Sie sind auf Seite 1von 357

Groundwork for the

Metaphysics of Morals

Edited by
Christoph Horn
Dieter Schnecker

Walter de Gruyter
Groundwork for the Metaphysics of Morals


Groundwork for the
Metaphysics of Morals
Edited by
Christoph Horn and Dieter Schnecker

in cooperation with
Corinna Mieth

Walter de Gruyter Berlin New York


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

ISBN-13: 978- 3-11-017707-7


ISBN-10: 3-11-017707-2

Bibliographic information published by Die Deutsche Bibliothek


Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data is available in the Internet at "http://dnb.ddb.de#.

Copyright 2006 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin
All rights reserved, including those of translation into foreign languages. No part of this
book may be reproduced or transmitted in any form or by any means, electronic or
mechanical, including photocopy, recording or any information storage and retrieval
system, without permission in writing from the publisher.
Printed in Germany
Typesetting: OLD Media, Neckarsteinach
Printing and binding: Druckhaus Thomas Mntzer, Bad Langensalza
Cover design: Martin Zech, Bremen
Contents

Christoph Horn and Dieter Schnecker


Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Notes on Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Groundwork, Preface
Nico Scarano (Tbingen)
Necessity and Apriority in Kants Moral Philosophy: An Inter-
pretation of the Groundworks Preface (GMS, 387392) . . . . . 3

Groundwork I
Allen Wood (Stanford)
The Good Without Limitation (GMS I, 393394). . . . . . . . . . . . 25
Christoph Horn (Bonn)
Kant on Ends in Nature and in Human Agency:
The Teleological Argument (GMS I, 394396) . . . . . . . . . . . . . . 45
Marcia Baron (Indiana)
Acting from Duty (GMS I, 397401) . . . . . . . . . . . . . . . . . . . . . . 72
Harald Khl (Berlin)
The Derivation of the Moral Law (GMS I, 402403) . . . . . . . . . 93

Groundwork II
Marcus Willaschek (Frankfurt a. M.)
Practical Reason. A commentary on Kants Groundwork of
the Metaphysics of Morals (GMS II, 412417) . . . . . . . . . . . . . . 121
Bernd Ludwig (Gttingen)
Kants Hypothetical Imperatives (GMS II, 417419) . . . . . . . . . 139
vi Contents

Mark Timmons (Memphis)


The Categorical Imperative and Universalizability
(GMS II, 421424). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Samuel Kerstein (Maryland)
Deriving the Formula of Humanity (GMS II, 427437) . . . . . . 200

Groundwork III
Klaus Steigleder (Bochum)
The Analytic Relationship of Freedom and Morality
(GMS III, 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Corinna Mieth and Jacob Rosenthal (Bonn)
Freedom must be presupposed as a property of the will of all
rational beings (GMS III, 2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Marcel Quarfood (Stockholm)
The Circle and the Two Standpoints (GMS III, 3) . . . . . . . . . . . 285
Dieter Schnecker (Siegen)
How is a categorical imperative possible? Kants deduction of
the categorical imperative (GMS III,4) . . . . . . . . . . . . . . . . . . . . 301

A Bibliography on Kants Groundwork . . . . . . . . . . . . . . . . . . . . 325


General Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
Preface

Kants Groundwork for the Metaphysics of Morals is one of the most


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

ological considerations to support his idea of a good will? Does Kant


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

Bonn and Siegen, April 2006


Christoph Horn and Dieter Schnecker
Notes on Contributors

Marcia Baron is Rudy Professor of Philosophy at Indiana University


(Bloomington, Indiana). She is the author of Kantian Ethics Almost
without Apology (1995), and with Philip Pettit and Michael Slote,
Three Methods of Ethics (1997), as well as articles on a variety of
topics in ethics and philosophy of law, including friendship and im-
partiality, patriotism, justifications and excuses, sexual consent, and
manipulativeness.
Christoph Horn is Professor of Philosophy at the University of Bonn.
He works on various aspects of ancient philosophy (Plato, Aristotle,
Plotinus, Augustine) as well as modern moral philosophy. Publica-
tions: Plotin ber Sein, Zahl und Einheit (1995), Augustinus (1995),
Antike Lebenskunst (1998), Politische Philosophie (2003), Grundle-
gende Gter (forthcoming). He edited the following volumes: Augus-
tinus, De civitate dei (1997), (with Ch. Rapp) Wrterbuch der antiken
Philosophie (2002), (with N. Scarano) Philosophie der Gerechtigkeit.
Texte von der Antike bis zur Gegenwart (2002).
Samuel Kerstein is Associate Professor of Philosophy at the Univer-
sity of Maryland, College Park. He is the author of Kants Search for
the Supreme Principle of Morality (2002) and is working on a recon-
struction and defense of Kants Formula of Humanity.
Harald Khl is Lecturer at the University of Saarbrcken. He was
Guest Professor of Philosophy at the Humboldt University (Ber-
lin). His most important publications are: Kants Gesinnungsethik
(1990), Abschied vom Unbedingten. ber den heterogenen Charak-
ter moralischer Forderungen (2006). Papers on practical philosophy
(Kant, Schopenhauer, Williams, Rorty), theory of knowledge and
semantics.
Bernd Ludwig is Professor of Philosophy at the University of Gt-
tingen. His main research areas are the philosophy of I. Kant, moral
philosophy and the theory of Natural Law in the early modern period.
He is editor of Kants Metaphysics of Morals. Most important publica-
x Notes on Contributors

tions: Kants Rechtslehre (22005), Die Wiederentdeckung des Epikure-


ischen Naturrechts [on Th. Hobbes] (1998).
Corinna Mieth (MA 1999 and PhD 2002 at the University of T-
bingen) teaches ethics and political philosophy at the University of
Bonn. Her research interests are focussed on duties of assistance, on
the foundation of human rights, especially social rights, on the politi-
cal philosophy of John Rawls and theories of global justice. She is co-
author of a commented edition of Kants Groundwork with Christoph
Horn and Nico Scarano.
Marcel Quarfood is Senior Lecturer in Philosophy at Sdertrn
University College, Huddinge, Sweden. He translated Kants Pro-
legomena into Swedish, and published Transcendental Idealism and
the Organism. Essays on Kant (2004), as well as articles on Kant and
on the philosophy of biology.
Jacob Rosenthal graduated with a Diploma in Mathematics from
the University of Wrzburg. Afterwards he studied philosophy at the
University of Konstanz and received a PhD in 2002 with a disserta-
tion on objective interpretations of probability (published as Wahr-
scheinlichkeiten als Tendenzen (2004)). Currently he teaches ethics
and philosophy of science at the University of Bonn. His research in-
terests lie in the fields of epistemology, general philosophy of science,
theory of action and moral philosophy.
Nico Scarano is Assistant Professor at the University of Tbingen.
He works on metaethics, moral philosophy and political philosophy.
His dissertation Moralische berzeugungen. Grundlinien einer anti-
realistischen Theorie der Moral has been published in 2001. He edited
the following books: (with Ch. Horn) Philosophie der Gerechtigkeit.
Texte von der Antike bis zur Gegenwart (2002), Modelle politischer
Philosophie (2003) and Ernst Tugendhats Ethik. Einwnde und Er-
widerungen (2006).
Dieter Schnecker is Professor of Moral Philosophy at the Univer-
sity of Siegen. His main research areas are Kants ethics, metaethics,
medical ethics, hermeneutics, epistemology. Publications: (with Gre-
gor Damschen) Selbst philosophieren (2006), (in cooperation with
Stefanie Buchenau and Desmond Hogan) Kants Begriff transzen-
dentaler und praktischer Freiheit. Eine entwicklungsgeschichtliche
Studie (2005),(with Allen W. Wood) Immanuel Kant: Grundlegung
zur Metaphysik der Sitten. Ein einfhrender Kommentar (22004),
Notes on Contributors xi

Kant: Grundlegung III. Die Deduktion des kategorischen Imperativs


(1999). Co-editor of: (with Heiner F. Klemme and Manfred Khn)
Moralische Motivation. Kant und die Alternativen (2006), Internati-
onales Jahrbuch des Deutschen Idealismus / International Yearbook
of German Idealism (20032005), (with Gregor Damschen) Der mo-
ralische Status menschlicher Embryonen. Argumente pro und contra
Spezies-, Kontinuums-, Identitts- und Potentialittsargument (2003),
(with Niko Strobach) Einfhrungen Philosophie, 15 vols. (2002 ff.),
(with Thomas Zwenger) Kant verstehen/Understanding Kant. ber
die Interpretation philosophischer Texte (22004) and Immanuel Kant:
Grundlegung zur Metaphysik der Sitten. Neu herausgegeben und ein-
geleitet von Bernd Kraft und Dieter Schnecker (1999). Numerous ar-
ticles and reviews in journals and dictionaries.

Klaus Steigleder is Professor of Philosophy at the University of Bo-


chum, Germany. His publications include: Kants Moralphilosophie
[Kants moral philosophy] (2002) and Grundlegung der normativen
Ethik [The Foundation of Normative Ethics] (1999). He is co-editor
of: Medizinethik. Geschichte, Theorie und Ethik der Medizin [Medi-
cal Ethics. History, Theory and Ethics of Medicine] (forthcoming),
Die Aktualitt der Philosophie Kants [The Timelessness of Kants
Philosophy] (2005), Bioethik. Eine Einfhrung [An Introduction into
Bioethics] (2003).

Mark Timmons is Professor of Philosophy at the University of Ari-


zona and author of Morality without Foundations (1999) and Moral
Theory: An Introduction (2002) and editor of Kants Metaphysics
of Morals: Interpretative Essays (2002). He and Terry Horgan ed-
ited Metaethics after Moore (2006) and they are currently working
on philosophical issues associated with the phenomenology of moral
experience.

Marcus Willaschek is Professor of Philosophy at the University of


Frankfurt/M. His publications include: Praktische Vernunft. Hand-
lungstheorie und Moralbegrndung bei Kant, Stuttgart/Weimar 1992;
Der mentale Zugang zur Welt. Realismus, Skeptizismus und Inten-
tionalitt, Frankfurt/M. 2003; Kant: Kritik der reinen Vernunft, Ber-
lin 1998 (ed. with Georg Mohr); Hilary Putnam und die Tradition des
Pragmatismus, Frankfurt/M. 2002 (ed. with M.-L. Raters); numerous
articles in journals and essay-collections on Kant, ethics philosophy
of action, metaphysics and epistemology, the philosophy of mind and
language.
xii Notes on Contributors

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


ford University. He has also taught at Cornell University and Yale
University, and has held visiting appointments at the University of
Michigan, University of California at San Diego, and Oxford Uni-
versity. He is author of nine books, most recently Kant (2005) and
is general editor (with Paul Guyer) of the Cambridge Edition of the
Writings of Immanuel Kant, as part of which Guyer and Wood col-
laborated on a new translation of the Critique of Pure Reason. Wood
has also translated Kants Groundwork (2002).
Groundwork
Preface
Nico Scarano

Necessity and Apriority in Kants Moral Philosophy


An Interpretation of the Groundworks Preface
(GMS, 387392)

Kants Preface to the Groundwork for the Metaphysics of Morals


can be divided into three parts. First of all, the text begins with a
topology of the philosophical disciplines, in which the project of a
metaphysics of morals is delimited from other forms of theory (GMS,
387,1388,14).1 Secondly, it contains a compactly reasoned appeal for
the indispensability of such a pure moral philosophy, that is, a theory
that proceeds completely a priori. Simultaneously, Kant emphasizes
the uniqueness of this project (GMS, 388,15391,14). And thirdly, the
Preface offers a prospective of the task, the structure, and the method
of the Groundwork, which does not itself already contain such a met-
aphysics of morals, but, as Kant stresses, represents a preliminary
work, laying the foundation upon which the comprehensive theory
will be developed in the future (GMS, 391,16392).
In the following I will go into all three parts of the Preface. The
focus of my analysis, however, concerns Kants central argument for
the indispensability of a pure moral philosophy. At decisive points,
Kant clearly inserts modal considerations. For the reconstruction of
the argument, it is therefore important to define precisely which form
of necessity is intended. Kant uses the terms necessary and necessity
in quite disparate contexts and with different meanings. He nowhere
explicitly says what necessity in the domain of morality can even mean.
For that reason, before I move onto the analysis of Kants argument,
I will address more closely the logical structure of ethical principles.
One must first understand the logical form that such a principle exhib-
its in order to be able to work out the basis of Kants central argument

1
This paper was translated by Aaron Looney.
4 Nico Scarano

for a pure moral philosophy. It will become evident that Kants argu-
ment is comprehensible if the decisive concept necessity is understood
as a modal operator in the sense of modern logic. The modal status,
necessity, allows moral principles to guide our counterfactual, practi-
cal reflections. This aspect of logical form can be clearly distinguished
from epistemological connotations, on the one hand, and the prescrip-
tive or imperative character of normative propositions, on the other.

1. What is a Metaphysics of Morals? (GMS, 387,1388,14)

Kant does not introduce the Preface of the Groundwork with a charac-
terization of the works content; rather he attempts first of all to define
the place of a metaphysics of morals within philosophy. For this task,
he makes use of three criteria. Kant first differentiates philosophical
theories by whether they are formal or material. Formal philoso-
phy, according to Kant, is equated with logic. It possesses no specific
object; rather it concerns itself, without distinction among objects,
with the universal rules of thinking in general (GMS, 387,10 f.). In
contrast, every material philosophy has to do with determinate ob-
jects and the laws to which they are subjected (GMS, 387,12 f.).
This formulation already offers an indication of the second crite-
rion. Kant subdivides material theories, in turn, into two classes. Kant
distinguishes them by with reference to the laws to which the objects
that the theories deal are subjected. He seems to assume that there
are exactly two kinds of laws. And, correspondingly, he differentiates
between two types of material philosophy: on the one hand, phys-
ics, or doctrine of nature, or, alternatively, natural wisdom; and,
on the other hand, ethics, or doctrine of morals, or, alternatively,
moral wisdom. It is a matter of the laws of nature, in the one case,
and of the laws of freedom, in the other, that each theory is respec-
tively concerned (GMS, 387,14 f.). What can Kant mean by this? The
expression laws of nature seems to be relatively unproblematic. But
what is to be understood by the expression laws of freedom?
From Kants elucidation one can infer a more exact interpretation.
Laws of nature are therefore laws in accordance with which every-
thing happens, while the laws of freedom are those in accordance
with which everything ought to happen (GMS, 387,25388,1).2 Ob-
viously, one can draw the parallel here to the modern terminologi-

2
See also the parallel passage in the Critique of Pure Reason, A 802/B 831.
Necessity and Apriority in Kants Moral Philosophy 5

cal distinction between descriptive and normative statements. At the


center of a philosophical doctrine of nature, there would then stand
general descriptive judgments; that means, more precisely, statements
of law that describe how objects of nature act. And at the center of a
doctrine of morals there would stand normative judgments that exhib-
it a comparable degree of generality and modal status. I will later ad-
dress more thoroughly what it could mean that some normative judg-
ments exhibit a comparable modal status to statements of law in the
natural sciences. Such a parallel is far from trivial. On the contrary,
it is one of the critical challenges for the interpretation of Kants ethi-
cal writings. 3 The first step toward a unified interpretation of the two
kinds of laws consists in seeing both kinds equally as propositionally
structured entities that demonstrate a clearly identifiable logical form
(see 2.1 through 2.3 below).4
Next to formality versus materiality and the two types of laws, Kant
introduces a third distinctive characteristic of theories. He seems to
understand material theories as complex systems that can be split into
individual parts. Every material philosophical theory accordingly
contains one part that he characterizes as pure, rational, or as
a priori (GMS, 388,414), and it may contain a second, empirical
part. Kant presupposes that in both the doctrine of nature and the
doctrine of morals the two parts can be isolated from each other. It
follows, therefore, that there are two a priori types of theory: on the
one hand, a metaphysics of nature, and, on the other, the sought-
after metaphysics of morals (GMS, 388,10).
Thus, a metaphysics of morals would be the following type of the-
ory: (a) It is not purely formal, but rather deals with definite objects.
(b) It deals with objects in so far as these are subsumed under laws
of freedom. (c) It contains no empirical elements. At this juncture
Kant does not yet presuppose that such a theory actually exists or that
it would be possible for a philosophy to work out convincingly such
a theory. He has not even argued yet that it is advantageous or even
necessary for philosophy to treat the a priori part separately from the
empirical part. Kant pursues these lines of questions in the subse-
quent passages.

3
Thus Tugendhat (1993, 99 f.), for example, doubts whether the use of the term ne-
cessity in Kants theoretical philosophy and in his practical writings can be under-
stood in a unified manner.
4
For the concept of law in Kants ethics see also Klotz (2001). The concept of neces-
sity as a modal quality that stands at the center of this essay is mentioned by him
without, however, further explication (58).
6 Nico Scarano

2. The Indispensability of a Pure Moral Philosophy


(GMS, 388,15391,15)

The Kantian appeal for a pure moral philosophy begins with a refer-
ence to the usefulness of the division of labor. Kant raises the ques-
tion whether philosophy would not benefit from such a division. Ul-
timately, he is not concerned with whether the burden of elaborating
philosophical theories should be distributed among the shoulders
of a multitude of persons. Rather with this rhetorical consideration,
he introduces an explicitly methodological train of thought. He asks
whether the nature of the science does not require the empirical part
always to be carefully separated from the rational, (GMS, 388,33 f.)
and he confines this question to the sub-question whether one is not
of the opinion that it is of the utmost necessity to work out once a pure
moral philosophy which is fully cleansed of everything that might be
in any way empirical and belong to anthropology (GMS, 389,69).
The passage (GMS, 389,523) that is introduced by this question
contains, at the same time, a positive response the thesis that there
must be a pure moral philosophy is self evident. And indeed for
Kant this evidence arises out of the common idea of duty and the
moral law (ibid.). Kants cursory argumentation for this thesis is dif-
ficult to grasp. Undoubtedly, it constitutes one of the key argumenta-
tive passages of the Groundwork. How can the argument be recon-
structed?
The point of departure for Kants line of argumentation lies in the
common idea of duty and the moral law. Kant seems to assume that
we (everybody: GMS, 389,11) may convince ourselves of the cor-
rectness of his thesis by reflecting upon what we understand by moral-
ity. What, however, is contained in our concept of morality? And with
which elements of the concept is he concerned in his argumentation?
In this context he makes use of an example. The moral command you
ought not to lie refers to a law that, according to Kant, is valid not
merely for human beings but also for other rational beings. That
moral laws have validity for all rational beings, therefore, is the first
step of argumentation towards a grounding of the thesis that there
must be a pure moral philosophy.
But what does Kant mean by his statement that a moral law is valid
for all rational beings? And how can the argument for it be recon-
structed, that is, be made as comprehensible as possible in the terms
of todays common philosophical concepts? Kant speaks in this con-
text of an absolute necessity of these laws. From this necessity, it
Necessity and Apriority in Kants Moral Philosophy 7

follows that these laws also must be valid for other rational beings.
And, in a further step, Kant concludes the apriority of moral philoso-
phy from this necessity. The decisive term for Kants argumentation,
necessity, especially requires interpretation. In order to obtain more
exact concepts of interpretation, I would like to examine more closely
in the following section the logical structure of ethical principles.

2.1. The Logical Structure of Ethical Principles

At the center of every ethics are normative statements. Ethics of prin-


ciple, like Kants, presuppose that there are ethical principles that
can be formulated with the aid of normative or evaluative statements,
which have a high degree of generality and therefore allow the moral
judgment of a plurality of cases. Such principles have the goal of guid-
ing our moral judgments and therewith our actions. They formulate
criteria by which we can morally judge objects, for example actions or
institutions.
Ethical principles can have different logical strengths. They offer
either a necessary or a sufficient or a both necessary and sufficient cri-
terion for determining when an object receives a corresponding moral
predicate. The categorical imperative, as defined by Kant, or, more
precisely, the founding law of the categorical imperative, appears to
be a quite demanding principle, one which prescribes both a necessary
and a sufficient condition for morality. Kant compares this principle
with a compass in the hand by which one knows [his] way around
very well in all the cases that come before [him], how to distinguish
what is good, what is evil, what conforms to duty or is contrary to
duty (GMS, 404,13).
Normative propositions in which principles with such a logical
rigor come to expression exhibit, by means of a first approach, the
following logical form:
(P1) For all objects x:
if, and only if, x satisfies the criterion C, does x have the moral
quality M.

Such propositions contain criteria for determining when an object re-


ceives a certain moral predicate or when it is denied it. In P1 it is a
matter of the conditions for the assignation of the moral predicate M.
According to the all-quantified biconditional, every object that satis-
fies the condition C simultaneously receives the moral predicate, and
vice versa. General propositions that exhibit such a form consequently
8 Nico Scarano

assist us in morally evaluating individual cases. If such principles are


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

Without going into the difficult questions of interpretation here, the


criterion CI can be articulated with slightly greater precision. Accord-
ing to Kant, the quality of the maxim from which an action results is
decisive for the moral valuation of an action. For Kant, a concrete
maxim for action appears to correspond to the demands of the cat-
egorical imperative only if the actor can will at the same time that
her maxims become a general law. Regardless of the familiar diffi-
culties in applying this criterion and regardless of the question about
the equivalence of the three formulas depicted by Kant, the follow-
ing formulation could be a preliminary rendering of the fundamental
principle of his ethics:
(P3) For all maxims x and all actions y:
if, and only if, the underlying maxim x of the action y has the qual-
ity that the actor of y can will at the same time that x becomes a
general law, does y satisfy the demand of morality.

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

5
I interpret the categorical imperative as a test not only for the legality but also for the
morality of an action. Such an interpretation is not uncontested in the contemporary
literature on Kant (s., for example, Khl, 1990, 66 f. and Wood in this volume); it
does not, however, play a decisive role in my further reflections. The term moral-
ity in (P2) would, accordingly, simply have to be replaced by the term legality.
Necessity and Apriority in Kants Moral Philosophy 9

pect of the logical form refers to the necessity of the law is not, as yet,
explicitly worked out.

2.2 The Modal Status of an Ethical Principle

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

Applied to Kantian ethics, this thought results in the following formu-


lation of Kants fundamental principle:
(P2') Necessarily, for all actions x:
if, and only if, x satisfies the criterion CI, does x satisfy the demand
of morality.
10 Nico Scarano

The proposition (P3) is also to be completed accordingly:


(P3') Necessarily, for all maxims x and all actions y:
if, and only if, the underlying maxim x of the action y has the qual-
ity that the actor of y can will at the same time that x becomes a
general law, does y satisfy the demand of morality.

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

6
Actually, the type of necessity spoken of here has to be further specified. Is it a mat-
ter of logical, conceptual, nomological, or metaphysical possible worlds? In
Scarano (2001, chapter 3.2), I argued that our moral principles have a comparable
status to the metaphysical necessity introduced by Saul Kripke (1980). To Kant
has to be ascribed the view that it is herewith a matter of conceptual necessity.
I see an indication of this interpretation in the method he applies in the fi rst and
second sections. He presupposes that the content or the formula of the categorical
imperative can be found solely through the means of the conceptual analysis of our
Necessity and Apriority in Kants Moral Philosophy 11

2.3. Necessity, Normativity, and Apriority

A possible but easily avoidable equivocation in the expression necessi-


ty can be cleared up at this juncture. Sometimes the expression is used
in the realm of morality as a synonym of normativity or prescriptivity.
Consequently, actions are necessary if they connote a should or if it
is our duty to carry them out. This type of usage is also found in Kant.
In the central third proposition of the first part of the Groundwork,
this usage is clearly expressed: Duty is the necessity of an action
from respect for the law (GMS, 400,18 f.). While the necessity ana-
lyzed previously refers to the moral principles, it is here a matter of
the necessity of the action itself. However this aspect is terminologi-
cally classified, whether as normative, prescriptive, evaluative,
or whether one speaks of the imperative character of moral judgments,
it may be distinctly distinguished from the modal-logical concept of
necessity responsible for the counterfactual variations.
In the propositions (P1') through (P3') this aspect that is, neces-
sity in the sense of normative, prescriptive, or evaluative is
indeed contained. There it is connected, however, to the moral predi-
cate, not the operator of necessity. When I speak of necessity in the
following sections, I mean a modal quality of judgments and not the
specificum of normativity.
Kant uses the expression necessity with yet other meanings. Every
interpretation depends on the clarification in each particular context
of what Kant exactly intends in those corresponding places and of
how each particular argument is to be reconstructed. Next to (a) the
type of usage as a modal operator that makes counterfactual consid-
erations possible and (b) the usage in the sense of an imperative char-
acter, thus in the sense of normativity or prescriptivity, there is (c)

moral concepts. At the beginning of the decisive argumentation in the second sec-
tion, he writes, Regarding this problem we will first try to see whether perhaps the
mere concept of a categorical imperative does not also provide us with its formula
(GMS, 420,1820). And approximately twenty pages later, he asserts in retrospect:
Yet that the specified principle of autonomy is the sole principle of morals may
well be established through the mere analysis of the concepts of morality (GMS,
440,2830). He, therefore, assumes that he actually was able to extract the formula
of the supreme moral principle solely through a conceptual analysis. In my opinion,
such a proceeding allows only one conclusion: If the moral principle can be pro-
duced solely through an analytical procedure on our concept of morality, then it
would have the status of conceptual necessity. According to Kant, the founding law
of the categorical imperative is valid in all conceptually possible worlds. The ques-
tion of which type of necessity moral principles exhibit, however, is not essential for
the ensuing reflections.
12 Nico Scarano

an often encountered usage with an epistemological meaning. If the


expression is used in this sense, then it means as much as necessary
know-ability, that is, the independence of knowledge from contin-
gent, empirical factors. Typically, Kant uses the expression a priori
for judgments that exhibit this characteristic. Since this type of usage
also can be clearly distinguished from the modal one, I will speak of
apriority to designate this epistemological aspect. I will use necessity
solely in the first sense (a).
Kant doubtlessly sees a close connection between necessity and
apriority. He often moves quickly from the one concept to the other
without grounding the transition. The two concepts, however, origi-
nate from varying spheres. While necessity is a matter of the modal
status of judgments, apriority is an epistemological concept. In the
former case, the concern is the application of predicates to objects of
other possible worlds. In the later case, it is a matter of the know-abil-
ity of the relevant judgments. Between the two concepts there does
not seem to be a close conceptual connection. In particular cases it
must be explicitly argued for that apriority follows from necessity.7
Even if Kant sees a very close connection between the two con-
cepts, he does not appear to assume that necessity and apriority are
exchangeable concepts. In the central passage of the Preface, he for-
mulates rather an argument for their connection. In the following sec-
tion, I will sketch out a possible reconstruction of the argumentations
structure on the basis of the conceptual differences just worked out.

2.4. The Central Argument in the Preface (GMS, 389,535)

In the central passage of the Preface on the indispensability of a pure


moral philosophy, Kant does not simply presuppose that from the ab-
solute necessity (GMS, 389,13) of moral laws the apriority of moral
philosophy follows. Rather he delivers a detailed, even if difficult, line
of argumentation. First of all, he announces his aim of argumentation
and designates his point of departure: that it is of the utmost neces-
sity to work out once a pure moral philosophy or that there must be
such is self-evident from the common idea of duty and of moral laws
(GMS, 389,711). The aim of his argumentation, therefore, is the proof
of the apriority of moral philosophy. And his point of departure for
the argumentation lies in the analysis of our concept of duty or of our

7
Compare this point with Kripke (1980).
Necessity and Apriority in Kants Moral Philosophy 13

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

al beings have the possibility of acting. In such a thought experiment,


we imagine a world in which there are other beings capable of action.
Since the moral laws accepted by us first of all contain the modal sta-
tus of necessity that is, they apply to all possible worlds and sec-
ondly, involve a proposition about all actions, it is to be concluded that
even those actions of these merely imagined rational beings fall under
these laws. The second intermediate thesis, therefore, states that (2)
the duties arising from the moral laws are also relevant for the action
of these merely imagined rational beings. In Kants terms: Everyone
must admit that [] the command You ought not to lie is valid not
merely for human beings, as though other rational beings did not have
to heed it (GMS, 389,1115).
At this juncture the decisive transition from necessity to apriority
takes place. Since the applicability of obligation is valid for all imagi-
nable rational beings, the cognition of these obligations or the laws
that are responsible for them cannot depend on contingent features
that arise from the fact that the thought experiment involves different
rational beings. The laws, therefore, may solely depend on character-
istics that all rational beings capable of action have in common. What
all imaginable rational beings have in common, however, is solely the
characteristic of being rational beings. Every other characteristic is
contingent. The cognition of laws cannot depend, therefore, upon the
empirical cognition of these contingent features. Kant seems to as-
sume that a cognition contains the characteristic of apriority if it is
not dependent upon empirical perceptions. Consequently, according
to the next intermediate thesis (3), the cognition of moral laws must
be a priori. Kant himself expresses this decisive intermediate conclu-
sion in the following terms: Among the things to which everyone must
admit is the fact that the ground of an obligation, that is, the moral
law, is to be sought not in the nature of the human being or the cir-
cumstances of the world in which he is placed, but a priori solely in
concepts of pure reason (GMS, 389,1619).
At the beginning of the ensuing paragraph, Kant reiterates this
intermediate conclusion: Thus [] are moral laws together with their
principles essentially distinguished [] from everything else in which
there is anything empirical (GMS, 389,2426). And he continues
with his last step of argumentation, which leads him to the concluding
thesis (4) that there has to be a pure moral philosophy. This step is
constituted simply by the consequences drawn from the epistemologi-
cal status of moral laws for the construction of philosophical theories.
If moral laws can only be known a priori, then moral theory must have
Necessity and Apriority in Kants Moral Philosophy 15

a foundational part that proceeds purely a priori, that is, without ref-
erence to anything empirical. In Kants words, all moral philosophy
rests entirely on its pure part, and when applied to the human being it
borrows not the least bit from knowledge about him (anthropology)
(GMS, 389,2629).
This is one of the possible approaches for a systematic reconstruc-
tion of the central argument in the Preface. Unfortunately, Kant of-
fers very few suggestions of how he can assume that the thesis that
there has to be a pure moral philosophy is self-evident. Consequently,
much interpretation is required. And much of the interpretive work
revolves around which systematic concept of necessity is implemented
in the interpretation. The interpretation presented here bases itself
upon a purely modal-logical concept of necessity, which entails nei-
ther an epistemological meaning nor an equivalence of normativity
or prescriptivity. If one chooses other systematic basic concepts, the
structure of argumentation also has to be interpreted in other ways. 8

2.5. The Practical Use of a Pure Moral Philosophy


(GMS, 389,36390,18)

Although the aim of the argumentation appears to be complete, Kant,


in the ensuing paragraph, suggests another form of reasoning for the
necessity of a pure moral philosophy. This second argument, how-
ever, takes a completely different direction. Kants original question
concerning the necessity of a moral philosophy is not free of ambigui-
ty. For him, it was at first a matter of a necessity that referred to cogni-
tion (speculation: GMS, 389,37) as its goal. It could be summarized
as follows: If one wants to obtain knowledge of moral principles, then
a pure moral philosophy is an indispensable means.
In Kants second argument, a pure moral philosophy is indispen-
sably necessary (GMS, 389,36) not only for the correct cognition but

8
A somewhat different interpretation is suggested by Schnecker/Wood (22004, 22
29). They see in the necessity of moral laws primarily their categorical character
(Kategorizitt). Similarly, Timmermann (2004, 86) believes that it is decisive for
the Kantian argument that moral prohibitions do not allow any exceptions. And
Bittner (1989, 2428) sees the starting point of Kants argument in the inviola-
bility (Unverbrchlichkeit) of moral laws. All of these approaches can support
themselves on comments made by Kant. In their approaches, though, Kants presup-
posed connection between necessity in the theoretical sphere and necessity in the
practical sphere is easily lost. In particular, with such a starting point, it remains
unclear what role Kants counterfactual considerations about other rational beings
have for the argument.
16 Nico Scarano

also for the correct action. It could be said, If one wants to act cor-
rectly, then one needs a pure moral philosophy. Kant argues for this
thesis: because morals themselves remain subject to all sorts of cor-
ruption as long as that guiding thread and supreme norm of their cor-
rect judgment is lacking (GMS, 390,2 f.). The implication contained
in this argument seems, at first sight, quite strong, but it is not clear
whether Kant actually advocates it so strongly. Do we really need a
philosophical moral theory in order to act correctly? Are we com-
pelled to rely on a kind of philosophical expertise when dealing with
moral questions?
Such an implication seems hardly congruent with the rest of Kants
remarks. There remains nonetheless the possibility of a weaker in-
terpretation. In this context, the expression moral philosophy does
not entail only the explicit theory, whose formulation the scientific
discipline of philosophy has as one of its tasks. Rather every human
or rational being possesses a moral philosophy if she has the capac-
ity to act from principle (for the sake of: GMS, 390,5). Accordingly,
every being capable of action carries at least implicitly a moral phi-
losophy in itself.
Therefore, fleshed-out scientific theories can help us make explicit
the implicit philosophical knowledge that is present in every person
and therewith isolate the a priori part from the empirical part. Such
an explicit knowledge can serve as a reliable guiding thread for ac-
tion. According to Kant, one can speak of a purity of morals (GMS,
390,17) only if the normative principles with which we ground our ac-
tions are pure and contain no empirical elements.
Thus, the scientific discipline of philosophy has a supporting but
not a constitutive function for the morality of our conduct. A meta-
physics of morals, as an explicitly fleshed-out philosophical theory,
can, however, offer a supporting contribution in making the idea of
a pure practical reason, which, according to Kant, lies in each of us,
effective in concreto (GMS, 389,35).

2.6. A Methodological New Beginning (GMS, 390,19391,15)

Kant asserts that the process of working out such a metaphysics of


morals is an entirely new field [] to be entered on (GMS, 390,22 f.).
He therewith contrasts the type of theory he introduces with all other
approaches that had been previously worked out in the history of phi-
losophy. Kant is not humble in his claims. Not only does he claim that
the philosophy before him had not found the right moral theory. It is
Necessity and Apriority in Kants Moral Philosophy 17

his opinion that all previous approaches were on the wrong track from
their very starting points.
For an interpretation of this passage in the text, the fact that moral
principles, according to Kant, can serve not only as criterion for judg-
ment but also as genuine grounds of action is significant. For Kant,
actions are truly moral solely if not only the principles of valuation
but also the particular grounds of action for the acting persons are
free of all empirical content. Moral principles must be our grounds
of action. Kant intimates therewith one of his central theses that he
will more thoroughly develop in the course of the Groundwork. Kant
accuses Christian Wolff and with him the entire moral philosophical
tradition of not clearly recognizing this. In their moral-philosophies,
they did not investigate the idea and principles of a possible pure
will but began their investigations with the actions and conditions
of human volition in general, which (and this is where Kant sees the
methodological mistake) are for the most part drawn from psychol-
ogy (GMS, 390,3437). According to Kant, exclusively a metaphysics
of morals, free from all empirical content and based on the analysis
of a pure will, is an appropriate foundation for the working out of a
normative theory of morality.

3. The Method and the Relation of the Groundwork to the Meta-


physics of Morals (GMS, 391,16392)

At this point, Kant departs from his general reflections on a theory to


be provided someday (GMS, 391,16) and dedicates himself to the
work that follows. His remarks are extremely terse and without a total
interpretation of the Groundwork almost impossible to comprehend.
In particular, his comments on the method he uses are hardly under-
standable for a reader who is not already familiar with the work.
He assigns three transitions to the three sections of the work al-
though it is unclear whether the sections also contain the starting
points and termini named in the title or whether they ultimately
have only the transitions themselves as topics. Kants mention of an
initially analytically and subsequently synthetically proceeding
is also unclear and disputed among interpreters (GMS, 392,1922). A
possible interpretation consists in directly connecting the expressions
analytically and synthetically to the word method. This would
mean that Kant refers here to the two different methods of instruc-
tion (Lehrarten) mentioned in the Prolegomena, which appeared
18 Nico Scarano

in 1783. In the two-year older work, this process is explained: the


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

9
See, for example, Brandt (1988, 171174), Bittner (1989, 29 f.), Freudiger (1993, 61
70), Schnecker (1996, 1997) and Milz (1998).
Necessity and Apriority in Kants Moral Philosophy 19

(a) The first problem arises from the fact that Kant thinks two types
of a groundwork for a pure moral philosophy are possible: on the one
hand, a critique of pure practical reason, and on the other hand, the
present work. Although he claims there is really no other foundation
than a critique (GMS, 391,17 f.), he believes that one can do without
such a critique and lists three reasons for this (see GMS, 391,20392,2).
It remains nevertheless unclear how the relation of the Groundwork to
the intimated critique of practical of reason is to be defined.
(b) Equally unclear is the precise relation of the Groundwork to
the envisioned metaphysics of morals. Is the groundwork of a the-
ory itself to be considered a part of this theory or is what Kant calls
the preliminary work of laying the ground (GMS, 391,37) a part
separate from the actual metaphysics? Undoubtedly, the parts of the
theory developed in the Groundwork raise the claim of being inde-
pendent of empirical knowledge and therefore belong to a pure moral
philosophy. Thus, at least parts of the Groundwork have to be identi-
cal with parts of a metaphysics of morals.
(c) Making matters more difficult in the identification of the rela-
tion between the Groundwork, a metaphysics of morals, and the men-
tioned critique of pure practical reason is the fact that Kant later
published monographs with these or slightly modified titles. In any
case, it may not simply be assumed that the later produced works ac-
tually deal with the types of theories mentioned in the Preface to the
Groundwork.
How can the relations of these types of theory to each other be
more precisely defined? Is Kants expression metaphysics of morals
a unified concept or are there different meanings combined here in
opaque ways?
An approach toward the resolution of these difficulties offers
perhaps a more precise interpretation of what Kant understands by
a metaphysics of morals. In the opening passages of the Preface,
Kant had defined the metaphysics of morals with the help of three
criteria:10 fi rst of all, it is a theory that is not purely formal but refers
to particular objects. It refers, secondly, to objects insofar as these are
under laws of freedom. And thirdly, it is perfectly free of empirical
content. Therefore, a metaphysics of morals is a pure moral philoso-
phy. One question, however, remains unanswered. What actually is
a moral philosophy? Or articulated more precisely: what is the onto-
logical status of such a theory?

10
Compare with above 1.
20 Nico Scarano

Kants arguments for the practical use or the necessity of such a


philosophy offer a suggestion for answering this question. A moral
philosophy is not only an object for the scientific discipline of phi-
losophy. Rather, according to Kant, every rational being, even if at
times confused, carries in itself a moral philosophy.11 That raises the
question about what conception of theory can fulfill both of these
functions simultaneously.
For the clarification of this question I believe the conceptual dis-
tinction between a theory, understood as an abstract system of ordered
propositions,12 and the presentation of such a theory, for example in
the form of a philosophical text, can be of further assistance. A moral
theory, understood as a system of propositions that stand in a particu-
lar relation of grounding to each other and whose center is made up
of ethical principles, is an abstract object, which can simultaneously
fulfill both functions. On the one hand, such a system can be brought
to expression with the help of a philosophical text. On the other hand,
it is also possible to carry such a system in oneself, that is, to accept
the corresponding propositions.
If one takes such a conception of theory as a basis, the confusing
terminologies and exact relations of the types of theories mentioned
by Kant could be newly formulated and made transparent in the fol-
lowing way: (i) A metaphysics of morals or a pure moral philosophy
is an abstract object, a system of propositions complete in itself and
containing all a priori judgments that are concerned with the laws of
freedom. (ii) A work with the title Metaphysics of Morals contains
a presentation of this system. The text expresses the corresponding
propositions and their connections. (iii) If a person carries such a mor-
al philosophy in herself, it means that she accepts the corresponding
propositions and their connections. And if this person is rational, the
propositions accepted by her will comprise the grounds of her action.
(iv) A work with the title Critique of Pure Practical Reason could
be a philosophical presentation of parts of this system of propositions
(and for this reason is suitable as a kind of groundwork). But it con-
tains additionally a presentation that relies on philosophical theories
on other spheres of investigation. In Kants words, in part I require
for a critique of a pure practical reason that if it is to be completed,
its unity with the speculative in a common principle must at the same
time be exhibited (GMS, 391,2427). (iv) In contrast, a work with the

11
Compare with above 2.5.
12
Compare, for example, with Nida-Rmelin (2002, 36 f.).
Necessity and Apriority in Kants Moral Philosophy 21

title Groundwork for the Metaphysics of Morals can restrict itself


entirely to the presentation and mediation of the foundational part of
the metaphysics of morals. This contains, as Kant expresses it, noth-
ing more than the search for and the establishment of the supreme
principle of morality (GMS, 392,3 f.).

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesammelte


Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902). Kants Groundwork
for the Metaphysics of Morals (GMS) is quoted according to the edition and
translation by Allen Wood (New Haven, CT: Yale University Press, 2002).

Other works

Bittner, Rdiger (1989): Das Unternehmen einer Grundlegung zur Meta-


physik der Sitten. Zur Vorrede der Grundlegung, in: Otfried Hffe (ed.):
Grundlegung zur Metaphysik der Sitten. Ein kooperativer Kommentar,
Frankfurt/M., 1330.
Brandt, Reinhard (1988): Der Zirkel im dritten Abschnitt von Kants Grund-
legung zur Metaphysik der Sitten, in: Hariolf Oberer/Gerhard Seel (ed.):
Kant. Analysen Probleme Kritik, Wrzburg, 169191.
Freudiger, Jrg (1993): Kants Begrndung der praktischen Philosophie. Sys-
tematische Stellung, Methode und Argumentationsstruktur der Grundle-
gung zur Metaphysik der Sitten, Bern/Stuttgart/Wien.
Klotz, Christian (2001): Gesetzesbegriffe in Kants Ethik, in: Volker Ger-
hardt/Rolf-Peter Horstmann/Ralph Schumacher (ed.): Kant und die Berli-
ner Aufklrung. Akten des IX. Internationalen Kant-Kongresses, Berlin/
New York, Vol. 3, 5562.
Khl, Harald (1990): Kants Gesinnungsethik, Berlin/New York.
Kripke, Saul A. (1980): Naming and Necessity, Oxford.
Milz, Bernhard (1998): Zur Analytizitt und Synthetizitt der Grundle-
gung, in: Kant-Studien 89, 188204.
Nida-Rmelin, Julian (1994): Begrndung in der Ethik, in: Nida-Rmelin:
Ethische Essays, Frankfurt/M. 2002, 3247.
Scarano, Nico (2001): Moralische berzeugungen. Grundlinien einer antire-
alistischen Theorie der Moral, Paderborn.
Schnecker, Dieter (1996): Zur Analytizitt der Grundlegung, in: Kant-
Studien 87, 348354.
Schnecker, Dieter (1997): Die Methode der Grundlegung und der ber-
gang von der gemeinen sittlichen zur philosophischen Vernunfterkennt-
nis, in: Hariolf Oberer (ed.): Kant. Analysen Probleme Kritik. Vol.
III, Wrzburg, 8198.
22 Nico Scarano

Schnecker, Dieter / Wood, Allen W. (22004): Kants Grundlegung zur Me-


taphysik der Sitten. Ein einfhrender Kommentar, Paderborn u. a.
Timmermann, Jens (2004): Kommentar, in: Timmermann (ed.): Immanuel
Kant: Grundlegung zur Metaphysik der Sitten, Gttingen, 84152.
Tugendhat, Ernst (1993): Vorlesungen ber Ethik, Frankfurt/M.
Groundwork
I
Allen Wood

The Good Without Limitation (GMS, 393394)

The opening pages of Kants Groundwork for the Metaphysics of


Morals are some of the most discussed pages in the history of ethics.
But they are also some of the easiest pages in all philosophy to mis-
understand, and I think also some of the most commonly misunder-
stood. Kants aim in these pages is to win support for a revolutionary
way of looking at the principle of morality by appealing to a set of
judgments about the special goodness of the good will that he thinks
belong to common rational moral cognition. But some of these judg-
ments are not only much more controversial, but also much harder to
understand correctly than he takes them to be. The result is that those
who approach these pages with theoretical preconceptions opposed
to Kants are not only likely to reject what he is saying, but also likely
to misunderstand it in ways that rationalize that rejection. Even those
who sympathize with Kant are very apt to misread him, by projecting
on the text a set of preconceptions about the focus of the discussion
that are at odds with Kants actual aims.
In the Groundwork Kants argument from these claims about the
good will to his conclusions about the foundations of ethics are also
slow in developing, and it is difficult for readers to see exactly what
propositions about the good will are crucial to his argument. By the
time he begins to reach his fi rst conclusions late in the First Section,
readers are very likely to have been distracted from the overall argu-
ment by controversies surrounding his initial claims, some of which
may also involve serious misunderstandings of those claims. Also,
Kant makes a number of distinct claims about the good will and its
special goodness which he is not careful to distinguish from one an-
other, and which make it harder to see exactly what he is claiming and
how it is supposed to support his theoretical conclusions.
I say all these things to prepare the reader for an interpretation of
the text which will to a considerable extent go against the grain it
26 Allen Wood

will be at odds not only with what people have usually thought Kant is
saying but also with what I freely admit it is probably easiest to think
he is saying. My argument, however, will be that the common and
easy way of reading the text leads us into serious difficulties. It cannot
make good sense of what Kant is talking about or of what he is saying,
and it attributes to him a set of claims that are simply not there in the
Groundwork sometimes because Kant does not accept them, and
sometimes because they are claims about topics that Kant does not
intend to address at all in this passage. I ask only that the reader try
on the reading I will suggest, and see for herself or himself whether
my reading is not in the end more satisfactory than what seems at first
to be the easy and obvious reading.
This paper is going to focus chiefly on the claim made for the good
will in the opening sentence of the First Section: There is nothing it
is possible to think of anywhere in the world, or indeed anything at
all outside it, that can be held to be good without limitation, excepting
only a good will (GMS, 393). Further, the paper will focus less on
what Kant means by a good will than on what he means in claiming
that such a will is good without limitation. I want to distinguish this
claim both from other claims Kant makes about the goodness of the
good will in the opening pages of the Groundwork, and from some
claims that he is often thought to have made but is not making. And
finally, I want to try to say something about how the unlimited good-
ness of the good will is supposed to function in the overall argument
of the Groundwork.
Kant brings out the distinctiveness of the good will by comparing its
goodness with that of other goods, which he divides into two main cate-
gories: gifts of nature, which he further subdivides into talents of the
mind and qualities of temperament, and gifts of fortune, among
which he distinguishes the three main objects of human passions and
competitiveness (power, honor and wealth), and also two more general
and encompassing goods: health and that entire well-being and con-
tentment with ones condition, under the name of happiness. Kant
thinks the unlimited goodness of the good will makes it different from
any of these other goods. And he tries to persuade us of the unlimited
goodness of the good will both by comparing it with these other goods
and by considering how we regard both the good will and the bad will
in the context of complexes made up of good or bad volition and some
of these other goods (or of the bad things opposed to them).
Yet if we are to understand the claim that the good will is good
without limitation, we must see clearly that it is a claim neither about
The Good Without Limitation 27

its worth in comparison to other goods nor a claim about the relative
goodness of various complexes in which a good or a bad will might
figure. Something might be good without limitation even if it were a
rather minor good as compared with others, and something might also
be good without limitation even if many of the complexes in which it
figured were less good than complexes involving a bad will. Thus Ross
(1954, 10) is mistaken when he says that Kants claim here entails that
the good will must never unite with anything else to produce a bad
whole. The claim that the good will is good without limitation is rath-
er a claim about the constancy and unchangeability of the degree of
goodness possessed by the good will itself (specifically, its immunity
to having its goodness diminished) when it is combined with anything
else, and the uniqueness of the good will in this respect when com-
pared to any other good whatever.
Kant asserts, and expects his readers to agree, that for every other
kind of thing that is in general good, the goodness of its particular
instances varies with the circumstances in which those instances are
found, and the way they are combined with other things, especially
with a will that is not good. Thus talents of the mind, such as under-
standing, wit and the power of judgment, or qualities of temperament,
such as courage, resoluteness and persistence in an intention, although
they are without doubt in some respects good and to be wished for
[] can also become extremely evil and harmful, if the will that is to
make use of these gifts of nature, and whose peculiar constitution is
therefore called character, is not good (GMS, 393). Even happiness is
good, Kant claims, only when it is enjoyed by a person having a good
will, since such a will appears to constitute the indispensable condi-
tion even of the worthiness to be happy (GMS, 393).
In contrast to all other goods, however, every instance of the good
will is good, and equally good, under all conditions, in all combina-
tions, and whatever its effects or consequences. The good will is good
not through what it effects or accomplishes, not through its efficacy
for attaining any intended end, but only through its willing, i. e. good
in itself (GMS, 394). Even if through the peculiar disfavor of fate,
or through the meager endowment of a stepmotherly nature, this will
were entirely lacking in the resources to carry out its aim, if with its
greatest effort nothing of it were accomplished, and only the good will
were left over (to be sure, not a mere wish, but as the summoning up
of all the means insofar as they are in our control): then it would shine
like a jewel for itself, as something that has its full worth in itself
(GMS, 394). The good will of course aims at good results, and with
28 Allen Wood

good fortune, achieves them. But they form no part of its own worth,
and do not add the least bit to it. It would be only the setting, as it
were, to make it easier to handle in common traffic, or to draw the at-
tention of those who are still not sufficiently connoisseurs, but not to
recommend it to connoisseurs and determine its worth (GMS, 394).
Kants claim that the good will is good without limitation good
thus amounts to what we might call the nondiminishability thesis:
The goodness of the good will is not diminished by any circumstance
in which it is found, by any of its effects, or by any combination with
other things, however good or bad, in which it may be involved. Kant
also holds a parallel nonincreasability thesis to the effect that the
goodness of the good will cannot be increased by any of its circum-
stances or effects or by any combination with other goods. Kant some-
times states the claim that the good will is good without limitation by
saying that it is absolutely good (GMS, 394; 402). I think he means
this in the sense of absolute that he explicates in the Critique of
Pure Reason when he says that we apply a predicate to something ab-
solutely when we mean either that the predicate applies to the thing
in itself or internally (apart from any relation the thing may have
to other things) or else when it applies to the thing in all respects or
in every relation (KrV A324 f./B380 f.). The good will is absolutely
good, in this sense, because its goodness does not vary with its relation
to any other thing, and is therefore possessed entirely in itself or apart
from any relation that the good will may stand to other goods.
There is another thesis that Kant asserts in the same paragraph
that obviously has some close relation to these two theses. This is what
we can call the unconditionality thesis: The good will is uncondi-
tionally good, or equally good under any and all conditions, whereas
all other goods are good only conditionally. Gifts of nature are good
when used by a good will but bad when used by a bad will for its ends.
Gifts of fortune are bad when they lead (as through arrogance) to
a bad will, or when they are enjoyed by a being that lacks the good
will which is the indispensable condition for the worthiness to enjoy
them. I think we may view the nondiminishability and nonincreas-
ability theses as consequences of the unconditionality thesis, in the
following way: If we consider all possible instances of some kind of
thing, then that kind of thing is good without limitation only if all
of these possible instances are good. If some of them are bad, then
this limits the goodness of that kind of thing. Since the goodness of
all good things except a good will is conditioned, those instances of
other kinds of things in which the condition is not fulfilled are not
The Good Without Limitation 29

good, and they constitute a limitation on its goodness. By contrast,


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

posite of what B aims at is realized, and this is a very good result, per-
haps exactly the result that A, with the good will, would have willed to
come about. Which complex, considered as a whole, should we value
more: As good will, with its bad results, or Bs bad will, with its good
results? Nothing in the nonincreasability thesis, or the nondiminish-
ability thesis, implies any answer to this question.
In other words, the nondiminishability and nonincreasability the-
ses are theses are only about the goodness of the good will as a possi-
ble kind of good, and also theses only about the good will itself. There
is no reason to suppose that Kant would deny that the complex of a
good will plus the good results at which it aims is a better whole than
the combination of the same good will with bad results that might
come about despite the good wills fruitless attempts to prevent them.
Nor do the the nondiminishability and nonincreasability theses them-
selves make any claims about the goodness of the good will in relation
to the goodness of other goods. Something whose goodness can be
neither increased nor diminished might be a relatively minor good,
not nearly as good as many of the things whose goodness is increased
or diminished by their circumstances, or effects or by their combina-
tion with other things.
As a matter of fact, however, Kant does hold that the good will is
also a greater good than any of the goods that are subject to being in-
creased or diminished: considered for itself, without comparison, it is
to be estimated far higher than anything that could be brought about
by it in favor of any inclination, or indeed, if you prefer, of the sum
of all inclinations (GMS, 394). And this seems also to be something
that Kant sometimes intends to say when he claims that the good will
is absolutely good. Here, however, the claim that the good will is
absolutely good does not mean that its goodness is independent of its
relations to other things, but rather makes the quite distinct claim that
it is a greater good, in comparison to other things, than anything else.
This distinct claim, which we may call the higher worth thesis, does
entail that when a person with a good will achieves the happiness of
which she has made herself worthy, then it is the good will and not the
happiness that is the greater good.
The higher worth thesis may also imply though this seems to me
less certain that in our example of A and B above, that the complex
of As good will and its unfortunate bad results is to be preferred to
the complex of Bs bad will and its fortunate but unintended good
results. This remains uncertain because it might be that the goodness
of the unintended good results in the case of B, or the badness of the
The Good Without Limitation 31

unintended bad results in the case of A, or especially both together,


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

the formula of autonomy. He mentions the worth of the good will


again only at GMS, 437, after that development has been completed,
and merely in order to connect up the results of the Second Section
with those of the First Section.
In order to say how Kants claims about the unique value of the
good will are related to his argument, we need to say at least a little
more about what Kant means by the good will. This is not as easy as
one might think, since despite the emphasis he places on the goodness
of the good will, Kant never gives us a direct and clear account of what
the good will is. Instead, he immediately narrows his focus to a special
case of the good will, namely, the concept of acting from duty, which,
according to Kant, contains that of a good will, although under cer-
tain subjective limitations and hindrances, which, however, far from
concealing it and making it unrecognizable, rather elevate it by con-
trast and let it shine forth all the more brightly (GMS, 397). It is not
immediately clear how much farther Kant intends the concept of the
good will to extend, beyond the concept of acting from duty. Clearly
the divine will is thought of as a good will, but God utterly lacks the
subjective limitations that would be necessary for the concept of
duty to apply to him. Many readers of the Groundwork tend simply to
identify the good will, at least for humans, with the will that acts from
duty. But I think that is a mistake.
Kant famously, if cryptically, identifies the will with practical rea-
son (GMS, 412). Reason, however, is described in the Critique of Pure
Reason as the faculty of principles (KrV, A299/B356). This means
that the will is the faculty of practical principles that is, the faculty
determining the normative policies that guide a persons actions. Kant
distinguishes two kinds of principles: subjective principles or maxims,
and objective principles or laws (GMS, 401 f.; KpV, 19). Maxims are
principles adopted contingently by a particular subject to guide its ac-
tions: they are normatively valid only for the subject that adopts them,
and they remain valid only as long as this subject continues to accept
them. Laws are principles objectively valid for all rational beings. It is
an important Kantian thesis that rational beings are capable not only
of adopting maxims, valid subjectively for themselves, but are also
autonomous, capable of giving laws that are valid objectively for all
rational beings (GMS, 431). Will is the faculty that gives both kinds of
principles, and derives actions from them (GMS, 412).
Every rational will gives laws, so the distinction between a good
will and bad will can have nothing to do with this objective legislation.
It can have to do only with the subjective principles or maxims the will
The Good Without Limitation 33

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

merchant whose maxim, adopted from prudence, is to deal honestly


regardless of whether he can get away with it. The latter maxim makes
for a good will, because it accords with the moral law, while the former
maxim, by allowing for cheating when you can get away with it, does
not accord with the moral law.)
I think this latter interpretation, which allows that a person may
have a good will even when they do not act from duty, is clearly pref-
erable to the former interpretation. For it makes sense to think that
Kants examples in the Groundwork (of the merchant, the man pre-
serving his life, the beneficent man, the gout-sufferer who resists the
temptation to indulge in unhealthy food or drink) are examples in
which Kant is imagining two people (or the same person at differ-
ent times) following the same dutiful maxim, one time from prudence
or inclination, the other time from duty. Moreover, Kant says of be-
neficent or honorable actions done from inclination that they deserve
praise and encouragement (GMS, 398) something Kant should not
be expected to say if such maxims exhibit a will that is not good. Fur-
ther, Kant holds that we have a duty to cultivate our sympathetic feel-
ings (MdST, 456) and even our love for others as inclination (EaD,
337 f.) on the ground that these feelings and inclinations will motivate
us to do our duty. Yet surely this is inconsistent if Kant thinks any
maxim adopted or followed from inclination would deprive the will
of its goodness.
Finally, Kant describes duty as necessity of an action from respect
for the law (GMS, 400). From other passages, we know that Kant
understands this necessity, also called necessitation (Ntigung) as
a kind of constraint (Zwang), namely an inner constraint or self-
constraint through respect for the law (MdST, 379). In other words,
to act from duty is to constrain oneself to do an action that accords
with the moral law. However, it is obvious that such self-constraint is
not always needed to achieve compliance with the law. Prudence may
give the merchant a sufficient ground to deal honestly with custom-
ers, self-love may be sufficient to move us to preserve our lives or to
adopt a healthy diet, and sympathy may lead to beneficence. It also
makes sense to think that where self-constraint is not necessary, it
is also properly speaking not even possible. For what sense would it
make to say that one person who enjoys making others happy does so
freely, while another does it with self-constraint? Those who wish to
limit the good will to the will that acts from duty must apparently say
such things. Still worse, they must say that only the person who acts
with self-constraint has a good will, while the other equally beneficent
The Good Without Limitation 35

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

1
Many think that for Kant it is possible to act from duty even when one has an incli-
nation to do the act which is strong enough that no self-constraint is required. To
sustain such a reading of the text, however, one is forced to do either of two things:
On the one hand, one might attribute to Kant a highly mechanistic conception of
action and motivation, according to which, when one has two different incentives to
perform the same action, there must be a fact of the matter about which incentive
caused the action, much as, if a light bulb were hooked up to two different wires,
there would have to be a fact of the matter about which wire it was that delivered the
current to the lighted bulb. Or, on the other hand, one would have to interpret claims
in such cases about ones real motive as claims not about that particular action at
all, but rather about ones character or volitional dispositions, the structure of ones
volitional priorities in general to act from duty is to give duty general priority over
inclination in making ones decisions. But neither of these options is the least bit
attractive. The first requires us to ascribe to Kant, on the basis of no explicit textual
support whatever, a highly implausible and unappealing theory of action. The sec-
ond transforms what he evidently intends to be a claim about the moral worth of this
action into a general claim about the agent, or about what the agent would do under
counterfactual circumstances. (Would the agent still act in a dutiful way if the co-
operating inclination were removed?) But it also does not seem that there need be
any corresponding general truth about the agents character, or about what the agent
would do, for this action to have moral worth in the case Kant is imagining -- where
the agent must constrain himself through respect for the law if the dutiful action is
to be performed at all. In fact, we might think that an action would have more moral
worth in such a case if it were performed against the agents general dispositions,
and in conflict with his (generally bad) character. For that would give the agents
act of self-constraint an even more heroic quality. The more one reflects on the in-
conveniences and implausibilities we are required to digest in order to maintain the
usual interpretation of Kant here, the more attractive becomes the idea that in this
passage, all Kant means by acting from duty is acting with self-constraint through
respect for law something that need not, and therefore could not, occur in any case
where there were a co-operating inclination that made self-constraint unnecessary
(hence impossible).
2
Another temptation to think that a will can, in the sense relevant here, act from duty
even when it need not constrain itself to do so is drawn from Kants claim that only
the action done from duty has moral worth. It might seem to follow from this that
a dutiful action requiring no self-constraint (due to co-operating inclinations, such
as sympathy) is morally worthless. Kant has sometimes been accused of holding
such views, but it is easy to see how we could avoid that reading. For it is easy enough
to avoid it even while understanding acting from duty as acting with self-con-
straint out of respect for the law once we realize that an action which lacks moral
worth in the sense Kant means here need not be a morally worthless act, or an act
of no value to morality. Obviously, an act that conforms to duty, especially when
done with a good will (that is, from a maxim that conforms to duty) would have some
kind of value to morality (and deserve praise and encouragement as Kant says it
does), even if it is not done from duty. Notice that when speaking of moral worth in
36 Allen Wood

ciple of morality (the formula of universal law) by arguing that duty


is necessity of an action only from respect for the law, and hence that
the law to which this respect is directed can be limited by no condi-
tion arising from any end or inclination, so that conformity to this law
must be nothing but the conformity of actions as such with universal
law (GMS, 402). Later, to be sure, he also relates the way in which
the law is unlimited in its content by any end or inclination to the
unlimited goodness of the good will, when he says about respect for
the moral law estimation of a worth which far outweighs everything

these passages, Kant often appends an adverb to moral: he speaks of: true moral
worth (GMS, 398) genuine moral worth (GMS, 398) authentic moral worth
(GMS, 399), and then speaks of that so pre-eminent good we call moral (GMS,
401). I think his talk about moral worth is best understood as follows: Morality
cares about many things, but not all of them have true or genuine or properly moral
worth. Morality approves of the happiness of a good person, for instance. But hap-
piness, even deserved happiness, is not properly a moral good. Neither, thinks Kant,
is even an action conforming to duty, when this conformity serves self-interest or
inclination. It has properly moral worth and deserves, as Kant says, not merely
praise and encouragement but also esteem only when it would not have happened
but for the agents self-constraint out of respect for the moral law.
This point is closely related to a way in which Kants entire discussion in these pag-
es, and most especially his comparison of beneficent actions done from inclination
and from duty, is very commonly misunderstood. Kant describes a man who fi rst (a)
makes others happy because he enjoys doing so, and then, after his inclination to be-
neficence is eclipsed by his own sorrows, (b) tears himself out of deadly insensibility
and is beneficent not from inclination but from duty (GMS, 398 f.). There are different
questions one might ask in comparing the two cases (a) and (b). The question Kant
means to ask is: In which case does the action call forth from us the most properly
moral esteem for the action? This is not at all the same question as: Which situation
would we most want to be in? or even Which sort of beneficent action would we, as
moral educators, most try to encourage? There may be moral theories (and moral
psychologies) that deliver the same answer to all these questions: that say: the act
deserving the most properly moral esteem is the one we would most want to perform
and most want to encourage. To people who are committed to such theories (or who,
out of inattention, dont even see the difference between the questions), Kants insist-
ence that we should feel properly moral esteem for the act only in case (b), not in case
(a), is often taken to imply that he would also favor case (b) in response to the other
two questions which would, of course, be perverse. But there is actually no good
reason to think that Kant would not agree with us that we should prefer to do benefi-
cent actions from inclination rather than having to constrain ourselves to do them
from respect for the law, or that he would dissent from the thought that we should try
to educate people to be beneficent from inclination wherever possible, rather than
having to make themselves be beneficent out of respect for moral laws. His point (to
which I think many readers remain simply oblivious) is that despite this, it is the man
who must constrain himself to be beneficent whose action is most to be esteemed as
having a worth that is properly (truly, genuinely, authentically) moral. Perhaps even
when they understand the question, not all readers will agree with Kants answer.
But if we look carefully at most of the critics who scorn what he says in this passage
(and they are legion), I think we will find virtually none who has actually addressed
precisely the question Kant intends here.
The Good Without Limitation 37

whose worth is commended by inclination, and that the necessity of


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

who come away from the Groundwork with the impression that his
opening claim about the goodness of the good will is fundamental to
his entire ethical theory are profoundly mistaken.
Kant sometimes speaks (though only here in the Second Section,
never in the First Section) not merely of the good will but of the
absolutely good will. That will is absolutely good which cannot
be evil, hence whose maxim, if it is made into a universal law, can
never conflict with itself (GMS, 437; cf. 426, 439, 444, 447). This is
not merely a repetition of the claim that the good will is absolutely
good (either in the sense that it is good unconditionally, regardless of
any of its relations to other things, or in the sense that it has a higher
worth than any other possible good thing). Rather, it identifies a spe-
cial case of the good will, where not only does its maxim conform to
the moral law, but where the principle of this will is itself conformity
to moral laws. 3 This makes sense, however, only after we have identi-
fied the moral law, and can therefore specify the will whose goodness
is absolute in the sense that it conforms to that law. It must not be
thought that all along Kant had this absolutely good will in mind,
and that it was only of it that he said that it is good without limitation.
Rather, any will that adopts a good maxim, in whatever respect and
from whatever incentive, is, to that extent, a good will, and good with-
out limitation.4
3
In the Preface, Kant says: For as to what is to be morally good, it is not enough that
it conform to the moral law, but it also must happen for the sake of the law (GMS,
390). This is usually read as saying the same as that an action has moral worth only
when it is done from duty. But it does not say the same thing. For a holy will (to which
the very concept of duty, and of acting from duty, could not apply) presumably also
conforms itself to the moral law and acts for the sake of the law. It is better to under-
stand Kant to be saying here that properly moral good is to be found in a will which
performs dutiful actions on principle, rather than (as he puts it here) contingently
and precariously from co-operating inclinations. In other words, we come closer to
a worth that is properly moral when we are dealing with an absolutely good will, a
will whose basic principle is to conform its actions to the moral law, than a good will
which is merely acting on some principle or other that conforms to the law. Kant no
doubt would deny that it would be possible for a finite and imperfect human will to
do this always without sometimes having to act from duty (with self-constraint). But
in a person whose temperament has been well cultivated to dispose it to morality,
it seems that it might be possible to act quite often for the sake of the law without
having to constrain oneself through respect for the law. So I do not think that what
Kant is talking about here in the Preface precludes the possibility that a good will,
possessing a good that is properly moral, especially if it should be an absolutely good
will, might often act not only in conformity with the moral law, but also for the sake
of the law, without having to act from duty.
4
A will that acts on the principle that all its maxims should conform to the moral
law is, I think, what Kant means by a will that acts for the sake of the law. But, as
has been pointed out in the previous note, this is a will that need not act from duty
The Good Without Limitation 39

It may help here to realize that a will (as practical reason or the
capacity to act on principles) is an abstraction. A good will is not the
same as a good person. Every (finite, imperfect) persons will is always
in some respects good and in some respects bad, assuming that all of
us adopts some good maxims and some bad ones. So a bad person
may (sometimes, in some respects) have a good will, and a good per-
son may (sometimes, in some respects) have a bad will. For example,
the merchant who, out of prudence and with a self-serving aim, adopts
the maxim of dealing honestly with his customers, still has a good will
as far as that maxim is concerned. But if his maxims are generally
self-serving, then on the whole he would doubtless be a bad man, and
would even have (on the whole) a bad will. We must not think that
because he is a bad man on the whole, his will regarding this maxim is
a bad will, as long as his maxim does indeed conform to morality.
Moreover, a merchant who also had the maxim (or meta-maxim)
of adopting maxims in conformity with the moral law for the sake
of the law (GMS, 390) would have on the whole a better will than
this merchant, who adopts the dutiful maxim from prudential consid-
erations. Even this latter merchant, however, would still not be acting
from duty in cases where prudential considerations gave him sufficient
grounds for following the maxim of honesty, since no self-constraint
would be needed in those cases to get him to follow the maxim of
honest dealing. Thus we should not confuse acting from duty with
acting for the sake of the law. Nor should we think that anyone who
is honest, or beneficent on principle (as distinct from being honest out
of prudence or beneficent out of sympathetic inclination) is thereby
acting from duty. Such thoughts, which are admittedly all too easy to
entertain, misconstrue what Kant means by acting from duty and
contribute to many common confusions about what he is saying in the
opening pages of the First Section of the Groundwork.
A good will must also be distinguished from a persons good char-
acter or virtue, which is the strength of the persons character in act-

(this is especially obvious in the case of a holy will, but might be true in the case of a
finite human will wherever its conformity on principle to the moral law comes about
without the need for self-constraint through respect for the law). If that is right, then
an absolutely good will might exemplify the properly (truly, genuinely, authentically,
pre-eminently) moral good, yet without acting from duty. In the First Section, how-
ever, Kant never considered the case of the absolutely good will, so he took no posi-
tion on this. All the cases he considered had to do with the conformity of particular
actions to particular duties or dutiful maxims. In those cases, it was only through
exhibiting self-constraint through respect for the law (acting from duty) that the
properly moral good could be exemplified.
40 Allen Wood

ing on good maxims (MdST, 408 f.). Good character implies a certain
good will, which is why Kant speaks of the two together in Section
One of the Groundwork (GMS, 393; 398); but as he makes explicit
elsewhere, a good will can co-exist with lack of virtue, when a person
adopts good principles, but is too fragile or weak of character to carry
them out reliably (RGV, 29; GMS, 407 f.). Goodness of will might also
be combined with a serious lack of wisdom or judgment, so that the
actions performed may even be contrary to the good principles that
make the will good. In both these ways, a person with a good will can
be a person who often does what is morally wrong. When it forms
part of a syndrome of moral weakness or bad moral judgment, the
good will (the adoption of good principles) might be thought to lose
some of its goodness, and to be itself less good than it would be when
found in a person of greater virtue and moral wisdom. The more we
reflect on this point, the more reason we might find to doubt that Kant
is even correct in regarding the good will as good without limitation.
But even if we do doubt the unlimited goodness of the will on these
grounds, that doubt need not call into question the fundamental val-
ues on which Kantian ethics rests: the autonomy of reason, the univer-
sal self-legislation of the rational will, the dignity of rational nature as
an end in itself.
The fact that goodness of will is an abstraction that all of us
doubtless have a good will in some respects and a bad will in others
may account for the fact that in the early pages of the Groundwork
Kant chooses to narrow his focus from the good will to the special
case of acting from duty. For every instance of acting from duty dis-
plays a good will, even though the converse of this does not hold, and
acting from duty also displays a good will under adverse or trying cir-
cumstances (under certain subjective limitations and hindrances,
as Kant puts it) so that in acting from duty the good will shows itself
with a kind of special heroism not present in many cases of the good
will (the adverse circumstances elevate it by contrast and let it shine
forth all the more brightly.) For the same reason, Kant prefers the
person who acts from duty to the one who is dutiful from prudence
or inclination only in the sense that the person who acts from duty is
due greater esteem, not in the sense that we should prefer to be like
him in the sense of being in his (adverse and unfortunate) situation.
All too many readers of the Groundwork probably because they do
not distinguish the question Who is most morally admirable? from
the question Whom would I most want to be like? have badly mis-
read what Kant is saying here, as though he thought we should prefer
The Good Without Limitation 41

having interests or inclinations adverse to the moral law over having


interests and inclinations in harmony with it. There are, as I said at the
outset, many different ways in which it is easy to misread this famous
but treacherously brief discussion. It is no exaggeration to say that a
large part of Kants reputation as a moral philosopher is based on such
serious but widespread misunderstandings.
The (perhaps surprising) conclusion that the goodness of the good
will is not foundational for Kantian ethics, though in one sense true,
may nevertheless be exaggerated, at least if it is stated in this unquali-
fied form. For if we look beyond Kants explicit arguments I think it
is nevertheless possible to see some very suggestive connections be-
tween his striking claims about the good will and some of the main
theses for which he is arguing in the Groundwork. Perhaps the clear-
est case is the relation between Kants unconditionality thesis about
the good will and his claim that the supreme principle of morality is a
categorical imperative.
An imperative is an objective principle or law of reason, or more
precisely, a command of reason, that is, an objective principle or law
that is directed to a being whose will does not necessarily accord with
that law, but which may need to be rationally constrained to accord
with it (GMS, 413 f.). A holy will, such as the divine will, necessarily
acts according to laws or objective principles, and therefore no impera-
tives or rational commands apply to it. A categorical imperative is a
command of reason whose rational bindingness on the will is inde-
pendent of any end external to that imperative itself. A technical im-
perative that directs us to take the necessary means to an arbitrary end
is rationally binding only on those who have set that end. A prudential
imperative that directs our pursuit of happiness is rationally binding
on us only because we are assumed to have our happiness as an end.
These imperatives are therefore hypothetical rather than categorical
(GMS, 414). But a moral imperative directs us to perform actions that
are morally required, or at least morally meritorious, and it is ration-
ally binding on us irrespective of any ends we may have. Hence a moral
imperative is a categorical imperative. Of course, a categorical impera-
tive might also rationally direct us to set certain ends Kant argues
that it requires us to set as ends both our own perfection and the happi-
ness of others (MdST, 385388). But its rational bindingness is neither
conditional nor dependent on our setting those ends.
A categorical imperative therefore rationally directs us to do cer-
tain things without reference to any ends we may already have, and
therefore, its bindingness is not conditional on the consequences of
42 Allen Wood

our actions for any actual or possible ends given independently of the
imperative itself. More precisely, however, a categorical imperative ra-
tionally directs us not so much to do certain things, as to adopt and
act on certain principles (maxims), which may in turn include or entail
the performance of certain actions, or the setting of certain ends. The
basic effect of a categorical imperative, therefore, is to declare certain
maxims of action binding on us irrespective of their relation to any
independently given ends, or therefore, of their relation to any conse-
quences (good or bad) of acting on them.
The will, as we have seen above, is for Kant the rational faculty of
practical principles. That is, it is the faculty through which a rational be-
ing gives itself such principles, whether subjective principles (maxims)
or objective principles (laws). A will is good when its maxims accord
with the laws it gives itself. For a holy will, such as the divine will, there
are only objective principles, or else, what would come to the same
thing, its subjective principles or maxims are necessarily identical to
the objective principles it rationally legislates to itself. (This is why no
rational commands or imperatives apply to such a will.) For a finite and
imperfect will, however, goodness of will consists in the contingent and
voluntary accord between its maxims and its laws. In other words, as
we have already seen, what makes a good will good are its principles,
or, in the case of a finite and imperfect will, its maxims.
Now there are two possible ways in which maxims, hence wills,
might be considered good. They might be good (at least in part) be-
cause of the consequences of acting on them, hence good relative to
certain independently given ends. Or they might be good simply in
themselves, irrespective of any such consequences or their conducive-
ness to any independently given ends. In the former case, the good-
ness of the good will itself would necessarily be increased if its con-
sequences were good, and diminished if its consequences were bad
(relative to those ends).
According to Kants unconditionality thesis, however, and the non-
diminishability and nonincreasability theses that follow from it, the
goodness of the good will is unaffected by its consequences, whether
these be good or bad. Therefore, the maxims constituting the good-
ness of a will are not good only relative to certain consequences or in-
dependently given ends, but must be regarded as good in themselves,
irrespective of their consequences or the effect of these consequences
for any independently given ends. That means that the principle that
specifies which maxims are good must not do so by reference to any
independently given ends, but only by regarding certain maxims as
The Good Without Limitation 43

good (as rationally required or rationally meritorious) in themselves,


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

attention to the moral importance of willing on principle might be


enough for Kants purposes, even if we decide on reflection that it
contains more hyperbole or exaggeration than it does literal truth.

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesammelte


Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902) (abbreviated as
AA). The translation of the Groundwork will be my own, taken from Kant,
Groundwork for the Metaphysics of Morals (New Haven: Yale University
Press, 2002). The Critique of Pure Reason will be cited according to the A/B
pagination from the first and second editions.
GMS Grundlegung zur Metaphysik der Sitten, AA, IV
KpV Kritik der praktischen Vernunft, AA, V
KrV Kritik der reinen Vernunft, AA III, IV
MdST Metaphysik der Sitten, Tugendlehre, AA,VI
EaD Das Ende aller Dinge, AA, VIII
RGV Die Religion innerhalb der Grenzen der blossen Vernunft, AA, VI

Other works

W. D. Ross, Kants Ethical Theory: A commentary on the Grundlegung zur


Metaphysik der Sitten (Oxford: Clarendon Press, 1954),
Christoph Horn

Kant on Ends in Nature and in Human Agency


The Teleological Argument (GMS, 394396)

Many readers of the Groundwork feel some irritation about the fact
that Kant directly combines, in the First Section, his concept of a
good will with teleological reflections and even with quite strange
ones. Immediately after having introduced his starting thesis that
good will is the only good without restriction (GMS, 393,5394,31),
Kant turns to a discussion of the teleology of human nature (GMS,
394,32396,37). In this passage, he opposes a view of teleology which,
in his eyes, is thoroughly mistaken. Moreover, he develops what he
thinks to be the correct account of this topic. This is surprising since
Kant, at least in his critical period, is well known as a philosopher
who rejects traditional teleological claims, in particular the famous
teleological argument for Gods existence. Therefore, what we find in
our passage seems to be a sort of theoretical anachronism, a return
to bad metaphysics. And indeed there is, at least prima facie, little to
say in favor of Kants teleological considerations. Consequently, in the
vast literature on the GMS, the passage received relatively little atten-
tion and some bad press.1
The line of thought developed in the teleological passage can be
divided into four sections:
(a) Kant fi rst tells us why he introduces his considerations on tel-
eology anyway (GMS, 394,32395,3). The reason is that he wishes to
reject a serious suspicion which can be raised against his concept of
good will: it could be said to be a mere fantasy or a philosophical fig-
ment. What he wants to make plausible is that the aim of nature in
assigning reason to govern our will consists precisely in the moral-
ity of good will.

1
See e. g. Schnecker/Wood (2002, 524).
46 Christoph Horn

(b) After that, Kant provides a more extensive rejection of the


claim that happiness (Glckseligkeit) is the true end of human nature
(GMS, 395,427). He makes use of an argument which takes the form
of a modus tollens: The natural dispositions of each animal are opti-
mally purposive or suitable for the goals for which the animal is des-
tined. (Let us call this the principle of suitability.) Now, if happiness
were the goal for humans we would have to find, according to this prin-
ciple, some indications for the alleged destination within mens natural
endowment. In truth, however, there is no sufficient ground to accept
the idea that happiness is our natural goal. What we in fact find is that
we are deeply determined by our possession of practical reason. If the
natural goal of men consisted in the pursuit of happiness, then the pos-
session of reason would be a relatively dysfunctional means, since it
would have been more appropriate to endow humans with stronger
instincts. Therefore, given the principle of suitability, the fact that we
possess practical reason is not consistent with the idea of happiness as
our natural end. Hence, according to Kant, the assumption that happi-
ness is mens natural destination should be relinquished.
(c) In the next, smaller passage (395,28396,13), Kant offers us
some considerations which are intended to support the thesis that
practical reason is a dysfunctional instrument to gain happiness and
contentment of life (Zufriedenheit des Lebens). In Kants opinion, we
can find an evidence for it in the empirical fact that highly cultivated
persons who possess extraordinary mental capabilities tend to show
a certain misology, i. e. a hatred of reason. When these people retro-
spectively reflect on the gains and losses of having cultivated their in-
tellectual abilities within their biographies, they typically come to the
conclusion that this development didnt lead them to a larger amount
of happiness, but to an increase of hardship. So they feel even envy for
ordinary people and their non-intellectual, purely sensual way of life.
Kant affirms that there is some truth in their judgment, but not in the
way they put the problem: in fact, these people involuntarily validate
the thesis that our practical reason is not made to bring about hap-
piness. Hence, what Kant claims, is that their conviction is covertly
founded on the idea of another aim for their existence, possessing
much greater dignity.
(d) Kant hereafter develops his own, positive view of the teleol-
ogy of human nature (GMS, 396,1437). According to him, the true
vocation of practical reason is not to produce volition as a means to
some other aim, but rather to produce a will good in itself. The deci-
sive argument in Kants account is founded, as in section (b), on the
Kant on Ends in Nature and in Human Agency 47

principle of suitability: nature has endowed us with practical reason,


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

The line of argument advanced by Kant in the teleological passage


of GMS I seems to be extremely questionable. At least, we have to
reproach him for the very elliptic character and the unclear form of
his argument. But I think there are many other problems to be solved.
48 Christoph Horn

As an interpreter of section (a), one has to make sense of the fact that
Kant, as I already mentioned, surprisingly hopes to support his view on
good will by such an unclear argument anyway. Concerning the crucial
section (b), some important questions one would like to ask Kant may
be these: What precise type of teleology is it that he has in mind? How
can Kant hope to defend his idea of good will by a kind of doctrine
which must be, in his own eyes, obscure and illegitimate? Why does he
introduce such serious and far-reaching claims in such a superficial and
provisionary way? How would Kant legitimize what we called the prin-
ciple of suitability? Apparently, he does nothing at all here to make it
plausible. Further questions are: Is it really convincing to say that the
natural dispositions of each animal are indications of the goal or end
towards which it is directed? What may be a typical example for such
a natural goal? Who or what constituted this goal? Do there exist bio-
logical species for the sake of this end? Are men determined by nature
to follow these alleged ends? Or is it rather a sort of natural inclination
which humans feel with regard to such ends? Hence, are we capable of
choosing to act contrary to nature, and is Kant demanding us not to
do so, but to follow our true nature? If that were the case, we would
have to accuse him of committing a naturalistic fallacy. Next, we have
to raise the question of which natural dispositions of biological species
may be regarded, from a teleological point of view, as relevant or tell-
ing. What does count here as important? Would it be persuasive to ask
why men possess a cecum? It is a well-known problem of Aristotelian
teleology that in describing biological species, e. g. humans, it takes
into consideration no more than few selected features, called proper
features (idia) and leaves out all the others.2
Apparently, there are more or less dysfunctional features or dispo-
sitions to be found in biological species. But given this fact, couldnt
it also be true that human reason has been generated by nature with-
out any goal, namely as being produced at random or as a feature
which has been useful under former environmental conditions which
nowadays no longer persist? To put the same point in a slightly dif-
ferent way: Why should one consider humans as a biological species
having this or that set of natural characteristics? Wouldnt it be more
attractive to separate the biological and cultural features of humanity,
and to put practical reason primarily on the second side? As a con-
sequence of this, doesnt it seem more persuasive to assume that hu-
man goals always exist in a culture-dependent context and that all are,

2
Cf. the new comprehensive monograph of Johnson (2005).
Kant on Ends in Nature and in Human Agency 49

consequently, relative to it? In order to raise the problem the other


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

rightly indicate that Kant is in a sort of dilemma: Kant himself per-


mits that reason can be appropriately used as an instrument to sup-
port our needs and interests, but at the same time, he suggests it to be
a non-instrumental power which exercises a profound moral influence
on human will. 3
Concerning section (d), there are some further questions to be
added. The most important one may be: why should the influence of
practical reason generate the good will which has been characterized,
some pages before, as the only good without restriction? What kind of
impact does Kant have in mind when he claims that practical reason
is able to produce a good will? One of the difficulties which this claim
raises is how a good will can be said to be a goal anyway. Is it the kind
of object someones actions can be directed to in the same sense as it
is the case with happiness? Can it serve as a goal in the way in which
the majority of ancient and medieval philosophers meant happiness
to be a goal, namely as a final end? Suppose that this is what Kant
wants to tell us. How can, then, our natural dispositions indicate such
a moral end? If it were correct to say that a naturalistic argument is
not appropriate to challenge morals, why should it be suitable to sup-
port a moral point of view? A further question of some importance is
why Kant takes into consideration only these two possible fi nal ends,
as if such an alternative was exhaustive, and why he takes them to be,
at least in this sense, mutually exclusive. We find no explicit justifica-
tion for these claims in our text.
The list of problems seems so overwhelming that there remains lit-
tle hope to making sense of our passage. A simple strategy to acquit
Kant on all these points is, now, to suppose that he did not ground
his teleological considerations on his own philosophical position, but
reacted on the convictions of some unnamed adversaries. By this line
of interpretation, it would be superfluous to ascribe to him such a
strange position. At least prima facie, it looks promising to claim that
Kant simply argues on the basis of a view adopted in this context and
for the sake of argument. At the beginning of our text, Kant seems to
say that he is merely testing the idea of a good will in the light of tele-
ological assumptions: Hence we will put this idea [i. e. that of a good
will] to the test from this point of view [i. e. teleological thinking]
(GMS, 395,13). Taken this way, what we have in our text would be an
argument ad personam, not Kants own position. And in fact, if one
compares our text with some passages to be found in his handwritten

3
Cf. Schnecker and Wood (2002, 54).
Kant on Ends in Nature and in Human Agency 51

reflections and in the transcripts of his lectures, it may seem as if Kant


responded to a precisely identifiable group of 18th century philoso-
phers, namely to naturalistic or materialist authors such as Helvtius,
LaMettrie, and Mandeville who in his eyes rejected precisely what
he meant by a good will and did this by adopting teleological argu-
ments. This view has been suggested by M. Forschner.4
But on a closer look, this line of interpretation can be excluded. If
one takes into consideration other Kantian texts on teleology, one will
find again all of the substantial claims of our passage. The central ele-
ments are present, e. g., in the Third proposition of the Idea toward a
Universal History with a Cosmopolitan Aim, a text published in 1784,
just one year before the Groundwork. In considering the purposes of
nature regarding mankind, Kant there unmistakeably maintains the
validity of the principle of suitability. 5 Furthermore, he claims that
humans are determined by nature to pursue their happiness inde-
pendently of their instincts, namely by activating their reason. Nature
equipped men only with a minimal set of means in order to force them
to develop their rational capabilities. The final aim of nature is, ac-
cording to Kant, to urge humans to develop their rational freedom.
All achievements of mankind should be, according to natures pur-
pose, the result of its own efforts. Kant says:
The invention of his food, his shelter, his external security and defense
(for which it [i. e. Nature] gave him neither the horns of the bull nor the
claws of the lion nor the teeth of the dog, but only hands), all pleasure
which can make life agreeable, no less insight and cleverness and even
the goodness of his will should entirely be his own handiwork. (IGA,
19,2934)

As the quotation shows, it is again Kants opinion that, according to


the plan of nature, goodness of will is produced by human insight.
We have no reason to think that this opinion functions as an argu-
mentum ad personam in the Idea. What is different here compared
with our text from the GMS, is that Kant does not stress the funda-
mental distinction between the things that man has brought about by
reason: human successes in producing food and shelter, in protecting
oneself and in attaining pleasure count as a rational achievements in

4
This and other interesting accounts of the sources and implicit presuppositions of
our passage are provided by Forschner (1988 and 1989).
5
Cf. Ak. 8.19,234: Nature does nothing in vein and is not sumptuous in using the
means towards its ends (Die Natur tut nmlich nichts berflssig und ist im Ge-
brauche der Mittel zu ihren Zwecken nicht verschwenderisch).
52 Christoph Horn

the same way as the generation of the moral attitude does. Leaving
the differences between our two texts aside, it is remarkable that the
text quoted from the Idea confirms that Kant considers nature to be a
kind of quasi-divine power developing a rational plan for all biologi-
cal species and providing them even if in an extremely parsimonious
way the means necessary for their survival and their happiness.
Before we come to the question of Kants overall view of teleology,
I would like to re-examine the problem of the possible adversaries of
our passage in the GMS. As I already mentioned, one crucial problem
of our text consists in the fact that it doesnt really contribute to the
analytical line of argument which leads, in GMS I, from common ra-
tional moral cognition, centered around the concept of a good will, to
the philosophical moral cognition, focused on the idea of a categori-
cal imperative. Now, one of the possible explanations for Kants intro-
duction of teleology could be that he thought of certain adversaries
which he wanted to refute in this very context. Who may be the phi-
losophers that Kant perhaps had in mind? What makes this question
difficult to answer is that these adversaries have to defend two differ-
ent claims which do not fit very well together. The first is that nature
does nothing in vain, since it is a divine order, and that our natural
dispositions are indicators of our natural goal. The second is that man
is determined by nature to strive for happiness which seems to be un-
derstood in a hedonistic and instinct-guided way. Whereas the former
seems typical for intellectualist conceptions of teleology such as that
of Aristotle and the Stoics, the latter belongs to the tradition of Epi-
curean philosophy. But as is well known, the former is not hedonistic,
and the latter is strictly a-teleological. Therefore Forschners solution
which I already mentioned regains, at least temporarily, a certain at-
tractivity. It may seem as if Kant had in mind the Neo-Epicurean bon
sauvage-philosophers Helvtius, LaMettrie, and Mandeville.
There is, however, a major obstacle for Forschners identification
of Kants adversaries. When Kant characterizes the kind of happi-
ness which they have in mind he makes use of the terms preser-
vation (Erhaltung) and welfare (Wohlergehen, GMS, 395,89). If
these authors were hedonists, Kant should have spoken of pleasure
or lust. Preservation and welfare are not subjective mental states, but
clearly objective goods. The absence of a hedonistic and subjective
vocabulary and the occurrence of the key word preservation makes
it highly plausible that the actual Kantian adversaries are the Sto-
ics. The ancient Stoics described self-preservation (sustasis heautou,
conservatio sui) as one of the provisional primary goods of nature,
Kant on Ends in Nature and in Human Agency 53

and they put it into sharp contrast to Epicurus cradle argument


which tried to identify pleasure as the first natural good.6 Its Kants
own concept of happiness which can be characterized, at least partly,
as being hedonistic and subjective, as we will see below (in part iii.).
For example, shortly before our passage, he describes happiness as
contentment with ones own condition (GMS, 393,156). Two fur-
ther indications for a Stoic background of Kants observations lie in
his concept of an instinct 7 and in the constatation that the persons
feeling misology are by no means morose or ungrateful toward the
kindness of the worlds government (GMS, 396,89). The expres-
sion worlds government is apparently Stoic and does not make any
sense with regard to the a-theological authors of the 18th century.
If we accept that the philosopher who hates reason is, according to
Kant, Rousseau8 (and perhaps authors like Helvtius, LaMettrie, and
Mandeville)9, the material collected by Forschner makes a new and
better sense. What we see, then, is that Kant partly defends his con-
temporaries against Stoic intellectualism without sharing the hedon-
istic naturalism of these philosophers. Kant thus opposes the kind of
teleology proposed by the Stoics who regarded reason and a rational
version of happiness as the ends towards which human nature is di-
rected. Kants strategy hence is a double one: On the one hand, he
accepts the idea that human reason, like any other natural feature of
mankind, possesses a teleological significance according to what we
called the principle of suitability. Insofar as this is the case, he shares
the Stoic view. What he, however, tries to make plausible, against the
Stoics, is that the real purpose of human reason is not happiness, but
the generation of a good will. Kant here somewhat illegitimately mix-
es his own hedonistic and subjective concept of happiness with the
Stoic one which is intellectualist and objectivist; this last one speaks
of preservation and well-being as provisionary goods.10 On the other
hand, he accepts the view of the a-theological Naturalists who reject
6
Cf. e. g. Chrysippus, SVF III, 178. 181. 184.
7
Instinctus is the Latin standard translation of the Greek horm which plays a fun-
damental role in Stoic theory of natural impulses.
8
Rousseau claims in the Discours sur lorigine et les fondements de lingalit as well
as in the mile that a full degree of happiness has only been real for our uncivilized
ancestors in a pre-rational state. For the passages in Rousseau which Kant might
have had in mind see J. Ferrari (1979, 17188).
9
In this respect, I follow Reich (1935) and Himmelmann (2003, 1302).
10
According to the KpV, the Stoic view of happiness is that the wise man attains it
by the consciousness of the moral way of thinking (im Bewusstsein der sittlichen
Denkungsart: KpV, 127,145). Kant misinterprets Stoic eudaimonism in terms of a
subjective fulfillment of desires.
54 Christoph Horn

reason insofar as it is a useless instrument to gain happiness. Kant


takes them as witnesses for the falsity of the Stoic kind of teleology,
but he immediately adds that there exists a more valuable kind of
destination for human reason: the rise of morality.
One expression which should be taken seriously as a signal is the
word misology. It hints at a precisely identifiable historical source,
namely Platos Phaedo (89d1 and d4). In Platos account of misology,
the hatred of those contemptuous of rationality is caused by their own
former love of reason and the disappointment which results from a
rough and uncritical use of it. To despise rationality is thus a conse-
quence of an intense, but disappointed enthusiasm for it. Somebody
may become an enemy of arguments in the same way in which one can
take the identity of a misanthrope: namely by putting too much trust in
unchecked arguments, as another person may practice a naive form of
philanthropy until she becomes repeatedly disappointed. Kant seems
to have read the Phaedo inspired by Moses Mendelssohn, who pub-
lished his own writing Phaedo or On the Immoratlity of Soul in 1767
as a selection and amplification of Platos dialogue. What is particu-
larly interesting is the observation that not only does the expression
misology stem from Platos Phaedo, but perhaps also the principle
of suitability. The Phaedo seems to be the first text in Western phi-
losophy which develops the idea of teleological explanation of nature.
There, it is Socrates who reports his deep disappointment when he
read as a young man the treatise of Anaxagoras whom he expected
to develop a philosophy of nature according to teleological principles.
As the Platonic Socrates says:
And I rejoiced to think that I had found in Anaxagoras a teacher of the
causes of existence such as I desired, and I imagined that he would tell me
first whether the earth is flat or round; and then he would further explain
the cause and the necessity of this, and would teach me the nature of the
best and show that this was best; and if he said that the earth was in the
centre, he would explain that this position was the best, and I should be
satisfied if this were shown to me, and not want any other sort of cause.
And I thought that I would then go and ask him about the sun and moon
and stars, and that he would explain to me their comparative swiftness,
and their returnings and various states, and how their several affections,
active and passive, were all for the best. For I could not imagine that when
he spoke of mind as the disposer of them, he would give any other account
of their being as they are, except that this was best; and I thought when
he had explained to me in detail the cause of each and the cause of all, he
would go on to explain to me what was best for each and what was best for
all. (Phaedo 97d598b3; transl. by B. Jowett).
Kant on Ends in Nature and in Human Agency 55

For the young Socrates, it is Universal Reason which structures the


world according to the principle that everything is organized in the
best way available. Unfortunately, Anaxagoras did not do him the
favour to provide a philosophy of nature based on this fundamental
principle. We are not precisely told what it could possibly mean that
something is organized in such a perfect form. But we find, in Pla-
tos text, a clear-cut opposition between mechanistic causes which do
not even count as causes and teleological explanations which alone
are seen as acceptable. In our terminology, Socrates distinguishes
between necessary and sufficient causes in explanation and assigns
mechanistic causes to the former kind, teleological causes to the latter
(Phaedo 99a4b8; transl. by B. Jowett):
There is surely a strange confusion of causes and conditions in all this. It
may be said, indeed, that without bones and muscles and the other parts
of the body I cannot execute my purposes. But to say that I do as I do
because of them, and that this is the way in which mind acts, and not
from the choice of the best, is a very careless and idle mode of speaking. I
wonder that they cannot distinguish the cause from the condition, which
the many, feeling about in the dark, are always mistaking and misnaming.
And thus one man makes a vortex all round and steadies the earth by the
heaven; another gives the air as a support to the earth, which is a sort of
broad trough.

For our purpose it is interesting that the whole autobiographical pas-


sage in the Phaedo seems to be devoted to the introduction of a new
and revolutionary way of explaining the natural world: by teleologi-
cal explanation.11 Apparently, there is little distance between this So-
cratic teleological thought and the principle of suitability which Kant
formulates as follows (GMS, 395,47):
In the natural predispositions of an organized being, i. e., a being arranged
purposively for life, we assume as a principle that no instrument is to be
encountered in it for any end except that which is the most suitable to and
appropriate for it.

Nevertheless, we find one major difference between the principle de-


veloped in the Phaedo and the Kantian formula: Kant restricts his
principle to living organisms. These are said to have instruments
which are organized in the best way possible. There is no such restric-
tion in Platos writings and his Socrates does not confine the principle
11
See e. g. the commentary of Gallop (1975) ad 97b898b6: The present passage
marks the transition from a mechanistic to a teleological conception of the natural
order that was to dominate European science for the next two thousand years.
56 Christoph Horn

to living beings. Yet, we also find a striking similarity: As in Plato, we


find in Kant the claim that living organisms are purposively organized
in a way that is best for a certain objective end. Kant does not mean
that they are perfectly organized for survival or reproduction, i. e. for
their own best; apparently, his kind of teleology is the objective one
inaugurated in the Phaedo.

II

It may be helpful now to clarify the most relevant kinds or aspects of


teleology before characterizing Kants observations in GMS I. One
fundamental dichotomy is that between a universal teleology which is
based on the thesis that the entire universe shows a coherent sort of
purposiveness directed towards one comprehensive end and special
or individual forms of teleology which claim that there exist final ends
only for certain (natural) species or individuals respectively. It is typi-
cal for traditional theological approaches to defend the first version,
and typical in many early modern positions to find the second form
of teleological thinking. Concerning our passage in the Groundwork,
we are confronted with the first form; Kant is apparently speaking
of a teleological design of the entire universe. Furthermore, it makes
a considerable difference if someone defends an external teleology,
maintaining that, in the outer world, an entity or event exists or hap-
pens for the sake of something else, e. g., that moonshine is made for
lovers, or an internal teleology which concerns the inner structure,
function, or organisation of something, especially that of an organism;
so we might say that the heart exists for the sake of blood circulation
and thereby for the survival of animals. In 63 of the Critique of the
Power of Judgment, Kant distinguishes between an external purpo-
siveness (ussere Zweckmigkeit) which he calls relative (since it
happens between two entities) and an internal one which signifies a
teleological structure within an entity. To explain this innere Zweck-
migkeit, he famously introduces the concept of an organism (or as
he says, an organisirtes Product der Natur) that possesses such an
internal purposiveness. An organism shows an internal structure of
reciprocity and interdependence in which every part is in the same
time means and end, cause and effect with regard to all other parts
(cf. especially KU 66). Nevertheless, it is obvious that Kant, in our
passage, is not thinking of this intriguing phenomenon. His issue here
clearly is based on external teleology.
Kant on Ends in Nature and in Human Agency 57

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

of Zweckmigkeit, if we reflect on the mere form of something e. g.


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

12
A useful overview on Kants concept of teleology and purposiveness is provided
by Frank and Zanetti (1996, 1184204). Marc-Wogau (21938, 71) differentiates be-
tween ten aspects of Kants concept of purposiveness.
Kant on Ends in Nature and in Human Agency 59

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

capabilities. What makes our impression so profoundly ambiguous is


that we seem to have enough reason to defend both anthropological
pessimism and optimism. Kant tries to solve the problem by assum-
ing that evil conditions in nature and human history are destined to
bring about civilization and, as a last consequence, even morality. It
is surprising to see how far he is going in this direction: e. g. even the
existence of mosquitoes and other pests is interpreted as an impulse
for human development:
Thus one could say, e. g., that the vermin that plague humans in their
clothes, hair, or bedding are, in accordance with a wise dispensation of
nature, an incentive for cleanliness, which is in itself already an important
means for the preservation of health. Or the mosquitoes and other sting-
ing insects that make the wilds of America so trying for the savages are
so many goads to spur these primitive people to drain the swamps and let
light into the thick, airless forests and thereby as well as by the cultivation
of the soil to make their abode more salubrious (KU, 67; 379,2231).

Only in the third area of traditional teleology, that of human agency,


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

13
Cf. KpV, 59: nihil appetimus nisi sub ratione boni, nihil aversamur nisi sub ratione
mali. See on this point Herman (1993) ch. 9.
Kant on Ends in Nature and in Human Agency 61

and comprehensive purpose of human agency.14 But on the other hand


there are passages in which rational freedom or autonomy seems to be
summum bonum, as Kant formulates explicitly.15 Be that as it may, if
human agency is a well-structured whole of non-arbitrary fi nal ends
and necessary means, it is tempting to think there may even exist a
connection to the teleologies of nature and history.
Now, what is the subject matter of GMS I, 3946? Apparently, what
is lacking is historical teleology. But we are confronted both with natu-
ral teleology and with considerations concerning human agency. As
soon as this thematic ambiguity becomes clear, one can express the
most fundamental difficulty of our passage as follows: what it seems
to make so extremely weak is its connection (and, as one might think,
confusion) of natural and personal teleology. Natural teleology (if le-
gitimate) is universal and external, whereas personal teleology is in-
dividual and internal. The first one may be judged as a pre-modern,
theological relic in our common consciousness without relevance for
scientific explanations of nature, whereas the second is relevant, but
should be confined to intentional human behavior. Therefore, the cru-
cial question for our further interpretation of the text has to be how the
two topics may, in Kants eyes, fit together. The situation is worsened by
the insight that these two themes of teleology are completely independ-
ent from one another and therefore of no reciprocal importance even
in the case that both versions are sound and acceptable. Suppose that
we had good reasons to accept the principle of suitability: every natu-
ral species might be considered as optimally adapted to its objective
purpose. Would that have any consequences for agents and the struc-
ture of their intentions, goals, and ends? Of course, not. Reasons to act
can only be justified by a process of internal considerations, by practi-
cal deliberation, not by clarifying the structure of the external world.
Whether humans are by nature (or by divine providence) destined for
this or that purpose does not matter on the level of an agents practical
reasons. If it were e. g. highly plausible that the human set of teeth is
that of a carnivore species, it would nevertheless incorrect to see in this

14
Cf. GMS,415,28416,1: There is one end, however, that one can presuppose as
actual for all rational beings [] and thus one aim that they can not merely can
have, but of which one can safely presuppose that without exception they do have t
in accordance with a natural necessity, and that is the aim at happiness. The hypo-
thetical imperative that represents the practical necessity of the action as a means
to furthering happiness is assertoric. One may expound it as necessary not merely
to an uncertain, merely possible aim, but to an aim that one can presuppose safely
and a priori with every human being, because it belongs to its essence.
15
Cf. Moralphilosophie Collins: Ak. 27/1. 344.
62 Christoph Horn

fact a reason to refuse vegetarianism. Precisely from a Kantian point


of view, it would be completely excluded to accept a sort of moral natu-
ralism in which we are told what is morally right or wrong with regard
to basic facts of nature even if they might be beyond any doubt. The
even more absurd consequences would be the result if we tried to infer
from our intentions, goals, and ends to the sphere of natural teleology.
What we are striving for or aiming at needs not to be something that
turns out to be, at the same time, fundamental for the basic structure
of nature, even if it were convincing to say that these aims and ends
are not only arbitrary, but essential ones. It is of crucial importance to
distinguish between objective and subjective teleology, the view from
outside and the view from inside, and Kant seems to commit the disas-
trous mistake of bringing these two aspects closely together.
What should we make of this precarious situation? Can we rightly
accuse Kant of such a profound confusion? Or is it preferable to inter-
pret our text as a well-considered connection between natural and per-
sonal teleology? To understand our passage in the Groundwork, we
now have to investigate the Kantian attitude towards teleology devel-
oped since his writing On the Various Races of Men (1775), continued
in the Idea (1784), in Determination of the Concept of a Human Race
(1785), in On the Use of Teleological Principles in Philosophy (1788),
and finally in the second part of the third Critique (1790). Kant repeat-
edly raises the question if we are entitled to follow our impression that
there exists an objective teleology in nature and in history. What he
proposes is a solution which lies in between the alternative of objective
and subjective teleology. Nature itself is not an intelligent being, and
history is not designed by a divine ordering hand, but nevertheless, we
find in it a high degree of suitability, organization, coordination, and
providence. As we perceive, lots of entities are in a harmonious or co-
operative relationship with each other, many things exist for the sake
of others in a sort of hostile and yet balanced relationship, and many
evils are suited to generate useful consequences. Despite this impres-
sion, Kant claims that it would be inadmissible to use final causes in
our explanation of the natural world or history. We are not allowed to
transgress the borderlines of our usual scientific explanation, namely
to explain things in terms of efficient causes. The use of fi nal causes
for the understanding of natural phenomena would be an illegitimate
transfer from the sphere of human behavior. Natural suitability is, in
contrast to the mechanistic nexus efficiens, not a constitutive or sub-
stantial principle in nature. According to his Transcendental Idealism,
Kant maintains that we can legitimately make use of final causes only
Kant on Ends in Nature and in Human Agency 63

in the reductive sense of what he calls a regulative principle. On one


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

be vindicated on the basis of practical teleology. Already in his writ-


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

16
On the Use of Teleological Principles in Philosophy: Ak. 8.182,35183,9: But since
a pure practical teleology, i. e. a morals, is determined to make real its purposes in
the world, it will not be allowed to neglect its possibility within it [i. e. the world],
both regarding the final causes implied in it, and regarding the appropriateness of
the supreme cause of a totality of all purposes as effect, hence a natural teleology,
as well as the possibility of a Nature at all, i. e. a transcendental philosophy, in order
to ensure that pure practical teleology has objective reality concerning the possibil-
ity of the object in practice, namely concerning the purpose whose realization is
prescribed by it [i. e. practical teleology].
Kant on Ends in Nature and in Human Agency 65

makes him alone capable of being a final end, to which the whole of na-
ture is teleologically subordinated. (KU, 435,15436,2).

To resume, we find two substantial reasons given in the third Critique


that Kants connection between natural and personal teleology is not
an involuntary confusion or mistake, but a well-considered argumen-
tative step. According to Kant, it is (a) the teleological structure of
an organism and (b) the teleology of rational agency which renders
plausible that man is the end-in-itself for the sake of which nature is
entirely designed. Personal teleology thus provides the decisive argu-
ment for the acceptance of natural teleology which otherwise would be
anachronistic, even if its status is confined to that of a regulative prin-
ciple. Natural teleology must not be transferred to the level of scientific
research or to speculative metaphysics. Historical teleology cannot be
taken as an objective law of history. But here we should look again at
to our text and raise the question if we discover here, too, any evidence
for a Kantian strategy of conscious fusion between the two topics.
Let us now return to the Groundwork. Note that only in the First
Section of the Groundwork do natural teleology and personal teleol-
ogy seem to be illegitimately mixed. There are no traces of natural
teleology in the rest of the text. In GMS II and III, Kants use of the
concept of nature is strongly reduced and only appears dependent on
or in contrast to personal teleology. On page 421, he expresses the
formula of universal law in terms of a universal law of nature: So
act as if the maxim of your action were to become through your will a
universal law of nature. In this alternative version of the categorical
imperative, Kant demands that one should, by a thought experiment,
imagine that ones own maxims should be valid in the universe as laws
of nature. The character of a thought experiment becomes manifest
by the words as if were which have no parallel in the formula of
universal law. The concept of a law of nature clearly implies here the
impossibility to act against it. The agent, thus, is asked to imagine
that her maxim might become a rule without exception. Also on page
455 f., we find the usual Kantian contrast of nature identified with
necessity and determinism and human freedom. When Kant uses
examples like that of the mosquitoes (and similar cases in 67 of the
KU), all he wants to claim is that teleology is valid for the reflecting
power of judgment (reflektierende Urteilskraft), not for determining
power of judgment (bestimmende Urteilskraft). The teleologies of na-
ture and history must not be understood in a constitutive, but in a
regulative sense (KU, 379,1016).
66 Christoph Horn

III

Towards the end of our teleological passage, Kant makes this claim
about the relationship of good will and happiness (GMS, 396,2430):
This will may therefore not be the single or entire good, but it must be the
highest good, and the condition for all the rest, even for every demand
for happiness, in which case it can be divided with the wisdom of nature,
when one perceives that the culture of reason, which is required for the
former, limits in many ways the attainment of the second aim, which is
always conditioned, namely of happiness, at least in this life [].

We are familiar with the doctrine of the highest good developed here
in nuce from the more detailed version of the Critique of Practical
Reason (KpV, 10814): The summum bonum for human beings must
be conceived as a combination of morality and happiness in the sense
that morality is the end for the sake of which one should act, whereas
happiness is the reward that God distributes in the afterlife according
to someones worthiness to be happy. According to this view, it would
be morally inappropriate if one acted for the sake of his own happi-
ness, but it would disappoint our natural inclinations if we couldnt
fulfill our desire for happiness. By combining morality as the end and
happiness as the reward and by setting happiness under the condition
of morality Kant escapes from the dilemma that either the demands
of moral law or our natural desires are missed. The first mistake is,
according to Kants account in the second Critique, that of Epicurean-
ism, the second that of Stoicism. In our passage of the Groundwork,
we encounter of the first short drafts of this doctrine. It presupposes
far-reaching assumptions concerning the understanding of happiness,
and it entails a harsh criticism of eudaemonism in moral philosophy.
As we see from Kants remarks on happiness in the Groundwork, he
is fully conscious of these consequences.
In his pre-critical period, Kant himself defended a sort of eudae-
monism which has been founded on a perfectionist account of human
life. We find this account e. g. in his General History of Nature and
Theory of Heaven (1755) where Kant claims that human happiness
consists in the future community with the infinite being; at that
point of his development, he believed that human nature is directed
towards its perfection. Nature thus prepares man for a perfect exist-
ence with God.17 A second stage of his thought on happiness might

17
Cf. Ak. 1.322 and Himmelmann (2003, 10).
Kant on Ends in Nature and in Human Agency 67

be identified in his lectures on anthropology prior to 1777. Following


Rousseaus sentiment of existence, Kant there developed an opti-
mistic view of happiness based on the principle of the perfection of
pleasure via aesthetic taste; in his account, he tries to unify the Stoic
and the Epicurean account of pleasure, as S. Meld Shell (2003) has
made plausible. Beginning with his lectures after 1777, and especially
in the Doctrine of Method of the fi rst Critique, we find the kind of
serious criticism of happiness known in more detailed versions from
the Groundwork and the second Critique. Kant now interprets happi-
ness as the comprehensive fulfilment of all the wishes and inclinations
of a person: Happiness is the satisfaction of all our desires; exten-
sive, in regard to their multiplicity; intensive, in regard to their degree;
and protensive, in regard to their duration (KrV, A 806/B 834). In
his lectures known as Praktische Philosophie Powalski [dating prob-
ably from the end of the 1770s] he defines happiness as satisfaction
of all pleasures anyway (Ak. 25.101). Happiness is thus reduced to
its subjective and hedonist components, and it is criticized precisely
for those aspects. Kant now characterizes all approaches in moral
philosophy which are based on the general principle of self-love or
someones own happiness as material principles (KpV, 22,68) and
claims that they are principally mistaken (5.39). In his late Anthro-
pology (1798), Kant famously maintains that all eudaemonists are
practical egoists (Anthro, 130).
In all of his treatises on moral philosophy dating from the criti-
cal period, the concept of happiness plays a thoroughly negative role
in Kants view. A very pointed statement is that of the Doctrine of
Virtue where Kant says that if eudaemonism (the principle of hap-
piness) is adopted as the principle instead of eleutheronomy (the
principle of freedom of the inner legislation), the consequence is the
euthanasia (quiet death) of all morality (MS, 378,157). At a closer
look, there are two different objections that Kant raises against hap-
piness as the key concept of morals. The first depends, as we saw, on
his conviction that the pursuit of happiness contradicts the principle
of morality. According to Kant, the only way in which an agent can
be adequately motivated is when she is acting from the viewpoint of
practical reason. Solely practical rationality can both rule out egocen-
tric motives in human agency, and guarantee a stable and permanent
form of motivation. The second objection is an epistemological one.
It says that all individual opinions concerning happiness are based on
experience and are thus vague and transitional; they do not permit
of more than a strategic approximation to a maximum of welfare,
68 Christoph Horn

since nobody is capable of determining with complete certainty, in


accordance with any principle, what will make him truly happy, be-
cause omniscience would be required for that (GMS, 418,224). We
can only act in accordance with empirical counsels, e. g., of diet, fru-
gality, politeness, restraint etc (GMS, 418,257). In GMS I 399, Kant
takes both objections together; but in our context, it is the fi rst one
that he has in mind.18
But despite his principal rejection of happiness as a practical and
theoretical basis of morals, Kant doesnt contest, neither in our con-
text nor in his general view, the fundamental significance of the hu-
man longing for happiness. Happiness is for Kant even the most
powerful and inward inclination (GMS, 399,8). As he underlines in
the second edition of the Groundwork, happiness is an aim that they
[i. e. all rational beings] not merely can have, but of which one can
safely presuppose that without exception they do have it in accord-
ance with a natural necessity (GMS, 415,302). How can we describe
the kind of synthesis which Kant proposes in order to connect the two
antagonistic concepts of morality and happiness? Roughly spoken,
there exist four strategies in moral philosophy to interpret the rela-
tionship of morality and happiness: [a] an identity thesis, [b] a recon-
ciliation thesis, [c] a conflict thesis, and [d] an incompatibility thesis.
Kant himself discusses thesis [a] in the Critique of Practical Reason
and characterizes it by its assumption that happiness and virtue are
analytically connected to each other, i. e. connected by conceptual
implication (KpV, 1134). But he rejects [a] both in its Epicurean form
(where moral virtue is an implication of happiness being a necessary
instrument to it) and in its Stoic variant (where the attainment of full
moral virtue is simply identified with the state of perfect happiness). If
a philosopher, as Kant does, accepts that both morality and virtue are
irreducibly founded in human nature, it clearly follows that he cannot
adopt the incompatibility thesis [d]. Hence we would expect that he
had to choose between [b] and [c]. But Kants position is a double one:
His doctrine of postulates advanced in the second Critique amounts

18
Both objections dont do justice to ancient (or medieval) Eudaemonism. As has
been shown by, e. g., Irwin (1996) and Weidemann (2001), Kant firstly neglects the
fact that ancient eudaemonism is based on an objective account of happiness, and is
hence able to discuss the necessary and sufficient conditions in an intersubjectively
comprehensible way. Secondly, Kant ignores that moral virtue doesnt figure, at
least in most of these accounts, as a means which is instrumentalized towards a
selfish form of happiness. On the contrary, adopting virtue implies the transforma-
tion of a moral personality. Furthermore, it doesnt seem adequate to characterize
ancient Eudaemonism generally in terms of hedonism.
Kant on Ends in Nature and in Human Agency 69

to the reconciliation thesis [b], whereas his remarks on the principal


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

The value of life would be thoroughly destroyed if one would describe


it in terms of happiness interpreted as the satisfaction of desires. Ac-
cording to Kant, no one would choose, if he could, to lead a hedon-
istic life for a second time. The value of life is based on what can be
actively done by an agent and, as Kant continues, by the sort of action
that confers purposiveness to nature.19

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesam-


melte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902) (abbreviated
as AA). The Critique of Pure Reason will be cited according to the A/B
pagination from the first and second editions.
19
I am grateful to Dieter Schnecker for his valuable comments on an earlier version
of my paper and to Alexander Cotter for checking my English.
70 Christoph Horn

All quotations of the Grundlegung are taken from Immanuel Kant,


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

Other works

Ferrari, Jean (1979): Les sources franaises de la philosophie de Kant. Paris.


Forschner, Maximilian (1988): Moralitt und Glckseligkeit in Kants Refle-
xionen, in: Zeitschrift fr philosophische Forschung 42, 351370.
Forschner, Maximilian (1989): Guter Wille und Ha der Vernunft, in: O.
Hffe (ed.), Grundlegung zur Metaphysik der Sitten. Ein kooperativer
Kommentar, Frankfurt a. M., 4565.
Frank, Manfred and Zanetti, Vronique (1996): Immanuel Kant. Schriften
zur sthetik und Naturphilosophie, Frankfurt a. M.
Gallop, David (1975): Plato, Phaedo. Translated wit Notes, Oxford.
Guyer, Paul (2000): Kant on Freedom, Law, and Happiness, Cambridge/New
York.
Guyer, Paul (2002): Ends of Reason and Ends of Nature: The Place of Tel-
eology in Kants Ethics, in: The Journal of Value Inquiry 36, 161186.
Herman, Barbara (1993): The Practice of Moral Judgment, Cambridge/
Mass.
Himmelmann, Beatrix (2003): Kants Begriff des Glcks, Berlin/New York.
Irwin, Terence H. (1996): Kants Criticism of Eudaemonism, in: S. Eng-
strom/J. Whiting, (edd.), Aristotle, Kant, and the Stoics. Rethinking Hap-
piness and Duty. Cambridge, 63101.
Johnson, Monte Ransome (2005): Aristotle on Teleology, Cambridge.
Korsgaard, Christine M. (1996): Creating the Kingdom of Ends, Cambridge/
New York.
Langthaler, Rudolf (1991): Kants Ethik als System der Zwecke. Perspekti-
ven einer modifizierten Idee der moralischen Teleologie und Ethikothe-
ologie, Berlin/New York.
Marc-Wogau, Konrad (21938): Vier Studien zu Kants Kritik der Urteilskraft.
Uppsala.
McFarland, J. D. (1970): Kants Concept of Teleology, Edinburgh.
McLaughlin, Peter (1990): Kants Critique of Teleology in Biological Expla-
nation: Antinomy and Teleology, Lewiston.
Reich, Klaus (1935): Kant und die Ethik der Griechen, Tbingen.
Kant on Ends in Nature and in Human Agency 71

Schnecker, Dieter / Wood, Allen (2002): Kants Grundlegung zur Metaphy-


sik der Sitten. Ein einfhrender Kommentar, Paderborn u. a.
Shell, Susan Meld (2003): Kants True Economy of Human Nature, in:
B. Jacobs/P. Kain (edd.), Essays on Kants Anthropology, Cambridge,
194229.
Weidemann, Hermann (2001): Kants Kritik am Eudmonismus und die Pla-
tonische Ethik, in: Kant-Studien 92, 1937.
Marcia Baron

Acting from Duty (GMS, 397401)

1. Kants discussion of acting from duty and his remarks about moral
worth are often read and taught to students as if they constituted
an independent essay that just happened to be published where it is
in the Grundlegung (or even: just happened to be published in the
Grundlegung). So read, it may seem that the whole point of that dis-
cussion is to put forward a test for the moral worth of actions. This
misconception then nourishes the perception that Kants ethics is dis-
turbingly moralistic. Why, it is asked, would we want a test for the
moral worth of actions? Rightness, perhaps, but moral worth?1 Ar-
guably, only those eager to keep the moral score on their neighbors or
(recognizing perhaps that they do not know their neighbors or even
their own motives well enough to apply the test) to dream about be-
ing the gatekeepers of heaven would be interested in such a test. Not
everyone puzzles over the (supposed) fact that Kant is putting forward
a test for the moral worth of actions. But those who do not puzzle
over why a test of moral worth is something we need are frequently
troubled by the content of the apparent test. Moral worth seems to be
doled out incorrectly.2
To a considerable extent these misgivings arise because GMS,
397401, are read in isolation from the rest of Section I. To appreci-
ate the import of GMS, 397401, we need to attend to Kants avowed
purpose in discussing acting from duty. He introduces the discussion
by saying that to develop the concept of a good will, he will put before
us the concept of duty, which contains that of a good will, though

1
Thanks to Vann McGee for raising this question in conversation with me around
1977, when we were graduate students at the University of North Carolina at Chapel
Hill.
2
The thought is that actions done from admirable feeling are morally worthier, or at
least as morally worthy, as actions done from duty. Among the many people who
have so remarked are Annas (1984); Blum (1980); Oakley (1992); Sidgwick (1981,
223); Stocker (1976).
Acting from Duty 73

under certain subjective limitations and hindrances, which, however,


far from concealing it and making it unrecognizable, rather elevate it
by contrast and let it shine forth all the more brightly (GMS, 397).
The end of the discussion is also to be borne in mind: it leads to Kants
first statement of the Categorical Imperative. Through a discussion of
what it is to act from duty, he asks what sort of principle would guide
a good will. The answer tells us what, if there is a supreme principle of
morality, that principle is. And as he said in the Preface, The present
groundwork is [] nothing more than the search for and establish-
ment of the supreme principle of morality (GMS, 392).
Harald Khl is providing the chapter that addresses GMS, 402
406, so I will not enumerate the steps that take us from acting from
duty to the Categorical Imperative. But we need to remember that the
discussion of acting from duty and moral worth does not stand alone.
It leads us from the idea of a good will to an explication of what a su-
preme principle of morality, if there is one, would have to be.

2. I turn now to the details of Kants discussion of acting from duty.


Kant says that he passes over all actions that are already recognized
as contrary to duty, even though they might be useful for this or that
aim; for with them the question cannot arise at all whether they might
have been done from duty, since they even conflict with it. It looks as
if he is saying that actions contrary to duty cannot be done from duty.
That would be a strange claim: why could I not mistakenly think X
was my duty, and act accordingly, when it in fact is contrary to duty?
But there is no need to read Kant as ruling out this possibility, given
his already recognized (schon [] erkannt). 3 Although this is not
explicit, it is reasonable to suppose that by already recognized he
means already recognized by the agent as contrary to duty (and prob-
ably also by those of us thinking about what sorts of actions might be
done from duty).4 And of course if the agent recognizes the action as
contrary to duty she cannot be performing such actions from duty.
Kant also sets aside actions done from self-interest, i. e., actions
to which human beings have immediately no inclination, but never-
3
I am grateful to Mark Timmons for drawing my attention to Kants already recog-
nized.
4
In saying that it is reasonable I do not mean to imply that it is clear that already rec-
ognized does mean already recognized by the agent [and probably others]. There
is room for reasonable disagreement here. Indeed, Kant does seem to think it quite
easy to distinguish what is good, what is evil, what conforms to duty or is contrary
to duty (GMS, 404). Still, he does not claim that honest mistakes are impossible,
and it would be uncharitable to attribute this view to him if we do not have to.
74 Marcia Baron

theless perform them because they are driven to it through another


inclination. Why set them aside? In this case, he explains, it is easy
to distinguish whether the action [] is done from duty or from a self-
seeking aim. By contrast, it is much harder to notice this difference
where the action is in conformity with duty and the subject yet has
besides this an immediate inclination to it. [Weit schwerer ist die-
ser Unterschied zu bemerken, wo die Handlung pflichtmssig ist und
das Subjekt noch berdem unmittelbare Neigung zu ihr hat.]5 What
exactly is Kant saying here? This much is clear: He takes the latter ac-
tions to be the cases that, for his purposes, merit attention. But a lot is
not clear. The puzzles include the following:

(1) Why do the latter actions those in accordance with duty and to which
the agent has an immediate inclination merit so much attention?
(2) Kant compares (a) actions done from self-interest with (b) actions to
which the agent has an immediate inclination. [(a) Handlungen [] zu
denen [] Menschen unmittelbar keine Neigung haben, sie aber dennoch
ausben, weil sie durch eine andere Neigung dazu getrieben werden; (b)
die Handlung pflichtmssig ist und das Subjekt noch berdem unmittel-
bare Neigung zu ihr hat.] Why the asymmetry? Why are the latter not
done from inclination (or why are the former not actions for which
the agent has a motive of self-interest)? If the answer is No reason; the
asymmetry is insignificant, contemporary commentators myself in-
cluded who mark a distinction between acting aus Neigung and acting
mit Neigung are misguided. For, thus understood, Kant would not be ob-
serving any such distinction.
(3) Just what is it that is much harder to notice?

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

5
A comment is in order concerning Neigung. Should this be understood narrowly,
to refer only to one particular kind of desire, or more generically? At GMS, 413n,
Kant writes, The dependence of the faculty of desire on sensations is called incli-
nation. and this always therefore proves a need. Strictly speaking, then, inclination
is only one type of desire. But I follow Henry Allison in thinking that Kant intends
inclination broadly in this discussion, to encompass, as Allison puts it, momen-
tary desires, instincts, passions, fears and disinclination more generally, any
stimulus to action that stems from our sensuous, as opposed to our rational nature
(Allison, 1990, 108). This reading is borne out by the variety of things in the exam-
ples at GMS, 397399, that are loosely classified under the heading of immediate
inclination.
Acting from Duty 75

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

6
One reason for my assumption is that he clearly is marking the distinction between
aus Neigung and mit Neigung at GMS, 400, where he writes, Only that which is
connected with my will merely as a ground, never as an effect, only what does not
serve my inclination but outweighs it, or at least wholly excludes it from the reckon-
ing in a choice, hence only the mere law for itself, can be an object of respect and
hence a command (italics added). The part I italicized allows for the possibility of
an inclination (though not necessarily a supporting inclination) being present yet
not influencing the decision.
7
It is troubling that Kant seems by this to be treating all actions done from immedi-
ate inclination as actions with a self-seeking aim. I will not here enter the debate on
whether Kant is a psychological hedonist with respect to all actions not done from
duty. For discussion of that question, see Reath (1989), Herman (2001), and Kerstein
(2002).
76 Marcia Baron

Third, it does not fit Kants purposes in that paragraph. Attention to


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

8
I will not keep repeating immediate; the idea is simply that one has an inclination
to do the action in question, and immediate serves to contrast that with acting
from self-interest (or as Kant has it, a mediate inclination). Note too that my clumsy
wording is due, at least in part, to my efforts to preserve the asymmetry noted above.
Otherwise I would write acting from duty is easier to differentiate from acting from
self-interest than from acting from immediate inclination.
Acting from Duty 77

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

9
For yet another interpretation, see Wood (1999, Ch. 1).
78 Marcia Baron

On my interpretation, Kants reason for attending so closely to ac-


tions to which the agent lacks any inclination (mediate or immediate) is
to bring home the difference between acting from duty and acting from
inclination, a difference less apparent than that between acting from
duty and acting from self-interest and a difference vital for under-
standing the idea of an unconditionally good will and why there is noth-
ing except the universal lawfulness of the action in general that can
serve the will as its principle (GMS, 402). But as it is easy to suppose
that Kant instead is aiming to convince us that an action cannot count
as done from duty if one is also inclined so to act, I will devote the next
section to assessing evidence thought to support that reading.

3. Does Kant hold that an action to which one has an immediate incli-
nation is thereby precluded from counting as done from duty (and
thus from having moral worth)?
Before we address this, a clarification is in order. The question
should not be confused with the following: Does Kant hold that an
action done from inclination cannot at the same time be done from
duty? I. e., does the fact that an action is done from inclination preclude
its also being done from duty? I believe the answer to this question is
Yes, but that the answer to the former question (that posed in the
previous paragraph) is No. One can act from duty mit Neigung but
cannot act aus Pflicht und Neigung.10 Some resist this distinction be-
tween (a) acting from duty, where one also has an inclination to do the
action in question, and (b) acting from both duty and inclination. They
claim that to have an inclination is by definition to be moved, at least
a little, to act accordingly. That being the case, they argue, one cannot
be inclined to do X, and then do X, without doing X at least partly from
the inclination. But the definitional claim is not one to which Kant
would, or should, accede. It presupposes a different (and less robust)
view of agency than he holds. His picture is not a mechanistic one of in-
10
As I explain in Baron (1995, Ch. 5), and more recently in Overdetermined Ac-
tions and Imperfect Duties (forthcoming), I do not consider this to be an embar-
rassment for Kant and his defenders. I think it is correct to say that one in fact
cannot do X simultaneously from duty and inclination, if by doing X from duty
we mean doing it because it is morally required. I also note in Overdetermined
Actions Allen Woods suggestion (primarily by e-mail and in conversation, but
briefly indicated in Wood (1999, 44) that acting from duty should be understood
more broadly, to include not only doing X because it is morally required but any
action that is the result of self-constraint through moral considerations. Thus un-
derstood, acting from duty and inclination might indeed be possible. I do not be-
lieve, however, that this conception of self-constraint is Kants, for reasons I briefly
indicate in that essay.
Acting from Duty 79

ternal forces moving us to act; so there is nothing in the nature of action


or motivation, on his theory of agency, that would render nonsensical
the idea that one can have an inclination to do X and do X (from duty)
without actually being moved by that inclination.11
Returning now to the question with which I opened this section:
Does Kant hold that an action to which one has an immediate inclina-
tion is thereby precluded from counting as done from duty? I will
first note inadequate bases for thinking the answer has to be Yes.
All of Kants illustrations in Grundlegung I of actions done from
duty are actions to which the agent is not inclined (and in some cases
is strongly disinclined). This has been regarded by some as strong evi-
dence that Kant holds that an action cannot count as done from duty
(and therefore, as morally worthy) if the agent is inclined so to act.
But it is not strong evidence, for there is another explanation of this
feature of his examples. Kant needs to provide examples of actions
done from duty, and wants readers to be clear that the actions really
are done from duty. That they are done from duty is most apparent if
no alternative account is available. So he purposely describes conduct
to which the agent is not inclined.
Examples of acting from duty without any inclination so to act
serve another purpose, as well. He is concerned to convince readers
that actions done from duty do not need the assistance of inclination
indeed, that one can act against all felt inclination, and against ones
own interests. He therefore presents examples of acting from duty
where no assistance from the inclinations is available.
Now the foregoing does not purport to show that Kant in fact does
hold that one can do X from duty while also being inclined to do X; it
purports only to show that the passages cited as evidence against that
view in fact are not evidence against it. That his examples are of ac-
tions to which the agent has no inclination (or on balance is more dis-
inclined than inclined) is thus compatible with both views: that an ac-

11
Of course it would be very hard to know that one has not acted from that inclina-
tion, but that is a different matter. (And clearly the fact that I cannot tell whether
I have acted from a motive I have when I do X is no reason to conclude that I must
have done so.)
For more on the distinction between acting mit Neigung and acting aus Neigung,
see my Freedom, Frailty, and Impurity (1993), a discussion of some issues in Al-
lison, Kants Theory of Freedom (1990). Allison replies in the same issue, and his
article is reprinted in Allison (1996). See also Baron (1995, 151 f.). For a sustained
argument that it makes no sense to suppose that one can act mit Neigung with-
out acting aus Neigung (an argument that I will not be able to address here), see
Latham (1994).
80 Marcia Baron

tion done from duty excludes inclination, and that it is consistent with
the agent being inclined so to act. (It is also compatible with the view
that an action can be done from both duty and inclination, though in
fact I do not think that Kant holds that to be possible.)12
Let me expand on why Kants choice of examples does not show
that he believes that an action from duty excludes inclination.13 Con-
sider how he sets up his examples. He speaks first of people so sym-
pathetically attuned that, even without any other motive of vanity or
utility to self, they take an inner gratification in spreading joy around
them, and can take delight in the contentment of others insofar as it is
their own work. In such a case, he asserts, the action [] has no
true moral worth for its maxim lacks moral content (GMS, 398). He
then asks us to imagine that the mind of that same friend of humanity
were clouded over with his own grief, extinguishing all his sympathetic
participation in the fate of others, thus emphasizing that the motives
available to him earlier are now no longer available to him. We are
asked to suppose that now, where no inclination any longer stimulates
him to it, he tears himself out of this deadly insensibility and does the
action without any inclination, solely from duty; only then does it for
the first time have its authentic moral worth (GMS, 398). The exam-
ple brings to the fore that unless he acts from duty, his action lacks
moral worth, no matter how beneficial the action and how amiable the
12
GMS, 400, provides strong evidence that Kant does not believe that to be possi-
ble: An action from duty is supposed entirely to abstract from [absondern] the
influence of inclination. The evidence is strong but not quite decisive. It is a little
unclear what absondern means here, and for that reason the evidence is not deci-
sive, but it is difficult to imagine anything absondern might mean that would not
entail that Kant believes an action cannot be done from both duty and inclination.
Some additional evidence comes just before GMS, 402: Nothing other than the
representation of the law in itself, which obviously occurs only in the rational being
insofar as it, and not the hoped-for effect, is the determining ground of the will,
therefore constitutes that so pre-eminent good which we call moral. The relevance
of this passage is apparent if we assume that that so pre-eminent good which we
call moral is equivalent to moral worth, and bear in mind that an action has moral
worth if and only if it is done from duty (GMS, 398). Additional evidence comes
from recognizing what it is to act from duty for Kant, and more generally, what
the relation is between incentives and action on Kants view of agency. Since the
incentive does not operate as a mechanical force, and instead as a reason, to say
that someone acts from both duty and from x is to say that duty alone is not the sole
reason for which he acts. But some explanation is needed here as to why it is not the
sole reason, and I can think of no explanation that avoids attributing to the agent an
insufficient regard for the fact that the action in question is ones duty. Acting from
duty is incompatible with acting from desire, since it involves regarding duty as a
less than decisive consideration. I explain this more fully in Baron (1995, Ch. 5) and
in Baron (forthcoming). See also Allison (1990, Ch. 6).
13
This paragraph and the following two paragraphs overlap with Baron (1995, 148 f.).
Acting from Duty 81

inclination. (It also is designed to remove any doubt that we have this
incentive the motive of duty and that it can be effective even in the
absence of any cooperating inclination.)
Dissenters those claiming that Kant holds that one cannot act from
duty mit Neigung charge that for the first time, in the sentence just
quoted, makes it plain that Kant is saying that the mans action lacks
moral worth unless he has no inclination to it. The only way around it,
they claim, is to take the liberty of inserting into the proposition the
qualification that it is the first time we can know that an action has
moral worth (Baker, 1986, 460).14 But as Barbara Herman points out,
it matters that Kant is not speaking at the end of his example (at GMS,
398, 513) of a different person from the one described earlier in the
example. He is not comparing two people, one who acts from inclina-
tion, and another who lacks inclination and acts from duty. He is saying
of someone who formerly acted on others behalf from inclination and
not from duty that only now, when the inclination to act on their behalf
is not available, does his action of aiding them have moral worth. Her-
man explains: Of him it is [] said: only when the inclination to help
others is not available does his helping action have moral worth. For of
him it was true that when he had the inclination he did not act from the
motive of duty (Herman, 1993, 1819). Kants use of first does not
signify that only when one acts without inclination does the persons
action have moral worth. He is saying of someone who, when he acts
with inclination, does not act from duty, that only when he lacks the in-
clination to help others does his action of helping have moral worth, for
only then does he act from duty. So the critics are wrong to claim that
if we are to reject the if inclinations, no moral worth interpretation,
we have to insert the qualification that only now can we know that an
action has moral worth.
But what about the end of GMS, 398, where Kant writes, if nature
had put little sympathy at all in the heart of this or that person, []
nevertheless would he not find a source within himself to give him-
self a far higher worth than that which a good-natured temperament
might have? This looks bad for those of us who hope to avoid the if
inclinations, no moral worth interpretation. For is not Kant saying
here that those who lack sympathy are able to give themselves a far
higher worth than are those whom nature has endowed with plen-
ty of sympathy? The idea, apparently, is that only the person lack-
ing sympathy, or a heartfelt inclination to help others, can act from

14
See also Henson (1979).
82 Marcia Baron

duty. Schillers (in)famous jabs do not seem so clearly unwarranted


after all. Kant seems to be saying that those with fellow-feeling are
precluded from attaining the moral worth available to those who by
temperament are cold and indifferent toward the sufferings of oth-
ers. Maybe Schiller was right: ones only choice, if one wants to be
morally worthy, is to try to rid oneself of fellow-feeling.15
Looking only at the wording in that passage, the reading is plausi-
ble enough. But it would be in conflict with Kants theory of freedom
to hold that the very having of fellow-feeling debars one from acting
from duty. Kant holds that no matter what our inclinations, we are al-
ways capable of acting from duty.16 Fortunately the un-Kantian read-
ing of the passage is not the only one possible. A much more Kantian
reading is simply that in acting from duty, the man gives himself a
higher worth than does the person who never acts from duty (the per-
son with the good-natured temperament). The latter could act from
duty, but does not. The former man acts from duty, and thus gives
himself a higher worth than the latter does.
Further support for understanding Grundlegung, 39899, not to
rule out the possibility of acting from duty mit Neigung comes from
the many passages in which Kant emphasizes the importance of sepa-
rating the pure from the empirical. These passages shed light on the
motivation for choosing examples in which the agents inclinations are
obstacles to doing his or her duty. See GMS, 388 f., and recall GMS,
397, where Kant explains that he will develop the concept of a good
will in the following manner: we will put before ourselves the con-
cept of duty, which contains that of a good will, though under cer-
tain subjective limitations and hindrances, which, however, far from
concealing it and making it unrecognizable, rather elevate it by con-
trast and make it shine forth all the more brightly. The limitations

15
I am modifying Schillers suggestion only slightly. It went like this (Schiller, 1981,
221):
Scruples of Conscience
I like to serve my friends, but unfortunately I do it with inclination
And so often I am bothered by the thought that I am not virtuous.
Decision
There is no other way but this! You must seek to despise them,
And do with repugnance what duty bids you.
I am using, with very slight alteration, Allen Woods translation of the passage. See
Wood (1999, 28). For a discussion of Schillers differences with Kant on duty and
inclination, see Reiner (1983).
16
One might argue that Kants view is that one is always capable of doing ones duty,
not that one is always capable of acting from duty. I will not try to address that pos-
sibility here.
Acting from Duty 83

and hindrances inclinations that conflict with duty make the good
will more evident. It should come as no surprise, then, that in each of
Kants examples, the agents inclinations hinder rather than help the
agent to act as morality requires. That they do should not be taken to
entail that inclinations, and affect in general, are always or typically,
in Kants view, moral hindrances.17
Kants discussion and commendation in Kritik der praktischen
Vernunft of a method of isolation further buttress this explanation
of his inclusion, in each example, of subjective hindrances to doing
ones duty. He likens the philosopher who arranges an experiment to
distinguish the moral (pure) determining ground from the empirical,
to the chemist (KpV, 92). Later in the same work he writes that the pu-
rity of the moral principle [] can be clearly shown only by removing
from the incentive of the action everything which people might count
as a part of happiness (KpV, 156, substituting people for men).

4. The arguments presented thus far undermine support for the claim
that Kant holds that an action cannot be done from duty (and thus can-
not have moral worth) if the agent is inclined so to act. Can we do better
than that? Is there evidence that Kant in fact does hold that an action
can be done from duty even when the agent is inclined so to act? I be-
lieve that GMS 397 provides very strong evidence, though not as deci-
sive as I used to take it to be. After saying that it is easy to distinguish
whether the action in conformity with duty is done from duty or from
a self-seeking aim in the case of an action to which the agent has no
immediate inclination but is driven to it [] through another inclina-
tion, Kant says that it is much harder to notice this difference where
the action is in conformity with duty and the subject yet has besides this
an immediate inclination to it. I reasoned that if absence of inclination
were necessary for the action to qualify as an action done from duty,
it would not be difficult at all to know whether an action to which the
agent had an inclination was done from duty. It would be clear that it
was not done from duty. Since Kant is quite explicit in saying that it is
not clear, it cannot be the case that absence of inclination to X is neces-
sary for X to qualify as an action done from duty.18

17
I discuss in Baron (1995, Ch. 6) a nexus of questions concerning whether there is
any positive role for affect in Kants ethics. Lara Denis defends Kant against some
of my claims in Denis (2000). See also Guyer (1993) and Sherman (1990).
18
So I wrote in Baron (1995, 150). That Henry Allison had independently interpreted
the passage in this way increased my confidence that I was reading the passage
correctly. See Allison (1990, 111). When I revisited the matter while writing Act-
84 Marcia Baron

I now think the evidence, though strong, is not decisive. It looks


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

ing from Duty for Allen Woods edition of Groundwork, I registered some doubt
(Baron, 2002, n. 4).
19
Additional evidence can be found in the Tugendlehre. Since Kant emphasizes in
that work the importance of cultivating various sentiments, it would be odd if he
held that if we have sentiments that support doing what is in fact our duty senti-
ments that would incline one to so act we cannot act from duty.
Acting from Duty 85

5. Assuming that I am right to say that acting from duty is compatible


with being inclined so to act, but not compatible with acting from in-
clination, some questions remain as to just what it is to act from duty.
Could the following actions count as actions done from duty?

(1) An action that is permissible but not obligatory. Let us suppose that
the agent would not have performed the action had she thought it im-
permissible, and that she reflected on whether or not it was permissible
before so acting.
(2) An act of helping another or of doing something by way of develop-
ing ones talents (practicing the violin, making arrangements to resume
violin lessons, buying new strings for the violin). Let us assume that the
agent opts to practice violin (or to arrange to take violin lessons) in part
because she recognizes a duty to develop her talents, and that she helps
another in the recognition that others needs make a moral claim on her
(and on everyone else, as well).

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

20
Unless, perhaps, one mistakenly thought it was morally required; but I will not
consider that possibility here.
86 Marcia Baron

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

21
I am ignoring here what I believe are exceptions (though Kant does not say that
they are): aiding where the need is grave and the aid needed is easy for the agent to
provide. See Baron (forthcoming).
22
This is clear both from his remarks at GMS, 421, and in the fact that the Categorical
Imperative applies differently to imperfect duties than to perfect duties. For further
discussion, see Baron (1995), Baron (forthcoming), Hill (1971), and Hill (2002).
Acting from Duty 87

at least some perfect duties. I have a duty not to lie to others, but just
how I fulfill this is, to some degree, up to me. Just as my duty to help
others may be discharged in many ways, my duty not to lie to others
can be discharged in a number of ways. Admittedly it is not exactly
the same; there is less latitude in the latter than in the former; but the
difference is not as huge as is sometimes supposed.23 I might discharge
the duty not to lie by telling the truth (though this avenue may not
be open to me if, say, it entails divulging anothers secret), by deftly
changing the subject, or saying something outrageous at just the right
time so as to distract or disarm my interlocutor, or more straightfor-
wardly by indicating that I do not want to discuss that topic (though
if the timing is not quite right, this may amount to giving the answer
away, something I may be obliged not to do).
But are they really so unlike merely permissible actions? The situ-
ation seems fairly similar to the case where there are, in a particular
situation, only a few permissible actions; it is my duty to opt for one of
those in order to avoid doing what is impermissible. The difference
is this: the maxims of the merely permissible actions are not instan-
tiations of a general maxim which is itself obligatory (except at a very
abstract level: do not do what is impermissible). They have in common
only that they are not impermissible. Unlike acts of helping others, or
of developing ones talents, they are not attached to an obligatory end.

6. Reflection on these sorts of cases leads to the thought that per-


haps there is something askew in asking whether an individual action
is done from duty or not. Although Kant does indeed use the word
Handlung, which suggests individual actions, acting from duty is re-
ally a matter of acting a certain way, not performing this action from
this motive. It concerns conduct over time, conduct guided by certain
maxims. Likewise, moral worth seems to reside not primarily in indi-
vidual actions, but in conduct and more specifically, in the maxims,
or principles, by which one conducts ones life.
That moral worth does not reside primarily in individual actions is
evident in Kants second proposition: an action from duty has its
moral worth not in the aim that is supposed to be attained by it, but

23
I am highlighting the similarities in order to make sense of what I take to be Kants
position (insofar as he had a clear position). In my Overdetermined Actions
(forthcoming, sect. 5), I highlight the differences so as to indicate why I think one
should take a different position, viz. that one can adopt a maxim of beneficence
from duty but cannot help from duty in any particular instance (except where aid
actually is obligatory, because it is desperately needed and easy to give).
88 Marcia Baron

rather in the maxim in accordance with which it is resolved upon, i. e.,


in the principle of the volition, in accordance with which the action is
done (GMS, 400). I am drawing attention not to the negative part;
I think Kants readers generally are clear that moral worth does not
reside in the aim, or intended result. What seems not to be sufficiently
heeded is the positive part of the assertion. The moral worth resides
in the maxim. This suggests two matters that are inadequately attend-
ed to (especially by Kants critics, but sometimes by his defenders, as
well). First, since maxims guide conduct, and are not simply attached
to individual actions, moral worth is much more naturally attributed
to conduct than to isolated actions. Secondly, whereas Kants views
are often summarized as the moral worth of an action lies in its mo-
tive, that is not really accurate. It resides in the maxim. The mistake
of thinking that moral worth lies in the motive is pernicious, because
the word motive carries some theoretical (roughly, empiricist) bag-
gage: it suggests a moving force that impels the agent so to act. This is,
of course, quite far from Kants view of agency, but even those of us
who are fully aware of that may still be laboring with a distorting im-
age when we speak of acting from duty (or especially, acting from
the motive of duty). If we are, then the assertion that moral worth re-
sides in the motive is all the more misleading. Recognizing that moral
worth resides in the maxim changes the picture.
That Kant is far more concerned with character, and conduct over
a long stretch of time, than with the moral worth or lack thereof of iso-
lated actions is something I have argued elsewhere (Baron, 1995), fo-
cusing especially on later works (Metaphysik der Sitten and Religion).
But even if we confine our attention to GMS, 397401, it is difficult to
maintain the view that Kant is offering an account of the moral worth
of individual actions. First, there is the fact that he shifts from talking
about the moral worth of actions to the moral worth of character
(at the end of GMS, 398, and start of GMS, 399) and later (towards the
end of GMS, 399) speaks of conduct (Verhalten) having moral
worth. We would not expect such a shift (particularly from action to
character) if he were concerned to provide an account of the moral
worth of actions. Second, as noted, the fact that moral worth resides
in maxims indicates that the value lies in the principle by which one
conducts oneself, not in each individual action, and that evaluating
individual actions for their moral worth would (quite apart from the
epistemic problems) be misguided.
It is nonetheless somewhat perplexing that Kant speaks of moral
worth as attaching to a Handlung (unless Handlung in late 18th
Acting from Duty 89

c. German could signify conduct, rather than simply action). But it


helps to make sense of this if we bear in mind the role that discussion
of the moral worth of actions plays in Kants development of the idea
of a supreme principle of morality, purified of anything empirical. To
elucidate the good will and to develop therefrom the principle that
guides the good will, he needs to zero in on the goodness of actions
done from duty so as to bring out the following: (a) their goodness
comes only from the will itself; and (b) the principle of a good will is
the universal lawfulness of the action in general, i. e., I ought never
to conduct myself except so that I could also will that my maxim be-
come a universal law (GMS, 402).

7. In this essay I have focused on two interpretive questions. The first


was very specific: What exactly is much harder to notice at GMS,
397? What contrast is Kant drawing? My answer was that Kants
much harder to notice this difference concerned comparative dif-
ficulties in differentiating concepts, not in discerning motives. Act-
ing from duty is easier to differentiate from acting from self-interest
than from an action to which one has an immediate inclination. The
second question was more general, and covered more of the text; and
an answer to it provided a context for the first question. The second
question was: Why do actions which accord with duty, and to which
the agent has an immediate inclination, receive so much attention in
Kants discussion of acting from duty? I argued that it is not because
Kant is trying to show that such actions cannot be done from duty,
despite appearances that he is arguing that. The idea, rather, is this.
Kant is, in the interests of illuminating the concept of a good will and
ultimately locating the principle that would guide such a will, explain-
ing what it is to act from duty. To be clear on what it is to act from duty,
it is necessary to isolate acting from duty from conduct that might be
confused with it. There is no risk of mistaking for actions done from
duty actions that we and the agent recognize as contrary to duty, so
there is no need to discuss actions recognized as contrary to duty.
Nor is there any danger of confusing actions from duty with actions
from self-interest. But actions to which the agent has an immediate
inclination where the agent does not act from duty might be con-
fused with actions done from duty. Hence the need to discuss actions
to which the agent has an immediate inclination. Kant then presents
examples both of actions to which the agent has an immediate inclina-
tion, where the agent acts from that inclination, and actions which we
can see are not done from inclination and are done from duty. So that
90 Marcia Baron

it is clear that the actions really are done from duty, Kant builds into
the case that the agent is not inclined to so act.
Actions done from duty differ from actions that otherwise resemble
them in that the maxims of the former have moral content. Exactly what
this means is not spelled out very fully in Grundlegung, and readers in-
terested in this question would do well to read Part I of Religion.24 But it
is clear that what differentiates them is not that actions from inclination
spring from one sort of motive, and actions from duty spring from anoth-
er;25 rather, the idea is that only actions from duty are guided by a commit-
ment to doing whatever morality requires.26 Just as understanding, wit,
the power of judgment, and like talents of the mind, though good and
to be wished for can also become evil if the will that is to make use
of [them ] is not good (GMS, 393) likewise, inclinations such as fellow-
feeling need to be guided by a commitment to doing what is right. 27

Literature

Kants writings
Kants writings will be cited according to the pagination of Kants gesammelte
Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902).
GMS Grundlegung zur Metaphysik der Sitten, AA, IV
Groundwork for the Metaphysics of Morals (2002), Wood, Allen W.
(ed., trans.), New Haven, Yale University Press.
KpV Kritik der praktischen Vernunft, AA, V
The Critique of Practical Reason (1956; 3rd ed.1993), Beck, Lewis
White (ed., trans.), New York, Macmillan Publishing Co.
MdST Die Metaphysik der Sitten. Metaphysische Anfangsgrnde der Tu-
gendlehre , AA,VI
The Metaphysics of Morals, Part II (1996) in Gregor, Mary J. (ed.,
trans.): Immanuel Kant: Practical Philosophy, Cambridge, Cam-
bridge University Press.
RGV Die Religion innerhalb der Grenzen der blossen Vernunft, AA,VI
Religion within the Boundaries of Mere Reason (1998) in Wood, Al-
len / di Giovanni, George (eds., trans.): Religion within the Bound-
aries of Mere Reason and Other Writings, Cambridge, Cambridge
University Press.
24
See especially RGV 2930 and 36.
25
See Korsgaard (1989).
26
For further discussion, see Baron (1995, Ch. 5).
27
A draft of this paper was presented in Bonn, in July 2004, as part of the confer-
ence on Kants Grundlegung held at Rheinische Friedrich-Wilhelms-Universitt.
I am grateful to the conference organizers and to the participants for their helpful
discussion, to Lara Denis, Dieter Schnecker, and Allen Wood for their written
comments, and to Elizabeth Tropman for editorial assistance.
Acting from Duty 91

Other Works

Allison, Henry E. (1990): Kants Theory of Freedom, New York, Cambridge


University Press.
Allison, Henry E. (1996): Idealism and Freedom: Essays on Kants Theoreti-
cal and Practical Philosophy, New York, Cambridge University Press.
Allison, Henry E. (2001): Ethics, Evil, and Anthropology in Kant: Remarks
on Allen Woods Kants Ethical Thought, Ethics 111, 594613.
Annas, Julia (1984): Personal Love and Kantian Ethics in Effi Briest, Phi-
losophy and Literature 8, 1531, reprinted in Badhwar, Neera Kapur (ed.)
(1993): Friendship: A Philosophical Reader, Ithaca, Cornell University
Press.
Baker, Judith (1986): Do Ones Motives Have to Be Pure? in Grandy, Rich-
ard / Warner, Richard (eds.): Philosophical Grounds of Rationality, Lon-
don, Oxford University Press.
Baron, Marcia (1993): Freedom, Frailty, and Impurity, Inquiry 36, 431
441.
Baron, Marcia (1995): Kantian Ethics Almost without Apology, Ithaca, Cor-
nell University Press.
Baron, Marcia (2002): Acting from Duty, in Wood, Allen W. (ed.): Ground-
work for the Metaphysics of Morals, New Haven, Yale University Press, 92
110, published (2004) in German as Handeln aus Pflicht, in Ameriks, Karl
/ Sturma, Dieter (eds.): Kants Ethik, Paderborn, Mentis Verlag, 8097.
Baron, Marcia (forthcoming): Overdetermined Actions, Imperfect Duties,
and Moral Worth, in Klemme, Heiner F. / Khn, Manfred / Schnecker,
Dieter (eds.): Moralische Motivation. Kant und die Alternativen, Ham-
burg, Felix Meiner Verlag.
Blum, Lawrence (1980): Friendship, Altruism, and Morality, London,
Routledge and Kegan Paul.
Denis, Lara (2000): Kants Cold Sage and the Sublimity of Apathy, Kan-
tian Review 4, 4873.
Guyer, Paul (1993): Kant and the Experience of Freedom: Essays on Aesthet-
ics and Morality, Cambridge, Cambridge University Press.
Henson, Richard G. (1979): What Kant Might Have Said: Moral Worth
and the Overdetermination of Dutiful Action, Philosophical Review 88,
3954.
Herman, Barbara (1993): The Practice of Moral Judgment, Cambridge, Har-
vard University Press.
Herman, Barbara (2001): Rethinking Kants Hedonism, in Byrne, Alex /
Stalnaker, Robert / Wedgwood, Ralph (eds.): Fact and Value: Essays on
Ethics and Metaphysics for Judith Jarvis Thomson, Cambridge, MA, MIT
Press.
Hill, Thomas E., Jr. (1971): Kant on Imperfect Duty and Supererogation,
Kant-Studien 62, 5576, reprinted in Hill (1992): Dignity and Practical
Reason in Kants Moral Theory, Ithaca, Cornell University Press.
Hill, Thomas E., Jr. (2002): Human Welfare and Moral Worth: Kantian Per-
spectives, Oxford, Oxford University Press.
Kerstein, Samuel J. (2002): Kants Search for the Supreme Principle of Mo-
rality, Cambridge, Cambridge University Press.
92 Marcia Baron

Korsgaard, Christine M. (1989): Kants Analysis of Obligation: The Argu-


ment of Foundations I, The Monist 72, 311340, reprinted in Korsgaard
(1996): Creating the Kingdom of Ends, New York, Cambridge University
Press.
Latham, Noa (1994): Causally Irrelevant Reasons and Action Solely from
the Motive of Duty, Journal of Philosophy 94, 599618.
Oakley, Justin (1992): Morality and the Emotions, New York, Routledge.
Reath, Andrews (1989): Hedonism, Heteronomy, and Kants Principle of
Happiness, Pacific Philosophical Quarterly 70, 4272.
Reiner, Hans (1983): Duty and Inclination: The Fundamentals of Morality
Discussed and Redefined with Special Regard to Kant and Schiller, Santos,
Mark (trans), The Hague, Martinus Nijhoff Publishers.
Schiller, Friedrich (1981): Xenien in Trunz, Erich (ed.): Goethes Werke,
vol. 1, Munich, Beck, 208234.
Sherman, Nancy (1990): The Place of Emotion in Kantian Morality, in
Flanagan, Owen / Rorty, Amelie Oksenberg (eds.): Identity, Character,
and Morality, Cambridge, MA, MIT Press, 149170.
Sidgwick, Henry (1981): The Methods of Ethics, Indianapolis, Hackett.
Stocker, Michael (1976): The Schizophrenia of Modern Ethical Theories,
Journal of Philosophy 68, 453466.
Wood, Allen W. (1999): Kants Ethical Thought, Cambridge, Cambridge
University Press.
Wood, Allen W. (2003): The Good Will, Philosophical Topics, 457484.
Harald Khl

The Derivation of the Moral Law (GMS, 402, 420f.)

No new principle of morality is set forth in


it [in the Groundwork] but only a new for-
mula. But whoever knows what a formu-
la means to a mathematician, which deter-
mines quite precisely what is to be done to
solve a problem will not take a formula
that does this with respect to all duty in gen-
eral as something that is insignificant and
can be dispensed with.
Kant, Critique of Practical Reason, Preface

I. Posing the Question and a Common Prejudice

A supreme moral principle is an essential part of an ethical theory.


This, at least, was Immanuel Kants conviction. Thus he had to be
able, concerning his moral law, to state its content (GMS, 420).1
It is common knowledge that Kant made several attempts to derive
the content of his moral law: in the Groundwork of the Metaphysics
of Morals and the Critique of Practical Reason (GMS, 402 and 420 f.;
KpV, 17).
Kants interpreters have not appreciated his efforts in this area. His
attempts at deriving the formula of his moral law (the Categorical
Imperative) are commonly seen as difficult to understand or as being
incomprehensible; or, if his derivations are believed to be understood,
they are characterized as implausible or fallacious. A few examples of
such critique are representative of many others:
1
What is here at issue is the basic formulation of Kants moral law (or his Categorical
Imperative) that can be found (with small variations) in GMS (402, 421) and in KpV,
7. One could also call it the Universal Law-Formulation of the Categorical Impera-
tive. I will not take into account the End-in-itself-, the Autonomy-, and the Kingdom
of Ends-formulations of the Imperative.
94 Harald Khl

H. J. Paton, not famous as a Kant critic, sees a weakness in


Kants argument: If we interpret the argument [in the way Paton
does] it is simply misleading. Further, in response to Kants attempt
at a proof in GMS I, Paton makes the general remark: Kant has the
unfortunate tendency to conflate and skim over the important transi-
tions in his proofs (Paton 1958, 71). One of the most distinguished
contemporary German philosophers, Ernst Tugendhat, thinks that
Kants argument in GMS I is of course invalid, the transition
to the conclusion is simply arbitrary (Tugendhat, 1993, 130). The
argument that Kant gives [in GMS II] is surely false (ibid. 136 f.;
my insertion). so we can state that his attempt to give the very
idea of a categorical imperative [the Categorical Imperative] a par-
ticular meaning has failed (ibid. 137 f.; my insertion). Henry E. Al-
lison (1996, 143) fi nds a gap in the proof of GMS I and transfers this
diagnosis to the argument in GMS II. Nonetheless, Allison thinks of
the corresponding argument in the KpV that it doent contain a gap.
Allen W. Wood and Dieter Schnecker also think Kants proof to de-
termine the content of the moral law in GMS I is highly problematic
because it contains a gap. The argument in GMS II is open to the
same objections (Schnecker / Wood, 22004, 91, 93, 126).
In what follows, I will try to show that, contrary to common opin-
ion, Kants derivations of the basic formula of the moral law in the
GMS can be plausibly reconstructed (parts IIIV of this essay).2
Consequently, other reconstructions are targets for critique (part V).
Nonetheless, the possibility of a comprehensible explication of Kants
derivation of the Categorical Imperative does not mean that his proofs
are themselves beyond criticism (part VI).

II. From outside in

It seems to me useful to begin with a self-made argument (albeit in-


spired by Kant): before reconstructing the Kantian texts. My own ar-
gument to determine the content of a moral law runs as follows:
1. If there were something like a moral law, what would it say? This
is the initial question.
2. Even if we dont (yet) know the wording of the moral law, we know
that we are searching for a law.

2
In connection with my critique of Allison, in part V of this paper, I discuss the cor-
responding argument in the KpV.
The Derivation of the Moral Law 95

3. From the connotations of the term law we can conclude that a


moral law is a rule that encompasses a wide range of creatures (for
Kant: all rational beings). Such a law is a universal law. 3 Let us
assume that as a moral law it carries with it an expectation that all
addressees can be made to insightfully understand it. Hence the
universality of our law means the expectation of its being univer-
sally valid.
4. Further, we assume that we know what the formula of the moral
law cannot contain. We presume that it cannot be a rule that de-
mands certain actions because they are useful to achieve contin-
gent aims.4 (On this point many of us have Kantian intuitions.)
5. What else could our moral law contain? In any case, its formula-
tion must express its normative character (for beings like us): it is
a universal rule. As such, it demands a particular type of action,
actions in conformity with the rule. In the same sense we can speak
of action that is demanded by a law and in conformity with a law
(or of maxims that are in conformity with a law). In this sense, I
will speak of lawfulness1 (Gesetzmigkeit1), which is demanded
by the (moral) law; if an action or maxim is in this sense lawful
(gesetzmig), it is lawful1. It is necessary to use an index here,
because the Kantian notion Gesetzmigkeit has a second, cru-
cially different meaning to which I will refer as lawfulness2 and
which means something like lawlikeness; more on this later.
6. The previous point expresses the banality that a moral law, as a
rule, demands action in accordance with the law. But which kind
of action? What property of the law could it be that the demanded
behaviour should conform with? If the law does not contain the
demand to pursuit certain aims, whose realization affords the ex-
ecution of certain actions (cf. above step 4.): then the only thing
that remains in the law of which it can sensibly be said that actions
must conform to, is the legal form, its having the character of a law,
i. e. the pretension of being universally valid that belongs to the
concept of a law (cf. above step 3).
For the content of the moral law, this means that the demanded be-
haviour must be able to be prescribed in a legal formulation, i. e. in

3
If practical laws are to be laws, they must, as Kant puts it, be valid for the will
of every rational creature (KpV, 19). A practical law that I cognize as such must
qualify for a giving of universal law: this is an identical proposition and therefore
self-evident. (KpV, 27)
4
Even hapiness is a contingent aim in this sense, since particular conceptions of it
cannot be made universally binding.
96 Harald Khl

the form of a law, because it is universalizable and can thus justifiably


be demanded of all addressees. In other words, the actions demanded
must be grounded in intentions or subjective principles (maxims) that
can serve as universal laws (or: can be reformulated into such laws). 5
Here now is the place for speaking of lawfulness2 (Gesetzmigkeit2).
To act in accordance with a moral law in this sense means to constitute
ones actions in a lawful 2 manner, i. e. to act in a way that ones action it-
self is lawlike (and hence carries the expectation of being universally
valid with it). The content of the moral law thus demands action that is
lawful2 (gesetzmig2) or maxims that are lawful2.
By my lights, the foregoing argument is intelligible and conclu-
sive. The conclusion and so the Kantian proofs, if my considerations
should substantially coincide with his therefore can only be ques-
tionable if one or more premises are false. I will examine this possibil-
ity in the concluding part of my paper.

III. Kants Proof in Groundwork II

I will now turn to those passages in Kants texts where he himself at-
tempted to derive the content of his moral law. The proof he presents
in GMS II (420 f.) for the basic formula of the Categorical Imperative
is, in my opinion, easier to understand than its counterpart in GMS I
(402). In any case, the latter proof can be understood with the help of
the deliberations developed in the preceeding part of my paper. For
this reason, I take it to be methodologically necessary to begin with
the latter proof.
Kant surprises us here with the idea of whether the mere concept
of a categorical imperative may not also provide its formula contain-

5
Strictly speaking, a morally qualified maxim cannot become a moral law. Maxims
express general intentions of individual persons, while laws (at least for beings like
us) are normative, i. e. are rules. So, strictly speaking one must say that the moral
qualification of being universalizable that a maxim may possess is a good reason to
formulate a corresponding norm (law). Kant himself invites the reading just criti-
cized when he writes: Now I want only to know whether a maxim could also hold
as a universal practical law. (KpV, 27) But he also expressly writes: maxims are
indeed principles but not imperatives. (KpV, 20) Charitable readers orient them-
selves at the best formulations that an author has found for his claims. See also
the corresponding interpretation of Bruce Aune (1979), 23 f.: Kant characterizes
practical laws in a way implying that they could not be [] maxims. But it would,
acccording to Aune, be okay to say: if a given maxim is universally valid, a cor-
responding, fully general, lawlike statement is also valid (my emphasis). Such a
statement is, in other words, a candidate for the relevant law.
The Derivation of the Moral Law 97

ing the proposition which alone can be a categorical imperative (420;


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

I will start my interpretation of this passage with sentence (3), in which


Kant presents the basic formulation of his Categorical Imperative. Two
meanings of law are here at work: the single categorical imperative
refers to the Categorical Imperative (i. e. the moral law for morally
imperfect beings). The searched-for law (singular) prescribes maxims
to have a constitution so that they could become a universal law. The
last formulation implies the assumption of a potential plurality of laws
that could be generated from maxims that would meet the require-
ments of the law. Thus, (3) implies a hierarchy of laws. Consequently
I will call the law the supreme moral law, or the fundamental moral
law.6 Kants formulation of that law: act only in accordance with that
maxim through which you can at the same time will that it become a
universal law (my emphasis) must be understood in the sense of my
footnote 5 that the universalizability of maxims is a reason to estab-
lish a corresponding norm.
In (2.1) the term law, which the Categorical Imperative contains,
refers to the law-character which the supreme law posesses as a law:
6
See the title of KpV, 7.
98 Harald Khl

It is accompanied by the expectation of its universal validity. (Cf. II, 2


and II, 3.) Kants formulation of the necessity of the maxim (still in
2.1) to conform with that law, expresses the prescriptive character of
the supreme law. According to Kant, every norm expresses a practi-
cal necessity (and categorical imperatives an unconditional practical
necessity); in this context that means that each maxim must conform
to this law: i. e. to the law-character of the Categorical Imperative.
I understand the idea of conformity to the law (Gesetzmigkeit)
which is contained in the previous formulation in the sense of lawful-
ness1 (Gesetzmigkeit1; cf. II, 5), i. e. in the sense of a demand to
follow a categorically formulated rule. Note that if Kant meant here
already lawlikeness (Gesetzmigkeit2) the prescription that the
maxims themselves should have legal form there would have been
no reason for him to continue his proof. But more importantly, in that
case he would, in (2.1), have presupposed what should result only as
the conclusion (2.3).
In (2.2) it is established what the searched-for formulation of the
Categorical Imperative (Kant: the law) cannot contain: no condi-
tion to which it would be limited. (This corresponds to II.4.) After
all, the Categorical Imperative, whose content is to be ascertained, is a
categorical imperative. Such imperatives, qua being un-conditional,
are characterized negatively7 in the following way: they must not, in
contrast to hypothetical imperatives, be followed under the condi-
tion of the utility of the hypothetically demanded actions. Note that
this and only this point of his proof justifies the Kantian suggestion
that hes deriving the content of the Categorical Imperative from the
character of a categorical imperative.
Ad (2.3): [N]othing is left with which the maxim of action is to
conform but the universality of a law as such; and this conformity
alone is what the imperative properly represents as necessary. Now
Kant has arrived at his conclusion. In II,3 I explained the universal-
ity of a law as such by our expectation of its universal validity. That
the maxim of action is to conform to the universality of a law as
such a conformity the imperative properly represents as neces-
sary (represents = formulates) I read in the sense of lawfulness2

7
The positive descriptions that Kant gives in the GMS about the character of categor-
ical imperatives are often overlooked. A categorically demanded action is neces-
sary of itself and good in itself (414); good through the form and the principle
(416) of the action itself. Its form is the form, which consists in universality (436)
that is demanded by the principle of Kantian ethics. See H. Khl (2001), part IV.
See also Khl (2004), 128, as well as Khl (2006), Chapter 5.
The Derivation of the Moral Law 99

= lawlikeness (= Gesetzmigkeit2) (my emphasis). This means the


demand that a maxim, if it is to be considered as moral, can (nor-
matively re-formulated) be a law. (See II.6.) The moral law has this
demand as its content. The formulation of the Categorical Impera-
tive in Kants (3) is simply an alternative expression of what Kant has
already concluded in (2.3).
In the antecedent explication of Kants proof in GMS II, I believe
the difference between lawfulness1 (= Gesetzmigkeit1) and law-
fulness2 = lawlikeness (Gesetzmigkeit2) has proofed useful. Of
course, what has been shown thus far has only been an explication
of the component parts out of which Kants reasoning consists. We
are still missing an insight into the plot of his argument. This can be
presented as follows:
What knowledge is now available to us that we could use to deter-
mine the content of the Categorical Imperative? First, we know that
we are dealing with an imperative, a rule (see (2.1) and (2.3)): with
which the maxim of action is to conform). Second, we mean to know
that this imperative should be categorical,8 not hypothetical (see (2.2)
of Kants proof). Third, since the Categorical Imperative is meant to
formulate a supreme law, it possesses a legal character: in the sense of
being associated with an expectation to be valid for all X.9
So, when Im thinking of a categorical imperative, I also know
what it contains: one can say in allusion to Kants sentence (1) in 420.
It contains a universal, supreme, categorical rule for action. Moreover
it is clear that the searched-for imperative, qua rule, should prescribe
something for actions or maxims: presumably a property that they
should possess; an aspect in which they should conform to the im-
perative.
Now Kants question seems to me to have been the following: as-
suming we know nothing about the Categorical imperative at the out-
set except the three aforementioned aspects, which of these contains
something that could be prescribed for actions or maxims of action?
Not the first aspect, the prescriptive character. That would be point-
less, because maxims have, in expressing intentions, a different logical
form than prescriptions. Not the second aspect, since the application
of the predicate categorical to maxims is nonsensical. Thus only the
third aspect remains: the legal character. So the supreme law can at
8
[L]aws must be categorical: otherwise they are not laws. Cf. KpV, 20. this
rule is a law because it is a categorical imperative. Cf. KpV, 21.
9
With the lawgiving of reason the rule must be objectively and universally valid.
Cf. KpV, 20 f.
100 Harald Khl

most stipulate that maxims must themselves have a lawlike character,


that is, be able to establish (lower-level) moral laws (plural).
And is this not precisely Kants proof? For, since the imperative
contains, beyond the law [my third: legal character], only the neces-
sity that the maxim be in conformity with this law [my first: prescrip-
tive character], while the law contains no condition to which it would
be limited [my second: the categorical character], nothing is left with
which the maxim of action is to conform but the universality of a law
as such [the result of my process of elimination]; and this conformity
alone is what the imperative properly represents as necessary. There
is, therefore, only a single categorical imperative and it is this: act only
in accordance with that maxim through which you can at the same
time will that it become a universal law. Now, we can say in more
than simply an allusion, along with Kant that, when I think of a cat-
egorical imperative, I immediately know what it contains.
By my lights, this explication of Kants proof in GMS II is both
intelligible and conclusive. And the moral of this story? There is only
hope for a supreme moral principle, to be actually a law (that is to be
universally valid), if it prescribes nothing for maxims other than this:
they themselves must be universal. Stated negatively: if the fundamen-
tal moral law permitted other maxims, the laws own claim to univer-
sality would be destroyed. For such maxims would not be compatible
with its character of being a law.
Of course, Kants proof only gives a formulation of a supreme moral
principle. Justification of its content (that is of the demand expressed
by it) is another matter, which Kant will address in GMS III. Even if
the demand, to act according to universally valid rules of action, were
justified, one could still ask if the resulting law is a constructive, or
fruitful principle.

IV. Kants Proof in Groundwork I

Kants first attempt to prove the content of his moral law (402) is,
in my opinion, more difficult to understand than his later attempt
420 f. The difficulty comes from the perspective from which he ap-
proaches the matter. Kants starter for the proof is now the previously
conducted search for the moral motive. For him, a morally motivated
action (for morally imperfect beings) consists in acting from duty.
Only then does an action have moral worth (397 ff.). Kant takes up
these theses again immediately preceding his proof 402, as well as
The Derivation of the Moral Law 101

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

Sentence (1) poses the initial question (what kind of law can that
be) and makes clear from the beginning that Kant wants to perform
a proof via an analysis of the genuine moral motive. First, Kants talk
about the representation of the law which must determine the will
(my emphasis) is clearly about a motive. Second, because that rep-
resentation is a representation of the moral law, Kant can only be
referring to the moral motive to act for the sake of the law (see 390),
or in the case of human beings, out of duty. This is the motive that
must be present for a will to be called good (401).11 Kants initial
question is thus: what is the content of moral law itself, if the motive
to act according to it is the motive to act for the sake of the law? In
10
Here, for the first time in GMS I, Kant mentions the (moral) law. In the Preface
to the GMS, Kant already mentions (moral) laws (plural; in 387 f.) and soon there-
after for the first time the moral law (390; my emphasis).
11
With his talk of the representation (of the law) that must determine the will in
moral action, Kant takes up again his previous discussion about the representation
of the law (401). There already he claimed that this representation insofar as it is
the determing ground of the will, can constitute the preeminent good of the will we
call moral (ibid.).
102 Harald Khl

other words, what is implied, for the wording of the moral law, by this
description of the moral motive?
Ad (2.1): The moral motive to act for the sake of the law refers
ultimately to the moral law. Thus it cannot consist in the impulse to
achieve something beyond obeying the law (no regard is present
for an effect expected (see (1)). But then (2.2) the moral motive to
act for the sake of the law can only refer to the law-character of the
law, and be the motive for generating actions with this character. Act-
ing this way is consequently demanded by the moral law. (2.3)
This last step in the proof takes place by a shift in perspectives. If
out of the perspective of the moral motive to act for the sake of the
law this motive can only refer to the law-property : then seen
from the perspective of the searched-for law-formulation a law that
demands the exercise of so motivated actions accordingly can only
contain the precept: Carry out actions with property (or act accord-
ing to maxims with this property).
The conformity of actions with universal law, in (2.2), and the
mere conformity to law, in (3), is that which I earlier referred to as
lawlikeness (Gesetzmigkeit2):
The moral law consists in the command to carry out actions that
are capable of being laws (or whose underlying maxim can be trans-
formed into a law). A good action (with moral worth) is one done for
the sake of the law: that is (solely12) because it is capable of being a
law. I can here leave the question unanswered whether for Kant it is
part of the content of the moral law that actions that occur because of
their lawlikeness must be done from that motive.13
What differentiates the proofs that Kant presents in 402 and 421 f.
to determine the content of the moral law? The first proof, an ex-
amination from the perspective of the moral motive, seems to me to
rely on the crucial move (not explicit in Kants text) of assuming that
the moral motive must ultimately refer to the law. Thus, it cannot be
for the sake of profits that may result from following the law. Conse-
quently, it can only aim at that which defines the law qua law: that is
the expectation associated with it of being universally valid.
In the proof 421 f., I see the crucial move as the transition from
lawfulness1 (Gesetzmigkeit1) to lawfulness2 = lawlikeness (Gesetz-

12
I argue for this strict reading in Khl (1990), 84 ff.
13
If one reads into the formulation of the the Categorical Imperative the demand
of genuinely morally motivated action it is impossible to explicate the difference
between the law-conformity (Legalitt) and morality (Moralitt) of an action.
I have argued for this thesis in Khl (1990), 67, Fn. 9.
The Derivation of the Moral Law 103

migkeit2). Gesetzmigkeit1 appears only implicitly in the proof


402.14 One easily gets the impression that here Kants introduction of
lawlikeness (in (2.2) and (3)) is somewhat abrupt, thereby skipping a
step in the argument which I believe is a false impression.
Kant could have carried out the proof by explicitly working with
Gesetzmigkeit1 as follows:
(1) Good actions have to be done for the sake of the law (i. e. out of the
moral motive: morality).
(2) To act for the sake of the law implies wanting to act lawfully1 (= ge-
setzmig1; law-conformity), that is in accord with the legal precept
even if the content of said precept is for now unknown.
(3) That at which the lawfully1-oriented action cannot aim is some profit
resulting from obeying the law.
(4) Thus it must aim at conforming to the legal character of the law, that
is according to its claim to universal validity (lawlikeness = Gesetz-
migkeit2). To act for the sake of the law thus means to act for the
sake of the lawlikeness of ones own actions.
(5) The content of the moral law thus consists in the demand for actions
that are lawlike (lawful 2).

V. Three Critics of Kant, and an ally

In my discussion of secondary literature I concentrate on a thread that


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

14
It occurs insofar as the thought that the searched-for law contains the notion that
it should be followed, that is, the prescribed actions should be in accord with the
law.
15
See for the following: Aune (1979), 28 ff.
104 Harald Khl

Aune took up this text as follows:


(S1) Kants idea here is that since a good and dutiful will does not con-
form to a practical law because the consequences of doing so will satisfy
some purpose or inclination, the only rational basis for conforming to it
must be supplied by the intrinsic character of the law itself [i. e.] the
character of being a universal law. (28; my insertion.)

Out of this Aune develops supposedly Kants formula of the prac-


tical respectively moral law:
L: Conform your actions to universal law. (29)

From this Aune distinguishes the following, Kants fi rst formulation


of the moral law in 402 and the formulation of the Categorical Im-
perative 421, the latter equivalent with Aunes
C1: Act only on that maxim through which you can at the same time will
that it should become a universal law. (Ibid.)

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

Based on these considerations, Aune poses the question: Which is


really the fundamental practical law, L or C1 ?, and answers it for
the benefit of C1 (29). For Aune, the supposed gap in Kants proof
consists in the absence of an argumentative bridge from L (which he
presumes to have found in Kants text) to C1. What should we make
of Aunes interpretation?
(S1) does not make it clear that Kant wanted to reach a formulation
of the moral law by considering the motive for moral action. For Kant
The Derivation of the Moral Law 105

this is clear in the discussion prior to sentence (2.1). By contrast, Aune


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

proof at 402, first wanted to discover the formulation of the moral law
(out of which concrete laws could be derived), he had, on his way to
that formulation, to avoid talk of such laws. On my reading there is no
mention of them.
If Aune were correct in attributing L to Kant, he would also be
correct in his assertion that C1 is more likely to be the searched-for
fundamental moral principle than L. But since L is not held by Kant
at all, Aunes question regarding the relationship between L and C1
does not even arise. Of course, there can be no gap between two one
of which doesnt exist. Kants proof in 402 may be faulty, but for this
gap he cannot be held responsible.
I want to move from my last point of contention with Aune to my
criticisms of both Woods and Allisons interpretations. Aune had
formulated that Kants good and duitiful will does not conform
to a practical law because and the only rational basis for con-
forming to it (see (S1)), and also that L is a higher-order princi-
ple telling us how to conform to certain lower order laws (see (S3),
my emphasis). Through this conformity-talk a misplaced notion has
been imported into Kants Text on 402 (and took place in the his-
tory of this passages interpretation): what I have termed lawfulness1
(Gesetzmigkeit1). In this sense one acts lawful if ones own action
corresponds with a/the law, i. e. one does what one is commanded to
do. As I have tried to show in part IV of my paper, this notion of Ge-
setzmigkeit1 is not present in 402 (in contrast to the proof in 420 f.).
Rather, the conformity of actions as such with universal law, which
alone must serve the will as its principle and the mere conformity
with the law as such that should be present in morally motivated ac-
tion serving the will as its principle (present in 402) is captured by
what I have called lawlikeness. This means the capacity of an action
or a maxim to serve as the epistemic basis for a norm that can be
thought to be universally valid.
The presentation of proof 402 that Allen Wood gives in his book
on Kants Ethical Thought (KET)16 is too brief to be considered a
serious effort at comprehending the argument. His critique is just as
brief. Wood obviously thinks the errors in Kants proof are so evident
that it is unnecessary to explicate them. Kants reflections are based
on fallacious reasoning (or at least it is an argument with a large gap
remaining to be filled) (KET: 48). Wood sees the same gap in the
proof that Aune saw between his L and C1.

16
See Wood (1999), 47 f.
The Derivation of the Moral Law 107

Woods treatment of proof 420 f. is more thorough. His criticism


of Kant is beset by a misreading similar Aunes. Thus, Woods recon-
struction also utilizes the notion of lawfulness1 (Gesetzmigkeit1)
understood as conformity with moral laws (plural):
The argument [420 f.] infers that the conformity of maxims to universal
laws as such follows from the mere concept of a categorical imperative.
The first and most abstract version of the categorical imperative could be
expressed as follows:
CI: Adopt only maxims that conform to universal law as such. (KET: 78)

Out of this Wood generates the objection we have encountered in


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

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

Wood cannot have generated his interpretation out of Kants formu-


lation of the Categorical Imperative, according to which one should
only act following according to maxims that one can simultaneously
desire to become a universal law (my emphasis). Even though the
idea of a plurality of laws is implicit here it would be incoherent to
speak of a maxim conforming to a law that the maxim eventually is
fi rst becoming17 .
(3.) There is, therefore, only a single categorical imperative and it is this:
act only in accordance with that maxim through which you can at the
same time will that it become a universal law. (420 f. my emphasis; origi-
nal text italicizes the entire Categorical Imperative.)

17
On the problem of Kants claim that a maxim can become a universal law, cf. my
footnote 5.
108 Harald Khl

Wood could have discovered his interpretation only in the formulation


of Kants sentence (2.3) that states: nothing is left with which the max-
im of action is to conform but the universality of a [singular] law as such;
and this conformity alone is what the imperative properly represents as
necessary (my emphasis and insertion). Wood obviously reads a law
as one of a potential plurality of laws and the conformity to law in
the sense of lawfulness1 (Gesetzmigkeit1). But this sentence is not
speaking of many potential laws, but of the character of a law as such:
its universality. This is conveyed by lawlikeness: maxims should evince
the same universality as a law, so that they may become laws. That
which maxims should conform to, are, on my reading of 420 f., not laws
(as Wood contends) but the universality that signifies a law qua law.
The difficulty that Wood had in reading Kants text is apparent in his
vacillation between different formulations. He speaks of the the con-
formity of maxims to universal laws [plural] as such; formulates as CI:
Adopt only maxims that conform to universal law [sing.] as such; und
formulates that our maxims should conform to universal law [sing.] (See
the Wood quotes above in my text; my emphasisses and insertions.)
Wood even imports the notion of conformity with moral laws (plu-
ral) into his understanding of the basic formulation of the Categorical
Imperative 421 (as demonstrated above this idea is there not present):
it [the basic formula] proposes a test to use on any given maxim
to determine whether that maxim conforms to these universal laws
(KET: 79; my emphasis and insertion Wood italicizes test). However
Kants thought that maxims eventually can become laws is different
from the notion that they should conform to laws in the sense of law-
fulness1 (Gesetzmigkeit1).
In his explicitly critical treatment of The fallacy in Kants deduc-
tion (KET: 81 f.) Wood claims, like Aune, that the basic formula of
the Categorical Imperative does not follow out of his CI. (Here again
he wavers between speaking of the conformity [of a maxim] to the
universality of law in general and saying: this tells us no more
than that our maxims ought to conform to whatever universal laws
(plural) there are (KET: 81; emphasis and insertions mine). But since,
by my lights, Kant does not hold CI, there can be no gap in moving
from CI to the basic formula.

Henry Allison follows Aunes and Woods criticism of Kant in his pa-
per already mentioned.18 Allison leans on Woods work on Hegels

18
See Allison (1996), footnote 6.
The Derivation of the Moral Law 109

Ethical Thought (HET),19 in which Wood previously examined Kants


arguments about the content of the Categorical Imperative. Allison
writes, like Aune:
since Kant presents this imperative [the Categorical Imperative] in a
parenthetical clause intended to explicate the notion of conformity to uni-
versal law as such, it would appear that he regarded these principles as
equivalent. At the very least, he must be taken as claiming that the fi rst
somehow entails the second.
Here, then, is the crucial gap in Kants argument. (145 my insertion.)

According to Allison, Wood wanted to show the impossibility of de-


riving a viable moral principle from the mere concept of universal
practical law (143). The way in which Allison relates the misreadings
of Aune and Wood is remarkable in at least one point. To make this
clear, I will introduce a distinction that Wood used in HET:
Wood distinguished the universality of applicability and the
universality of concern as two desirable properties of practical
laws (HET: 135 f.; my emphasis; see also 165 f.). [T]hat laws should
have universality of applicability means [e]verybody should be
equally subject to them. (HET: 165; my emphasis.) The universal-
ity of concern that we desire of good laws rests on the assumption
that everyones following those laws will result in a system of collec-
tive behaviour that is rational and generally beneficial (ibid.). Allison
recapitulates Woods distinction (which Wood uses as an objection
against Kant), noting that Kants mistake consisted in failing to dis-
tinguish between these two meanings of universality in his proofs to
determine the content of the Categorical Imperative.
Now the astonishing point I alluded to above is Allisons claim that
Woods universality of applicability (Everybody should be equally
subject to them = the laws) is equivalent to Aunes L, which the read-
er will recall is the formulation of the moral law that Kant supposedly
holds: conform your actions to universal law. Aune elaborated this
claim as follows in his (S3): L is a higher-order principle telling us
how to conform to certain lower-order laws. Rather, Woods univer-
sality of applicability roughly corresponds to what I have called the
legal character of a practical principle (in part II of this paper). Out
of the connotations of the notion law one can deduce, as I claimed
there, that a moral law is a precept encompassing a broad circle of
creatures.

19
See Wood (1990), 161 ff.
110 Harald Khl

In my opinion, the prospects of Woods criticism are not encourag-


ing if his universality of applicability means the same as Aunes L.
This is because, as I argued above, Kant not at all held L. Thus if Alli-
sons equation would be correct, Wood lacks something to set against
his universality of concern.
Regardless of whether or not Kants proofs to determine the con-
tent of his Categorical Imperative are correct, I find Woods distinc-
tion between universality of applicability (as he understands it) and
a universality of concern very helpful. But is it, as he understands
it, a useful critical tool against Kant? What Wood characterizes as
concern in his phrase universality of concern sometimes is called
by him collective benefit or collective rationality. He then uses
this terminology to claim that collective benefit [] is not identical
with Kants FUL test [that is the basic formulation of the Categorical
Imperative], which says that it must be possible to will without con-
tradiction that the law will be universally followed (HET: 165 my
emphasis). Because Kants conception of the law contains [only] uni-
versality of applicability, it does not contain anything like collective
benefit (ibid. my insertion). True. But what looks to be a factual
argument for the relative emptiness of content, or unfruitfulness, or
the morally indeterminate character of Kants formulation of the law
itself, Wood transforms into a criticism of Kants derivation of this
formulation (without good reason, as Id argue): We cannot infer (as
Kant does) that a principle with universality of applicability is eo ipso
a principle with collective rationality. Such an inference is invalid.
(HET: 166; my emphasis.) Im unable to understand how Kant could
have made this mistake in his derivation of the Categorical Imperative
if he (according to Wood) should have been satisfied with a principle
that merely possessed universality of applicability.
Before I return to Allison on a different point, Id like to sum-
marize the hitherto given critique of all three of Kants critics. They
have allowed themselves to be mislead by the ambiguity of the term
Gesetzmigkeit (and related ways of speaking). Gesetzmigkeit1
is the conformity (of an action or a maxim) to a (or the) Moral Law,
while Gesetzmigkeit2 means the legislative form of a maxim.
Kants critics failed to clarify this ambiguity, consequently taking cru-
cial places in Kants argument to mean Gesetzmigkeit1, when Kant
was actually working with the notion of Gesetzmigkeit2. Thus they
were able to import the idea of a plurality of laws into Kants deriva-
tion of the Categorical Imperative, a plurality to which maxims needed
to conform. Kant was subsequently criticized for this idea. In fact, it is
The Derivation of the Moral Law 111

unintelligible where these laws could have originated. Such laws can
only be generated by the application of the Categorical Imperative to
maxims. But this Imperative can only be used after its formulation
has been derived. The aforementioned laws cannot appear during the
derivation.
The gap in the derivation of the moral law that Kants critics believe
they have found in GMS 402 and 420 f. cannot be found in the corre-
sponding proof in the KpV ( 17). This, at least, is Allisons claim.
According to him, in the KpV the gap has successfully been bridged
by means of the introduction of transcendental freedom (144). Of
course, Kant would have done himself, by introducing transcenden-
tal freedom at that place, a disservice. Who, except staunch Kantians,
would still believe in the existence of a freedom which implies inde-
pendence from causal determination by everything empirical? (151)
Even Allison knows that if Kants so-called metaphysical deduction
(L. W. Beck) of the moral law rests on a thick conception of transcen-
dentally free agency, that is for better or worse (154; my emphasis).
If that is what grounds moral law! I take the assumption to be ques-
tionable that Kant used transcendental freedom as a premise in his
proof in the KpV. I read this proof as follows: Kant goes from the the-
sis of 4 to his conclusion in 7. 5/6, where transcendental freedom
is introduced, is an interlude:
In the Theorem of 4, Kant moves from the assumption [i]f all
material of a law, , is abstracted from it, nothing remains except the
mere form of a universal legislation to the conclusion: a rational
being must suppose that their mere form [= the form of maxims],
through which they are fitted for giving universal laws, is alone that
which makes them a practical law. (KprV, 27; my insertion.) This form
is prescribed for maxims in the Fundamental law of pure practical
reason ( 7) by the demand One ought to act a certain way (KpV,
30; my emphasis). The transition from 4 to 7 would be the transi-
tion that was, concerning the corresponding GMS-passages, criticized
by Aune, Wood and Allision an objection I challenged.
So, on my view, what function does the interlude in 5 and 6 have?
For Kant a will to wich only the legislative form of the maxim can
serve as a law is a free will: free in the transcendental sense (KpV
29, 5). Granted that a will is free [which I read as: if we can assume
that ] then we can manage Problem II, which is to find the law
which alone is competent to determine it necessarily. Because the
lawgiving form, insofar as it is contained in the maxim, is therefore
the only thing that can constitute a determining ground of the [free]
112 Harald Khl

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

20
See my reconstruction of Kants Deduction of the categorial imperative in: Khl,
2004, 129 ff.
The Derivation of the Moral Law 113

If therefore, the freedom of the will is presupposed, morality together


with its principle follows from it by mere analysis of its concept. Free-
dom must be presupposed as a property of the will of all rational beings.
(GMS, 447)
Now, in the KpV, Kant no longer believes this is legitimate. My im-
pression is that Kant was here so bent on correcting his position of the
GMS that he inserted the subject of the (now believed to be impos-
sible) deduction of the moral law in an argument that was concerned
with something else: not to ground the Categorical Imperative (the
demand it expresses) but to first of all discover its formulation.

In his search for an understanding of Kants Search for the Supreme


Principle of Morality (so the title of his illuminating book21), Samu-
el J. Kerstein unfolds what he calls a new, criterial reading (12) of
Kants derivation of the moral law. His critique of the gap-theorists
(Aune, Wood, Allison) is preceeded by the following reconstruction
concerning both of the Kantian proofs in GMS I and II:
First, Kant develops criteria that any viable candidate for the supreme
principle of morality must fulfil. [] Second, Kant tries to establish that no
possible rival to the Formula of Universal Law fulfils all of these criteria.
Finally, Kant attempts to demonstrate that the Formula of Universal Law
remains as a viable candidate for a principle that fulfils all of them. With
these three steps, Kant strives to prove that if there is a supreme principle
of morality, then it is this formula (12, cf. Kersteins Summary: 93).
The criteria developed by Kerstein in step one constitute (cf. 12) what
he calls Kants basic concept of the requirements that have to be
fulfilled by a supreme moral principle: It would be practical, abso-
lutely necessary, binding on all rational agents, and would serve as a
supreme norm for the moral evaluation of action (1). The first three
of these criteria are in accordance with my own claim, concerning
Kants proof in Groundwork II, that the moral law contains the high-
est categorical prescription for action, which is simultaneously univer-
sal (in part III of this paper).
Concerning Kants derivation in GMS I, Kerstein contends (as I
did above in part IV): Kant holds his discussions of the good will, [of
acting from] duty, and moral worth to be essential for his proof, add-
ing that he (and Kant) believe this discussion to be necessary for the
development of criteria for the supreme principle of morality (86; my
emphasis and insertion). Considering Kants proof in Groundwork II
21
Cambridge 2002. References in the following paragraphs are to this book.
114 Harald Khl

Kerstein finds it elliptical (93). Surprisingly, he introduces an ad-


ditional criterion as a constituent of the derivation: The derivation is
not complete unless all imperatives of duty can be derived from the
imperative Kant proposes as the only viable candidate for the supreme
principle of morality (87; cf. 88). Compared with this interpretation,
my own reconstruction in part III has been more text-immanent. Fur-
ther analysis of the differences between Kersteins and my analysis
of Kants attempts to derive the correct formulation of the moral law
would be beyond the scope of this paper.

VI. Critical Considerations

Let us assume for the sake of the argument that I have succeeded in
parts IIIV of my paper to plausibly reconstruct Kants arguments to
determine the content of the Categorical Imperative into a conclu-
sive argument. Then why does Kant fail to persuade us? Which of his
premises are problematic?
1. Surely Kants notions of the moral law and moral laws are al-
ien to us. For his predecessors22 and contemporaries this was different.
Kants talk of such law(s) becomes problematic when he exploits the
connotations of these expressions for certain philosophical purposes.
So, for example, the argument that he presents in the Preface to the
GMS for the benefit of a non-empirical understanding of ethics (GMS,
389) trades on his rhetoric about law. Because only the talk of a law
makes it seem compulsory, that it carr[ies] with it absolute necessity
does not hold only for human beings [but also] for other rational be-
ings (ibid. my insertion). By comparison, Kants mention of a com-
mand (e. g. thou shall not lie, ibid.) does not so clearly express an
(unconditioned) necessity to act in accord with it (my emphasis).
2. Kants presumption of the scope of validity for a moral law is
questionable as well. I would like to admit that the connotations of the
expression law, and the description of the law as the highest moral
principle, imply that this precept is valid for an encompassing circle of
creatures (see II.3). But whether this circle consists out of all rational
creatures is doubtful. It is even problematic that this law does not
hold for humans only but for all humans (ibid.). At most, it could
apply to all rational or competent in an xy-way or born with capaci-

22
The natural law theorists thought only a morality built around a specific concept
of law and obligation would be serviceable. Cf. Schneewind (1998, 518).
The Derivation of the Moral Law 115

ties xy humans. Since the use of rational is not without alternatives,


Kants move to other rational creatures (and hence to all rational
creatures) qua addresses of the law is also not necessary.
3. The way in which Kant introduces the definition of a law in
GMS I Duty is the necessity of action out of respect for law
makes clear that he presumes that there can only be one highest moral
principle. His description of the function of the law as the highest cri-
terion of moral judgment, and his distinction between the Categorical
Imperative and categorical imperatives (= laws) make clear that Kant
thinks of a hierarchy of moral principles with one supreme principle.
The corollary of this conception is a deductive model of moral justifi-
cation (from top to bottom).
This hierarchical conception of an ethical theory is optional. Some
philosophers have tried to plausibly argue that we can do without a
highest moral principle.23 Beyond this, there is a more plausible alter-
native to Kants conception of ethics. I am speaking of a holistic eth-
ics, such as Ludwig Siep24 has argued for recently, and as it has been
introduced to me by J. B. Schneewind. His essay Moral Knowledge
and Moral Principles25 is an ethical counterpart to W. V. O. Quines
epistemological holism in the concluding part of his Two Dogmas of
Empiricism.26
Kants hierarchical model of moral principles and his modell of
deductive reasoning in ethics raises a question about the justification
of a highest principle (the law). It cannot be grounded in a higher prin-
ciple. How then? It is doubtful that Kants justification of the moral
law in GMS III was successful or that his doctrin of the fact of pure
reason in the KpV does away with the problem.27 If, furthermore, one
is not impressed by attempts of ultimate justification (like those by K.-
O. Apel), an ethical holism will be even more attractive. According
to this position, moral laws do not necessarily need to be grounded
in higher laws, and ultimately in a or the law. Because principles, in-
cluding a highest principle, can also be justified from the bottom:
by demonstrating that they summarize lower-level principles, moral
commands and assessments in a fruitful way. Speaking more gener-
ally, moral principles, including the highest, can be justified by ap-

23
See Larmore (1987), Chapter VI, as well as Patzig (21983).
24
See Siep (2004), as well as Khl (2006), Chapter 6.4.
25
See Schneewind (1970).
26
See Quine (1953).
27
For a criticism of the Doctrine of the fact of pure reason, see Khl (2001), part VI,
as well as Khl (2006), Chapter 6.3.
116 Harald Khl

pealing to their inferential relations to the total inventory of an ethical


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

28
See Khl, 2006.
29
I am grateful to Christoph Horn and Dieter Schnecker for their efforts to realize
the translation of my paper.
The Derivation of the Moral Law 117

Literature

Kants writings

Kants Grundlegung zur Metaphysik der Sitten (= GMS; engl.: Groundwork


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

Other works

Allison, Henry E. (1996): On a presumed gap in the derivation of the cat-


egorical imperative, in: Idealism and Freedom, Cambridge, 143154.
Aune, Bruce (1979): Kants Theory of Morals, Princeton.
Khl, Harald (1990): Kants Gesinnungsethik, Berlin/New York.
Khl, Harald (2001): Moral und Klugheit. Rortys Kritik an einer kantischen
Unterscheidung, in: Deutsche Zeitschrift fr Philosophie 49, 1941.
Khl, H. (2004), Kants Grundlegung: neu ediert, kommentiert, interpre-
tiert in: Philosophische Rundschau 51.
Khl, H. (2006), Abschied vom Unbedingten. ber den heterogenen Charak-
ter moralischer Forderungen, Freiburg i. Br.
Larmore, Charles (1987): Patterns of Moral Complexity, Cambridge.
Paton, Henry. J. (1971): The Categorical Imperative: A Study in Kants Moral
Philosophy. Philadelphia: University of Pennsylvania Press.
Patzig, Gnther (21983): Ein Pldoyer fr utilitaristische Grundstze in der
Ethik, in: Patzig: Ethik ohne Metaphysik, Gttingen, 127147.
Quine, Willard V. O. (1953): Two Dogmas of Empiricism, in: Quine, From
a logical point of view, Cambridge: Harvard University Press, 2046.
Schneewind, Jerome B. (1970): Moral Knowledge and Moral Principles, in:
G. A. Vesey (Ed.), Knowledge and Necessity, London / New York.
Schneewind, Jerome B. (1998): The Invention of Autonomy, Cambridge.
Schnecker, Dieter / Wood, Allen W. (22004): Kants Grundlegung zur Me-
taphysik der Sitten. Ein einfhrender Kommentar, Paderborn.
Siep, L. (2004), Konkrete Ethik, Frankfurt a. M.
Tugendhat, Ernst (1993): Vorlesungen ber Ethik, Frankfurt a. M.
Wood, Allen. W. (1990): Hegels Ethical Thought, Cambridge.
Wood, Allen W. (1999): Kants Ethical Thought, Cambridge.
Groundwork
II
Marcus Willaschek

Practical Reason
A commentary on Kants Groundwork of the
Metaphysics of Morals (GMS, 412417)1

1. Introduction

On pages 412417 of the GMS, Kant introduces his conception of the


will as practical reason and the closely related distinction between
hypothetical and categorical imperatives in a rather concise manner.
In the following, I will address the much discussed definition of the
will (4: 412), the concepts of both the holy will (4: 412 und 414) and the
imperative (4: 413414), as well as the differences among the different
kinds of imperatives (4: 414417), while explaining Kants central con-
cepts and theses. Prior to this, however, I will present a short overview
of Kants general line of argument.

2. Overview

Practical reason is the ability to act rationally that is, the ability of a
person to rationally coordinate her goals and ends and to orient her ac-
tions according to these rationally set ends. Kant identifies this ability
with the will, or with the ability of a rational being (more precisely, a
being possessing reason) to be the cause of its conduct through its own
representations (mental states). In other words, having a will means
being able to act in accordance with ones own rational representations.
Within the concept of practical reason, Kant distinguishes pure
from empirically qualified practical reason a distinction that he first
1
For their valuable advice and criticism, I would like to thank Alexander Bagattini,
Steffi Schadow, Dieter Schnecker as well as the participants of the preparatory
conference in Bonn.
122 Marcus Willaschek

makes explicit in the KpV, and that is already anticipated (and im-
plicitly presupposed) in the GMS (cf. 4: 389, where Kant speaks of a
practical pure reason as well as of the question of how pure rea-
son can be practical, 4: 458; cf. 461). God and the angels (if they ex-
ist) possess exclusively pure practical reason; their decisions are in no
way influenced by subjective factors, and particularly not by sensible
(sinnliche) inclinations, but solely by objective, and specifically moral,
rational grounds. Each of their wills is a holy will, which necessarily
and exclusively seeks the moral good.
Humans, on the other hand (as the only rational beings of whose
existence we have knowledge), are influenced by their naturally and
socially conditioned wishes and inclinations in many ways. Their wills
are, therefore, not holy, for when it accommodates their interests they
can decide to act immorally. Their practical reason is therefore em-
pirically qualified: as instrumental rationality, it serves to satisfy em-
pirically-given inclinations in the most effective way possible. If, on
the other hand, as Kant supposes, moral principles hold without ex-
ception and necessarily for all beings who possess reason, thus also
for humans, then it must be rational for every rational being to follow
these principles, independent of its respective inclinations or wishes.
Accordingly, humans must not only have empirically qualified reason,
but also pure practical reason; that is, they must be able to orient their
action according to moral principles when this would not further, or
would even contradict, their own interests.
The human will is thus distinguished from a holy will in two ways:
first, our decisions do not necessarily accord with that which would be
rational to do, and second, rational for us does not only mean moral-
ly good, but also useful for the satisfaction of subjective inclinations.
The first difference leads to the concept of an imperative, the second to
the distinction between hypothetical and categorical imperatives.
Imperatives are propositions in which what is good and reasonable
to do is expressed in the form of a command. Such commands are not
directed at a being with a holy will that necessarily does what is ration-
al, but rather at beings such as humans, who can act rationally but who
do not necessarily do so (i. e., those for whom both rational and irra-
tional actions are possible in any given situation). Most people are thus
capable of doing something because they have recognized it as rational
(e. g. taking care of their health or paying their taxes), but this rational
insight does not necessarily motivate them to a corresponding course
of action, because their inclinations (e. g. towards indolence) or their
self-interest may oppose it. The principles of rational action, therefore,
Practical Reason 123

appear for human beings as commands that express themselves in an


ought: (From a rational point of view) you ought to live healthily or
(From a rational point of view) you ought not cheat on your taxes.
This ought is not an expression on the part of a power or convention
(as is the case when one says that someone ought to follow the com-
mand of a superior or ought to greet ones neighbors), but an expres-
sion of a specifically rational form of motivation, which Kant describes
as rational necessitation (Ntigung). This necessitation serves as a
motive for doing what is rational even when we find it difficult (and
must therefore force ourselves to do it). The more rational one is in a
practical respect, the more fully developed will this motive be.
By prescribing an action, imperatives distinguish it as in some way
good. Corresponding to the distinction between a pure and a practi-
cally qualified reason, there are also two kinds of imperatives: those
that demand an action because it is good as a means to the effec-
tive satisfaction of inclinations and needs, and those that prescribe
an action because it is rational per se or good in itself, without
taking into account the ends of the agent. The former are hypotheti-
cal imperatives and the latter, categorical imperatives. Kant divides
the hypothetical imperatives again into two classes, depending on
whether they prescribe an action as a means to a possible or an actual
end. Thus Kant arrives at a tripartite division of imperatives, giving
several alternative descriptions of each of the three classes, depend-
ing on which of their characteristics he wants to emphasize: (1) prob-
lematically practical principles (= imperatives/rules of skill, technical
imperatives), (2) assertorically practical principles (= rules/counsels
of prudence, pragmatic imperatives), (3) apodictically practical prin-
ciples (= imperatives/laws of morality, moral imperatives). (1) and (2)
are the hypothetical, (3) the categorical imperatives.

3. The definition of the will (4: 412) 2

Kant defines the will as the capacity to act in accordance with the
representation of laws, that is, in accordance with principles, and he
identifies this capacity with practical reason. This passage has given
rise to a variety of different interpretations (cf. Laberge 1989; Tim-

2
The following translations of passages from Kants works are based on the Cam-
bridge Edition of the Works of Immanuel Kant (New York: Cambridge University
Press, 1992-).
124 Marcus Willaschek

mermann 2003, 9), and even if one takes the context into account
the passage remains ambiguous. Kant begins with this observation:
(1) Everything in nature works in accordance with laws.

This claim follows from the regularity conception of causation that


Kant inherited from Hume (cf. Willaschek 1992, 3443). According
to this view, a at time t0 is the cause of b at time t1, if a is of type A
and b is of type B, such that from things or events of type A at t0 there
always follow things or events of type B at t1. In that case, If A at t0,
then B at t1 is a universally valid law. Thus the claim holds analyti-
cally that all things in nature work in accordance with laws, or that
their causality functions according to laws of nature.
(2) Only a rational being has the capacity to act in accordance with the
representation of laws, that is, in accordance with principles, or has a
will.

The will is thus a capacity (i. e., an ability) of a rational being to act
in a specific way, namely in accordance with the representation of
laws. But what does it mean to act in accordance with the represen-
tation of laws? Which laws are meant here? It is exactly with regard
to this question that interpretations diverge.
In order to understand Kant correctly here, it will be helpful to
review the connections among the concepts life, faculty of desire (Be-
gehrungsvermgen), and will (cf. with regard to this and what follows
Willaschek 1992, 8290). Life, as Kant writes in the KpV, is the
faculty of a being to act in accordance with the laws of the faculty of
desire. The faculty of desire is the faculty of this being to be, through
its representations, the cause of the actuality of the objects of these
representations (5: 9; cf. already in 2: 327 as well as 4: 544). As is
briefly stated in the MdS, The capacity of a being to act in accord-
ance with its representations is called life (6: 211). Here, to act has
the broad meaning of the Latin agere: it means the same as to bring
about or to cause (cf. Gerhardt 1986). Kant understands life as the
purposeful spontaneous activity (Selbstttigkeit) of an organism (or
in the case of a living God, cf. 3:421, the activity of a non-material
entity) in accordance with the Aristotelian-Leibnizian concept of en-
telechia. This spontaneous activity, however, is conceived (at least in
the case of physical beings) in a distinctively modern way as a lawful
and purely causal association. A living being also works according
to laws: it has a representation of what it desires and this representa-
tion motivates it to act in a way that leads to the realization of the
Practical Reason 125

desired object or goal. The causal connection between representation


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

the repetition of the words according to (nach) creates difficul-


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

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

The expression practical reason takes up the Aristotelian language


of nous praktikos, but has no modern-language predecessor in the
18th century (cf. Beck 1974, 265, note 14). Hence Kant was free to
define this expression according to his own needs. Nevertheless, Kant
apparently does not want to give a stipulative definition, as he justi-
fies the identification of the will and practical reason with the remark
that reason is required for the derivation of actions from laws. One
might grant Kant that derivation is the task of reason; at another
point he expressly identifies reason with the faculty of the determi-
nation of the particular through the universal (the derivation of prin-
Practical Reason 127

ciples) (20: 201). However, that means only that one cannot have a
will without also having reason (cf. 4: 427: The will is thought as a
capacity to determine itself to action in conformity with the represen-
tations of certain laws. And such a capacity can be found only in ra-
tional beings.) This does not mean, however, that the will is nothing
other than practical reason. For this, rather, two claims would have
to hold, claims that Kant implicitly presupposes: first, that the will (as
the capacity to act according to the representation of laws of reason)
is nothing other than the capacity of the derivation of actions from
laws, and second, that for this capacity, practical reason is not only
necessary (Kant: is required), but also sufficient.
The talk of a derivation of actions refers to Aristotles practical
syllogism in which a general major premise (Everything sweet must
be tasted) and a minor premise (This is sweet) are followed direct-
ly by the execution of the rationally commanded action (cf. EN 1147a
25 ff.). Kant also assumes that practical reason does not only consist
in deriving propositions or opinions about actions but also the actions
themselves: practical reason is the ability to orient ones behavior ac-
cording to the laws of reason. That is, it is not only the ability to rec-
ognize what is rational (the so-called principium diiudicationis), but
to also do it (principium executionis). However, Kants conception
differs from the Aristotelian view in one important respect: practi-
cal reason does not produce the derived actions directly (i. e., sim-
ply through insight into their rationality), but indirectly, via a feeling.
This feeling, however, which in the case of action according to moral
laws Kant terms respect for law (cf. 4: 400, 440) or moral feeling
(4: 460), is itself of rational origin, for it arises through insight into the
rationality of moral laws (cf. 4: 460, 5: 76). In Kants view practical
reason delivers both objective reasons and the subjective incentives
of rational action and is thus nothing other than the will or the ability
to act rationally.
This identification of practical reason and the will raises a prob-
lem, however, when Kant directly afterwards (and in many other pas-
sages) speaks of reason determining the will (more or less effec-
tively). How can this be if the two are identical? The answer lies in
the ambiguity of the expressions will and practical reason (cf. on
this and the following Willaschek 1992, 4853). In the GMS, as we
have already seen, Kant defines will and practical reason as the
ability to act rationally. What is at issue there, however, is a complex
ability. It includes on the one hand the ability to bring the totality of
ones wishes and convictions into a rational order (which guarantees
128 Marcus Willaschek

consistency and completeness). Kant calls this ability reason and, in-
sofar as it concerns primarily wishes, ends, and grounds for action,
practical reason or will: The will is therefore the faculty of desire
considered not so much in relation to action (as choice (Willkr) is),
but rather in relation to the determining ground of this choice of ac-
tion. The will itself actually has no determining ground, but insofar as
it can determine choice it is instead practical reason itself (6: 213, my
emphasis). In this passage from MdS Kant does not understand will
and practical reason as he does in GMS in the broader sense, as the
ability to act rationally, but rather in the narrower sense, as the abil-
ity to produce rational determining grounds for action. On the other
hand, practical reason in the broad sense also includes the ability to
actually act according to these rational grounds of action. This ability
is what Kant calls free choice (freie Willkr). Choice is the capacity
to act according to ones own representations, which are connected
with pleasure and displeasure (Lust and Unlust) (6: 413); choice is free
insofar as pleasure and displeasure do not determine behavior but
rather serve as a basis for the rational establishing of ends (cf. KrV A
802/ B 830 and, with a slight change of emphasis, 6: 413).
Practical reason or will in the broader sense as the ability to act
rationally includes, first, practical reason or the will in the narrower
sense (that is, the ability to bring ones wishes and ends into a con-
sistent scheme of grounds for action) and, second, free choice (that
is, the ability to do what is rationally desired). The will, choice, and
practical reason are therefore not distinct causal instances or sub-
ject-like homunculi, but instead aspects of the complex ability to act
rationally. When Kant speaks of the fact that reason determines the
will or choice, he means that practical reason in the narrower sense
(the capacity to set rational ends) provides the determining grounds
for the will in the broader sense (the ability to act rationally) or free
choice (the ability to act rationally). That reason determines the will
thus means depending on emphasis either that one actually does
what is rationally desired, or that what one does is actually rational.
When Kant, however, equates reason and the will, then he means ei-
ther (in the broader sense) the ability to act rationally or (in the nar-
rower sense) the ability to establish ends rationally. Although Kants
use of words is confusing and not always uniform, his conception of
the will as practical reason is in fact uniform and thoroughly compre-
hensible.
A further meaning of the term will in Kant has not yet been
mentioned: namely, according to Kant the will is not only the ability
Practical Reason 129

to act according to laws or principles and thus rationally, but rather, as


free will, also the law-giving faculty (cf. 6: 226). This means that a free
rational being can not only orient its decisions and actions according
to the rational principles that it is given and adherence to which serves
the satisfaction of subjective inclinations, but that it also has the abil-
ity to give itself these principles according to which it acts and de-
cides. Kant calls this ability autonomy or self-legislation (Selbstge-
setzgebung) (4: 440 ff.) The fact that a rational being gives itself the
principles according to which it acts is obviously not intended to mean
that the recognition of these principles is the result of a conscious
and deliberate decision that could also have turned out otherwise. It
means that the rationality of a rational agent consists in the recogni-
tion of these principles; they are not given to her by a foreign authority
(heteronomy), but are rather an expression of her own reason. While
Kant in the GMS takes a heteronomous will to be conceivable (4: 441),
in the MdS he seems to almost identify the will with the ability of self-
legislation (6: 226).

4. The concept of a holy will (4: 412 and 414)

Kants transition from the definition of the will to his conception of


the imperative is constituted by the concept of the holy will, which he
introduces in two steps: If reason infallibly determines the will, the
actions of such a being that are recognized as objectively necessary
are also subjectively necessary, that is, the will is a capacity to choose
only that which reason independently of inclination recognizes as
practically necessary, that is, as good. (4: 412). Soon afterwards, this
necessarily good will turns out to be a holy will: A perfectly good
will would, therefore, stand just as much under objective laws (of the
good), but it could not on this account be represented as necessitated
to actions in conformity with law, since it of itself, according to its
subjective constitution, can only be determined by the representation
of the good. Hence no imperatives hold for the divine will and in gen-
eral for a holy will: the ought is out of place here, because volition
is already of itself necessarily in accord with the law (4: 414). A holy
will is thus a will which of itself, according to its subjective constitu-
tion can do nothing other than the rational and good. This concep-
tion raises a number of questions, only two of which I want to consider
more closely here: (1) Is the conception of a will that wills only the
objectively rational, i. e., the morally good, consistent? (2) And how
130 Marcus Willaschek

does this concept of a perfectly good will relate to that of the good
will that is the topic of the beginning of the fi rst section?

(1) A holy will, in contrast to a human will, is a capacity to choose


only that which reason independently of inclination recognizes as
good; such a being can be determined only through the representa-
tion of the good. These formulations suggest that a being can only
have a holy will if its will is not affected by inclinations, but in-
stead acts exclusively according to the objective principles of reason.
This becomes clear, too, through comparison with the human will
for which, in contrast to the holy will, reason alone does not ade-
quately determine the will and thus which is also subject to subjec-
tive conditions (certain incentives) (4: 412). Thus the holy will would
be subject to no subjective conditions, such that reason alone would
determine it sufficiently. Similarly, Kant writes in the KpV that the
moral law has for humans the character of an imperative, because
in them, as rational beings, one can presuppose a pure will but, as
beings affected by needs and sensible motives, not a holy will, that is,
a will that would be incapable of maxims conflicting with the moral
law (5: 32). In this passage as well it sounds as if a being with a holy
will may not be affected by needs and sensible motives, because
it is this state of being affected that prevents humans from attaining
this holy will.
Such a view would be awkward, however, since a being without
inclinations would be incapable of action. Rational principles alone
are not sufficient to establish how one should act in a certain situa-
tion. With regard to the principle of instrumental action (He who
wills the end, also rationally wills the necessary means to that end)
this is obvious: it presupposes that one already has ends; what these
ends are does not only depend on rational considerations, but also on
that to which one has an inclination. Even moral principles, and
above all the categorical imperative, are by themselves not sufficient
to be taken to show a certain action to be rationally commanded. Af-
ter all, one is supposed to test, using the categorical imperative, if the
maxim by which one intends to act could be a universal law. Maxims,
however, are subjective rules of action, which reason determines ac-
cording to the conditions of the subject (more often his ignorance
or also his inclinations) (4: 421; my emphasis). A being that has no
inclinations (and is also not subject to any other subjective condi-
tions) has also, therefore, no maxims, and thus nothing to which
it could apply the categorical imperative. Thus, even a being with a
Practical Reason 131

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

irrational and immoral. (The inclinations of a divine being divine


love, for example would not be sensibly conditioned like ours, but
would have non-sensible origins.)

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

3
Cf. on this subject Allen Woods contribution to this volume.
Practical Reason 133

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

(in the form of inclinations) to do so, but rather does the good, under
any conceivable circumstances, gladly. (To what extent holiness of
the will can be demanded of humans, as Kant claims in the KpV and
the MdS, is an additional problem that cannot be treated here; cf. on
this Allison 1990, ch. 9).

5. The concept of the imperative and the division of the imperatives


(4: 413417)

The representation of an objective principle, insofar as it necessitates


a will, is called a command (of reason), and the formula of that com-
mand is called an imperative (4: 413). According to this definition,
an imperative is the linguistic expression (formula) of a command
of reason, which itself is the representation of an objective prin-
ciple or law that necessitates the will. (Kant does not always hold
himself to this usage, though, and often uses the terms imperative,
command, law, and objective principle synonymously). All im-
peratives, as Kant continues, are expressed by an oughtor, more
precisely, can be expressed by an ought, since Act in such a way
that the maxim of your action could be a universal law is an impera-
tive in Kants sense, even though it is not expressed by an ought. It,
however, could easily be reformulated accordingly (You ought to act
in such a way, so that the maxim of your action ).
The necessitation of the will that is expressed in the imperative
is that specific rational form of normative motivation which consists
in the awareness that one ought to do what is rational to do. Kant
does not try to ground the connection between reason and normativ-
ity further: he presupposes here that a being that is endowed with
reason but not always rational ought to be rational (cf. Kants discus-
sion of the usefulness of reason at the beginning of GMS I; 4: 395/6).
He does, however, explicate the normative and evaluative dimension
of reason a bit later by highlighting the connection between ought
and good: imperatives say that to do or to refrain from something
would be good []. Practical good, however [in contrast to the pleas-
urable; M. W.], is that which determines the will by means of rep-
resentations of reason, hence not by subjective but objective causes,
that is, from grounds that are valid for every rational being as such
(4: 413; cf. 414). According to Kant, every imperative implies a value
judgment, but one that is rationally justifiable and thus claims to hold
for all rational beings. (Here we are reminded of the famous thesis of
Practical Reason 135

the KpV concerning the priority of the moral law over the concept of
the good; cf. 5: 5771.)
Imperatives, however, do not hold for all rational beings, because,
as we have already seen, a holy will is not subject to any normative
necessitation: the ought is out of place here, for Kant, because
volition is already of itself necessarily in accord with the law (4:
414). Kant is famous for the thesis that ought implies can: some-
one who ought to do A, also can do A (cf. 5: 30). In contrast, his
claim here is: ought implies being able to do otherwise. Someone
who ought to do A, can also not do A. (It does, for example, seem as
senseless to command someone to be identical with himself, as it is
to command him to not be identical with himself; he cannot stop do-
ing the first, and he cannot do the second. In both cases the ought
is out of place.)
Kants view can be summarized by saying that imperatives are
rational directives (or directives of reason). Their normative binding
force is based on the fact that it is rational to follow them. Their pre-
scriptive character is based on the fact that they are directed toward
beings that can act rationally, but do not necessarily always do so.
Kant continues: All imperatives command either hypothetically or
categorically. The former represent the practical necessity of a possi-
ble action as a means to achieving something else that one wants (or
that it is at least possible for one to want). The categorical impera-
tive would be one that represented an action as objectively neces-
sary in itself, without reference to another end (4:414). On the basis
of the analytic relationship between ought and good, Kant can
express the distinction between hypothetical and categorical impera-
tives in the following way as well: Now, if the action would be good
merely as a means to something else, the imperative is hypothetical;
if the action is represented as in itself good [], then it is categorical
(4:414).
The distinction between hypothetical and categorical imperatives
thus has an impact on the question of why it is good and rational to act
in the way commanded by the imperative. This distinction does not
necessarily show in the linguistic form of the imperative: Invest your
money cautiously! is, viewed linguistically, a categorical sentence but
a hypothetical imperative, for a cautious investment is obviously good
merely as a means and is not sought after for its own sake. If you
owe money, you ought to pay it back is linguistically a conditional
sentence but nevertheless a categorical imperative, because it repre-
sents an action as objectively necessary in itself, without reference to
136 Marcus Willaschek

another end (on the problem of the distinction between hypothetical


and categorical imperatives cf. Brinkmann 2003, 1951).
As mentioned above, Kant divides hypothetical imperatives fur-
ther into rules of skill and counsels of prudence and thus arrives
at a tripartite division of the imperatives:
(1) Rules of skill prescribe an action under the condition of an end
concerning which it is left open whether one actually follows the end
and whether it is rational to do so: If you want A, then you ought to
do B (as a necessary means thereto)! Nothing changes in this hy-
pothetical structure when the condition is met and one thus actually
pursues the end in question: Kant does not provide for the derivation,
using modus ponens, of a categorical imperative, You ought to do
B, from the hypothetical imperative If you want A, you ought to do
B and the empirically-contingent observation You want A. The
imperative thus derived, You ought to do B, remains a hypotheti-
cal one, as its validity depends upon the condition that the person in
question wants A.
(2) For this very reason counsels of prudence are also merely hy-
pothetical imperatives. They recommend an action as a means to an
end that the addressee of the imperative actually has: Because you
want A, you ought to do B (as a necessary means thereto)! Because
imperatives are expressions of principles of reason and thus have uni-
versal validity, the only end in question here is one that all rational be-
ings actually pursue (insofar as, as Kant notes, imperatives apply to
them as dependent beings, 4: 415). The only such end is that of ones
own happiness: every person wants to be happy. (Kant defines hap-
piness as a state in which everything goes according to ones wish and
will (5: 124)). This pursuit of happiness is a natural fact it is part of
what constitutes humans as rational living beings; thus the purpose
of happiness is something we have by a natural necessity (4: 415).
It can thus be presupposed to exist in every human being. Neverthe-
less, this purpose cannot serve as a basis for a categorical imperative,
for different people find happiness in very different things. No means,
therefore, can be identified that would be actually necessary for every
person in order to become happy: Because you (like every rational
natural being) want to be happy, you ought to live healthily, or, more
concisely, Live healthily! is a rational counsel, but with some luck
(Glck, fortuna) one can also be happy (glcklich, beatus) if one does
not heed it.
(3) Only moral (categorical) imperatives hold without exception
for all (dependent, i. e., capable of being sensibly motivated) rational
Practical Reason 137

beings. They do not concern the material of the action (thus do not
depend on which ends one is pursuing), but concern rather the form
and the principle from which it [the action] itself follows (4: 416). The
form of the action is, as has already been shown in the first section
of the GMS, the conformity to law as such, which serves the will as
its principle (4: 402). In this manner, Kant arrives at the categorical
imperative in the singular (never to act except in such a way that
I could also will that my maxim should become a universal law, 4:
402; cf. 421). According to Kant, out of this universal categorical im-
perative, as their principle, all imperatives of duty, i. e., all particu-
lar categorical imperatives such as the prohibition on lying, stealing,
etc., are derived (cf. 4: 402). What unites these distinct categorical
imperatives is the fact that one ought to act in the prescribed way not
for an actual or possible purpose, but simply because the action is
good in itself. To prove that there actually are such actions (and
thus also the categorical imperative that commands them) is the aim
of the third section of the GMS.

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesam-


melte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902) (abbreviated
as AA). The Critique of Pure Reason will be cited according to the A/B
pagination from the first and second editions. All textual references are to
The Cambridge Edition of the Works of Immanuel Kant (Cambridge Univer-
sity Press, 1992).
GMS Grundlegung zur Metaphysik der Sitten, AA, IV
KdU Kritik der Urteilskraft, AA, V
KpV Kritik der praktischen Vernunft, AA, V
KrV Kritik der reinen Vernunft, AA III, IV
MAN Metaphysische Anfangsgrnde der Naturwissenschaft, AA, IV
MdSR Metaphysik der Sitten, Rechtslehre, AA, VI
MdST Metaphysik der Sitten, Tugendlehre, AA,VI
RGV Die Religion innerhalb der Grenzen der blossen Vernunft, AA, VI
TG Trume eines Geistersehers, AA, II

Other works
Allison, Henry E. (1990): Kants Theory of Freedom, Cambridge.
Beck, Lewis White (1974): Kants Kritik der praktischen Vernunft. Ein
Kommentar, Mnchen.
138 Marcus Willaschek

Brinkmann, Walter (2003): Praktische Notwendigkeit. Eine Formalisierung


von Kants Kategorischem Imperativ, Paderborn.
Gerhardt, Volker (1986): Handlung als Verhltnis von Ursache und Wir-
kung. Zur Entwicklung des Handlungsbegriffs bei Kant, in G. Prauss
(Hg.), Handlungstheorie und Transzendentalphilosophie, Frankfurt/M.
Laberge, Pierre (1989): La dfinition de la volont comme facult dagir se-
lon la reprsentation des lois, in O. Hffe (Hg.), Grundlegung zur Meta-
physik der Sitten. Ein kooperativer Kommentar, Frankfurt/M., 8396.
Timmermann, Jens (2003): Sittengesetz und Freiheit. Untersuchungen zu Im-
manuel Kants Theorie des freien Willens, Berlin.
Willaschek, Marcus (1992): Praktische Vernunft. Handlungstheorie und Mo-
ralbegrndung bei Kant, Stuttgart/Weimar.
Wood, Allen (2006): The Good Without Limitation, in this volume.
Bernd Ludwig

Kants Hypothetical Imperatives (GMS, 417419)

I.

In German language (and also in many others) there is a grammati-


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

1
This kind of insertion is, for instance, used when trying to take into account the
peculiarities of the addressee, like: If you want to play the piano well, YOU have to
practice (in opposition to your advanced brother).
140 Bernd Ludwig

as to others. But as statements these latter imperatives do not contain


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

II. Kants Terminology

The conception of an objective principle, in so far as it is obligatory for a


will, is called a command (of reason), and the formula of the command is
called an imperative. (GMS, 413)

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

Any given imperative can be examined not only with respect to the
content, but also with respect to the conditions under which it is ac-
tually commanding: If certain conditions would not apply anymore,
would the necessitation, the ought, vanish and hence the imperative
as such turn invalid or void? These conditions may either be changed
by the addressee of the ought, or not. In the second section of the
Groundwork, Kant is only interested in the former kind of conditions.
Consequently, he distinguishes imperatives as follows:
Kants Hypothetical Imperatives 141

Now all imperatives command either hypothetically or categorically. The


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

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

There are corresponding formulations in the Reflections, too (cf. R


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

Whoever recognises himself as the addressee of the two imperatives


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

matical imperative actually is an imperative, and, since statement (1)


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

Hence, applied to practical rules, the terms categorical and hypo-


thetical do generally not refer to a possible (judgment-)form of an
imperative, but to the mode of the determination of will only. For this
reason, grammatically the two terms are no adjectives which belong
to imperative, but they determine the verb command attributively.2
There are neither hypothetical imperatives nor categorical im-
peratives sensu stricto, 3 but only hypothetically- or categorically-
commanding statements respectively, which, in so far as they com-
mand, are imperatives (or are called so) and which all have the same
form: Do X!. Hypothetical and categorical are not placeholders
in an imaginary Kantian table of the imperatives4, but just name
the relevant relation in a judgment about the practical character of
a statement given as a practical one. Strictly speaking, hypothetical
and categorical are terms that belong to the meta-theory of impera-
tives. They dont classify imperatives, but they appear in judgements

2
Also KpV, 31: represented a priori as a categorical proposition, KpV, 32, and
MdST, 404: commands categorically, similar GE, 282 and 416. In German, the
terms kategorisch and hypothetisch stand for adjectives and adverbs both.
3
That is to say, the semantic relations fundamentally differ here from the ones in the
case of hypothetical judgments: There it is not the character of the judgment which
is contingent, but solely the asserting of (one of) the singular judgments. Hence, hy-
pothetical judgments are always judgments, but hypothetical imperatives sensu
stricto were only imperatives if the antecedent were true.
4
In analogy to the Table of Judgements (in KrV, B95) M. Moritz f. i. misses disjunc-
tive imperatives (s. 9 ff.).
Kants Hypothetical Imperatives 143

which state that and why a particular statement is an imperative


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

5
The hypothetical judgment is not made of one judgment and one imperative but of
two judgments (KrV, B98).
6
See Patzig (1973, 209 f.).
7
The inventor of this example, Hare (1971), shows the trouble one gets into, if one
sticks to this misguided (and not Kantian) analysis. A simple reduction ad absur-
dum shows that any expression of the form EO(M) wont work as a formal ver-
sion of Who wants the end, ought to want the necessary means, with E standing
for Person p wants the end , M for Person p wants the means and O( ) for
something like It has to be the case that . We only have to apply contraposi-
tion twice: let ~m~e stand for the proposition [underlying EO(M)] that the
means is necessary for the end in question. Since ~m~e is equivalent to ~~e
~~m, ~MO(~E) as well has to be the expression of an imperative (by the very
same reason as for which EO(M) is one). But ~MO(~E) is equivalent to
~O(~E)M again and this latter is unquestionably nonsense: nobody wants the
means in all cases where there is no obligation/necessity to abandon the end. Since
the only non-trivial step in the argument above is the transition from ~m~e to
EO(M), this must be inadequate.
144 Bernd Ludwig

non datur). However, then there only came off an infinite regress of
hypothetical conditions but never a necessitation, not to mention that
it remains entirely clouded what could be the content of the antecedent
of each single one of the hypothetical imperatives in question.
Therefore, according to the opening considerations and to Kants
proper definitions, a statement like: If you want to be happy, avoid pain!
is not an imperative at all,8 but simply a truth-apt judgment whichcontains
an imperative. At the most, it is a kind of advice which one usually uses
to argue for the relevance of the imperative-formula Avoid pain! And
further, it is not even a hypothetical judgement (that is, a conditional),
since it does not answer the question: Under which conditions shall I
avoid pain?, but rather: For what reasons shall I avoid pain?. Only if
one kept on asking after having received the answer (Because you want
to be happy.) and thus discovered being free of pain to be a necessary
condition of human happiness, eventually a hypothetical judgment came
into play: If you dont avoid pain, then you cannot be happy.
Although at first sight Kant himself speaks a bit loosely in this con-
text and seems to equate hypothetical imperatives with counsels
(GMS, 416,19; but then 24 f.), one should be aware of the fact that
he gave a clear and consistent exposition of the basic concepts afore,
which make it possible for the reader to specify such stenographical
formulations and thus explicate them in a coherent and adequate man-
ner. We will see that Kant isnt very particular about the terminology
in this respect when looking at the alleged analyticity of hypotheti-
cally-commanding imperatives. But first, the basic idea of the proof of
the possibility of hypothetical imperatives must be recapitulated.

III. The Possibility of hypothetical Imperatives

Lets start again with some elementary considerations independent of


Kants writings:
1. A man9 who wants to realise an end (E) which cannot be real-
ised without any help on his part, is necessitated to do10 something
definite, viz. to apply the necessary means (M). This is completely
8
G. Patzig pointed that out in 1971 already.
9
Although the following does apply to women as well, I dont consider this in the
main text, just to avoid grammatical redundancy or discrimination by using the
male pronouns pars pro toto.
10
But not to want anything definite: if I want to step through the door, I must (rebus
sic stantibus) push the door handle. But I can push the door handle without wanting
to push it (e. g., if I push it accidentally).
Kants Hypothetical Imperatives 145

independent of his being rational in a narrower sense. If he doesnt


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

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

11
Who does not want to practice the piano and knows that it cannot be done without,
simply cannot do anything he himself could understand as expression of his will to
become a pianist, i. e. it is not clear what is meant by he wants to become a pianist
in this case.
12
For reasons of clarity, the three assertions are represented in an abridged way.
146 Bernd Ludwig

If the means-end-formula is analytic, its truth can be understood


by the analysis of the meaning of the words contained in the judg-
ment, i. e. of the Kantian concepts means, end and willing/will.
Moreover, this insight should result without any additional idiosyn-
cratic assumptions according to Kants assurance that there is no
need of any exceptional discussion in this context. In the Groundwork
one can find the following about the first of these concepts:
On the other hand, that which merely contains the ground of possibility
of the action of which the effect is the end, this is called the means. (GMS,
427)

Using the explicit definitions of the two other concepts provided in


the Metaphysics of Morals from 1797, one reaches the goal directly.
An end is an object of choice (of a rational being) through the representa-
tion of which choice is determined to an action and to bring this object
about. (MdST, 381)

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

This Willkr (choice) corresponds to Wille in the means-end-for-


mula of the Groundwork. To answer the question of the analyticity of
this formula one has to be aware that Kants concept of Willkr (1785:
of Willen, GMS, 449) includes consciousness of the ability to bring
about its object (MdSR, 213). But under this presupposition the will-
ing of the end almost naturally comprises the willing of every ground
of possibility of the action (GMS, 427) identified as necessary.13 For if

13
See also the quotation from GMS, 417, above: but if I know that it is only by this
process Surely, one feels already necessitated if one only believes. As expected,
Kants Hypothetical Imperatives 147

I desired an end without wanting the means identified as necessary, I


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

Kant does nowhere in his argument assume that the relevant causal judgments are
true.
14
However, it is hard to judge singular cases: I cannot live that way usually just
means: I do not want to live that way, for the price is too high still, the bystanders
sometimes consent that one cannot live that way. But then it is all about what we
want to expect of each other and not about where the natural necessities begin.
148 Bernd Ludwig

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

Second, ad 3): Necessitation and Necessity


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

The necessitating character of the hypothetically-commanding im-


peratives, which goes beyond the insight that who wants the end also
wants the means, as long as the reason decisively determines his actions,
shows itself in two cases: 1. If the one, whose will is not (yet) determined
decisively by the reason in the relevant respect, is necessitated to find
the means (without he cannot realise the end), i. e. get his reason the
relevant influence. Or 2.: If the one, who already knows about the nec-
essary means, feels necessitated to decide whether to do without the
realisation of the ends or to use the means. What he surely cannot do
(and hence, which he cannot be ordered to omit) is both to want the end
and at the same time not want the means already known to be neces-
sary. Hence in this second case he is not necessitated to do something,
but only to make a decision: He is necessitated to give up his end
(GMS, 420), or to use the means in question. Though, who wants end
and means likewise is not necessitated at all, just like Kants saints: For
those only laws apply which are no imperatives (GMS, 414).
In this sense, a hypothetically-commanding imperative indeed
necessitates, regardless of the fact that everyone who wants the end
wants the means as well as long as the reason decisively determines
his willing. For, who wants the end and knows what the necessary
means is (sc. as long as ) is not necessitated to want the means, he
just wants it. Who does not want the means is not necessitated not to
want the end, he (as long as ) just does not want it. The necessita-
tion only shows itself when incompatible wishes face each other. In
that case only one of them can become a will, and if that happens, the
other wish looses its force and reason has gained decisive influence.
Consequently, the ought, which is expressed by the hypothetical
imperative in view of the means in question, is always a hypothetical
Kants Hypothetical Imperatives 149

willing at the same time. However, it is misleading to talk about ra-


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

15
This is approximately what Schnecker/Wood (22004, 119) propose as an interpre-
tation.
150 Bernd Ludwig

egorical imperatives respectively.16 The necessitation in hypothetical-


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

16
This lack of clarity fi nds an echo in Kants ambiguous use of the word Gebot. In
the beginning he defi nes all imperatives as formulas of Gebote (GMS, 413), but
later (GMS, 416) he suggests that only categorical imperatives are such.
17
Or muss (must), as in Patzig (1973, 215). At best Who wishes the end, shall will
the means would render an understandable assertion. In Kant we mostly fi nd: I
ought to do something, because I want something (GMS, 441,11; 444, 4 and 11; the
only exception: GMS, 419,9).
18
Though Downie (1984, 486) consents to that, he tries to show that there is a moral
obligation of practical consistency alike to a liability behind hypothetical impera-
tives.
Kants Hypothetical Imperatives 151

(KdU, 173; MdSR, 222; see also AA XX, 200), which might contain
a determination of will. As one can easily see now, they just neces-
sitate us, when we know about the means to an end, to decide for the
means or against the end.19 Simply and solely categorical imperatives
necessitate us to decide for or against something definite, hence, to do
something definite.

IV. Hypothetical vs. Absolute Necessitation Achenwall

Still, the perhaps surprising discovery that hypotheticallynecessitat-


ing imperatives do not make particular actions necessary but rather
decisions is not just a modern interpretation of Kant, but can already
be found in Gottfried Achenwalls (whom Kant thought so highly of)
terminological fixing of an obviously ready-established usage:
Si determinatur alter ad unicum; coactio vocatur absoluta: si ad unum ex
quibusdam determinatis eligendum; dictur hypothetica. (Elementa Iuris
Naturae, 1750, 70)20

Achenwalls remark sheds new light on the distinction of imperatives,


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

19
Accordingly something like O(~M ~E), that is O(EM) might be closer to
the point than EO(M) [compare note Fn. 6 above]. But unfortunately this does
not give any opportunity for fruitful formal fi nger exercises.
20
If someone is necessitated to do a specific act this is an absolute coercion [coactio
absoluta]. If he is necessitated to choose one amongst others [eligere] this is a hypo-
thetical coercion.
21
In Achenwall we find that absolute and categorical are usually closely related
in the context in question: si voto voluntas declaratur [] hypothetica, hypo-
theticum, si absoluta, categoricum [dicitur.] AA XIX, 341 (= Ius naturalis, pars
posterior, 26). For further details see Schwaiger (1999, especially 168 ff.).
152 Bernd Ludwig

absolutem und hypothetischem Naturrecht in Achenwalls writings


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

On the other hand, at least in Kants publications there is no talk of


unconditional judgments as opposed to hypothetical ones.
The factual differentiation between absolute/categorical necessita-
tion on the one hand and hypothetical necessitation on the other is
thus at least genetically independent from the special terminology of
the Table of Judgements from 1781. And Kants assimilation of im-
peratives and judgements in the Groundwork could par consquent
be nothing but a quite powerful artifice of suggestive presentation
without having a real fundamentum in re. Also, the hypothetical
necessity was sometimes called conditional by Kant (e. g., R 6639).
Correspondingly, one can find the distinction of necessitatio [] cat-
egorica vel conditionalis (R 6463). Thence, the Kantian innovation
was less of the discovery of the distinction itself, than of its specific
usage: maybe this is the reason why Kant can do without a system-
atic introduction of this distinction in the Groundwork, and rely-
ing on its familiarity simply introduces it with the formula: Now,
all imperatives command either hypothetically or categorically. (See
above), then interpreting this quite traditional distinction in his own
and new sense: as the distinction concerning the dependency of the
necessitation on a subjectively set end. And this sense, again, is not a
specifically critical one, but was set out already in Kants pre-criti-
cal Untersuchung ber die Deutlichkeit der Grundstze der natrli-
chen Theologie und der Moral from 1764 (see UDG, 298 ff.).

V. Are hypothetical imperatives analytic?

We saw above that the term hypothetical in the expression hypo-


thetical imperative is no grammatically independent adjective, but
an elliptic adverb to commanding. Given imperatives are hypotheti-
cally- or categorically-commanding respectively. A similar distinction
should be drawn as well for the adjectives analytic and synthetic.
Kants Hypothetical Imperatives 153

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

22
Cf., e. g., Seel (1989, 160 f.).
23
Otherwise we had to assume that some piece of text had dropped out.
24
It has to be kept in mind that in Kants writings the analytic/synthetic-distinction is
defi ned for categorical judgements exclusively, that is, it makes only sense for judge-
ments of the form S is P. A judgement of this kind is analytical if and only if the
predicate(-term) is contained in the subject(-term) (see e. g. KrV, B11). No matter
how this might be spelled out more precisely: there is no natural way to bring any
imperative (be it categorical or hypothetical) to the form: S is P and we dont
have any evidence that Kant himself ever thought he could accomplish this.
154 Bernd Ludwig

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

Hence this is as in the case of hypothetical and categorical not about


transferring the distinction between analytical and synthetical from
categorical judgments to imperatives. Rather, it is about the syntheti-
cal character of the very categorical judgment which ties the concept
of a definite willing to the concept of the will of a rational being. To
cut a long story short: synthetical in synthetical-practical proposi-
tion refers to the syntheticity of the very judgment which expresses
or explains the practical character of the imperative in question. This
insight involves, firstly, that logic and grammar are in complete cor-
respondence (like in analytic-practical proposition), and secondly,
that at least Kant himself can be released from the reproach of having
overlooked that imperatives (hypothetical as well as categorical ones)
are not (categorical) judgments.
Since nearby the formulation synthetic practical proposition
(GMS, 444) there is no evidence that Kant wanted to say anything
different here than in GMS, 420, one has to take this formulation as
elliptic. The same goes for the third and last relevant passage:
We cannot prove that this practical rule is an imperative, i. e., that the will
of every rational being is necessarily bound to it as a condition, by a mere
analysis of the conceptions which occur in it, since it is a synthetical prop-
Kants Hypothetical Imperatives 155

osition; we must advance beyond the cognition of the objects to a critical


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

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

25
Thus, those proposals of correction which claim against Kant that hypothetical im-
peratives were not analytical become obsolete, cf. e. g. Schnecker (1999, 96).
156 Bernd Ludwig

just not enough, and the inquiry has to be postponed until the third
Section of the Groundwork.
Since this interpretation is compatible with all passages in the
Groundwork26 in which Kant makes use of the analytic/synthetic-dis-
tinction, it does not need any demanding theses about tacit reinter-
pretations of the concept of analyticity, a concept which was already
well-known from his other writings.

VI. The End

Who does not only wish but also acts, i. e. who wills anything in the
Kantian sense, is necessitated to obtain adequate insights, realise them
and, as a consequence of these insights, to decide between the options
which are identified as being incompatible. Kants theory of hypotheti-
cally necessitating imperatives which are analytically-practical state-
ments (but neither hypothetical judgments nor analytic judgments) is
the subtle explication of this central but not very enigmatic fact. It is
no hidden theory of practical rationality and it does not presuppose
any idiosyncratic doctrines from transcendental philosophy. Hence,
if Kant himself says that the possibility of hypothetical imperatives
does not need any exceptional discussion, the aequitas hermeneutica
requires to refute any interpretation as inadequate which urges us to
suppose that there is anything to make out but an ingenious conceptual
frame for a picture rather familiar to his contemporaries. Whether or
not this conceptual frame is adequate has not been answered this way27
but it seems that smart alternatives are quite rare up to present.

Literature

Kants writings

Kants writings are cited according to the pagination of Kants gesammelte


Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902) (abbreviated as
AA). The Critique of Pure Reason will be cited according to the A/B pagi-
nation from the first and second editions. All textual references are to The

26
Kant does not take up the distinction between analytical and synthetical im-
peratives later again.
27
See further Ludwig 1999.
Kants Hypothetical Imperatives 157

Cambridge Edition of the Works of Immanuel Kant (Cambridge University


Press, 1992).
Ge ber den Gemeinspruch: Das , AA, VIII
GMS Grundlegung zur Metaphysik der Sitten, AA, IV
KdU Kritik der Urteilskraft, AA, V
KpV Kritik der praktischen Vernunft, AA, V
KrV Kritik der reinen Vernunft, AA III, IV
MdSR Metaphysik der Sitten, Rechtslehre, AA, VI
MdST Metaphysik der Sitten, Tugendlehre, AA,VI
R Reflexionen, AA, XVIXX
UDG Untersuchung ber die Deutlichkeit der Grundstze , AA, II

Other works

Achenwall, Gottfried/Ptter, Johann Stephan (1750): Anfangsgrnde des


Naturrechts (Elementa Iuris Naturae). Frankfurt a. M. 1995.
Downie, R. S. (1984): The Hypothetical Imperative, in: Mind vol. 43, 481
490.
Hare, R. Melvin (1971): Wanting: Some Pitfalls, in: R. Blinkley et al. (eds.),
Agent, Action, and Reason, Oxford, 8197.
Moritz, Michael (1960): Kants Einteilung der Imperative, Lund.
Paton, Herbert James (1953): The Categorical Imperative: A Study in Kants
Moral Philosophy, New York et.al.: Hutchinson.
Patzig, Gnther (1973): Die logischen Formen praktischer Stze in Kants
Ethik, in: G. Prauss (Hrsg.): Kant. Zur Deutung seiner Theorie von Er-
kennen und Handeln, Kln, 207222.
Schnecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kate-
gorischen Imperativs, Freiburg i. Br./Mnchen.
Schnecker, Dieter/Wood, Allen W. (22004): Immanuel Kant Grundlegung
zur Metaphysik der Sitten: ein einfhrender Kommentar, Paderborn.
Schwaiger, Clemens (1999): Kategorische und andere Imperative: zur Ent-
wicklung von Kants praktischer Philosophie bis 1785, Stuttgart-Bad Cann-
stadt.
Ludwig, Bernd (1999): Warum es keine hypothetischen Imperative gibt,
und warum Kants hypothetisch-gebietende Imperative keine analytischen
Stze sind, in: H. Klemme, B. Ludwig, et al. (eds.), Aufklrung und Inter-
pretation, Wrzburg, 105124.
Mark Timmons

The Categorical Imperative and


Universalizability (GMS, 421424)

Kants moral theory and, in particular, his theory of right conduct,


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

1
Kants rationale for this re-casting is elaborated in the Critique of Practical Reason,
chapter II, in a section entitled, On the Typic of Pure Practical Judgment.
2
Kant writes: I understand by canon the sum total of the a priori principles of the
correct use of certain cognitive faculties in general (KrV, A796/B824).
The Categorical Imperative and Universalizability 159

emphasis in moral philosophy on providing a theory of right conduct,


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

3
Even those sympathetic to Kants universalization tests will agree with Otfried
Hffe (2002, 144) that [Kants] manifold assurance that the universalization of
false promising runs into contradiction stands in peculiar contrast to the absence
of a precise demonstration of the point. And the same point holds with respect to
Kants use of the test in connection with other duties as well.
4
A point forcefully argued by Allen Wood (1999, 97110).
160 Mark Timmons

wrong; a solution to the problem of relevant descriptions (which can


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

5
Here are a few of relatively recent vintage: Korsgaard (1985), Herman (1993), Baron
(1997), Wood (1999, ch. 3), and Kerstein (2002, ch. 8). In the appendix below, I have
attempted to organize the various interpretations of how contradictions are sup-
posed to be generated from an application of the CI.
The Categorical Imperative and Universalizability 161

I. Kantian Universalizability as a Decision Procedure:


A Few Basics and Many Questions

The standard reading of FLN has to do with its role: it is supposed


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

6
For an excellent discussion of the main varieties, see Narveson (1985).
162 Mark Timmons

a maxim as the subjective principle of acting [] [which] contains


the practical rule determined by reason conformably with the condi-
tions of the subject (often his ignorance or also his inclinations), and is
therefore the principle in accordance with which the subject acts [].
Although in formulating sample maxims to be tested by FLN, Kant
includes sometimes more and sometimes less information about the
agents action and circumstances, it is clear that a maxim, when fully
stated, includes a specification of: (1) circumstances of choice, (2) the
action, and (3) the end, or what Kant calls the matter of action the
purpose or motivating reason for which one is proposing to undertake
the action in question.
So, the first step in the deliberative process is:
1. Formulate ones maxim (or the maxim which one is contemplat-
ing acting upon) having the form:

If/whenever _________ I will _________ in order to _________,

where the blanks are to be filled respectively with a description of


ones circumstances, action, and end. The next step involves asking
the question: how would it be if my maxim were to become universal
law? (GMS, 422). Thus, one is to:
2. Raise ones maxim to the status of universal law by reformulat-
ing it to have the form:

Of necessity, if/whenever _________ everyone will ________ in order to


_________.

Next, the consistency tests are applied:


3. Consider whether, in willing this imagined law of nature to hold,
one is thereby committed to a contradiction either in conception or
in volition.
Finally, given the connection between the universalizability of the
maxim and the deontic status of actions, we reach a moral verdict:
4. If the maxim is universalizable (it can without inconsistency be
conceived and willed as a universal law of nature) then action on the
maxim is morally permitted. If, however, the maxim is not universal-
izable, then action on the maxim is morally forbidden (wrong) and,
depending on which test the maxim fails, the action mentioned in the
maxim is either a violation of perfect or of imperfect duty.
In relation to this step-wise sketch, let us take a brief tour of the
main interpretative questions that arise at each step, and then in later
sections, we shall consider some of these questions in more detail.
The Categorical Imperative and Universalizability 163

Step 1: formulating ones maxim

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

Rosss example might prompt the thought that the more detail the bet-
ter, since in his example the description lying to a would-be murderer
is both the most detailed and intuitively correct maxim for moral as-
sessment. But in many circumstances, there will be many details (e. g.,
ones age) which are not morally relevant and whose mention in ones
maxim will affect its universalizability and thus lead to intuitively mis-
taken moral verdicts. So, presumably, a solution to the problem of rele-
vant descriptions would specify which features of ones choice situation
are morally relevant and thus ought to be included in a formulation of
ones maxim and, by implication, which features lack moral relevance
and ought not be included. We return to this issue in IV.

From step 1 to 2: raising ones maxim to universal law

Going from step 1 to step 2 is a matter of raising ones maxim to its


universalized counterpart, but there are questions about how this
is to be done. The dominant interpretation (reflected in the above
scheme) is that one imagines ones maxim raised to the status of a
164 Mark Timmons

psychological law according to which everyone who recognizes or at


least believes that she is in circumstances specified in the maxim, will
necessarily perform the specified action for the purpose in question.
We thus are to consider the universal conformity to the maxim. In
opposition to this interpretation, Thomas Pogge (1989) has defended
the view that we are to raise our maxim to universal law by imagining
its universal availability: we are to imagine that everyone feels mor-
ally free to adopt and act on the maxim in question, thus allowing
that not all agents will in fact do so.7 (In step 2, this would require re-
placing Necessarily everyone will with Everyone may). The main
philosophical importance of this interpretative issue concerns the aim
of making Kants tests yield as wide a range of intuitively acceptable
moral verdicts as possible. We will return to the issue in III.

Step 3: the consistency tests

This step has generated perhaps the most discussion about Kants de-
liberative procedure, and in the secondary literature we find a variety
of interpretations of exactly how contradictions can be generated in
connection with non-universalizable maxims. In the Groundwork,
Kant explains that the general canon of moral appraisal, encapsulated
in FUL and FLN is that we must be able to will that the maxim of
our act become universal law (GMS, 424). He goes on to describe
and illustrate two apparently distinct consistency requirements im-
plicit in the general canon: the so-called contradiction-in-conception
(CC) and contradiction-in-the-will (CW) tests.8 Those fascinated by
the question of how contradictions are supposed to be generated in

7
Pogge notes that in some places (e. g., GMS, 403) Kant poses the crucial universaliz-
ability question by asking whether one can consistently will that everyone may (as
opposed to will) act on her maxim; hence, the textual basis of the universal avail-
ability interpretation. In his 1954 book, Ross argued (in effect) that the universal
conformance to the false promising maxim featured in Kants example would not
destroy itself since if people acted always on this maxim every promise could be
relied upon to be broken, and this would in many (though not in all) cases serve as
well as if it could be relied upon to be kept (p. 30). Since he does think the universal
availability interpretation would, no doubt, lead to chaos, he too favors this lat-
ter interpretation. Allen Wood (1999, 80) claims that Kants presentations of FUL
and FLN in the Groundwork express the universal availability and universal con-
formity interpretations respectively. More recently, ONeill (1996, chs. 2 and 6) has
proposed what she calls a thin modal interpretation of the universalizability test in
terms of universal adoptability, though ONeill is less concerned about interpreting
Kant than she is about making use of a plausible understanding of this sort of test.
8
These labels were introduced by ONeill (Nell) in her 1975 book and have since
caught on in the secondary literature.
The Categorical Imperative and Universalizability 165

connection with non universalizable maxims, have mainly focused on


the CC test, and the various interpretations of it can be usefully or-
ganized and compared by considering how they stand with respect to
two questions:
(1) What kind of ancillary information information in addition to the
maxim and whatever is conceptually involved in its being raised to univer-
sal law is allowed to figure in generating contradictions in conception?
(2) What is the epistemic status of this additional information?

With regard to these questions, interpretations are more or less aus-


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

9
Galvin (1991), for example, defends an extremely austere interpretation.
10
The most famous of these is in the preface to the Groundwork at 389, but also occur
at GMS, 408 and GMS, 410n.
166 Mark Timmons

Step 4: inferring a moral verdict

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

11
In MdST, 393, Kant refers to inner actions, by which I take him to be referring to
such actions as consciously adopting ends and which is reflected in his discussion of
maxims of ends (adopting ends) and contrasted with maxims of (overt) action.
12
See, for instance, ONeill (1975) for an interpretation of Kants tests that restricts
deontic verdicts to ones about subjective rightness. See Timmons (1997) for a de-
fense of the claim that Kants tests are not restricted to yielding verdicts about the
subjective rightness of actions.
13
See ONeill (1985). In one place, Kant himself suggests that appeal to universaliz-
ability can be used for deriving both deontic verdicts and verdicts about worth.
After presenting FUL in section I of the Groundwork and illustrating its use in
connection with making a lying promise, Kant remarks that with this compass in
hand, [common human reason] knows very well how to distinguish in every case
that comes up what is good and what is evil, what is in conformity with or contrary
to duty (GMS, 404).
The Categorical Imperative and Universalizability 167

the moral verdicts in question are primarily about the rightness and
wrongness of actions.14
This concludes, then, a first pass over the main steps in the deliber-
ative procedure associated with FLN and some of the main questions
that arise at each step. We now proceed to examine in more detail
the very idea of FLN as a decision procedure, which will lead us to
make contact with some of the recent (and not so recent) secondary
literature.

II. FLN as a Decision Procedure:


Core Assumptions and the Strong Model

Fundamental moral principles are often cast in the dual role of pro-
viding both a decision procedure for moral deliberation and a moral
criterion. As I am using the term, a moral principle represents a moral
criterion of right action when it purports to specify those fundamen-
tal morally relevant (nonmoral) features (one or more), possession of
which by an action, makes it morally right. According to classical he-
donistic versions of act utilitarianism, for instance, the fact that a con-
crete action (open to an agent in some circumstance) would produce
as much net happiness as would any other available alternative action,
is what makes the action in question morally right; alternatives lack-
ing this feature also lack the feature of moral rightness, and so (given
a standard understanding of deontic concepts) are morally wrong.
Understood in this way, the principle of utility represents a moral cri-
terion of right action. It may also be understood as representing a
moral decision procedure, according to which in moral deliberation
we are to estimate the utilities associated with various alternative ac-
tions and decide accordingly. But it need not be taken as a decision
procedure, and some defenders of the utilitarian theory have argued
that it should not be understood in this manner.15 The point is that a
moral principle need not be cast in the dual role of decision procedure
and moral criterion; it may play one role without playing the other in
the overall economy of some moral theory.

14
Doing so still leaves open questions about, for example, whether the verdicts rep-
resent claims about the objective deontic status of actions or whether they only
represent verdicts about subjective rightness.
15
See, for example, Sidgwick (1907, book IV, ch. 1, 1), Bayles (1971), and Brink
(1989, ch. 8).
168 Mark Timmons

Let me suggest, then, that we simply focus on the hypothesis that


FLN represents a decision procedure, putting aside the question of
whether it is also a moral criterion. I distinguish between those core
assumptions of FLN as a decision procedure that are most central to its
functioning in this role from what seem to be additional desiderata that
may (arguably) be rejected without giving up this principle what I call
the strong model of FLN. In what immediately follows, I lay out the
core assumptions and then proceed to describe the strong model.

A. FLN: core assumptions

1. Self-sufficiency
Kant claims that the CI is the sole fundamental moral principle and
this seems to imply that FLN, in its role as a decision procedure, can be
employed as the sole moral premise in an argument that (together with
other nonmoral premises) can be used to derive deontic verdicts about
actions. That is, the four step deliberation procedure set out above pro-
vides the basis for a Kantian moral argument of the form: (1) An action
A is right only if (and perhaps if)16 its maxim M is universalizable; (2)
Maxim M is/is not universalizable; thus, (3) A-ing (acting on maxim M)
is right/wrong. Of course, this simple argument form will require that
premise 2 be supported by a further argument about Ms universaliz-
ability, but the idea of FLN as self-sufficient is that it is supposed to be
the only substantive moral premise in the three step argument.17

2. Coherence
Kant also claims that because the CI is implicit in ordinary moral think-
ing thinking which becomes even subtle (GMS, 404) it can be used
to derive a battery of common sense moral judgments, assumed to be
correct. In short, the moral verdicts about actions that can be reached
by using FLN as a decision procedure are supposed to cohere with ad-

16
Whether FLN is plausibly understood as expressing both a necessary and sufficient
condition of rightness has to do with what I call below its deontic strength. The
issue of self-sufficiency, notice, is separate from the question of deontic strength.
Even if FLN at best expresses a necessary condition of rightness, it will be self-suf-
ficient if it can serve as the lone moral premise in the sort of moral argument we are
here concerned with.
17
For example, Nelson Potter writes: A successful interpretation of the categorical
imperative consists of an argument having only one premise (the categorical im-
perative), and whatever factual and causal empirical premises are needed, from
which a conclusion concerning the moral rightness or wrongness of some particular
kind of action follows (1973, 397).
The Categorical Imperative and Universalizability 169

mitted correct common sense moral verdicts about those actions. This
requirement, note, will be met if, for whatever reason, the principle,
together with relevant nonmoral information, yields correct moral ver-
dicts in all actual and possible cases in which the test yields verdicts.18

3. Relevance
Not only must an adequate moral principle satisfy the coherence cri-
terion, but it must do so for the right reason in the sense that as a de-
cision procedure the moral principle either expresses or otherwise is
internally related to facts about actions in virtue of which they have
the deontic status they have. Assume that FLN is not a moral criterion
but that the humanity formulation is. Then, for FLN to satisfy the rel-
evance condition, the fact that a maxim is not universalizable must be
somehow internally connected to the fact that the action in question
fails to properly respect humanity as an end in itself.19
To summarize: self-sufficiency articulates what is required for FLN
to play the role of a decision procedure; coherence represents perhaps
the most basic condition of adequacy of a decision procedure; and rele-
vance serves to help explain why it is that a self-sufficient decision pro-
cedure does satisfy the coherence criterion. Giving up on some or all of
these seems tantamount to giving up on FLN as an adequate decision
procedure; they represent core assumptions. By contrast, the following
four additional theses each represent ways in which the core may be
strengthened thus making FLN a very potent decision procedure.

B. FLN: the strong model

4. Deontic power
There is textual evidence that Kant understood FLN as expressing
both a necessary and sufficient condition of moral rightness.20 If this
principle does have this kind of deontic power, then not only will it be
18
My formulation here is meant to capture a very weak coherence requirement ac-
cording to which FLN never leads to deontic verdicts that conflict with common
sense moral convictions. Strictly speaking, this requirement will be met even in case
FLN simply fails to yield any deontic verdicts. In this case, the complaint would be
that FLN is useless because completely indeterminate.
19
Although I do not take this desideratum to be quite as central as are the other two
to the success of FLN as a decision procedure, it certainly represents an assumption
that one would expect it to satisfy if it indeed satisfies self-sufficiency and coher-
ence. The importance of this desideratum is stressed by Herman (1989 and 1993).
20
See Pogge (1989, 189 f.) on this point who cites GMS, 421, and other passages as
evidence that it expresses a necessary condition and GMS, 403, as evidence that
Kant took it to express a sufficient condition as well.
170 Mark Timmons

able to rule out actions and omissions that are morally wrong, but it
will be able to rule in actions and omissions as obligatory. This would
mean, further, that FLN would be powerful enough to generate nega-
tive duties and positive duties.

5. Act-level determinacy
According to this desideratum, FLN is powerful enough to yield con-
clusions about the deontic status of a wide range of reasonably specific
act tokens concrete doings (or omittings) that are or might be per-
formed by a person at a time in a particular set of circumstances.

6. Classificatory accuracy
The classification at issue here is between importantly different types
of duty. Kant claims that his two tests are sufficient for accurately clas-
sifying morally required actions and omissions into the categories of
perfect and imperfect duty.

7. Maximum scope
The scope of a moral principle is a matter of the width or generality
of the class of agents to whom the principle is properly addressed. A
principle has wide scope vis--vis humans when it holds for all ac-
countable human agents. A principle has maximum width of scope
if it holds for all finite accountable agents, human and non-human.
Certainly Kant thinks that the CI has maximum scope. The more
interesting and controversial thought is that derived mid-level moral
generalizations of the sort featured in the Tugendlehre system of du-
ties (what I will call moral rules) also have maximum scope.

As I will proceed to explain further in the next section, these latter


four desiderata are not a package deal: one might accept some of them
and reject others (what I am calling a strong model comes in degrees
of strength). But let us pause before moving on in order to bring to-
gether the key ideas from this and the previous section.
We now have before us what I take to be the main interpretative
questions about Kants universalizability tests ( I) and in this sec-
tion we have set out the various core assumptions implicated in tak-
ing FLN as a decision procedure as well as made note of various de-
siderata each of which represents one way of strengthening FLN as
a decision procedure. In the next section, I will briefly indicate why
many recent interpreters (some of them sympathetic to FLN as some
kind of adequate decision procedure) deny one or more of the strong
The Categorical Imperative and Universalizability 171

theses. This will help us focus on how those who are sympathetic to
FLN as a decision procedure understand its deliberative power. But
then in IV and V, I want to get down to business and argue that:
(1) contrary to what some believe, the promise of FLN as a decision
procedure, whether strong or weak, cannot be saved (as perhaps some
think) by appeal to a theory of relevant descriptions, that therefore (2)
we ought to reject FLN as an adequate decision procedure. All this is
negative. But I then go on to argue, in a more positive vein, that (3) the
significance of universalizability in Kants ethics need not rest with
the CC and CW tests, rather (4) it is best understood as a metaethi-
cal constraint on moral reasons, and finally that (5) if we view Kants
Groundwork applications of FLN to maxims in the context of the
books overall strategy, then we should view Kants tests as represent-
ing ad hominem arguments addressed to morally conscientious agents
who are capable of experiencing moral demands but who are tempted
by considerations of self-interest to violate those demands.

III. Troubles for FLN as a Decision Procedure

In this section, then, I continue our investigation of FLN by first ex-


plaining the move away from strong interpretations by interpreters
who remain (somewhat) optimistic about this principle as a decision
procedure. However, I go on to explain why I think that a pessimistic
verdict about FLN is warranted; its importance within Kants ethics is
not primarily that of a decision procedure.

A. Chipping away at the strong model

Let us consider the grounds for rejecting each of the four strong the-
ses, taking them up in reverse order.

7. Wide Scope
In the preface to the Groundwork, Kant famously claims that:
Everyone must grant that a law, if it is to hold morally, that is, as a ground
of obligation, must carry with it absolute necessity; that, for example, the
command thou shalt not lie does not hold for human beings, as if other
rational beings did not have to heed it, and so with all other moral laws
properly so called; that, therefore the ground of obligation here must not
be sought in the nature of the human being or in the circumstances of the
world in which he is placed, but a priori simply in concepts of pure reason
(GMS, 389).
172 Mark Timmons

How properly to interpret this passage, especially in light of Kants


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

21
Galvin (1991) and Swoyer (1983) defend the idea that only a priori ancillary as-
sumptions are allowed in Kants tests, many others, including, Gram (1967), Pot-
ter (1975), Harrison (1957, 1958), ONeill (1985), Timmons (1984), Wood (1999),
interpret Kants tests as allowing empirical ancillary assumptions to play a role in
generating contradictions.
22
Here I am compressing two distinct levels: (1) the level where Kant appeals to very
general information about human beings in deriving the basic obligations that are
at the same time ends (self-perfection and happiness of others), and (2) the level of
more specific duties to oneself and duties to others that is elaborated in Parts I and
II of the Doctrine of Elements of Ethics in The Metaphysics of Morals.
The Categorical Imperative and Universalizability 173

terpretations only allow broadly conceptual truths to figure into the


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

6. Classificatory accuracy
The classificatory accuracy of the CC and CW tests is suspect. On this
point, I believe that most all contemporary commentators agree that

23
For an overview of the empirical assumptions operative in Kants tests see Wood
(1999, ch. 3).
24
Among those who allow information about human beings qua human as a legiti-
mate part of Kants derivation of duties disagree over the epistemic status of such
information. See Swoyer (1983), Pogge (1989), and Herman (1993).
174 Mark Timmons

the correspondence thesis (as Wood calls it) cannot be defended.25


One would think that at the very least acts of murder would fail the
CC test seemingly they are prime candidates for violations of perfect
duty. But Barbara Herman (1989), for instance, argues that instances
of murder and mayhem at best only fail the CW test and thus would
have to classified as violations of imperfect duty. And, indeed, no mat-
ter which particular interpretation of Kants CC test one accepts, it is
just not plausible to suppose that all maxims upon which one might
murder someone will be somehow inconceivable as a law of nature.26
This is bad news for the thesis of classificatory accuracy.

5. Act-Level Determinacy
To say that a decision procedure is determinate is to say that it is pow-
erful enough to yield deontic verdicts about a full range of particular
duties duties at the third level where questions about what a particu-
lar person ought to do, on some particular occasion (given the morally
relevant details of her situation) are in focus. After all, it may be that
the deliberative procedure prescribed by the FLN can be used to gen-
erate verdicts about our general duties moral rules at the second lev-
el but the procedure may not be of use in going from general duties
to concrete cases. Perhaps, for instance, Kants CC and CW tests can
be used to generate a set of moral rules, but applying them requires
moral judgment (both in determining to which situation those rule ap-
ply and how conflicts of rules are to be adjudicated). This would limit
the determinacy of FLN as a decision procedure.
Again, I think that there is a trend away from thinking that Kants
tests satisfy this desideratum. For instance, ONeill (1996), who is
sympathetic to Kants universality tests, restricts their use to that
of generating general moral rules of justice and virtue, and Herman
(1993), at least as I understand her deliberation model, agrees with
ONeill on this point.27 But again, rejecting this desideratum does not
25
See Wood (1999, 97101).
26
Christine Korsgaards so-called practical interpretation of the CC test works so
long as the murderers maxim includes some end of action (which the killing is sup-
posed to promote or bring about) such a getting a job that would otherwise go to
ones only rival. In a world in which everyone adopts and (when relevantly situated)
acts on this maxim, one would fail to achieve the getting of the job by acting on
ones maxim. But, as Korsgaard notes, her practical interpretation fails to show any
contradiction in conception in cases of murder for revenge. See Korsgaard (1985).
27
In reaching all thing considered judgments about duties in concrete circumstances,
Herman claims that the general moral rules generated by Kants tests must involve
what she calls a value translation. As I understand her view, we are to use moral
judgment in considering the relative normative force of possibly competing moral
The Categorical Imperative and Universalizability 175

mean that Kants tests are complete failures as decision procedures.


Their fertility may lie simply in generating a system of general moral
rules, and that would be significant.

4. Deontic Power
Even if one rejects the strong claim about act-level determinacy, one
might still hold that FLN and its tests are deontically powerful if one
holds that universalizability is both a necessary and sufficient condi-
tion for an act type being right. And here we come to what is by far the
most common way of criticizing FLN: describing examples in which
a maxim passes Kants tests but the action in question is nevertheless
wrong and thus falsely gives a positive moral evaluation of the action
(a false positive), and describing examples in which a maxim fails one
or another of Kants tests but where the action is not wrong and thus
falsely gives a negative moral evaluation of the action (a false nega-
tive). False positives challenge the sufficiency of universalizability as
a test for rightness; false negatives challenge its necessity.
Many recent (and not so recent) FLN optimists have been willing
to grant (in light of numerous false positives) that universalizability
is not sufficient for rightness. They thus reject the thesis of deontic
power. Nevertheless, they maintain the core idea that it represents at
least a necessary condition of rightness, perhaps to be supplemented
in Kants theory by his other formulations of the CI. Again, we take
this up in the next section.
So suppose we were to agree that Kants tests: (1) yield moral ver-
dicts that are limited in scope to human beings; (2) cannot be used to
discriminate perfect from imperfect duties; (3) yield moral verdicts
of limited determinacy; and (4) represent only negative tests of mor-
al rightness. All of this would weaken the deliberative procedure of
FLN, but it might still be useful in this role. Suppose that FLN can be
used, apart from other moral assumptions, in moral arguments yield-
ing moral verdicts which, though limited in the ways just conceded,
nevertheless cohere with common sense in the verdicts it does yield,
and also can be understood to yield those verdicts for the right rea-
son. Then this principle would still be self-sufficient, coherent, and
relevant the more central requirements that a moral decision proce-
dure must satisfy.

considerations (represented by the rules) in coming to an all-things-considered


moral verdict about a particular case. If this is right, the use of value translation
does not involve application of Kants tests in going from moral rules to concrete
verdicts. (I say more about this issue in the next section.)
176 Mark Timmons

B. Attacking the core

As I have said, by far the most common form of objection to FLN


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

28
Of course, this claim would have to be defended by examining the various inter-
pretations of the CC and CW tests since what works as a counterexample to one
interpretation may be avoided by one of its rival interpretations; a project I cannot
undertake here. However, for recent, good overviews that substantiate my pessi-
mism see Wood (1999, ch. 3) and Kerstein (2002, ch. 8).
The Categorical Imperative and Universalizability 177

position of being able to save an innocent persons life in order to pro-


tect that person from being harmed. Again, were we to have in hand a
principled account of morally relevant features, we would have a basis
for arguing that the second maxim is to be used in Kants tests since it
includes morally relevant information about ones circumstances and
purposes that the former lacks.

IV. Will a Theory of Relevance Save FLN?

I myself, from time to time, have optimistically suggested that alleged


counterexamples to FLN can be dealt with in this way.29 My current
skepticism is not due to the fact that Kants theory does not include
a principled account of moral relevance this is a quite absurd sug-
gestion, anyway. Rather, as I will explain shortly, making use of Kant
theory of moral relevance undermines FLN as self-sufficient and it
cannot save Kants tests from all obvious counterexamples. Let us
explore this issue by first bringing into focus Kants theory of moral
relevance, and then proceed to examine how his theory might be of
use in his universality tests.

A. Kants solution to the problem of relevant descriptions

So here is the hypothesis under consideration. Were Kant to have a


principled account of moral relevance, then we could use it to deter-
mine which features of ones circumstances are morally relevant and
should be included in an expression of ones maxim for moral testing,
and we could also use it to determine which features are not morally
relevant and should not be included in a specification of ones maxim.
Now Kant does have a theory of moral relevance that is expressed
in humanity formulation of the CI and is elaborated in the Tugendle-
hre system of moral duties. Let me explain. 30 A moral theory (whether
of right conduct or value) is in the business of specifying those fun-
damental features or properties of actions, persons, and their circum-
stances in virtue of which an action or other item of evaluation has
the moral properties it has. 31 The fundamental moral principles fea-
tured in a moral theory that purport to pick out which features, most
29
Timmons (1997, 2002b, ch. 7).
30
For more detail about Kants theory of moral relevance, see Timmons (1997).
31
I am helping myself to property talk strictly for convenience; it thus should not be
read in any metaphysically heavyweight manner.
178 Mark Timmons

fundamentally, make an action right or wrong, represent (what I am


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

32
This is not the whole story, however. The principle of autonomy, for reasons that I
will explain below in section V, is also deeply implicated in Kants theory of moral
relevance.
33
Talk of bearing here is meant to include not only how our actions causally affect
the rational autonomy of persons, but also considerations about what various ac-
tions express about rational autonomy. For more on this, see Wood (1999, 141 f.;
147 f.) and Timmons (2002a).
The Categorical Imperative and Universalizability 179

and promoting those essential capacities and tendencies that are es-
sential to the exercise of human rational autonomy.
How then do we get a theory of moral relevance from the enriched
conception of humanity? There are two points to be made. First,
think of the various Tugendlehre negative duties to oneself duties
to refrain from suicide, immoderate consumption of food and drink,
improper use of ones sexual organs as indicating general types of
action which, given basic anthropological facts about human nature,
necessarily interfere in some way with the maintenance of human-
ity in ourselves. Again, think of the various positive duties to oneself
duties to develop our (specifically human) powers of body, mind,
and spirit as types of activity the principled omission of which would
interfere with the maintenance and promotion of our humanity. Simi-
lar remarks apply to the various negative and positive duties to others.
So the first idea here is that the various duties of Kants Tugendlehre
system represent an anthropologically grounded specification (better:
an interpretation) of the (initially) thin notion of rational autonomy.
But what does all this have to do with the issue of relevant descrip-
tions? This brings us to the second point. To consider an action as of a
certain type is, in effect, to consider it under some description or other.
For many such descriptions, because of their importance in moral and
social life, we have single terms that serve to pick out actions under a
description. To classify an action as a case of suicide is short for clas-
sifying it under a more cumbersome description. The same goes for
terms such as, gluttony, lying, avarice, servility, ingratitude,
malice, and envy that are featured in Kants system of duties.
I am suggesting, then, that the system of Tugendlehre duties repre-
sents the outline of a principled account of moral relevance. In cases
where some action of mine can be correctly classified as an instance
of one or more of the various actions mentioned in the system of du-
ties, then this fact about it is morally relevant. Any other fact about an
action that is morally relevant in some context will be derivatively rel-
evant in the sense that its relevance (in that context) can be explained
by appeal to one or more of the more basic relevant features that are
expressed in the system of duties.

B. Why the solution does not save FLN

Recall the three core assumptions associated with FLN as a decision


procedure: self-sufficiency, coherence, relevance. I will now proceed
to explain why Kants theory of moral relevance will not vindicate
180 Mark Timmons

either the self-sufficiency or the coherence assumptions. Of course, if


FLN does not satisfy the coherence assumption, then it cannot possi-
bly indicate some feature of actions that is internally related to what
makes an action right or wrong.34

1. Self-sufficiency
In connection with Kants theory of moral relevance and its bearing
on the plausibility of the universality tests, there are two observations
worth recording here. First, this proposal vindicates Rosss complaint
(see the above quote) that Kants universality principle itself does not
specify the level of abstractness appropriate for maxims to then be
tested by the principle. 35 And so the current proposal represents a
rejection of the self-sufficiency thesis, because the moral principle of
humanity is now playing a crucial role in step 1 of the FLN procedure
it provides a basis for characterizing ones maxim so that it is suit-
able for moral testing. 36 Second, an immediate implication of all this is
that FLN cannot be used to derive the various moral rules featured in
Kants Tugendlehre; as we have explained, those moral rules involve
terms that are presumed to pick out morally relevant features of ac-
tions. 37 So if the reliability of the tests associated with FLN depends
on maxims including all and only morally relevant descriptions, and if
those descriptions are taken from the moral rules featured in Kants
ethical system, then it would seem that FLN cannot be used in argu-
ments to derive those very rules; it rather presupposes them.
But this is not the end of the story because it leaves open the possi-
bility that FLN, though not self-sufficient as so defined, can be used in
reasoning from general moral rules to specific moral verdicts in con-
crete circumstances. Specifically, it might serve an adjudicative role
34
However, in section V I consider what I take to the philosophical and practical rel-
evance of Kants universality tests.
35
ONeills 1975 solution to the problem of relevant descriptions was simply that
whatever information ones maxim (assuming that ones formulation of it is sincere)
does contain, is relevant for use of Kants tests. This means, in effect, that FLN does
contain a solution to the problem of relevance, contrary to Ross. See ONeill (1975,
13). One implication of this solution is that the tests are restricted to yielding only
judgments about the subjective rightness of actions. See ONeill (1975, 129).
36
This, incidentally, is to reverse the view, expressed by many interpreters, that be-
cause the concept of humanity (and the associated requirement of treating human-
ity as an end in itself) is excessively vague, in order to judge whether some action
would constitute a case of failing to treat someone as an end in themselves, one
must first use FLN to determine whether ones maxim is universalizable. See, for
example, Singer (1961, 235).
37
In the Tugendlehre, Kant only uses the universal law formulation of the CI in argu-
ing for the duty of beneficence. See MdST, 393, 451, and 453.
The Categorical Imperative and Universalizability 181

in moral reasoning. The idea would be that in judging whether some


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

38
This proposal would, of course, embrace the act-level determinacy thesis.
39
In Smart and Williams (1973, 97 f.).
182 Mark Timmons

M In circumstances where: I am unemployed, I have (for health rea-


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

I submit that on either of the two most plausible interpretations of the


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

This example (and many more can be provided) strongly suggest that
the CC test cannot help us reach conclusions about concrete cases
(in which we need deliberative help) when morally relevant detail is
specified in the maxims being tested.
What about the CW test? We have not said much about this test. But
from Kants Groundwork examples of letting ones (natural) talents
rust and refraining from helping others in need, the blue print for its
use is fairly clear. The maxims in these cases can be consistently con-
ceived as laws of nature (they pass the CC test), but they cannot be con-
The Categorical Imperative and Universalizability 183

sistently (rationally) willed to be such laws. This is so because (1) qua


rational agent, there are what we may call rational maxims maxims
which one ought to adopt (if one has not already done so) given ones
nature as a rational agent and which (2) are inconsistent with the im-
plications of the maxim being tested functioning as a universal law of
nature. In Kants writings we find appeal to two rational maxims in his
Groundwork arguments for the duties to develop ones own talents and
to help others in need. If the maxim to develop ones own (natural) tal-
ents is a rational maxim, then arguably, so is the maxim to preserve and
develop ones perfection as a moral being. Indeed, in the Tugendlehre,
the maxim to preserve and promote ones moral powers is a fundamen-
tal positive duty to oneself. And, M (insofar as it represents a sacrifice
of moral integrity) conflicts with this positive duty of moral self-perfec-
tion. But, of course (as many critics have pointed out) universalization
does not play a role in Kants argument regarding the duties of self-
perfection.40 Since we are examining the idea that Kants universality
tests (involving raising ones maxim to the status of a universal law of
nature) might be of use in deriving concrete verdicts given a general
theory of moral relevance, appealing to rational maxims concerned
with self-perfection will not help us out.
Let us therefore consider the rational maxim of willing to be helped
by others in cases of need. Unfortunately, this seems hardly relevant
to the case at hand because one is not proposing to refuse help to
others in need. Granted, if George acts on M (and takes the job) his
action will express a certain lack of proper regard for the welfare of
others; he will not be acting with the greatest possible care. But, here
we must be careful about how we interpret the duty of beneficence,
otherwise this duty will be implausibly demanding.41
The CW test, just like the CC test, does not seem to be of much
help in arguing for conclusions about the deontic status of concrete
actions even if one appeals to Kants theory of moral relevance. The
upshot is that the idea that a theory of moral relevance will save FLN
and the associated universality tests from counterexamples is false.
More precisely, I have argued that (1) given Kants theory of moral rel-
evance, FLN cannot stand as a self-sufficient test for generating moral
rules nor (2) can it be used in reasoning from general moral rules to
conclusions about concrete cases.
40
But see Glasgow (2003) for an intriguing attempt to explain how considerations of
temporality play a role in Kants universality tests and thus how it is that considera-
tions of universalizability can be made to do work in deriving duties to oneself.
41
For a discussion of this matter, see Rawls (1989).
184 Mark Timmons

2. Coherence
As I mentioned earlier, the coherence desideratum has been central in
evaluating the plausibility of Kants CC and CW tests. And I do think
that Kants theory of relevant descriptions does help to explain why
certain maxims particularly those that include intuitively morally
irrelevant information are not appropriate for moral testing. But,
unfortunately, Kants theory of moral relevance cannot save the tests
from all counterexamples, even from some of the well-known ones.
For instance, the maxim, I will buy toy trains but not sell them, cannot
be consistently universalized, but this maxim does not seem to include
morally irrelevant information about the action, nor does it seem to
omit morally relevant information about the agent, her act, or her cir-
cumstances. Here we have a false negative.
I think it is also clear that Kants tests yield false positives. With
false positives the problem seems to be that fairly specific maxims
pass the tests. In the above example featuring George, both M and
its contrary M* pass the tests. Now, of course, we could just conclude
that Georges taking the job and his turning it down are permissible
actions a case where the moral pros and cons are fairly evenly bal-
anced. But notice that if it is the specificity of maxims that get them
through Kants tests (as seems to be the case on both the logical and
practical interpretations) then in the many cases like Georges in
which a number of competing moral considerations are relevant, both
the maxim and its contrary will pass Kants tests which implies that
the action in question and its omission are both morally permissible. I
do not believe it is plausible to suppose that in all such cases the cor-
rect moral verdict is that both the action and its omission are morally
permissible this amounts to massive moral indeterminacy.
So, I conclude that Kants theory of moral relevance, used as a con-
straint on formulating maxims, does not save Kants universality tests
from counterexamples and thus that the coherence desideratum per-
haps the most important of the core assumptions is not satisfied.

C. Rejecting FLN as a moral decision procedure

We began with a very strong conception of a decision procedure and,


following the train of a long history of interpretation, have found
reason to conclude that (1) not only is FLN not a powerful decision
procedure, but (2) it is not self-sufficient, (3) it does not seem able to
play an adjudicative role in moral reasoning, and (4) even appealing
to a principled theory of morally relevant descriptions does not save
The Categorical Imperative and Universalizability 185

the tests associated with FLN from counterexamples. Suppose, then,


we agree that FLN is not a self-sufficient decision procedure, or even
part of one. One option for those interested in carrying on the project
of Kantian ethics is simply to jettison FLN and look to Kants other
formulations of the categorical imperative for inspiration in develop-
ing an adequate theory of right conduct.42 But surely, the very idea
of universalizability is philosophically important for Kants overall
ethical theory, and we ought now to ask what its true philosophical
significance is.

V. The Formal Constraint Interpretation of the Universal Law


Formulation of the Categorical Imperative

Let us suppose that the universality formulations of the CI do not rep-


resent moral criteria and that they do not represent adequate decision
procedures. What then? I am going to propose what I call the formal
constraint interpretation of the FUL (and FLN) according to which
(roughly) this formulation of the CI represents a formal constraint (or,
better, a system of interconnected formal constraints) on what can
count as a substantive moral reason. Thus, FUL plays an important
role in Kants ethics, but not that of a moral criterion or of a decision
procedure. So, let us make a fresh start by first focusing on Kants no-
tion of universalizability.
I mentioned earlier that the term universalizability has been
used to refer to a variety of importantly different ideas and claims in
ethics. Recall that what I have called substantive universalizability
theses purport to set forth conditions which, together with relevant
nonmoral information, have normative implications. According to
one brand of utilitarian universalizability, an action is right just in
case the following counterfactual supposition involving universaliza-
tion is true: if everyone who is relevantly similarly situated were to
perform the action in question, the utility thereby produced would
be at least as high as the universal conformance utility associated
with any alternative action one might perform instead. Kantian uni-
versalizability (as originally explained above) promised to represent
a self-sufficient moral principle with substantive moral implications.
This, as I have explained, we have reason to doubt.

42
Alan Donagan (1977), for example, develops a systematic moral theory on the basis
of the humanity formulation.
186 Mark Timmons

But universalizability is also used as a term to refer to meta-


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

43
Sidgwick (1907, 379; see also, 209 f.).
The Categorical Imperative and Universalizability 187

these same formulations do a great deal of justice to what the univer-


salizability test in Kants writings really does help reveal.
Let us now proceed to examine respectively the philosophical and
practical import of the universal law formulations of the CI.

A. Universality: it philosophical significance

Talk of the ground(s) of obligation can be usefully viewed as involv-


ing two related issues: an issue about the content of those features of
an action in virtue of which it is obligatory, and an issue about the
ultimate source of the kind of normativity that is characteristic of
moral obligation. The two issues are connected: to serve as a moral
reason, a consideration must have a kind of normative authority, in-
deed, for Kant, normative supremacy.44 Let us examine the matter
more closely.
I interpret the content question as having to do with a specification
of the kinds of considerations (their content) that represent moral rea-
sons considerations about actions, agents, and their circumstances,
that are a basis for (ground) moral requirements. The content ques-
tion, then, is: What kinds of contentful considerations qualify as
moral reasons? More precisely, if we assume that there is a single
underlying or fundamental consideration bearing on action in virtue
of which an action has the deontic status it has, then we can re-state
our question as: What sort of contentful consideration qualifies as a
fundamental right- and wrong-making feature?
To answer this question, we must specify some general constraints
that any kind of consideration must satisfy in order to count as a moral
reason, given our common sense notion of duty. My suggestion on
behalf of Kant is that the universal law formulations (in effect) en-
capsulate a set of formal requirements, based on the common notion
of moral obligation, which moral reasons, qua moral, must possess.
Doing so is a central part of Kants attempt in the Groundwork to
search for the supreme principle of morality understood as a mor-
al criterion that picks out a fundamental morally relevant reasoning
providing consideration that grounds moral rightness.
Here, then is a list of three formal constraints on what can count as
a basic moral reason, implicated in the universalization formulations
of the CI.

44
See T. M. Scanlon (1998, ch. 4) for discussion of these two questions about moral
reasons.
188 Mark Timmons

Constraint 1: Law-like character of moral reasons


In a number of places in the Groundwork, Kant says that the common
sense notion of duty essentially involves the idea that moral require-
ments hold necessarily for all rational beings as such.45 Let me con-
nect this idea of necessity with the idea of a moral reason.
For it to be true of someone that she is under a moral obligation, it
must be true that there are considerations about her circumstances in
virtue of which she has the obligation in question. If we think that un-
derlying the various and sundry morally relevant features in a particu-
lar case that bear on ones obligations there must be some underlying
right-making feature, then Kant is claiming that there must be some
underlying feature of actions that makes them obligatory. We can put
this in terms of moral reasons talk by saying that for Kant, underly-
ing our obligations is some fundamental consideration that provides a
(normative) reason for action. Now, according to Kants understanding
of morality, the basic moral requirements hold necessarily for all ra-
tional agents, by which I take him to mean that necessarily all rational
agents qua rational have reason to comply with those requirements.
Putting these ideas together, then, we can say that fundamental moral
reasons considerations in virtue of which, most fundamentally, a
rational agent is obligated to act in one way or another, must be such
that necessarily all finite rational beings have such reasons. To say
that necessarily all finite rational agents have such reasons for action
commits one to the claim that such reasons are universal in scope.

Constraint 2: Supremacy of moral reasons


In the Groundwork, Kant often uses the term supremacy to char-
acterize the kind of value possessed by actions done solely from duty
and the kind of will that is disposed to act from such a motive. But
in other places, Kant says that the moral law itself has normative su-
premacy or commanding authority (GMS, 426, 431), and, of course,
he does refer to this law as the supreme principle of morality. Now it
is a constant theme in Kants moral writings that given our ordinary
notions of moral obligation, moral reasons are normatively superior to
competing prudential considerations.46 So, according to this suprema-
cy constraint, moral reasons must be such that they represent consid-

45
See esp. GMS, 389 and 408.
46
The various contrasts between morality and prudence, including the superior nor-
mative force of the former over the latter, is particularly clear in the Critique of
Practical Reason in Remark II appended to Theorem IV (KpV, 35 ff.).
The Categorical Imperative and Universalizability 189

erations that provide reasons that are superior in normative force to


any competing, nonmoral reasons.

Constraint 3 Proper object of respect


In the Groundwork, Kant tells us tells us that respect is a represen-
tation of a worth that [not only] infringes upon my self-love (GMS,
401n), but far outweighs any worth of what is recommended by incli-
nation (GMS, 403). He also tells us that the proper object of respect
is the moral law. There are two things to note here. First, when Kant
says that the law is the proper object of respect, I read him as in effect
claiming that the sorts of considerations that necessarily provide rea-
sons for action to all rational beings is a proper object of the attitude
of respect. (More on this below.) Second, it is the supremacy of the law
or, more precisely, moral considerations that call forth or demand
respect47 and, in so doing our recognition of them checks our self-love
and strikes down what Kant calls self-conceit the latter a kind
of smug comfort one might take in herself without having made moral
considerations fundamental in ones estimation of self worth.
To sum up: moral reasons qua moral must satisfy an interconnected
set of formal constraints: they must be such that (1) all rational agents
have such reasons for action, (2) they must be supremely authorita-
tive, and (3) they must be a proper object of respect. I am suggesting
that these tightly interconnected ideas are (in effect) encapsulated in
Kants universality formulations of the CI. Looking at Kantian uni-
versality in this way, we can now ask what sort of fundamental right-
making consideration can play this role. I believe that Kants answer is
expressed in the humanity formulation of the CI. Let me explain.

B. The formal constraint argument

The task, then, is to identify some consideration that has a kind of worth
that necessarily counts as a supremely authoritative normative reason
for all rational beings something that is a proper or fitting object of
respect. As I understand Kants ethical theory, it is humanity as an end
in itself having what Kant calls a dignity which possesses nonrela-
tive, absolute worth (GMS, 435 f.), and whose normative significance is
expressed in the humanity formulation of the CI. In brief, the idea is
that if we assume that something of value can ground reasons for ac-
tion, then, of course, the dignity of humanity can ground reasons for

47
See, for instance, KpV, 80.
190 Mark Timmons

action. Moreover this dignity is something necessarily possessed by all


rational agents and possesses an absolute nonrelative worth which,
according to Kant, outweighs any reason grounded in inclination or
at least excludes it altogether from calculations in making a choice
(GMS, 400) and is thus a proper object of respect. That humanity as
an end in itself is a fundamental right-making feature of actions is, of
course, reflected in how Kant uses the humanity formulation of the
CI to argue for and explain the rightness and wrongness of the vari-
ous duties comprising the Tugendlehre. Assuming that the humanity
formulation can provide a basis for common sense moral judgments
(something Kant defends in the Tugendlehre), we can understand the
humanity formulation of the CI as expressing a moral criterion of right
action. This is roughly how Kants notion of universal law leads us from
formal criteria (that constrain what can count as a fundamental moral
reason bearing on action) to a substantive moral criterion.
So, on my formal constraint interpretation of the CI, we are to un-
derstand Kant as (in effect) arguing as follows:

The formal constraint argument


1. According to common rational cognition, in order for some con-
sideration bearing on the rightness of action to count as a moral
reason, it must satisfy formal constraints 13 listed above.
2. Humanity as an end in itself satisfies all the relevant formal con-
straints.
3. Humanity as an end is the only consideration that satisfies the for-
mal constraints.
Thus,
4. According to common rational cognition, humanity as an end in
itself is a fundamental moral reason.
Since a moral principle cast in the role of a moral criterion is in the busi-
ness of specifying some one (or more) fundamental morally relevant
features that provide normative reasons that purport to explain why an
action has the deontic status it does, the import of this argument is that
it takes us from formal constraints on moral reasons to a fundamen-
tal substantive moral principle the humanity formulation. Of course,
whether this argument and, in particular, its third premise, can be de-
fended cannot be examined here. However, let me conclude this section
with two observations about my formal constraint interpretation.
First, in GMS, 436, Kant tells us that Autonomy is [] the ground
of the dignity of human nature and of every rational creature. So,
although having dignity is the fundamental value that grounds moral
The Categorical Imperative and Universalizability 191

requirements, it is the autonomy of rational beings roughly, their


capacity for giving law that represents a feature of such beings
that is most fundamental in Kants theory of right conduct and val-
ue. At the beginning of this section, I mentioned that in addition to
philosophical questions about the content moral reasons, there are
questions concerning the normative authority of such reasons. I also
mentioned that the two issues are deeply related in Kants thought: in
order for a consideration to count as a moral reason bearing on action,
it must have a kind of normative supremacy vis--vis competing rea-
sons. Kants account of normative authority has to do with his notion
of our natures as law giving creatures, and so a full and complete
story about moral reasons in Kant would require exploration of his
notion of autonomy. This I leave for another occasion.
Second, one of the desiderata that any plausible interpretation
of Kants notion of universalizability must satisfy is that the kind of
constraint expressed by this notion be stronger than the logical the-
sis of universalizability. This desideratum would be nicely satisfied
were Kants universalizability procedures adequate tests of the de-
ontic status of actions. But, as explained earlier, we have reason to
doubt that FUL or FLN are adequate when cast in the role of decision
procedures. But now notice that my formal constraint interpretation
takes Kants conception of universality as a metaethical constraint on
reasons being moral reasons, and as such is supposed to have enough
power to collectively to pick out certain considerations, in terms of
their content, as being moral reasons. So, on the one hand, it is strong-
er than the logical principle of universalizability. But on the other
hand, understood as I propose, it constrains without serving as a self-
sufficient decision procedure from which we can derive an adequate
system of duties. The content that the humanity formulation adds to
the universal law formulation of the categorical imperative is signifi-
cant for purposes of reasoning ones way to a system of duties. And
so it is not surprising that when Kant turns his attention to deriving a
rich set of duties in the Tudgendlehre, he does so on the basis of the
humanity formulation.48
48
I might be accused at this point of false advertising. At the outset, I promised to
provide an interpretation of universalizability, but what I have really done is offer
an interpretation of the notion of universal law and what I claim are conceptually
associated notions (normative supremacy and respect as a fitting attitude). This bait
and switch was accomplished by vague talk of universality, so the complaint might
go. Reply: if one wants to use universalizability associated with Kants moral the-
ory to refer exclusively to a consistency test on maxims involving a hypothetical
scenario, I wont complain. In the next section, I explain what I take to be the true
192 Mark Timmons

C. Universalizability as an argumentum ad homimen

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

significance of this kind of thinking. What I have been doing, then, can be under-
stood as presenting what I take to be the philosophical significance of the central
concepts implicated in Kants CC and CW tests.
The Categorical Imperative and Universalizability 193

whether ones maxim can be willed as universal law typically operates


against the background of a rich set of moral assumptions. In the case
of suicide, the appeal to a teleological law seems to involve an appeal
to moral considerations about the (morally) proper use of a motive. In
the case of lying promises, the moral acceptability of the institution
of promising is taken for granted. And in the talents and beneficence
examples, appeals to rational maxims maxims that any non-fi nite
agent qua rational would necessarily adopt (and thus maxims that any
finite rational agent ought to adopt) involve background moral as-
sumptions. Without these assumptions, the universality arguments do
not get off the ground, and with them they help reveal the kind of ra-
tionalizing duplicity that is typically involved in morally wrong action.
It is in revealing this kind of duplicity that Kants universalizability
arguments function as ad hominem arguments.
So, even if the CC and CW tests do not represent self-sufficient
tests of moral rightness, they are useful in helping to confirm the dis-
tinct authority of moral reasons and, in the context of moral reason-
ing, help reveal a kind of duplicity in an agents attempt to justify im-
moral action.

VI. Concluding Remarks

I have argued that Kants universal law of nature formulation, the


typic of the more general formula of universal law, is not plausibly
understood as expressing a self-sufficient moral decision procedure;
rather, its chief philosophical significance lies in representing a formal
metaethical constraint (really, a set of related constraints) on moral
reasons. Because this constraint puts limits on what can count as a
moral reason, it is stronger than what is represented by resultance,
supervenience, and logical universalizability, as these logical con-
straints are currently understood. This generic formal idea of moral
reasons as universally valid reasons is systematically spelled out (un-
folded as it were) in the Groundwork through the various formula-
tions of the categorical imperative: from the universal law formula-
tions that express this formal constraint, to the humanity formulation
that identifies the content of fundamental moral reasons, to the for-
mula of autonomy which expresses the ultimate source of normative
authority of the moral reasons.
194 Mark Timmons

Appendix:
Sorting Out Interpretations of Kants Universalizability Tests49

There are two issues, partly (but not completely) orthogonal to one
another that figure in standard ways of classifying interpretations of
Kants CC test. 50 First, as mentioned above in section II, there is the
question of what sorts of ancillary information may legitimately play
a role in the CC test. Austere interpretations restrict such information
to necessary truths knowable a priori, while opulent interpretations
allow a posteriori claims to enter in. Second, there is the question of
the precise locus of the alleged contradiction that is revealed through
testing non-universalizable maxims. Here, so far as I can tell, there
are two main interpretative options. One of them attempts to locate
the contradiction either within the maxim qua universal law at-
tempting to make sense of Kants claims that certain maxims, when
universalized are self-contradictory or within a system of nature
one of whose laws is the maxim qua universal law. Within this camp
we find a variety of different views on what ancillary information may
be used in generating such contradictions, and hence we find both aus-
tere and opulent versions of this species of interpretation. According
to another interpretive strain of thought, the contradiction is not in
the very conception of a law or a system of nature but rather involves
a contradiction among ones intentions involved in the process of rais-
ing ones maxim to universal law.
Here, then, is a brief overview of the various interpretations, in-
cluding footnote references to each interpretations defenders.

A. Inconsistency in Law Interpretations of the CC test

1. Strict conceptual interpretation


On this interpretation the only auxiliary information allowed in the
test is represented by analytic truths that express the meanings of
those terms featured in the maxim. For example, some interpreters
have claimed that given the very concept of slavery, in order for this
practice to exist, there must be slaves and slave owners (who by defini-
tion cannot themselves be slaves). Thus, the very idea of a law accord-
ing to which everyone holds slaves is itself logically inconceivable. In
49
Special thanks here to Josh Glasgow. Ive benefited from his unpublished handout
on the various interpretations of the CC test.
50
There is far less controversy, so far as I can tell over how the CW test is supposed to
work. But maybe this is owing to comparative lack of attention.
The Categorical Imperative and Universalizability 195

restricting auxiliary information to a narrow set of conceptual truths,


this interpretation is necessarily austere. 51

2. Relaxed conceptual interpretation


Some interpreters have proposed that in addition to the narrow set
of analytic truths allowed by the strict conceptual interpretation, we
may allow analytic truths about finite rational agency. Permitting
this expansion presumably allows the CC test to generate a greater
number of contradictions covering a wider range of alleged violations
of perfect duty than does the stricter version. Again, this counts as an
austere interpretation. 52

3. Teleological law interpretation


Kants Groundwork suicide example (GMS, 421 f.) appeals to the al-
leged purpose of self-love and thus the idea of teleology. Some inter-
preters have attempted to make the appeal to teleological laws central
for understanding Kants CC test. The idea is that certain maxims, if
raised to the status of a teleological law would be inconsistent with
other teleological laws. Here, we seem to get an intra-world contra-
diction, so to speak, rather than a brute inconsistency in the very con-
ception of ones maxim raised to universal law. Whether this interpre-
tation counts as austere or opulent depends on how one understands
the modal and epistemic status of teleological claims. 53

4. Causal law interpretation


On this interpretation of the CC test one is to raise ones maxim to the
status of a casual law of nature and consider it together with other causal
laws that happen to govern the actual world. The idea is that some max-
ims are such that in trying to conceive of them as causal laws of nature
operating in a world like the actual world, a contradiction emerges: given
actual causal laws (plus the normal and predictable results of the op-
eration of such laws) the attempt to conjoin to those laws the universal-
ized maxim leads to contradiction. This interpretation is clearly opulent
in allowing auxiliary information knowable only a posteriori.54

51
Cf. Galvin (1991). Similarly, Hffe (2002, 149) argues that experience is not re-
quired to know that a false promise harbors two conceptually internally and mutu-
ally incompatible aims.
52
Cf. Swoyer (1991) and perhaps Pogge (1989).
53
Cf. Paton (1953) and Beck (1960).
54
Cf. Harrison (1957); Dietrichson (1964); Timmons (1984); Rawls (1989); Herman
(1993).
196 Mark Timmons

B. Inconsistency in Intention Interpretation of the CC test


Often called the practical contradiction interpretation of the CC test,
the idea here is that some maxims involve intended ends and actions
(as means) such that were one to will that everyone act on ones maxim,
the result would be that no one (including the agent) would be able to
achieve the intended end by performing the action (that is, a token of
the action type) in question. Thus, in willing ones maxim as a universal
law one would in effect be willing the frustration of the end in question
at least the frustration of the ability to achieve the end by the intended
means. More precisely, the contradiction emerges as follows. (1) First,
take ones maxim of obtaining or achieving end E by Aing (in C), and
then (2) will, as per the test, that ones maxim become universal law. In
so willing the universality of ones maxim, (3) one also wills the normal
and predictable results, including the result that it not be the case that
one obtain or achieve E by Aing (in C). Thus, (4) the intention expressed
by ones maxim (to E by Aing in C) is flatly inconsistent with what one is
committed to intending in willing ones maxim as universal law: not (to
E by Aing in C). Note that this interpretation makes the CC test similar
to the CW test, but does not collapse the former into the latter. The CW
test requires appealing to rational maxims maxims that it would be
irrational to fail to have in generating contradictions in the will.
Most writers who favor this interpretation seem to be committed
to an opulent set of auxiliary premises, since drawing out the im-
plications of willing ones maxim as universal law seems to appeal
to the same considerations involved in the causal law interpretation
namely, the normal and predictable results of the operation of the
hypothetical law in a world otherwise like the action world informa-
tion that can only be known empirically. Might one appeal only to a
priori knowable auxiliary information in developing this interpreta-
tion? Perhaps. But then if a priori auxiliary information is all that is
needed to generate contradictions, why not just opt for either the strict
or relaxed version of the conceptual contradiction interpretation?55
55
Cf. Ewing (1933); M. G. Singer (1961); Gram (1967); Potter (1975); ONeill (1975);
Korsgaard (1989). This paper was presented at the Beijing International Sympo-
sium on Kants Moral Philosophy in Contemporary Perspectives, May 2004, Peking
University, Beijing, China. It was also presented at Friedrich-Wilhelms Universitt,
July, 2004, Bonn, Germany at which the papers in this volume were presented and
discussed. I wish to thank audiences at both conferences for their valuable input. I
would especially like to thank Christoph Horn and Dieter Schnecker for organiz-
ing the Bonn conference and editing this volume. Special thanks to Corinna Mieth
for her kind help with conference arrangments. I am indebted to Josh Glasgow for
his comments and suggestions on an earlier draft of this paper.
The Categorical Imperative and Universalizability 197

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesam-


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

Other works

Aune, Bruce (1979): Kants Theory of Morals, Princeton, NJ: Princeton Uni-
versity Press.
Baron, Marcia. W. / Pettit, Philip / Slote, Michael. (1997): Three Methods of
Ethics: A Debate, Oxford: Blackwell.
Bayles, Richard E. (1971): Act Utilitarianism: Account of Right-Making
Characteristics or Decsion-Making Procedure? in: American Philosoph-
ical Quarterly, 8, 25765.
Beck, Lewis. W. (1960): A Commentary on Kants Critique of Practical Rea-
son, Chicago: Chicago University Press.
Brink, David O. (1989): Moral Realism and the Foundations of Ethics, Cam-
bridge: Cambridge University Press.
Donagan, A. (1977): The Theory of Morality, Chicago: Chicago University
Press.
Ewing, Alfred C. (1933): The Paradoxes of Kants Ethics, in: Philosophy,
XXI, 4056.
Galvin, Richard (1991): Ethical Formalism: The Contradiction in Concep-
tion Test, in: History of Philosophy Quarterly, 8, 387408.
Glasgow, Joshua (2004): Expanding the Limits of Universalization: Kants
Duties and Kantian Moral Deliberation, in: Canadian Journal of Phi-
losophy, 33, 2347.
Hare, Richard M. (1963): Freedom and Reason, Oxford: Oxford University
Press.
Harrison, Jonathan (1957): Kants Examples of the First Formulation of the
CI, in: Philosophical Quarterly, 7, 5062.
Harrison, Jonathan (1958): The Categorical Imperative, in: Philosophical
Quarterly, 8, 36064.
Hegel, Georg F. W. (1821): Elements of the Philosophy of Right, Nisbet, H. B.
(trans.), (1991), Cambridge: Cambridge University Press.
Herman, Barbara (1989): Murder and Mayhem, in: The Monist, 92, 411449.
198 Mark Timmons

Herman, Barbara (1993): Moral Deliberation and the Derivation of Du-


ties, in: Herman, B., The Practice of Moral Judgment, Cambridge: Har-
vard University Press.
Hffe, Otfried (2002): The Prohibition against False Promising, in: Hffe,
O., Categorical Principles of Law, Migotti, M. (trans.), University Park,
PA: Pennsylvania University Press. (Originally published as Katogorische
Rechtsprinzipien, Frankfurt a. M.: Suhrkamp Verlag, 1990.)
Kerstein, Samuel J. (2002): Kants Search for the Supreme Principle of Mo-
rality, Cambridge: Cambridge University Press.
Korsgaard, Christine M. (1985): Kants Formula of Universal Law, in:
Pacific Philosophical Quarterly, 66, 2447. Reprinted in Korsgaard, C.
(1996): Creating the Kingdom of Ends, Cambridge: Cambridge University
Press, and in Guyer, P. (ed.), (1998): Kants Groundwork of the Metaphys-
ics of Morals: Critical Essays, Lanham, MD: Rowman & Littlefield.
McNair, Ted (2000): Universal Necessity and Contradictions in Concep-
tion, in: Kant-Studien, 91, 2543.
Mill, John S. (1863): Utilitarianism. London.
Mulholland, Leslie (1978): Kant: On Willing Maxims to Become Universal
Law, in: Dialogue, 17, 92105.
Narveson, Jan (1985): The How and Why of Universalizability, in: Potter,
N./Timmons, M. (eds.): Morality and Universality: Essays on Ethical Uni-
versalizability, Dordrecht: Reidel.
ONeill (Nell), Onora (1975): Acting on Principle, New York: Columbia Uni-
versity Press.
ONeill, Onora (1985): Consistency in Action. in: Potter, N./Timmons, M.
(eds.): Morality and Universality: Essays on Ethical Universalizability,
Dordrecht: Reidel. Reprinted in: Guyer, P. (ed.), (1998): Kants Ground-
work of the Metaphysics of Morals: Critical Essays, Lanham, MD: Row-
man & Littlefield.
ONeill, Onora (1996): Towards Justice and Virtue, Cambridge: Cambridge
University Press.
Paton, Henry J. (1953): The Moral Law, London: Hutchinson & Co.
Pogge, Thomas V. (1989): The Categorical Imperative, in Hffe, O, (ed.):
Grundlegung der Metaphysik der Sitten: Ein Kooperativer Komentar,
Frankfurt a. M.: Vittorio Klostermann. Reprinted in Guyer, P. (ed.),
(1998): Kants Groundwork of the Metaphysics of Morals: Critical Essays,
Lanham, MD: Rowman & Littlefield.
Potter, Nelson (1975): How to Apply the Categorical Imperative, in: Philo-
sophia, 5, 395416.
Rawls, John (1989): Themes in Kants Moral Philosophy, in E. Frster (ed.):
Kants Transcendental Deductions, Stanford: Stanford University Press.
Ross, Sir William D. (1954): Kants Ethical Theory, Oxford: Oxford Univer-
sity Press.
Scanlon, Thomas M. (1998): What We Owe to Each Other, Cambridge: Har-
vard University Press.
Sidgwick, Henry (1907): The Methods of Ethics, 7th ed, Indianapolis, IN:
Hackett Publishing Co., 1981.
Singer, Marcus G. (1961): Generalization in Ethics, New York: Alfred A.
Knopf.
The Categorical Imperative and Universalizability 199

Smart, J. J. C. / Williams, Bernard (1973): Utilitarianism: For and Against,


Cambridge: Cambridge University Press.
Swoyer, C. (1983): Kantian Derivations, in: Canadian Journal of Philoso-
phy, 13, 409431
Timmons, Mark (1984): Contradictions and the Categorical Imperative, in:
Archiv fr Geschichte der Philosophie, 66, 294312.
Timmons, Mark (1997): Decision Procedures, Moral Criteria, and the Prob-
lem of Relevant Descriptions in Kants Ethics, in: Jahrbuch fr Recht und
Ethik (Annual Review of Law and Ethics), 5, 289417.
Timmons, Mark (2002a): Motive and Rightness in Kants Ethics, in: Tim-
mons, M. (ed.): Kants Metaphysics of Morals: Interpretative Essays, Ox-
ford: Oxford University Press.
Timmons, Mark (2002b): Moral Theory: An Introduction, Lanham, MD:
Rowman & Littlefield.
Wood, Allen W. (1999): Kants Ethical Thought, Cambridge: Cambridge
University Press.
Samuel J. Kerstein

Deriving the Formula of Humanity


(GMS, 427437)

In Groundwork II, Kant tries to establish that if there is a supreme


principle of morality, then it is (or is equivalent to) the Formula of
Humanity. He offers a derivation of the principle: So act that you
use humanity, whether in your own person or in the person of any
other, always at the same time as an end, never merely as a means
(GMS, 429, emphasis omitted).
In broad outline, Kants derivation (GMS, 427 ff.) is not hard to
discern. It takes shape against the background of his fundamental
tenet that the supreme principle of morality would have to be a
categorical imperative, that is, a principle binding on all of us no
matter what our particular inclinations might be. First, Kant con-
tends that if there is a supreme principle of morality (and thus a cat-
egorical imperative), then there is an objective end: something that
is unconditionally valuable. A categorical imperative requires an
unconditionally valuable ground. Second, Kant claims that this
unconditionally good thing would have to be humanity. In his view,
therefore, if there is a supreme principle of morality, then humanity
is unconditionally good. For Kant humanity does not refer to the
class of human beings, but rather to a set of capacities. In the Meta-
physics of Morals, Kant tells us that the capacity to set oneself
an end any end whatsoever is what characterizes humanity (as
distinguished from animality) (MdST, 392). So at the very least,
if a being has humanity, then it has the capacity to set ends. Kant,
it seems, uses humanity interchangeably with rational nature
(see, for example, GMS, 439). In doing so he suggests that having
humanity involves having certain rational capacities. Among them
are the capacity to act on maxims and hypothetical imperatives, as
well the capacity to act autonomously, that is, (roughly) to conform
to self-given moral imperatives purely out of respect for these im-
Deriving the Formula of Humanity 201

peratives.1 According to the third main claim in the derivation, if


humanity is unconditionally good, then we must always treat it not
merely as a means but also as an end. Therefore, if there is a supreme
principle of morality, then we ought to do just what the Formula of
Humanity says. So the supreme principle of morality, assuming there
is one, must be this formula, or at least something equivalent to it.
This paper focuses on the derivations first two steps, especially the
second. Assuming that a categorical imperative would require there
to be something unconditionally good, why must this unconditionally
good thing be humanity, rather than something else? As Part II below
points out, the argument Kant appears to invoke in response to this
question, namely an argument by elimination (GMS, 428), is very dis-
appointing. But Christine Korsgaard (1996, 106132) and, later, Allen
Wood (1999, 124132) have claimed that this argument does not rep-
resent Kants best effort at a derivation. Kant, they contend, actually
undertakes a regressive argument. I concentrate on Woods more
recent account of the argument. According to a crucial part of Woods
account, Kant claims that whenever an agent sets an end, he must as-
cribe objective goodness to it.2 But in ascribing objective goodness to
his ends, Kant contends, the agent commits himself (rationally speak-
ing) to holding his rational nature (that is, his humanity) itself to be
unconditionally good. So, in sum, since a rational agent exercises his
rational nature in setting ends, he must conclude that it is uncondi-
tionally valuable.
Part III questions whether Kant actually makes this argument.
Contrary to Wood, I do not, for example, believe that, according to
Kants Groundwork position, whenever an agent sets an end, he is
rationally compelled to ascribe objective goodness to it. I do not deny
that it was open to Kant to appeal to the regressive argument or that

1
Here I am following Thomas Hill, Jr. (1992, 3841). Allen Wood presents a slightly
different account of what Kant means by humanity (1999, 118 ff.).
2
As Allen Wood has pointed out to me, it is not obvious that Korsgaard attributes this
claim to Kant. But I take as some indication that she does her statement that in the
argument for the Formula of Humanity, as I understand it, Kant uses the premise
that when we act we take ourselves to be acting reasonably and so we suppose that
our end is, in his sense, objectively good (1996, 116). In any case, it is worth noting
that Korsgaards account of the regressive argument incorporates a robust notion of
an objectively good end. On her account, Kant tries to show that if an agent takes
himself to have objectively good ends, then he is committed to the unconditional
goodness of humanity. But in order for an end to be objectively good in Kants sense,
she claims, it must be the object of rational choice, provide reasons for action that
apply to every rational being (1996, 115) and be fully justified. All three criteria are
apparent in 1996, 114116.
202 Samuel J. Kerstein

the regressive argument is philosophically interesting. It is just not, I


think, an argument that Kant himself unfurls.
Part IV tries to bolster Kants argument by elimination with the
help of material that he presents later in Groundwork II. Kant points
out that rational nature is that which can possess a good will (GMS,
437). Since it is, he suggests, rational nature alone can be the uncon-
ditionally good ground of a categorical imperative. For all other
candidates are such that if they had this status, the good will would
be subordinate to them. But, claims Kant, such a will cannot with-
out contradiction be subordinated to any other object (GMS, 437).
Suitably interpreted, this reasoning strengthens Kants argument by
elimination, I contend. The argument turns out to be stronger than an
isolated reading of GMS, 427429, makes it appear. Of course, even
if this claim is correct, Kants argument is open to challenge. Part V
focuses briefly on one objection to it.
Before exploring Kants defense of the view that humanity must
be the unconditionally good ground of any categorical imperative, we
need to examine why he thinks a categorical imperative requires such
a ground in the first place. Why, if there is a categorical imperative,
must there be anything unconditionally good?

I.

In his initial step in the derivation of the Formula of Humanity, Kant


claims that if there is a supreme principle of morality (and thus a cat-
egorical imperative), then there is something of absolute (i. e., uncon-
ditional) worth. In something of absolute worth alone would lie the
ground of a possible categorical imperative, and if all worth were
conditional and therefore contingent, then no supreme practical prin-
ciple for reason could be found anywhere (GMS, 428). But why could
not a principle be unconditionally binding on us if nothing was uncon-
ditionally good?
According to Kant, an agent sets himself to do something, that is,
determines his will, on the basis of his idea that doing this thing will
enable him to secure some end. In Kants view, all acting has an end
(KpV, 34). Kant distinguishes between subjective and objective ends.
Objective ends, if there are any, would hold for all rational beings.
The idea of securing them would make available to all rational beings
a sufficient ground (motive) for acting. But subjective ends do not give
all rational beings grounds for securing them. These ends are such
Deriving the Formula of Humanity 203

that their mere relation to a specially constituted capacity of desire


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

II.

In the first step of his argument, Kant tries to show that if there is a
categorical imperative, then there must be some object (or objects) that
all rational agents must hold to be unconditionally good. In the sec-
3
I am not convinced that this argument is successful. For detailed criticism of it, see
Kerstein (2002, 4754).
204 Samuel J. Kerstein

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

4
For a defense of this sort of objection, see Gaut (1997, 176 f.).
Deriving the Formula of Humanity 205

springs quickly to mind, namely everyones being happy. On the face


of it, Kants argument by elimination does not seem very promising.

III.

Although he acknowledges shortcomings in Kants argument by elim-


ination, Wood does not despair of Kants establishing that humanity
must be the ground of any categorical imperative (1999, 124). For, ac-
cording to Wood (1999, 125), Kant offers a positive argument for this
claim in the very next paragraph.
A key step in Kants positive argument, suggests Wood (1999, 127),
is an attempt to establish the following: each rational agent must (is
rationally compelled to) agree that in setting himself ends, he com-
mits himself to the view that his rational nature is unconditionally
valuable. Wood suggests that Kants defense of this key step involves
a regress on conditions. Inseparable from an agents setting ends for
himself is his holding that they have a certain value. But an agent can,
rationally speaking, place this value on his ends only on condition that
he hold his own rational nature to be unconditionally valuable.
When we look more closely, Wood suggests, we find that this key
step in Kants argument rests on three claims. The first has to do with
the kind of value the agent is committed to placing on the ends he
sets. Whenever an agent sets himself an end, goes the claim, he must
ascribe objective goodness to it. The second claim is that the agent
must hold the source of the objective goodness of his ends to be his
setting of them. What makes his ends good, he is rationally compelled
to affirm, is that they are the objects of the exercise of a certain capac-
ity, namely his rational choice. 5 Finally, if the source of the goodness
of objectively good ends is his rational choice (rational nature), then
the agent must hold his rational nature itself to be unconditionally
good.6 So, in sum, since a rational agent exercises his rational nature

5
That Wood attributes the fi rst two claims to Kant is manifest in the following pas-
sage: Thus Kants argument is based on the idea that to set an end is to attribute
objective goodness to it and that we can regard this goodness as originating only
in the fact that we have set those ends according to reason. The thought is . . . that
rational choice of ends is the act through which objective goodness enters the world
(1999, 129). See also (1999, 127 and 129 f.).
6
Wood makes clear that he attributes this claim to Kant when he states that Kant
does infer from the premise that rational nature is the source or ground of the objec-
tive goodness of all ends to the conclusion that rational nature itself is the underiva-
tive objective good, an end in itself (1999, 130). See also (1999, 127 and 129).
206 Samuel J. Kerstein

in setting ends, he must conclude that it is unconditionally valuable.


Even if we embrace this conclusion, we might wonder how to get from
it to the further point that a rational agent must conclude that every-
ones rational nature is unconditionally good.7 We also might wonder
how the exclusivity of rational nature as a possible ground for a cate-
gorical imperative gets established. That humanity is unconditionally
valuable does not in itself seem to entail that nothing else is. But these
are not issues for us to pursue here.
Also absent from our agenda is evaluating the philosophical plau-
sibility of the argument that Wood (and Korsgaard before him) at-
tribute to Kant. I find the argument (in both Wood and Korsgaards
versions) to be engaging and important, but as I have argued in detail
elsewhere, I think it has serious shortcomings.8 Here I concentrate on
the question of whether there are adequate textual grounds for attrib-
uting the argument to Kant in the first place. The regressive argument
may be Kantian in spirit. But I doubt whether he actually employs it
in the Groundwork.
To begin, let us consider the first main claim that, according to
Wood, Kant makes in defending the view that humanity is uncondi-
tionally valuable. For Kant, it seems, all cases of an agents setting an
end are cases of his exercising his capacity of rational choice (human-
ity). Kant implies that having humanity involves having the capacity
to set ends.9 The first claim is that whenever an agent sets an end, he
must (is rationally compelled to) ascribe objective goodness to it. He
must, says Wood, regard it as something of value universally for all
rational beings (1999, 127). Let us assume that in Kants view, when
an agent sets an end, he must take it to be good in some sense. Why
should we agree that in the Groundwork Kant embraces the view that
he must take it to be objectively good? According to Wood, at GMS,
412414, Kant maintains that goodness, whether moral or nonmoral,
is that which reason represents as practically necessary, and hence as
an object of volition for all rational beings (1999, 127). Something is
an object of volition for all rational beings just in case all such beings
must regard it as valuable, Wood implies. So, he suggests, at GMS,
412414, Kant commits himself to the view that if an agent sets him-
self an end, whether she views the end to be good morally or non-mor-
ally, she is rationally compelled to hold it to be objectively good.

7
Korsgaard (1996, 123) and Wood (1999, 131) each discuss this point.
8
Cf. Kerstein (2002, 4672).
9
Cf. MdST, 392.
Deriving the Formula of Humanity 207

I do not see where in the specified pages Kant commits himself to


this view.10 Among the central claims Kant there makes are the follow-
ing: All actions prescribed by an imperative (command of reason) are
in some sense practically necessary, that is, good. Actions prescribed
by a hypothetical imperative are good as means to some end; actions
prescribed by a categorical imperative are good in themselves.11 But it
does not follow from these claims that if an agent sets herself an end,
she is rationally compelled to hold it to be valuable for all rational
beings. Suppose an agent sets herself an end that, she realizes, is an
action commanded by a hypothetical imperative, but not a categorical
imperative. It is a necessary means to some further goal she has. She
would then have to agree that everyone is rationally compelled to hold
the following: if someone, including the agent herself, has this further
goal, then, as a means to this goal, it is good for that person to realize
the end. But the agent would not be rationally compelled to hold that
this end is something of value for each and every rational agent.
Let me illustrate this rather abstract point with an example. Sup-
pose Sally sets herself the end of traveling to New York. This action,
she realizes, is commanded by a hypothetical imperative, but not by
a categorical imperative. Traveling to New York is a necessary means
for visiting the Empire State Building, which is something she has had
an inclination to do for a long time. (The hypothetical imperative at
issue would be something like this: If you want to visit the Empire
State Building, then you ought to travel to New York.) If we grant
the claims set out above, Sally would have to agree that everyone,
including her, must hold the following: if a person has the end of visit-
ing the Empire State Building, then it is good as a means to that end
for that person to go to New York. But Sally would not be rationally
compelled to agree that the end of traveling to New York is something
that all rational agents must value.
In GMS, 412414, Wood suggests, Kant embraces the view that if
an agent sets himself an end he is rationally compelled to hold it to
be objectively good. Some central claims Kant makes there do not
entail this view. If there is a passage in GMS, 412414, that seems to
license this conclusion, it is perhaps the following: Practical good,

10
Thomas Hill, Jr. argues (2002, 244274) that Kant does not embrace the view that
in setting an end an agent always commits himself to the ends objective goodness. I
focus more than does Hill on the issue of whether specific texts in the Groundwork
provide evidence that Kant adopts the view in question.
11
In this context Kant is using imperative as a success term; he is conceiving of
imperatives as valid.
208 Samuel J. Kerstein

however, is that which determines the will by means of representa-


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

12
At GMS, 428, Kant calls relative ends, such as Sallys end to visit the Empire State
Building, grounds of hypothetical imperatives.
Deriving the Formula of Humanity 209

New York. Nothing here implies that in setting herself the end of vis-
iting New York, Sally must embrace the view that the end is valuable
to all rational agents. The interpretation on the table of GMS, 413,
harmonizes well with what follows. In the very next sentence Kant
suggests that practical good is distinguished from the agreeable, as
that which has influence on the will only by means of sensation from
merely subjective causes, those which are valid only for the senses of
this or that one, and not as a principle of reason, which is valid for
everyone (GMS, 413). An objective practical principle is, of course, a
a principle of reason, which is valid for everyone. It seems a stretch
to claim that in Groundwork, 412414, Kant endorses the view that in
setting an end, an agent must hold it to be objectively good.
Actually, near the beginning of the derivation of the Formula of
Humanity there is evidence that he does not endorse this view. Kant
says:
The ends that a rational being proposes as effects of its action at its dis-
cretion (material ends) are all only relative; for only their relation to a
particular kind of capacity of desire of the subject gives them their worth,
which therefore can provide no necessary principles valid universally for
all rational beings and hence valid for every volition, i. e., practical laws.
(GMS, 427 f.)

Let us suppose that a particular agent, John, correctly believes these


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

promote Johns end in cases in which doing so would not interfere


with any of the agents morally required action or pursuit of his own
material ends. In the passage immediately preceding his derivation of
the Formula of Humanity, not only does Kant fail to embrace the view
that in setting an end, an agent must hold it to be objectively good, he
implies that he rejects it.
The text of the Groundwork fails, I think, to imply that Kant em-
braces the first claim in the regressive argument for humanity as the
ground of a categorical imperative. Kant does not, I contend, adopt
the view that whenever an agent sets himself an end he must ascribe
objective goodness to it.13 If this contention is correct, then it is unsur-
prising that we find little evidence that he makes the second or third
claims. Where does Kant actually make the argument that an agent
must hold that the source of the objective goodness of his ends is his
exercising his capacity of rational choice (rational nature) in setting
them (claim 2)? Where does he actually make the argument that if
an agent must hold this, then he must also hold his rational nature
itself to be unconditionally good (claim 3)? According to Wood, Kant
makes these arguments in order to defend his contention at GMS, 429,
that each of us necessarily represents his existence as that of an end in
itself, that is, as that of something that is unconditionally valuable. I
believe that it was open to Kant to make the arguments. But I do not
find compelling textual evidence that he actually does.
Indeed, Kants claim that each of us necessarily represents his ex-
istence as that of an end in itself, as well as the surrounding text, ad-
mits of an interpretation that does not invoke the regressive argument
at all.14 In the paragraph at the end of which Kant sets out the Formula
of Humanity, he says:
The ground of [the moral principle] is: Rational nature exists as end in
itself: The human being necessarily represents his own existence in this
way; thus to that extent it is a subjective principle of human actions. But
every other rational being also represents his existence in this way conse-
quent on just the same rational ground that also holds for me;* thus it is
at the same time an objective principle from which, as a supreme practical
ground, it must be possible to derive all laws of the will. (GMS, 429)
13
Both Korsgaard (1996, 115) and Wood (1999, 128 f.) appeal to passages outside of
the Groundwork for support of their view that, according to Kant, if an agent sets
an end, he holds it to be objectively valuable. Both invoke, for example, Kants
discussion of goodness and well-being in The Critique of Practical Reason (KpV,
5861). For an interpretation of this discussion according to which Kant is not there
committing himself to the view in question, see Hill (2002, 262 ff.).
14
This contention challenges Korsgaard as well as Wood. See Korsgaard (1996, 122 f.).
Deriving the Formula of Humanity 211

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

15
See Anth, 127.
16
See GMS, 447 f.
212 Samuel J. Kerstein

tive would have to be humanity. If I am correct, he simply does not


present the more philosophically interesting regressive argument at-
tributed to him by Wood. However, things might be brighter than they
appear. Well after he unfolds his argument by elimination (but still in
Groundwork II), Kant sets out considerations that supplement and
reinforce it. In particular he appeals to the notion of a good will in
arguing that if there is a supreme principle of morality, then human-
ity, but not anything else, is an end in itself. Reflection on this appeal
will help us to see that Kants argument by elimination is stronger and
more interesting than it might seem to be.

IV.

At GMS, 437, Kant begins a summary of a main line of argument he


has developed up to this point in the Groundwork. Now we can end,
he announces, at the place from which we set out at the beginning,
namely with the concept of an unconditionally good will. In the next
paragraph, he says the following:
Rational nature is distinguished from the rest of nature by this, that it
sets itself an end. This end would be the matter of every good will [. . .]
Now, this end can be nothing other than the subject of all possible ends
itself, because this subject is also the subject of a possible absolutely good
will; for, such a will cannot without contradiction be subordinated to any
other object. (GMS, 437)

At GMS, 428, let us recall, Kant seems to dismiss precipitately three


candidates for the unconditionally good ground of a categorical im-
perative: inclinations themselves, the objects of inclinations, and
beings the existence of which rests not on our will but on nature.
In the passage just cited, Kant suggests an argument that might bol-
ster his elimination of these candidates. The argument takes shape
against the background of Kants view of the value that common
rational moral cognition attributes to a good will. In our ordinary
moral thinking, claims Kant, we hold that such a will is uncondition-
ally good. An impartial rational spectator would, we believe, judge
it to be good in every possible context in which it exists.17 Moreover,
we believe that a good will is preeminently good.18 That we believe
this presumably implies that, in our view, nothing that is devoid of a
17
See GMS, 393.
18
See GMS, 394 and 401.
Deriving the Formula of Humanity 213

good will is as good as something that has such a will. The argument
unfolds as follows. Whatever constitutes the ground of a categorical
imperative must be compatible with the value we attribute to a good
will (i. e., unconditional and preeminent goodness). But let us suppose
that any of the rivals to humanity were the ground of a categorical
imperative. We would have to acknowledge that a good will would (in
some circumstances) be subordinated to the rival and that a good
will would thereby fail to have the value we attribute to it. We would
land in contradiction. Therefore, we have license to dismiss any of the
rivals as possible grounds for a categorical imperative. As is perhaps
already evident, the material that is to supplement Kants argument
by elimination is material Kant develops before his presents the argu-
ment. I will not appeal to any positions Kant develops after he does
sofor example his view that humanity has dignity, that is, uncondi-
tional and incomparable value.
Filling in the supplemented arguments details requires elabora-
tion of the concept of a good will as well as of the notion of one things
being subordinated to another. The latter task is in a sense easier,
since Kant simply does not tell us precisely what this notion amounts
to. But I think it reasonable to assume that he would endorse the fol-
lowing claim: If x is subordinated to y, then x is less valuable than y
and we thereby have sufficient grounds to use x in whatever way is
necessary in order to maintain y. So for example, if plants are subor-
dinated to rational beings, then they are less valuable than rational
beings, and we have sufficient reason to harvest the former in order to
preserve the latter.
Regarding a good will, it suffices for our purposes to take note of
two ways in which Kant seems to employ the notion as it applies to us,
agents who can indulge their inclinations and thereby act contrary to
the moral law. According to the first usage, a good will is a particular
sort of willing or, what for him amounts to the same thing, of acting.
Kant writes of the unqualified [uneingeschrnkten] worth of actions
(GMS, 411), presumably of actions done from duty, which he has pre-
viously stated to have unconditional and moral worth (GMS, 400).
Since, according to Kant, the good will is good without qualification
[ohne Einschrnkung], it appears that sometimes good will refers to
a certain kind of action, that is, action done from duty.19 According to
19
This reading of good will would have to be broadened to accommodate Kants
view that perfectly rational beings such as God cannot act from duty. To them the
ought of duty does not apply, since their willing is necessarily in accord with the
law. See GMS, 414. We might attribute to Kant the view that these beings have a
214 Samuel J. Kerstein

a second usage of good will, it refers not to a particular kind of ac-


tion an agent might perform but rather to a kind of character she might
have. An agent has a good will on this usage just in case she is com-
mitted to doing what duty requires, not just in this or that particular
action, but overall. Presumably if an agent has this commitment, then
she will sometimes act from duty. (For example, she will invoke duty
as her incentive to do what is morally required in cases in which she is
tempted by her inclinations to act contrary to what morality demands.)
Kant intimates that having a good will amounts to having a certain
kind of character in the first paragraph of Groundwork I. Right after
suggesting that a good will is good without qualification, he tells us that
certain qualities of temperament, for example, courage or resolution,
are undoubtedly good and desirable for many purposes, but they can
also be extremely evil and harmful if the will which is to make use of
these gifts of nature, and whose distinctive constitution is therefore
called character, is not good (GMS, 393).20 Sometimes Kant employs
what we might, following Karl Ameriks (1989, 5459), call the whole
character conception of a good will. I believe Kant to be employing
the whole character conception of a good will in the passage we are
discussing (GMS, 437). In any case, it is that conception of a good will
that might help him to bolster his argument by elimination.
Let us now return to our (GMS, 437based) supplement to Kants
argument by elimination. Suppose that beings the existence of which
rests not on our will but on nature, say, species of wild animals, were
the ground of a categorical imperative, namely one commanding us
never to eradicate currently existing species of such animals. We would
be committed to the view that such beings were not only uncondition-
ally but also preeminently valuable. Otherwise, we might sometimes
lack sufficient grounds to abide by the principle not to eradicate them.
And if we might lack such grounds, we cannot take the principle to be
a categorical imperative.
We would presumably lack sufficient grounds to abide by the prin-
ciple in question when doing so would conflict with maintaining some-

good will (engage in unconditionally good willing) just in case they act for the sake
of the law. Presumably such beings are capable of doing this. And Kant does not
seem averse to the idea that acting from duty is a species of acting for the sake of
the law.
20
Later Kant is discussing a man who is by temperament cold and indifferent to oth-
ers, but who, from duty, acts beneficently. It is just then, says Kant, that the
worth of character comes out, which is moral and incomparably the highest (GMS,
398 f.). This passage suggests that good will refers not merely to a particular kind
of action, but to a kind of character that can be expressed in action.
Deriving the Formula of Humanity 215

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

such as, perhaps, the existence of persons. And if we might be without


adequate reason to conform to a principle, then we cannot take it to
be a categorical imperative. But if everyones being happy has both
unconditional and preeminent value, then a good wills value is obvi-
ously subordinate to it. (For Kant, of course, not everyone who has a
good will is happy and not everyone who is happy has a good will. See,
for example, GMS, 442.) The object constituted by everyones being
happy is worth more than a good will, and so we have sufficient reason
to destroy the latter in order to promote the former. But this conclu-
sion forces us into a contradiction. We are assuming along with Kant
that a good will is unconditionally and preeminently good. Yet a good
will cannot both be less valuable than everyones being happy and, as
the notion of its preeminence implies here, more valuable than it is.
Of course, the supplemented argument by elimination just brought
to bear against a version of utilitarianism is far from invulnerable to
attack. It is obviously open to philosophers to disagree with Kants
view that, according to ordinary moral thinking, a good will is un-
conditionally and preeminently good. Moreover, it remains to be seen
how the supplemented argument avoids dismissing Kants own candi-
date for the ground of a categorical imperative. How is it that human-
ity itself does not get eliminated from contention?
Initially, it seems that it would. After all, if humanity is to be the
ground of a categorical imperative, then we must hold it to be not
only unconditionally good, but also preeminently good. Otherwise we
leave open the possibility of there being other, better, uncondition-
ally good thingsthings that we would, rationally speaking, have to
preserve even at the cost of disobeying the principle grounded by the
value of humanity. But if we hold humanity to be preeminently as well
as unconditionally good, then we must, it seems, conclude that a good
will is subordinate to it, contradicting our assumption that the good
will is preeminently good.
Fortunately, this argument ignores the special relationship that ob-
tains between humanity and a good will. Every being who has a good
will necessarily has humanity. For having a good will entails having
capacities constitutive of humanity, for example, the capacity to act on
categorical imperatives. An agent cannot have an overall commitment
to do what duty requires according to Kant unless she has the capacity
to act on principles that specify what it requires, namely moral rules.
Much as it would be impossible to be an excellent classical pianist
without being a pianist, so it would be impossible to have a good will
without having humanity. With this point in view, we are able to see
Deriving the Formula of Humanity 217

that the supplemented argument by elimination does not, in the way


the objection alleges, throw the baby (humanity) out with the bath wa-
ter (rival candidates for the ground of a categorical imperative). If we
hold humanity to be preeminently good, then we believe that nothing
that is devoid of humanity is as good as something that has it. A good
will, of course, is not devoid of humanity. So holding humanity to be
preeminently good does not force us to conclude that a good will is
subordinate to it, that is, has less value than it does.
Here it makes sense to object that, even if this last point is correct,
humanity would get eliminated as a candidate for the ground of a cat-
egorical imperative. For the idea that humanity is unconditionally and
preeminently good contradicts the notion that a good will is uncondi-
tionally and preeminently good. Two different things cannot both be
preeminently valuable.
Given the notion of preeminence that (at least on my reading) Kant
employs, it turns out that two different things can both be preeminent-
ly good. Humanity and a good will are indeed two different things; for
one can have humanity without having a good will. Not every being
possessed of rational nature (humanity) is committed overall to do-
ing what moral principle requires. Some of us do not have excellent
character. But there is no contradiction in maintaining that humanity
is preeminently good and all the while holding that a good will is. Ac-
cording to the former claim, nothing that is devoid of humanity is as
good as something that has it. According to the latter, nothing that is
devoid of a good will is as good as something that has it. The latter
claim implies that humanity without a good will is not as good as hu-
manity with a good will. But there is no contradiction in maintaining
both that nothing devoid of humanity is as good as something with it
and, at the same time, that humanity devoid of a good will is not as
good as humanity with one. In much the same way, it is not self-con-
tradictory (though it may be false) to maintain both that no frying pan
that is not copper is as good as any frying pan that is copper and, at
the same time, to hold that every copper frying pan that is 3 mm thick
is better than any copper frying pan that is less thick.

V.

Even when supplemented as I suggest, Kants argument by elimination


appears to be closer to Kants intentions concerning the Groundwork
derivation of the Formula of Humanity than the regressive argument
218 Samuel J. Kerstein

attributed to him by Korsgaard and Wood. For not only does Kant ex-
plicitly embrace the reasoning that bolsters the argument (namely at
GMS, 437), but this reasoning itself relies only on concepts that Kant
develops before the derivation is complete.
Of course, the supplemented argument from elimination and thus
the derivation as a whole is, at best, only as convincing as Kants view
that a good will is unconditionally and preeminently valuable. This
view is surely in need of defense.21
But let me conclude by considering a further challenge to the sup-
plemented argument. If the reasoning in Part IV is sound, the argu-
ment avoids two pitfalls. It does not descend into self-contradiction;
for there is nothing inconsistent in maintaining as the argument does
that both humanity and a good will are unconditionally and preemi-
nently good. Moreover, the argument does not imply that a good will
is subordinated to humanity. So it avoids thereby eliminating human-
ity as a possible ground for a categorical imperative.
However, a serious question remains: What justification does Kant
have for holding that it is beings with rational nature who constitute
the ground of a categorical imperative, rather than maintaining that
it is merely beings with a good will who do so? What, for example,
permits Kant to reject the view that it is not all of us, but rather only
those of us with a good will, who never ought to be treated merely as
means? Consider a principle that Kant does not discuss, namely what
I call the Formula of the Good Will: So act that you treat a good will,
whether in your own person or in the person of any other, always at
the same time as an end, never merely as a means. It makes sense to
stipulate that, according to this formula, having a good will necessar-
ily involves having an overall commitment to doing what Kants For-
mula of Universal Law requires. For Kants derivation of the Formula
of Universal Law takes place well before the argument by elimination.
The supplemented argument by elimination seems to leave open the
possibility that a good will serves as a ground for the Formula of the
Good Will. So why must we say that it is humanity rather than a good
will that remains viable as a ground for a categorical imperative?22
21
But the supplemented argument does not depend on a further controversial view
Kant holds (GMS, 393), namely that only a good will is good without qualification.
22
In reply, one might focus attention on the fact that Kants derivation of the For-
mula of Humanity occurs after his derivation of the Formula of Universal Law. One
might then argue that Kant implicitly holds the following: whatever is to serve as the
ground of a categorical imperative must serve as the ground of a principle equiva-
lent to the Formula of Universal Law, that is, a principle that requires or permits
just those actions that are required or permitted by the Formula of Universal Law.
Deriving the Formula of Humanity 219

Our response to this objection depends on our interpretation of


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

But the good will would serve as the ground of a principle, namely the Formula of
the Good Will, that is not equivalent to the Formula of Universal Law. Therefore,
concludes the argument, the good will is not the ground of a categorical imperative.
Perhaps Kant would embrace this argument, but I do not find it philosophically
promising. Kant, of course, suggests that the Formula of Humanity is equivalent to
the Formula of Universal Law (GMS, 436). But I know of no substantive interpre-
tation of the two principles according to which they turn out to be equivalent. So
I fear that this sort of argument would likely not only eliminate the good will as a
possible ground of a categorical imperative, but humanity as well.
23
See GMS, 407 f.
24
See RGV, 36 f.
25
See, for example, GMS, 404, and KpV, 36. This view seems to me to be unpersua-
sive, as I explain in Kerstein (2002, 119129).
26
At this point, it might be tempting to make the following argument. Even if we are
sure that Sue does not have a good will now, it is open to her to develop one in the
future. On the basis of this potential she has, the Formula of the Good Will would
forbid us from treating her merely as a means. But this argument misses the mark.
For the Formula of the Good Will does not forbid us from treating beings with the
potential for having a good will merely as means, but rather from treating beings
with a good will merely as means. And here we are assuming that we know Sue not
to have a good will.
220 Samuel J. Kerstein

and a good will would constitute competing grounds for categorical


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

27
See IV above.
28
I would like to thank the other contributors to this volume for very helpful discus-
sion of this paper.
Deriving the Formula of Humanity 221

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesammelte


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

Other works

Ameriks, Karl (1989): Kant on the Good Will, in: Hffe, Otfried (Ed.):
Grundlegung zur Metaphysik der Sitten; Ein kooperativer Kommentar,
Frankfurt a. M., 4565.
Gaut, Berys (1997): The Structure of Practical Reason, in: Cullity, Garrett/
Gaut, Berys (Ed.): Ethics and Practical Reason, Oxford, 161188.
Hill, Thomas, Jr. (1992): Dignity and Practical Reason, Ithaca: Cornell Uni-
versity Press.
Hill, Thomas, Jr. (2002): Human Welfare and Moral Worth, Oxford; Oxford
University Press.
Kerstein, Samuel J. (2002): Kants Search for the Supreme Principle of Mo-
rality, Cambridge: Cambridge University Press.
Korsgaard, Christine M. (1996): Creating the Kingdom of Ends, Cambridge:
Cambridge University Press.
Wood, Allen W. (1999): Kants Ethical Thought, Cambridge: Cambridge
University Press.
Groundwork
III
Klaus Steigleder

The Analytic Relationship of Freedom


and Morality (GMS III, 1)1

The first part of the Third Section of the Groundwork belongs to that
genre of philosophical texts that are relatively simple to comprehend
in their general orientation yet considerably more difficult to compre-
hend in their details. I will proceed in the following manner. As the
first part introduces the third and final section of the Groundwork, I
will begin 1. by situating this Third Section in the context of the entire
work, clarifying its task and briefly surveying the functions of the first
part. 2. I will then attempt to explicate the fi rst part of Groundwork
III in greater detail. Given its brevity, I will quote the text in its en-
tirety in the course of my reflections. 3. In conclusion, I will discuss a
divergent interpretation of the text.

1. On the Position and Task of Groundwork III

The task of the Third Section of the Groundwork is the justification


of the supreme principle of morality. According to Kant, this is not
to determine the content of the moral principle, or to ascertain how
it reads. Rather, the task is to demonstrate that the moral principle is
binding for us and that, correspondingly, we have a strict obligation to
comply with it. In the First and particularly in the Second Section of
the Groundwork, Kant has shown that we can develop the idea of an
unconditional ought and that this idea leads us to a nonarbitrary and
clearly defined content of a moral principle. If there are binding moral
norms of action for us, then these conform to a specific moral princi-
ple. Initially, this moral principle is nothing other than an idea, i. e., a
thought-entity, and it could, as Kant writes, constitute a mere phan-
1
Translated from the German by Jo Ann Van Vliet, Ph. D., Tbingen, and checked by
Alexander Cotter (Stonehill College).
226 Klaus Steigleder

tom (GMS, 407). It could even be the case that no unconditional ought
exists, and that actually all norms, which we generally address as moral
norms, are nothing other than norms of prudence. These would then
constitute not an unconditional ought but only a conditional ought.
In order to better comprehend the task of Groundwork III, i. e.
to demonstrate that morality is not a phantom but that we can and
must obey the moral principle, we should recall why we are capable of
formulating a nonarbitrary idea of an unconditional ought and what
this idea entails.
We are able to conceive the idea of an unconditional ought because
we are conscious of the reality of conditional demands, of conditional
norms of action. In Groundwork II Kant develops his theory of norms
governing action (Kant himself does not speak of norms of action but of
imperatives) as a theory of practical judgments from the perspective of
an acting subject. Conditional norms of action are so structured that, for
the agent, the practical necessity exists of pursuing an end or undertak-
ing an action in order to pursue an end, if or because he or she has a cer-
tain actual end. Because I want to lose weight, it is necessary for me to
want to eat less chocolate candy or Because I want to avoid destitution
in old age, it is necessary for me to want to put money aside for retire-
ment are two examples of such practical judgments.2 The intention to
do or achieve something necessitates the intention to do something else
which is a means of attaining the original end. In this context, wanting
means volition, the actual resoluteness, the real intention of pursuing an
end or undertaking an action in order to pursue an end.3 Since we do not
automatically want to do what it is necessary for us to want to do on the
basis of our actual ends (we would like to continue to enjoy chocolates
with abandon, although we want to lose weight, and we would prefer to
spend rather than to save money, even though we do want to avoid des-
titution in old age), the practical necessity of wanting to do something,
arising out of our actual ends, confronts us as a demand, a necessitation,
an ought. Kant differentiates two types of conditional norms of action,
namely, technical norms and prudential norms. Technical norms are de-
pendent on ends that we can have; prudential norms, in contrast, are
dependent on an end that all those beings that are under the influence
of feelings of pleasure and displeasure inevitably have, to wit, the end of
ones personal well-being, or, as Kant says, ones own happiness.
2
Kant speaks of propositions (Stze) rather than judgments. Propositions are,
however, in his opinion, assertoric judgments. See Ent, 193,33; Log, 30, note 3,
109,1122.
3
Kant designates such real intentions to act as maxims.
The Analytic Relationship of Freedom and Morality 227

Given the indeterminate nature of this end, we continually inter-


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

4
For a more extensive discussion see Steigleder (2002, 2358, Chapter 2: Hypo-
thetische Imperative).
5
Happiness is the satisfaction of all our inclinations (extensive, with regard to their
manifoldness, as well as intensive, with regard to degree, and also protensive, with
regard to duration) (KrV, B 834 /A 806).
228 Klaus Steigleder

ingly, attribute an absolute worth to himself or herself. Such a being


must assume that, on the basis of his or her faculty of pure practical
reason, he or she is capable of determining ends on his or her own
accord which take into account the unconditionally necessary end as
which he or she necessarily exists, because he or she, on the basis of
his or her faculty of reason, is able to establish ends on his or her own
accord. Correspondingly, there is an unconditionally practical neces-
sity for him or her to take into account the unconditionally necessary
end as which he or she, and every other being equally capable of ac-
tion, exists himself or herself in all action.
This is, so to speak, the quintessence of the aspects and considera-
tions through which Kant develops the idea of an unconditional ought
in the Second Section of the Groundwork. Kant does not, however,
speak explicitly of pure practical reason in this context.6 This is in
striking contrast to the beginning of the Second Section in which Kant
critically valorizes the conclusions of the First Section over the philos-
ophy of his day. Kant reserves the Third Section of the Groundwork
for this quintessence, which he presents to a considerable extent, as
we will see, in the first part of Groundwork III under discussion here.
To be sure, it is important to observe that the requisite considerations
have already been provided by Kant within the context of his expli-
cation of the idea of a categorical imperative. There, Kant preceded
his analysis of the norms of action, the hypothetical and categorical
imperatives, for example, with the description of his intention: []
we must follow and present distinctly the practical faculty of reason,
from its general rules of determination to the point where the concept
of duty arises from it (GMS, 412,2225).
Earlier, Kant had emphasized, in conjunction with the conclusions
of the First Section, [] that all moral concepts have their seat and
origin completely a priori in reason (GMS, 411,8 f.; my emphasis).
And in the course of the actual explication of the idea of an uncondi-
tional ought he writes:
Here, however, it is a question of objective practical laws and hence of the
relation of a will to itself insofar as it determines itself only by reason; for
then everything that has reference to the empirical falls away of itself,
since if reason entirely by itself determines conduct (and the possibility

6
That is, from GMS, 412,26 and particularly from GMS, 420,24 on. In both explicit
references to pure practical reason following GMS, 412,26 (namely, GMS, 440,25 f.
and 445,1115) the task of a critique of pure practical reason is mentioned and it
is thereby referred to GMS III.
The Analytic Relationship of Freedom and Morality 229

of this is just what we want now to investigate), it must necessarily do so a


priori (GMS, 427,1318).

The further determination of this self-direction of a will exclusively


determined by reason leads to the concept of a necessary end valid
for all rational beings:
Now, what serves the will as the objective ground of its self-determination
is an end, and this, if it is given by reason alone, must hold equally for all
rational beings (GMS, 427,2124; only end (Zweck) is emphasized in
the original).
But suppose there were something the existence of which in itself has an
absolute worth, something which as an end in itself could be a ground
of determinate laws; then in it, and in it alone, would lie the ground of a
possible categorical imperative, that is, of a practical law (GMS, 428,36;
original emphasis).

Kant develops this even further in the idea of a will which is giving
universal law and subject only to the law which it gives to itself, sum-
marizing that the autonomy of the will is the supreme principle of
morality.
But that the above principle of autonomy is the sole principle of morals
can well be shown by mere analysis of the concepts of morality (GMS,
440,2830).

As this quotation evinces, the idea of an unconditional ought (or of


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

7
Cf. GMS, 406,1425; 441,17 f.
230 Klaus Steigleder

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

8
For a more comprehensive discussion, see Steigleder (2002, 2439).
The Analytic Relationship of Freedom and Morality 231

According to Kant, three kinds of practical judgments are conceiv-


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

9
Here action is to be understood in the broad sense of voluntary action respectively
inaction directed toward an end. Maxims can also have inaction (omission) as their
subject.
10
Hypothetical imperatives can call for the rejection of certain maxims (to not do
things that I want to do) as well as the pursuit of certain goals. Here it is not my
intention to outline all the forms of hypothetical imperatives but simply to indicate
their basic structure or form.
232 Klaus Steigleder

truth of the moral principle as a synthetic practical judgment a priori,


does not, as in the Critique of Pure Reason, lie in the determination
of the scope of the use of theoretical reason independent of experi-
ence but first and foremost in the establishment of the reality of the
faculty of pure practical reason. Thus, it is not unforeseen when Kant
opens the Third Section of the Groundwork with a part entitled The
Concept of Freedom is the Key to the Explanation of the Autonomy
of the Will, and then in the second part attempts to defend the thesis
Freedom must be Presupposed as a Property of the Will of All Ra-
tional Beings.
Nevertheless, the prospects are not promising for Kants envisioned
proof of the validity of the moral principle in Groundwork III. For
the establishment of the reality of the faculty of pure practical reason
would appear to be tantamount to the attempt to establish freedom in
the strict sense. Yet one conclusion from the Critique of Pure Reason
is precisely that freedom in the strict sense or transcendental freedom
cannot be proven. Transcendental freedom is conceivable. As it can-
not be an object of experience, and as experienceable nature as such is
appearance and not a self-existent reality, it can without contradiction
be conceived together with the necessity that in experienceable nature
all alterations occur in accordance with the law of the connection of
cause and effect (KrV B 232, see also B 566586 [A 538558]). But
it is neither possible to prove nor disprove the existence of transcen-
dental freedom. If it were necessary to prove the existence of pure
practical reason in the course of the proof of the validity of the moral
principle, then, on the basis of the conclusions of the Critique of Pure
Reason, it is evident from the very beginning that the validity of the
moral principle cannot be proven.
For this reason, Kant does not make the slightest effort to prove the
existence of freedom in Groundwork III. Rather, he will attempt to
demonstrate that even a less ambitious undertaking is fully sufficient.
It is not requisite to establish that we are free in the strict sense (that
we possess the faculty of pure practical reason); rather it is requisite to
establish that we must hold ourselves to be free (that we must assume
that we possess the faculty of pure practical reason). The latter can
be true, even if the former (that we are free in the strict sense) were
false; or if we, as is the case, cannot ascertain its truth or falseness. If
it is, however, true that we must assume that we have the faculty of
pure practical reason, then we must further assume that we constitute
an unconditionally necessary end and that we can and must take this
unconditional end into account in all our action, in summary: we must
The Analytic Relationship of Freedom and Morality 233

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

2. The Argumentation of Part 1 in Groundwork III

The first part of Groundwork III can be divided into two segments.
The first of these, which is comprised of the first two paragraphs, has
the task of explicating the thesis in the title The Concept of Free-
dom is the Key to the Explanation of the Autonomy of the Will, and
culminates in the statement that hence a free will and a will under
moral laws are one and the same. (GMS, 447) The second segment,
which the third and final paragraph comprises, addresses the task and
the problem of the justification of the moral principle as a synthetic
practical judgment a priori.
Let us begin with the argumentation in the first segment. The first
paragraph reads:
Will is a kind of causality of living beings insofar as they are rational, and
freedom would be that property of such causality that it can be efficient
independently of alien causes determining it, just as natural necessity is
the property of the causality of all nonrational beings to be determined to
activity by the influence of alien causes (GMS, 446,712).

This sentence contains in condensed form essential elements from


Kants theory of action. What Kant calls will here, he calls practical
reason or free choice (freie Willkr) elsewhere. Will is initially to
be differentiated in two directions. On the one hand, it is to be distin-
guished from the unfree faculty of desire (arbitrium brutum) possessed
by animals, and on the other, from pure practical reason. Common to
both animals and human beings, Kant asserts, is their possession of
the faculty of desire (Begehrungsvermgen). Kant understands this
as the faculty entailing an inner causation of the behaviour of a being
through psychic respectively mental contents (representations). A bil-
liard ball rolls as a result of an external propulsion; an animal begins
234 Klaus Steigleder

to move out of an inner impulse, because it has, for example, seen or


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

11
Life is the faculty of a being to act in accordance with laws of the faculty of desire.
The faculty of desire is a beings faculty to be by means of its representations the
cause of the reality of the objects of these representations. (KpV, 9,1922; original
emphasis); see also MdSR, 211,79: The faculty of a being to act in accordance
with its representations is called life (Original emphasis).
The Analytic Relationship of Freedom and Morality 235

grounds of the causality of a being insofar as its existence is determina-


ble in time and therefore under the necessitating conditions of past time,
which are thus, when the subject is to act, no longer within his control
and which may therefore bring with them [] natural necessity [] (KpV,
96,1937; the phrase causality determined in accordance with a natural
law is not italicized in the original).

Yet at the beginning of the first part of Groundwork III Kant imme-
diately focuses on free choice, or the will, under the hypothesis that
this free choice possesses the faculty of freedom in the strict sense
(and freedom would be that property of such causality that it can be
efficient independently of alien causes determining it). If we think
of free choice as free in this sense, then we must not think of it as
pure practical reason, i. e., we must not equate free choice with pure
practical reason but must think of it as being endowed with the fac-
ulty of pure practical reason an insight which is significant for the
comprehension of the line of argumentation in Groundwork III in
general. If free choice were equatable with pure practical reason, a
being endowed with such a faculty would necessarily determine itself
to activity through pure practical reason (would necessarily do what is
morally good). This is clearly not the situation of finite living beings,
and, in any event, not our situation insofar as we possess the faculty
of pure practical reason, because otherwise the moral law would not
have to confront us as a demand, an imperative. If our capacity of
choice possesses the faculty of pure practical reason, then we are al-
ways capable of determining ourselves to activity independently of
alien [] determining causes.
If we thoroughly contemplate the implications of this negative de-
termination of freedom in the strict sense, we realize that we have
also been directed toward a positive content. In certain respects this
repeats the establishment of an unconditional ought in Groundwork
II. This originally consisted in negating the essential properties of a
conditional ought. However, it proved to have given grounds for posi-
tive determinations.
The preceding defi nition of freedom is negative and therefore unfruitful
for insight into its essence; but there flows from it a positive concept of
freedom, which is so much the richer and more fruitful (GMS, 446,1315;
original emphasis).

The positive concept of freedom is conveyed through the conception


of free choice as a causality. If, namely, free choice can be efficient
independently of alien causes determining it, then the capacity to be
236 Klaus Steigleder

the cause of such effects must be inherent without this cause itself in
turn being comprehensible as the effect of a foreign, preceding, and
determining cause. Since the concept of causality is not an experien-
tial concept and, consequently, is to be comprehended as entailing
a necessary relationship between cause and effect,12 we cannot con-
ceive the relationship between an original (uncaused) cause and its
effect as a lawless relationship replacing the naturalistic progression
of cause and effect but rather as a lawful relationship of its own kind.
A different law supplants natural law, namely, the law of pure practi-
cal reason.
Since the concept of causality brings with it that of laws in accordance with
which, by something that we call a cause, something else, namely an effect,
must be posited, so freedom, although it is not a property of the will in ac-
cordance with natural laws, is not for that reason lawless but must instead
be a causality in accordance with immutable laws but of a special kind;
for otherwise a free will would be an absurdity. Natural necessity was a
heteronomy of efficient causes, since every effect was possible only in ac-
cordance with the law that something else determines the efficient cause to
causality; what, then can freedom of the will be other than autonomy, that
is, the wills property of being a law to itself? (GMS, 446,15447,2)

The will has the capability to establish ends and to actively pursue
ends independent of foreign determining influences. Given the fact
that such independent establishment and pursuit of ends do not ap-
pear out of nowhere, but are brought about by reason, they must also
be determined by the characteristics of reason, specifically, universal-
ity and necessity. Reason that establishes ends necessarily observes
the law, which it is to itself, is autonomy. This thought can be devel-
oped further with various accentuations. It can be stressed that rea-
son, which establishes ends, can constitute an unconditionally nec-
essary end and, correspondingly, an absolute worth for itself. Kant
himself sets a different accent:
But the proposition, the will is in all its actions a law to itself, indicates
only the principle, to act on no other maxim than that which can also have
as object itself as a universal law. This, however, is precisely the formula
of the categorical imperative and is the principle of morality; hence a free
will and a will under moral laws are one and the same (GMS, 447,27).

12
We must assume (expressed in a somewhat simplified manner here) that in nature
every change is the effect of a cause, which necessarily produces this effect. Which
cause produces which effect is a matter which we, however, must attempt to ascer-
tain through experience.
The Analytic Relationship of Freedom and Morality 237

Before I briefly address the relationship between the various accen-


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

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

This thesis of the analytic relationship between the freedom of the


will and the moral principle is, as shown above, not surprising in light
of the conclusions of Groundwork II. These culminated in a non-arbi-
trary idea of the unconditional ought. As Kants considerations in the
first section are supposed to show, this idea is grounded in the idea of
a free will proper.
We are still proceeding on the level of mere thought. This is why
the justification of the moral principle consists, as also shown above,
in establishing the moral principle as a synthetic practical judgment
a priori. And we have already attempted to comprehend that this is
actually tantamount to the task of establishing the reality of freedom,
or, as we can now say, of a free will (in the strict sense), although this
is not directly possible (in the form of a proof of freedom). Ultimately,
this is exactly what Kant says in the final paragraph of Part 1. There
he continues with the following remark:
But the principle of morality that an absolutely good will is that whose
maxim can always contain itself regarded as a universal law - is never-
theless always a synthetic proposition; for, by analysis of the concept of
an absolutely good will that property of its maxim cannot be discovered
(GMS, 447,1014).

This assertion may appear to be rather curious at first glance. Is this


not exactly what Kant tried to show in Groundwork I, namely, that the

13
It is important to note in general that the investigation which has the task of devel-
oping the fundamental characteristics of the content of the moral principle is not
the Groundwork, but The Metaphysics of Morals of 1797. There, Kant establishes
that the moral principle justifies a number of orderly sub-principles. Correspond-
ingly, for Kant it is not at all the task to always directly judge our actions and max-
ims on the basis of the moral principle. See especially Steigleder (2002, 129274,
Teil B: Moralphilosophie als Rechts- und Tugendlehre).
The Analytic Relationship of Freedom and Morality 239

idea of an absolutely good will leads to the idea of a moral principle?


And are Groundwork II and the first two paragraphs of Groundwork
III not an eloquent testimonial to the richness of the idea of the un-
conditional ought, and the cross-referentiality of its various aspects?
The solution to this problem presumably lies in the fact that here Kant
does not only focus on the absolutely good will as an idea and, cor-
respondingly, that the determination of the capacity of the maxims of
this will does not hypothetically follow. The judgment (proposition)
differs on this point from the introductory sentence, which reads: and
freedom would be that property of such causality that it can be effi-
cient independently of alien causes determining it (GMS, 446,810;
original emphasis), as well as from the first sentence of the paragraph
under discussion, which opens with the stipulation If, therefore,
freedom of the will is presupposed [] What cannot be derived by
analysis of the concept of an absolutely good will is the real capacity
of a subjective determination of the will to take the moral law as such
into account, and therefore the reality of the faculty of pure practical
reason and consequently of the moral law.
Thus, it is requisite to prove the existence of the moral principle as
a synthetic practical judgment a priori. While an analytic (theoretical)
judgment in the simplest case of a categorical judgment makes a char-
acteristic (Merkmal) explicit in the predicate term which was already
conceived in the subject term, a synthetic judgment cannot remain on
the internal level between subject term and predicate term.14 For here,
namely in the predicate term, a new (not yet conceived) characteristic
is ascribed to the subject term, and so the question arises what (which
third entity in relation to the subject and predicate term) allows the
attribution of such a new characteristic to the subject term.15 In the
case of synthetic judgments a posteriori it is experience, in the case of
synthetic theoretical judgments a priori these are, according to Kant,
the forms of intuition, time and space. Kant draws parallels between
the problem of synthetic practical judgments a priori and of synthetic
theoretical judgments a priori and also inquires about the third cogni-
tion which is capable of overcoming the mere internal relationships
of conceived determinations of volition and allows the attribution of
reality to the determinations of unconditional volition and uncondi-
tional ought. A viable possibility for this third cognition is, on the
one hand, freedom, as shown above. On the other hand, freedom is

14
See KrV, B 10 f./A6 f.
15
See KrV, B 12 f./A9 f.
240 Klaus Steigleder

not a true parallel to the forms of intuition, time and space. For us,
freedom is something, the reality of which is neither ascertained nor
ascertainable.
Such synthetic propositions are possible only in this way: that the two cog-
nitions are bound together by their connection with a third in which they
are both to be found. The positive concept of freedom provides this third
cognition, which cannot be, as in the case of physical causes, the nature
of the sensible world (in the concept of which the concepts of something
as cause in relation to something else as effect come together) (GMS,
447,1420).

Kant, thus, does not speak of freedom as the third cognition being
sought but rather purports that the positive concept of freedom pro-
vides or creates (schafft) this third cognition. This third cognition,
as Kant recapitulates with reference to his preceding reflections on
the positive concept of freedom, is to be distinguished from the causal
relationships constituting experienceable nature (and thus independ-
ent of a time structure). Of interest here is Kants assertion that the
concept of freedom provides or creates this third cognition. This is
initially quite surprising since his task seems to be the departure from
the level of mere ideas. In actuality, however, as addressed above,
Kants solution will consist in establishing that the idea of freedom
cannot be a mere idea for us but that we must necessarily consider
ourselves to be free in the sense of the positive concept of freedom.
According to Kant, it will also be crucial to show that we must nec-
essarily comprehend ourselves as rational beings with wills, that we
cannot comprehend our sensible nature as our real nature, and that
we must attribute ourselves with the faculty of pure practical reason.
Consequently, Kant concludes the first section by maintaining:
What this third cognition is, to which freedom points us and of which we
have an idea a priori, cannot yet be shown here and now; nor can the de-
duction of the concept of freedom from pure practical reason, and with it
the possibility of a categorical imperative as well, as yet be made compre-
hensible, some further preparation is required (GMS, 447,2025).

It will eventually become evident that we must regard ourselves and


every other being capable of action as an unconditionally necessary
end because we cannot consider our practical evaluations and our
ends as being limited to interests which issue from feelings of pleas-
ure and displeasure. This affirms what Kant could only present as a
postulate in Groundwork II, namely, that every rational being must so
conceive his or her existence that he or she exists as an end in himself
The Analytic Relationship of Freedom and Morality 241

or herself.16 Kant had envisioned the proof as a task of Groundwork


III. The course of the appropriate line of argumentation is established
in its fi rst part.

3. A Critical Discussion of Dieter Schneckers Interpretation

Dieter Schnecker supports an interpretation of the fi rst part of


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

16
See GMS, 429,27, 35 f.
17
Schnecker (1999, 147195).
242 Klaus Steigleder

sensible-rational being, whose will is, however, considered merely


as an intelligible will necessarily obeys the moral law (Schnecker,
1999, 160).18 Thus Kant, in Schneckers interpretation, wishes to em-
phasize that a free will and a rational or moral will respectively a will
subject to the principle of morality are interchangeable concepts.
This understanding of the thesis of analyticity is, in his eyes, the
only reading [] which allows the placement of the text in a coherent
interpretive framework (Schnecker, 1999, 154).
This is, of course, a bold assertion. To begin with, neither the limi-
tation of the final step of the argumentation exclusively to Part 3 nor
the characterization of the aim of Part 4 (ostensibly the initial reali-
zation of deduction in Kants understanding) are as unequivocally
clear as Schnecker purports. There may conceivably be other read-
ings without having to impute total confusion to Kant.19 Moreover,
Schneckers interpretation of the thesis of analyticity is not di-
rectly suggested by the text. Schnecker expends considerable effort
in showing that another understanding of the introductory sentence
than the obvious one is possible: Living beings must not necesssar-
ily mean corporeal beings but can mean fully rational beings for Kant,
as Schnecker verifies on the basis of lecture transcriptions. Addition-
ally, Schnecker must acknowledge that Kant is not totally without
responsibility for the (in Schneckers opinion) faulty interpretation
of the thesis of analyticity.20 Kant himself speaks of the categorical
imperative in conjunction with the thesis of analyticity, although
it would have been better not to have done so. Finally, Schnecker
must assert that Kant presupposes that for a sensible-rational being
the proof of his or her possession of freedom in the strict sense would
not render the proof of the necessity for action in accordance with this
freedom unessential. Even if it is successfully established that we can
act freely and thus morally, it does not automatically follow that we
ought to act morally (Schnecker, 1999, 69).21
In my opinion, this thesis does not appear to be consistent with
Kants comprehension of the structure of imperatives. This structure,
as outlined above, can be summarized as follows: for a being capable
of action a) in accordance with his or her faculty of reason the practi-
cal necessity exists of acting in a certain manner; this being, however,
b) does not automatically conduct himself or herself as would be ap-
18
Cf. also Schnecker (1999, 157): If a being acts freely, then it acts morally.
19
For an example see my own interpretation in Steigleder (2002).
20
See Schnecker (1999, 161 f.).
21
See also Schnecker (1999, 79, 154, and 157, for example).
The Analytic Relationship of Freedom and Morality 243

propriate to his or her faculty of reason, and therefore, c) the practical


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

freedom terminologically in the distinction between Wille (will)


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

22
See MdSR, 213 f.
23
Cf. KrV, B830 / A802.
24
Cf. MdSR, 213 f.: Freedom of choice is this independence from being determined
by sensible impulses; this is the negative concept of freedom. The positive concept
of freedom is that of the ability of pure reason to be itself practical. But this is not
possible except by the subjection of the maxim of every action to the condition of
its qualifying as universal law.
The Analytic Relationship of Freedom and Morality 245

would, on the other hand, be senseless if one were to understand the


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

25
See, e. g., GMS, 331: Hence the will is not merely subject to the law but subject to
it in such a way that it must be viewed as also giving the law to itself [].
26
See GMS, 433 f.: A rational being belongs as a member to the kingdom of ends when
he gives universal laws in it but is also himself subject to these laws. He belongs to it
as sovereign when, as lawgiving, he is not subject to the will of any other.
A rational being must always regard himself as lawgiving in a kingdom of ends
possible through freedom of the will, whether as a member or as sovereign. He can-
not, however, hold the position of sovereign merely by the maxims of his will but
only in case he is [a] completely independent being and with unlimited resources
adequate to his will.
246 Klaus Steigleder

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesam-


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

Other works

Schnecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kate-
gorischen Imperativs, Freiburg i. Br.
Steigleder, Klaus (2002): Kants Moralphilosophie. Die Selbstbezglichkeit
reiner praktischer Vernunft, Stuttgart.
Corinna Mieth and Jacob Rosenthal

Freedom must be presupposed as a property of


the will of all rational beings (GMS III, 2)

Section III of Kants Groundwork is subdivided into five (or six, in-
cluding the Concluding remark) parts. The task of this section is to
show that a categorical imperative is possible, a task that has at least
two aspects. The first concerns justification: Kant aims to show that
and why the moral law is a valid principle for any rational will and
in its imperative form a binding principle for our will, and no chi-
merical idea or phantom (GMS, 445). The second aspect concerns
motivation: how our will and actions can be determined by the moral
law independently of any inclinations we may happen to have, how
it is possible that we take an interest in the moral law (GMS, 460).
The second question cannot be answered: [] it is impossible for us
to explain [] how pure reason can be practical (GMS, 461).1
As a first step, Kant focuses on a rational will as such (parts 1 and
2). Subsequently, he confronts the difficulty surrounding the fact that
our human wills are not purely rational and it is therefore not prima
facie guaranteed that what is valid for a rational will as such is in the
same way also valid for our wills (parts 35). In our commentary, we
focus on the interpretation of part 2 (GMS, 447448). To begin with,
we give a short summary of part 1 in order to sketch the background
of our passage. Then we carry out our task in two steps: the first is a
reconstruction of the argumentation of part 2, the second is a discus-
sion of five main problems of interpretation with regard to it. These
include the ambiguity of the concept of freedom, the significance of
the first-person perspective, the kind of reason to which Kants argu-
ment applies, the relation between freedom and indeterminism, and
last but not least the extent to which Kants argument is successful, in
other words, what it actually establishes.
1
For a thorough discussion of the meaning of the question how a categorical impera-
tive is possible see Schnecker (1999, ch. 2).
248 Corinna Mieth and Jacob Rosenthal

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

2
Cf. Allison (1990, 2002).
3
When we speak of rational being we follow Mary Gregors translation of ver-
nnftiges Wesen.
Freedom must be presupposed as a property of the will of all rational beings 249

mentative steps can be distinguished in our passage. In a first step


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

I. Reconstruction

Step 1:

It is not enough that we ascribe freedom to our will on whatever ground,


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

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

and that belongs to anthropology (GMS, 389), this road is closed,


and an a priori argument for freedom must be given that is valid for
any rational will. Freedom has to be proved as a property of beings
that are a) rational and b) have a will. So in part 1, Kant hypothetically
presupposed freedom and arrived at morality analytically. Now he
sets up the task of proving freedom from rationality, and so in the end
he moves from rationality to morality. An outline of the completed
argument would run as follows:
1. Every rational being has a free will.
2. A free will is not determined by alien causes or according to alien
laws.
3. A free will is an autonomous will.
4. Autonomy is the property of the will to be a law to itself.
5. The fact that a will is a law to itself means that it acts on a maxim
that can be willed as a universal law through itself, which is the
principle of morality.
6. An autonomous will is a will under the moral law.
7. The will of every rational being is a will under the moral law.

The argumentative program for the passage under discussion is:


1. Morality is valid for us insofar as we are rational beings.
2. Morality must be valid for all beings insofar as they are rational.
3. Morality follows from freedom as a property of the will of a ra-
tional being.
4. The reason to ascribe freedom to a will must be the same for all
rational beings.
5. Freedom must be proved as a property of beings that are a) rational
and b) have a will.

Step 2:

I say now: every being that cannot act otherwise than under the idea of
freedom is just because of that really free in a practical respect, that is, all
laws that are inseparably bound up with freedom hold for him just as if
his will had been validly pronounced free also in itself and in theoretical
philosophy. [In a footnote, Kant restates this in greater detail:] I follow
this route that of assuming freedom, sufficiently for our purpose, only
as laid down by rational beings merely in idea as a ground for their actions
so that I need not be bound to prove freedom in its theoretical respect
as well. For even if the latter is left unsettled, still the same laws hold for a
being that cannot act otherwise than under the idea of his own freedom as
Freedom must be presupposed as a property of the will of all rational beings 251

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

Step 2 consists in a claim expressed in the fi rst half-sentence, and its


following clarification and justification. Kant does not aim at what he
calls a theoretical proof for freedom: a straightforward argument to
the effect that any rational will is a free will. Such an argument would
be difficult or impossible to give, but that does not matter, because for
Kants purposes it is dispensable. Instead, he claims that any being
that has to view itself as free when acting (that has to act under the
idea of his own freedom) may be considered to be free in a practi-
cal sense. That is, the being may be considered to be subject to the
same laws as if it had been proved that it is free. What it means that
any rational agent has to view himself as free will become clear in
step 3, when Kant presents his argument for this claim. What is clear
by now is that Kants argument is bound up with the perspective of
the first person, because any rational agent has to presuppose his own
freedom. The topic is freedom [] as laid down by rational beings
[] as a ground for their actions. The practical respect mentioned
here as opposed to the theoretical respect characterizes the kind of
proof Kant is going to give for the freedom of any rational being. If
any such being has to view itself as free, it also has to admit that it is
subject to any law that goes along with freedom, in particular, it has to
accept the moral law as a valid law for its will.

Step 3:

Now I assert that to every rational being having a will we must necessar-
ily lend the idea of freedom also, under which alone he acts. For in such
a being we think of a reason that is practical, that is, has causality with
respect to its objects. Now, one cannot possibly think of a reason that
would consciously receive direction from any other quarter with respect
to its judgments, since the subject would then attribute the determination
of his judgment not to his reason but to an impulse. Reason must regard
itself as the author of its principles independently of alien influences; con-
sequently, as practical reason or as the will of a rational being it must be
regarded of itself as free, that is, the will of such a being cannot be a will of
his own except under the idea of freedom, and such a will must in a practi-
cal respect thus be attributed to every rational being. (GMS, 448)

As in step 2, Kant starts with a claim and gives the justification after-
wards. The claim reformulates the title of part 2: Freedom must be
presupposed as (not: is) a property of the will of all rational beings.
252 Corinna Mieth and Jacob Rosenthal

Presupposed by whom? In the first place, freedom must be presup-


posed by the rational being itself who is acting, as we have already
seen, i. e. freedom must be presupposed from the first-person perspec-
tive. What Kant additionally seems to claim is that we (from the
perspective of the third person) must ascribe (necessarily lend) the
idea of freedom to any agent who has to view himself as free (acts
under the idea of freedom).
There are two possibilities of interpretation for this, a weak and
a strong one. The weak one says that this is not really an additional
claim. We simply discover that the agent has to consider himself to be
free and in this sense lend him the idea of freedom. The strong one
says that we have to consider him to be free. We have to introduce a
distinction here, because there is a certain ambiguity in terminology.
If we speak of a rational being we might merely think (weakly) of a
being that engages in practical deliberations. Any such being has to
consider itself to be free, as Kant subsequently shows, but of course
we dont have to view it so. We can ascribe its practical judgments to
mere impulses, while the being itself assumes at the same time that it
is acting according to reason. But, if we speak of a rational being, we
might also think (strongly) of a being endowed with a purely rational
or perfect will, and such a being would of course have to be considered
to be free by any spectator and not just by itself. The point is that we
cannot know if we are rational beings in this strong sense. But what
we know is that we are rational beings in the aforementioned weak
sense: We engage in practical reasoning and view our practical judg-
ments as causing our actions. In doing so, we have to view ourselves as
being guided by reason and therefore as free. We do not have to think
that we are rational throughout (we know that we are not), but when
we deliberate what to do and come to a considered judgment, we have
to ascribe this judgment to reason and so hold ourselves to be free at
least in this respect.
Kant states that if we think of a rational being with a will, we have
to assume that it is provided with practical reason: a reason that has
causality with respect to its objects. In this assumption, Kant identi-
fies the will of a rational being with practical reason: We thereby think
of a will as determining what is chosen or what is done in a reasonable
way. The latter is important, because Kant does not think of a will
here in the sense that it has a freedom to choose whatever it likes in
an indeterminate way. The formulation that the will has causality
with respect to its objects rather means that we have a reason for an
action, and that the action is really due to this reason, and not due
Freedom must be presupposed as a property of the will of all rational beings 253

to some other external cause, e. g. an impulse. We have to remember


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

these deliberations are theoretical (what to believe) or practical (what


to do). As the passage Reason must regard itself as the author of its
principles independently of alien influences; consequently, as practical
reason [] (emphasis by us), and also the foregoing sentence Now,
one cannot possibly think of a reason that would consciously receive
direction from any other quarter with respect to its judgments []
indicate, Kant is aware of this. Of course, his main interest is practical
reason, consequently most of his formulations refer to the will, and so
it may be disputed that he has more in mind here than practical reson,
namely, reason in general. But at least the systematic fact remains that
Kants argument, if valid at all, is as valid for theoretical deliberations
and judgments as for practical ones. We will come back to this in the
third comment.
So, any rational (in the sense of deliberating) being has to consider
itself to be autonomous, to be free in the positive sense, at least as far
as its actual deliberations and judgments are concerned. In particular,
it has to view itself as acting according to the laws of reason, and it has
to view these laws as self-given. Therefore every deliberating being
has to admit that it is subject to the principle that expresses the idea of
autonomy, i. e. the moral law. In particular, we (as human beings) have
to view ourselves so and thus accept the moral law as a law for our
wills. It cannot be questioned that human beings are rational beings in
the sense that is required for Kants argument to go through, for such
a question, when put forward by a human being, presupposes what is
questioned. Kant here gives (or at least purports to give) an argument
in order to establish certain conclusions, and we try to comprehend
and judge it (or, at least, we view ourselves as doing this). In doing so,
we have to look upon ourselves as being subject to the laws of reason.
Whenever we think about what to do or what to believe, we are bound
to view ourselves as rational beings, as far as this very activity of de-
liberating is concerned.
Let us assume that Kants argument is successful. Let us assume
that in deliberating we have to view ourselves as free in the negative
as well as the positive sense. Then the question arises why he has not
thereby reached his goal. Hasnt he shown that any being that is ca-
pable of practical deliberation is subject to the moral law, in the sense
that any such being has to admit that it is subject to this law? Why
does Kant not simply stop at this point, why the further parts of Sec-
tion III? This is due to the fact that we are not only rational, deliberat-
ing beings, but also sensible beings. As rational beings, we are subject
to the self-given laws of reason, and in particular to the moral law, but
Freedom must be presupposed as a property of the will of all rational beings 255

as sensible beings, we are subject to the laws of nature. So, we have


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

4
Cf. GMS, 456.
256 Corinna Mieth and Jacob Rosenthal

simultaneously, at least not in any straightforward sense, because the


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

II. Discussion

1. There are several, and not only two, notions of freedom involved in
Section III of the Groundwork, which at the beginning of the discus-
sion should be distinguished clearly, even if either further or obvious
or well-known considerations show that these notions amount to one
and the same thing. What Kant actually proves in our passage is this:
A reasoning being has to assume that in its reasoning it obeys the
Freedom must be presupposed as a property of the will of all rational beings 257

principles of reason. This, and no more, is shown. It is not proved that


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

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

5
Cf. KdU, 77.
Freedom must be presupposed as a property of the will of all rational beings 259

The notion of autonomy is often traced back to Rousseaus political


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

6
This is what Dieter Henrich (1975, 64 f.) calls Vernunftfreiheit (freedom of rea-
son) or Urteilsfreiheit (freedom of judgment). He distinguishes it from transzen-
260 Corinna Mieth and Jacob Rosenthal

as freedom as autonomy if and only if the laws an autonomous being


gives to itself are the laws of reason. Why should this be the case? It
is only the case if we identify a rational being with its reason with its
rational part. The point of Kants argument here is that in reasoning a
rational being has to identify itself with reason, and so has to consider
itself to be free in the senses (i) and (iv) simultaneously. The self-given
laws are just the laws of reason. So, Kant seems to have proved suc-
cessfully that in deliberating any being has to view itself as free1, free2,
free3, and free4 simultaneously (whereas freedom5 is an absurdity).
This preliminary judgment will be further clarified and scrutinized in
the following comments.

2. Note that it is the fi rst-person perspective from which freedom as


rationality must be presupposed. That is why so many commentators
come to the conclusion that one of the main topics of GMS III is the
self-relation of the subject (e. g. Prauss, 1993, 255; Steigleder, 2002).
But then the problem arises of whether this insight about ones own
practical reason is necessarily to be transferred to others, and why.
You can easily say about somebody else: He judges this course of
action to be best, but his judgment is due to the fact that he had no
breakfast today and is therefore in a bad mood. But you cannot say
such a thing about yourself. If you do, you contradict yourself in an
outright way. Now, in making this statement about a third person, you
represent this persons will as irrational. This means that even from
the third-person perspective a rational will must be considered to be
subject to the laws of reason, otherwise it would simply be no rational
will. But this analytic truth is not Kants point here. He does not say:
From the notion of a rational will, it flows analytically that it is a
will under the laws of reason. If this were his point, he would have
given us a purely theoretical proof for the freedom of any rational
will. Kants concern here is not whether a rational will is a will that
follows the laws of reason, which is clear, but whether we actually have
rational wills in this sense. The point is that in practical deliberation
we have to represent ourselves as rational and therefore as free1. So,

dentale Freiheit (transcendental freedom) and claims that the further is relevant
for theoretical contexts and rational beings in general, whereas the latter is relevant
for practical contexts and for rational beings that are endowed with a will. In Hen-
richs interpretation, transcendental freedom is the presupposition for the principle
of morality. It is a mixture of what we call freedom 2 , freedom 3 and freedom4. Using
this terminology, Kants argument establishes freedom of reason in a straightfor-
ward way, whereas the case for transcendental freedom is problematic and remains
to be scrutinized.
Freedom must be presupposed as a property of the will of all rational beings 261

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

just because of this as first subject to the law (of which it can regard
itself as the author). (GMS, 431) From there we are lead, according
to Kant, to the idea of a kingdom of ends that is a systematic union
of rational beings through common objective laws (GMS, 434) that
can be understood as common and objective because they abstract
from the personal differences of rational beings as well as from all
the content of their private ends (GMS, 433). But this means to in-
terpret the others as free beings as well.
It seems that from the argument given in GMS III, 2 (that should
give a justification for the moral point of view developed in GMS II)
it follows that we have to consider ourselves to be free i. e. autono-
mous because we have to consider ourselves to be rational from the
first-person perspective. And we know that any other deliberating
being has to do so as well with respect to itself. But is this enough
to show that we have to interpret others as free i. e. autonomous be-
ings as well and therefore as ends in themselves that are an object
of respect for all our maxims? This does not seem to be implied by
Kants argument in III, 2, and is at the very least in need of addition-
al argumentation. The problem surfaces as early as Section II, when
Kant derives the formula of humanity (GMS, 429). There he writes:
The ground of this principle is: rational nature exists as an end in
itself. The human being necessarily represents his own existence in
this way; so far it is thus a subjective principle of human actions. But
every other rational being also represents his existence in this way
consequent on just the same rational ground that also holds for me;
thus it is at the same time an objective principle [] from which the
categorical imperative in the form of the formula of humanity fol-
lows. But again, the question is, even if every rational being has to
regard itself as an end in itself, why does he have to regard the other
rational beings as ends in themselves, too? This is the same kind of
problem as we face in our passage. To be sure, if every rational being
has to view itself as autonomous, and if the principle of autonomy is
to act on no other maxim than that which can also have as object
itself as a universal law (GMS, 447), every rational being is bound to
will according to this principle. But it is not clear what this principle
implies with respect to other rational beings, whether it dictates a
specific attitude towards them. In other words, if we look at the cat-
egorical imperative: It is not clear whether the formula of universal
law indeed implies or is equivalent to the formula of humanity. This
is claimed by Kant, of course, but to get to this result he has to make
the problematic step just mentioned.
Freedom must be presupposed as a property of the will of all rational beings 263

3. Although Kants argument aims at practical deliberation and prac-


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

7
Cf. Schnecker (1999, ch. 4.2.1).
264 Corinna Mieth and Jacob Rosenthal

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

8
Cf. Schnecker (1999, ch. 4.4.1).
Freedom must be presupposed as a property of the will of all rational beings 265

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

9
Cf. the Introduction to KdU.
266 Corinna Mieth and Jacob Rosenthal

tain whether to call an action free that is due to the representation of a


hypothetical imperative involving a natural desire. We will argue later
that contrary to what Kant thinks, by his argument in GMS III, part
2 he is only able to prove (in a practical sense) freedom as rationality,
but not the other kinds of freedom. We will expound this claim in our
fifth comment. Freedom as rationality as viewed from the first-person
perspective, to be sure, extends to any kind of deliberation, especially
to instrumental reasoning.
That Kants argument also applies to theoretical reasoning creates the
following problem. Kant wants to give no theoretical proof for freedom,
but a proof that we have to view ourselves as free. We cannot be sure that
we are free, but we have to look upon us as if we were free. But what kind
of standpoint is this? If you really have to consider yourself to be free, you
are bound to say I am free. You cannot say: Maybe I am not free, but I
have to view myself as free. The first part of this statement implies that
you need not view yourself as free, and so contradicts the second. If we re-
ally follow Kants argument, we have to take its conclusion seriously. If I
am really bound to consider myself to be free in any rational deliberation,
I cannot at the same time say: But this does not show that I am free.
The latter statement means that I entertain the possibility that I am not
free, and consequently, that I do not necessarily view myself as free. The
content of the judgment (perhaps I am not free) contradicts its presup-
position (in making judgments I have to consider myself to be free).
The position on freedom Kant describes is dialectically instable. If we
honestly mean that we have to view ourselves as free, then we really have
to view ourselves so and say we are free. The problem is independent
of what freedom means. It would not arise if Kants argument applied
only to practical deliberations and not to theoretical reasoning as well.
In that case I could easily say: In practical deliberations I have to con-
sider myself to be free, but maybe I am not free. I could say this without
contradicting myself, because this judgment is a result of a theoretical
deliberation, and so does not speak about its own presuppositions. If
theoretical reasoning would not be affected by the question of freedom,
I could entertain the possibility that I am not free in my theoretical as
well as my practical judgments without rendering such a piece of reason-
ing automatically instable. But this is not the situation. Kants argument
applies to theoretical reasoning as well, and so his argument, if it is valid
at all, should convince us that we are free period.
Kant shows convincingly that any deliberation on what to do or what
to believe, and consequently any action which is taken according to
prior deliberation, takes place under the presupposition that it is made
Freedom must be presupposed as a property of the will of all rational beings 267

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

4. We have seen that it is freedom1, freedom as rationality, that we


have to ascribe to ourselves according to Kants argument. In deliber-
ating and judging, we have to suppose that we deliberate and judge ac-
268 Corinna Mieth and Jacob Rosenthal

cording to the laws of reason. Any other supposition would amount to


the admission that our piece of reasoning is invalid and our judgment
cannot be upheld. In particular, when we reason what to do, come to
a conclusion and make a corresponding decision, we have to suppose
that our will is determined by the laws of reason. This kind of freedom
freedom as being subject to the laws of reason is obviously very dif-
ferent from another kind of freedom we might be interested in when
we are acting: freedom as a genuine, unconditioned ability to choose
between various options, to will otherwise as one actually does. This
would be freedom5, freedom as indeterminism. In addition, the fact
that a reason [] is practical, that is, has causality with respect to its
objects (GMS, 448), i. e. freedom as spontaneity, does not imply that
the being who is endowed with this kind of freedom could will and do
otherwise. Quite the contrary. If in deliberating about what to do, you
have to assume that you are rational, and if you additionally assume
that there is only one rational thing to do, you have to conclude that
there is only one action open to you. The assumption that there is only
one rational thing to do is not always accurate, of course, but accord-
ing to Kant it is at least accurate in any morally relevant situation in
which you have a certain obligation. So, contrary to what one might
think, neither freedom as spontaneity, nor freedom as autonomy, nor
freedom as rationality have anything to do with freedom as the genu-
ine possibility to will otherwise as one actually does.
This does not mean that we always act rationally, and in particular
morally, or that we have to assume that we do. We do not, but as far
as we dont, we are irrational. In these cases our reason does not have
causality with respect to its objects, our act is instead determined by
alien causes. This happens very often, and we know this. We make
mistakes in our deliberations, or we arrive at a rational judgment and
nevertheless do otherwise, i. e. perform a weak-willed act because of
the interference of an inclination. As Kant thinks that morality is re-
quired by reason, every immoral act is also irrational, and must there-
fore be a weak-willed act or due to a faulty deliberation. Provided that
your practical judgment determines what you do (and if it does not,
practical deliberation does not make sense), and provided that there
is only one rational course of action, you have to assume that there is
no genuine choice about what to do, that your will is predictable, that
there is only one action genuinely open to you. You just have to find
out which it is. You have to assume all this, because in practical delib-
erating and willing accordingly you have to view yourself as rational.
You might be wrong, of course, and this may be obvious from the per-
Freedom must be presupposed as a property of the will of all rational beings 269

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

10
Cf. Timmermann (2004, 146).
270 Corinna Mieth and Jacob Rosenthal

meta-choice between autonomy and heteronomy, because there is no


law according to which such a meta-choice could be made. It would
have to be a lawless choice and so an absurdity. It may be due to this
that the concept of decision (Entscheidung) plays no prominent role
in Kants practical philosophy. Either your will is determined by natu-
ral causes according to the laws of nature, or by rational grounds ac-
cording to the laws of reason.
All this is difficult to reconcile with ideas about the connection be-
tween moral responsibility and indeterminism, ideas that at least to
some extent seem to be shared by Kant. Concerning this, there are obvi-
ous tensions in his philosophy. In Religion within the Limits of Reason
Alone, for example, he says: The human being must make or have made
himself into whatever he is or should become in a moral sense, good or
evil. These two [characters] must be an effect of his free power of choice,
for otherwise they could not be imputed to him and, consequently, he
could be neither morally good nor evil (RGV, 44). On the other hand,
he says: Freedom does not consist in the fact that we might have pre-
ferred the opposite, but only in the fact that our preference was not ne-
cessitated passive (Nachlass, 4226, AA XVII, 465). Or: But freedom
of choice cannot be defined as some have tried to define it as the
ability to make a choice for or against the law (libertas indifferentiae),
even though choice as a phenomenon provides frequent examples of this
in experience. (MdSR, 226)11 Shortly after this, he says that freedom
can never be located in a rational subjects being able to choose in op-
position to his (lawgiving) reason, even though experience proves often
enough that this happens (though we still cannot comprehend how this is
possible). (MdSR, 226)12 The last passages are particularly important,
because in it Kant admits the (incomprehensible) reality of what we call
freedom as indeterminism, but he refuses to call it freedom.
Sometimes the distinction between Wille (will) and Willkr13
(faculty of arbitrary choice), that Kant explicitly introduces in the
11
Die Freiheit der Willkr aber kann nicht durch das Vermgen der Wahl, fr oder
wider das Gesetz zu handeln (libertas indifferentiae), definiert werden, wie es wohl
einige versucht haben, obzwar die Willkr als Phnomen davon in der Erfahrung
hufige Beispiele gibt. (MdSR, 226)
12
dass die Freiheit nimmermehr darin gesetzt werden kann, dass das vernnftige
Subjekt auch eine wider seine (gesetzgebende) Vernunft streitende Wahl treffen
kann, wenn gleich die Erfahrung oft genug beweist, dass es geschieht (wovon wir
doch die Mglichkeit nicht begreifen knnen). (MdSR, 226)
13
As e. g. Beck (1960, 177, note 1) points out, it is very hard to find an adequate Eng-
lish translation for Willkr. There exist as many different proposals as transla-
tions of Kants works. That is why we follow Becks solution and keep the German
expression in the following. Mary Gregors translation is power of choice. In her
Freedom must be presupposed as a property of the will of all rational beings 271

Metaphysics of Morals, is invoked to elucidate his reasoning in GMS


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

glossary (Gregor, 1996, 650), she refers to the latin difference between voluntas
(Wille) and arbitrium (Willkr). What is meant by Willkr is in our opinion most
adequately described by faculty of arbitrary choice, in order to emphasize the
aspect of arbitrariness.
14
Cf. Allison (1990, 225 f.); Beck (1960, 177 ff.); Wimmer (1980); Steigleder, (2002,
94 f., 109 ff.).
272 Corinna Mieth and Jacob Rosenthal

ability of pure reason to be of itself practical.15 (MdSR, 213 f.). So,


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

15
Die Freiheit der Willkr ist jene Unabhngigkeit ihrer Bestimmung durch sinnli-
che Antriebe; dies ist der negative Begriff derselben. Der positive ist: das Verm-
gen der reinen Vernunft, fr sich selbst praktisch zu sein. (MdSR, 213 f.)
Freedom must be presupposed as a property of the will of all rational beings 273

ture. It cannot, according to him, be subject to the laws of reason and


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

16
See, e. g., Schnecker (1999, ch. 4.2.2).
274 Corinna Mieth and Jacob Rosenthal

sis, he of course claims to do so for reasons. And he also claims (by


his very thesis) that there are natural causes that made him have these
opinions, make these statements etc. But this is no contradiction. His
grasping of the reasons is the cause for what he thinks and says, this
grasping in turn has other causes, and so on, arbitrarily far into the
past. The determinist has to presuppose that nature has built him such
that he is able to understand arguments and judge them in the right
way (most of the time, at least as far as the present discussion is con-
cerned). And he has the hope that his opponent, the indeterminist, is
structered alike, is structured such that when confronted with his, the
determinists, arguments, he recognizes their validity. This expecta-
tion or hope may prove futile or not, but there is nothing in the whole
szenario that would rule out the causal pre-determination of all the
mental processes involved. There are obviously several preconditions
that must be fulfilled before one can comprehend an argument, judge
its validity etc. A certain talent is needed, a certain upbringing and
education, and the like. Our understanding of arguments develops in
time. It is not clear if one could in principle spell out a full causal
story as to why a certain being at a certain point in time understands
a certain argument, but still less is it clear why such a story should be
impossible. Ones grasping of certain reasons could simply be part of
the course of nature.
A closer look reveals that we have to distinguish between two is-
sues here. One is causal determinism: Are there sufficient causes for
the fact that a certain being at a certain time makes a certain judg-
ment? Are there full causal explanations for mental states? If that
should be the case, the mental states (opinions, judgments, decisions
etc.) of the being would be pre-determined by causes arbitrarily far in
the past, they would in principle be predictable, and the being would
not have freedom as spontaneity. Another, further issue is naturalism:
Even if there are sufficient causes for mental states, can these causes
be addressed as or somehow reduced to natural causes? If they could,
the being would not have negative freedom. The point is trickier, be-
cause one might concede that ones grasping of a certain reason is the
cause for making a certain judgment, but deny that this cause can be
viewed as or be reduced to a natural cause that operates according to
the laws of nature. We can dimly imagine a full causal story as to why
someone makes a certain judgment or comprehends a certain argu-
ment. That story would be a psychological story, by and large, but it
is doubtful if it could be called naturalistic, or could in principle be
reduced to, or replaced by, a naturalistic one. This is because a psy-
Freedom must be presupposed as a property of the will of all rational beings 275

chological story for the most part consists of ascriptions of intentional


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

it would at bottom be nothing better than the freedom of a turnspit,


which, when once it is wound up, also accomplishes its movements of
itself. (KpV, 96 f.)
So, even if the determining grounds are [] thought by reason, as
is the case when a subject is built by nature so as to comprehend cer-
tain arguments, the subject is not free in the negative sense, because
his thoughts are part of a causal chain. How the links of the chain look
like does not matter for Kant: When the chain is there, the subject is
determined by the laws of nature. For Kant, the relation of cause to ef-
fect constitutes a natural necessity in any case, and consequently there
is no place in his conception for a causal pre-determination that is not
naturalistic. The causal pre-determination as such essentially reduces
a being to the status of a turnspit: like the turnspit, it is a mere autom-
aton. But Kants argument in our passage does only give us freedom
as rationality. It is not shown, although intended by Kant, that the be-
ing cannot at the same time view itself as being subject to the laws of
nature, or that the being has to look upon itself as spontaneous. There
is no successful argument in GMS III as to why rationality should
be incompatible with either determinism or naturalism. Kant himself
concedes, in the passage quoted above, that determining grounds can
drive a subject through representations thought by reason, and we
have no reason to think that this is not our situation. We are given no
reason to think that we are no automata spirituale. Kants true motive
to discard this possibility is that he thinks it excludes moral responsi-
bility. And in the Critique of Practical Reason he indeed argues from
the reality of the moral law to the reality of transcendental freedom
(which is the same as freedom as spontaneity). But in the Groundwork
he wants to have it the other way round: He wants to prove the moral
law as binding for our wills by establishing transcendental freedom, at
least in a practical sense. But this kind of freedom does not come out
of his argument. At the end we will discuss in how far or in what sense
the argument is nevertheless successful.
Kant repeatedly emphasizes that every action may on the one hand
be viewed as an event determined by the laws of nature, as a link in a
causal chain, and on the other hand as spontaneously caused by the
agent (Prol, 53; GMS, 455 ff.; KpV, 94 ff.; KrV, A 538 ff., B 566 ff.).
Both perspectives seem indispensable to him, the former to guarantee
the unity of appearances, the latter as a necessary condition for moral
responsibility. He claims that these perspectives and the corresponding
assertions can be reconciled: The former concerns the phenomenal, the
latter the noumenal world. So he is a compatibilist in a certain sense.
Freedom must be presupposed as a property of the will of all rational beings 277

But his brand of compatibilism requires the distinction between phe-


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

But inconceivable or not, the demanding kind of resolution to this


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

17
Cf. the teleological arguments in Toward Perpetual Peace.
Freedom must be presupposed as a property of the will of all rational beings 279

Kant, required to secure moral responsibility, but it is not established


by his argument in GMS III, 2.
Our diagnosis may be corroborated by looking at Kants Schulz-
Review he published two years before the Groundwork. Johann Hein-
rich Schulz had written an Attempt at an Introduction to a Doctrine
of Morals, in which he advocates the complete causal pre-deter-
mination of all human actions. Kant calls this a general fatalism
that turns all human conduct into a mere puppet show and has the
consequence that what is left us is only to await and observe what
sort of decisions God will effect in us by means of natural causes,
but not what we can and ought to do of ourselves, as authors. (AA
VIII, 13) This, Kant claims, is a self-refuting thesis, as: Although he
[Schulz] would not himself admit it, he has assumed in the depths
of his soul that understanding is able to determine his judgment in
accordance with objective grounds that are always valid and is not
subject to the mechanism of merely subjectively determining causes,
which could subsequently change; hence he always admits freedom to
think, without which there is no reason. (AA VIII, 14) Again, this
line of argument is only convincing if one already presupposes what
Kant purports to establish by it. The idea that if we were causally pre-
determined, we could only await and observe what sort of decisions
are effected in us by natural causes simply assumes that our (practi-
cal) deliberations that lead to our decisions are no part of the course
of nature. The idea that if Schulz is able to determine his judgment
in accordance with objective grounds, he cannot at the same time be
subject to the mechanism of merely subjectively determining caus-
es, simply assumes incompatibilism instead of justifying it. He who
argues has to presuppose freedom to think in the sense of deter-
mination of his judgment in accordance with objective grounds this
much is clear. But Schulz need not at all deny that, he just has to say
that his judgment is also determined by natural causes according to
the laws of nature, because nature has created himself as a (partially,
at least as far as the present argument is concerned) rational being,
as a (partially) rational automaton spirituale, if one wants to put it
that way. That subjectively determining causes could subsequently
change, whereas objective grounds are always valid, is besides the
point. We know very well that we are not always in the state of mind
to judge the validity of arguments correctly, e. g. when we are tired
or drunk or distracted, when we were children or when we become
feeble-minded in old age. We know that the subjectively determin-
ing causes that make us grasp the objective grounds that are always
280 Corinna Mieth and Jacob Rosenthal

valid can and will subsequently change. So, again, the freedom to
think that we have to claim for ourselves when we deliberate does
only amount to freedom as rationality, but does not imply negative
freedom or freedom as spontaneity.
The case of freedom4 (autonomy) is most difficult to judge. In a
weak sense, it is implied by Kants argument. In deliberating and judg-
ing, we not only have to assume that we follow whatever laws of reason
there are, but also approve of our being subject to them. We could not
wish to violate them while deliberating, because this would render our
conclusions unjustified. In this sense the laws of reason are self-given.
They need not be self-given in the sense of literally being chosen by
ourselves, but we can certainly think of them as self-given. Upon re-
flection we realize that if we had the choice we would (have to) impose
these laws on ourselves, or negatively, that we could not wish to get rid
of them. But this sense of autonomy does not imply freedom2 or free-
dom3. We can further clarify this weak meaning of autonomy if we
ask why the laws of nature are alien laws. Insofar as we are subject
to them, we have, of course, not actually chosen to be so. But as we
have seen, this may also be the case with the laws of reason. The cru-
cial difference is that we are not bound to approve of our being subject
to the laws of nature. In deliberating, we need not identify with nature
and her laws, but we have to identify with reason and its laws. It is this
point that is successfully made by Kants argument.
One might agree with this and nevertheless object to the charac-
terization of the laws of nature as alien. Insofar as we are deliber-
ating beings, we have to identify with reason and its laws this we
can concede. But are we not also sensible beings? And insofar as we
are such beings, dont we have to view ourselves as part of nature and
identify with her laws? Doesnt Kant beg the question if he calls the
laws of nature alien and the laws of reason self-given? Why are we
entitled to prefer the view of ourselves as rational beings to the view of
ourselves as sensible beings? These questions point to the reason why
Kant has not reached his goal within the text passage under discussion,
even if his argument here were to be completely successful. They are
answered in the further parts of Section III by an appeal to the dis-
tinction between phenomenal and noumenal world, by attributing the
laws of nature to the former and the laws of reason to the latter, and
by claiming priority of the latter because the world of understanding
contains the ground of the world of sense (GMS, 453), and therefore,
what belongs to mere appearance is necessarily subordinated by rea-
son to the constitution of the thing in itself (GMS, 461).
Freedom must be presupposed as a property of the will of all rational beings 281

Perhaps there is a more straightforward argument as to why the


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

as spontaneity, it is not enough to secure our moral responsibility


according to Kant. If autonomy amounts to the simple claim that in
theoretical and practical deliberations we have to accept and approve
of the laws of reason and assume that we are guided by them, we
could be free positively without being free negatively. This sounds
paradoxical, and so confirms again that Kant had more in mind than
this kind of autonomy.
But all this does not mean that Kants argument in GMS III, 2
is a complete failure. To be sure, we have to view ourselves as being
subject to the laws of reason. In deliberating, we have to presuppose
that we follow them, and upon reflection, we have to approve of this
situation. Insofar they are self-given. We not only have to assume that
we are subject to the laws of reason, but we have to endorse them re-
flectively. This is a weak sense of autonomy, but it constitutes a view
we necessarily take of ourselves and our relation to reason, and thus,
if we accept the connection between autonomy and the moral law that
is established in GMS III, 1, Kant has shown successfully that we have
to view ourselves as being subject to the moral law. We have to accept
the moral law as a law for our wills. But in this line of reasoning, free-
dom as the absence of natural determination or as spontaneity or as
autonomy in a strong sense plays no role.

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesam-


melte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902) (abbreviated
as AA). The Critique of Pure Reason will be cited according to the A/B
pagination from the first and second editions. English translations, except
where otherwise noted, are from Groundwork of the Metaphysics of Morals,
translated and edited by Mary Gregor, Cambridge: Cambridge University
Press 1997, or from The Cambridge Edition of the Works of Immanuel Kant
Practical Philosophy, translated and edited by Mary Gregor, Cambridge:
Cambridge University Press 1996.
Ge Gemeinspruch, AA, VIII
GMS Grundlegung zur Metaphysik der Sitten, AA, IV
KdU Kritik der Urteilskraft, AA, V
KpV Kritik der praktischen Vernunft, AA, V
KrV Kritik der reinen Vernunft, AA, III, IV
MdSR Metaphysik der Sitten Rechtslehre, AA, VI
Prol Prolegomena, AA, IV
RGV Die Religion innerhalb der Grenzen der bloen Vernunft, AA, VI
Freedom must be presupposed as a property of the will of all rational beings 283

Other works

Allison, Henry E. (1990): Kants Theory of Freedom, Cambridge: Cambridge


University Press.
Allison, Henry E. (1993): Kants Preparatory Argument in Grundlegung
III, in: Hffe, Otfried (ed.): Grundlegung zur Metaphysik der Sitten. Ein
kooperativer Kommentar, Frankfurt a. M.: Klostermann, 314324.
Allison, Henry E. (2002): Morality and Freedom: Kants Reciprocity The-
sis, in: Pasternack, Lawrence (ed.): Immanuel Kant Groundwork of
Metaphysics of Morals in Focus, London: Routledge, 182210.
Beck, Lewis White (1960): A Commentary on Kants Critique of Practical
Reason, Chicago: University of Chicago Press.
Beck, Lewis White (1975): The Actor and the Spectator, New Haven/London:
Yale University Press.
Bennett, Jonathan (1974): Kants Dialectic, Cambridge: Cambridge Univer-
sity Press.
Eisler, Rudolf (1994): Kant-Lexikon. Nachschlagewerk zu Kants smtlichen
Schriften, Briefen und handschriftlichem Nachla, Hildesheim u. a.: Olms.
Henrich, Dieter (1975): Die Deduktion des Sittengesetzes. ber die Grn-
de der Dunkelheit des letzten Abschnittes von Kants Grundlegung zur
Metaphysik der Sitten, in: Schwan, Alexander (ed.): Denken im Schatten
des Nihilismus. Festschrift fr Wilhelm Weischedel zum 70. Geburtstag,
Darmstadt, 55112.
Korsgaard, Christine M. (1996): Morality as freedom, in: Korsgaard, Chris-
tine M.: Creating the Kingdom of Ends, Cambridge: Cambridge Univer-
sity Press, 159187.
Korsgaard, Christine M. (1998): The Sources of Normativity, Cambridge:
Cambridge University Press.
ONeill, Onora (1989): Reason and Autonomy in Grundlegung III, in:
ONeill, Onora: Constructions of Reason. Explorations of Kants Practi-
cal Philosophy, Cambridge: Cambridge University Press, 5165.
Prauss, Gerold (1993): Fr sich selber praktische Vernunft, in: Hffe,
Otfried (ed.): Grundlegung zur Metaphysik der Sitten. Ein kooperativer
Kommentar, Frankfurt a. M.: Klostermann, 253263.
Schnecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kate-
gorischen Imperativs, Freiburg i. Br./Mnchen: Alber.
Schnecker, Dieter / Wood, Allen (22004): Kants Grundlegung zur Meta-
physik der Sitten. Ein einfhrender Kommentar, Paderborn: Schningh
(UTB).
Schneewind, Jerome B. (1992): Autonomy, obligation, virtue, in: Guyer,
Paul (ed.): The Cambridge Companion to Kant, Cambridge: Cambridge
University Press, 309341.
Steigleder, Klaus (2002): Kants Moralphilosophie. Die Selbstbezglichkeit
reiner praktischer Vernunft, Stuttgart: Metzler.
Sturma, Dieter (2004): Kants Ethik der Autonomie, in: Ameriks, Karl /
Sturma, Dieter (ed.): Kants Ethik, Paderborn: mentis, 160177.
Timmermann, Jens (2004): Kommentar, in: Jens Timmermann (ed.): Im-
manuel Kant. Grundlegung zur Metaphysik der Sitten, Gttingen: Van-
denhoeck & Ruprecht, 84158.
284 Corinna Mieth and Jacob Rosenthal

Wimmer, Reiner (1980): Praktische Vernunft und Freiheit. Kants doppelter


Freiheitsbegriff als Voraussetzung fr die Doppelung von Recht und Mo-
ral, in: Wimmer, Reiner: Universalisierung in der Ethik. Analyse, Kritik
und Rekonstruktion ethischer Rationalittsansprche, Frankfurt a. M.:
Suhrkamp, 124144.
Marcel Quarfood

The Circle and the Two Standpoints (GMS III,3)

At the end of Section II of the Groundwork, the content and presup-


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

possible to proceed along the lines of the straightforward argument.


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

1
Cf. Schnecker (1999, ch. 5.2).
2
Cf. Allison (1990, 227 f.).
The Circle and the Two Standpoints 287

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

1. Is the circle a circle?

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

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

of will.3 As we know that this is supposed to be a piece of circular


reasoning, there is a strong pull to read the two parts of this passage as
arguing in opposite directions, roughly along the following lines:
(1) We take ourselves as free because we think ourselves under moral
laws.
(2) We think ourselves under moral laws because we have taken our-
selves as free.

Simplifying a little bit more, it seems that (1) is a derivation of freedom


from morality, whereas (2) is a derivation of morality from freedom.
On this construal we certainly have a piece of circular reasoning.
In view of the expectation to find a logical circle, this way of inter-
preting the passage is natural and it is not uncommon, but it is hardly
adequate to the text.4 We take ourselves as free [] in order to [um]
think ourselves under moral laws should rather be taken to mean that
freedom is adopted as a premise for deriving morality. By taking our-
selves as free we can reach the conclusion that we stand under moral
laws. A problem with this reading, however, is that now we do not get
a genuine circle, but only a linear derivation starting from the assump-
tion of freedom. And this should be a bit worrying, as Kant after all
calls it a circle. I shall return to this issue below.
The need to find a circle drives one of Kants earliest commenta-
tors to offer another reading of the same kind, but this time reversed.
In an exposition of the Groundwork from 1788, the anonymous au-
thor (writing under the signature T-h) cites the first circle passage and
then says: thus it appears that from freedom we infer autonomy, and
then from autonomy in turn we infer freedom.5 Quite plausibly, T-h
took the beginning of the passage to state that we assume freedom
as a premise from which we reach the conclusion that we are autono-
mous.6 But then, in order to obtain a circle, T-h had to construe the
second half of the passage as describing an inference from autonomy
to freedom. Kant, on the other hand, states that we think ourselves as
3
The German reads: Wir nehmen uns in der Ordnung der wirkenden Ursachen als
frei an, um uns in der Ordnung der Zwecke unter sittlichen Gesetzen zu denken, und
wir denken uns nachher als diesen Gesetzen unterworfen, weil wir uns die Freiheit
des Willens beigelegt haben.
4
For a recent example of such an interpretation, see Kim (2002, 71). Schnecker
(1999, 349 f.) refers to several other instances. My treatment of this issue owes much
to Schnecker (1999).
5
T-h (1788, 271): wir scheinen also aus der Freiheit auf die Autonomie, und dann
wieder aus der Autonomie auf die Freiheit zu schliessen.
6
In the present context it does not matter that T-h construes the argument in terms of
autonomy rather than morality.
The Circle and the Two Standpoints 289

subjects to these laws because we have ascribed to ourselves freedom,


which can hardly be interpreted otherwise than that the inference de-
scribed has freedom as premise rather than as conclusion.
We cannot yet conclude that these interpretations are definitely
wrong, however. Perhaps there is a wider context that can justify a re-
construction of the text in terms of a circular argument structure. But
at least when looking closely at the first circle passage, we do not fi nd
a circle in the strict sense. Let us now see how Kants second circle
passage is constructed. It occurs at the end of the subsection Of the
Interest Attaching to the Ideas of Morality. Just before, Kant claims to
have removed the suspicion that a hidden circle was contained in our
inference from freedom to autonomy and from the latter to the moral
law (GMS, 453). Then comes the second circle passage, which speci-
fies what kind of circle was suspected to be hidden in the straightfor-
ward argument:
namely that we perhaps took as a ground the idea of freedom only for the
sake of the moral law, so that we could afterwards infer the latter in turn
[wiederum] from freedom. (GMS, 453)7
Perhaps clearer than in the corresponding part of the first circle pas-
sage, here it is stated that the idea of freedom is taken as a ground,
that is, a premise. The last part of the passage says that once freedom
is assumed, we can infer the moral law.8 It seems that here, like in the
first circle passage, we have a non-circular argument starting from the
assumption of freedom and having morality as its consequence.
Kant goes on to explain what kind of mistake we would have made
had we argued in the manner described in the circle passages. It was
suspected that we were unable to furnish any ground at all for the
moral law but could put it forward only as a petitio principii [Erbit-
tung eines Prinzips] disposed souls would gladly grant us, but never
as a demonstrable proposition (GMS, 453).9 We would thus have put

7
The German reads: da wir nmlich vielleicht die Idee der Freiheit nur um des
sittlichen Gesetzes willen zum Grunde legten, um dieses nachher aus der Freiheit
wiederum zu schlieen.
8
The words in turn might be taken to indicate a reversal of argumentative order,
suggesting a genuine circle after all. Perhaps Allen Woods translation (Kant 2002)
of wiederum as once again gives a slightly different tone.
9
Alternatively, the translation could be unable to furnish any ground at all for our
inference. Compare Woods rendering (Kant 2002): we could not offer any ground
for the former. In its context the former (jenem) might refer to our inference
(unserem Schlusse). Here is the complete German sentence: Nun ist der Verdacht,
den wir oben rege machten, gehoben, als wre ein geheimer Zirkel in unserem
Schlusse aus der Freiheit auf die Autonomie und aus dieser aufs sittliche Gesetz
290 Marcel Quarfood

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

enthalten, da wir nmlich vielleicht die Idee der Freiheit nur um des sittlichen
Gesetzes willen zum Grunde legten, um dieses nachher aus der Freiheit wiederum
zu schlieen, mithin von jenem gar keinen Grund angeben knnten, sondern es
nur als Erbittung eines Prinzips, das uns gutgesinnte Seelen wohl gerne einrumen
werden, welches wir aber niemals als einen erweislichen Satz aufstellen knnten
(GMS, 453; I have italicized the words under discussion). For a discussion of what
jenem refers to, see Schnecker (1999, 340). Anyway, even if jenem refers to the
moral law, so that the moral law is what is put forward as a petitio, this can be taken
as resulting from the groundlessness of the premise for inferring the moral law,
namely freedom (cf. Schnecker, 1999, 340 f.).
10
See also Kim (2002, 70) for references to Kants other lectures on logic treating
these two fallacies.
The Circle and the Two Standpoints 291

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

11
Kim (2002, 70) considers the possibility that the circle in GMS III may take the
form of a petitio. But first he only discusses an argument in which morality is as-
sumed without ground and used for deriving freedom, which is surprising since
he argues that the goal of the deduction in GMS III is to establish the principle of
morality. When Kim then (on p. 73) briefly treats the inverse argument, in which
freedom is assumed and morality inferred, his rejection of this interpretation (as
presented in Allison 1990, 221) seems to rely on taking the circle as a fallacy Kant
himself is to have committed. Moreover, because he construes the fi rst circle pas-
sage (GMS, 450) as a description of a strict circle (p. 71), he looks for a circulus in
probando even when he considers a petitio, so that his initial insight that the circle
might be a petitio is lost.
12
The circular train of thought could perhaps be presented quasi-formally in the fol-
lowing way (letting M represent morality and F freedom):
(1) M (groundless assumption, psychologically based on conviction)
(2) FM (assumption on the basis of the reciprocity between these concepts)
(3) F (groundless assumption, psychologically based on the wish to infer M)
(4) M (conclusion from (2) and (3), but not from (1), as that would be unbear-
ably trifling.)
292 Marcel Quarfood

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

It might here seem as if Kant refers to inferences in both directions,


from freedom to morality and from morality to freedom. If one of the
members of a pair of reciprocal concepts cannot be used to explain
the other or to furnish a ground for it, this to be sure goes for both of
the concepts. But, as Schnecker (1999, 335) points out, the text does
in no way imply that both of the possible inferences are relevant for
Kants discussion. For when freedom is assumed as the ground for
morality, one of the reciprocal concepts is used to furnish a ground
for the other one, irrespective of whether an inference in the other
direction is made or not.

2. The circle in context

The previous discussion does not seem to have achieved much. We


do not yet know how there might be a petitio principii involved in the
straightforward argument from freedom to morality. This argument
has two components: the analytical connection between freedom and
the moral law, and the necessity for rational beings to regard their
judgments as free. But if we now have reason to think that the circle
pertains to the insufficient ground for positing freedom, we can con-
clude that the threat should concern the second step of the straightfor-
ward argument. Its first step merely claims that there is a conceptual
link between the concepts of a free will and an autonomous will. It is
in connection to the second step that we might be enticed to ascribe
freedom to ourselves, so it is here that a petitio can be committed.
But why should there be a problem in our ascribing freedom in a
practical respect to ourselves? One possibility is that there is some-
thing wrong with the argument to the effect that reason must regard
itself as the author of its principles and therefore must stand under
the idea of freedom (GMS, 448), so that clinging to its conclusion
would be entirely groundless. However, nothing indicates that Kant re-
jected the argument as such. There are two other possibilities that are
more plausible. One is that we have not yet (that is, at the point where
the threat of a circle occurs) done enough to justify the assumption
that we are rational beings, so that the application of the second part
of the straightforward argument to the case of human beings amounts
to a petitio. Another possibility is that even if the argument for the
The Circle and the Two Standpoints 293

practical freedom of rational beings applies to us, this argument leads


to inadmissible consequences. As will be seen, these interpretations
share one common feature, as both of them stress the importance of
making a distinction between the moral law (as the necessary course
of action for an ideally rational being) and moral obligation in the
form of the categorical imperative (as applying to beings who are both
rational and sensibly affected).
To begin with the first line of interpretation, when Kant in the sec-
ond subsection argues that a rational being has to assume its own free-
dom, he points out that the proof concerns beings that are rational
and endowed with a will (GMS, 448). The requirement that a will is
needed in addition to rationality opens for the possibility that we may
possess theoretical reason but yet lack practical reason (that is, will), so
that the petitio one might commit here would be to merely assume with-
out further ado that we have a will.13 Or, to go one step further in this
direction, perhaps even assuming theoretical rationality in the case of
human beings is to go too far, as long as independent grounds have not
been given.14 If this is the right way to view the argument establishing
practical freedom from the capacity to judge, the argument concerns
ideally rational beings, and is therefore of no immediate consequence
for non-ideal human beings. I cannot pursue the many questions that
arise in the interpretation of the second subsection of GMS III, so I
merely note here that this reading fits well with the circle considered as
a petitio with respect to assuming freedom, as well as with Kants turn-
ing to transcendental idealism in order to solve the problem. For when
the two standpoints doctrine of transcendental idealism is deployed, a
new sphere is opened where human rationality (theoretical as well as
practical) may be found, providing the straightforward argument with
an independent reason for the assumption of freedom.
Another significant feature of this interpretation is that it points to
the crucial distinction between the moral law as such and the categori-
cal imperative.15 This distinction is sometimes (e. g. in GMS, 420; 440;
447) marked by Kant in terms of analytic versus synthetic a priori con-
nections. Simply put, the contrast is between a being for which morality
is a necessary law of action, so that it cannot fail to act morally, and the

13
Cf. Allison (1990, 218).
14
See the discussion in Schnecker (1999, e. g. 208, 233, 319).
15
The radical contrast between the categorical imperative as applying to beings that
are both rational and affected by sensibility on the one hand and the moral law
pertaining to purely rational beings on the other is emphasized time and again in
Schnecker (1999).
294 Marcel Quarfood

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

16
Cf. Prol, 267.
17
I think contains in GMS, 397, is used in the exact sense it has in Kants theory
of conceptual containment, but for present purposes it suffices to consider it as an
analogy.
18
Allen Woods translation, in Kant (2002).
The Circle and the Two Standpoints 295

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

stressing the admission in the statement that it must be freely admit-


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

19
Brandt (1988, 186). For criticism of Brandt, see Allison (1990, 220), Schnecker
(1999, 353), and Kim (2002, 74 f.).
The Circle and the Two Standpoints 297

tian point of view, a thinker who draws such analytic connections, but
lacks the resources of critique by means of which synthetic a priori
claims can be justified, is a dogmatist par prfrence. The circle might
be said to represent the route the ideal typical Wolffian dogmatist (as
seen by the lights of Kant) would have to follow, even though no Wolf-
fian actually reasoned along the lines of the circle.20

3. The two standpoints

Immediately upon introducing the circle in 450 Kant presents the way
out. It consists in attending to the difference between two standpoints
we can adopt with respect to ourselves. On the one hand we think of
ourselves by means of freedom as causes efficient a priori, on the
other hand, we conceive of ourselves in terms of our actions as effects
that we see before our eyes. The two standpoints are linked to the dis-
tinction between things in themselves and appearances. In these pages
(GMS, 450 f.), Kant describes an observation possible to make even
for the commonest [gemeinste] understanding (GMS, 450). But he
states that a reflective human being would come to the same conclu-
sion (GMS, 451), and the whole discussion is conducted in the technical
language of transcendental philosophy, so it is endorsed also on this
level of reflection. Bringing the unsophisticated persons understand-
ing into the discussion accords with the starting point of the Ground-
work, namely that the content of morality is already implicitly assumed
in everyday moral thinking. The distinction is not just imported from
philosophical theory, but is corroborated in human experience (though,
as we will see, it does certainly not have its source in experience).
The drawing of a distinction between appearances and things in
themselves is traced back to an observation [Bemerkung] made even
by the commonest understanding in the form of an obscure discrimi-
nation of judgment [Urteilskraft] which it calls feeling (GMS, 451), to
the effect that representations over which we have no control present
things to us only with respect to how they affect us, that is, how they
appear, but not as to how they are in themselves. The passivity of the
reception of representations leads us to view them as appearances,
and thereby we are committed also to assuming that there is behind
appearances something else that is not appearance (GMS, 451).21 In
20
Which Brandt, incidentally, construes as a circulus in probando.
21
Compare KrV, B xxvi, on the absurd proposition that there is an appearance with-
out something that appears.
298 Marcel Quarfood

thus noticing the difference between passively received representa-


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

22
Compare KrV, e. g. A 682/B 710.
The Circle and the Two Standpoints 299

dence for our possession of reason, so that we can ascribe freedom


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

is that both before and after introducing the distinction between the
two points of view, Kant urges the reader not to look for a theoretical
proof of freedom. Early on in GMS III he states that he is not bound
to prove freedom in its theoretical respect, but only in a practical re-
spect (GMS, 448n.). And at a later stage of his exposition, he says that
the concept of a world of understanding is [] only a standpoint that
reason sees itself constrained to take outside appearances in order to
think of itself as practical (GMS, 458).

Literature

Kants writings

Kants writings will be cited according to the pagination of Kants gesam-


melte Schriften, Akademie Ausgabe (Berlin: deGruyter, 1902) (abbreviated
as AA). All textual references are to The Cambridge Edition of the Works
of Immanuel Kant (Cambridge University Press, 1992), except where oth-
erwise noted. The German text of the Grundlegung will be quoted accord-
ing to the edition by Bernd Kraft and Dieter Schnecker (Hamburg, Felix
Meiner Verlag, 1999). The Critique of Pure Reason will be cited according to
the A/B pagination from the first and second editions.
GMS Grundlegung zur Metaphysik der Sitten, AA, IV
KrV Kritik der reinen Vernunft, AA III, IV
Log Logik-Jsche, AA, IX
Prol Prolegomena, AA IV
Kant, Immanuel (2002): Groundwork for the Metaphysics of Morals, New
Haven and London: Yale University Press.

Other works

Allison, Henry E. (1990): Kants Theory of Freedom, New York, Cambridge


University Press.
Brandt, Reinhard (1988): Der Zirkel im dritten Abschnitt von Kants Grund-
legung zur Metaphysik der Sitten, in: H. Oberer, Hariolf / G. Seel (Hrsg.),
169191.
Kim, Halla (2002): Has Kant Committed the Fallacy of Circularity in Foun-
dations III?, in: Journal of Philosophical Research, 27, 6581.
Schnecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kate-
gorischen Imperativs, Freiburg i. Br./Mnchen.
T-h (1788): Vierter Versuch ber die Kantsche Grundlegung zu einer Meta-
physik der Sitten, in: Deutsches Museum, 13, 264293.
Dieter Schnecker

How is a categorical imperative possible?


Kants deduction of the categorical imperative (GMS, III,4)

Kants deduction of the categorical imperative is the answer to the fol-


lowing question: How is a categorical imperative possible? The an-
swer is given in subsection 4 (Sec. 4) of chapter three of the Ground-
work. It is impossible to understand this answer, and hence impossible
to understand Kants deduction of the moral law, without taking into
account the overall context of Groundwork III (GMS III). However,
here I can only sketch the overall structure of GMS III, and therefore
only present a sketch of what I call Kants thesis of analyticity.1 This
thesis is developed in Sec. 1 of GMS III; however, it appears time and
again in GMS III, and it deserves special attention (part 1). Part 2,
then, offers a close reading and analysis of Kants deduction.
The purpose of this paper is not to criticize or make philosophical
use of Kants deduction. As I have argued elsewhere, one of the ma-
jor obstacles of serious (i. e. historical) Kant-research is the inability
(and unwillingness) to distinguish between the question of what a text
means and and the question of whether what it manifests is true.2

1. Structure and task of GMS III: What does Kant want to achieve?

In the preface of the GMS, Kant claims that his Groundwork is noth-
ing more than the search for and establishment of the supreme princi-

1
For a detailed analysis of Groundwork III, cf. Schnecker (1999); this article is
based upon that book. Here I will not address the secondary literature (I did so
extensively in Schnecker, 1999); as far as I can tell, little has been published since
that pays close attention to the text. However, I will make some brief comments on
Steigleders interpretation of what I call the thesis of analyticity. Many thanks to
Richard Capobianco and Alexander Cotter for checking my English.
2
Cf. Schnecker (2001); cf. also Damschen / Schnecker (2006).
302 Dieter Schnecker

ple of morality (GMS, 392). This supreme principle is the categorical


imperative (CI). It is generally agreed upon that the search for the
CI, i. e. its conceptual analysis, takes place in GMS I/II, whereas its
establishment is to be found in GMS III. To be more precise, this
establishment is in the answer that Kant provides to the question of
how is a categorical imperative possible? This question is first raised
in GMS I (GMS, 417 ff.), and then again in GMS II, 425; its answer,
however, is postponed to Sec. 4 of GMS III, the heading of which is
again that question of how a categorical imperative is possible. There
is no doubt that the answer to this question is what Kant himself calls
a deduction (GMS, 447,22; 454,21; 463,21); the second paragraph of
Sec. 4 begins with the formulation And thus categorical imperatives
are possible (GMS, 454,6, m. e.).
Before we move on, it is important to realize two crucial structural
elements of the text. First, Kant clearly states at the end of Sec. 1 that
he cannot yet or immediately provide an answer to the question of how
a categorical imperative is possible; rather, he says, this still needs
some preparation (GMS, 447,25). Since the central question How is
a categorical imperative possible? is indeed the heading of Sec. 4, it is
only natural to assume that Sec. 2 and Sec. 3 provide that preparation,
whereas the actual answer, then, is given in Sec. 4. (As we will see, this
also fits very well with what really happens in those sections.)
The second structural element relates to what I call Kants thesis of
analyticity. In Sec. 1, Kant puts it as follows:
a free will and a will under moral laws are the same. [paragraph] Thus if
freedom of the will is presupposed, then morality follows together with its
principle from mere analysis of its concept (GMS, 447,610).
This thesis has been widely misunderstood; as a result, Kants overall
argument (deduction) in GMS III has been misunderstood too (often
not recognized at all, as a matter of fact). His overall argument in
GMS III always has been reconstructed as follows: A free will is a will
under the moral law; freedom must be presupposed as a quality of the
will of all rational beings; human beings are rational beings; therefore,
the human will as a free will is under the moral law, which is to say
the categorical imperative is valid. 3 Since premise 1 is proven in Sec. 1,
premise 2 is argued for in Sec. 2 and premise 3 in Sec. 3,4 the answer

3
Cf. for instance Wood (1999, pp. 171176).
4
As a matter of fact, the claim that freedom must be presupposed as a quality of the
will of all rational beings is the heading of Sec. 2; Sec. 3 relates this to human beings
in particular.
How is a categorical imperative possible? 303

to the question of how is a categorical imperative possible? and thus


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

perfectly rational beings always act morally. There are many passages
that prove this point (some of which can actually be found in GMS
III). A famous one in GMS II reads as follows:
If reason determines the will without exception [unausbleiblich], then the
actions of such a being, which are recognized as objectively necessary,
are also subjectively necessary, i. e. the will is a faculty of choosing only
that which reason, independently of inclination, recognizes as practically
necessary, i. e., as good (GMS, 412,3035).

Since human beings dont have such a perfectly rational being, to them
the moral law is a categorical imperative or, as Kant puts it, a synthetic
proposition a priori. The categorical imperative is
a practical proposition that does not derive the volition of an action analytically
from any other volition already presupposed (for we have no such perfect will),
but is immediately connected with the concept of the will of a rational being,
as something not contained in it (GMS, 420,3235, Fn., m.e.).

Willing the good action is not necessarily contained in the volition of


a sensuous-rational being, i. e. it cannot be analytically derived from
the volition of such a being. That is why, in contrast to a pure will, the
CI does not follow by mere analysis (GMS, 447,9) of the concept of
freedom of such a sensuous-rational being. So when Kant in his thesis
of analyticity states that a free will and a will under moral laws are the
same and that morality follows together with its principle from mere
analysis of the concept of the freedom of the will (GMS, 447,610),
all he really says is this: A perfectly rational and free will always wills
morally. Such a will must be understood as the will of an actually holy
being; or (as we will see) it must be understood as the intelligible will of
a being that is both a member of the empirical and intelligible world.
The argumentative structure of GMS III strongly supports this
reading. After Kant has presented his solution to the notorious circle
in Sec. 3, he concludes:
For now we see that if we think of ourselves as free, then we transport
ourselves as members into the world of understanding and cognize the
autonomy of the will, together with its consequence, morality; but if we
think of ourselves as obligated by duty, then we consider ourselves as be-
longing to the world of sense and yet at the same time to the world of
understanding (GMS, 453,1115).

It is highly remarkable that Kant does not contrast the formulation if


we think of ourselves as free with if we think of ourselves as not free.
Rather, he contrasts it with if we think of ourselves as obligated by
How is a categorical imperative possible? 305

duty. The counterpoint to freedom is not the absence or lack of free-


dom, but duty. Since morality without freedom is not thinkable for
Kant, freedom understood as a counterpoint to duty cannot be identi-
cal with the freedom presupposed by morality for beings that are both
sensuous and rational. Freedom as the counterpoint to duty must be
understood as the quality of a will that is completely rational and free.
However, a being that is obligated by duty must be free, too. There-
fore, to think of oneself as obligated by duty implies to consider one-
self as belonging to the world of sense, but it also means to consider
oneself at the same time (zugleich, GMS, 453,15) as belonging to the
world of understanding. 5 Hence the first part of the passage quoted
above (GMS, 453,1113) really is nothing but a reformulation of the
thesis of analyticity. The second part after the semi-colon, however,
makes clear why for sensuous-rational beings the moral law is an im-
perative and hence duty.6
The thesis of analyticity also shows up in the deduction found in
Sec. 4 which we will address in great detail later: As a mere member
of the world of understanding, all my actions would be perfectly in
accord with the principle of the autonomy of the pure will (GMS,
453,2527, m. e.). A bit later it says: And thus categorical imperatives
are possible through the fact that the idea of freedom makes me into a
member of an intelligible world, through which, if I were that alone, all
my actions would always be in accord with the autonomy of the will
(GMS, 454,69, m. e.). Again, at the end of the Sec. 4 Kant concludes:
The moral ought is thus his own necessary volition as a member of

5
On the zugleich cf. also GMS, 454,9 and 462,31.
6
The passage just quoted (For now we see world of understanding, GMS, 453,11
15) is Kants final answer to the problem of the notorious circle (GMS 450,18;
453,4) in Sec. 3. This circle is yet another issue I cannot address here (cf. Sch-
necker, 1999, pp. 317358). However, let me state that its correct understanding must
begin with the insight that for Kant there is a difference between a petitio principii
and a circulus in probando, and that the aforementioned circle is an Erbittung
eines Prinzips (GMS 453,9), not a vicious circle (Erbittung eines Prinzips is Kants
translation of the Latin phrase petitio principii, i. e. begging the question). What
is still missing in Sec. 3 (just begged for) is the rationale for the human beings belief
in the idea of freedom as well as the justification for the categorical imperative. So
what Kant actually does to remove the suspicion of the petitio is, first, to show how
and why the human being can understand himself as a member of the intelligible
world and therefore as free. Second, Kants discussion of the petitio and his fi nal
solution of it at the end of Sec. 3 also draws our attention to a problem which is
equally important, to wit, that by the thesis of analyticity the validity of the categori-
cal imperative has not been demonstrated; as a matter of fact, the common misun-
derstanding of Kants overall argument in GMS III is exactly what Kant warns his
readers against. On the circle also cf. Quarfoods paper in this volume.
306 Dieter Schnecker

an intelligible world and is thought of by him as an ought only inso-


far as he at the same time considers himself as a member of the sen-
sible world (GMS, 455,79, m. e.). And to quote yet another passage:
Under the presupposition of freedom of the will of an intelligence, its
autonomy, as the formal condition under which alone it can be deter-
mined, is a necessary consequence (GMS, 461,1417).
Kant describes an intelligence as a rational being that considers
itself solely as a member of the world of understanding. Thus, the will
in the context of the thesis of analyticity cannot simply be understood
as the will of a sensuous-rational being. It must be understood either
as the will of a perfectly rational being, whose free will always is a
good will; or, with an eye on human beings, it must be understood as
the will of a sensuous-rational being, whose will is both part of the
world of sense (Sinnenwelt) and of the world of understanding
(Verstandeswelt), whose will, however, as part of the latter world, is
the idea of reason, which would have full control over all subjective
motivations (GMS, 420,31, m. e.). It is exactly in this sense that in
the context of the deduction proper (Sec. 4), Kant writes that the cat-
egorical ought represents a synthetic proposition a priori by the fact
that my to will affected through sensible desires there is also added
the idea of precisely the same will, but one belonging to the world of
understanding, a pure will, practical for itself (GMS, 454,11, second
emphasis mine). In this perspective the human being, too, considers
his will as free, and such a free will always wills the good (all actions
would be perfectly in accord with the principle of the autonomy of the
pure will). This is the meaning of Kants thesis of analyticity.7
Thus Kant writes right after stating the thesis of analyticity:
Nonetheless, the latter is always a synthetic proposition: an absolutely
good will is that whose maxim can always contain itself considered as uni-
versal law, for through analysis of the concept of an absolutely good will
that quality of the maxim cannot be found. (GMS, 447,1014)

With the nonetheless (Indessen) Kant sets off the syntheticity


of the CI from the analyticity of the principle of morality (GMS,
447,6) just mentioned in the sentence before. Whereas this principle
follows [] from mere analysis (GMS, 447,89) of the concept of
freedom, from the analysis of the concept of an absolutely good will
it does not follow (cannot be found) that its maxim can always con-
7
I cannot discuss here Kants repeated claim that the categorical imperative is a syn-
thetic proposition; in any event, I would hold that, strictly speaking, it does not make
sense.
How is a categorical imperative possible? 307

tain itself considered as universal law, i. e. that such a will is always


good. And this means, first, that in the last passage quoted above the
sentence after the colon (an absolutely good will ) is a synthetic
proposition; for after the colon the reason is given for why the analy-
sis cannot take place which implies that we are, indeed, dealing with
a synthetcial proposition rather than an analytic one. Second, one has
to behold that the concept of an absolutely good will here does not
refer to the will of a perfectly rational being; for if one analyzes such
a concept, it does follow that its maxim can always contain itself con-
sidered as universal law. 8
In light of these considerations, what then does the question of how
a categorical imperative is possible really mean? Lack of space does
not allow to go into any analysis of the quite obvious parallel that
Kant likes to draw between the famous theoretical question of how
synthetic propositions a priori are possible. In any event, a closer look
reveals that the parallel question How is a categorical imperative
possible? actually includes three questions or aspects:
1. Why is the CI valid (binding, obligatory)?
2. How can freedom be understood, and why may we consider our-
selves to be free?
3. How can pure practical reason bring about an interest in the moral
law?
The third question cannot be answered.9 The second question is ans-
wered in Sec. 2 and Sec. 3 of GMS III: In Sec. 1, Kant first argues that
a perfectly rational and free will always wills (acts) morally. Sec. 2
demonstrates that because of its ability to think spontaneously (free-
ly), a rational being must also consider itself practically free; Sec. 3
then refers to the difference between the world of understanding and
8
From early on this passage and especially the (formulation of the) concept of an
absolutely good will has tremendously contributed to the misunderstandings of the
thesis of analyticity (cf. the report on the secondary literature in Schnecker, 1999,
168171). For Kant does indeed avail himself of this concept of an absolutely good
will to refer to a perfectly rational (holy) will (cf. GMS, 439,2934). However, he
also refers to an imperfect will as absolutely good: That will is absolutely good
which cannot be evil, hence whose maxim, if it is made into a universal law, can nev-
er conflict with itself (GMS, 437,6). In this passage and context Kant clearly does
not talk about perfectly rational (holy) beings. He refers to the will as absolutely
good only insasmuch its (particular) maxim can be universalized; cf. GMS, 426,10;
437,24; 437,32; 444,28.
9
In Sec. 5 Kant provides a lengthy justification for his claim that the question of how
a categorical imperative is possible can only be answered partially: Thus the ques-
tion, How is a categorical imperative possible? can be answered to this extent: :
(GMS, 461,7, m. e.); what cannot be answered is the question of how pure practical
reason indeed can be practical (cf. GMS 458,37; 459,34; 460,10; 461,25, 461,32).
308 Dieter Schnecker

the world of sense in order to argue that the human being also must
understand himself as practically free. But then, it seems, Kant still
thinks that the answer to that crucial question is still not answered;
recall that Kant in Sec. 3 still asks from whence the moral law obli-
gates (GMS, 450,16), a question that is not answered in Sec. 3 itself.
For now, I will leave it open what exactly it is that question is asking
for; as we will see later, Kant himself seems not to be entirely clear
about its meaning. In any event, what is needed is somehow the proof
that the CI is really valid, i. e. really binding on us. As long as this has
not been shown, morality could very well be a figment of the mind
(GMS, 407,17, 445,8); I will come back to this later.

2. The deduction of the categorical imperative

Sec. 4, then, raises our crucial question again and finally answers it.
This section is broken down into three paragraphs. In the first par-
agraph, Kant offers the argument proper. In the second paragraph,
Kant provides the final answer to the question of how categorical im-
peratives are possible (And thus categorical imperatives are possible
GMS, 454,6). Paragraph three comes back to common rational
moral cognition (something mentioned at the end of the preface); I
will not discuss it here.

2.1 Presuppositions of the deduction

Let me now quote the first four sentences of Sec. 4:


The rational being counts himself as intelligence in the world of 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
appearances of that causality are encountered, but whose possibility from
the latter, with which we have no acquaintance, is something into which
we can have no insight, but rather in place of that we have to have insight
into those actions as determined through other appearances, namely de-
sires and inclinations as belonging to the world of sense. As a mere mem-
ber of the world of understanding, all my actions would be perfectly in ac-
cord with the principle of the autonomy of the pure will; as a mere piece of
the sensible world, they would have to be taken as entirely in accord with
the natural law of desires and inclinations, hence with the heteronomy of
nature. (The former would rest on the supreme principle of morality, the
second on that of happiness.) (GMS, 453, 17)
How is a categorical imperative possible? 309

These lines are basically a summary of what Kant has said about
freedom and morality in Sec. 13, and so they are also a summary of
his thesis of analyticity. Using Kants own words of paragraphs 912
(GMS, 452,7453,15), the gist of it is this: Now the human being ac-
tually finds in himself a faculty, and this faculty is reason as a faculty
of pure spontaneity (self-activity). On account of this, a rational be-
ing has to regard itself as an intelligence (thus not from the side of its
lower powers), as belonging not to the world of sense but to the world
of understanding. As a rational being, hence one belonging to the in-
telligible world, the human being can never think of the causality of
his own will otherwise than under the idea of freedom. Now, with the
idea of freedom the concept of autonomy is inseparably bound up, and
with the latter the universal principle of morality. If we think of our-
selves as free, then we transport ourselves as members into the world
of understanding and cognize the autonomy of the will, together with
its consequence, morality; but if we think of ourselves as obligated by
duty, then we consider ourselves as belonging to the world of sense
and yet at the same time to the world of understanding. Again, par-
agraph 1 of Sec. 4 is nothing but a recapitulation of these basic ideas.
I will now present them in several theses (D14).
Thesis 1 states that every human being must understand itself as a
rational being:
(D1) The human being finds in itself the faculty of reason, which, as an
epistemic faculty, is a faculty of pure spontaneity.
It is important to see that Kants argument both in Sec. 3 and Sec. 4
begins with reason as a theoretical (epistemic) faculty. Only on pain
of self-contradiction, he argues, one can deny that one has reason, for
denying that one does presupposes that one does.10 The activity of
reason is self-activity, spontaneity, freedom, and these qualities justify
the human being to understand himself as an intelligence. And as an
intelligence, the human being must understand himself as a member
of the world of understanding.
(D2) As a rational being, a human being must understand himself as an
intelligence and, in this perspective, as a member of the world of under-
standing.

It is important to say in this perspective because only as an intel-


ligence, and inasmuch a human being is an intelligence, he may con-
10
Cf. Schnecker (1999, pp. 196316); here it is important to take into account Kants
Review of Schulzs Attempt at Introduction to a Doctrine of Morals (1783).
310 Dieter Schnecker

sider himself as a member of the world of understanding. Once the


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

Here it shows how important a proper understanding of the thesis of


analyticity is. For what that thesis claims is exactly that the principle
of morality is inseparably bound up with freedom.
(D4) Since the moral law is analytically bound up with freedom as a prop-
erty of the will of a rational being that is a member of the world of under-
standing, the human being also may understand his autonomy and the
moral law as the law of his rational volition, inasmuch as he understands
himself as such a rational being that is a member of the world of under-
standing.

Again, it is of utmost importance to realize that with this move the


validity of the moral law as a categorical imperative has not been dem-
onstrated. All that has been shown so far is that the will of a human

11
Cf. GMS, 448,1322.: Now one cannot possibly think a reason that, in its own
consciousness, would receive steering from elsewhere in regard to its judgments;
for then the subject would ascribe the determination of its power of judgment not
to its reason but to an impulse. It must regard itself as the author of its principles
independently of alien influences, consequently it must, as practical reason or as
the will of a rational being, be regarded by itself as free, i. e. the will of a rational
being can be a will of its own only under the idea of freedom and must therefore
with a practical aim be attributed to all rational beings (m. e.). How are we to read
consequently? The transition from the thinking I to the willing (acting) I only
appears plausible if the reason, the freedom of which cannot be denied without a
performative contradiction, is the very same reason that also is practical. That Kant
actually has something like this in mind shows in a thought of his from the preface.
There he says that it can in the end be only one and the same reason that is dis-
tinguished merely in its application (GMS, 391,27, m. e.). Kant calls this a unity
(GMS, 391,25) between theoretical and practical reason in a common principle
(GMS, 391,26). Only if this unity is comprehensively exhibited, Kant continues, a
critique of pure practical reason can be ventured and this is something, he says, he
could not do in the Groundwork. However, the truth of the matter is that in chapter
three a transition to such a critique of pure practical reason does take place, if only
its main feature (GMS, 445,15) is exhibited. But there is no argument whatsoever
for that alleged unity or identity of theoretical and practical reason. How this unity
is to be understood, how Kant moves from the freedom to think to the freedom to
will, and how from the concept of an intelligence to the concept of an intelligence
with a will we are left in the dark.
How is a categorical imperative possible? 311

being, considered merely as a member of the world of understanding,


is analytically bound up with the moral law. This is why Kant writes in
Sec. 4 of GMS III (453 f.) that all actions of a human being, if he were
only a member of the world of understanding, would be perfectly in
accord with the principle of the autonomy of the pure will; if a hu-
man being were alone a member of the world of understanding, all
his actions would always be in accord with the autonomy of the will.
That autonomy and hence morality is a consequence, as Kant puts
it several times, is only true under the presupposition of freedom of
the will of an intelligence (GMS, 461,14, m. e.).12 For only as intel-
ligence (GMS, 453,17, m. e.) does a human being understand himself
as a member of the world of understanding. In Sec. 5 Kant argues that
the human being must think of him[self] as intelligence, also as thing
in itself (GMS, 459,22) and hence of himself as the authentic self
(eigentliches Selbst, GMS, 457,34, m. e.). As an authentic self, a hu-
man bing gives himself the law. And so morality is that which the hu-
man being, in some sense and respect, authentically wills. Thats why
the moral ought is eigentlich ein Wollen (really a volition, GMS,
449,16, m. e.;), and thats why it is his own necessary volition as a
member of an intelligible world (GMS, 455,7, m. e.); I will come back
to this latter thought momentarily.

2.2 The ontoethical principle

Let me now quote the passage (both in German and in English) that
is at the center of the deduction:
Weil aber die Verstandeswelt den Grund der Sinnenwelt, mithin auch der
Gesetze derselben, enthlt, also in Ansehung meines Willens (der ganz
zur Verstandeswelt gehrt) unmittelbar gesetzgebend ist, und also auch
als solche gedacht werden mu, so werde ich mich als Intelligenz, obgleich
andererseits wie ein zur Sinnenwelt gehriges Wesen, dennoch dem Ge-
setze der ersteren, d. i. der Vernunft, die in der Idee der Freiheit das Ge-
setz derselben enthlt, und also der Autonomie des Willens unterworfen
erkennen, folglich die Gesetze der Verstandeswelt fr mich als Impera-
tive und die diesem Prinzip gemen Handlungen als Pflichten ansehen
mssen. (GMS, 453,31454,5)
But because the world of understanding contains the ground of the
world of sense, hence also of its laws, hence is immediately legislative in
regard to my will (which belongs wholly to the world of understanding),
and hence must also be thought of wholly as such, therefore as intelli-

12
The words of an intelligence are missing in Woods translation.
312 Dieter Schnecker

gence I will cognize myself, though on the other side as a being belonging
to the world of sense, as nevertheless subject to the laws of the first, i. e., to
reason, which in the idea of freedom contains the law of the understand-
ings world, and thus to autonomy of the will; consequently I must regard
the laws of the world of understanding for myself as imperatives and the
actions that accord with this principle as duties. (GMS, 453,31454,5)
An analysis of this sentence is quite tiresome; it is most important,
however, because it contains what I call the ontoethical principle
(OP). So please bear with me; also, keep in mind that we have to do
this referring to the English translation.
Lets first understand the grammatical structure of this elusive sen-
tence. The elements because-therefore break down the sentence into
two parts, the first of which obviously provides a reason (because)
for a conclusion drawn in the second (therefore).13 Abstracting as
much as we can from the actual content of what is being said, the first
passage of the first part says, grammatically speaking, this:
(OP1) The world of understanding contains the ground of the world of
sense.
Grammatically speaking, there is no problem with this part. Recon-
structing the next element of the sentence is also rather compelling
(although the hence will deserve more attention later):
(OP2) Hence the world of understanding also contains the ground of the
laws of the world of sense.
Now, one might think that because of the relationship between the
world of understanding and the world of sense laid out in OP1 and
OP2, the world of understanding is also immediately legislative in
regard to my will and hence must also be thought of wholly as such
(i. e. as immediately legislative). As we will see, that cannot be true.
Rather, a correct understanding is as follows:
(OP3) In regard to my will, which belongs wholly to the world of under-
standing, the world of understanding is immediately legislative and must
also, in regard to my will, be thought of such that it (the world of under-
standing) contains the ground of the world of sense and its laws.
Already for grammatical reasons, the first part of this reconstruction
cannot be disputed: In regard to my will, which belongs wholly to

13
The translation of the German folglich (GMS, 454,4) with consequently (and
the preceding colon) can be misleading. The folglich functions as an explanation
or elucidation rather than a consequence in any stricter sense; Ill come back to
this later.
How is a categorical imperative possible? 313

the world of understanding, the world of understanding is immedi-


ately legislative. What is problematic here is the relation of this part of
OP3 to OP1 and OP2, or to be more precise, problematic is both the
function of the conjunction because at the beginning of OP1 (But
because ) as well as the function of the adverb hence at the begin-
ning of what we have reconstructed as OP3 (hence is immediately
legislative ). Clearly, it would make no sense to interpret Kant as
arguing that the world of understanding contains the ground of the
world of sense and the ground of its laws and therefore (hence) were
immediately legislative for the will; it would not, because this very will
already belongs wholly to the world of understanding anyway and
not to the world of sense. However, this problem of interpretation can
be solved by pointing out that the because at the beginning of the
whole sentence must be put (and read) also at the beginning of OP3,
such that one must read: Because the world of understanding also in
regard to my will, which belongs wholly to the world of understand-
ing, is immediately legislative . The will is part of the world of un-
derstanding, and for this world of understanding it is true (according
to OP1 and OP2), that it contains the ground of the world of sense and
the ground of its laws. In what follows, Kant again avails himself (as
he did with the because) of an ellipsis; what is left out is in regard
to my will. The alleged fact that the will belongs wholly to the world
of understanding entails that the world of understanding must also
be thought of wholly as such; that is, however, not as a world which is
immediately legislative for this is clear anyway since the will belongs
to the world of understanding but as a world for which it is also
true in regard to the will as part of it, that it contains the ground of
the world of sense and its laws. Thus we get: It is also true in regard
to the will as part of the world of understanding that this will contains
the ground of the world of sense and its laws. As we will see later, it
is indeed the crux of Kants deduction that the pure will as a member
of the world of understanding contains the moral law as an imperative
for this very will as a member of the world of sense.
So we must reconstruct the first part of that sentence as follows:
(OP13) Because the world of understanding contains the ground of the
world of sense; and because hence the world of understanding also con-
tains the ground of the laws of the world of sense; and because hence the
world of understanding also in regard to my will, which belongs wholly to
the world of understanding, is immediately legislative and must also, in
regard to my will, be thought of such that it (the world of understanding)
contains the ground of the world of sense and its laws
314 Dieter Schnecker

The second part of the crucial sentence (GMS, 453,35454,5) is easier


to digest, yet still not a piece of cake. Lets look at it again: therefore
as intelligence I will cognize myself, although on the other side as a
being belonging to the world of sense, as nevertheless subject to the
laws of the first, i. e. to reason, which in the idea of freedom contains
the law of the understandings world, and thus to autonomy of the will;
consequently I must regard the laws of the world of understanding for
myself as imperatives and the actions that accord with this principle
as duties. It is broken down into two parts: The first part (in its Eng-
lish translation, of course) runs from therefore as to autonomy of the
will; the second from consequently to the end (as duties). One might
think that Kant in the first part Kant says mainly something like this:
therefore as intelligence I will cognize myself [] as [] subject to the
laws of the first, i. e. to reason, which in the idea of freedom contains
the law of the understandings world, and thus to autonomy of the will.
Thus one might think that what is stated in the omitted part (although
on the other side as a being belonging to the world of sense, as never-
theless ) only becomes relevant in the later second part, because the
human being as such a being (that belongs to the world of sense) re-
gards the laws as imperatives. However, I submit that the human being
in the entire second part of the ontoethical principle (sentence) must
be understood as a being that does not as intelligence cognize himself
subject to the law of the world of understanding and the autonomy of
the will; rather it does so as intelligence, although on the other side as a
being belonging to the world of sense. Three reasons speak in favor of
this reading: First, thats what it says why put that insertion at the end
of the sentence? In a number of parallel passages Kant also emphasizes
the simultaneity of both perspectives with regard to the imperative
character of the moral law. Secondly, inasmuch as the human being
considers himself merely as intelligence, he does not cognize himself
as subject to the moral law; thats part of the meaning of the thesis of
analyticity. This concept of a being that is subject to the moral law is,
strictly speaking, reserved for sensual-rational beings; inasmuch as the
human being considers himself as intelligence, he considers oneself
merely as a member of the world of understanding. Thirdly (and re-
lated to the last point), it is striking that Kant says that the human being
cognizes himself as nevertheless (obgleich) subject to the moral law.
For a being that considers itself only as intelligence, there is no need
to consider itself as nevertheless subject to the moral law for what
kind of difference could the nevertheless indicate if not the difference
between being a member of the world of understanding and the world
How is a categorical imperative possible? 315

of sense?14 Also, the consequently (folglich) only makes sense if it is


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

The rest of the sentence is easy. It, too, makes clear that the moral law,
which describes the volition of perfectly rational beings, is an impera-
tive for sensuous-rational beings:
(OP5) I must regard the laws of the world of understanding for myself as
imperatives and the actions that accord with this principle as duties.
Connecting these elements and adding the conjunctive particles, we
thus get the ontoethical principle:
(OP) Because the world of understanding contains the ground of the
world of sense; and because hence the world of understanding also con-
tains the ground of the laws of the world of sense; and because hence the
world of understanding also in regard to my will, which belongs wholly to
the world of understanding, is immediately legislative and must also, in
regard to my will, be thought of such that it (the world of understanding)
contains the ground of the world of sense and its laws, I cognize myself
understanding myself as a being belonging both to the world of under-
standing (intelligence) and to the world of sense as subject to the law of

14
Cf. the elucidating parallel in GMS, 450,1213 ( consider ourselves as free in
acting and thus nevertheless take ourselves to be subject to certain laws , m. e.).
316 Dieter Schnecker

the world of understanding, i. e. to reason, which in the idea of freedom


contains the law of the world of understanding, and thus as subject to the
autonomy of the will. Consequently, I must regard the laws of the world of
understanding for myself as imperatives and the actions that accord with
this principle as duties.15

Thus the deduction of the categorical imperative is completed in just


one sentence. This deduction is Kants answer to the question of how
a categorical imperative is possible. Therefore, it is no surprise that
right after stating OP, Kant officially gives the answer to that question
as well:
And thus categorical imperatives are possible through the fact that the
idea of freedom makes me into a member of an intelligible world, through
which, if I were that alone, all my actions would always be in accord with
the autonomy of the will; but since I intuit myself at the same time as
member of the world of sense, they ought to be in accord with it, which
categorical ought represents a synthetic proposition a priori by the fact
that to my will affected through sensible desires there is also added the
idea of precisely the same will, but one belonging to the world of under-
standing, a pure will, practical for itself, that contains the supreme condi-
tion of the first in accordance with reason (GMS, 454,615, m. e.).

Clearly, the deduction is not a deduction in any strict (deductive) sense


(at least not as long as the reconstruction sticks to the original text).
What then, exactly, is the key idea in Kants argument? The human
being is aware of himself as an intelligence due to the spontaneous
epistemic activities of his reason and understanding; as such an intel-
ligence, the human being is the eigentliche Selbst (authentic self;
GMS, 457,34; 458,2; 461,4). From there, Kant goes on (folglich;
GMS, 448,18) to the intelligence with a will. Considering himself as
an intelligence, the human being understands himself as a member of
the world of understanding and thus as a thing in itself; its rational
will, then, constitutes the eigentliche Selbst as a practical being. This
idea again, that the will as an intelligible faculty is the eigentliche
Selbst of the human being, as opposed to the human being inasmuch
as he is only appearance of himself (GMS, 457,35, m. e.), i. e. only a
phenomenon in the world of sense (GMS, 457,13) is the core of OP.
Kant argues for the validity of the categorical imperative as a moral
15
Cf. R 5086: In der Verstandeswelt ist das substratum: intelligentz, die Handlung
und Ursache: Freyheit, die Gemeinschaft: Glckseligkeit aus Freyheit, das Urwe-
sen: eine Intelligentz durch idee; die form: moralitaet, der nexus: ein nexus der
Zweke. Diese Verstandeswelt liegt schon itzt der Sinnenwelt zum Grunde und ist
das wahre selbstandige (m. e., andere Hervorhebungen getilgt).
How is a categorical imperative possible? 317

law for sensual-rational beings with the superiority of the ontic status
of the world of understanding. The human being as a thing in itself and
hence the eigentliche Selbst and its law is of higher ontic value then
the human being as an appearance; and this is why the law of the world
of understanding (the moral law) is binding upon the human being (as
a categorical imperative) who is a member both of the world of under-
standing and the world of sense. Thats the basic idea behind OP.
That this is, indeed, the basic idea behind OP can hardly be seen
just by reading OP itself (that sentence in GMS, 453). We have to look
at other passages. An external characteristic may lead the way: In that
passage in GMS, 453, OP1 and OP2 are emphasized, an emphasis that
at this length, as far as I know, hardly ever can be found in Kants writ-
ings (if at all). Exactly parallel to this, Kant again provides the answer
to the question of how a categorical imperative is possible in GMS,
461. And just as in GMS, 453 f., he again avails himself of a lengthy
emphasis to stress that the moral law
is valid [!] for us as [!] human beings, since [!] it has arisen from our will as
intelligence, hence from our authentic self; but what belongs to the mere
appearance is necessarily subordinated by reason to the constitution of
the thing in itself (GMS, 461,26).

The law of the world of appearances is the natural law of desires and
inclinations (GMS, 453,28); these desires and inclinations, as ap-
pearances (GMS, 453,24, m. e.), determine human actions. However,
this law just belongs to the mere appearance and therefore is neces-
sarily subordinated by reason to the constitution of the thing in itself.
Here again it becomes clear that the whole force of Kants argument
depends on the ontic superiority of the authentic self.
And there is yet another passage that provides textual evidence for
this interpretation. The human being, Kant says, as a rational being is
a member of the world of understanding, and
since in that world he himself only as intelligence is the authentic self
(as human being, by contrast, only appearance of himself), those laws [of
the world of understanding] apply to him immediately and categorically
(GMS, 457,3336, m. e.)

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

as a member of the world of understanding which elevates a human


being above himself (as a part of the sensible world) (KpV, 86, m. e.);
hence it is not to be wondered at that a human being, as belonging
to both worlds, must regard his own nature in reference to his second
and highest vocation only with reference, and its laws with the highest
respect (KpV, 87, m. e.). This, Kant says, is the origin (KpV, 86) of
duty. This axiological position is also reflected in the last paragraph
of Sec. 4 of the GMS, where Kant writes that even the most wicked
scoundrel [] transports himself in thoughts into entirely another or-
der of things (GMS, 454,2131) in which he finds a greater inner
worth of his person (GMS, 454,37): This better person, however, he
believes himself to be when he transports himself into the standpoint
of a member of the world of understanding (GMS, 454,37455,2)
Thus we can reconstruct the ontoethical principle as follows:
(OP*) The world of understanding and thus the pure will as a member of
this world of understanding are ontically superior to the world of sense,
and therefore the law of this world and will (the moral law) is binding as
a categorcial imperative for beings that are both members of the world of
understanding and the world of sense.16
16
Steigleder (2006, this volume, pp. 225246) rejects my interpretation of Kants the-
sis of analyticity and therefore also my interpretation of Sec. 4. Although it would
be most interesting to have a detailed discussion, it is simply impossible to have it
here for lack of space. Hence just three brief comments: First, no interpretation can
be satisfying that is not comprehensive; its always easy, too easy indeed, to make
claims about what a text means by ignoring those passages that dont fit. (Thus,
Steigleder is right to point out that there are elements in Sec. 1 that pose a prob-
lem for my interpretation; I have addressed these elements in Schnecker, 1999.)
This being said, Id like to reply, second, that Steigleder correctly points out that
my interpretation is partly based on the overall structure of GMS III (that at least
part of GMS II and GMS III are preparation and that there is a Sec. 4 to say the
very least). However, it is certainly not sufficient to reply to this crucial element of
my interpretation by just asserting that there may conceivably be other readings
(Steigleder 2006, 242, m. e.) without actually providing an alternative reading that
is also in a position to account for the overall structure of GMS III. As did many
before him, Steigleder still argues that it is only Kants intention that we must
necessarily see ourselves as rational beings, which have a will (Steigleder, 2006,
p. 243). Clearly, however, this has been demonstrated no later than in Sec. 3. But
what then is the purpose of Sec. 4 and that long and complicated sentence that is (or
includes) what I call the ontoethical principle? Steigleder provides no answer, and
it is therefore no surprise that in his book (2002, 6796), Steigleder shows no inter-
est whatsoever in that sentence either (he does, however, very briefly mention that
the Gesetz unserer Vernunft und das Verlangen unserer Bedrfnisse [] nicht
auf gleicher Stufe stehen, p. 89, m. e.). There may very well be other readings of it;
but whether there are any, we will not figure out by ignoring it. Leaving aside that
Steigleder, as I see it, pays no sufficient attention to Kants repeated claim that there
is no categorical imperative for perfectly rational beings and does not sufficiently
distinguish between the moral law as an synthetic imperative and as an analytical
How is a categorical imperative possible? 319

2.3 Validity and Motivation

I already pointed out that, according to Kant, the good is something


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

law (though this distinction is, obviously, at the very heart of Kants question of how
a categorical imperative is possible), Id like to note, third, that Steigleder does not
have an explanation for Kants repeatedly posed question from whence the moral
law obligates. I do, by the way, agree that it is certainly worthwhile to consider the
option that Kant in the Groundwork already distinguishes, as Steigleder suggests,
between Wille and Willkr; but again, this issue I cannot possibly address here.
17
In the preface Kant already mentions the Grund der Verbindlichkeit of the CI
(ground of an obligation; cf. GMS, 389,12; 389,16; 391,11; 432,31; 439,31; 439,33;
448,34). He also speaks of the Realitt of the CI (reality; GMS, 425,14; 449,26), of
its Wirklichkeit (reality; cf. vgl. 420,1; vgl. 406,15), Geltung (validity; cf. GMS,
389,12; 389,14; 403,7; 408,18; 412,3; 424,35; 425,18; 442,8; 447,32; 448,6; 448,32;
449,29; 460,25; 461,3; 461,12), Richtigkeit (correctness; cf. GMS, 392,13), objek-
tiven Notwendigkeit (objective necessity; cf. GMS 442,9; 449,26; 449,30); Kant
says that the CI gibt (is; cf. 419,18), that it wirklich stattfinde (is; cf. GMS, 425,9)
and that the human being ought to unterwerfen himself to it (subject himself to it;
cf. GMS, 449,12). All this concepts and formulations are probably best subsumed
under the idea of the Gltigkeit dieses Imperativs (validity of this imperative;
cf. GMS 461,12). Also note that Kant not only speaks of the deduction of the CI,
but also of its Beweis etc. (proof; cf. GMS, 392,4; 392,13; 403,27; 412,28; 425,8;
425,15; 427,17; 431,33; 440,2028; 445,1; 447,30448,4; 449,27).
18
For our context, I assume that answers which could be classified as formalistic
(such as Karl-Otto Apels or Habermas) fail from the word go; but they would clas-
sify as a third possible answer. On the Moral Question, cf. Schnecker (2006).
320 Dieter Schnecker

ing of morality developed in GMS I/II, this question, if understood as


Why be moral?, obviously cannot be asking for the utility of the CI;
it cannot ask for how the CI might serve ones interests. It is important
to see, though, that although Kant goes out of his way to argue that
the question of how pure practical reason can be practical cannot be
answered, he is not only interested in the validity of the CI, but also in
its power to motivate. In this respect, he points out that pure practial
reason (the pure will) is still ones own will. Kant argues that, if this
will and its law were not ones own will, and hence the law were not
ones own law, one still were to find a nonmoral motive to actually
comply with this law, i. e. a motive to want (or to do, respectively) what
one would not want to do: For if one thought of him only as subject to
a law (whatever it might be), then this would have to bring with it some
interest as a stimulus or coercion, because as a law it did not arise
from his will, but rather this will was necessitated by something else to
act in a certain way in conformity with the law (GMS, 432,32433,3).
In the metaphysics of morals as part of GMS II, Kant had already
asked for the possibility (GMS, 427,17) of a will that is merely deter-
mined by reason and thus asked for the ground of a possible categori-
cal imperative, i. e. of a practical law (GMS, 428,5). This ground, he
says, can only be something whose existence in itself has an absolute
worth (GMS, 428,3) a rational and free being as an end in itself.
Since the absolute worth of rational beings as ends in themselves is
grounded in their autonomy,19 Kant then in order not to beg the ques-
tion (for the law of this autonomy is the CI the very validity of which
is suspicious) argues in GMS III with the ontoethical principle and
the ontic superiority of the pure will of the eigentliche Selbst. In the
context of GMS II, Kant sets forth the postulate (GMS, 429,35) that
every rational being must understand his or her own being as such an
end in itself (a postulate that is then demonstrated in sections 2 and
3 of GMS III). He already provides arguments that this status justifies
why one must, and how one can, subject oneself to the moral law as a
categorical imperative: The will is thus not solely subject to the law,
but is subject in such a way that it must be regarded also as legislating
to itself, and precisely for this reason as subject to the law (of which
it can consider itself as the author) (GMS, 431,21, second emphasis
D. S.). It is because the human being gives himself the universal law

19
Cf. Kants lecture Feyerabend (1319 ff.). I cannot get into this but one should not
forget that an important variant of the CI is based on the idea of the human being
as an end in itself; cf. Schnecker / Wood (22004, 140153).
How is a categorical imperative possible? 321

that this law can motivate us: The moral ought is thus his own nec-
essary volition as a member of an intelligible world and is thought by
him as an ought only insofar as he at the same time considers himself
as a member of the sensible world (GMS, 455,7, m. e.); this is why
Kant so often emphasizes that this ought is really a volition (GMS,
449,16, m. e.). Thus Kant not only attempts to demonstrate that there
is a good reason to be moral, but at the same time and this is partly
responsible for the confusion GMS can easily cause the deduction of
the moral law provides an incentive to be moral: Das Gute ist immer
das, was ieder Mensch will, und er wrde es auch immer thun, wenn
es ihm nur nicht schwer wrde, es auszuben, und wenn unsere Natur
so beschaffen wre, da wir immer nach dem Begriff des guten han-
delten, so wren wir recht frei. 20 The moral insight that Kant wants
to induce with the addressee of the moral law is that he, the addressee
himself, in some sense already wants what he ought to do.

2.4 Some brief critical points

As mentioned before, it is not the purpose of this paper to criticize (or


further develop) Kants deduction. However, let us briefly look at some
critical points which will also help us to better understand Kants argu-
ment. I dont find it convincing at all, neither from an external nor from
an internal point of view. Externally speaking, I would criticize Kants
axiology as much too narrow because it allows only rational beings to
have worth; but thats a long story and, in any case, not the issue here.
Internally speaking (i. e. presuming Kants own critical philoso-
phy), it is quite obvious that Kant avails himself of an ontological in-
terpretation of his own distinction between thing in itself and appear-
ance that otherwise is merely a epistemological distinction. Lets have
a look at OP one more time. The first part of the ontoethical princi-
ple (OP13), we said, must be reconstructed as follows: Because the
world of understanding contains the ground of the world of sense; and
because hence the world of understanding also contains the ground of
the laws of the world of sense; and because hence the world of under-
standing also in regard to my will, which belongs wholly to the world
of understanding, is immediately legislative and must also, in regard
to my will, be thought of such that it (the world of understanding) con-
tains the ground of the world of sense and its laws . As seen earlier
in our detailed analysis, it seems that Kant has a general principle in

20
MM, 903.
322 Dieter Schnecker

mind according to which the world of understanding is the ground


of the laws of the world of sense which is only applied to the special
case of the will; hence what Kant would be saying is that the pure will,
which is a member of the world of understanding, contains the ground
for the laws of the will as a member of the world of sense. And as a
matter of fact, right after the deduction in Sec. 4, Kant says that the
good will of an evil man constitutes by his own admission the law for
his evil will as a member of the sensible world (GMS, 455,46).
But how are we to understand this general principle? Kant him-
self suggests in Sec. 4 reading it in light of his epistemological asser-
tion that the logical subject is the ground for the laws of nature. How-
ever, that the world of understanding is the ground of the laws of the
world of sense does not follow, as suggested by the hence in OP2,
from Kants own fundamental claim, that one must assume behind
the appearances something else that is not appearance, namely the
things in themselves (GMS, 451,1214). On the other hand, even if
one accepted that somehow the logical subject is the ground of the
laws of the world of sense in terms of that epistemological assertion of
Kants, how are we to understand the application of that general prin-
ciple on the will? At best, it seems, there is some kind of resemblance
between the legislation of the logical subject and the legislation of
the practical subject. And yet it is this resemblance that Kants seems
to have in mind. For he closes his official answer to the question of
how a categorical imperative is possible by drawing a parallel between
OP13 and the aforementioned epistemological claim, saying that it
is approximately in this way that concepts of the understanding, which
for themselves signify nothing but lawful form in general, are added
to intuitions of the world of sense and through that make possible syn-
thetic propositions a priori on which rests all cognition of a nature
(GMS, 454,1519, m. e.). Well, Kant himself admits that this parallel
is only ungefhr (approximately).
All the more we have to rely in our interpretation of OP on later
passages in which Kant emphasizes the ontic superiority of the world
of understanding (GMS, 457; 461). However, not only is this alleged
superiority in itself dubious, but it is also not in harmony with Kants
overall understanding of the distinction between things in themselves
and appearances. According to this understanding, there is, in itself,
one world. This world (of things in themselves) is what it is; this very
world as understood (interpreted) by us, is called the world of sense.
The latter is not, however, in any sense inferior to the former unless
one understands the fact that appearances, except from being appear-
How is a categorical imperative possible? 323

ances, do not exist as establishing their inferiority; if so, one would


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

Literature

Kants writings

All textual references of the Groundwork are according to Allen Woods


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

Other works

Damschen, Gregor / Schnecker, Dieter (2006): Selbst philosophieren, Stutt-


gart.
Quarfood, Marcel (2006): The Circle and the two standpoints, in: This vol-
ume, pp. 285300.
Schnecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kate-
gorischen Imperativs, Freiburg i. Br./Mnchen.
Schnecker, Dieter (2001): Textvergessenheit in der Philosophiehistorie,
in: Schnecker, Dieter / Zwenger, Thomas (Hrsg.): Kant verstehen/Under-
standing Kant. ber die Interpretation philosophischer Texte, 159181.
Schnecker, Dieter (2006): Warum moralisch sein? Eine Landkarte fr
Moralische Realisten, in: Heiner F. Klemme / Manfred Khn / Dieter
Schnecker (eds.): Moralische Motivation. Kant und die Alternativen. Fe-
lix Meiner Verlag, Hamburg.
324 Dieter Schnecker

Schnecker, Dieter / Wood, Allen W. (22004): Immanuel Kant: Grundlegung


zur Metaphysik der Sitten. Ein einfhrender Kommentar, Paderborn.
Steigleder, Klaus (2002): Kants Moralphilosophie. Die Selbstbezglichkeit
reiner praktischer Vernunft, Stuttgart.
Steigleder, Klaus (2006): The Analytic Relationship of Freedom and Moral-
ity On GMS III, 1, in this volume, pp. 225246.
Wood, Allen W. (1999): Kants Ethical Thought, Cambridge University Press.
A Bibliography on Kants Groundwork

The following bibliography is an updated version of the bibliography that was


made available by Bernd Kraft and Dieter Schnecker in their edition of the
Grundlegung zur Metaphysik der Sitten with Felix Meiner Verlag, Hamburg
(1999). This updated bibliography takes into account literature published be-
tween 2000 und 2005. Like Krafts and Schneckers, it is committed to the
idea to include only publications that somewhat specifically refer to Kants
Groundwork rather than to Kants ethics in general.1

Acton, Harry B. (1970): Kants Moral Philosophy. New York.


Allison, Henry E. (1990): Kants Theory of Freedom. Cambridge.
Allison, Henry E. (1996): On the Presumed Gap in the Derivation of the Cat-
egorical Imperative. In: Ders.: Idealism and Freedom. Essays on Kants
Theoretical and Practical Philosophy, Cambridge, 143154.
Alqui, Ferdinand (1974): La morale de Kant. Paris.
Alqui, Ferdinand (1985): Prsentation des Fondements de la Mtaphysique
de Murs. Tome II, Paris.
Albrecht, Michael (1994): Kants Maximenethik und ihre Begrndung. In:
Kant-Studien 85, 129146.
Ameriks, Karl (22000): Kants Theory of Mind. An Analysis of the Paralo-
gisms of Pure Reason. Oxford.
Ameriks, Karl (2001): Zu Kants Argumentation am Anfang des Dritten Ab-
schnitts der Grundlegung. In: Hans-Ulrich Baumgarten / Carsten Held
(Hrsg.): Systematische Ethik mit Kant, Freiburg i. Br./Mnchen, 2454.
Ameriks, Karl / Sturma, Dieter (Hrsg.) (2004): Kants Ethik. Paderborn.
Atwell, John E. (1969): Are Kants First Two Moral Principles Equivalent?
In: Journal of the History of Philosophy 7, 273284.
Aubenque, Pierre (1975): La Prudence chez Kant. In: Revue de Metaphy-
sique et Morale, 156182.
Aune, Bruce (1979): Kants Theory of Morals. Princeton.
Baker, Judith (1988): Counting Categorical Imperatives. In: Kant-Studien 79,
389406.
Barnes, Gerald W. (1971): In Defense of Kants Doctrine of the Highest
Good. In: Philosophical Forum 2, 446458.
Baron, Marcia (1995): Kantian Ethics almost without Apology. Ithaca.
Baumanns, Peter (1982): Kants kategorischer Imperativ und das Problem der
inhaltlichen Pflichtbestimmung. In: Herta Nagl-Docekal (Hrsg.): ber-

1
Thanks to Raoul Simon Weber (Bonn) for important help in updating the bibliog-
raphy.
326 Bibliography

lieferung und Aufgabe. Festschrift fr Erich Heintel zum 70. Geburtstag.


2. Teilband, Wien, 165179.
Baumanns, Peter (2000): Kants Ethik: Die Grundlehre. Wrzburg.
Baumgarten, Hans-Ulrich / Held, Carsten (Hrsg.) (2001): Systematische
Ethik mit Kant, Freiburg, i. Br./Mnchen.
Benton, Robert J. (1978): The Transcendental Argument in Kants Ground-
work of the Metaphysic of Morals. In: The Journal of Value Inquiry (The
Hague) 12, 225237.
Bittner, Rdiger (1974): Maximen. In: Gerhard Funke (Hrsg.): Akten des 4.
Internationalen Kant-Kongresses, Berlin/New York, 485498.
Bittner, Rdiger (1980): Hypothetische Imperative. In: Zeitschrift fr phi-
losophische Forschung 34, 210226.
Bittner, Rdiger (1983): Moralisches Gebot oder Autonomie. Freiburg i. Br./
Mnchen.
Bittner, Rdiger (1986): Handlungen und Wirkungen. In: Gerold Prauss
(Hrsg.): Handlungstheorie und Transzendentalphilosophie, Frankfurt a. M.,
1326.
Bittner, Rdiger (2000): Wer frei ist, ist gebunden: Kants Argument aus dem
dritten Abschnitt der Grundlegungsschrift. In: Philosophiegeschichte und
logische Analyse / Logical Analysis and History of Philosophy Bd. 3 (Phi-
losophie der Neuzeit / From Descartes to Kant), 209221.
Bollnow, Otto Friedrich (1966): Als allein ein guter Wille . Zum Anfang
der Grundlegung zur Metaphysik der Sitten. In: Wolfgang Ritzel (Hrsg.):
Rationalitt Phnomenalitt Individualitt. Festgabe fr Hermann
und Marie Glockner, Bonn, 165174.
Brandt, Reinhard (1988): Der Zirkel im dritten Abschnitt von Kants Grundle-
gung zur Metaphysik der Sitten. In: Hariolf Oberer / Gerhard Seel (Hrsg.):
Kant. Analysen Probleme Kritik, Wrzburg, 169191.
Brinkmann, Walter (2001): Die Goldene Regel und der Kategorische Im-
perativ: Rationalitt und praktische Notwendigkeit. In: Volker Gerhardt /
Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.): Kant und die Berliner
Aufklrung, Bd. 3, Berlin/New York, 1320.
Bruton, Samuel V. (2000): Establishing Kants Formula of Humanity. In:
Southwest Philosophy Review 16, 4149.
Bubner, Rdiger (2001): Another Look at Maxims. In: Predrag Cicovacki
(Hrsg.): Kants Legacy: Essays in Honour of Lewis White Beck, Roches-
ter, 245259.
Burri, Alex / Freudiger, Jrg (1990): Zur Analytizitt hypothetischer Impera-
tive. In: Zeitschrift fr philosophische Forschung 44, 98105.
Canivet, Michel (1980): Enseigner les Fondements de la mtaphysique des
murs. In: Revue Philosophique de Louvain 78, 365384.
Carnois, Bernard (1973): La Cohrence de la Doctrine Kantienne de la Lib-
ert. Paris.
Cicovacki, Predrag (2001): Zwischen gutem Willen und Kategorischem Im-
perativ. In: Hans-Ulrich Baumgarten / Carsten Held (Hrsg.): Systema-
tische Ethik mit Kant, Freiburg i. Br./Mnchen, 330354.
Copp, David (1992): The Possibility of a Categorical Imperative Kants
Groundwork, Part III. In: Philosophical Perspectives 6, 261284.
Bibliography 327

Craemer-Ruegenberg, Ingrid (1983): Logische und andere Eigenschaften des


kategorischen Imperativs. In: Neue Hefte fr Philosophie 22, 4561.
Cramer, Konrad (1972): Hypothetische Imperative? In: Manfred Riedel
(Hrsg.): Rehabilitierung der praktischen Philosophie. Bd. 1, Freiburg
i. Br., 159212.
Cramer, Konrad (1991): Metaphysik und Erfahrung in Kants Grundlegung
der Ethik. In: Neue Hefte fr Philosophie 30/31, 1568.
Croitori, Rodica (1995): Kant and Totalitarianism from the Point of View of
the Groundwork of the Metaphysics of Morals. In: Hoke Robinson (Hrsg.):
Proceedings of the Eighth International Kant Congress (Memphis), Mil-
waukee. Volume II, Part 2, 813820.
De Bolt, Darian C. (2002): Pasternack on Intrinsic Value and Moral Over-
ridingness in Kants Groundwork. In: Southwest Philosophy Review: The
Journal of the Southwestern Philosophical Society 18, 121125.
Delbos, Victor (1969): La philosophie practique de Kant. Paris.
Delbos, Victor (1993): La moral de Kant (Nachwort zu: Immanuel Kant:
Fondements de la mtaphysique des murs. Traduction et notes par Vic-
tor Delbos). Paris, 153202.
Delfosse, Heinrich P. (2000): Kant-Index, Band 15: Stellenindex und Kon-
kordanz zur Grundlegung zur Metaphysik der Sitten, Stuttgart/Bad
Cannstatt.
Denis, Lara (2001): Moral Self-Regard: Duties to Oneself in Kants Moral
Theory. New York.
Duncan, Alistair Robert Campbell (1957): Practical Reason and Morality.
A Study of Immanuel Kants Foundations for the Metaphysics of Morals.
London/New York.
Dye, James W. (1978): Kant as Ethical Naturalist. In: The Journal of Value
Inquiry 12, 111125.
Ebbinghaus, Julius (1986): Deutung und Mideutung des kategorischen Im-
perativs. In: Ders.: Gesammelte Schriften. Band I, Praktische Philosophie
19291954, herausgegeben von Hariolf Oberer und Georg Geismann,
Bonn, 279295.
Ebbinghaus, Julius (1988): Die Formeln des kategorischen Imperativs und
die Ableitung inhaltlich bestimmter Pflichten. In: Ders.: Gesammelte
Schriften. Band II, Praktische Philosophie 19551972, herausgegeben von
Georg Geismann und Hariolf Oberer, Bonn, 209229.
Engstrom, Stephen (1997): Kants Conception of Practical Wisdom. In: Kant-
Studien 88, 1643.
Ertl, Wolfgang (2001): Das Freiheitsproblem bei Kant: Zwischen Tradition
und Innovation. In: Jens Kulenkampff / Thomas Spitzley (Hrsg.): Von der
Antike bis zur Gegenwart: Erlanger Streifzge durch die Geschichte der
Philosophie, Erlangen, 7996.
Fleischer, Margot (1963): Das Problem der Begrndung des kategorischen
Imperativs bei Kant. In: Paulus Engelhardt (Hrsg.): Sein und Ethos. Un-
tersuchungen zur Grundlegung der Ethik, Mainz, 387404.
Fleischer, Margot (1964): Die Formeln des kategorischen Imperativs in Kants
Grundlegung zur Metaphysik der Sitten. In: Archiv fr Geschichte der
Philosophie 46, 201226.
328 Bibliography

Forschner, Maximilian (1974): Gesetz und Freiheit. Zum Problem der Au-
tonomie bei I. Kant. Mnchen/Salzburg.
Frster, Eckart (1992): Was darf ich hoffen? Zum Problem der Verein-
barkeit von theoretischer und praktischer Vernunft bei Immanuel Kant.
In: Zeitschrift fr philosophische Forschung 46, 168185.
Fouille, Alfred (1904): Kant a-t-il tabli lexistence du devoir? Revue de la
Mtaphysique et de Morale 12, 493523.
Freudiger, Jrg (1993): Kants Begrndung der praktischen Philosophie. Sys-
tematische Stellung, Methode und Argumentationsstruktur der Grundle-
gung zur Metaphysik der Sitten. Bern/Stuttgart/Wien.
Galay, Jean-Louis (1977): Philosophie et invention textuelle. Essai sur la
potique dun texte kantien. Paris.
Gass, Michael (1994): Kants Causal Conception of Autonomy. In: History of
Philosophy Quarterly 11, 5370.
Geismann, Georg (2002): Die Formeln des kategorischen Imperativs nach
H. J. Paton, N. N., Klaus Reich und Julius Ebbinghaus. In: Kant-Studien
93, 374384.
Geismann, Georg (Hrsg.) (2002): Julius Ebbinghaus Die Naturgesetz-
Formel des kategorischen Imperativs und die Ableitung inhaltlich bestim-
mter Pflichten. In: Kant-Studien 93, 371373.
Genova, A. C. (1978): Kants Transcendental Deduction of the Moral Law.
In: Kant-Studien 69, 299313.
Glass, Ronald (1971): The Contradictions in Kants Examples. In: Philosoph-
ical Studies 22, 6570.
Green, Ronald M. (2001): What Does It Mean to Use Someone As A Means
Only: Rereading Kant. In: Kennedy-Institute of Ethics Journal 11, 247
261.
Grnewald, Bernward (1988): Praktische Vernunft, Modalitt und transzen-
dentale Einheit. Das Problem einer transzendentalen Deduktion des Sit-
tengesetzes. In: Hariolf Oberer / Gerhard Seel (Hrsg.): Kant. Analysen
Probleme Kritik, Wrzburg, 127167.
Gunkel, Andreas (1989): Spontaneitt und moralische Autonomie. Kants
Philosophie der Freiheit. Bern/Stuttgart.
Gupta, R. K. (1971): Kants Groundwork of Morality. In: Studi Internazionali
di Filosofia 3, 111161.
Gupta, R. K. (1973): Kants Problem of the Impossibility of the Categorical
Imperative. In: Kant-Studien 64, 4955.
Guyer, Paul (1995): The Possibility of the Categorical Imperaive. In: The
Philosophical Review 104, 353385.
Guyer, Paul (Hrsg.) (1998): Kants Groundwork of the Metaphysics of Mor-
als: Critical Essays. Totowa.
Guyer, Paul (2001): The Form and Matter of the Categorial Imperative. In:
Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.):
Kant und die Berliner Aufklrung, Bd. 1, Berlin/New York, 131150.
Guyer, Paul (2002): The Derivation of the Categorical Imperative: Kants
Correction for a Fatal Flaw. In: Harvard Review of Philosophy 10, 6480.
Haardt, Alexander (1982): Die Stellung des Personalittsprinzips in der
Grundlegung zur Metaphysik der Sitten und in der Kritik der praktischen
Vernunft. In: Kant-Studien 73, 157168.
Bibliography 329

Harbison, Warren (2000): Self-Improvement, Beneficence, and the Law of


Nature Formula. In: Kant-Studien 91, 1724.
Harris, N. G. E. (1988): Imperfect Duties and Conflicts of Will. In: Kant-Stu-
dien 79, 3342.
Harrison, Jonathan (1957): Kants Examples of the First Formulation of the
Categorical Imperative. In: The Philosophical Quarterly 7, 5062.
Heidemann, Ingeborg (1956): Person und Welt. Zur Kantinterpretation von
Heinz Heimsoeth. In: Kant-Studien 48, 344360.
Henrich, Dieter (1954/55): Das Prinzip der Kantischen Ethik. In: Philoso-
phische Rundschau 2, 2038.
Henrich, Dieter (1973): Der Begriff der sittlichen Einsicht und Kants Lehre
vom Faktum der Vernunft. In: Gerold Prauss (Hrsg.): Kant. Zur Deutung
seiner Theorie von Erkennen und Handeln, Kln, 223254.
Henrich, Dieter (1975): Die Deduktion des Sittengesetzes. ber die Grnde
der Dunkelheit des letzten Abschnittes von Kants Grundlegung zur Me-
taphysik der Sitten. In: Alexander Schwan (Hrsg.): Denken im Schatten
des Nihilismus. Festschrift fr Wilhelm Weischedel zum 70. Geburtstag,
Darmstadt, 55112.
Henrich, Dieter (1982): Ethik der Autonomie. In: Ders.: Selbstverhltnisse.
Gedanken und Auslegungen zu den Grundlagen der klassischen deut-
schen Philosophie, Stuttgart, 656.
Herman, Barbara (1993): The Practice of Moral Judgement. Cambridge.
Hill Jr., Thomas E. (1992): Dignity and Practical Reason in Kants Moral
Theory. Ithaca.
Hill Jr., Thomas E. / Zweig, Arnulf (2002): Editors Introduction: Some Main
Themes of the Groundwork; Analysis of Arguments. In: Immanuel Kant
(1785): Groundwork for the Metaphysics of Morals, translated by Arnulf
Zweig and edited by Thomas E. Hill Jr. and Arnulf Zweig, Oxford, 19108
und 109177.
Hochberg, Gary M. (1973): A Re-Examination of the Contradiction in Kants
Examples. In: Philosophical Studies 24, 264267.
Hoerster, Norbert (1974): Kants kategorischer Imperativ als Test unserer
sittlichen Pflichten. In: Manfred Riedel (Hrsg.): Rehabilitierung der
praktischen Philosophie. Bd. 2: Rezeption, Argumentation, Diskussion,
Freiburg i. Br., 455475.
Hffe, Otfried (1977): Kants kategorischer Imperativ als Kriterium des Sit-
tlichen. In: Zeitschrift fr philosophische Forschung 31, 354384.
Hffe, Otfried (1985): Introduction la philosophie pratique de Kant. La
morale, le droit et la religion. Albeuve.
Hffe, Otfried (Hrsg.) (1989): Grundlegung zur Metaphysik der Sitten. Ein
kooperativer Kommentar. Frankfurt a. M.
Hgemann, Brigitte (1980): Die Idee der Freiheit und das Subjekt. Eine Un-
tersuchung von Kants Grundlegung zur Metaphysik der Sitten. Meisen-
heim am Glan Knigstein/Ts.
Horn, Christoph (2004): Die Menschheit als objektiver Zweck Kants Selbst-
zweckformel des Kategorischen Imperativs. In: Dieter Sturma (Hrsg.),
Kants Ethik, Paderborn, 195212.
Hunter, Ian (2002): The Morals of Metaphysics: Kants Groundwork as Intel-
lectual Paideia. In: Critical Inquiry 28, 908929.
330 Bibliography

Hutchings, Patrick A. (1972): Kant on Absolute Value. A Critical Examina-


tion of Certain Key Notions in Kants Groundwork of the Metaphysics of
Morals and of his Ontology of Personal Value. Detroit.
Ilting, Karl-Heinz (1972): Der naturalistische Fehlschlu bei Kant. In: Man-
fred Riedel (Hrsg.): Rehabilitierung der praktischen Philosophie. Bd. 1:
Geschichte, Probleme, Aufgaben, Freiburg i. Br., 113130.
Irrlitz, Gerd (2002): Kant-Handbuch: Leben Werk Wirkung. Stuttgart.
Jacobsen, Mogens Chrom (1992): ber das Verhltnis zwischen Immanuel
Kants Rechts- und Moralphilosophie. In: Danish Yearbook of Philosophy
27, 7292.
Johnson, Darrell (1995): Analytic and Synthetic Method and the Structure of
Kants Grounding. In: Hoke Robinson (Hrsg.): Proceedings of the Eighth
International Kant Congress (Memphis), Milwaukee. Volume II, Part 2,
613620.
Kaulbach, Friedrich (1988): Immanuel Kants Grundlegung zur Metaphysik
der Sitten. Interpretation und Kommentar. Darmstadt.
Kemp, J. (1958): Kants Examples of the Categorical Imperative. In: The
Philosophical Quarterly 8, 6371.
Kerstein, Samuel J. (2002): Kants Search for the Supreme Principle of Mo-
rality. Cambridge.
Kersting, Wolfgang (1983): Kann die Kritik der praktischen Vernunft pop-
ulr sein? ber Kants Moralphilosophie und pragmatische Anthropolo-
gie. In: Studia Leibnitiana 15, 8293.
King, J. Charles (1968): Bradleys Duty for Dutys Sake and Kants Ethics.
In: Kant-Studien 59, 309317.
Kirchmann, Julius Hermann v. (1874): Erluterungen zur Grundlegung der
Metaphysik der Sitten. Leipzig.
Klamp, Gerhard (1949): Zum Problem der Textverschiebung in Kants
Grundlegung zur Metaphysik der Sitten. In: Zeitschrift fr philosophische
Forschung 4, 574578.
Klein, Sherwin (1988): Kants Methodology in Grundlegung, Section One.
In: The Modern Schoolman 65, 227244.
Khl, Harald (1990): Kants Gesinnungsethik. Berlin/New York.
Konhardt, Klaus (1986): Faktum der Vernunft? Zu Kants Frage nach dem
eigentlichen Selbst des Menschen. In: Gerold Prauss (Hrsg.): Handlung-
stheorie und Transzendentalphilosophie, Frankfurt a. M., 160184.
Korsgaard, Christine M. (1996): From Duty and for the Sake of the Noble:
Kant and Aristotle on Morally Good Action. In: Stephen Engstrom / Jen-
nifer Whiting (Hrsg.): Aristotle, Kant, and the Stoics: Rethinking Happi-
ness and Duty, Cambridge, 203236.
Korsgaard, Christine M. (1996): Creating the Kingdom of Ends. Cam-
bridge.
Korsgaard, Christine M. (1998): Einleitung zu: Immanuel Kant: Ground-
work of the Metaphysics of Morals, translated and edited by Mary Gregor.
Cambridge.
Kraft, Bernd / Schnecker, Dieter (1999): Einleitung. In: Immanuel Kant,
Grundlegung zur Metaphysik der Sitten, mit einer Einleitung hrsg. von
Bernd Kraft und Dieter Schnecker, Hamburg, VIIXXXIX.
Bibliography 331

Kraft, Bernd / Schnecker, Dieter (2000): Kants Grundlegung zur Meta-


physik der Sitten in der Edition der Akademieausgabe. In: Kant-Studien
91 (Sonderheft), 2022.
Krausser, P. (1968): ber eine unvermerkte Doppelrolle des kategorischen
Imperativs in Kants Grundlegung zur Metaphysik der Sitten. In: Kant-Stu-
dien 59, 318332.
Kupperman, Joel J. (2002): A Messy Derivation of the Categorial Impera-
tive. In: Philosophy 77, 484502.
Laberge, Pierre (1980): Du Passage de la Philosophie Morale Populaire la
Mtaphysique des Murs. In: Kant-Studien 71, 418444.
Laberge, Pierre (1982): LEspce de cercle dont, ce quil semble, il ny pas
moyens de sortir. In: Dialogue 21, 745753.
Liddell, Brendan E. A. (1970): Kant on the Foundation of Morality. A Mod-
ern Version of the Grundlegung. Translated with Commentary. Bloom-
ington, London.
Liddell, Brendan E. A. (1972): Kants Deduction in the Grundlegung. In:
Lewis White Beck (Hrsg.): Proceedings of the Third International Kant
Congress, Dordrecht-Holland, 401406.
Lhrer, Guido (1995): Menschliche Wrde. Wissenschaftliche Geltung und
metaphorische Grenze der praktischen Philosophie Kants. Freiburg i. Br./
Mnchen.
Maliandi, Ricardo (1974): Bedeutung und Zweideutigkeit des als-ob in
der Naturgesetz-Formel des kategorischen Imperativs. In: Gerhard Funke
(Hrsg.): Akten des 4. Internationalen Kant-Kongresses, Berlin/New York,
540549.
Marshall, John (1982): Hypothetical Imperatives. In: American Philosophi-
cal Quarterly 19, 105114.
Marshall, John (1985): The Syntheticity of the Categorical Imperative. In:
Gerhard Funke / Thomas M. Seebohm (Hrsg.): Proceedings of the Sixth
International Kant Congress. Vol. II/2, 185200.
Matson, Wallace I. (1961): Foundations of the Metaphysics of Morals. In:
Frank N. Magill (Hrsg.) / Ian P. McGreal (Associate Editor): World Phi-
losophy. Essay-Reviews of 225 Major Works. Volume 3, Englewood Cliffs,
N. J., 11311141.
Mautner, Thomas (1971): Kants Grundlegung. In: The Philosophical Quar-
terly 22, 255257.
McCarthy, Michael H. (1976): Analytic Method and Analytic Propositions in
Kants Groundwork. In: Dialogue 15, 565582.
McCarthy, Michael H. (1979): Kants Application of the Analytic/Synthetic
Distinction to Imperatives. In: Dialogue (Canada) 18, 373391.
McCarthy, Michael H. (1979): Patons Suggestion that Kants Principle of
Autonomy might be Analytic. In: Kant-Studien 70, 206224.
McCarthy, Michael H. (1982): Kants Rejection of the Argument of Ground-
work III. In: Kant-Studien 73, 169190.
McCarthy, Michael H. (1984): Kants Groundwork Justification of Freedom.
In: Dialogue (Canada) 23, 457473.
McCarthy, Michael H. (1985): The Objection of Circularity in Groundwork
III. In: Kant Studien 76, 2842.
332 Bibliography

Melches Gibert, Carlos (1994): Der Einflu von Christian Garves berset-
zung Ciceros De Officiis auf Kants Grundlegung zur Metaphysik der Sit-
ten. Regensburg.
Melnick, Arthur (2002): Kants Formulations of the Categorical Imperative.
In: Kant-Studien 93, 291308.
Miller, R. D. (1993): An Interpretation of Kants Moral Philosophy. Harrogate.
Milz, Bernhard (1998): Zur Analytizitt und Synthetizitt der Grundlegung.
In: Kant-Studien 89, 188204.
Moritz, Manfred (1972): Die Probleme der Deduktion des kategorischen Im-
perativs (ein Vorbericht). In: Lewis White Beck (Hrsg.): Proceedings of
the Third International Kant Congress (Rochester), Dordrecht-Holland,
424433.
Mulholland, Leslie (1978): Kant: On Willing Maxims to Become Laws of
Nature. In: Dialogue 17, 92105.
Oberer, Hariolf (1983): Kants praktische Philosophie. In: Karl Brthlein
(Hrsg.): Zur Geschichte der Philosophie. Bd. 2, Von Kant bis zur Gegen-
wart, Wrzburg, 1530.
ONeill, Onora (1989): Agency and Anthropology in Kants Groundwork.
In: Yirmiyahu Yovel (Hrsg.): Kants Practical Philosophy Reconsidered.
Papers presented at the Seventh Jerusalem Philosophical Encounter, De-
cember 1986, Dordrecht/Boston/London, 6382.
ONeill, Onora (1989): Constructions of Reason: Explorations of Kants
Practical Philosophy. Cambridge.
Pasternack, Lawrence (Hrsg.) (2002): Immanuel Kant: Groundwork of the
Metaphysics of Morals. London/New York.
Pasternack, Lawrence (2002): Intrinsic Value and Moral Overridingness in
Kants Groundwork. In: Southwest Philosophy Review: The Journal of
The Southwestern Philosophical Society 18, 113121.
Paton, Henry James (1947): The Categorical Imperative. A Study in Kants
Moral Philosophy, London.
Paton, Henry James (1958): The Aim and Structure of Kants Grundlegung.
In: The Philosophical Quarterly 8, 112130.
Patt, Walter (1997): The Synthetic Character of the Moral Law According to
Kant. In: Hariolf Oberer (Hrsg.): Kant. Analysen Probleme Kritik. Bd.
III, Wrzburg, 2139.
Patzig, Gnther (1973): Die logischen Formen praktischer Stze in Kants
Ethik. In: Gerold Prauss (Hrsg.): Kant. Zur Deutung seiner Theorie von
Erkennen und Handeln, Kln, 207222.
Patzig, Gnther (1980): Der kategorische Imperativ in der Ethik-Diskussion
der Gegenwart. In: Ders.: Tatsachen, Normen, Stze. Aufstze und Vor-
trge, Stuttgart, 155177.
Patzig, Gnther (1986): Principium diiudicationis und Principium execu-
tionis: ber transzendentalpragmatische Begrndungsanstze fr Ver-
haltensnormen. In: Gerold Prauss (Hrsg.): Handlungstheorie und Tran-
szendentalphilosophie, Frankfurt a. M., 204218.
Pelegrinis, Theodosius N. (1980): Kants Conceptions of the Categorical Im-
perative and the Will. London.
Petrus, Klaus (1994): Beschrieene Dunkelheit und Seichtigkeit. Histor-
isch-systematische Voraussetzungen der Auseinandersetzung zwischen
Bibliography 333

Kant und Garve im Umfeld der Gttinger Rezension. In: Kant-Studien


85, 280302.
Philonenko, Alexis (1972): Luvre de Kant. La philosophie critique. Tome
II, Paris.
Pistorius, Hermann Andreas (1786): Rezension der Grundlegung zur Meta-
physik der Sitten. In: Rdiger Bittner / Konrad Cramer (Hrsg.): Materialien
zu Kants Kritik der praktischen Vernunft, Frankfurt a. M. 1975, 144160.
Pogge, W. (1994): Freudigers Grundlegung. In: Grazer Philosophische Stu-
dien 47, 223239.
Potter, Nelson (1974): The Argument of Kants Grundlegung, Chapter One.
In: Canadian Journal of Philosophy. Supp. 1, 7391.
Potter, Nelson (1993): What is wrong with Kants four examples? In: Journal
of Philosophical Research 18, 213229.
Prauss, Gerold (1983): Kant ber Freiheit als Autonomie. Frankfurt a. M.
Quarfood, Marcel (2001): Kants Practical Deduction of Moral Obligation
in Groundwork III. In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph
Schumacher (Hrsg.): Kant und die Berliner Aufklrung, Bd. 3, Berlin/
New York, 7279.
Rawls, John (1989): Themes in Kants Moral Philosophy, in: E. Frster (ed.),
Kants Transcendental Deductions, Stanford, 81113.
Rawls, John (2000): Lectures on the History of Moral Philosophy, ed. by B.
Herman, Harvard.
Reath, Andrews (1989): The Categorical Imperative and Kants Conception
of Practical Rationality. In: Monist 72, 384410.
Reath, Andrews (2003): Immanuel Kant, Groundwork of the Metaphysics
of Morals (1785): Duty and Autonomy. In: Jorge J. E. Gracia (Hrsg.): The
Classics of Western Philosophy: A Readers Guide, Malden, 346356.
Reboul, Olivier (1967): Prescription ou Proscription? Essai sur le sens du
devoir chez Kant. In: Revue de Metaphysique et Morale 72, 295320.
Reich, Klaus (1935): Kant und die Ethik der Griechen. Tbingen.
Reiner, Hans (1949): Eine Textverschiebung in Kants Grundlegung zur Meta-
physik der Sitten. In: Zeitschrift fr philosophische Forschung 4, 123124.
Reiner, Hans (1949): Stellungnahme zu den vorgelegten Bemerkungen. In:
Zeitschrift fr philosophische Forschung 4, 579581.
Rickless, Samuel C. (2004) From the Good Will to the Formula of Universal
Law. In: Philosophy and Phenomenological Research 68, 554577.
Riedel, Manfred (1989): Kritik der moralisch urteilenden Vernunft. Kants
vorkritische Ethik und die Idee einer Grundlegung zur Metaphysik der
Sitten. In: Ders.: Urteilskraft und Vernunft. Kants ursprngliche Frage-
stellung, Frankfurt a. M., 6197.
Rodriguez Lopez, Blanca (2002): Sobre el caracter hipotetico de los imperativos
morales. In: Telos: Revista Iberoamerica de Estudios Utilitaristas 11, 99119.
Rollin, Bernard E. (1976): There is Only One Categorical Imperative. In:
Kant-Studien 67, 6072.
Ross, Sir William David (1954): Kants Ethical Theory. A Commentary on
the Grundlegung zur Metaphysik der Sitten. Oxford.
Rossvaer, Viggo (1979): Kants Moral Philosophy. An Interpretation of the
Categorical Imperative. Oslo.
334 Bibliography

Sandermann, Edmund (1989): Die Moral der Vernunft. Transzendentale


Handlungs- und Legitimationstheorie in der Philosophie Kants. Freiburg
i. Br./Mnchen.
Sartiaux, F. (1917): Morale Kantienne et Morale humaine. Paris.
Schaller, Walter E. (1995): From the Groundwork to the Metaphysics of Mor-
als: What happened to Morality in Kants Theory of Justice? In: History of
Philosophy Quarterly 12, 333345.
Schneewind, J. B. (1992): Autonomy, Obligation, and Virtue: An Overview
of Kants Moral Philosophy. In: Paul Guyer (Hrsg.): The Cambridge Com-
panion to Kant, Cambridge, 309341.
Schnoor, Christian (1989): Kants Kategorischer Imperativ als Kriterium der
Richtigkeit des Handelns. Tbingen.
Schndorf, Harald (1985): Denken-Knnen und Wollen-Knnen in
Kants Beispielen fr den kategorischen Imperativ. In: Zeitschrift fr phi-
losophische Forschung 39, 549573.
Schnecker, Dieter (1996): Zur Analytizitt der Grundlegung. In: Kant-Stu-
dien 87, 348354.
Schnecker, Dieter (1997): Die Methode der Grundlegung und der ber-
gang von der gemeinen sittlichen zur philosophischen Vernunfterkenntnis.
In: Hariolf Oberer (Hrsg.): Kant. Analysen Probleme Kritik. Bd. III,
Wrzburg, 8198.
Schnecker, Dieter (1997): Gemeine sittliche und philosophische Vernunf-
terkenntnis. Zum ersten bergang in Kants Grundlegung. In: Kant-Stu-
dien 88, 311333.
Schnecker, Dieter (1997): Die Art von Zirkel im dritten Abschnitt der
Grundlegung. In: Allgemeine Zeitschrift fr Philosophie 22, 189202.
Schnecker, Dieter (1999): Kant: Grundlegung III. Die Deduktion des kate-
gorischen Imperativs. Freiburg i. Br./Mnchen.
Schnecker, Dieter (2001): What is the First Proposition Regarding Duty in
Kants Grundlegung? In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph
Schumacher (Hrsg.): Kant und die Berliner Aufklrung, Bd. 3, 8995.
Schnecker, Dieter (2001): Kants Grundlegung zur Metaphysik der Sitten le-
sen. In: Information Philosophie, Heft 5, 5056.
Schnecker, Dieter / Wood, Allen W. (22004): Immanuel Kant: Grundlegung
zur Metaphysik der Sitten. Ein einfhrender Kommentar. Paderborn.
Scott, J. W. (1924): Kant on the Moral Life. An Exposition of Kants Grundle-
gung. London.
Sherline, Edward (1995): Heteronomy and Spurious Principles of Morality in
Kants Groundwork. In: Pacific Philosophical Quarterly 76, 3246.
Singer, Marcus G. (1975): Verallgemeinerung in der Ethik. Zur Logik mora-
lischen Argumentierens. Frankfurt a. M.
Singer, Marcus G. (1983): Reconstructing the Groundwork. In: Ethics 93,
566578.
Skorpen, Erling (1981): Making Sense of Kants Third Example. In: Kant-
Studien 72, 415429.
Staege, Roswitha (2002): Hypothetische Imperative. In: Kant-Studien 93,
4256.
Bibliography 335

Stattler, Benedikt (1788): Anhang zum Anti-Kant in einer Widerlegung der


Kantischen Grundlegung zur Metaphysik der Sitten. Mnchen, wiederab-
gedruckt in: Aetas Kantiana. Brssel.
Steigleder, Klaus (2001): Hypothetische Imperative als reflexive Urteile. In:
Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schumacher (Hrsg.),
Kant und die Berliner Aufklrung, Bd. 3, Berlin/New York, 113121.
Steigleder, Klaus (2002): Kants Moralphilosophie. Die Selbstbezglichkeit
reiner praktischer Vernunft. Stuttgart/Weimar.
Steinberger, Peter J. (1999): The Standard View of the Categorical Impera-
tive. In: Kant-Studien 90, 9199.
Storheim, E. (1979): Kants und Fichtes Begrndung der Moral. In: Klaus
Hammacher / Albert Mues (Hrsg.): Erneuerung der Transzendental-
philosophie im Anschlu an Kant und Fichte. Reinhard Lauth zum 60.
Geburtstag, Stuttgart-Bad Cannstatt, 411424.
Stratton-Lake, Philip (1993): Formulating Categorical Imperatives. In: Kant-
Studien 84, 317340.
Stratton-Lake, Philip (2000): Kant, Duty and Moral Worth. London/New
York.
Sullivan, Roger J. (1989): Immanuel Kants Moral Theory. Cambridge.
Sullivan, Roger J. (1994): An Introduction to Kants Ethics. Cambridge.
Swoyer, Chris (1983): Kantian Derivations. In: Canadian Journal of Philoso-
phy 13, 409431.
Teale, A. E. (1951): Kantian Ethics. London.
Tenbruck, Friedrich (1949): Bemerkungen zu H. Reiners Textkorrektur. In:
Zeitschrift fr philosophische Forschung 4, 572574.
Terzis, George N. (1989): Autonomy without Indeterminism: A Reconstruc-
tion of Kants Groundwork, Chapter Three. In: The Modern Schoolman
67, 113.
Timmermann, Jens (2000): Kant und die Lge aus Pflicht: Zur Auflsung
moralischer Dilemmata in einer kantischen Ethik. In: Philosophisches
Jahrbuch 107, 267283.
Timmermann, Jens (2001): The Dutiful Lie: Kantian Approaches to Moral
Dilemmas. In: Volker Gerhardt / Rolf-Peter Horstmann / Ralph Schu-
macher (Hrsg.): Kant und die Berliner Aufklrung, Bd. 3, 345354.
Timmons, Mark (1984): Contradictions and the Categorical Imperative. In:
Archiv fr Geschichte der Philosophie 66, 294312.
Timmons, Mark (1989): McCarthy on Practical Necessitation in Kant. In:
Kant-Studien 80, 198207.
Timmons, Mark (1992): Necessitation and Justification in Kants Ethics. In:
Canadian Journal of Philosophy 22, 223261.
Tugendhat, Ernst (1993): Vorlesungen ber Ethik. Frankfurt a. M.
Vialatoux, J. (1956): La morale de Kant. Paris.
Wagner, Hans (1993): Kants schwierige Bestimmung des Verhltnisses
zwischen Freiheit und Moralgesetz. (Zum 3. Abschnitt der Grundle-
gung zur Metaphysik der Sitten. A.-A. IV 446463). In: Reinhold Breil
/ Stephan Nachtsheim (Hrsg.): Vernunft und Anschauung. Philosophie
Literatur Kunst. Festschrift fr Gerd Wolandt zum 65. Geburtstag,
Bonn, 315.
336 Bibliography

Wagner, Hans (1994): Kants Konzept von hypothetischen Imperativen. In:


Kant-Studien 85, 7884.
Walsh, Dorothy (1971): The Dilemma of the Grundlegung. In: The Philo-
sophical Quarterly 21, 149157.
Ward, Keith (1972): The Development of Kants View of Ethics. Oxford.
Wartenberg, Thomas E. (1992): Reason and the Practice of Science. In: Paul
Guyer (Hrsg.): The Cambridge Companion to Kant, Cambridge, 228
248.
Wenzel, Uwe Justus (1992): Anthroponomie. Kants Archologie der Auton-
omie. Berlin.
Westphal, Kenneth R. (1995): How Full is Kants Categorical Imperative?
Jahrbuch fr Recht und Ethik 3, 465509.
Wike, Victoria S. (1987): The Role of Happiness in Kants Groundwork. In:
The Journal of Value Inquiry 21, 7378.
Wilde, Leo Henri (1975): Hypothetische und kategorische Imperative. Eine
Interpretation zu Kants Grundlegung zur Metaphysik der Sitten. Bonn.
Willaschek, Marcus (1992): Praktische Vernunft: Handlungstheorie und
Moralbegrndung bei Kant. Stuttgart/Weimar.
Williams, T. C. (1968): The Concept of the Categorical Imperative. A Study
of the Place of the Categorical Imperative in Kants Ethical Theory. Ox-
ford.
Wimmer, Reiner (1980): Universalisierung in der Ethik. Analyse, Kritik und
Rekonstruktion ethischer Rationalittsansprche. Frankfurt a. M.
Wolff, E. (1966): Les trois Imperatifs Categoriques et les trois Postulats de la
Raison Pratique chez Kant. In: Archives de philosophie 29, 3758.
Wolff, Robert P. (Hrsg.) (1969): Foundations of the Metaphysics of Morals.
Translated by Lewis White Beck, with Critical Essays Edited by Robert
Paul Wolff. New York.
Wolff, Robert P. (1973): The Autonomy of Reason. A Commentary on Kants
Groundwork of the Metaphysic of Morals. New York.
Wood, Allen W. (1972): Kant on False Promises. In: Lewis White Beck
(Hrsg.): Proceedings of the Third International Kant Congress, Dor-
drecht-Holland, 614619.
Wood, Allen W. (1976): Kant on the Rationality of Morals. In: P. La Berge
/ F. Duchesneau / B. E. Morrisey (Hrsg): Proceedings of the Ottawa Con-
gress on Kant in the Anglo-American and Continental Traditions, Ot-
tawa, 94109.
Wood, Allen W. (1999): Kants Ethical Thought. Cambridge.
Wood, Allen W. (2002): Immanuel Kant, Groundwork for the Metaphysics
of Morals. Edited and translated by Allen W. Wood. With Essays by J. B.
Schneewind, Marcia Baron, Schelly Kagan, Allen W. Wood, New Haven.
General Index

a priori, apriority 3, 5, 7, 1116, 61, 142, Baumgarten, Hans-Ulrich 126, 326


154 f., 158, 165, 171 ff., 186, 194, 196, Bayles, Richard E. 167, 197
228233, 238 ff., 243, 249 f., 293 f., Beck, Lewis White 111, 126, 138, 195,
296 f., 304, 306 f., 316, 322 197, 261, 270 f., 277, 283, 326, 331 f., 336
Achenwall, Gottfried 151 f., 157 Bennett, Jonathan 277, 283
Acton, Harry 325 Benton, Robert J. 326
agency 58, 60 f., 63, 65, 67, 78 ff., 88, 111, Bittner, Rdiger 15, 18, 21, 326, 333
165, 172, 195 Blum, Lawrence 72, 91
Albrecht, Michael 325 Bollnow, Otto-Friedrich 326
Allison, Henry 74, 79 f., 83, 91, 94, 103, Brandt, Reinhard 18, 21, 296 f., 300,
106111, 112 f., 117, 134, 138, 248, 271, 326
283, 286, 291, 293, 296, 300, 325 Breil, Reinhold 335
Alqui, Ferdinand 325 Brink, David O. 167, 197
Ameriks, Karl 214, 221, 325 Brinkmann, Walter 136, 138, 326
analytic/synthetic 17 f., 145 f., 152 f. Bruton, Samuel 326
analyticity, thesis of 241 f., 301, 303 Bubner, Rdiger 326
307 Burri, Alex 326
Anaxagoras 54 f.
Annas, Julia 72, 91 Canivet, Michel 326
anthropology 6, 15, 67, 250 Capobianco, Richard 301
Apel, Karl-Otto 115, 319 Carnois, Bernard 326
appearance 286, 297, 316 f., 321 f. categorical imperative 7 f., 10 f., 18,
argument by elimination 201 ff., 211 ff. 41 ff., 63, 65, 73, 86, 94100, 104, 107
Aristotelian biology 57 116, 125, 130 f., 137, 141 ff., 150, 154,
Aristotelian teleology 48 158196, 200221, 228 f., 231, 236,
Aristotle 52, 127, 330 241243, 247 f., 262, 293 ff., 301323
Atwell, John E. 325 categorical imperative, deduction of
Aubenque, Pierre 325 108, 111, 113, 241 ff., 291, 294 f., 301
Aune, Bruce 96, 103106, 108111, 113, 323
117, 197, 325 categorical imperative, formula of
automaton spirituale 275 f., 279 autonomy 32, 37, 178 f., 190 f., 193
autonomy 61,129, 131, 178 f., 190 f., 229, categorical imperative, formula of hu-
232 f., 236, 243, 245, 248, 250, 253 f., manity 31, 169, 177179, 189191,
257, 259265, 268272, 275, 278, 200221, 261 f.
280 ff., 320 f. categorical imperative, formula of uni-
versal law 36, 65, 93, 113, 158, 185,
Bagattini, Alexander 121 193, 218 f., 262
Baker, Judith 81, 91, 325 philosophical significance of 187191
Barnes, Gerald W. 325 practical significance of 192 f.
Baron, Marcia 7880, 83 f., 86, 88, 90 f., as a decision procedure 161, 167177
160, 197, 325, 336 core assumptions of 168 f.
Brthlein, Karl 332 formal constraint interpretation of
Baumanns, Peter 325 f. 185191
338 General Index

categorical imperative, formula of Duncan, Alistair R. Campbell 327


universal law of nature (FLN) 158 duties of self-perfection 172, 181, 183
185, 193 duty 6 ff., 11 ff., 34, 36 f., 47, 72 f., 82, 87,
formal constraint interpretation of 97, 101, 114 f., 133, 137, 158, 160 ff.,
FLN 160 170174, 177, 179183, 187 f.,192, 195,
strong model of 167, 169171 216, 219, 228, 294 f., 299, 304 f., 318
categorical imperative, possibility of duty, acting from 3240, 7290, 100
229 f., 240, 301323 103, 105, 113, 132 f., 213 f.
categorical imperatives (pl.) 97 f., 116 Dye, James W. 327
causality 58, 63 f., 124, 145, 147, 233 ff.,
239, 244, 248, 251 ff., 257 ff., 268, 275, Ebbinghaus, Julius 327
277 f., 308 ff. Eisler, Rudolf 283
character 27, 35, 39 f., 88, 214, 270 end, necessary 227229, 236, 240 f.
Chrysippus 53 end, objective 56, 200, 202 f.
Cicero 331 end, subjective 202 f.
Cicovacki, Predrag 326 end-in-itself 31, 40, 64 f., 169, 178, 180,
circle 285299, 304 f. 189 f., 204 f., 210 ff., 229, 261 f., 320, 323
circulus in probando 290 f., 305 Engelhardt, Paulus 327
Collins 61, 141 Engstrom, Stephen 327, 330
conceptual necessitiy 10 f. Epicureanism 52, 6668
consistency test 164 f. Epicurus 53
contradiction-in-conception 164, 173 f., Ertl, Wolfgang 327
182, 194198 ethical principles 710, 17, 20 f.
contradiction-in-will 164, 173 f., 182 f., Eudaemonism 49, 53, 66
196 Ewing, Alfred 196 f.
Copp, David 326 faculty of desire 74, 124126, 128,
Cotter, Alexander 69, 225, 301 146 f., 233 f., 244, 271
Craemer-Ruegenberg, Ingrid 326 Ferrari, Jean 53, 70
Cramer, Konrad 327, 333 Fichte, Johann Gottlieb 335
Critique of practical reason 19 f., 60, Fleischer, Margot 327
228, 231, 310 Forschner, Maximilian 327
Croitori, Rodica 327 Frster, Eckart 327
Fouille, Alfred 328
Damschen, Gregor 301, 323 Frank, Manfred 58, 70
Darwinism 57 Frankfurt, Harry 259
De Bolt, Darian C. 327 free choice 233 ff., 237, 244 f. (see also
deduction (see categorical imperative, freie Willkr)
deduction of) freedom 51, 61, 64 f., 67, 84, 232, 235 f.,
Delbros, Victor 327 239 f., 243 f., 247282, 285, 287288,
Denis, Lara 83, 91, 327 292, 297, 300, 304 f., 310
derivation 37, 93116, 126 f., 172 f., freedom of the will 60, 111 f., 129, 233,
200 ff., 209 ff., 217 f., 288 236 ff., 242, 245, 247282, 287 f., 291 f.,
Descartes, Ren 326 302 ff., 306, 309 ff.
determination of judgement 251, 253, freedom, negative 235, 248, 253 f., 256
263, 279, 310 259, 264 f., 267, 269, 271276, 280 ff.
determination, natural 111, 234 f., 256, freedom, positive 235 f., 240, 248, 253 f.,
275, 278, 282 (see also pre-determina- 258 f., 269, 271 f., 275, 282
tion) freedom, transcendental 111 f., 232,
determination, of the will 142, 150 f., 258, 260, 265, 275 f.
237, 239, 244 f., 269 Freudiger, Jrg 21, 326
dignity 40, 46, 189191, 213 Funke, Gerhard 326, 331
Donagan, Alan 185, 197
Downie, R. S. 150, 157 Galay, Jean-Louis 328
Duchesneau, F. 326 Gallop, David 55, 70
General Index 339
Galvin, Richard 165, 172, 195, 197 Held, Carsten 326
gap theorists 94, 103 ff. Helvtius 5153
Garve, Christian 331, 332 Henrich, Dieter 259, 283, 329
Gass, Michael 328 Henson, Richard G. 81, 91
Gaut, Berys 204, 221 Herman, Barbara 60, 70, 75, 81, 91, 160,
Geismann, Georg 327 f. 169, 173 f., 195, 197 f., 329
Genova, A. C. 328 heteronomy 129, 253, 257, 259, 264,
Gerhardt, Volker 124, 138, 326, 328, 269272, 278, 308, 323
333335 Hill, Thomas E. jr. 84, 86, 91, 201, 207,
gifts of fortune 26, 28 221
gifts of nature 2629, 214 Himmelmann, Beatrix 53, 66, 70
Glasgow, Joshua 183, 194, 196 f. history 5863, 65
Glass, Ronald 328 Hochberg, Gary M. 329
Glockner, Hermann 326 Hoerster, Norbert 329
Glockner, Marie 326 Hffe, Otfried 159, 195, 198, 329
God 10, 32, 122, 124, 131133, 213, 279 Hgemann, Brigitte 329
Gods existence 45, 66 holism (ethical) 115 f.
good will 2543, 4553, 66, 72 f., 77 f., holy will 3742, 121 f., 129134, 304,
82 f., 89 f., 101, 104 ff., 113, 129133, 307
202, 212220, 238 f., 244, 278, 294, Horn, Christoph 116, 196, 329
306 f., 322 Horstmann, Rolf-Peter 326, 328, 333
good, absolute 2830, 38, 133, 204 , 335
212, 220, 238 f., 307 (see also good will) humanity 31, 80, 178 f., 200 ff., 261 f.
good, highest (summum bonum) 47, 61, humanity, as biological species 49
63, 66 Hume, David 58, 124
good-natured temperament 81 f. Hunter, Ian 329
goodness, objective 52, 201, 205 ff., Hutchins, Patrick 329
210 ff. hypothetical/categorical 97 ff.,122 f.,
goodness, preeminent 101, 212218 134 ff.,141 ff., 150 ff., 207, 230 f.
goodness,unconditional 201, 204, 206, hypothetical imperative 41, 43, 61, 98 f.,
210 ff., 215 116, 122 f.,135 f., 136, 139156, 230 f.,
Gracia, Jorge J. E. 333 243, 264 ff.
Green, Ronald M. 328 counsels of prudence 123, 136 f., 141,
Gregor, Mary 248, 270 f. 144, 153
Grnewald, Bernward 328 rules of skill 123, 136, 141, 153
Gunkel, Andreas 328
Gupta, R. K. 328 Ilting, Karl-Heinz 330
Guyer, Paul 70, 83, 91, 328 imperative 41 f., 121123, 125, 129131,
134137, 139146, 153, 155, 207 f.,
Haardt, Alexander 328 226 f.
Habermas, Jrgen 319 imperatives, as judgements 226, 231
Hammacher, Klaus 335 imperatives, structure of 136, 242 f.
happiness 20, 27, 29 f., 36, 41, 43, 4654, imperfect duties 8587, 175
60 f., 6669, 83,116, 136 f.,144, 153, 167, incentive 33, 38, 60, 80 f., 83, 127, 130,
172, 204, 226 f., 278, 308 214, 294, 321
Harbison, Warren 328 inclination 30, 33 ff., 48, 66 ff., 7385,
Hare, Richard 143, 157, 186, 197 89, 122 f., 129 ff., 140, 149 f., 162, 189 f.,
Harris, N. G. E. 328 200, 204, 207, 212 ff., 219, 227, 229, 234,
Harrison, Jonathan 172, 195, 197, 329 247, 253, 255 f., 259, 264 f., 268 f., 271 f.,
hedonism 53, 67, 69 317
Hegel, Georg F. W. 108, 158 f., 197 indeterminism 247, 252, 257, 259, 268
Heidemann, Ingeborg 329 272
Heimsoeth, Heinz 329 instinct 46, 49, 5153, 74, 234
Heintel, Erich 326 intelligence 306, 309 ff., 314
340 General Index

interests 41, 47, 79, 122, 234, 240, 319 f. laws of reason 33, 125, 127, 131 f., 253
Irrlitz, Gerd 330 258, 260, 265, 267270, 272 f., 278,
Irwin, Terence H. 68, 70 280 ff.
legality 8
Jacobsen, Mogens Chrom 330 Leibniz, Gottfried Wilhelm 275
Johnson, Darrell 330 Leonard (character of Memento) 173
Johnson, Monte Ransome 48, 70 Lidell, Brendan E. A. 331
Jowett, B. 54 f. life 124, 234
judgement 5, 7 ff., 11 f., 16 ff., 20, 40, 65, living being 124 ff., 234, 242, 244
115, 142 ff., 152156, 168, 172, 174, 181, logical structure of ethical principles
190, 251 f., 259 ff., 263, 265 ff., 273 f., 3 f., 711
278 f., 285, 292, 297, 310 Lhrer, Guido 331
judgement, synthetic a priori 230 ff., Looney, Aaron 3
239 f., 243 Ludwig, Bernd 156 f.
judgement, analytic 153, 156, 230, 238 f.
judgement, practical 226233, 252 ff., Magil, Frank N. 331
266, 268 f., 273 Maliandi, Ricardo 331
judgement, table of 142, 152, 155 Mandeville, Bernard 5153
Marc-Wogau, Konrad 58, 70
Kagan, Schelly 336
Kaulbach, Friedrich 330 Marshall, John 331
Kemp, J. 330 Matson, Wallace I. 331
Kerstein, Samuel J. 75, 91, 113 f., 160, Mautner, Thomas 331
176, 198, 203, 206, 219, 221, 330 maxim 8, 33 f., 39, 42, 65, 8688, 96 ff.,
Kersting, Wolfgang 330 131134, 137, 163 f., 226, 231, 237
Kim, Halla 288, 290 f., 296, 300 McCarthy, Michael H. 331, 335
King, J. Charles 330 McFarland, J. D. 70
kingdom of ends 262 McGee, Vann 72
Kirchmann, Julius Hermann v. 330 McGreal, Ian P. 331
Klamp, Gerhard 330 McLaughlin, Peter 70
Klein, Sherwin 330 McNair, Ted 198
Klotz, Christian 5, 21 means-end-formula 145153, 264
Khl, Harald 8, 21, 73, 98, 102, 112, mechanism 55, 275
115117, 330 Meiklejohn, J. M. D. 58
Konhardt, Klaus 330 Melches, Gilbert Carlos 331
Korsgaard, Christine 70, 90 f., 160, 174, Melnick, Arthur 332
196, 198, 201, 206, 210, 218, 221, 261, Mendelssohn, Moses 54
283, 330 metaphysics of morals 4 f., 16 f., 1921
Kraft, Bernd 325, 330 metaphysics of nature 5
Krausser, P. 331 method 17 f.
Kripke, Saul 10, 12, 21 Mieth, Corinna 196, 247
Kulenkampff, Jens 327 Mill, John S. 158 f., 198
Kuppermann, Joel J. 331 Miller, R. D. 332
Milz, Bernhard 18, 21, 332
Laberge, Pierre 123, 138, 331, 336 misology (contempt of rationality) 54
LaMettrie 51, 52, 53 modal status 5, 9 f., 13
Langthaler, Rudolf 70 moral criteria 167 f., 178 f., 190
Larmore, Charles 115, 117 moral law 31, 34, 37 ff., 41, 63 f., 66, 93
Latham, Noa 79, 92 116, 125, 130135, 188, 211, 213, 219,
lawfulness 78, 89, 95 f., 98 f., 102 f., 106 235, 237, 239, 242, 245, 247 f., 250,
108, 110, 294 254 f., 259, 263, 271, 276, 282, 286299,
laws of freedom 4 f., 10, 19 f., 267 301323
laws of nature 4, 10, 65, 124, 174, 182 ff., moral law, respect of 3336, 127 (see
195, 248, 255 ff., 261, 269, 273 f., 276 also duty, acting from)
281, 322 moral laws (pl.) 6, 12 ff., 165
General Index 341
moral motive (acting for the sake of the ONeill, Onora 164, 166, 172, 174, 180,
law) 38 f., 100103, 214 (see also 196, 198, 263, 283
duty, acting from) Oakley, Justin 72, 92
moral principle 83, 106, 114 ff., 161, Oberer, Hariolf 326328, 332, 334
168 f., 172, 180, 185, 190, 210, 217, obligation 13 f., 165, 171 f., 181, 187 f.,
225 ff. (see also supreme principle of 203, 225, 268 f., 293299, 303309, 319
morality; categorical imperative) ontoethical principle 311318
moral principle, justification of 225 organism 55 f., 58 f., 63, 65, 124 f.
243, 301323 ought 123, 130, 134137, 139 ff., 148 ff.,
moral principles 4, 10 f., 13, 15, 17, 43, 208, 213, 225229, 231, 265, 271,305 f.,
115, 122, 130, 167, 177 f. 311, 316, 321
moral question 307, 319 ought, conditional 226229, 235
moral worth 35 f., 7273, 7778, 8081, ought, unconditional 227231, 233, 235,
8788, 100, 113 238 f.
morality 3, 611, 16 f., 31, 47, 54, 60, 63
69, 83 f., 102, 112, 131 f., 158, 188, 219, Pasternack, Lawrence 327, 332
225243, 248 ff., 268, 285 f., 288 f., 291 Paton, Henry James 94, 117, 153, 157,
300, 305, 308311 (see also moral 195, 198, 331 f.
motive) Patt, Walter 332
Moritz, Manfred 332 Patzig, Gnther 115, 117, 143 f., 150,
Moritz, Michael 142, 157 157, 332
Morrisey, B. E. 336 Pelegrinis, Theodosius N. 332
motivation 35, 67, 79, 82, 123, 134, 173, perfectionism 66 f., 161
229, 234, 247, 256, 291, 306, 319321 permissible actions 8587, 184
Mrongovius, Christoph C. 141 petitio principii 289 f., 292, 295, 305
Mues, Albert 335 Petrus, Klaus 332
Mulholland, Leslie 198, 332 Philonenko, Alexis 333
Pistorius, Hermann Andreas 333
Nachtsheim, Stephan 335 Plato 5457
Nagl-Docekal, Herta 325 pleasure 49, 51 ff., 60, 67, 86, 128, 178,
Narveson, Jan 161, 198 226 f., 240
naturalism 53, 62, 274, 276 Pogge, Thomas 164, 169, 173, 195, 198
naturalists 53 postulate 68, 211, 240, 320
nature 5, 26 f., 28 f., 45, 47 ff., 51 f., 54 f., Potter, Nelson 168, 172, 196, 198, 333
5766, 69 f., 81, 124, 158, 194, 204, 212, practical deliberation 61, 252 ff., 260 ff.,
214, 232, 236, 240, 248, 261, 273 f., 275, 265 ff., 278 ff., 282
277280, 308, 322 practical principles 32, 42 f., 104, 123,
nature, gifts of 20, 27 f., 214 208
nature, human 14, 45 f., 53, 59, 66, 68, practical syllogism 127
171 f., 178 f., 190, 211, 318 Prauss, Gerold 260, 283, 326, 329 f.,
necessitation 34, 123, 134 f., 139 ff., 332 f.
144 f., 148153, 226, 231, 243 pre-determination 257, 273276, 279,
necessity 315, 34, 36 f., 61, 97 f., 101, 281
107, 114 f., 133, 135, 141, 145, 148, 152, prescriptive 4, 11, 13, 98 ff., 135, 299
162, 171, 188, 226 ff., 230 f., 236 f., principium diiudicationis/executionis
242 f., 255, 269 bonitatis 127
necessity, natural 61, 65, 68, 137, 233
237, 253, 258, 275 f. Quarfood, Marcel 285, 305, 323, 333
Nida-Rmelin, Julian 20, 21 Quine, W. V. O. 115, 117
Nolan, Christopher 173
nondiminishability thesis 2830, 42 radical evil 133
nonincreasability thesis 2830, 42 rational being 6, 10, 14 ff., 20, 32, 42, 61,
normativity 11, 13, 134, 187 63 f., 68, 80, 95, 104, 111, 113 f., 121 f.,
norms 225 ff. 124 ff., 134 f., 146, 149, 154, 170, 188 f.
342 General Index

191, 201 f., 204, 206 ff., 213, 229, 232 f., Schadow, Steffi 121
240 ff., 247255, 260264, 267, 278 Schaller, Walter 333
281, 285, 290, 292 ff., 302307, 315, Schiller, Friedrich 82, 92
318 Schneewind, Jerome B. 114117, 259,
rational nature 31, 40, 74, 183, 200 ff., 283, 334, 336
205 f., 210, 212, 217 f., 240, 262, 299 Schnoor, Christian 334
rationality 54, 67, 105, 110, 122, 127, Schndorf, Harald 334
129, 145, 148 f., 150, 156, 250, 257, Schnecker, Dieter 15, 18, 21 f., 45, 49 f.,
259 ff., 264268, 272, 276, 278, 280 f., 69, 71, 90, 94, 116 f., 121, 149, 155, 157,
293, 295 196, 241 f., 245247, 263 f., 273, 283,
Rawls, John 183, 195, 198 286, 288, 290, 292 f., 296, 300 f., 305,
reason, law of 41, 125, 315 307, 309, 318320, 323, 325, 330, 334
reason, practical 32, 39, 43, 46 ff., 61, 67, Schulz, Johann Heinrich 279, 309
121 ff., 126128, 149, 186, 233 f., 244, Schumacher, Ralph 326, 328, 333335
249, 251254, 260 f., 263 f., 271, 275 f., Schwaiger, Clemens 151, 157
278, 293, 304, 307, 309 f., 318 ff., 323 f. Schwan, Alexander 329
reason, pure practical 16 , 111, 122, 155, Scott, J. W. 334
227 f., 231 ff., 235 ff., 249, 294, 307, 320 Seebohm, Thomas M. 331
law of 111, 236 f. Seel, Gerhard 153, 326, 328
reason, theoretical 232, 266 f., 293, self, authentical 311
309 f. self-activity see spontaneity
Reath, Andrews 75, 92, 333 self-determination 229, 258, 264, 281
Reboul, Olivier 333 self-interest 36, 7378, 84, 89, 123, 171,
regressive argument 201 f., 206, 210, 319
212, 217 self-preservation 52
regulative principle 63, 65 Shell, Susan Meld 67, 71
Reich, Klaus 53, 70, 333 Sherline, Edward 334
Reiner, Hans 82, 92, 333, 335 Sherman, Nancy 83, 92
relevant descriptions 159 f., 163, 171, Sidgwick, Henry 72, 92, 167, 186, 198
177184 Siep, Ludwig 115, 117
representations 121, 124 ff., 135, 147, Singer, Marcus G. 180, 186, 196, 198,
208, 233 f., 258, 264, 275 f., 287, 292, 334
297 f. Skorpen, Erling 334
respect 34 ff., 75, 86, 93, 101, 104 f., 115, Smart, J. J. C. 181, 198
123, 126 f., 133, 140, 144, 148, 150, 159, Socrates 54 f.
165, 169, 182, 189, 249 ff., 261 ff., 278, Spitzley, Thomas 327
285 f., 290, 292 f., 295, 297, 300, 311, spontaneity 253, 256259, 265, 268,
318 ff. 272, 274278, 280 ff., 298299, 309
Rickless, Samuel 333 Staege, Roswitha 334
Riedel, Manfred 329 f., 333 standpoint, problem of 293, 261, 286,
Ritzel, Wolfgang 326 296300, 318
Robinson, Hoke 327, 330 Stattler, Benedikt 334
Rodriguez Lopez, Blanca 333 Steigleder, Klaus 227, 230, 238, 246,
Rollin, Bernard E. 333 260, 271, 283, 301, 318 f., 324, 334
Rosenthal, Jacob 247 Steinberger, Peter J. 335
Ross, Sir William D. 27, 44, 163 f., 176, Stocker, Michael 72, 92
180, 198, 333 Stoicism 52, 54, 66, 68
Rossvaer, Viggo 333 Stoics 52 f.
Rousseau, Jean-Jacques 53, 67, 259 Storheim, E. 335
Stratton-Lake, Philip 335
Sandermann, Edmund 333 Sturma, Dieter 259, 283, 325, 329
Sartiaux, F. 333 subject, logical 322
Scanlon, Thomas M. 187, 198 Sullivan, Roger J. 335
Scarano, Nico 10, 21 superiority of the ontic status 316 f.
General Index 343
supreme law 9799, 115, 133, 210 Wike, Victoria S. 336
supreme principle of morality 11, 16, Wilde, Leo Henri 336
18, 21, 31, 37, 41, 43, 73, 89, 93, 100, will 32, 35, 38 f., 41 ff., 47, 50, 65, 75, 78,
113 f., 172 f., 187 f., 200 ff., 204, 211 f., 80, 89 f., 95, 101, 103, 105 f., 121137,
225, 229, 301, 308 140 ff., 145 f., 148 ff., 153 f., 158, 178,
Swoyer, Chris 172 f., 195, 198, 335 188, 202, 204, 208 ff., 228 f., 232 f.,
sympathy 34 f., 8082, 181 235 ff., 241, 244, 247284, 293, 304,
306, 308, 319 (see also good will)
talents of the mind 26 f., 90 will as practical reason, problem of 32,
Teale, A. E. 335 39, 121123, 126 ff., 149, 235, 237, 244 f.,
teleological argument 45, 51, 278 252, 263, 271, 293, 310, 320
teleology 4569, 195 will, autonomy of 229, 232 f., 248, 250,
Tenbruck, Friedrich 335 259, 285, 291 f., 302, 304, 309, 311 f.,
Terzis, George N. 335 314 ff.,
theology 56, 57 will, intelligible 242, 299, 304, 316, 320
thing-in-itself 28, 256, 277, 297311, will, pure 17, 130, 304 f., 308, 311, 313,
316 f., 321323 316321
things-in-themselves 297 will, rational 31 f., 40, 149, 247, 250 f.,
though experiments 10, 13 f. 260 f., 263, 316
Thucydides 59 Willaschek, Marcus 124, 128, 133, 138,
Timmermann, Jens 15, 22, 123, 138, 336
269, 283, 335 Williams, Bernard 181, 198
Timmons, Mark 73, 166, 172, 177 f., Williams, T. C. 336
195, 198, 335 Willkr 146, 244, 270 ff., 319 (see also
transcendental idealism 6264, 286 f., free choice)
293, 296299 Willkr, freie 128, 233, 244 f., 270 ff.
Tropman, Elizabeth 90 Wimmer, Reiner 259, 271, 283, 336
Tugendhat, Ernst 5, 22, 94, 117, 335 wisdom 4, 40, 66
wish 27, 67, 122, 128, 136, 145148, 150,
unconditionality thesis 28 f., 37, 42 156, 231
universality 37, 95, 97 f., 100, 105, 107 Wolff, Christian 17, 297
110, 133, 158 ff., 187, 189, 191, 193, 196, Wolff, E. 336
236 f. Wolff, Robert Paul 336
universalizability 97, 158199 Wood, Allen 8, 15, 45, 49 f., 77 f., 82, 90,
92, 94, 103, 106, 107111, 113, 117, 132,
Vialatoux, J. 335 138, 149, 159, 164, 173 f., 176, 178, 198,
virtue 39 f., 68, 116, 174 201, 205 f., 210, 212, 218, 221, 289, 294,
Vliet, JoAnn van 225 302, 311, 320, 324, 334, 336
world of sense (sensible world) 256,
Wagner, Hans 335 280, 286, 298 f., 304309, 312322
Walsh, Dorothy 335 world of understanding (intelligible
Ward, Keith 335 world) 256, 280, 286, 298300,
Wartenberg, Thomas 335 304 ff., 309322
Weber, Raoul Simon 325 worth, absolute 136, 189 f., 202 ff.,
Weidemann, Hermann 68, 71 228 ff., 236 f., 320
Weischedel, Wilhelm 329
welfare 49, 52, 67, 183, 215 Yovel, Yirmiyahu 332
Wenzel, Uwe Justus 336
Westphal, Kenneth R. 336 Zanetti, Vronique 58, 70
Whiting, Jennifer 330 Zweig, Arnulf 329

Das könnte Ihnen auch gefallen