sympathetic to Kant but not to a radically hedonistic account of nonmoral
action, all might not be lost. In my view, although Kant™s assertions permit a
reading on which an agent™s own pleasure constitutes her only motive in act-
ing nonmorally, they do not require it. They also permit the interpretation
that, whenever an agent acts from inclination, she has her own pleasure
as one, but not necessarily as her only, motive. I call this the “alternative
interpretation.”28
The alternative interpretation seems more attractive than the traditional
one. According to the former, if, from inclination, an agent writes a short
story or practices the piano, one of her motives must be her own pleasure.
Yet at the same time she might have other motives: the desire to exercise her
creativity or to play beautifully: desires the agent does not strive to satisfy for
the sake of pleasure. On the alternative, Kant avoids the suspect reduction of
all nonmoral motives to one. He can acknowledge some of the complexity of
acting in ways other than from duty. As we will see, however, the traditional
interpretation ¬ts more naturally with some of Kant™s claims in the second
Critique than does the alternative.


1.7 Acting from Inclination in the Groundwork
and in the Metaphysics of Morals
To construct an interpretation of Kant on nonmoral action, we must engage
in close reading of some dif¬cult passages. To begin, in an often overlooked

. 1
( 8)



>>