The Standpoint of Practical Reason

Citation
Korsgaard, Christine. 1981. The Standpoint of Practical Reason. Doctoral dissertation, Harvard
Graduate School of Arts and Sciences.

Permanent link
https://nrs.harvard.edu/URN-3:HUL.INSTREPOS:37366804

Terms of Use
This article was downloaded from Harvard University’s DASH repository, and is made available
under the terms and conditions applicable to Other Posted Material, as set forth at http://
nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAA

Share Your Story

The Harvard community has made this article openly available.
Please share how this access benefits you. Submit a story .

Accessibility
 THESTANDPOINT
OFPRACTICAL
REASON

A thesis presented

by

Christine Marion Korsgaard

to

The Departmentof Philosophy

in partial fulfillment of the requirements
for the degree of
Doctor of Philosophy
in the subject of
Philosophy
Harvard University
Cambridge,Massachusetts

August, 1981
 ACKNOWLEDGEMENTS

More people than I can possibly mention here have helped me with
their questions, comments,criticisms, and encouragementin the various
stages of this project's development. Myoriginal intention was to
write a thesis that covered the treatment of practical reason by
Aristotle as well as Kant, and a numberof people have read and helped
with parts of the thesis that was to have been as well as the one that
finally is. I would especially like to thank Hilary Putnamand
Amelie Rorty for commentson the Aristotle chapters. Someof the
ideas in Chapters Four and Five appeared in a much briefer form in a
paper entitled "Practical Reason and Rational Faith which was read
11

to several philosophy departments early in 1979. On those occasions

I received manyuseful commentswhich have contributed to the more ex-
tensive presentation of those ideas here. Almost everything in the
thesis has been tried out on the students in my seminars on Kant at
Yale in the Spring of 1980 and at the University of California at
Santa Barbara in the Winter of 1981, and I owe a great deal to the
questions and to the commentsand to the patience of those students.
WayneBuck at Yale and Susan Purviance at Santa Barbara wrote papers
for those courses which I have profited from reading.
In the early stages of writing Peter Hylton was a constant reader,
and set a standard for philosophical friendship, amongother things by
his extensive, useful, and immediatecomments. His criticisms and his

ii
encouragementhave been invaluable throughout. In the spring tenn of
myyear at Yale, Charlotte Brownand I spent long Friday afternoons
talking about Hume,Mill, Ross, and the foundations of ethics generally.
To those conversations I owemuchof what I say about these philosophers
in Chapter One and muchof the way in which I nowconceive that issue.
The Mrs. Giles Whiting Foundation provided me with a year of
support in 1978-1979for work on this thesis, for which I am deeply
grateful. The final draft was typed by June Kelley, Meredith Sedgwick,
and AnnWitkower, and proofread by Susan Purviance, and the vigilance
of all four has saved it from a great manyerrors. I would like to
thank myMother and Father, whose confidence and encouragementhas been
a source of support that I could always rely on.
Myadvisors, John Rawls and Martha Nussbaum,have contributed to
this project in very manyways. Martha Nussbaumhas given me extensive
and valuable written commentson large portions of what I have written.
I would especially like to express my thanks for her commentson the
teleology section of Chapter Two, which helped me greatly to clarify my
thinking about that topic. Of the very manyessential things that I
have learned about moral philosophy and about Kant from the teachings
and writings of John Rawls, there is one for which I am especially
grateful and from which I most aspire to benefit: an attitude, which
his work inspires, of respect for and a willingness to be instructed
by the tradition of moral philosophy.
Fromthe first beginnings of this project Timothy Gould has been
both its keenest critic and its most comprehendingsupporter. He has
taught me to believe that since philosophy is knowingwhat one is doing,
the presuppositions of a philosophical project are always as important

iii
as the assumptions behind a philosophical argument. For this, as well
as the careful reading and detailed criticism that every page of this
thesis has received from him, I would like to express mymost grateful
appreciation.

iv
 Note on Citations

Citations of Kant's works are given in the text, using the following
abbreviations.
G Foundations of the Metaphysics of Morals, Beck
translation.
Cl Critique of Pure Reason, KempSmith.
C2 Critique of Practical Reason, Beck.
C3 Critique of Judgment, Bernard.
DV Doctrine of Virtue, Gregor.
DJ Metaphysical Elements of Justice, Ladd.
MM General Introduction to the Metaphysics of Morals,
Gregor. Page numberscited are from Gregor's
Doctrine of Virtue.
R Religion Within the Limits of ReasonAlone, Greene
and Hudson. In this \'Klrk, the translators have used
11
will II to transl ate 11Wi11e 11 and have indicated in
the text where an English word or phrase such as
11
will 11 , 11choice 11, "power of choice", etc., is used
to translate 11WillkUr Accordingly, in quotations
11
•

from this book I have simply substituted 11Wi11e and

11

11
Willkur11 (omitting the umlaut} for the English
phrases as indicated.
ANTH Anthropology Froma Pragmatic Point of View, Gregor.
OH: Essay Title On History, edited by Beck.
OH:CBHH Conjuctural Beginning of HumanHistory 11 in On History.
11

SRTL 0n a SupposedRight to Lie FromAltruistic Motives",

11

Beck.
Theory and "On the ConmanSaying: 'This Maybe True in Theory,
Practice But it does not Apply in Practice, 111 Nisbet.
In citations of the Foundations, the Critique of Practical Reason, both
parts of the Metaphls,cs of Morals, the Anthropology, 0 on a Supposed
Right to Lie FromA truistic Motives," and the essays in On History, the
second page numbergiven is that of the Prussian Academyof Sciences
edition of Kant's works, as indicated by the translators. The Volume
references to the Academyedition are as follows:

V
 G IV
C2 V
MM VI
ANTH VII
OH VII-VIII
SRTL VIII
Morecomplete infonnat1on is supplied in the Bibliography.

vi
 TABLE
OFCONTENTS

Acknowledgements. . . . . . . . • • • . . . . • . . . . . . . . . . ii
Note on Citations ...........•.........•.... v
Chapter One: Introduction: The Objectivity of Ethics ........ 1
I. The Objectivity of Ethics. . . . • . . ........ 1
II. Attempts at a Theoretical Foundation .... 7
III. The Idea of a Practical Foundation. . ..•.... 16
IV. Kant on Practi ca1 Reason . . . . . . . . . . 23
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Chapter Two: Universal Law ............•........ 34

I. The Argumentfor the Formulaof Universal Law ....•.. 36
II. Interpreting the Formulaof Universal Law .......•. 39
III. The Theoretical Contradiction Interpretation .•...... 42
IV. The Terrible ConsequencesInterpretation .......... SO
V. The Teleological Contradiction Interpretation ....... 57
VI. Teleology in Kant's Moral Philosophy ........•... 65
VII. Practical Reasonand Instinct ......•......•. 77
VIII. Conclusion ......................... 84
Notes . . . . . . . . . . . . . . . . . . . . . . . . .. 86

Chapter Three: A World for Action ................. 91

I. Practical Contradictions in Universalized Maxims...... 91
II. TwoObjections to the Practical Contradiction Interpreta-
tion . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
II I. Practical Contradictions in the Wi11 . . . . . . . . . • . 112
Notes ............................ . 135

Chapter Four: The Leibnizian Interpretation ............ 138

I. Formal Implications of the Formulaof the Lawof Nature .. 141
II. The Leibnizian Interpretation ............... 150
III. Practical Reasonand Justice ............... 173
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

Chapter Five: Humanityas an End in Itself ..••........ 183

I. TwoDistinctions in Value ....•.....•...... 185
II. The Argumentfor the Formulaof the End in Itself ..... 189
III. Conferring Value ...•................. 204
Notes. . . . . . • . . . . . ..... 214
Chapter Six: Acting FromDuty .....•..•.......•.. 216
I. The Division of Duties ..........•.•..... 221
II. Acting FromDuty I: Obligatory Ends .•........• 231
II I. Acting FromDuty II: Juri di cal Action . • . . . . . . . • 242
IV. Moral Perfection and Happiness ••.•.....•.... 255
V. Humanityas an End: The Duty to Tell the Truth ...... 265
Notes . . . . • . . • . . . . . . . . . . . . . . . . . . . . . 276

vii
Chapter Seven: Conclusion: Autonomy....•••••..•.•• 278
Notes . . • . • . . . • . . . • . . • • • • . 302

Bibliography. . . . . . . . . . . . . . . . . . . . . . . 303

viii
... are we to say that absolutely and in truth the
good is the object of wish, but for each person the
apparent good; that that which is in truth an object
of wish is an object of wish to the good man, while
any chance thing may be so to the bad man,... since
the good man judges each class of things rightly, and
in each the truth appears to him? For each state of
character has its own ideas of the noble and the
pleasant, and perhaps the good man differs from
others most by seeing the truth in each class of things,
being as it were the norm and measure of them.
Aristotle

The noble type of man experiences itself as determining

values; it does not need approval; 1t judges, 'what
is harmful to me is harmful in itself'; it knows itself
to be that which first accords honor to things; it is
value-creating.
Nietzsche

And what is it that justifies the morally good

disposition or virtue in making such lofty claims? It
is nothing less than the participation it affords the
rational being in giving universal laws. He is thus
fitted to be a memberin a possible realm of ends to
which his own nature already destined him. For, as an
end in himself, he is destined to be legislative in the
realm of ends, free from all laws of nature and obedient
only to those which he himself gives .... Autonomyis thus
the basis of the dignity of both humannature and every
rational nature.
Kant
 Chapter l
Introduction: The Objectivity of Ethics

This thesis is an essay into the most fundamental issue of moral

philosophy, which might be called the problem of the objectivity or the
foundation of ethics. Mypurpose is to elucidate and to defend a type
of solution to this problem which is found in the ethical works of both
Aristotle and Kant. The solution is that ethics is founded on practical
reason, and it is Kant's version of the solution that I present. In
this chapter, I try to explain what the problem of objectivity in ethics
is and why it presents itself as a problem. I sketch out the fonn that
treatments of it usually take - a form dictated by the way in which the
problem presents itself - and indicate what I take to be the basic dif-
ficulty with those treatments. I then explain why I believe that a
foundation in practical reason is free of that difficulty, and there-
fore the most promising line to pursue. Finally, I describe the line
of argument which I will follow in my defense of the Kantian version of
ethics founded on practical reason.

I. The Objectivity of Ethics

The problem of the objectivity of ethics can be and has been posed
in a great manydifferent ways. It has been envisioned as the search
for criteria, perhaps factual ones, in tenns of which we can assess or
verify judgments that employethical concepts as true or false. It has
taken the form of the question whether words like good11 and 11ri ght 11 can
11
 2

be defined in tenns that do not invoke the subjective preferences of the

speaker or the cultural norms of his or her social group. It has been
conceived as the metaphysical project of locating special objects or
properties of objects to which ethical concepts apply; or as the more
psychological project of unveiling the universal features of humannature
that account for our obligation to behave morally or our propensity to do
so. The specific motives for asking such questions or undertaking such
projects are of various kinds, but the questions and projects are all
concerned with settling doubts about whether morality can be reconciled
with other things that we knowor believe or wish to believe or with
other standards that we hold. They express a worry about reconciling
moral motivation with the insistent demandsof self-interest; or a doubt
about whether the metaphysics of ethics can find a place in a physicalis-
tic or detenninistic universe; or an uncertainty whether the language of
ethics can meet our standards of meaningful use. Yet these rather dif-
ferent sources of philosophical concern all spring from a corranonroot,
at least where they are genuinely sources of concern and not mere excuses
for the glib dismissal of the difficult problems and more difficult
demandsof morality from the "objective" realm. I want to start by
describing what I take the commonroot of these various "reconciliation"
worries to be. For if there is a corranon
source from which all these
questions spring, then a solution which answers any one of them in a way
that really reaches this source should satisfy those who ask the others
as well. Such a solution would meet all of the various needs that are
expressed by these questions, although not necessarily by an answer in
their own terms. It would somehowteach us how to use the concepts of
ethics or show us why we use them as we do; help us assess the correct-
ness of utterances that employ them; license their use in advice and
 3

imperative; illuminate our motive for caring about, and demandingthat

others care about, the things and actions to which we apply them.
The concepts of ethics - good, right, virtue, and so on - play a
practical and urgent role in our lives: practical, because they are
the standards that guide our choices and decisions; and urgent, because
we use them to think about the things we believe are worth living for,
and sometimes even worth dying for. It is natural to wonderwhat pos-
sible account of these concepts or their objects could makesense of,
or justify, the role that they play in our lives. Here we find the root
of objectivity problems. The question may take an epistemological form:
a demandthat it be shownthat we all have access, the same access, to
the truths of ethics. It may take a metaphysical form, as a requirement
that the truths of morality be given someontological foundation, in the
"nature of things. 11
It may take a psychological form in inquiries into
the motive for being moral, the search for an intelligible and convinc-
ing reason why people makethemselves liable to the demandingand some-
times dangerous conditions which morality imposes. But in fact, so-
called non-objectivist views can give answers of some sort to all such
questions. The real threat of such views - subjectivism, the various
forms of relativism, emotivism - lies not in any failure to provide
morality with some sort of foundation, but rather in their failure to
provide it with a foundation which can support the full weight of our
11 11

commitmentto the moral life. Subjectivism, for example, might at its

best be taken to be the view that we are capable of sustaining our moral
outlook without the assistance of reason or the concurrence of others.
In that case the refutation of it would be that we are not.
It is not really our theoretical needs at all but our practical and
moral ones that make us unsatisfied with non-objectivist views. It is
 4
not that our questions about metaphysics or language or psychology go
unansweredby these views, but rather that the answers provided, to
various extents, undermine our interest in and our commitmentto ethical
considerations. Most of the phenomenaassociated with morality probably
can be explained, in a theoretically satisfactory way, in terms of the
expedience of keeping society in good order and the usefulness of being
able to predict one another's behavior and to depend upon contracts and
promises; in terms of the historical accident 11 of generations of care-
11

fully indoctrinated customs, preferences, and prejudices; or even in

terms of a genetic urge to do what promotes or at some stage of our bio-
logical history promoted the survival of our species. One could express
the point somewhatfancifully by saying that an amoral but theoretically
rational creature observing us would most likely be perfectly satisfied
with some pastiche of such explanations. So the dissatisfaction is
rather a practical one, resulting from a sense that this kind of explana-
tion would vitiate the ethical interests of those who came to understand
and believe them. Wedo not want just any explanation, but a justifying
one.
This does not mean that those who have therefore gone on searching
for an objective foundation for morality can be accused of an irrational
partisanship or a stubborn refusal to follow perfectly good explanations
where they lead. The point is rather that the search for objectivity in
something other than an historical or empirically-psychological or bio-
logical explanation will be undertaken only from a certain standpoint,
and only by a philosopher who in a certain way takes his or her own
conmitmentas a fact that is special in that it needs not merely a
theoretically adequate explanation but a practically adequate one as
well - an explanation that will not merely account for the existence
 5

and substance of that commitmentbut will justify it and keep it alive.

Such a philosopher refuses to stand wholly outside of her or his own
ethical interests and view them on the same level with other psychologi-
cal phenomenain need of an explanation. It is precisely because the
demandfor objectivity in ethics is made from such a standpoint that it
is a philosophical rather than an historical or psychological or biologi-
cal issue.
This, I would like to say, is an approximation to a universal human
standpoint from which objectivity questions in various forms continually
arise. The discomfort which they all express comes from a sense that
the accounts of morality satisfactory from a theoretical standpoint,
from 11outside 11 of the moral humanbeing, are not satisfactory from the
practical standpoint created by one's own ethical interests. Perfectly
satisfactory theoretical accounts undennine our practical commitments;
those in turn make demandson theory that it cannot sustain. In a re-
flective person, who can take up either of these standpoints with res-
pect to herself, the clash becomesan internal one. The theoretical and
practically-moral perspectives seem to be at a standoff, neither satis-
fying the demandsof the other.
As a result, the quest for objectivity usually takes either of two
forms. Either we start with an explanation, and then see whether it has
the power to justify; or we start with a justification, and see whether
it has the power to explain. In the first case, a theoretical question
is asked from the practical standpoint, that is, with the standards set
by that perspective in mind: and then we look for objects or properties
corresponding to ethical concepts, or for objective sources of ethical
knowledge, which will provide a theoretical or external account of moral
phenomenathat somehowis also sufficient to sustain the practical point
 6

of view. In the second case, the practical question is posed with the
standards of theoretical intelligibility in mind in the form of the
question 11Whyshould I be moral?11 , where that is a request for some
reason outside of morality, and accessible from the standpoint of theory,
for entering into the moral realm. This is generally understood as a
question about the availability of self-interested reasons for being
moral, and this is precisely because we think of reasons springing from
self-interest as perspicuous from the standpoint of theoretical reason.
This in turn is because we think of self-interest as a well-known psycho-
logical feature of humanbeings, and because we are tempted to think of
self-interested reasoning as causal reasoning with respect to a given
end. (This sense of the theoretical pespicuousness of self-interested
practical reasons may itself be a sort of illusion and can certainly be
challenged - but that is not to the purpose here. Illusory or not, the
sense exists and has a great deal to do with the way the objectivity
problem presents itself to us.) It will perhaps seem paradoxical to say
that the search for a metaphysical or epistemological basis for ethics
is undertaken from a practical point of view and the search for a motive
for being moral is undertaken from a theoretical point of view. But the
idea is that the metaphysical and epistemological project is conjoined
with the demandthat the answer be one appropriate to practical ethical
commitment,and the motivational project is conjoined with the demand
that the motive be one intelligible from a theoretically-psychological
point of view. Either way, the aim is to produce a better 11
fit 11
between
these two perspectives from which we can view our own ethical interests
and actions.
In the next section I try to explain why the attempt to support
ethics by producing a 11fit 11
between the theoretical and practical
 7
perspectives cannot work. The point of founding ethics on practical
reason is to prevent this need from arising: ethical principles do not
require any sort of theoretical support, because they have their own
justification in an employmentof reason that is purely practical. The
standpoints of theory and of practice are hannonized in quite a different
way: not by finding an account of ethics that is at once theoretically
creditable and practically sustaining, but by showing that the principles
that govern theoretical understanding and the principles that govern
practical choice and decision share a commonground in reason. Ethics
does not require a support perspicuous from the standpoint of theoreti-
cal reason, for the standpoint of practical reason is grounded directly
in reason itself.

II. Attempts at a Theoretical Foundation

Whenour sense of the mismatch between the theoretical and practical
perspectives, as two ways we can view our ethical actions and interests,
takes the form of a demandupon theoretical reason to provide us with an
explanation of morality which will sustain those interests, then the
objectivist effort will be metaphysical or epistemological in character.
It is supposed that ethical interest could be supported in the way
commitmentto beliefs or scientific theories is supported if ethical
values or ethical judgments were provided somehowwith the same sort of
rootedness-in-the-nature-of-things that we suppose facts or physical
objects to have. Ethics is supported by making the good or the right
an object of knowledge. A wide range of ethical philosophies, especi-
ally in this century, can be described without too much distortion as
following this path; some with a metaphysical, some with an epistemo-
logical emphasis. G. E. Moore is perhaps the boldest in simply adding
 8

goodness into his ontology, as a non-natural property; he also provides

us with a faculty of intuition with which to discern it. 1 W. 0. Ross
gives us a partial way of knowingwhat is right, by making truths about
what is right 11self-evident 11 • 2 Ross• approach seems to spare us the
need for ontological additions, objects or properties for ethical judg-
ments to be about, at any rate for those philosophers whose general
practice is to spare themselves objects for self-evident truths to be
about. In moral sense theory we might say that emotional responses are
used for detecting what is right or good. Even in empiricist ethics,
where the good is thought to be some plain and obvious thing like plea-
sure or happiness or the satisfaction of desire, perhaps generalized
over the largest possible group of people or taken in the largest pos-
sible degree, the good is treated primarily as an object of knowledge,
and the emphasis in argument is on showing the reader either that this
thing is in fact The Good (in fact what we mean by 11good11 , the argument
may try to run) or that what is good in a given case can be readily
calculated or determined. 3
Classical utilitarianism provides the best examples of this. Mill's
proof of the principle of utility is an attempt to prove and to defend
the claim that the general happiness is good by reference to psychologi-
cal fact - to what people actually do desire. 4 Bentham's calculus is
meant to show that it is in principle possible to ascertain, by normal
epistemic procedures, what action will do the most good. Perhaps the
best example is Mill 1 s argument about the qualities of pleasures. Mill
defends the existence of qualitative differences amongpleasures by
showing us that differences of quality can be ascertained by the normal
epistemic procedures of empiricism. Indeed, he points out, the pro-
cedure used to ascertain difference of quality (judgment of the
 9
experienced) is the very same one we must use to ascertain difference of
quantity. 5 The emphasis throughout is on the applicability of normal
epistemic procedures: this is what the classical utilitarians felt was
11
necessary to make ethics scientific.
11

In all of these various ways, the good or the right is made an ob-
ject of knowledge - of theoretical reason - and having established that
the good or the right, once properly understood, can be ascertained by
some normal epistemic procedure or - as in intuitionism - by something
sufficiently analogous to a normal epistemic procedure to make us feel
at ease, the philosopher feels that theory has done its job in providing
a theoretical account of ethics that will sustain our commitmentto
ethical practice. Of course, we may have added something to our ontology
or perhaps a new faculty of discernment to our theory of epistemological
psychology; and our willingness to do this must after all hark back to
our sense of what would be necessary to support our practical commit-
ments. And then one might begin to wonder why, if our practical conunit-
ments can sustain demandsmade upon our ontological or epistemological
theories, they seem not to be able to sustain themselves. I will come
back to this point. The issue now is whether making the good or the
right an object of knowledge has given the ethical life the support that
we sought for it at all.
Having established that the distinction between good and evil is
well-grounded in the nature of things, have we thereby established a
reason or a motive for giving our allegiance to the good? Thoughwe may
be asked in the name of that allegiance to sacrifice our heart's desire,
our whole happiness, even our lives? Is it to be automatically supposed
that no matter what goodness or rightness turn out to be, they are wor-
thy of such a commitmentso long as they can be identified objectively?
 10

Where the good is identified with something as mysterious as Moore's

non-natural properties appear in certain moods to be, the answer to this
question is not at all clear. Whyafter all should we govern our lives,
or anywayrestrict our private interests, for the sake of the promotion
or maximization of things possessing a non-natural property about the
character of which we are able to say very little? Is it really enough
to say that after all this is the good we are talking about, not just
any old property? If it is a natural state or object, happiness or
utility or pleasure or satisfaction, suitably maximized or generalized
over groups of people, we might feel somewhateasier about the intelligi-
bility of a commitmentto it. But even if the best thing in the world
were, say, happiness, or the only good thing in the world were happiness,
and you were asked to sacrifice your own in the name of someoneelse's
or indeed of everyone else's, would you have to be some sort of moral
monster to feel puzzled as to why, after all, you should do so?
Of course, not all philosophers cower before Occam's Razor, and just
as one can add to one's ontology to solve metaphysical problems or add
on extra faculties, intuitions and moral senses, to cover the epistemo-
logical angle, we can add on extra motives to take care of this diffi-
culty. Ross, for example, supposes that we act from what he calls the
sense of right, and suggests that
... it may be maintained that there is no more mystery
in the fact that the thought of an act as one's duty should
arouse an impulse to do it, than in the fact that the
thought of an act as pleasant, or as leading to pleasure,
should arouse an impulse to do it.6
It may indeed, from a sufficiently theoretical standpoint. What counts
as a mystery depends largely on one's perspective, and on what seems
obvious from that perspective. Ross' perspective here is, to put it
fancifully, that of some extraterrestrial and scientifically-minded
 11

creature, recording observations of its first specimen of the earth

creature homosapiens: it is movedto action chiefly by two kinds of
impulses, which it calls expectation of pleasure and awareness of duty.
This is a perspective from which it would surely be more proper to main-
tain that there is no less of a mystery about pleasure than about duty,
one from which they are both as yet unaccounted for.
The unhelpfulness of Ross' remark illustrates my point that our
demandfor objectivity in ethics is born in a practical rather than a
purely theoretical context. The reason that his remark is so unsatis-
fying is that it is made from the wrong perspective - from there, I want
to say, you cannot see what the person who, faced with a difficult moral
decision, asks, "why should I do this?" is asking to be shown. Ross is
not exactly wrong; it is simply not the dissolution of a mystery of
theoretical anthropology that we are after. All of this tends to show,
I think, that when theoretical reason responds to the demandsof morality
by providing us with an object of knowledge for ethics, it does not
thereby automatically answer the question which we might say that it, in
turn, addresses to practice: then why should I be moral? It is impor-
tant to keep in mind that the answer to this must be satisfactory from
the practical perspective as well, because as Ross' move shows, it is
not hard to take up a theoretical viewpoint from which the problem fails
to arise, or anywayfails to exhibit its difference from other motiva-
tional mysteries.
This question, Whyshould I be moral?, has itself been the subject
of great controversy, since it is so unclear just what sort of answer
might be required. The first possibility always considered is that it
is a demandto be shownwhy it is in my interest to be moral. I have
already mentioned the reason for this. Of course it has rightly been
 12
pointed out by just about everyone that where "my interest" is constitu-
ted by my present and prospective needs and desires, and the demandis
to be shown that the satisfaction or promotion or protection of those is
an object to which mora1i ty is ins trumenta1, the demandcannot be met and
would be irrelevant to genuinely ethical behavior if it could. Whatever
morality is, it is something different from the intelligent pursuit of
my own interests whatever they may be; and however much it promotes my
interest to follow the precepts of morality, I am not being moral or
noble when I follow those precepts only under that condition or for that
reason. So much is trite. This very standard response to the question
why one should be moral does not however silence it, and should not.
The question need not be made from the standpoint of self-interest, and
a request to be shown the point of the ethical life can be perfectly
legitimate. If we are asked to give moral considerations the precedence
in every decision, to sacrifice to them our dearest wishes, to restrict
our interests in accord with them, then surely we have a right to some
account of the point - for ourselves or for humanity - of committing
ourselves to a policy or to a way of life that will require such things
of us. To seek for such a point is not to show that one is heartless or
selfish or that one has altogether misunderstood the nature of morality. 7
For ethics must after all have some sort of a point; there must be some
way in which its criteria for action and choice and living generally
make our lives better or richer or at the very least more tolerable. 8
It would be insane to adopt a set of standards and restrictions which
had no effect or meaning but an addition to the burdens of our already
too often cramped and narrow lives. This is why the question why one
should be moral keeps resurfacing even in the face of the most plausible
theoretical accounts of the metaphysics or epistemology of ethics. It
 13

is not enough to make the good an object of knowledge. Wemust also

make it something whose point is transparent, something that is clearly
an object of value, something that we can intelligibly value. It must
be justified.
This is a main source of the appeal of utilitarianism; for the ob-
vious temptation is to give the good its point by constructing it out of
the objects of natural desire. An appeal to natural and psychologically
familiar motives such as benevolence or sympathy can then be made. Con-
temporary utilitarians, such as J. J.C. Smart, may appeal directly to
something like benevolence. 9 Although Mill's account of the moral sanc-
tion is more subtle than that, he too emphasizes as a point in utili-
tarianism's favor that it can appeal to the natural desires. Mill's
view is that any moral system can be inculcated so that it becomes part
of the motives and feelings of a child, but that without "a natural basis
of sentiment" the adult's loyalty to the system may not survive "intel-
lectual culture" and "the dissolving force of analysis." Utilitarianism,
however, finds a "basis of powerful natural sentiment" in "the social
feelings of mankind", and so the utilitarian conscience not only survives
examination and the discovery that it is acquired but indeed is likely to
be strengthened by both individual and historical development.10
The difficulty now is that the relationship between the goodness of
the good, its point, and our motivation to promote it is too contingent
(in the case of the so-called sense of right, almost coincidental). The
recognition of something as good or right does not of itself provide the
motive; the motive must be supplied. But this leaves a gap which cannot
be filled. For if the reason why I ought to be moral is provided by my
sympathy or my benevolence or my fairness, then there is no reason why I
ought to be sympathetic or benevolent or fair; these are not part of the
 14

right or the good but merely desires aimed in their direction. If I do

not have these motives I amnot an evil but merely an amoral creature.
Wecannot say that I ought to have these motives, for why ought I? Be-
cause they are right or good? These were the motives in terms of which
my allegiance to these values was supposed to be explained in the first
11
place. This is an argument familiar from Kant's Foundations. It is
also the basis of Nagel's argument for the raional version of 11internal-
ism11in The Possibility of Altruism. 12 The point of it might be taken
to be this: the reason that appeals to self-interest fail as supports
of morality is not just that self-interest is a morally inappropriate
motive. Benevolence, sympathy, and loyalty to the social order fail in
the same way. Our motives, seen as the objects of empirical psychology,
have no power to justify the actions to which they give rise.
The attempt to provide ethics with an objective foundation from the
resources of theoretical reason necessarily fails, because knowledgedoes
not motivate or automatically justify motivation, and empirical psycho-
logy, being irreducibly contingent, has no resources of justification.
The attempt to put together epistemological and psychological supports
of objectivity within the confines of a single ethical philosophy cannot
make a connection that is tight enough to fulfill the real purpose of
objectivist ethics. The knowledgedoes not make it necessary to have
the desire; and the desire cannot make it imperative to act on the know-
ledge. Wecannot get a theory of objective value by getting metaphysics
to provide us with the element of objectivity and psychology to provide
us with the element of valuability; for no matter how it is done, the
seam shows.
I have described the problem of the objectivity of ethics as arising
from a mismatch between, on the one hand, our practical sense of our
 15

ethical interests and commitments, of their seriousness and worth, and,

on the other hand, the accounts of these interests and commitmentswhich
can be given from an external and theoretical perspective which we can
always take up in thinking about those interests and commitments. The
mismatch arises because what we can say from a theoretical perspective
about our values and our relation to them cannot capture or explain the
urgency that underlies ethical choice and action. The quest for objecti-
vity becomes a search for an account made from a theoretical perspective
that could sustain our ethical commitment. This gives rise to the two
projects I have described: the attempt to show that the right and the
good are objects of theoretical knowledge, and the attempt to show that
our concern for them can be explained in the terms of theoretical natural
psychology.
But if the point of these accounts was to explain ethics and make
its standards objective in a way that would sustain ethical commitment,
they cannot do that job. The very best that a theoretical account can
do is to be sufficiently attractive not to vitiate commitment.13 This
is why the humanity and commitmentto liberty that characterize Mill's
version of utilitarianism make his such a forceful presentation of that
view. Though the arguments cannot be proofs, as the utilitarians them-
selves insisted, the system is morally attractive. 14 But it is one thing
for an account of ethics merely to avoid leaving us cynical or disheart-
end about our ethical commitmentsand another for it actually to enable
us to explain and justify those commitmentsto ourselves. This is some-
thing an epistemological and psychological approach to ethics simply
cannot do.
If pure reason is practical, however, these projects are as unneces-
sary as they are hopeless. For they are based on the notion that
 16

objectivity must be sought in the realm of theory, and that ethics must
therefore be supported from outside. But the notion that objectivity
belongs especially to the realm of theory is in turn based on the idea
that only the realm of theory is governed and regulated by the rational
standards that make objectivity both possible and important to us. If
pure reason can be practical, that is just false: there will be objec-
tive standards that apply directly to, and can govern, action and choice.
A foundation in practical reason obviates the need to take up a theoreti-
cal perspective with respect to our ownethical interests and moral
commitmentsin order to justify or understand or explain them. The
account of ethics remains resolutely within the realm of the practical.
Once it is shown how practical reason generates its own standards and
its own object (the good), we need not go rummagingaround in the realms
of theory in search of objects or properties or motives to support ethi-
cal concepts. Weneed not provide any theoretical proofs that something
is right or good: it is rather to practical reason that we appeal to
settle such questions. Practical reason is put forth as a separate
source of reason and "argument" and justification, whose special domain
is ethics itself. There is and needs to be no extraneous motive for
acknowledging this source of reasons; it has the same power to justify
that theoretical reason does to explain. Morality or Ethics is reason
brought directly to life in humanaction, and it can be accounted for
only in its own terms.

III. The Idea of a Practical Foundation

The idea of practical reason is to supply us with a practical
foundation for ethics. It is possible that, because of the way in which
objectivity problems arise, the enterprise of producing a foundation in
 17

practical reason seems at first to be a sort of second choice, a thing

to try when the theoretical foundation has failed. After all, one may
protest, if the problem of objectivity rests in a mismatch between our
practical sense of our ethical commitmentsand a theoretical view we can
take of them, how does a solution that remains, as it were, wholly within
the realm of the practical help?
The answer to that worry comes in three stages. The first is the
reminder that we took up a theoretical position precisely for the sake
of finding a basis or foundation for our moral views and feelings. If
we can do this from another point of view, this other point of view is a
suitable replacement for the theoretical foundation - we need never have
made the move to the theoretical view in the first place. If this still
makes the idea of a practical foundation sound like a second-best substi-
tution, the next stage in the answer is to correct that impression. When
we see what the practical foundation for ethics is, we will see that the
attempts at a theoretical foundation were implicitly dependent upon it
all along. For we will see that whatever theoretical claims we were and
are willing to make for the sake of grounding ethical truth or motiva-
tion, the basis of our confidence for making those claims lies in practi-
cal reason. If a philosopher feels prepared to say that goodness must
be a real property of all good things, or that there must be some sort
of moral facts, or that we must have the motivational capacities that
account for action from duty somewherein our psychological apparatus
... it will always turn out that the necessity expressed in these
musts11 does not spring from the need to explain a theoretically
11
well-
knownfact that cannot be otherwise accounted for, but from a conviction
which in turn draws its sustenance from practical reason. Our very will-
ingness to make revisions in our theoretical views for the sake of ethics
 18

shows how little necessary it is to do so: if these commitmentscan

sustain metaphysical or psychological beliefs, then they can sustain
themselves.
Still, these remarks may leave it sounding as if the practical
realm, though firmly grounded of itself, remains opaque to theoretical
reason, and it might be supposed that the gap between theoretical ex-
planation and ethical commitmentthat gave rise to the problem of objec-
tivity still remains: nowwhat we have looks like a division in our
mental life, two separate realms, one set of objective standards for
knowledgeand explanation, another for choice and action, with no connec-
tion between them. Wecan tell those who insist that we must either have
a theoretical foundation or hold our ethical convictions unfounded that
they are mistaken, but that is all. The third stage is to show that
this is not so - that the gap between theory and practice is bridged,
though in a manner that is different from the one envisioned by those
who attempt a theoretical foundation. It is bridged by the fact that
the standards that govern the realm of the practical are standards of
reason. It is reason in its practical employment, to be sure, but never-
theless reason, the same reason that governs our quest for understanding
and explanation in the realm of theory.
The success of the enterprise of founding ethics on practical reason
depends on the possibility of bridging the gap between the theoretical
and practical realms in this alternative way. That is to say, it re-
quires showing that the standards that, according to a practical-reason
view, are supposed to govern choice and action, are indeed standards of
reason - of something that is recognizably reason. It is on this issue
that this thesis is focussed.
After all, it is hardly news that Kant believed himself to have
 19

produced a view of the sort that I have been describing: that he sought
for, and thought he had found, a principle of pure practical reason that
could govern action and choice. And if his principle is one of pure
practical reason and can in a substantive way govern action and choice,
then he has solved the problem of objectivity. If Kant's view - or at
least, Kant's approach to the problem of objectivity - has not found
complete acceptance amonghis readers, I believe that it is for two
reasons. One is that they have failed to see how this principle could
really give substantive direction to choice and action, 15 and the other
is that they have failed to see what the claim that it is a principle of
reason really comes to. 16 These complaints have been related in the fol-
lowing way: it is thought that if it is a principle of reason it must
be in some sense formal and if it is "merely" formal it cannot possibly
guide choice and action. Kant addresses both of these issues in his
ethical writings, and addresses them in a way that is both substantial
and direct. Certainly, it has not always been clear to his readers when
he is doing so, but these blindnesses in the critical reception of Kant
are not to be attributed entirely either to his difficult methods of
presentation or to his critics' impoverished reading habits. The prob-
lem in the reception of Kant's ethics is a deeper one, lying in our at-
tachment to a view of reason in general that we have inherited from the
empiricist tradition. On this view reason cannot be practical - the
idea just makes no sense, leaves us blank. Kant's view cannot be compre-
hensible, much less convincing, unless the spell of this inherited view
of reason - not the sense of its truth, but the sense of its inevit-
ability - can be broken. This is why anyone who now attempts to present
Kant's arguments for the rational status of the categorical imperative
must directly address this empiricist challenge.
 20

This challenge finds its classic statement in Hume's Treatise of

HumanNature, in the section entitled 11
Of the Influencing Motives of the
Will. 1117 There Humeclaimed that practical reason can be nothing more
than theoretical reason as we use it in the service of action - that is,
in the service of our passions, needs, and desires.
Reason is, and ought only to be, the slave of the passions,
and can never ~retend to any other office than to serve
and obey them.18
Reason does not, on his view, dictate to us our ultimate ends, although
it may dictate to us the adoption of somemediate ends for the sake of
other ends, as instrumental to those others or as more particular reali-
zations of them. Our ultimate ends are all given by needs and desires,
either those imbeddedin humannature or those we simply happen to have
at a given moment. What is called practical reason is simply reasoning
about how to bring about the desired end or perhaps maximal fulfillment
with respect to some complicated system of needs, desires, and prospec-
tive goals. The actual reasoning itself is ''theoretical" in character;
it is for the most part, as Humeinsisted, reasoning about causes and
effects, and is no different in method or principle than what is found
in other instances of theoretical reasoning.
Hume's challenge cannot be read as an argument against the existence
of pure practical reason, for such argument as he gives depends directly
on the assumption that reason has only theoretical functions - mathemati-
cal and logical relations, and the discernment of causal relations. Hume
simply invites us to contemplate the impossibility that either of these
should ever give rise to an action: reason, we are told, is concerned
with truth, and truths do not motivate us in the absence of desires
whose satisfaction or frustration might be influenced by these truths.
As for the desires themselves, they cannot be true or false, as they are
 21

"original existences" rather than copies of anything. From this Hume

draws two conclusions: reason can never dictate an action, since know-
ledge never moves us unless it touches upon something we care for; and
reason can never by itself forbid an action, for passions, not being
true or false, can only be opposed by other passions. It is not sur-
prising that Humemakes this argument in such short order, for to say
that reason cannot really be practical if its only functions are theore-
tical is easy. The real argument behind Hume's glib treatment of the
role of reason in action lies in his view of what reason is - a faculty
concerned only with the relation amongideas. This view limits reason
to (at most) theoretical functions by its very nature. For us, the im-
portance of the Humeanchallenge is that despite the lack of argument in
his treatment of the influencing motives of the will, Hume's challenge
has been found completely convincing by manyphilosophers. Our view of
the nature of reason has been sufficiently similar to Hume's to make it
seem to us, as it did to him, rather obvious that reason could not
govern action, but could only show us how to satisfy already existing
desires. Or rather, our views of reason have had the same effect of
making the idea of an objective rational standard for action a blank one
for us. Hume's formulation has set the problem which the proponent of
a practical reason based ethics must answer. Howcan reason be practi-
cal, unless it can direct us to ends? And howcould it ever direct us
to ends?
Wherea passion is neither founded on false suppositions,
nor chooses means insufficient for the end, the under-
standing can neither justify nor condemnit. It is not
contrary to reason to prefer the destruction of the whole
world to prevent the least uneasiness of an Indian, or
person wholly unknownto me. It is as little contrary
to reason to prefer even my ownacknowledgedlesser good
to my greater, and have a more ardent affection for the
former than the latter.19
 22

This comes out clearly if we think of an objective rational standard for

action as one in terms of which an action would be fully justified. When
we ask someoneto justify an action, to say why she is doing something,
to give us a reason - not a mere cause - there seems to be a definite
limit on what she can say. If she is doing it as a means to or an in-
stance of some other end or for the sake of some other end, she can tell
us that. Wecan push forward, of course, and demanda reason for that.
But eventually it looks as if we will reach a point where she has nothing
more to say about why she does or wants something. Perhaps it is just
the thing she wants most in all the world, or something she needs to
live, or something she needs just because she is living. (For living
cannot very naturally be described as an end, and when we do things
simply because we are living it is certainly not always because we are
in the grip of an activated survival instinct.) Eventually (this thought
goes) there must be things you just choose, allowing yourself - as if
you had any choice - to be governed by instinct and passion and neces-
sity. No reason can be given for these things.
So it seems to be here, at the level of the choice of one's ulti-
mate ends, that the question of practical reason arises. For if the
choice of these ultimate ends is in some way open to rational assess-
ment, then pure reason can indeed be practical, but if not, then - as
Humethought - it cannot. On the other hand, if there are rational
principles of choice, or rational ways of assessing principles of choice;
if there are rational standards for the goodness of ends and the right-
ness of actions, requiring some, and rejecting others; if there are
rational motives or rational ways of being motivated - if rational cri-
teria can be made to apply even to the ultimate ends and first princi-
ples from which we act, only then can pure reason be practical. If this
 23

is so, actions are open to complete justification, all the way down.
They are not rooted in arbitrary features of our psychology that we can
only hope we will think well of. Objective rational standards apply to
what we do and what we take as our ultimate aims and ends. It is to
these standards that we would look for an account of important concepts
of ethics.
What the Hwneanchallenge makes so clear is this: the task of show-
ing that pure reason can be practical is the task of showing that reason
by itself can govern actions and set ends. And to accomplish this task,
what is needed will be nothing short of a completely worked out idea of
what reason, in both its theoretical and practical employments, is and
does. A practical foundation for ethics depends upon a general theory
of reason. This is what we find in Kant.

IV. Kant on Practical Reason

Myproject in this thesis is to present and explain the argwnents
by which Kant establishes that the categorical imperative is a principle
of reason, and that the sort of action and choice which it governs is a
quintessentially rational kind of action and choice. The result is a
general exposition of Kant's moral philosophy, drawing on all of the
major ethical writings of the critical period, but so organized as to
make the structure of the various arguments for the rational status of
the moral law stand out. At this point I will try only to outline the
general strategy of the argument that is used.
It is obvious that it is no easy or straightforward matter to
"prove" that a given principle is in fact a fundamental principle of
reason. For if there is doubt or disagreement about the nature and sub-
stance of rational principles themselves, in terms of what is a proof
 24

to proceed? Views about what reason is and does are of the most funda-
mental kind in philosophy: they are the souls of philosophical eras,
traditions, and movements, endowing them with such individual identity
or separateness as they admit of. One cannot prove that a given princi-
ple is a fundamental principle of reason: one can only hope to produce
in one's audience some recognition and acknowledgementof the reasonable-
ness, necessity, or rationally compelling character of what the principle
says.
Wewould have an orderly way to proceed if we had at the start some
clear, detailed, conception of reason in general - of what it is, and of
why it plays the role in the establishment of objectivity and the genera-
tion of conviction that it does. (It cannot be an accident that the
first modern philosopher to make an explicit project of this was the
first modern philosopher to see the possibility of a practical founda-
tion for ethics.) In the absence of such a developed conception, we
must take another way: we can try to pick out those features of recog-
nized rational principles in virtue of which we acknowledge them to be
such and in virtue of which we allow them to play the role of reason in
our lives. And we can show that the principles claimed to be those of
practical reason share those same features. The route followed here is
a comparison of the purported practical principles with those of theore-
tical reason, since we start by supposing that we are more confident of
those. If the practical principles have the features that we find com-
pelling in the theoretical ones, then we may come to believe that they
claim the same allegiance. And the very task of comparing the two kinds
of principles and isolating these features may enable us to arrive at
that more general conception of reason - the reason that is behind both
theory and practice - that we lack at the start. This is what Kant felt
 25

had happened to him when he arrived at the main arguments of the

Critique of Practical Reason. (C2 110/106)
The strategy, then, is to show that the purported principles of
practical reason are indeed principles of reason by concentrating on
features that they share with those of theoretical reason and which seem
to be what give theoretical reason its authority. Nowit is obvious to
everyone that this is what Kant was doing in the case of one of these
distinguishing features: that of universality. Rational principles
have a certain form, the form of universality, and so the test of the
rationality of a principle is universalizability. It is in terms of this
feature that Kant first and most frequently formulates the categorical
imperative; and I suspect that it is commonlybelieved that because it
is on this formulation that his claim for the principle's formality 11
11

rests, it must be primarily on this feature of universalizability that

his claim for the principle's rationality also rests. This is a mistake,
and misses the richest and most evocative features of Kant's argument
for the identity of morality and practical reason.
Universality is only one of several key ideas in the Second Section
of the Foundations of the Metaphysics of Morals, all of which are used
in formulating the categorical imperative. The others are Humanityas
an End in Itself, and Autonomyin the Kingdomof Ends. Mypurpose is to
show howeach of these ideas (and the formulations that can be based on
it) highlights the categorical imperative's embodimentof one of the main
features of a pure rational principle. This is not precisely the use
Kant himself suggests for the various formulations:
The three aforementioned ways of presenting the principle
of morality are fundamentally only so manyformulas of the
very same law, and each of them unites the others in itself.
There is, nevertheless, a difference in them, but the
 26

difference is more subjectively than objectively practical,

for it is intended to bring an idea of reason closer to
intuition (by means of a certain analogy) and thus nearer
to feeling. All maximshave:
1. A fonn, which consists in ~niversality; and in this
respect the fonnula of the moral imperative requires that
the maximsbe chosen as though they should hold as univer-
sal laws of nature.
2. A material, i.e., an end; and in this respect the
formula says that the rational being, as by its nature an
end and thus as an end in itself, must serve in every maxim
as the condition restricting all merely relative and
arbitrary ends.
3. A complete determination of all maximsby the fonnula
that all maximswhich stem from autonomouslegislation ought
to harmonize with a possible realm of ends as with a realm
of nature.
There is a progression here like that through the cate-
gories of the unity of the form of the will (its univer-
sality), the plurality of the material (the objects, i.e.
the ends), and the all-comprehensiveness or totality of the
system of ends. (G 54-55/436)21
But bringing the categorical imperative closer to intuition and to feel-
ing is not the only purpose that these different fonnulations have, and
the basis of the account I will give is present in this passage. Rational
principles have not only a characteristic form, but a characteristic
matter and a characteristic source of authority. The fact that the cate-
gorical imperative can be formulated in tenns of Humanityas an End In
Itself and Autonomy shows that it has these features: and an argument
for the rational status and authority of the categorical imperative can
be based upon each of them. It is these three arguments that I present.
Mytask is focused on showing that these features do characterize
principles of reason, and that each of them can serve as the basis of an
argtonent for the rational status of the categorical imperative and for
the rational character of action and choice governed by the categorical
imperative. While one of my purposes is certainly the explanation and
defense of Kant's moral philosophy, I am equally interested in trying to
showwhat a certain sort of argument - an argument for the practicality
 27

of pure reason, or for the full rationality of a certain modeof action

and choice - might look like, howit might proceed. If I can break the
spell of the empiricist sense of the unimaginability of such an argument,
one of mymajor aims will be achieved. If I can show howKant's cate-
gorical imperative can completely govern action - can assess and some-
times determine our ends, as well as assessing and determining the means
we take to pursue them - and if I can show that this same categorical
imperative is characterized by features that characterize rational prin-
ciples and have something to do with their rational force, then I will
have shownthat Kant has provided us with something that is certainly a
recognizable and plausible candidate for a principle of pure practical
reason. If we can find a principle of pure practical reason we can find
the source of a purely practical kind of objectivity. Such an objecti-
vity could sustain our ethical commitmentsin a world in which theory can
no longer pretend to be their true source of support.
 28

NOTES

1. G.E. Moore, Principia Ethica. That the good must be a simple pro-
perty is argued in the first chapter; the denial that that property
is a "natural" one is based on the denial that by "good" we could
mean any natural property. Fromthis it is supposed to follow, as
the argument is usually read, that the assertion that any natural
thing or property is good is always a substantive, synthetic, claim.
I am unfair about the faculty of intuition; what Mooresays is that
we are left by his analysis of the subject matter of ethics with
"the hope of determining one another's intellect" and that various
intuitions about what is good are to be "submitted to our verdict"
(p. 77}. It is not, I think, quite clear how this verdict is to
be given. Moore seems more interested in heading off errors and
confusions we are inclined to get into when we try to give it than
in setting forth any method. The path taken by Moore's argument
with respect to Kant is exceedingly curious: because "good" must
be a simple property, we cannot say what we mean by it; because we
cannot say what we mean by it, we cannot say that what we mean by it
is what can be rationally willed. At most, seeing what can be
rationally willed might be taken to be a reliable way of spotting
which things are good - but in order to decide that, we would first
have to get up a reliable list of things that are good and then see
howregularly this list coincided with what can be rationally willed.
So we must knowthe good before we can knowwhether rational willing
is a reliable way of finding out what is good. The odd thing is that
this argument could be used about any proposed method of finding out
what is good - and so the non-natural property view El. itself cuts
off epistemic access to the good. This is what makes it so hard to
see how, on Moore's view, we could give~ verdict about what is
good (see especially pp. 131-138).
2. Again I am unfair: it is only the truths about what is prima facie
right that are self-evident. Thus Ross:
That an act,~ fulfilling a promise, or .9!I!, effecting a just
distribution of good, or~ returning services rendered, or
.9.!:@.promoting the good of others, or~ promoting the virtue
or insight of the agent, is prima facie right, is self-evident;
not in the sense that it is evident from the beginning of our
lives, or as soon as we attend to the proposition for the first
time, but in the sense that whenwe have reached sufficient
mental maturity and have given sufficient attention to the pro-
position it is evident without any need of proof, or of evidence
beyond itself. It is self-evident just as a mathematical axiom,
or the validity of a form of inference, is evident. The moral
order expressed in these propositions is just as much part of
the fundamental nature of the universe (and, we may add, of any
possible universe in which there were moral agents at all) as is
the spacial or numerical structure expressed in the axioms of
geometry or arithmetic. In our confidence that these proposi-
tions are true there is involved the same trust in our reason
 29

that is involved in our confidence in mathematics; and we should

have no justification for trusting it in the latter sphere and
distructing it in the former. In both cases we are dealing with
propositions that cannot be proved, but that just as certainly
need no proof. The Right and The Good, pp. 29-30.
I quote the passage at length for the sake of pointing out the odd
response to Kant that it embodies: it reads like a scolding to those
who engage in Critique. That the propositions that seem self-evident
to our reason express anything about the fundamental nature of the
universe is of course what Kant undertook to doubt - so Ross seems
to be trying to struggle back to a pre-Kantian position. That strug-
gle seems to me to be a general feature of 20th century moral philo-
sophy; and I think that an explanation of it might be given.
3. The reference is of course to Jeremy Benthamin The Principles of
Morals and Legislation, especially Chapter IV. Benthamis a good
case because he seems prepared to insist that this is what it is to
makemorals and legislation "scientific" - for him, the point of the
calculus is that under its aegis we can settle moral and legal ques-
tions in a way that is objective in two familiar and related senses:
(a) such questions are settled by reference to (ordinary) facts, such
as facts about what people enjoy, etc. (b) any two people who use the
calculus with respect to the same body of facts will get the same
result. Benthamis inclined to think that principles other than
utility will be used as cover-ups for mere subjective and tyrannical
biases; Mill seems to share that view sometimes. See Bentham, chapter
II; and Mill, in On Liberty, the Introductory chapter. Mill's
accounts of opposing views in Utilitarianism are more generous.
4. J.S. Mill, Utilitarianism, Chapter IV. Mill announces at the outset
that questions about ultimate ends cannot be settled by reasoning,
but are questions of fact - from here he moves to the thesis that
something is an end is established by its desirability, and that in
turn is established by its being desired. It is worth noting that,
given its ownpresuppositions about the nature of ethics, there is
absolutely nothing wrong with this proof. Or rather, if there is
anything wrong with it, the difficulty must lie not in the relation
between desire and desirability, but in the relation between desir-
ability and good - in the idea that questions about ultimate ends
are simply questions of fact, facts about what is desired, or about
what can be desired, or even facts about what would be desired by
those with the right sort of experience of the options, which is
what Mill has in mind.
5. Utilitarianism, Chapter II, pp. 261-262, in Warnock. The passage is:
And there needs be the less hesitation to accept this judgment
respecting the quality of pleasures, since there is no other
tribunal to be referred to even on the question of quantity.
Whatmeans are there of determining which is the acutest of two
pains, or the intensest of two pleasureable sensations, except
the general suffrage of those who are familiar with both?
I do not think it is often noticed that the general argument about
the preferability of pleasures which Mill uses to argue for the
quality difference is taken directly from Plato's Republic, Book IX,
 30

580d-583b. Mill's own, very characteristic, contribution is to see

the possibility of using this argument from classical philosophy in
the context of empiricism. That the higher quality pleasures are in
fact taken in the exercise of our faculties comes from Aristotle,
NicomacheanEthics, BookVII, chapter 12, and in general.
6. The Right and The Good, pp. 157-158. WhenRoss makes the remark,
he is criticizing Kant for being puzzled at the idea that pure rea-
son can becomepractical. "In comparison with this," says Ross,
"actions from desire seemed to him easily intelligible." This,
however, is not right, exactly, although it comes near the truth
about Kant's views in the Foundations. The problem with it is that
it turns out that no humanactions are simply from desire, though
somemight be done on the principle of doing what one desires.
7. As seems to be argued by Prichard in "Does Moral Philosophy Rest
on a Mistake?" and "Duty and Interest".
8. I rely here on the arguments of Anscombein Intention to the effect
that we can give an action (or, I think, a principle or form of
action or whatever) a point or a "desirability characterization"
without making it a means to a separate and possibly ulterior end.
See especially sections 35-41.
9. In Utilitarianism For and Against, by J.J.C. Smart and Bernard
Williams. On p. 7 Smart says:
In setting up a system of normative ethics, the utilitarian
must appeal to some ultimate attitudes which he holds in
commonwith those people to whomhe is addressing himself.
The sentiment to which he appeals is generalized benevolence,
that is, the disposition to seek happiness, or at any rate, in
some sense or other, good consequences, for all mankind, or
perhaps for all sentient beings. (p. 7)
In one sense Mill's view and Smart's are diametrically opposite here.
Smart claims to be a non-cognitivist: his ethical theory is meant
to appeal to those who share the ultimate attitude of generalized
benevolence; and to address objections; it cannot be proved. Mill,
on the other hand, thinks that the principle of utility does admit
of a certain sort of proof, but that proof has nothing to do with
our motives for obeying the principle of utility at all. (The
movefrom the aggregate desiredness of the general happiness - each
part of it desired by the person whose happiness it is - to the
objective desirability of the general happiness could not be made
if the acknowledgementthat the general happiness is desirable were
intended to carry any motivational weight; you can aggregate the
happiness but not the desire, for there is no aggregate person to
have the desire. But the proof only leads to the acknowledgement
that the general happiness is desirable, not to any motives: Mill
is a pure externalist. See note 12.) Mill makes this clear by
treating the issue of the moral sanction separately and indeed, by
regarding it as sanething that we must simply acquire through a good
upbringing and education.
10. See Note 9, and Utilitarianism, Chapter III.
 31

11. G 37-38/420. Hypothetical Imperatives cannot be laws, because

"what is necessary merely for the attainment of an arbitrary purpose
can be regarded as itself contingent, and we get rid of the precept
once we give up the purpose."
12. The internalism/externalism was introduced by Falk in '11 0ught' and
Motivation" and taken up by Frankena in "Obligation and Motivation
in Recent Moral Philosophy". Nagel discusses it in Chapter II of
The Possibility of Altruism. The argument in question appears on
p. 28 of Chapter V.
13. A very good example of this is found in Hume's conclusion to Book
III of the Treatise. After summingup the evidence that all of our
moral judgment and attitudes can be sufficiently explained by his
theory of sympathy, Humeadds:
Were it proper in such a subject to bribe the readers assent,
or employ any thing but solid argument, we are here abundantly
supplied with topics to engage the affections. All lovers of
virtue (and such we all are in speculation, however we may
degenerate in practice) must certainly be pleas'd to see moral
distinctions deriv'd from so noble a source, which gives us a
just notion both of the generosity and capacity of humannature.
It requires but very little knowledgeof humanaffairs to per-
ceive, that a sense of morals is a principle inherent in the
soul, and one of the most powerful that enters into the composi-
tion. But this sense must certainly acquire new force, when
reflecting on itself, it approves of those principles, from
whence it is deriv'd, and find nothing but what is great and
good in its rise and origin. Those who resolve the sense of
morals into original instincts of the humanmind, may defend
the cause of virtue with sufficient authority; but want the
advantage, which those possess, who account for that sense by
an extensive sympathy with mankind. According to their system,
not only virtue must be approv'd of, but also the sense of
virtue: And not only that sense, but also the principles,
from which it is deriv'd. So that nothing is presented on any
side, but what is laudable and good. Treatise, p. 619.
All along, Humehas supposed that the task of moral philosophy is
to explain the origin and use of "moral distinctions" and especially
the praise and blame of the moral spectator. That the view that
results is an attractive one is (i) just a bonus and (2} supposed
to be motivating, and make it possible to defend the cause of virtue.
A closely related passage appears in Part II of the conclusion of
the Enguir~, where Humeis actually talking about "our interested
obligation to the practice of virtue
And though the philosophical truth of any proposition by no
means depends upon its tendency to promote the interests of
society, yet a man has but a bad grace who delivers a theory,
however true, which, he must confess, leads to a practice
dangerous and pernicious.... But what philosophical truths
can be more advantageous than those here delivered, which rep-
resent virtue in all her genuine and most engaging charms and
makes us approach her with ease, familiarity, and affection?
The dismal dress falls off.... Aiken, pp. 257-258.
 32

14. This is clear in Mill: because the natural feelings support the
utilitarian conscience, the utilitarian is not sorry to have this
conscience, even though it is wholly acquired.
It does not present itself to their minds as a superstition
of education, or a law despotically imposed by the power of
society, but as an attribute which it would not be well for
them to be without. This conviction is the ultimate sanction
of the greatest happiness morality. This it is which makes
any mind of well-developed feelings, work with, and not
against, the outward motives to care for others •...
(Utilitarianism, Chapter III, p. 287 in Warnock)
15. As in the famous passage from Mill discussed in Chapter two, p. 43.
16. See the opening section of Onora Nell's Acting on Principle for a
discussion. She has christened the two issues I mention fertility
and formality; she discusses there the appeal and the purported
difficulty of combining them.
17. David Hume,Treatise of HumanNature, Book II, Part III, Section
III. pp. 413-418 in Selby-Bigge.
18. Hume,Treatise, p. 415.
19. Hume,Treatise, p. 416.
20. Only the success of this thesis can prove that the formulations in
tenns of Humanityas an End in Itself, and Autonomyin the
Kingdomof Ends have a more important role than bringing the
formula closer to intuition, if that is taken, as usual, to mean
ordinary moral intuition. It is an odd thing for Kant to say, for
the analysis of ordinary rational knowledgeof morality led him
most immediately to the universal law formulation, which more or
less appears in section one. (G 18/402) As an example of the
attitude that sometimes finds its justification in this unfortu-
nate remark of Kant's, I offer the following passage from A.R.C.
Duncan's book with the promising title Practical Reason and
Morality.
Analyzing this motive, Kant finds that it may be expressed in
a formula of universality of the maxim..•.
Having reached the principle of morality in this sense,
Kant's next step is to connect this analysis with the idea
of practical reason ....
If Kant had at once gone on to deal with that problem,
instead of postponing it to Section III, muchmisunderstanding
of his argument might have been avoided.••• Unfortunately at
that point he interpolates a lengthy passage in which he sug-
gests three new fonnulae for the categorical imperative. As
we shall see, this passage does not advance the main argument
and is open to serious misinterpretation. Eventually, however,
after introducing the concept of the autonomyof the will to-
wards the end of Section II, Kant returns in section III to
his main critical theme. (pp. 52-53)
I do not think that this could possibly be more wrong, and that is
 33
what I hope to show. The Duncanexample is a good one because he
is actually writing about Kant. It is needless to say that most
writers on ethics who have treated Kant as an example of one kind
of view have supposed that the connection with rationality rested
entirely on the universality fonnulae. As an arbitrarily selected
example, here's Frankena: "AndKant and Hare take some such view
of the principle of justice, holding that the place of reason, in
ethics at least, is an insistence that we universalize our maxims."
(p. 203 of Pers ectives on Moralit, in 11Sidgwick and the Dualism
of Practical Reason'. This is quite typical.
 Chapter 2

Universal Law

The Second Section of the Foundations of the Metaphysics of Morals

contains three arguments concerning the categorical imperative, each
leading to a new way of fonnulating it. Each of these arguments has
the fonn: if there were a categorical imperative, this is what it
would have to be like. None of them is meant to establish that there
really is a categorical imperative, for the establishment of the cate-
gorical imperative, as a synthetic! priori law of reason, waits upon
the methods of critique, and requires a "deduction." 1
In the first of these arguments, leading to the Fonnula of Univer-
sal Law, Kant claims, "if I think of a categorical imperative, I know
irnnediately what it contains." (G 38/420) The second argument, leading
to the Formula of the End in Itself, establishes that, "if there is
to be a supreme practical principle and a categorical imperative for
the humanwill, it must be one that forms an objective principle of the
will from a conception of that which is necessarily an end for everyone
because it is an end in itself." (G 47/428} And the third, leading to
the ideas of Autonomyand the Kingdomof Ends, is that, "if there is a
categorical imperative (a law for the will of every rational being}, it
can only commandthat everything be done from the maximof its will as
one which could have as its object only itself considered as giving
universal laws.'' (G 50/432) In each case, Kant argues that we can see
what a practical law of reason - a law for the will of every rational

34
 35

being - would have to be like, if there were one. This means that each
of these arguments should focus our attention on some important feature
of principles of reason, somecharacteristic they have to have. A law
of reason, according to Kant, must be universal, unconditional, and
autonomous. The arguments leading to the three formulations are meant
to establish that the categorical imperative has these three character-
istics, and so that it is the law of practical reason, if there is one.
Mytask here is to describe these arguments, and show howKant uses
them to establish that the law of practical reason would be the cate-
gorical imperative.
In this chapter and the next two, I am concerned with the idea of
universal law. Because Kant himself said that "it is better in moral
evaluation to follow the rigorous method and to make the universal
fonnula of the categorical imperative the basis" (G 55/436), the
formulations involving universal law have received the most attention
from commentators, who have naturally been interested in the use of the
categorical imperative in guiding concrete moral decision. In this
chapter, I will consider someof the more commoninterpretations and
criticisms of the use of the universal law formulations in the identifi-
cation of duties. In Chapter Three, I describe and defend the interpre-
tation of the universal law formulations which I think correct. In
Chapter Four, I consider someof the ramifications, in light of the
interpretation I defend, of Kant's idea that in order for an agent to
detennine whether his maximcould be willed as a universal law, he
should consider whether he would be able to will it as a law of nature.
Whenthese arguments are in place it should be possible to see why it is
rational to act only on maximsthat can be willed as universal laws.
 36

I. The Argumentfor the Formula of Universal Law

The first formulation of the categorical imperative in the Second
Section of the Foundations is
Act only according to that maximby which you can at the same
time will that it should becomea universal law. (G 39/421)
Kant arrives at this idea, he claims, from "the mere concept of a cate-
gorical imperative" (G 38/420):
But if I think of a categorical imperative as such, I know
inmediately what it contains. For since the imperative con-
tains besides the law only the necessity that the maximshould
accord with this law, while the law contains no condition to
which it is restricted, there is nothing remaining in it
except the universality of law as such to which the maximof
the action should conform; and in effect this conformity alone
is represented as necessary by the imperative. (G 38-39/420-
421)
I will not attempt to reconstruct this very condensed argument.2 It is
based on, and to some extent recapitulates, the analysis of "conman
rational knowledgeof morals" in the First Section of the Foundations,
which leads to the conclusion that the principle of an absolutely good
will is
I should never act in such a way that I could not also will
that mymaximshould be a universal law. (G 18/402)
There the argument runs something like this: The person to whomwe
would attribute a good will is one who does her duty from the motive of
duty, simply because it is a duty. This must mean that in selecting
her maxim, the good-willed person does not merely consider its matter,
but looks to, and is motivated by, its form: its legality. The one
thing we can knowabout duty (in advance of substantive moral theory)
is that it is a law for everyone - a universal law. The person who acts
from duty therefore must be motivated by the conception of her maxim
as a law for everyone.
 37

It has been argued that the idea of "doing duty for the sake of
duty 11 cannot possibly have, by itself, any content. This argument shows
up before Kant, in Hume's Treatise of HumanNature; 3 and after Kant,
4
indeed, as a criticism of Kant, in Ross's The Right and The Good.
In both cases the argument is in essence this: until you knowwhat
your duty is, you cannot set about doing it for its own sake, and the
commandthat you do your duty for its own sake does not seemto tell you
what to do. Humesays that it is a circle, and Ross, that it is an in-
finite regress, to claim that such a commandtells us what to do. Now
it is worth pointing out that - to put it in Kantian language - if a
maximhad only a matter and not also a fonn, this would be true. A
maximis a subjective principle for the will, the principle upon which
the agent acts. The matter of a maximis very closely related to
its purpose, for the maximgives the purposive structure of the action -
its relation to its purpose. If maximshad only matter, the only way
to pick out which ones were done from duty would be by providing, in
advance of the selection, some criterion for judging the morality of
the matter of a maxim, and such a criterion would necessarily be a sub-
stantive one, requiring independent justification of some kind. One
might, for instance, simply provide a list of duties (known by intuition,
revelation, or what have you) or a list of morally good purposes. A
difficulty with this, which I think Kant is at some pains to make clear
in his examples, is that there is almost no purpose which might not be
adopted for either moral or non-moral reasons. Wecannot illuminate the
feature of actions and choices which especially constitutes their moral
worth by providing a list of purposes of which any one might be adopted
either for moral or completely non-moral reasons.
 38

Thus the moral worth of an action does not lie in the effect
which is expected from it or in any principle of action which
has to borrow its motive from this expected effect. For all
these effects (agreeableness of my own condition, indeed even
the promotion of the happiness of others} could be brought
about through other causes and would not require the will of
a rational being, while the highest and unconditional good
can be found only in such a will. (G 17/401)
In order to understand this correctly the first requisite is to see that
Kant distinguishes the motive of an action from its purpose. The pur-
pose lies in the matter of the maxim, but the maximof the action may
be adopted either because of its form or because of its matter (because
of the purpose). But in either case anyone who adopts the maximcer-
tainly adopts its purpose.
Take the notorious benevolence example. One person helps another
because, being sympathetic, she 11find[s] an inner satisfaction in
spreading joy, and rejoice[s] in the contentment of others 11 (G 14/398}
which she has made possible. Another person helps others because it
is a duty. Nowit is sometimes supposed that Kant is claiming that
the benevolent person s purpose is only to please herself (the benevo-
1

lence is just a pleasure-producing lever she pushes); whereas the dutiful

person's purpose is to do his duty. This gives the unfortunate impres-
sion that neither of them has the purpose of helping others for its own
sake. But that is quite mistaken - they both have the purpose of help-
ing others for its own sake. and in both cases it is a laudable, unsel-
fish purpose. The difference between them lies not in the purpose but
in the motive - the motive under the influence of which they adopt the
maximin which this purpose is embodied. The benevolent person, as
described, adopts the maximbecause of an inclination that interests
her in its purpose, its matter. She helps others because she likes
 39

helping others. Kant agrees with the view that this is the very opposite
of being selfish; he describes this person as being devoid of vanity
and selfishness. (G 14/398) The dutiful person, on the other hand,
adopts this maximbecause of its form. He may not be doing what he
likes, but he could not will a world in which no one has this maxim,
and so he does his part. But in adopting the maxim, he adopts its pur-
pose. And this purpose is not a means to the end of doing duty, but an
end which it is a duty to have. The two people differ, not in their
purpose, which is the same, but in their motive, which is quite differ-
ent.5 This is precisely why the distinguishing feature of a good will
cannot be given in terms of purposes, or indeed in terms of the matter
of maximsat all. It must, then, be given in tenns of their fonn.
Since duty is law, that fonn is the form of law, which is universality:
Since I have robbed the will of all impulses which could
come to it from obedience to any law, nothing remains to
serve as a principle of the will except universal conformity
of its action to law as such. That is, I should never act
in such a way that I could not also will that mymaximshould
be a universal law. (G 18/402)6

II. Interpreting the Fonnula of Universal Law

The Fonnula of Universal Lawsays that one is to act only on a
maximthat one could at the same time will to be a universal law. In
order to determine whether it is morally pennissible to perform some
action which you propose to yourself, you must submit your maximto a
test: you must see whether you could at the same time will it to be
a universal law. Nowhow is this test to be carried out? Kant tells
us that
(By analogy] ... the universal imperative of duty can be ex-
pressed as follows: Act as though the maximof your action
 40

were by your will to becomea universal law of nature.

{G39/421)
This is the Formula of the Lawof Nature, and it tells us that we are
to act on such maximsas we could will to be laws of nature. The test
for this, in turn, is a purely formal test. A maximcannot be willed
to be a law of nature if it cannot be willed to be a law of nature with-
out contradiction. Kant explains:
Wemust be able to will that a maximof our action becomea
universal law; this is the canon of the moral estimation of
our action generally. Someactions are of such a nature
that their maximcannot even be thought as a universal law
of nature without contradiction, far from it being possible
that one could will that it should be such. In others this
internal impossibility is not found, though it is still
impossible to will that their maximsshould be raised to
the universality of a law of nature, because such a will
would contradict itself. (G 41-42/424)
A maximfails the first contradiction test if it cannot be thought as
a universal law of nature; it fails the second contradiction test if
it cannot possibly be willed as a law of nature. It is not pennissible
to act on a maximwhose universalization generates a contradiction in
either of these ways.
It has been a subject of extensive discussion what sorts of con-
tradiction Kant has in mind. Different views about this have led to
very different interpretations of the two contradiction tests associated
with the Formulae of Universal Lawand the Lawof Nature. In trying to
answer this question, we have various things to go on. First of all,
we have the examples, and Kant's ownaccounts of the examples. Most
interpretation has depended almost entirely on these. Second, since
the contradiction is to be in somecases in even thinking a maximas a
universal law, and in some cases only in willing a maximas a universal
law, we can appeal to what we knowabout maxims, to see what sort of
 41

contradictions their universalized versions admit of. Of course to

some extent what we say about maximswill be informed in turn by the
interpretation we give of the contradiction tests - a maximmust be some-
thing whose universalized version gives rise to that kind of contradic-
tion, whatever it may be. But maximsalso have a specific psychological
or even ontological role - they are what we act on - and it may be pos-
sible to draw some conclusions from that about what they and their
universalized counterparts must be like. Not enough advantage has been
taken of this source of guidance.
Finally, we must allow ourselves to be guided by general considera-
tions of what Kant's moral philosophy is all about. The laws generated
by this test are supposed to be the laws of an Ideal community- a
Kingdomof Ends. This kingdomis an Ideal in the Kantian sense, which
he himself associates with the Platonic sense; 7 it is a rationallx
perfect community. Nowsince on a Kantian view an Ideal must be genera-
ted from an idea we cannot knowthe nature of the Kingdomof Ends inde-
pendently of an analysis of the moral law itself. It is not an end to
which the moral law is a means, but rather an end that would be generated
by the moral law if that law truly governed the world. Nevertheless,
we can work with some sense of what sort of Ideal communitythe Kingdom
of Ends is supposed to be. Eventually, I will argue that the Kingdom
of Ends is a world for action. And I will allow the idea of moral laws
as the laws of a rationally perfect world for action to guide my inter-
pretation of the categorical imperative.
Most interpretations of the Formula of Universal Laware based on
readings of the four examples Kant uses in the Foundations. Although
a great manyslightly varying readings of these can be found in the
 42

literature on Kant, it seems to me that they resolve into four basic

types. For ease of handling I will give these names, as follows:
i) The Theoretical Contradiction Interpretation. On this view
there is some logical or physical impossibility in the univer-
salization of the maxim,or in the law of nature corresponding
to the maximor the system of nature in which that is a law,
at least in the case of the first contradiction test.
ii) The Terrible ConsequencesInterpretation. On this view, the
maximcannot be universalized or madea universal law because
that would have very terrible consequences - consequences so
terrible as to be unacceptable to a rational being.
iii) The Teleological Contradiction Interpretation. On this view,
the law of nature in which the contradiction emerges is a
teleological law, or anywaythe universalized maximwould not
be a law fit for a teleologically organized system of nature.
iv) The Practical Contradiction Interaretation. On this view,
the sort of contradiction involve is a practical contradic-
tion - a thwarted purpose or a self-defeating way of acting.
In the case of the first test, the contradiction is that the
agent would be unable to achieve the purpose in her maximin
the world in which her maximwas a universal law. In the
case of the second test, somepurpose or purposes that belong
to rational agents as such must be thwarted in the world of
the universalized maxim.
The rest of this Chapter will be devoted to an examination of the first
three of these interpretations. I will describe each and try to show
what can be said for and against it. Mycritical discussion of the
Teleological Contradiction Interpretation will lead to some important
considerations about Kant's views concerning the nature of practical
reason, which I will treat of in the last section of this Chapter. In
Chapter Three I will set out and defend the Practical Contradiction
Interpretation, which I believe to be correct.

III. The Theoretical Contradiction Interpretation

Mill famously accuses Kant of wanting this view, but of having to
fall back on the terrible consequences view instead:
 43

But when he begins to deduce from this precept any of the

actual duties of morality, he fails, almost grotesquely, to
show that there would be any contradiction, any logical (not
to say physical} impossibility, in the adoption by all ration-
al beings of the most outrageously immoral rules of conduct.
All he shows is that the consequences of their universa~
adoption would be such as no one would choose to incur •
•.. he virtually acknowledgesthat the interest of mankind
collectively, or at least of mankind indiscriminately, must
be in the mind of the agent whenconscientiously deciding on
the morality of the act. Otherwise he uses words without a
meaning: for, that a rule even of utter selfishness could
not possibly be adopted by all rational beings - that there
is any insuperable obstacle in the nature of things to its
adoption - cannot be even plausibly maintained. To give any
meaning to Kant's principle, the sense put on it must be,
that we ought to shape our conduct by a rule which all ration-
al beings might adopt with benefit to their collective
interest.9 -- -
I suppose hardly any of this interpretation's proponents have held it
in the pure form that Mill describes - what they have looked for is
something very like a logical or physical impossibility, and it tends
to flow over into the Practical Contradiction view. Part of the reason
for this is that it is clear that this cannot be the right interpreta-
tion for the second contradiction test - that is, the case of maxims
whose universalizations Kant says cannot be willed but certainly can
be thought. But there is no question that some of the language Kant uses
in the case of the first contradiction test, and some very plausible
readings of one of Kant's examples - the false promising example - tend
to favor this interpretation. For example:
He inmediately sees that it could never hold as a universal
law of nature and be consistent with itself; rather it must
necessarily contradict itself. (G 40/422}
Thus mymaximwould necessarily destroy itself as soon as it
was made a universal law. (G 19/403}
Someactions are of such a nature that their maximcannot
even be thought as a universal law of nature without contra-
diction, far from it being possible that one could will that
 44

it should be such. (G 41-42/424)10

This talk of self-consistency, self-destruction, and of a maxim's not
being able even to be thought as a universal law without contradiction,
certainly sounds, especially to philosophers' ears, as if Kant had
something like a physical impossibility or logical inconceivability in
mind.
So evidently, it seemed to Mill and so it has seemed to others.
For example, Paul Dietrichson seems to have something like this inter-
pretation in mind when he says:
Myreason tells me that if it were morally legitimate for
me to act on my present subjective rule in the type of cir-
cumstances I am in at present, it would be morally legitimate
for everyone to act on the same type of rule in the same type
of circumstances. It should therefore be possible at least
to conceive the idea that everyone would do so; for it should
be possible to conceive the idea that everyone in circumstances
like mine would, with the regularity of events confonning to
natural law, act in a morally legitimate manner. If the
principle of the maximon which I act is such, then, that I
cannot consistently conceive of it as becomingraised to the
universality of the law of nature, I am acting in a morally
illegitimate manner: exempting myself from t?f categorical
demandwhich ml~ reason makes clear to me.
The false promising example, which is the one that best supports this
view, goes like this. A man in financial difficulties considers borrow-
ing moneywhich he knowshe can never repay. His maximfails the first
contradiction test in the following way:
... the universality of a law which says that anyone who be-
lieves himself to be in need could promise what he pleased
with the intention of not fulfilling it would make the pro-
mise itself and the end to be accomplished by it impossible;
no one would believe what was promised to him but would only
laugh at any such assertion as vain pretense. (G 40/422)
Dietrichson says, accordingly:
The principle of the maximis, of course, not self-contradic-
tory; but if the principle were to becomea universal law
of nature, such a law would be a self-contradictory law.
 45

Because the very principle of the law would contradict itself,

the law in question would be no law at all. - If the world
were to be such that everybody whomakes promises in order to
get out of a difficult situation would, according to a natural
law of voluntary action, be absolutely incapable of keeping
those promises, no one would take their so-called promises
seriously. Such 11promises11 would be knownto be ~orthless
and would therefore not be regarded as promises. 1
The difficulty with Dietrichson's line shows up in his own paper.
He mentions the dubious case of a womanwho has "arbi trari ly 11 (his tenn)
decided to consider the maxim11If I give birth to a baby weighing
1113
less than six pounds, I shall do everything in my power to kill it.
Dietrichson admits that it is certainly possible to conceive the idea
of every mother behaving according to this rule. (Wecould bring this
silly case closer to a real one if we imagine that the maximis "killing
off every girl-child after the first or second, because of the needless
and fruitless economicdrain of raising girls, 11and it is obvious that
this could be a universal practice, since in some places at some times
it almost has been.) Dietrichson's solution at this point is to appeal
to the second contradiction test, and to place this amongthe maxims
whose universalization cannot be willed although it can be thought. 14
Oietrichson's reading depends in general on making an extensive use of
the second contradiction test. But even in the Foundations, Kant as-
sociates the second contradiction test with wide, meritorious duties
(G 42/424), such as the duty of helping others, about which the law en-
joins only that we do something, but not what or howmuch. The first
test is supposed to cover the narrow or perfect duties, and it is ob-
vious that the duty not to murder one's children is of this type. 15
(At a later stage in the argument, this distinction between the two
kinds of duties becomesmore definite and more important. See
 46

Chapter 6.)
I think that it is fairly clear what has gone wrong here. There
is a class of actions for which a logical contradiction interpretation
works very well. These are actions that are in some way conventional
or institutional in their very nature - actions associated with what
Rawls ca 11s "practices. 1116 Promising, fol lowing the rules associated
with conventions of property (e.g. not stealing), and refraining from
various forms of cheating, are examples of duties that seem to be easily
established by the logical contradiction approach because they are
associated with practices. This is because practices have fairly
definite functions and are associated with definite rules, and - impor-
tantly - are more or less defined in terms of those rules. Rawls says:
•.. the rules of practices are logically prior to particular
cases. This is so because there cannot be a particular
case of an action falling under a rule of a practice unless
there is the practice •... Wemay think of the rules of a
practice as defining offices, moves, and offenses. Now
what is meant by saying that the practice is logically
prior to particular cases is this: Given any rule which
specifies a form of action (a move), a particular action
which would be taken as falling under this rule given that
there is this practice would not be described as that sort
of action unless there was the practice. In tne case of
actions specified by practices it is logically impossible
to perfonn them outside the stage-setting provided by those
practices, for unless there is the practice, and unless
the requisite properties are fulfilled, whatever one does,
whatever movementsone makes, will fail fOcount as a form
of action which the practices specifies. 7
If one wants to do an action which a certain practice speci-
fies then there is no way to do it except to follow the rules
which define it. Therefore, it doesn t make sense for a per-
1

son to raise the question whether or not a rule of a practice

correctly applies to his case where the acfion he contemplates
is a form of action defined by a practice. B
Obviously, when we imagine the universalization of 'false promising to
get out of a difficulty', we imagine that the practice of promising is,
 47

when that maximis universalized, quite different than it actually is

and, indeed, that it has a kind of contradiction built into it: your
promises must be true but can be false. This contradictory practice of
promising would not work; for under these conditions promises would not
be accepted. As Dietrichson said, 11
Such promises would be knownto be
worthless and would therefore not be regarded as promises." The person
who tries to will the universalization of the maximof false promising
in effect wills that the practice of promising will self-destruct. So
the theoretical contradiction interpretation might run: to picture a
world in which everybody in a difficulty makes false promises is in fact
to picture a world in which nobodymakes any promises, because this prac-
tive would not exist in that form. Therefore, such a world obviously
cannot be conceived - everyone is making false promises, nobody is mak-
ing promises - it has a contradiction in it, and cannot be conceived.
This is the place to interpose mention of an objection that seems
to be based on just this sort of interpretation. Kant gives an example
in the Critique of Practical Reason which is very like the false pro-
mising case. A man has in his possession a deposit for which there was
no receipt. The owner of the goods has died, and the man is considering
whether he may keep the goods. Kant says: "I immediately realize that
taking such a principle as a law would annihilate itself, because its
result would be that no one would make a deposit. 11
(C2 27/27) Discus-
sing this case H.B. Acton reports:
It is against this argument that Hegel brings one of his
objections to Kant's position. In an essay entitled On the
Scientific Treatment of Natural Law(1803), Hegel saysthat
all Kant's argument shows is that a system without deposits
is contradicted by a system with deposits, but not that there
is any contradiction in a system without deposits. Kant makes
there seem to be a contradiction in a system without deposits
because he assumes that everyone would want there to be deposits,
 48

and this, says Hegel, shows that Kant was assuming the system of
property and was arguing that if everyone kept what belongs to
others then there would be no system of property. The interesting
question, Hegel goes on, just is why there should be property,
and about this Kant says nothing. In his Hegelian Ethics Professor
W.H.Walsh applies this objection to Kant's discussion of lying
promises and says that if everyone made lying promises no promises
would be made, but that it does not follow that a world without
promises would be morally inferior to the existing world. All
that Kant shows, therefore, is that it is impossible to retain the
institution of promising §nd at the same always [sic] go against
what is essential to it.1
Acton's own response to the difficulty is to argue that a world without
promises would indeed be worse, both morally worse and worse off, than
one with them. This has the tendency to push Kant back to the Terrible
Consequencesview. Later, discussing the fourth example, Acton reads
Kant's argument against the maximof not helping anyone as saying that
the universalization of that maxim"leads to self-contradiction, or
would result in forms of society which are more or less unacceptable to
thinking and feeling humanbeings." 20
But there is no need for this, for the objection is not to the
point at all. On the theoretical contradiction interpretation, the
contradiction lies not in envisioning a society without deposits and
promises, but in envisioning a society in which people both are and are
not making deposits and promises. It is a result of universalizing the
maximthat no promises are made, but it is also a result of the univer-
salization of the maximthat everyone holds this maxim,and therefore
in certain situations is making (or trying to make) false promises. On
the practical contradiction interpretation - which in the case of the
practices is very close to the theoretical contradiction interpretation -
the answer to this objection can be given more directly. The person
whotries to will the universalization of this maximis not only willing
 49

a situation in which the practice of promising does not exist. He is

also, because he wills the maximitself, willing to make~ of the
practice of promising. His maximcommits him to the use of the practice,
and its universalization commits him to the non-existence of the prac-
tice. (To a world in which it has died of abuses like his.) Thus he
wills to achieve his purpose by a method that he wills not to work, and
so wills the thwarting of his ownpurpose. He has chosen his method of
getting out of the difficulty, but universalization commits him to the
failure of the method. The same goes for the other example - the man
is proposing to use the system of deposits to get what he wants, and
then willing the system of deposits out of existence. Thus the answer
to the Hegelian objection is that Kant need not be assuming that every-
one wants there to be deposits. The man in the example wants there to
be deposits, because he is using deposits to achieve his aim.
The theoretical contradiction interpretation can cope with this
sort of objection, I think. The real problem with it lies elsewhere -
in its difficulties in dealing with what might be called natural acts,
acts that are not defined by conventions or the rules of practices.
Dietrichson s problem about the child-murder case arises because killing
1

a child is a natural act - killing is still killing no matter howmany

people engage in it. It is a sort of action that does not logically
require for its existence any practices or conventions, and which cannot
be destroyed as a type of action by some sort of misuse. In the world
in which false promising was supposed to be universalized, the contra-
diction was that people were to be conceived both as making promises
(because of the universality of the maxim)and as not making promises
(because of the inevitable failure of the practice). But where the
 50

action is of a wholly natural (non-conventional) kind, this sort of

thing cannot be said. Wecannot say that the people in the world where
the maximof killing off undersized children both are and are not kill-
ing. They just are. There is no contradiction in the universalization
of killing as such, just as a kind of action. If we want to include
someof these natural actions amongthe obvious injustices that violate
narrow or perfect duty, we cannot rest with Dietrichson's approach.
And of course we do. Murder and assault and rape and violent crimes
generally are amongthe most obvious examples of things that are strict-
ly forbidden by considerations of justice. There is no theoretical
contradiction - no physical or logical impossibility - in the universa-
lization of these activities. If we are to cover this kind of case,
we need some other interpretation.

IV. The Terrible ConsequencesInterpretation

I have already mentioned Mill, and perhaps, Acton, as indicating
that Kant might be forced back on the Terrible ConsequencesInterpreta-
tion by the failures of the Theoretical Contradiction Interpretation.
The effect of the Terrib1e ConsequencesInterpretation is to make
Kant a kind of rule-utilitarian. Or, to put the same point another way,
its effect is to make Kant a proponent of what Singer calls the
Generalization Principle rather than the Universalization Principle. 21
Singer himself does not take this view, arguing finnly that the two
principles are different, although anything that fails a universaliza-
tion test will naturally fail generalization as well.
The generalization argument states that if the consequences
of everyone's acting in a certain way would be undesirable,
then no one has the right to act that way - without reason
or justification. The principle of universality, as I have
 51

been interpreting it, states that if, as a result of every-

one's acting or trying to act in a certain way, no one would
be able to, then no one has the right to act in that way -
without a reason. This is in fact the essential difference
between them.1122
Except for that last "without a reason" and once 11in a certain way" is
properly specified, this is nearly correct.
I do not knowof anyone who actually undertakes to defend Kant's
view who favors a Terrible ConsequencesInterpretation. Usually, it
is regarded as a position Kant is forced back to after the theoretical
contradiction method has been shownto fail. It is perfectly clear that
Kant wants something more than this, and it is perfectly clear why -
this interpretation will not sustain the claim that immoral action is
in any newsworthysense irrational. One can of course say that no
rational being will be in favor of terrible consequences, and that a
rational being could not will the universalization of any maximthat
fails the generalization argument. But this is precisely what Kant is
trying to show, and therefore to use this in interpreting the imperative
is to defeat the project; to make it circular. For on this reading we
have already put forth a heavily normative view of what it is to be a
rational being, and therefore assumedthe connection between moral good-
ness and rationality that Kant is trying to establish. The effect of
the terrible consequences view is to give up the idea that there is any
contradiction, unless it is this contradiction between being rational
and willing something with terrible consequences that makes the theory
circular.
The terrible consequences view is thus a defeat for Kant, and so
as a possible interpretation it is not of much interest. It is interes-
ting, I think, that so manyphilosophers have been prepared to see this
 52

as the only alternative to the theoretical contradiction view; I will

say more about that in the section on practical contradiction.
But it is important to see what drives people to suppose Kant is
thrown back on the Terrible Consequencesview, because here we come to
something that any account of Kant must deal with. This is the problem
of interpreting the second of the two contradiction tests: where the
contradiction is not in the universalized maximor in conceiving a
world in which the maximis universalized, but in the will if the
universalized maximis willed. The two examples given of this way of
testing a maximdirect us to the adoption of what will later, in the
Metaphysics of Morals, be declared to be the two "obligatory ends11 :
one's own perfection and the happiness of others. Since these turn out
to be the two obligatory ends it is a little misleading for Kant to
put these forward as examples in the Foundations, for in a sense these
two examples exhaust the use of this test, or at least encompassall
the uses of this test. In the Metaphysics of Morals, these duties are
derived from the duty of treating Humanityas an End in Itself. In the
first set of examples in the Foundations, the derivation of these duties
takes place in the same way as the derivation of the strict duties -
through the rejection of a maxim- except in this case it is because
the universalized maximcontradicts the will rather than because it
contradicts itself. Kant argues that a man cannot will the 11maximof
neglecting his gifts" because:
... he cannot possibly will that this should becomea univer-
sal law of nature or that it should be implanted in us by a
natural instinct. For, as a rational being, he necessarily
wills that all his faculties should be developed, inasmuch
as they are given to him for all sorts of possible purposes.
(G 41/423)
Similarly, the man who attempts to will the maximof indifference, or
 53

of not helping anyone else, in universal form finds that:

... although it is possible that a universal law of nature
according to that maximcould exist, it is nevertheless
impossible to will that such a principle should hold every-
where as a law of nature. For a will which resolved this
would conflict with itself, since instances can often arise
in which he would need the love and sympathyof others, and
in which he would have robbed himself, by such a law of
nature springing from his ownwill, of all hope of the aid
he desires. {G41/423)
Nowthe difficulty is this. It is perfectly obvious that nothing
like the theoretical contradiction view will help us with this test, for
Kant himself says that we can certainly conceive systems of nature in
which these are laws. The contradiction is "in the will." But this
must mean that there is something in the will for the world of the
universalized maximto conflict with, to contradict. And this something,
be it what it may, must be in the will of every rational being. For
every rational being must, it seems, be able to carry out the test with
the same results. And:
... it is extremely important to rememberthat we must not let
ourselves think that the reality of this principle can be de-
rived from the particular constitution of humannature. For
duty is practical unconditional necessity of action; it must,
therefore, hold for all rational beings (to which alone an
imperative can apply), and only for that reason can it be a
law for all humanwills. Whatever is derived from the parti-
cular natural situation of manas such or from certain feelings
and propensities, or even from a particular tendency of the
humanreason which might not hold necessarily for the will of
every rational being (if such a tendency is possible), can
give a maximvalid for us but not a law; that is, it can give
a subjective principle by which we might act only if we have
the propensity and inclination, but not an objective princi-
ple by which we would be directed to act even if all our
propensity, inclination, and natural tendency were opposed
to it. (G 43/425)
Kant's ownanalyses of the purported contradiction lead even those
who favor the Theoretical Contradiction Interpretation of the first
test to suppose that something like the Practical Contradiction Inter-
 54

pretation is on Kant's mind in the case of this test. In both cases

Kant emphasizes the idea that in willing the universalization of these
maximsa person wills the non-existence of something she will need or
want: love and sympathyin manyan instance, developed faculties for
all sorts of possible purposes. But, it is urged, this can only generate
even a practical contradiction if we can attribute these needs or wants
necessarily to every rational being, or at least to everyone to whom
these duties are supposed to apply. This raises two objections. Either
it is argued that we cannot attribute to every rational being the ends
to which these needs or wants apply, or, even if we can suppose that a
rational being has the ends, we cannot suppose that she will necessarily
require just these means. For instance; C.O. Broad complains:
The example of the second case is equally unfortunate,
though in a different way. If it is to be relevant at all
we must suppose that the principle of Egoismis accepted as
a categorical imperative, and not as a mere rule for gaining
maximum personal happiness. Myacceptance of the principle
therefore does not presuppose a desire for my own happiness
or a belief that this is the most effective way to secure
it. Nowall that Kant shows is that the acceptance of this
principle by others would be likely to lead to consequences
detrimental to my happiness. Thus he only shows that my de-
sire that everyone should accept and act on the principle of
Egoismwouldbeinconsistent with my desire for my ownmaximum
happiness. Andthis is wholly irrelevant. For we ought to
be testing the claims of Egoismas a categorical imperative;
and, as such, it does not presuppose the existence of a
desire for my own happiness.23
Bruce Auneconcludes:
Here again Kant's argument does not quite work. To be
sound his argument needs the premise that a rational being
(or person to the extent that he is rational) knowsthat,
in the system of natu , he will sometimeneed the unselfish
help of someoneelse. 24
AndHenry Sidgwick points out:
NowI cannot regard this reasoning as strictly cogent. In
the first place, that every man in need wishes for the aid of
 55

others is an empirical proposition which Kant cannot know

! priori. Wecan certainly conceive a man in whomthe spirit
of independence and the distaste for incurring obligations
would be so strong that he would choose to endure any priva-
tions rather than receive aid from others. But even granting
that everyone, in the actual momentof distress, must neces-
sarily wish for the assistance of others; still a strong man,
after balancing the chances of life, may easily think that he
and such as he have more to gain. on the whole, by the general
adoption of the egoistic maxim;benev~ience being likely to
bring them more trouble than profit. .
This sort of objection may seem even stronger in the case of the
first of these two examples, the duty of perfection, where the end in
question is not the actual end of happiness, or help in time of need,
but simply 11all sorts of possible ends. 11
In Chapter Three I will discuss howthis problem is dealt with by
the Practical Contradiction Interpretation. The point of bring it up
here is to show that this problem, certainly a serious one for Kant's
theory, has played an important role in driving people to suppose that
Kant is left with nothing except a terrible consequences view. Unless
we assume something that the person wants or needs, it seems impossible
to say that willing the universalized maximis in any conflict with his
(pure rational} will. Yet insofar as he is purely rational it seems
impossible to attribute any specific end to his will: specific ends or
the need of specific means seem to be attributable to us only empirically.
Sometimesthis is raised even as an objection against the first test
(as in the case of the Hegelian objection described by Acton}, but in
the case of the first test, properly understood, it is easy to get
around. In the case of the first test the maximitself was formulated
for a specific end or requires the use of a specific means, and it is
that end or that means that is unavailable in the world of the universa-
lized maxim: we can attribute the desire that there be promises to any
 56

rational being who undertakes to makea promise without making any

assumptions at all. But the maximsfrom which Kant here generates the
obligatory ends - to neglect one's gifts" "never to help anyone" - do
11

not seem to provide a basis for assigning to the person whoconsiders

them any particular end or need that can be contradicted in the world
of the universalized maxim. In these cases it looks as if we must
assign to the rational being someend or need he or she will have simply
as a rational being. But this looks as if it is to say that we must
already knowsomeobligatory {rationally necessary) ends in order to
generate the obligatory ends, and that of course will not do. Obviously,
this is the familiar Humeanproblem: ends can only be rational or ir-
rational in the light of other, previously given ends.
The one thing that we can surely say about the maximsof neglecting
one's gifts and of indifference is that the consequences of their
universalization would be terrible. I think sometimes people interpre-
ting the test in this way fail to see howstrong it is even in this
form. Whatone is to will is that everyone always neglects their gifts,
that no one ever helps anyone, except indirectly as a means to someother
end. Humansociety might not be possible under these conditions:
certainly, it would be left in a primitive state. Kant's comparison
of the world in which neglect of natural gifts is universal to the
South Sea Islands (G 41/423) is perhaps intended to make this point.
Those who suppose Kant is left with a Terrible ConsequencesInterpreta-
tion are unable to agree that we cannot will such a state of affairs,
but are happy to agree that, granting sane rational decency, we don't.
This is not enough for a foundation in practical reason ..
 57

V. The Teleological Contradiction Interpretation

According to the Teleological Contradiction Interpretation, when
we test our maximby the two contradiction tests under the Formula of
the Lawof Nature, we are to consider whether we could will the univer-
salized maximas a possible teleological law of nature - or as a possible
law in a teleologically organized system of nature. The contradiction
emerges only when the universalized maximis considered as a teleologi-
cal law, or a law for a teleological system. Very often, this view
slides over into another which is closely related to it - namely, that
the contradiction emerges whenan action or instinct is used in a way
that is inconsistent with its natural purpose, or is not used in a way
that its natural purpose calls for. This slide occurs, I think, because
it is sometimesthe only way to find a teleological contradiction in
a universalized maxim; but also because manyof Kant's ownapparent
appeals to teleological considerations seem to be of this nature.
The chief proponent of this view is H.J. Paton, and his version can
be found spelled out in Chapter XVof The Categorical Imperative. 26
Paton argues that using the causal law as the test for the moral law
cannot work.
In its strictest sense Kant usually takes a law of nature
to be a causal law, and it is essential to such a law that it
should have no exceptions: the same cause must always produce
the same effects. Hence it might be thought that if we are
to find a contradiction in a maximwhen it is conceived as a
law of nature, this must be because it would assert that the
same cause could produce different effects.
Kant is right in saying that there can be no arbitrary
exceptions either to a moral law or to a law of nature: but
it is manifestly impossible to find, by this method, breaches
of a moral law. Weassert no breach of a law of nature if
we say that food, which ordinarily causes life, may in special
circumstances - for example, in certain sorts of illnesses -
cause death. Nor need there be any breach of causal law if
 58

self-love, which ordinarily causes life, should in special

circumstances cause death. Indeed we may say generally
that any attempt to make the causal la~ of nature a test
of moral law is foredoomedto failure. 7
Paton thinks that in any case this is an incorrect interpretation, and
that it is clear that Kant always 11
appeals to teleological considera-
tions.1128 He claims that the laws of nature he has in mind are not
11

causa1, but tel eol ogical. 1129 Paton thinks that this is appropriate
to the case,
... for it is a maximof action, and action as such (quite
apart from any moral considerations) is essentially purposive.
Furthermore, we are asked to conceive it primarily as a law
of humannature, even if we are setting it against the back-
ground of nature as a whole; and humannature must be re-
garded as essentially purposive. All this was apparently
so much taken for granted by Kant that he fails
explicitly, and so tends to mislead his readers.
J8
state it

Paton believes that "the test or criterion, but not the essence 1131of
moral action is therefore to be given in terms of a systematic harmony
of purposes both for the species and within the individual. Weare to
test maximsby asking whether 11a will which aimed at a systematic
harmonyof purposes in humannature could consistently will this parti-
cular maximas a law of humannature. 1132
There are a numberof difficulties with Paton's argument here. One,
trivial but still perhaps worth mentioning, concerns one of the bases of
Paton's argument, given in the passage quoted above. It is certainly
true that we are supposed to conceive our universalized maximsas laws
of humannature. Kant says, for example, that we are to imagine the
maxim"implanted in us as by a natural instinct." (G 41/423) But the
fact that we are considering laws of humannature and supposing those
laws to govern humanaction does not necessarily imply that we are con-
sidering teleological laws; Paton is wrong to draw any conclusions from
 59

this fact alone. A teleological theory of the humaninstincts would

be one according to which those instincts themselves had natural pur-
poses - for example, that the purpose of the sexual instincts was to
ensure the reproduction of the species, or the purpose of the instinct
to blink whensomething is near the eye is to ensure its protection.
A complete teleological theory of the instincts would assign purposes
to all of them, and, perhaps, overarching purposes to the system. But
one can certainly imagine a maximas a law of humaninstinct without
any conmitmentto a teleological theory of the instincts. One can ima-
gine that humanbeings inevitably behave a certain way in certain cir-
cumstances without any commitment
to the idea that there is some natural
purpose for which they behave that way. Paton's real difficulty, I
think, is that he cannot see howto get a contradiction in a law of
humaninstinct unless there is somepurpose for the law to contradict,
to thwart.
In fact, the view that Paton describes as "making the causal law
of nature a test of the moral law" and as "foredoomedto failure" seems
to be closer to a genuinely teleological view than Paton's own. The
best evidence that Kant intends a Teleological Contradiction Interpre-
tation in the first set of examples is the suicide case:
His maxim, however, is: For love of myself, I make it my
principle to shorten my life when by a longer duration it
threatens more evil than satisfaction. But it is questionable
whether this principle of self-love could becomea universal
law of nature. One inmediately sees a contradiction in a
system of nature whose law would be to destroy life by the
feeling whose special office is to impel the improvementof
life. (G 39-40/422) .
Paton, however, reads this as the best case for the causal law inter-
pretation:
 60

It is perhaps possible to interpret Kant's discussion of

suicide on this basis. Here the maximis supposed to be
'I will commitsuicide, if life offers me more pain than
pleasure.' This is a principle of self-love, and we assume
empirical knowledgethat self-love ... has the 'detennination'
(Bestimmung)to work for the furtherance of life. Hence if
we universalize this maximof a law of nature, we are suppo-
sing that self-love, which is the cause of life, should in
certain circumstances be the cause of death. Weare, in
short, conceiving a law of nature to admit of arbitrary
exceptions, and so we are falling into a contradiction ..• 33
NowKant offers a general principle of teleological judgment in the
opening lines of the first teleological argument of the Foundations:
"we assume as an axiom that no organ wi11 be found for any purpose
which is not the fittest and best adapted to that purpose." (G 11/395)
This is itself a regulative principle of judgment, and what we would
derive from it would be the assignment of purposes or functions to or-
gans, instincts, and other organic arrangements. Kant uses it to esta-
blish that the attainment of happiness is not the natural purpose of
practical reason - the argument being that since instinct would be a
better guide to happiness than reason is, reason is not the "fittest
and best adapted" thing for that purpose. Wecould get a contradiction,
in accordance with the above principle, if we proposed an organ or in-
stinct for a certain purpose which was not the fittest and best for
that purpose, or if we proposed to use an organ or instinct in a way
that would involve its failure to be the fittest and best for its
purpose. Kant's discussion of the suicide case in the Foundations
strongly suggests this reading: self-love is for the purpose of self-
preservation; the contradiction would then be that in the system of
nature resulting from the universalization of the maximof suicide
from self-love, self-love would clearly not be the fittest and best
adapted to its purpose. The problem would not be, as Paton thinks,
 61

merely that the same cause can produce different effects, but that a
purposive arrangement was clearly not the best possible arrangement to
do the job in question. 34 Self-love would not be the best adapted
instinct for the individual's preservation if it often led to the in-
dividual's demise. Thus the principle of teleological judgment would
be contradicted by the proposed law.
Nowan immediate difficulty with this reading is that the suicide
himself is not supposed to be able to will the teleological system based
on the universalization of his maxim. It is important to rememberthat
the general formulation of the categorical imperative, not just the
second test, is given in terms of what we are able to will. Nowof
course it will be said that the suicide certainly cannot will the teleo-
logical system resulting from the universalization of his maxim, since,
gua teleological system, it has a contradiction in it. But this is a
curiously abstract way to make a case against suicide. The contradic-
tion in the teleological system is, after all, that a mechanismdesigned
for the protection of life is malfunctioning. But the suicide wants
to die. He doesn't want this mechanismto function well in his own
case. Thus if the suicide cannot will the teleological system in ques-
tion because of the contradiction in it, we must simply regard him as
rationally committed to the idea of a well-functioning teleological sys-
tem. His own purposes do not come into it at all. Thus it is like
telling someonethat they must not commit suicide because suicide
(supposing we could prove it) implies a round square. In the promising
case we do not have this odd sense of abstractness, for the man who is
unable to will the universalization the maximof false does have a stake
in the well-functioning of promising - he wants to make a promise. The
 62

last thing in the world the suicide has a stake in is the well function-
ing of the self-preservation instincts. Onemay continue to insist on
believing that in the case of the first test Kant's only point is that
one cannot will a contradictory object. But there are reasons, some
already discussed, not to do that. There are reasons to think that
the agent's own purposes must play a role.
For instance, unless the agent's ownpurposes play a role you
can get the sort of difficulty that Acton described: the Hegelian
criticism that the disappearance of certain practices from the scene is
not obviously any sort of contradiction. Acton was quoting Bradley,
and Bradley, of course, was working with the theoretical contradiction
idea, but we find even Paton running into this difficulty. Paton reads
the teleological interpretation into the promising case by suggesting
that the purpose of promises is to produce trust and so get out of
financial difficulties; 35 false promises do not produce trust and there-

fore universalization makes the purpose of (false) promising impossible.

A false promise is therefore not a good teleological device for the pro-
duction of trust. Paton goes on to remark:
WhatKant says is true enough so far as it goes, but it
does not offer a satisfactory basis for moral judgement unless
we make the further assumption that the keeping of such pro-
mises and the mutual confidence thereby aroused are essential
factors in the systematic harmonyof humanpurposes.36
That is, we have to presuppose that the teleological system needs
promises. Then, as in the suicide case, because we are rationally com-
mitted to the teleological system we cannot will a universal law that
would eliminate promises.
It is easy to see what is happening here. In order to find out
what we can and cannot will as part of a teleological system, as a law
 63

in a teleological system, it is not enough to say that in such a system

a systematic harnionyof purposes would be achieved, and each device would
be best suited to do what it does. The attempt to work with just these
two notions is a laudable attempt to keep the system formal, but in
fact it always turns out that we need some specific purpose, some con-
tent, to generate a real contradiction. No matter howmuch he tries,
Paton's analyses always end up depending on the importation of some
actual purpose - either an actual "purpose of nature'' or an actual pur-
pose associated with a practice, such as the maintenance of trust asso-
ciated with promising. Otherwise it is possible to say that the system
can do without the teleological device because it can do without the
purpose (or, in the case of the theoretical contradiction interpretation,
that there is simply no contradiction in the non-existence of the device
unless we assume certain purposes). Just as Bradley and his friends
complain that the world might not require promises and trust, so the
suicide might reply that the world does not require a self-preservation
instinct to keep people alive unless one supposes it is better that
people be alive. But this is what a suicide undertakes not to suppose.
A great attraction of the teleological view is that it looks at
first glance as if it could resolve the problem that we found in the
case of the theoretical contradiction view - as if it would have some-
thing to say about the problem of natural acts. The reason that it is
not hard to find a contradiction in the universalization of false pro-
mising, promise-breaking, stealing, and the various forms of cheating
is that in these cases we refer to practices which have a purpose as-
signed to them by the conventions in the context of which they exist -
if this purpose is regularly violated, the practices disappear and can-
 64

not be used at all. If everyone broke promises, they would not be

accepted. If everyone cheated on entrance examinations, other criteria
would be found. If everyone stole to acquire things for their ownuse,
no one would really have anything for their ownuse - somearrangement
other than private property might have to be found for the disposition
of the means of action and production. Nowthe teleological view seems
to put even the problematic cases of natural actions on the level of
these cases by assigning to natural acts specific purposes - natural
purposes. The force of teleology is to solidify natural acts or the
use of natural devices {instincts, organs) into "practices" by assigning
them a definite purpose, just the way that promising and property and
entrance examinations have a definite purpose. It turns out, however,
that the movedoes not work. There are two reasons. One is that even
where we can assign a definite purpose to a natural device, it is not
usually possible to show- as it is in the case of conventional prac-
tices - that the abuse of the device spoils it for its assigned purpose.
But the other is that the potential abuser may not care about its as-
signed purpose. It is not his purpose. {Unless we can showthat it is
a purpose that belongs to him as a rational being - but that's a differ-
ent sort of argument). So the Hegelian objection seems to work, here.
In the case of a practice, to spoil the practice for its assigned pur-
pose is to spoil it altogether, and the person who proposes to use the
practice must care about that. Yet the teleological view begins to show
us what we will need to deal with the case of natural acts. Wewill
need to assign them purposes, to regard them as available procedures in
which the agent can be shownto have some interest. The trouble with
bringing in teleological considerations in order to find assigned pur-
 65

poses for natural acts or the use of natural devices such as instincts
and organs is that these purposes may have nothing to do with what the
agent wants or indeed with what any humanbeing wants.

VI. Teleology in Kant1 s Moral Philosophy

Another difficulty that Paton s reading of the test generates is
1

that it seems in somecases to lay Kant open to the charge of depending

upon actual teleological knowledgein the derivation of duties - the
charge that Kant appeals to actual natural purposes. For instance, in
discussing the suicide case, Paton says:
WhenKant considers duties toward oneself, he does not
test maximsby their fitness to produce a systematic harmony
of purposes amongmen if they were to becomeuniversal laws
of nature. He does, however, test them by reference to
hannonyof purpose, a hannonybetween the ends proposed by the
maxim,whenuniversallzed as a law of nature, and what he calls
•purposes of nature.• 7
Again, in the case of the duty of self-cultivation:
Here his teleology is more explicit, and he even puts it
almost in the language of religion; for he says that a
rational being must necessarily will the development of his
powers, not only because they serve, but also - in the second
edition - because they have been given, for all sorts of
possible purposes. This, he holds, remains true, although
there is no contradiction in conceiving (as opposed to will-
ing} a law of nature such that all men live the life of lotus-
eaters. Apart from the theological language, if I conceive
myself as having created menwith all sorts of talents, I
should certainly feel myself to be willing inconsistently
if I willed it to be a law ~f nature that these talents should
never be developed or used. 8
Respondingto this sort of analysis of certain of the examples, Bruce
Auneremarks:
I think it is fair to say that if the use of C2 [the na-
tural law formula] requires us to adopt a teleological view
of nature, then C2 must be considered as obsolete as a teleo-
logical view of nature clearly is at the present time. A
defender of Kant might wish to reply that using C2 does not
require us actually to assume that nature is a teleological
 66

system; it merely requires us to view nature~ if it were 1

such a system. Although this reply no doubt expresses Kant s
actual view of the matter it prompts another charge that
Kant has no way of avoiding - namely, that if we are merely
to view nature as if it were governed by teleological laws,
we could have norational way of settling disputes about
nature's supposed purposes, or about what function a natural
thing is supposed to have. Consider, for example, the supposed
function of self-love. Kant views self-love as if it had the
function of stimul ating the furtherance of 1ife7 But sup-
11

pose I view its function as that of stimulating the further-

ance of a life in which pleasure outweighs pain. Whatcould
Kant possibly say to prove myview erroneous?•... If we are
merely viewing nature as if it operated according to purposes,
there seems to~§ no lTriiiton the variety of purposes we could
credit it with.
This is a common
criticism of Kantian ethics and I think that it
is important to consider the extent to which it is justified. The
Critique of Judgmentand the writings on history leave it unquestionable
that Kant sees an important relation between his teleological view and
his ethical ones: man under moral laws is the purpose of nature teleo-
logically conceived, and the historical writings describe the process
by which nature can be seen as movingthe humanspecies towards the
ultimate achievement of moral goodness. I will comeback to this. What
I want to discuss now is the extent to which teleological considerations
are invoked in the course of arguments in the ethical writings and used
to establish conclusions essential to the ethical theory.
There are five arguments in the ethical writings in which Kant's
reasoning depends upon teleological considerations. One is the argument
about the function of reason in the first section of the Foundations.
That argument is certainly teleological - Kant carefully sets it teleo-
logical basis forth - but that argument is not of course a derivation
of duty. I will discuss this later. Of the other four, two appear in
the Foundations, in connection with the first set of examples. Those
are the duty not to commitsuicide, to avoid a life that has more pain
 67

than pleasure; and the duty of self-cultivation. The teleological re-

marks are:
One immediately sees a contradiction in a system of nature
whose law would be to destroy life by the feeling whose
special office is to impel the improvementof life.
(G 40/422)
For, as a rational being, he necessarily wills that all his
faculties should be developed, inasmuch as they are given
to him for all sorts of possible purposes. (G 41/423)
Whenthe suicide example is treated under the Formula of the End in It-
self, there is no teleological dependence. NowKant says, very much
more forcefully,
If, in order to escape from burdensomecircumstances, he
destroys himself, he uses a person merely as a means to main-
tain a tolerable condition up to the end of his life. Man,
however, is not a thing, and thus not something to be used
merely as a means; he must always be regarded in all his
actions as an end in himself. (G 47/429)
The teleological emphasis does, however, reappear in the treatment of
self-cultivation under the Formula of the End in Itself:
Nowin humanity there are capacities for greater perfection
which belong to the end of nature with respect to humanity
in our own person; to neglect these might perhaps be consis-
tent with the preservation of humanity as an end in itself
but not with the furtherance of that end. (G 48/430)
In the Metaphysics of Morals, both of these duties fall under the rubric
of duties of virtue, although the first is a perfect duty and the second
an imperfect one. The duties of virtue are all presented by Kant in
that work in terms of the idea of Humanityas an end in itself. There.
both of these duties are presented without any reference to teleological
considerations, in terms of what a person owes to humanity in her own
person. The argument against suicide is:
To destroy the subject of morality in one1 s own person is to
root out the existence of morality itself from the world,
so far as this is in one's power; and yet morality is an end
in itself. (DV85/423)
 68

And the duty of self-cultivation exists because:

The power to set an end - any end whatsoever - is the
characteristic of humanity (as distinguished from animality).
Hence there is also bound up with the end of humanity in our
ownperson the rational will, and so the duty, to makeour-
selves worthy of humanity by culture in general, by procuring
or promoting the power to realize all possible ends, so far
as this power is to be found in man himself. (DV51/392)
Earlier Kant has asserted that:
An end is an object of free choice, the thought of which de-
termines the power of choice to an action by which the object
is produced. Every action, therefore, has its end; and since
no one can have an end without himself making the object of
choice into an end, it follows that the adoption of any end
of action whatsoever is an act of freedom on the agent's
part, not an operation of nature. (DV43/384-384)
These passages are from the Introduction. In the body of the text the
connections amongthese ideas are fairly explicit:
Manowes it to himself (as a rational being) not to leave
idle and, as it were, rusting away the natural dispositions
and powers that his reason can in any way use. Even supposing
that the native scope of his powers is sufficient for his
natural needs, his reason must first show him, by principles,
that this meager scope of his powers is sufficient; for, as
a being who is able to set ends (to make objects 1nto his
ends), he is indebted for the use of his powers not merely
to natural instinct but rather to the freedom by which he
determines this scope. Hencethe ground on which man should
develop his powers is not regard for the advantage that can
be gained by cultivating them;•.. Rather it is a command of
morally practical reason and a duty of man to himself to
cultivate his powers... and to be, from a pragmatic point of
view a man equal to the end of his existence. (DV111/444-
445)
What I have said so far is that the two cases of teleological de-
pendence in the Foundations are treated without such dependence, in
terms of the formula of the End in Itself, in the Metaphysics of Morals.
Certainly this does not establish that a teleological law is not what
Kant has in mind in the Lawof Nature formulation. It does show that
the teleological references in the Foundations have a specific location:
they occur in the cases of duties to oneself as presented under the
 69

Formula of the Lawof Nature. Whenthey are presented in tenns of the

Fonnula of the End in Itself there is no need for any teleological ap-
peal - and this is the way Kant eventually presents them. Of course if
the two fonnulations are equivalent there must be someway to argue
for them in terms of the Fonnula of the Lawof Nature, so this is where
the difficulty in these two cases lies.
The two other cases are the two that appear in the Metaphysics of.
Morals. These are the cases of carnal self-defilement and of lying.
The teleological argument against carnal self-defilement is:
Just as the natural function of the love of life is to
preserve the individual, so the natural function of sexual
love is to preserve the species: in other words, both of
these are natural purposes. By a natural purpose I mean such
a connection of cause with an effect that, without attributing
intelligence to the cause, we must yet conceive it by analogy
with an intelligent cause and so as if it produced the ef-
fect purposefully. The question now is whether the agent's
use of his sexual power comes under a limiting law of duty
with respect to his own person, or whether he can, without
violating a duty to himself, use his sexual power for mere
animal pleasure, without regard for its purpose. (DV87/424)
The second sentence of that passage that is often appealled to as evi-
dence that Kant believed that the sort of teleological thinking whose
inevitability and use he discusses in the Critique of Judgment could
be used in moral decision as if it were real information. In answering
the question he has raised, Kant distinguishes two kinds of cases. One
is where the person acts contrary to the end of nature; the other where
the person acts without regard to the end of nature. They are treated
differently. Kant forbids us to call of the first of these two vices
by its real name; but it occurs when a person's lust is aroused
••. not by its real object, but by his imagination of this ob-
ject, and so in a way contrary to the purpose of the desire,
since he himself creates its object. For in this way the
imagination brings forth an appetite contrary to nature's
purpose, and indeed an appetite that is still more important
 70

than love of life itself, since it aims at the preservation

of the whole species and not only of the individual. (DV87-
88/424-425)
The second case, which is not "contrary" but only "without regard" (DV
89/426) to nature's purpose, is simply sex without a reproductive aim.
Kant is very muchless certain of the wrongness of that; placing it
amongthe casuistical questions, and warning us to avoid purism and
pedantry in the observance of duty. (Kant suggests here that this is
only all right within marriage, and the reason for that given in the
Lectures on Ethics is that of avoiding conditions where one of the par-
ties is taken advantage of by, and so surrenders self-respect to, the
other. 40) In either case:
- The ground of proof, of course, is that man surrenders his
personality (throwing it away) by using himself merely as
a means to satisfy his animal instincts. But this does not
explain the high degree of the violation of humanity in one's
ownperson involved in unnatural vice .•.
... unnatural vice, which is complete abandonmentof oneself
to animal inclination, makesman not only an object of enjoy-
ment but, still further, an unnatural thing, i.e. a loathsome
object, and so deprives him of all reverence for himself.
(DV88-89/425)
The lying case has notable similarities:
A man who himself does not believe what he tells another ••.
has even less worth than if he were a mere thing. For a
thing, as something real and given, has the property of being
serviceable and another person can make someuse of it. But
the manwhoconmunicates his thoughts to someonein words
which yet (intentionally) contain the contrary of what he
thinks on the subject has a purpose directly opposed to the
natural purposiveness of the power of communicatingone's
thoughts and therefore renounces his personality and makes
himself a mere deceptive appearance of a man, not man him-
self.... (DV93/429)
Manas a moral being (homonoumenon)cannot use his
natural being (homophenomenon)as a mere means (a speaking
machine), as if.....,.-r-were
not bound to its intrinsic end (the
communicationof thought) .... (DV94/430)
Again, too, Kant declares that the violation makes the person contemp-
11
 71

tible in his owneyes". (DV93/430)

It is not my intention, with any of these cases, to try to read out
the teleological argument, which is certainly there. They are, on the
contrary, powerful evidence that Kant sometimesused the idea of possible
or even actual teleological relations in the derivations of duties. I
think, however, that these are the only cases in which Kant really seems
to be dependent on teleological importations. The Foundations cases can
be given an account, at least in tenns of the Fonnula of the End in It-
self, in which nothing is appealled to except the implications of making
humanity (freedom, the power to set an end) one's end. And that is sup-
posed to be the guiding idea of the two Doctrine of Virtue arguments
as well: the point is not just that the person uses his faculties con-
trary to their natural purposes, but that it follows (somehow)from the
contrary use that the person is using his humanity as a means and not
at the same time as an end. While it might seem clear that the person
who engages in carnal self-defilement or who lies is using himself
as a means (a speaking machine), that is not wrong unless it is esta-
blished that he is using himself merely as a means, not also at the
same time as an end. Kant's idea here seems to be, in a rough way,
that using oneself as this sort of means automatically implies not
treating oneself as an end. In the first case he says one is a loathe-
someobject, in the second case he says one is not even serviceable.
In any case, the use of teleology in the Metaphysics of Morals cases,
though troublesome enough, is not quite of the sort that Paton and Aune
are concerned about in their different ways. Teleological considera-
tions in these two cases are used to establish that the person is not
treating herself as an end, not doing justice to her humanity, but
 72

teleological laws are not being used as models for moral laws. Thus,
there are two kinds of teleological reference: in explaining the two
duties to oneself in the Foundations in terms of the Formula of the
Lawof Nature; and in explaining whycertain kinds of behavior involve
a failure to treat oneself as an end in the presentations in the
Metaphysics of Morals.
Nowin order to understand the proper place of teleological consi-
derations in the ethical writings correctly, we must turn our attention
to the initial teleological argument in the Foundations. That argument
appears almost at the very beginning of the book - immediately after
the unique status of the good will as the one unconditionally good
thing in the world or even beyond it has been declared. I am going to
quote it at length:
But there is something so strange in this idea of the abso-
lute worth of the will alone, in which no account is taken of
any use, that, notwithstanding the agreement even of common
sense, the suspicion must arise that perhaps only high-flown
fancy is its hidden basis, and that we may have misunderstood
the purpose of nature in its appointment of reason~ the ruler
of our will. Weshall therefore examine this idea from this
point of view.
In the natural constitution of an organized being, i.e.,
one suitably adapted to life, we assume as an axiom that no
organ will be found for any purpose which is not the fittest
and best adapted to that purpose. Nowif its preservation,
its welfare - in a word, its happiness - were the real end of
nature in a being having reason and will, then nature would
have hit upon a very poor arrangement in appointing the reason
of the creature to be the executor of this purpose. For all
the actions which the creature has to perform with this inten-
tion, and the entire rule of its conduct, would be dictated
muchmore exactly by instinct, and that end would be far more
certainly attained by instinct than it ever could be by
reason. And if, over and above this, reason should have been
granted to the favored creature, it would have served only to
let it contemplate the happy constitution of its nature, to
admire it, to rejoice in it, and to be grateful for it to
its beneficent cause. But reason would not have been given
in order that the being should subject its faculty of desire
to that weak and delusive guidance and to meddle with the
purpose of nature. l!!..! word, nature would have taken care
 73

that reason did not break forth into practical use nor have
tlie'"""presumpt1on,~thl"'ts weak ins,ght, to thinkout for
itself the plan of happiness and the means of attaining it.
Nature would have taken~ not only the choice of ends but
also that of means, and with wise foresight she would have
entrusted both to instinct alone. (G 10-11/394-395; my em-
phases) ---
Kant goes on from here to conclude that the function of reason must be
to produce a will good in itself, not as the sole and complete good
but as the condition of all other goods. (G 11-12/395-396)
It is of the first importance to understand the very clear contrast
that Kant is pointing out to us in this passage. Actions - the choice
of ends and the employmentof means - might have been entirely deter-
mined by instinct. In that case, nature would have chosen for us both
ends and means - just as nature teaches the other animals both to want
an appropriate dwelling place and howto build one. Instead, we have
been provided with a will governed by reason in its practical employment:
we choose both ends and means for ourselves, using reason. Kant even
fantasizes that nature might have given us a theoretical reason without
a practical use; then at most, reason would enable us to admire our own
instinctual constitution and perhaps to be grateful for it. 41 But this
is not the case: we have been given practical reason to govern our
choices of means and ends, and it takes the place of instinct.
I do not mean that Kant views usas having practical reason in
place of instincts, as if we did not have any instincts. Kant speaks,
for instance, of
... the commonrun of menwhoare better guided by mere natural
instinct and who do not permit their reason muchinfluence on
their conduct. (G 12/396)
These people may be envied by those who have tried to seek happiness
through a cultivated reason and become11misologists 11
• And I of course
 74

do not mean to deny that Kant agrees that we have instincts such as self-
preservation and the sexual instincts. Yet Kant thinks that we are not
instinctual in an extremely important sense - we are not under the con-
trol of our instincts. If we follow them, we follow them deliberately;
we allow them control - notice the "do not permit" in the passage above,
which Kant uses advisedly. Andsomeof our instincts, such as those of
self-preservation or sex, are certainly the sort of thing that, when
nature is viewed teleologically, are supposed to have natural purposes.
But it would be odd to suppose that morality could be understood in
general in terms of not contradicting these natural purposes, for it
would be odd to suppose that nature had released us from the control of
instinct and placed us under the governmentof practical reason merely
so that we could return to the control of our instincts voluntarily.
The idea of the developmentof the good will as a purpose of nature
reappears in two important contexts, as I have already mentioned. The
first is in the Critique of Judgment, where the argument goes like this.
Whenwe conceive nature as a teleological system - not just as something
pervaded with arrangements that seem intelligently designed for various
purposes, but something systematically ordered with everything for the
sake of something else - then we must suppose it to have a final purpose.
Nowin the Critique of JudgmentKant makes it clear that whenwe conceive
nature in this fully teleological way, as a system ordered to a final
purpose, that final purpose must be humanity as the subject of morality,
as the unconditioned end in itself:
... man is the final purpose of creation, since without him
the chain of mutually subordinated purposes would not be com-
plete as regards its ground. Only in man, and only in him as
subject of morality, do we meet with unconditioned legislation
in respect of purposes, which therefore alone renders him
 75

capable of being a final purpose, to which the whole of

nature is teleologically subordinated. (C3 286)
I will give the argument for this conclusion later on (Chapter Five).
For now, the important point is that the production of the good will
is not only the natural purpose of practical reason, as in the Founda-
tions argument, but the final purpose of the whole \...arld, and so the
thing around which a teleological system in general is centered.
These arguments form the basis of Kant's teleological historical
writings, in which he describes hownature is arranged so as to force
humanity to use reason, to develop talents and powers, and eventually
to realize freedom and, along with it, moral goodness. As in the argu-
ment in the Foundations, Kant often says that in this process humanbe-
ings are not likely to be made very happy through the use of reason,
and takes that to be evidence that it is not our happiness that is
nature's chief care. And in "Conjectural Beginning of HumanHistory",
Kant identifies as a source of this unhappiness a conflict between
two different kinds of natural purposes for humanbeings. Weare unfit
for happiness because
... of this conflict betweenman's striving towards his moral
destiny, on the one hand, and, on the other, his unalterable
subjection to laws fit for the uncivilized and animal state.
(OH:CBHH 61n/116n)
In the course of the note in which that remark appears, Kant tells us
that manners and morals, and the aim of nature, inevitably come to in-
11

terfere with each other":

For surely nature has not endowedliving beings with instincts
and capacities in order that they should fight and suppress
them. The disposition in question, then, [the appearance of
the sexual urge in boys too young to maintain families in
civilization] did not intend the civilized state, but merely
the preservation of manas an animal species ..•.
This, then, is one example to prove that nature has given
us two different dispositions for two different purposes,
 76

the one for man as an animal, the other for him as a moral
species. (OH:CBHH 6l-62n/ll6n}
The conclusion to the passage, in the regular text, is important and
revealing:
The real trouble is that, on the one hand, culture progres-
sively interferes with its [natural im~ulse's] natural func-
tion, by altering the conditions to which it was suited; while
on the other hand, natural impulse interferes with culture
until such time as finally art will be strong and perfect
enough to becomea second nature. This indeed is the ultimate
moral end of the humanspecies. (OH:CBHH 62-63/117)
At the very least these passages should serve to establish the complexity
of Kant's views about the "natural purposes" that are associated with
instincts. Thoughwe would be happier under instinctual control and the
guidance of nature's purposes, a return to such guidance is not being
advocated. Indeed, it is said to be impossible. Ours is another task.
Andour natural purpose as a moral species can come into clear conflict
with the natural purposes associated with our animal nature and in-
stincts. The prohibitions against suicide and carnal self-defilement
in the Doctrine of Virtue are classified as "duties to oneself as an
animal (natural) and at the same time moral being. 11
(DV82/420) It
is clearly stated that it is the preservation of the individual and the
species~ the bearers of morality, this other natural purpose, that is
at stake. Rememberthe argument against suicide: suicide is "rooting
out the existence of morality itself from the world." The appeal to
natural purposes cannot be straightforwardly: nature's purposes for us
are such-and-such, and so that is what we must do; or nature's purpose
for this instinct is such-and-such, and so that is howwe must use it.
To live according to nature's purposes in that sense is to live accor-
ding to instinct - but we have been given practical reason, because we
have a special task, quite different from living according to instinct.
 77

Wehave been released from the control of our instincts. This release
makes it possible for us to use our instincts in ways that animals
cannot; 42 and indeed to 11abuse11 them, but it is in cases where this in-
terferes with nature's other purpose for us - the existence of morality
and freedom in the world - that the violation occurs. And it is also
this other purpose of nature, in the form of the freedom to set ends,
that is associated with the cultivation of talents in the teleological
history writings. In these cases it is our "natural purpose" as the
moral species that is at stake, and the natural purposes that apply to
us as an animal species are appealed to only whenthe existence of
morality seems to depend upon them. This is the teleological argument
that informs Kant's ethical writings.

VII. Practical Reason and Instinct

The teleological argument at the beginning of the Foundations has
another implication which is extremely important for an understanding
of Kant1 s moral philosophy. This is the fact that it is instinct,
not passion or desire, which is the alternative to practical reason as
the source and guide of action. This maynot seem important - one may
be inclined to say that there is no doubt our desires and passions bear
somecausal relation to our instincts, that they are supplied by in-
stincts or developed out of them. Humesometimeswill speak of instincts
and passions almost interchangably. Yet the standard philosophical pic-
ture nowadaysof these alternatives looks like this: either reason
dictates an end (and, with Hume,we puzzle over howthis might be) or
desire (passion) dictates an end. An end is selected for its rational-
requiredness or it is selected because we want it.
 78

In either case, we suppose there is no trouble in seeing how reason

might dictate the means.
But this is not how Kant sees the alternatives at all. Suppose
that, in Humeanfashion, we trace a piece of practical reasoning back
to the end which is the purpose of the action, the thing that is
chosen for its own sake, that is not a means to some other end. For
Hume,the reasoning is over: our relation to this thing must be one of
desiring it or wanting it, for what else could it be? For Kant there
is a further question. Suppose this end is a thing that I desire; we
can still demandan account of my having chosen the thing that I desire.
There are two ways of choosing the thing I desire. If I am free, and I
choose the thing I desire, that is because I have made it~ principle
to take as my end the thing I desire. Then there are some further ques-
tions to ask about my having chosen this principle - further questions
that Kant will address in Religion Within the Limits of Reason Alone.
(See Chapter 7.) On the other hand, if I am determined by nature to
choose the thing I desire, then I am simply under the control of
instinct:
Since the conception of an object commensurateto our power
incites in the will an impulse according to the natural charac-
teristic of our person, this impulse belongs to the nature of
the subject (either to the sensibility, i.e., inclination and
taste, or to understanding and reason, which faculties, ac-
cording to the particular constitution of their nature, take
pleasure in exercising themselves on an object). It follows
that it would really be nature that would give the law.
{G 63/444)4 3

Nowsince an object can be chosen for the sake of its appeal to desire
either under the guidance of the free rational will or under the guidance
of instinct, it follows that it is instinct, not passion or desire,
that is the alternative to practical reason. Choosing the thing you
 79

most desire can happen either way: the real question is whether you
are so constituted by nature that you do this automatically.
A further consequence of the fact that Kant sets up the alternative
accounts of practice in terms of reason and instinct rather than in
tenns of reason and desire or passion is important. On the usual view,
according to which the issue is about reason and desire as the detenni-
nants of ends, it is more or less taken for granted that the means are
determined by reason: the question is entirely about the status of
ends. One can believe that the means can be determined by reason even
if one does not believe in practical reason, it is supposed, because
this is purely a matter of applied theoretical reason, of technique.
One simply observes cause-and-effect laws and puts them to work. Now
this is false already, in one sense. This sense is implied by Kant's
fantasy, in the teleological argument in the Foundations, about the
creature who has only theoretical reason admiring its purposive in-
stincts: nature takes over the choice both of ends and of means. In
fact, if we are to imagine a being whocan select even the means to
preordained ends, we must credit this being with a certain psychological
capacity: the capacity to transmit motivational energy from end to
means. Kant's principle that whoeverwills the ends wills the necessary
means is analytic, but only analytic of a being who has a will, as is
shownby the way Kant proves that principle. In an animal that is com-
pletely under the domination of instinct (like an insect), we must ima-
gine that motivational responses follow only absolutely preordained
paths: the creature's condition stimulates an end which stimulates a
certain means taken so automatically, so directly, that its behavior
could be described without reference to an end at all. But where the
 80

means are selected with the aid of reason, and of course alternative
strategies are available, we must imagine that the creature is able
to transmit motivational energy along any paths which are or seem to
be laid downby reason. If we change Kant's fantasy and the creature
who has only theoretical reason were not designed by a beneficent nature,
we could almost imagine it having an end and perceiving a means to that
end better than the one instinct taught it to take: it would still take
the path charted for it by instinct, for it could not be motivated to
take any other - its motives are rigidly channeled.
Once we recognize this - once we see that in order to determine
even means by reason a creature must be able to channel motivational
energy along paths laid downby rational calculation - we can see our
way clear to the beginning of an account of howit is we cometo adopt
different ends under the influence of reason. There is no need for us
to go to the labor of constructing this account, for Kant has provided
it for us. I am going to quote a long passage taken from Conjectural
Beginning of HumanHistory where it appears.
In the beginning, the novice must have been guided by in-
stinct alone, that voice of Godwhich is obeyed by all animals.
This permitted some things to be used for nourishment, while
forbidding others (Genesis 3:2,3). Here it is not necessary
to assume a special instinct which is nowlost. It could
simply have been the sense of smell, plus its affinity with
the organ of taste and the well-known relation of the latter
to the organs of digestion; in short an ability, perceivable
even now, to sense, prior to the consumption of a certain
foodstuff, whether or not it is fit for consumption.
... So long as inexperienced man obeyed this call of nature
all was well with him. But soon reason began to stir. A
sense different from that to which instinct was tied - the
sense, say, of sight - presented other food than that normal-
ly consumedas similar to it; and reason, instituting a com-
parison, sought to enlarge its knowledgeof foodstuffs beyond
the bounds of instinctual knowledge(3:6). This experiment
might, with good luck, have ended well, even though instinct
 81

did not advise it, so long as it was at least not contrary

to instinct. But reason has this peculiarity that, aided
by the imagination, it can create artificial desires which
are not only unsupported by natural instinct but actually
contrary to it. These desires, in the beginning called con-
cupiscence, gradually generate a whole host of unnecessary
and indeed unnatural inclinations called luxuriousness. The
original occasion for deserting natural instinct may have
been trifling. But this was man's first attempt to become
conscious of his reason as a powerwhich can extend itself
beyond the limits to which all animals are confined. As such
its effect was very important and indeed decisive for his
future way of life. Thus the occasion may have been merely
the external appearance of a fruit which tempted because of
its similarity to tasty fruits of which man had already par-
taken. In addition there may have been the example of an ani-
mal which consumedit because, for it, it was naturally fit
for consumption, while on the contrary, being hannful for man,
it was consequently resisted by man's instinct. Even so, this
was a sufficient occasion for reason to do violence to the
voice of nature (3:1) and, its protest notwithstanding, to
make the first attempt at a free choice; an attempt which, be-
ing the first, probably did not have the expected result.
But however insignificant the damagedone, it sufficed to open
man's eyes (3:7). He discovered in himself a power of choosing
for himself a way of life, of not being bound without alter-
native to a single way, like the animals. Perhaps the dis-
covery of this advantage created a momentof delight. But
of necessity, anxiety and alann as to howhe was to deal with
this newly discovered power quickly followed; for man was a
being whodid not yet knoweither the secret properties or
the remote effects of anything. He stood, as it were, at the
brink of an abyss. Until that momentinstinct had directed
him towards specific objects of desire. But from these there
nowopened up an infinity of such objects, and he did not yet
knowhowto choose between them. On the other hand, it
was impossible for him to return to the state of servitude
(i.e. subjection to instinct) from the state of freedom, once
he had tasted the latter. {OH:CBHH 55-56/111-112)
Those whowish to remain loyal to the usual empiricist model will want
to say that this is not a case of choosing a new end, but of choosing
a newmeans to an old end: satisfying hunger. But this is not what
Kant says: Kant says that comparing the objects of our desires to other
.
objects that resemble them can create new desires for these new objects.
Eve does not eat the apple because she is hungry. If the energy of de-
sire can follow any path of rational association, there is no limit,
 82

really,on what we might cometo desire - for its own sake: thus out of
this, we eventually develop the specifically humancapacity to choose
a "wayof life".
Wemight describe this by saying that libido unchained from the
rigid paths laid downby instinct can cathect anything with which the
original object can be associated rationally. Perhaps it will seem
inappropriate to use Freudian terms to describe Kant's views, but on the
page following the long passage I have quoted, Kant goes on to describe
howthe capacity for passions such as romantic love and the sense of the
beautiful arise from repressions (refusals) of the sexual instincts. To
me it seems Freud's and Kant's views on the nature of the specifically
humanare very close: both emphasize the flexibility of the desidera-
tive energies and the resulting humancapacity to take and even to
invent new objects of concern. 44 On a Humeanmodel, the only way to
account for the extraordinary variety of things that humanbeings desire
and care about, and for the variety found in humanways of life, is to
regard most of the things we care about as gratifications of an over-
arching desire for pleasure. Everything except our desire for the ob-
jects of original animal instincts must be accounted for this way. This
is because on this model passions and their objects are ineradicably
linked: a desire just is a desire for a particular end. Thus the energy
of that desire can be transmitted only to the means to that end: There
is no room in this psychological picture for the generation of desires
for new ends, so it turns out that everything not desired in a primitive
natural condition must fall under the rubric of things desired for the
sake of their pleasantness. Of course, this problem (assuming it is a
problem} only arises if one thinks of pleasantness as an end to which
 83

other things are a means. Mill, who thinks of pleasantness as a

feature of the things we desire as ends (desiring a thing for its own
sake and finding it pleasant are the same) gives an account of the gen-
eration of humandesires that resembles Kant's. After describing how
things desired as means and then found pleasant come to be desired for
their own sake, Mill renarks:
Life would be a poor thing, very ill provided with sources
of happiness, if there were not this original provision of
nature, by which things originally indifferent, but conducive
to, or otherwise associated with, the satisfaction of our
primitive desires, becomein themselves sources of pleasure
more valuable than the primitive pleasures, both in pennanancy,
in the space of humanexistence tgat they are capable of
covering, and even in intensity.
Howclose this view is to Kant's (or Freud's) depends upon howmuch
weight one can suppose Mill puts on "otherwise associated with. 11
It
is worth adding that once we begin to look at things from this perspec-
tive it becomesutterly mysterious whyHumethought that the only role
of reason in action was the detennination of means. After all, the
means/end relation is only one of the associative relations along which
thought travels in Hume's account; resemblance, to which Kant appeals
in the argument in Conjectural Beginning of HumanHistory, is another.
Whydid Humethink that theoretical association could travel along
all of these paths, but motivational energy only along one? There seems
to be no reason for making this sort of distinction between the theore-
tical and the practical. Even within the limits of Hume's theory, it
is arbitrary from the start.
The conclusion that is important for Kant's moral philosophy is
this: that reason enables humanbeings not only to find newmeans to
the ends that nature or passion directs, but to find new ends. When
 84

Kant says that the power to set an end is the characteristic of humanity
(DV51/392; G56/437), this is what he means. The specific characteris-
tic of a humanbeing is the power to take something as one's purpose,
to choose it, to make its realization one's project. Wehave this power
because we are released from the control of instinct; and it will turn
out that it is because we have this power that we are under moral laws.
In tenns of the teleological arguments, this will mean that nature's
purpose for us is to set us free from the domination of nature itself.
In terms of the moral arguments, this will turn out to be the source
of our autonomyand so of our obligation. These are points that I will
be arguing for throughout this thesis. Until it is understood that Kant
was not working with the empiricist model of reason and passion, it will
not be possible to get his arguments off the ground.

VIII. Conclusion
The difficulties with the interpretations of the Formulaof Univer-
sal Lawand the Formulaof the Lawof Nature that I have examinedin
this chapter for the most part resolve into a single problem. Nomatter
howone tries to interpret universalization, one can get no contradic-
tion in the universalization of a practical maximunless one can show
that in the world of the universalized maximsome purpose will be thwar-
ted, or some procedure unusable, or someend inaccessible; and that, at
the same time, the achievement of the purpose, the use of the procedure,
or the accessibility of the end is one that there is reason to want.
The universalization of a practical maximcan be shownto lead to the
thwarting of purposes and the destruction of the means that provide our
access to certain ends. But there will never be any contradiction in
 85

that unless the ends are themselves somehownecessary.

I do not, however, offer this as a reason to despair of making
the system fonnal. The system can be fonnal, and yet appeal to the
existence of actual ends and purposes, just as long as they are purposes
which we can assume to belong to the person who is considering acting
on a certain maximwithout adding any infonnation to the case. That the
person is considering acting on the maximin question and is rational
are built into any case under consideration. Purposes attributed to the
agent on the basis of these two facts alone do not in any way violate
the formality of the system. This means that so long as the purpose
is one that we can attribute to the agent on the basis of the maxim
(the purpose in the maxim, or the use of the procedure in the maxim),
or one that we can attribute to the agent simply on the basis of ra-
tionality, we can produce the sort of contradiction that is necessary
here: a contradiction in the will of a rational being who attempts to
will a certain kind of world - the world of the universalized maxim-
in which his or her ownpurpose must be thwarted. This is a practical
contradiction.
 86

NOTES

1. The scare quotes are there because the argl.DTient

by which Kant
eventually establishes the categorical imperative, in the Critique
of Practical Reason, is not by his lights a deduction. (C2 48-49/
47-48) Each of the three arguments I mention is followed or pre-
ceded by a reminder that it is not a proof or deduction. See
G 38/420; G 50/431; and G 64/445.
2. Lewis White Beck provides such a reconstruction in "Apodictic
Imperatives". See pp. 146-147 in Wolff.
3. David Hume,Treatise of HumanNature, pp. 477ff.
4. W.O. Ross, The Right and The Good, pp. 5-6.
5. I do not claim that Kant makes any terminological distinction
regularly corresponding to the distinction between purpose and
motive that I have described here, although it is possible that
a careful reading would show that he does.
6. Read this way, the argument for the Formula of Universal Lawis
very close to the argument for the Formula of Autonomy,where again
the issue is howwe are 11bound11 to a maxim,what brings us to
adopt it. See Chapter 7.
7. For these thoughts see Cl 485-495; A567/8595-A583/B611.For the
comparison to Plato, see Cl 486; A568/B595.
8. John Stuart Mill, Utilitarianism, p. 254 in Warnock.
9. John Stuart Mill, Utilitarianism, p. 308 in Warnock. It is neces-
sary to point out that in this passage Mill is a little unfair.
If the "rule of utter selfishness 11 is the maximrejected in Kant's
fourth example, it is in any case only one of those that Kant says
cannot be willed and therefore not even claimed to be one that
cannot be thought, as Mill's remark suggests.
10. In "Kant•s Examplesof the Categorical Imperati ve11 , J. Kempmakes
the following argument based on Kant's terminology:
What Kant actually says, however, is not that the maximof
making false promises could not exist as a universal law of
nature for very long... but that it could not exist at all as
a universal law of nature without contradiction. And if he
really means this, it is clear that the relation designated
by the verb machenmust be one of logical or quasi-logical,
not causal consequence. The argument has the effect of a
 87

reductio ad absurdum. If,~ imhossible, there were a uni-

versal law to this effect, then t ere would not be and would
never have been any promises (the 11if ... then 11 indicating an
entailment-relation); but the statement that there is such a
law of nature also entails that there are promises. Hence it
has contradictory implications - i.e. it is self-contradic-
tory.
It is worth noticing that Kant is careful to distinguish
the concept of self-contradiction from that of being self-
defeating or self-destructive. In the passages of the
Grundlegungunder discussion, whenever a putative law of
nature is rejected as being inconsistent with itself, Kant
says, (and, I have argued, means) that it is self-contradic-
tory, using always the phrase sich widersprechen. But when
he wishes to say, of a putative universal law of morality,
that it would be inconsistent with itself, he uses different
terminology - such a law would annihilate itself, sich selbst
vernichten or sich selbst aufreiben. (in Wolff, pp.238-239).
The difficulty with Kemp'sview is that Kant uses the latter
phrases in connection with the example of the manwho considers
keeping a deposit whose owner has died (see Criti ue of Practical
Reason (27/27) and the discussion of this example below ; and the
former phrases in connection with the false promising example.
But the false promising example and the deposits example have
exactly the same structure and admit of exactly the same sort of
argument.
11. Paul Dietrichson, "Kant's Criteria of Universalizability 11
, p. 184.
12. Uietrichson, p. 187.
13. Dietrichson, P: 188. Dietrichson's example here is in any case
not a maxim, since a child's being small is not a recognizable
reason for getting rid of it, unless something more is supposed -
such as that small children are likely to be unhealthy or unplea-
sant or unprofitable or some such thing. A maxim, as I mentioned
earlier and will argue later, must reveal a purposive structure
in the action in order to be susceptible to Kant's test.
14. Dietrichson, p. 188ff.
15. A negative duty could not in any case be a wide duty; only an end
can generate such a duty. If allowing children to live or keeping
them alive were a wide duty, we would only have to let some of
them live and it would be up to us to decide which ones. Obviously
we are derailed here.
16. John Rawls, "TwoConcepts of Rules", p. 3-32 in Gorovitz.
17. Rawls , p. 189.
18. Rawls, p. 190.
19. H.B. Acton, Kant's Moral Philosophy, pp. 24-25. Acton is quoting
 88

from Bradley in Ethical Studies. A curious thing about the objec-

tion, although not relevant here, is that Kant had a great deal to
say about property - the first part of the Metaphysical Elements
of Justice is about why there should be, in some sense, property.
In any case, while Bradley certainly seems to attribute such a
colim1entto Hegel, it is not very obvious from the passage Bradley
cites in Hegel's Phenomenologythat this is what is on Hegel's mind.
The deposit argument appears in the Phenomenologyin a section
called Reason as Testing Laws"p. 262. Hegel says:
11

Suppose something has been entrusted to me; it is the property

of someoneelse and I acknowledgethis because it is so, and
I keep myself unfalteringly in this relationship. If!
should keep for myself what is entrusted to me, then according
to the principle I follow in testing laws, which is a tau-
tology, I am not in the least guilty of a contradiction; for
then I no longer look upon it as the property of someoneelse:
to hold on to something which I do not regard as belonging to
someoneelse is perfectly consistent. Alteration of the
of view is not contradiction; for what we are concerned
.EQ.!..!!1
W'itn-is not the point of view, but the object and content,
which ought not to be self-contradictory. Just as I can -
as I do when I give something away - alter the view that it
is my property into the view that it belongs to someoneelse,
without becomingguilty of a contradiction, so I can equally
pursue the reverse course. It is not, therefore, because I
find something is not self-contradictory that it is right; on
the contrary, it is right because it is what is right. That
something is the property of another; this is fundamental.
I have notto argue about it, or hunt around for or entertain
thoughts, connections, aspects of various kinds; I have to
think neither of making laws nor of testing them. All such
thinking on mypart would upset that relation, since, if I
liked, I could in fact just as well make the opposite conform
to my indeterminate tautological knowledgeand make that the
law. But whether this or the opposite determination---:rs-the
right, that is determined in and for itself. I could make
whichever of them I liked the law7a°nd just as well neither
of them, and as soon as I start to test them I have already
begun to tread an unethical path.
This admits of Bradley's reading, to be sure; but Hegel may also
be more interested in what can happen to institutions such as
property once the autonomouspoint of view is taken up and we begin
to treat what from an earlier point of view would have seemed "an
unethical path". I am not saying that is Hegel's view, only that
this passage, which is clearly one of the sources of attributing
the objection to Hegel, does not unambiguouslysupport such a read-
ing. A passage that supports Bradley's reading does, however,
appear in the Philosophy of Right: and in this one the reference
to Kant is made openly:
Kant's further formulation, the possibility of visualizing
an action as a universal maxim, does lead to the more concrete
visualization of a situation, but in itself it contains no
principle beyond abstract identity and the 'absence of contra-
diction' already mentioned.
The absence of property contains in itself just as little
 89

contradiction as the non-existence of this or that nation,

family, &c., or the death of the whole humanrace. But if it
is already established on other grounds and presupposed that
property and humanlife are to exist and be respected, then
indeed it is a contradiction to committheft or murder; a
contradiction must be a contradiction of something, i.e. of
some content presupposed from the start as a fixed principle.
It is to a principle of that kind alone, therefore, that an
action can be related either by correspondence or contradic-
tion. But if duty is to be willed simply for duty's sake and
not for the sake of some content, it is only a fonnal identity
whose nature is to exclude all content and specification.
Philosophy of Right, p. 90. It is obvious that this passage simply
puts together the two challenges to practical reason already found
in Hume: the emptiness of duty for duty's sake and the impossibi-
lity of generating practical contradiction unless an end is presup-
posed. The reference to the death of the whole humanrace reads
like a reference to Hume's declaration that "'Tis not contrary to
reason to prefer the destruction of the whole world to the scrat-
ching of my finger." (Treatise, p. 416) Since these points are
already found in Hume,Bradley's loyal declaration that Kant's
system "has been annihilated by Hegel's criticism" (Ethical Studies,
p. 148n} seems a little misplaced.
20. Acton, p. 29
21. Marcus Singer, Generalization in Ethics.
22. Singer, p. 275.
23. C.D. Broad, Five Types of Ethical Theory, pp. 130-131.
24. Bruce Aune, Kant's Theory of Morals, p. 57.
25. Henry Sidgwick, The Methodsof Ethics, p. 389.
26. H.J. Paton, The Categorical Imperative, pp. 148-157.
27. Paton, pp. 148-149.
28. Paton, p. 149.
29. Paton, p. 148.
30. Paton, p. 151.
31. Paton, p. 151.
32. Paton , p. 151.
33. Paton, p. 148.
34. Paton comes around to this sort of view in some cases, for instance
in his discussion of promises, mentioned below. Seep. 153.
 90
35. Paton, p. 152.
36. Paton, p. 153.
37. Paton, p. 154.
38. Paton, p. 155.
39. Aune, p. 60. Later in the book, when summingup his discussion of
the Foundations, Aune remarks 11The fundamental weakness of Kant's
moral theory lies in its appeal to a teleological system of nature
in which things have certain natural purposes. 11 P. 120.
40. Lectures on Ethics, pp. 162-171.
41. This is an attitude which we find almost described in Sections Ill
and IV of Leibniz's Discourse on Metaphysics, although Leibniz does
not put it in terms of instincts. The baffling character of
Leibniz's remarks there actually can help us to think about what
Kant means in this passage: Leibniz urges we take the attitude of
believing that it was all for the best towards our past actions,
while ridiculing the idea that one should be a 11quietist 11 about
the future. Discourse on Metaphysics, p. 7.
42. That such an unnatural use (and so misuse) of one's sexual power
11

is a violation of duty to oneself and, indeed, one which is con-

trary to morality in thehighest degree occurs to everyone imme-
diately ...
... it is as if man in general felt ashamed of being able to treat
his own person in such a way, which degrades it beneath the
beasts. 11 (DV88/425)
43. However, not everything Kant says in the surrounding passages fits
very well with the claim I am making here. In the Foundations, Kant
was inclined to view the heteronomouswill - which he is discussing
here - as if it had slipped back under the control of nature, tak-
ing its laws from nature. By the time of the writing of the
Religion, Kant had seen that this is not really possible given his
overall view, and radical evil has taken the place of nature in the
account of the freely heteronomouswill.
44. SigmundFreud, Civilization and its Discontents, especially chap-
ters II and III; also see, for example, "Character and Anal Eroti-
cism"; and the various discussions of the "economic model".
45. Mill, Utilitarianism, p. 291 in Warnock.
 Chapter 3
A World for Action

In this chapter I describe and defend the Practical Contradiction

interpretation of the Formula of Universal Law. According to the
Practical Contradiction interpretation, the kind of contradiction
that should be employed in the two contradiction tests associated
with the Formula of Universal Lawis a practical contradiction, the
sort of contradiction that is involved in thwarting one's own purposes
or in willing in a way that is self-defeating or self-destructive. I
begin by spel 1i ng out how this interpretation \'lorks in the case of the
first contradiction test - the contradiction in the universalized
maxim. I then discuss some objections that are commonlyraised to
this way of interpreting the first contradiction test, and defend the
interpretation against those objections. Finally, I discuss a number
of ways in which the Practical Contradiction interpretation might be
applied to the second contradiction test, the contradiction in the
will.

I. Practical Contradictions in Universalized Maxims

According to the Practical Contradiction interpretation of the
first contradiction test, the contradiction that is involved in the
universalization of an unacceptable maximis that the agent would be
unable to act on the maximin a world in which the maximwere a
universal law without frustrating his own purpose - the purpose that

91
 92

is in the maxim. It is a feature of this interpretation that, in order

to detennine whether a maximfails the first contradiction test, we
must include in the proper formulation of the maxim both the action to
be performed and the purpose that is aimed at. 1 The test cannot go
through unless both elements are present, for the purpose that is to be
thwarted in the world of the universalized maximwill be the purpose
embodied in the maxim. This allows us to interpret the contradiction
as an ordinary practical contradiction without bringing any extraneous
11
empirical 11 information into the test. The purpose embodied in the
maximcan be attributed to the rational agent who is going to perfonn
the test because it is his maxim- he must have that purpose or the
issue would not come up. Nothing about his wants or needs or desires
is assumed. Nothing is assumed but his rationality. It is a feature
of his rationality that allows us to assume the purpose in the maximis
his purpose. We do not even have to assume that he desires or likes
the purpose. If he adopts the maximthen he wills the purpose. That
is enough.
The maximcontains an action and a purpose, which the action is
to achieve. The agent is then required, by the moral law, to attempt
to will the universalization of the maxim. What is the universaliza-
tion of the maxim? It is that everyone who has that purpose will use
that action in order to achieve it. By that purpose is meant that
~ of purpose, where, for example, my wanting a certain thing and
your wanting a certain thing are purposes of the same type. If I
want my daughter to get into law school and you want your daughter to
get into law school, our purposes are of the same type, and the test
will determine whether the action I undertake to achieve my purpose
 93

could also be undertaken by everyone who, like you, has a purpose of

the same type as mine. I will give a tag to the universalization of a
kind of action to achieve a kind of purpose: the action becomes, when
universalized, what I shall call the standard procedure for achieving
that purpose. That is what you are asked to will - a world in which
your action is standard procedure for anyone who has a purpose like
yours to use in bringing it about. The phrase "standard procedure" is
meant to carry an implication that I think Kant wants. It is not
just that others with purposes like yours may behave or find natural
the option of behaving as you propose to behave. That would not
capture the full force of "law of nature". A standard procedure is
the natural, obvious, automatic way of doing something. It is the
method used by the culture or society, the one learned by children,
the one graced by the authority of tradition and convention. A great
part of the technological and institutional side of a culture can be
seen as a set of standard procedures by which ordinary humanpurposes
are achieved. The institutions of a society are set up to be standard
procedures for various tasks. One of the odd features about the
philosopher's fantasy that reason is primarily used to ascertain the
means to ends is that most people almost never have to think about the
means to their ends. They need only choose their ends - the means are
laid out already, routed by the elaborate apparatus of institution
and technology. If you want something to be done it is obvious what
to do - you call someone listed in the phone book as the person who
does that sort of thing, or go to a certain store or business, or
fill out and send off a form: at most, you might have to choose
between standardized procedures available, for instance whether you
 94

will reach your destination by bus, train, or airplane. This is not

completely a digression. On Kant's view the purpose of the state
might be described as that of setting up, organizing, and protecting
some set of standard procedures for doing things, thus facilitating
humanaction, the free pursuit of ends. 2
You are asked to will a world in which your action is a standard
procedure for achieving purposes like yours. If it turns out that in
that world you could not achieve your purpose at all, that is a
contradiction - a practical contradiction. The easy cases are as
usual the practices and heavily conventionalized actions - not
accidentally. You are in a financial difficulty and propose to make
a false promise. The false promise is your action, getting out of the
financial difficulty is your purpose. In the world of the universalized
maxim, could you get the money? Certainly not. For you would be
acting on this maxim, trying to get the money by making the false
promise, but,
the universality of a law which says that anyone who
believes himself to be in need could promise what he
pleased with the intention of not fulfilling it would
make the promise itself and the end to be accomplished
by it impossible; no one would believe what was
promised to him but would only laugh at any such
assertion as a vain pretense. (G 40/422)
The same analysis will work smoothly for maxims such as that of
cheating on an entrance exam in order to get into a professional
school; any sort of faulting on contracts or agreements for financial
or material gain; ignoring the provisions of wills in order to put the
money or land to another use, and other such obvious things.
One reason these cases submit so easily to the contradiction
test was considered in the previous chapter: they involve the
 95

existence and use of conventions and practices which, if regularly

abused, would simply die out. Other means would be found for solving
the various problems of humanexistence that these conventional
devices are meant to solve - making economic exchanges across time
possible, detennining who should get into schools and other
institutions requiring entrance exams, determining the disposition of
your property after you are gone. A certain amount of abuse of such
devices can be and is contained by legal sanctions, but universal
abuse would simply eliminate them. Wealso saw that natural acts do
not seem susceptible to this sort of difficulty. A problem here, I
think, is that the real source of the contradiction in these cases
{the conventional ones) is covered up by the fact that these practices
actually are eliminated by abuse. Natural acts are not, so it looks
as if the test cannot handle them. But what generates the contradiction
is something short of the actual elimination of the procedure: what
matters is that the procedure loses its efficacy with respect to a
certain purpose. Practices lose their efficacy because they are
altogether eliminated; but it is the loss of efficacy, not the
elimination, that causes the purpose to be thwarted. I will also hang
a tag on this: I will say that a procedure cannot be universalized if
its efficacy depends on its exceptionalness. One does not need to go
as far as saying the cheating and promise-breaking and violations of
contract can only exist if most people do not engage in them. It is
enough to say that they only work because most people do not engage
in them. In the case of many actions that fail the first contradiction
test it is part of the actual workings of the methods that most
people use them a certain way. This is true for example of the case
 96

of lying. Humanspeech can be used to deceive only because most

people use it honestly (in appropriate contexts). The efficacy of
speech as a tool of deceit depends upon certain expectations, built
up around what people say, which are in turn based upon the fact that
most people are inclined to speak truly. In the language of the
Formula of the End in Itself, the liar uses everyone who tells the
truth as a means - for everyone who tells the truth is unknowingly and
unwillingly one of the builders of the liar's deceiving-machine.
If the test is to handle the problematic cases of natural
actions it is this analysis that will have to be applied to them.
They are not going to cease to exist; but it must be possible to say
that their efficacy for certain purposes depends upon their
exceptionalness. Here a great deal depends on what the purpose is
taken to be and how it is described. But some cases I think can be
made to work without too much difficulty. One case that is somewhat
borderline is stealing. Nowthat might seem wholly conventional,
relating to the institutions of private and personal property; but
it is hard to imagine an economic arrangement in which the means of
production and action were not assigned to the use of certain people
at certain times, even if there was no private or personal property
at all. And any violation of these assignments, however made, might
be thought of as stealing. Nowif the purpose of stealing is to
acquire something for your personal use or possession - to get something
you want when you want it - and you imagine that anyone in your
situation. anyone who wants a thing not assigned to him when he wants
it - steals it, by way of standard procedure - then you see that
under these conditions it is quite impossible to acquire something for
 97

your personal use or possession, to have it when you want it. For
anything you have can be stolen. The idea here is that what the
thief really wants is to make something his property, to have some
guarantee that he will have it when he wants it and so the efficacy of
his action does indeed depend upon its exceptionalness. 3
Wecan even make plausible cases involving thoroughly natural
acts, at least in the case of certain purposes. Suppose you are
second in line for a job, and are considering murdering your more
successful rival. Could this be universalized? Killing is a natural
act, not a conventional one. We cannot say that if this sort of
action is abused the practice will die out, for that makes no sense
whatever. Nor can we say that any amount or kind of use of this
practice will destroy its efficacy for its purpose, if we specify the
purpose as that of getting someoneout of the way. So here the test
will only work if the purpose is specified differently. Wemust say
that the purpose is to secure for yourself a job, that the action is
seen as a way of getting a job. Andwe must emphasize the fact that
if anyone else wants this job, or a job that you have, you will be
the victim. You cannot will this. Nowit may seem that the purpose
that is contradicted is not the purpose in the maxim- that of having
a job - but a different purpose, that of remaining alive. Of course,
one has to be alive in order to have a job. That might seem like a
silly response at first, but I do not think that it is. In
Utilitarianism, Mill argues that justice is specifically concerned
with a very special object of humaninterest - that of security.
Security is not merely one amongmany other goods, but a pervasive
condition of the possession of anything good.
 98

The interest involved is that of security, to every one's

feelings the most vital of all interests. All other
earthly benefits are needed by one person, not needed
by another; and many of them can, if necessary, be
cheerfully foregone, or replaced by something else;
but security no humanbeing can possibly do without;
on it we depend for all our immunity from evil, and
for the whole value of all and eve y gohd, beyond
6
the passing moment;sTnce notning ut t e gratification
of the instant could be of any worth to us, if we could
be deprived of anything the next instant by whoever
was momentarily stronger than ourselves. (my emphasis}
On this argument, security is, to cast it into Kantian language,
a condition of the goodness of a thing. To \·1ant something is to want
to be secure in the possession of it. Violent natural procedures for
obtaining good things cannot be universalized because that would leave
people insecure in the possession of these things, and without that
security the good thing is not a good thing at all. We can see that
there is a contradiction in willing the universalization of such a
maximif we include as part of the purpose that the agent wants to be
secure in the possession of the end. In fact, in the Metaphysical
Elements of Justice, Kant uses an argument very much like Mill 1 s in
explaining why there must be proprietary rights protected by the
state. (I describe that argument in Section III of Chapter Four.}
A harder case would be one of something like killing for revenge,
or out of hatred. These grim kinds of cases are managedwithout
difficulty by the Formula of the End in Itself, I think, but it is
hard to extract any sense of contradiction here. To find a
contradiction here, we must admittedly push the idea of the purpose
quite hard. Thus one might say that what the vengeful or furious
person wants is something else, not just the death of his enemy.
Perhaps what he wants is for something not to have happened - and so
there is no action that can really help him here. This is not a very
 99

satisfactory answer. But it is important to notice a sort of para 11el

that it brings out. It is very often contended that Kant's theory is
least helpful, least compelling, least plausible, when one is dealing
with a case where the other people around the agent are doing something
very wrong. Kant's debate with Benjamin Constant about whether one
has a right to lie to the murderer at the door about the victim
concealed upstairs (SRTL), and his insistence that there is never a
right to revolution (DJ 84/38lff; 138-140/370-372), are famous
examples of cases in which Kant's view seems to forbid us, in the name
of moral purity and of always treating even the worst man with the
respect due to an end in itself, from trying to prevent or to right
the wrongs committed by others. There is a similar sort of difficulty
in making out what Kant would say about cases where something has gone
wrong inside, where the purpose itself has gone wrong. I do not say
that Kant cannot give us an account of such cases. But the kind of
case around which the view is framed, and which he has most often in
mind, is temptation, the desire to make oneself an exception,
selfishness, advantage-taking, meanness, disregard for the rights of
others. It is this more everyday sort of thing, not the great and
sickening crime, that serves as Kant's model. I do not think we can
fault him on this, unless we are prepared to deny that this is the
sort of thing that commonmoral evil consists in. And in these kinds
of cases, it seems to me that the practical contradiction view can
handle the natural acts in the way I have described.
In any case where you propose to use a procedure whose efficacy
depends upon its exceptionalness, you will obviously not be able to
universalize the maximdescribing that procedure. As a standard
 100

procedure it does not work, and is no longer efficacious. Yet as the

standard procedure and as the one described in your maxim it is the
one you would use in the world of the universalized maxim. Therefore,
when you will to universalize your maxim, you find yourself willing the
use of a method that you knowperfectly well would not work. You find
yourself willing a state of.affairs in which your purpose cannot be
achieved, because the standard procedure for achieving your purpose is
a procedure that does not work. You are therefore willing the
thwarting of your own purpose, and contradicting yourself. The
contradiction is of the ordinary practical kind, willing in a self-
defeating manner. The moral intuition is obviously that it is not
fair to use a method whose efficacy depends on the fact that other
people with the same purpose do not use it. You are in a literal way
taking advantage of others. You are making them your tools, for they
make your method work, they fuel its engine. The Formula of Universal
Lawand the Formula of the End in Itself are seen, when it is put
this way, not to be very different. The grounds of their claimed
equivalence are obvious.
One of the advantages of this way of looking at the contradiction
test is that it seems to be the most natural way of doing so. In the
realm of practice, what could be a more obvious candidate for
contradiction than thwarting your own purpose? Willing as if you both
did and did not want the same thing is a perfect analogue for, say,
arguing, as if you both believed and didn't believe the same thing.
It also explains .,,hy Kant will insist that you will a maxim"by which
you can at the same time" will that it be universal law (G 39/421),
and the frequent repetitions that we must will through the maximor by
 101

the maxim(G 39/421) or will the maximat the same time as its
universalization (G 55/437}. All of this, I believe, is meant to
emphasize that the difficulty is in willing the maximand its
universalization together. The whole idea seems to emerge more clearly
and firmly in the fonnulation Kant uses in the Typic11 section of the
11

Critique of Practical Reason:

Ask yourself whether, if the action which you propose
should take place by a law of nature of which lQ.!!
yourself were.! part, you could regardit as possible
through your will. (C2 72/69; my emphasis}
The phrase "of which you yourself were a part" serves to emphasize
the fact that the contradiction arises because you cannot will to be
part of the world of your universalized maxim. You cannot \·lill this
because you have a purpose - the purpose in that maxim- which would
be thwarted in that world.

II. TwoObjections to the Practical Contradiction Interpretation

Though this interpretation may be the most natural one, it is
certainly not the one that has first occurred to Kant's readers; and
when it does occur it is usually promptly rejected. The reasons for
this seem to me to be of interest; they reveal again a certain
resistance to the idea of practical reason, or a certain blindness
to its possibility.
The sense that a practical contradiction must be a thing "weaker",
less objective, more vague than a theoretical or logical one turns up
over and over in writing about Kant1 even amonghis defenders. It is
often suggested by the approach taken by writers about Kant that the
closer one can bring the contradictions described in connection with
the fonnulae of universality to theoretical logical contradictions
 102

the more thoroughly one will have defended Kant. The implication of
this approach is that only a theoretical contradiction, a contradiction
between propositions or beliefs, is a real contradiction; a
contradiction in one's maximsor one's will or just amongone's
purposes must be something weaker and the best we can do is bring it
as close as possible to this "best case" of contradiction. Singer,
for instance, who offers many very accurate examples to illustrate
the workability of the contradiction tests, nevertheless says:
Nor does it make any difference whether the impossibility
of everyone acting in a certain way is or is not a logical
impossibility, determined solely a priori, with no
reference to empirical facts. It-is sufficient for it
to be impossible in ~ weaker sense. (my emphasis)5
And Nell, whose interpretation of the contradiction tests seems to me
to be a version of the practical contradiction test, remarks:
The concepts which can be applied appropriately to
practical principles, such as universalized maxims,
are those of being "self-defeating" or "self-frustrating".
But if the criteria for applying the concept of a
contradiction are clearer than those for determining
self-defeat or self-frustration, then it may be
desirable to test maximsby considering their UTC's.
["Universalized Typified Counterpart" - that is, the
idea of everyone acting on this maxim] There the sort
of maximwhich Kant wants to show incompatible with
duty will be represented by a law of nature which
contains a contradiction. Kant does not want to depend
on relatively opaque concepts like those of being
self-defeating or self-frustrating in his test of which
maximscannot be practical laws. Rather he wants to
provide a method for mapping instances of such concepts
onto the muchmore readily identifiable instances of the
concept of a contradiction.
I do not knowof anyone who argues for this idea that we have a
clearer notion of theoretical contradiction than we do of practical
contradiction. Whena person is chargeable with a theoretical
contradiction it is because her beliefs imply that she both holds and
does not hold some particular belief or tenet; when she is guilty of
 lro

a practical contradiction it is because her "willings" imply that she

both wills and does not will the same thing. It is supposed that the
first of these two things is more clear, sharp, and well-defined than
the second. It would be worth trying to determine why anyone is
inclined to suppose such a thing. For now it is perhaps enough to
point out that Kant emphasizes the idea of the will, which is practical
reason, and which has content, namely maxims, because he sees our
practical faculty as being not derivative from but on a level with our
theoretical faculty. This, after all, is what it means to believe
that there is practical reason, and not just theoretical reason
applied to action. The practical employmentof reason is to be
separate but equal; it will have its own sort of contradiction. Of
course we want it to be as analogous to the theoretical employmentof
reason as possible, because this will help us to understand the claim
that both are employments of reason. But heuristic points must be kept
in their proper place. To suppose at the outset that the theoretical
contradiction is the real one and the practical contradiction can only
be a weak approximation to it is to refuse to let Kant get his project
off the ground. For this project is that of showing us that there is
practical reason, and showing us that there can be a practical
contradiction is a sort of small first step in the project. If the
reader starts off by saying: well of course that cannot be a real
contradiction but perhaps it could be very like one, the reader is
flatly refusing to let Kant begin to make his argument. To insist
that the will must be a second best imitation of the theoretical
understanding is already to insist that reason has no thoroughgoing
practical employment. Therefore such readings beg the question.
 104

It is worth going into this a little further. I said above that

the will is practical reason and its content consists of maxims. This
can be compared to the understanding as governed by theoretical reason,
where its content consists, say, in beliefs. I do not wish to become
entangled in Kant's views of the understanding and its relation to
reason; but I think a helpful comparison can be made. On Kant's view
when we submit our beliefs to the discipline of reason there are two
sorts of things that we do. The one everyone agrees about is that our
beliefs are supposed to be logically consistent - they must not be
self-contradictory, and they must not contradict among themselves.
This is the "analytic" part of rationality. There are also the
synthetic principles that govern and regulate our systems of belief
in various ways, the necessities that go beyond the analytic to which
the rational knower must submit. Beliefs that are inconsistent among
themselves and beliefs that violate the synthetic.! priori principles
would represent different sorts of failures of rationality. I describe
these things as standards or norms that are applied to our beliefs in
order to bring out the analogy. In a parallel way, the will imposes
standards and norms on our maxims. The maxim is the basic unit of
willing that parallels the belief. The various standards in the
practical realm take the form of imperatives: the theory of imperatives
is a theory of practical reason. The maximthat is not in accord with
the technical imperative is like the belief that is inconsistent in
itself. The maxim that is not in accord with the imperative of prudence
is like the belief that is inconsistent with the other things that one
believes. The maxim that is not in accord with the categorical
imperative is like the belief that violates a synthetic principle.
 105

The categorical imperative is a synthetic principle; the hypothetical

imperatives, the technical imperative and the imperative of prudence,
are analytic.
It is obvious from this comparison that the sense of contradiction
appropriate to the standpoint of practical reason is the sense that is
associated with the hypothetical imperative. And the sense of
contradiction associated with the hypothetical imperative is certainly
practical contradiction. It is a mismatch between the willing of
means and the willing of ends. That is a contradiction, intuitively,
because you at once will an end but, by refusing to will the means, or
failing to will the means, or willing something that implies the
unavailability of the means, you imply" that you do not will the end.
11

The structure of your maxims implies that you both do and do not will
some particular end. Or rather, that you will that the end both be
and not be - to get the stronger formulation. The analogy is more to
believing and disbelieving the same proposition than to both believing
and merely failing to believe the same proposition, though this is a
rather fine point. In fact one might say that on the occasions when
it is appropriate to accuse someoneof a contradiction you point out
to them {as often as not} that their failure to believe something
amounts to explicit disbelief in that context. Something else to which
they are committed disallows the suspended judgment. In the practical
context, willing the end similarly disallows willing something that
implies the unavailability of the means. To apply that to what has
already been said about the categorical imperative test, I need only
point out that willing a situation in which the only means available
for achieving a certain purpose is a procedure which does not work is
 106

in effect willing the unavailability and inefficacy of the means by

which the purpose is to be achieved. If you will the purpose, you
cannot do this without a contradiction. And the contradiction is of
the sort described by the account of the hypothetical imperative.
For example, in the false promising case you will to get some ready
money, but when you try to will the world of the universalized maxim,
you will a state of affairs in which it is impossible to get any ready
money. The reason it is impossible is that the standard procedure for
getting ready money is a procedure that does not ~ark - and the reason
for that is that it is a procedure whose efficacy depends upon its
exceptionalness, and whose efficacy therefore disappears when it is
standardized. It is a willing-experiment, analogous to a thought-
experiment. You find you cannot will the world of the universalized
maximwithout a contradiction, because you cannot will a state of
affairs in which there is no available means to your willed end.
Of course, it may be objected that this only proves that the
practical employmentof reason is "weaker" and less thoroughgoingly
objective than the theoretical employment, on the grounds that the
hypothetical imperative is not as clearly analytic as a logical truth.
The argument shows that Kant must mean that the kind of contradiction
involved in the categorical imperative test is the same kind of
contradiction involved in the violation of a hypothetical imperative,
for the same notion of contradiction must be used throughout the domain
of practical reason. But that does not by itself show that the kind
of contradiction involved in the hypothetical imperative is not a
weaker and less clear sort than a logical contradiction. The hypothe-
tical imperative is supposed to be analytic, but it might be argued
 107

that this is not very obvious.

However, Kant has an argument to show that the hypothetical
imperative is analytic. We have two ways of judging the analyticity
of a principle. One is that its opposite is a contradiction -
obviously we cannot appeal to that here, since it is precisely the
reality and sharpness of this contradiction that has been challenged.
The other in the case of a theoretical proposition is that the
predicate is contained in the concept of the subject. The practical
analogue will be that the willing of the means is contained in the
willing of the end. This is in fact the argument that Kant makes in
the Foundations, when he is explaining how the various imperatives
are possible.
Whoeverwills the end, so far as reason has decisive
influence on his action, wills also the indispensably
necessary means to it that lie in his power. This
proposition, in what concerns the will, is analytical;
for, in willing an object as my effect, my causality as
an acting cause, i.e. the use of means, is already
thought, and the imperative derives the concept of
necessary actions to this end from the concept of
willing this end. (G 34-35/417)
Kant himself is guilty of theoreticizing his point by dragging in
concepts here, I think. It is better to say, not that the concept of
willing the means is involved in the concept of willing the end, but
simply that willing the means is involved in willing the end. However
that may be, the argument given here may be reconstructed to show how
hypothetical imperatives are analytic. The basis of this argument is
an idea that plays a crucial and central role in Kant's practical
philosophy. It is that willing involves, or is equivalent to,
regarding oneself as a cause. The important point is that the agent
who wills an end views himself as the cause of that end, as the one
 108

who is to bring it about. Willing is different from wanting or wishing

or desiring, because these attitudes may not involve regarding oneself
as the one who is to bring it about. But willing does. The agent who
wills an end, therefore, conceives himself as the cause of that end.
Conceiving oneself as the cause of the end involves conceiving oneself
as using the available causal connections to bring about the end: as
setting off a chain of causes that will lead to the end. To conceive
oneself as using causal connections to bring about the end is to
conceive oneself as using means to the end, for this is what means are.
So when you will a given end, willing the use of means to that end is
already part of that willing. From this it follows that whoever wills
an end wills the indispensable means: this principle is analytic, for
the willing of the means (or the concept of willing the means) is
contained in the willing of the end (or the concept of willing the end).
To will the end and at the same time to fail to will the means or
actually will the non-existence or unavailability or inefficacy of the
means is a straightforward contradiction. This contradiction is in no
way weaker or less definite or less clear than a theoretical
contradiction. It is to willing exactly what a theoretical
contradiction is to believing; the rule that it be avoided is to the
adoption of maxims exactly the same sort of discipline that the rule
against theoretical contradiction is to the believing of beliefs.
This argument depends upon the important idea of the agent
conceiving himself or herself as a cause. The importance of this idea
may be established by pointing out the central role it plays in Kant's
practical philosophy. Everyone remembers the famous passage that
initiates the discussion of imperatives and the general idea of
 109

rationality in action:
Everything in nature works according to laws. Only
a rational being has the capacity of acting according
to the conception of laws, i.e., according to
principles. This capacity is will. Since reason is
required for the derivation of actions from 1aws,
will is nothing else than practical reason. (G 29/412)
Here the will is identified as the capacity to act on one's conception
of a law. Later, in the initial paragraphs of the Third Section when
the critique begins, Kant says:
As will is a kind of causality of living beings so
far as they are rational, freedom would be that
property of this causality by which it can be
effective independently of foreign causes detennining
it ...
... Since the concept of a causality entails that of
laws according to which something, i.e., the effect,
must be established through something else which we
call cause, it follows that freedom is by no means
lawless even though it is not a property of the will
according to laws of nature. (G 65/445-446)
The initial argument for attributing autonomy to anyone with a
rational will proceeds by moving from the idea of the rational agent
regarding herself as a cause to the necessity that the rational
agent must regard herself as a first cause, the author of her own
principles. The idea of autonomy, central to the main critical
arguments for the rationality of acting on the categorical imperative,
is given in tenns of the idea of the rational agent as a cause, and
as having to regard herself as a cause. As the analysis of the
principle that whoever wills an end wills the means shows, however,
this theme runs throughout Kant's descriptions of rationality in
. 7
act ,on. If one takes seriously (as to follow the critical arguments
one must) the claim that willing~ regarding yourself as the cause
of an object - seeing it as a thing to be brought about through your
own use of means - then there is no grounds for the objection that
 110

practical contradiction is not a full-fledged sort of contradiction.

It is no weaker than its theoretical analogue.
There is one other objection to the practical contradiction
interpretation that I want to examine. The practical contradiction
argument with respect to false promises is nicely stated but then
rejected in the following passages from Allen Wood's 11 Kant on False
Promises":
Probably the most obvious and natural line of reasoning
it [the false promising example explanation in Foundations]
contains is this: If M [the maxim] were a universal law
of nature, it would coiiie to be general knowledge that
whenever someone borrows money and promises to repay
it, he may well have no intention of doing so. And thus
in the case in question, if the agent acting on M
tried to gain money by making a false promise, no one
would believe him. He thus would seem, in willing M
to be a universal law of nature, to be thwarting his
own purpose.
The most obvious thing wrong with this interpretation
of the argument is that it is plainly not the sort of
argument Kant claims to be giving, and is in fact an
argument of just the opposite kind. For the above
argument (if it is correct) still establishes only
that the agent cannot will Mto be a universal law
of nature without conflicting volitions. It does not
even claim that!:! cannot be thought as a universal
law of nature without contradiction.
It is noteworthy that this line of argument is
based on the alleged fact that if Mwere a universal
law of nature, the end involved in-M (gaining money
in time of need) would be impossible of attainment.
But what Kant says is that if!:! were a universal law
of nature, 11it would make the promise itself and the
end to be accomplished by it impossible. 11 Kant points
out not one thing which wguld be impossible but two.
Whydoes he mention both?
Wood's objection is based on two things Kant says: that in this
case the maximcannot be thought as a universal law, and that promising
itself, as well as its purpose, would be impossible. Similar objections
made by many others point to the language of the Critique of Practical
Reason, where Kant speaks of maximswhen universalized annihilating or
 111

destroying themselves, or the remarks in the Foundations about the

resulting systems of nature which could not even exist. None of this
seems to have anything to do with thwarted purposes. It might seem
as if only some other sort of contradiction could prevent something
from existing or cause it to self-destruct.
Something can be said in response to each of these points. First
of all, as I have already observed, the fact that promising itself
(as well as the purpose of promising) is threatened by universalization
can be explained by appeal to the fact that promising is a sort of
practice, and practices can simply die out under the stress of
universal abuse. But it can only cause confusion to emphasize this
interesting fact to the extent of concealing the more important fact
that it sometimes does conceal - that practices under the stress of
universal abuse cease to be effective. As I argued earlier, it is not
whether the practice still exists or not that matters but whether it
is still efficacious with respect to the purpose in question. The
ineffectiveness of the universalized procedure is enough to establish
the kind of contradiction that makes hypothetical imperatives analytic;
and as I have already said, it is reasonable to suppose that Kant
intended to employ the same notion of contradiction in the argument
for the analyticity of the hypothetical imperative and in the two
contradiction tests. And if we are to give an account of the injustice
of employing certain kinds of "natural" or non-conventional procedures
for certain purposes, which cannot possibly be made "impossible" the
way that promising could, it will not do to overemphasize the role of
establishing the impossibility of the universalized procedure itself.
The point about the universalized maximbeing inconceivable or
 112

unthinkable, annihilating or destroying itself, is a more forceful

point in Wood's favor. The practical contradiction view must
accommodatethese remarks by pointing out that in these cases there
is a sort of inherent instability in the world of the universalized
maxim. It is a world in which the standard procedure for achieving
some commonsort of purpose is a procedure that does not work, because
it is a procedure that loses its efficacy when standardized. It is a
world in which everyone runs around trying to achieve something by a
method which everyone must knowdoes not work. The situation is
unstable: at first one wants to say, some other method would be
found. But Kemp,for example, complains that what Kant says is not
that the situation could not last, but that it could not be. Here it
is helpful to insist on the fact that we are imagining our maximas a
law of nature. Anything that could not last as a law of nature could
not be a law of nature, for the laws of nature are permanent. At the
risk of putting it a bit fancifully, one might say this: that even if
it does take time for people to discover that everyone acts in the
objectionable fashion (as for example in the false promising case
where the failure of promises to "work" depends upon people's knowing
that they are false), since humanbeings are pictured as having acted
this way forever there is no time at which they would not already have
found out. So the universalized method would never work, and there
would always be a practical contradiction.

III. Practical Contradictions in the Will

Whena maximfails the second contradiction test, the contradiction
is found, not in the universalized maximor in the world of the
 113

universalized maxim, but in the will of the person who attempts to will
the world of the universalized maxim:
He sees that a system of nature could indeed exist in
accordance with such a law, even though man (like the
inhabitants of the South Sea Islands) should let his
talents rust and resolve to devote his life merely to
idleness, indulgence, and propagation - in a word, to
pleasure. But he cannot possibly will that this should
become a universal law of nature or that it should be
implanted in us by a natural instinct. For, as a
rational being, he necessarily wills that all his
faculties should be developed, inasmuch as they are
given to him for all sorts of possible purposes.
(G 41/423)
If such a way of thinking were a universal law of nature,
certainly the humanrace could exist, and without doubt
even better than in a state where everyone talks of
sympathy and goodwill, or even exerts himself occasionally
to practice them while, on the other hand, he cheats
when he can and betrays or otherwise violates the rights
of man. Nowalthough it is possible that a universal
law of nature according to that maximcould exist, it is
nevertheless impossible to will that such a principle
should hold everywhere as a law of nature. For a will
which resolved this would conflict with itself, since
instances can often arise in which he would need the
love and sympathy of others, and in which he would have
robbed himself, by such a law of nature springing from
his own will, of all hope of the aid he desires.
(G 41/423)
In Chapter Two, in the discussion of the terrible consequences
interpretation, I described a certain natural pattern of thought that
is provoked by the problems arising from these two arguments. It is
seen at once that the theoretical contradiction interpretation cannot
be applied in these cases, since Kant explicitly says that in these
cases the world of the universalized maximcould indeed exist.
Because the contradiction is said to be in the will, and because Kant
seems to speak directly of thwarted purposes here, the idea of the
practical contradiction at this point naturally arises. But so does
the standard objection to this test: that there is no purpose that
 114

can be attributed to a rational being as such that would be thwarted

in these worlds of the universalized maxims. Perhaps the agent will
not require the help of others or the use of his talents: these are
empirical assumptions that, according to Sidgwick and others, are
inappropriate here. And so there is supposedly nothing left of the
argument except a sort of rule-utilitarian thought that it would be
very bad indeed if no one developed any of their talents and powers
or if no one ever helped anyone else who was in need.
The sort of reading that the practical contradiction view gives of
the first contradiction test is unavailable here because the sort of
maximthat is tested seems to contain no reference to any purpose at
all. The two maxims are described as follows: in the first case, the
maximof neglecting his gifts; in the second case, Kant describes a
"way of thinking" which includes the line "I will not take anything
from him [anyone] or even envy him; but to his welfare or to his
assistance in time of need I have no desire to contribute." (G 41/423)
In neither case is any purpose specified. Although in the first case
Kant mentions that the maxim"agrees with his propensity to idle
amusement", in the second case there is not even this much motivational
explanation given: the man, for whomthings are going well, arrives
at this thought while contemplating others "whomhe could help and 11

who "have to struggle with great hardships•• and wondering "what concern
of mine is it?" (G 40-41/423-424). In neither case does the maxim
nor such motivational description as Kant gives provide any clear
grounds for assigning to the agent some particular purpose.which would
be thwarted in the world of the universalized maxim. In order to give
credence to Kant's claim that there is a contradiction in the will
 115

involved in willing the universalization of these maxims, what we need

to find is some purpose, thwarted in the worlds of the universalized
maxim, assignable to a rational being as such, but not especially
associated with the maximin question. That last requirement is there
because we must also make out the difference between the two tests, and
because Kant does say that in these cases the universalizations can
exist.
A number of suggestions of what sort of purpose might (more or
less) fulfill these criteria have been made in the literature on the
second contradiction test and related discussions. Although in the
end it is my sense that a clearer and more forceful account of the
duties arising from the second contradiction test can be given on the
basis of the Fonnula of the End in Itself, 9 I think that the
possibilities and their various merits can be helpfully reviewed here.
If we are to support Kant's claim that the various formulae are the
same, there certainly must be some sort of contradiction argument
available here that parallels the arguments based on the Formula of the
End in Itself, even if the latter are more accessible. Even if we
cannot decide which of these accounts is the best with any certainty
until after we have looked at the Formula of the End in Itself and its
arguments, we can at this point see how the various accounts work.
Singer, going on a suggestion from Jonathan Harrison, considers
the possibility that in the case of these maximswhich do not include
a purpose we can identify an implicit purpose - the natural or nonnal
purpose of adopting a maximof this kind. 10 The interesting feature
of this suggestion, as Singer himself points out, is that it tends to
bring the two tests close together:
 116

Thus, instead of speaking of someone adopting the maxim

"never to help others in distress", and then mentioning
his purpose in adopting it, we can say that he adopted
the maxim "never to help others in distress in order to
benefit from the help of others without doing anything
in return". No doubt to state a maximin this way will
often result in an excessively complicated and unrealistic
statement. But it can be done. One's purpose in adopting
a certain maximis his reason for adopting it, and to
specify the purpose is to specify the circumstances under
which the maximis applicable; thus all this is part of the
'llider pol icy of which the narrowly stated maxim is only a
part. Now, if the maximis stated in this more explicit
way, so that it not only states how one proposes to act
but also states when and why one proposes to act in that
way, then Kant's narrower criterion will be applicable in
every case in which the more inclusive one is. Not only
will it be impossible to will that everyone should adopt
the maxim, but it will be impossible for everyone to
adopt it or successfully to act on it - if everyone were
to adopt it then no one .,.,ould be ab1e to do so.
It appears, then, that these two criteria for
determining whether a maximcan be universalize~ are not
so distinct as we might at first have supposed. 1
The gain here is that, since the first test works, this device can
make instances of the second test work as well. The loss is that
Kant himself certainly thought that the two tests were different and
that the duties derived were of a different sort. If, as I believe,
it is correct for the most part to associate the first of the two
contradiction tests with the duties of justice and the second with the
duties of virtue, the distinction between the two tests becomesmore
rather than less important in light of the important doctrine of the
two kinds of duty set forth in the various introductions to the
Metaphysics of Morals (see Chapter 6).
A further disadvantage is that building these motivational
assumptions into the example tends to beg the question against
Sidgwick, whose argument was that it is only an empirical assumption
that the agent would want help. The duties of self-cultivation and
mutual helpfulness are positive duties, and as such, they are meant to
 117

be incumbent upon everyone. Of course in some sense all duties are

incumbent upon everyone, but negative duties only apply where
appropriate. Not everyone is tempted to make a false promise, keep a
deposit, cheat and trample upon humanrights. So all these issues need
not arise for everyone. But the duties of helping others and self-
cultivation are supposed to apply to everyone. Singer describes the
man in the example as wanting to benefit from the general condition
of mutual helpfulness but not wanting to contribute. Sidgwick1 s
question is about someonewho does not want to benefit from the general
condition of mutual helpfulness, so one cannot answer Sidgwick by
saying that this is the normal purpose of adopting such a maxim.
Andyet it seems to me that Singer is absolutely right in identifying
what Kant thinks is wrong here in terms of the man's willing to benefit
from a state of affairs to which he proposes to make no contribution.
All cases of violation of duty for Kant are cases of making an
exception of oneself. In cases where the violation can be argued in
terms of the first contradiction test, it is clear what the sense of
exception is: by using a method that not everyone with your purpose
could use, you claim for yourself an obviously exceptional privilege.
In the case of the second contradiction test, willing to use a state of
affairs to which you do not contribute seems like the right sort of
exception - like not paying taxes but using the public highways.
Everyone does in fact benefit from the conditions of mutual
cooperativeness; certainly everyone uses the fruits of the development
and cultivation of humantalents. This being so, everyone who refuses
to make a contribution to these resources is making himself or herself
an exception. Sidgwick's complaint, however, is that these are mere
 118

empirical facts. To answer Sidgwick, we need to do more than claim

that the formulators of these maximsdo, or usually do, have these
purposes. We need to show that rational beings must have these
purposes.
Singer, perhaps without quite realizing it, ends up trying to
establish just this point when he argues explicitly against Sidgwick.
His argument is roundabout and clever:
I can conceive of someonewho would rather starve than
receive any financial assistance from anyone (or who at
least would say that he would), but I cannot conceive
of anyone who, when set upon by a band of gangsters,
would prefer to be beaten, robbed, tortured, and killed,
rather than be assisted to escape. To suppose this is
to suppose someonewho would renounce all his ends rather
than be assisted in the attainment of any of them, and
this is impossible. It is possible for someoneto
renounce some particular end in order to avoid being
assisted in its attainment. But it is not possible
for anyone to renounce all his ends, for to suppose
that it is leads to contradiction. One of the ends of
such a person, surely, and it would appear to be a
paramount end, is not to be assisted by others. But
suppose that in order to achieve this end, that is, in
order to ward off the well-meant but unwanted assistance
of some, it is necessary for him to have the help of
others (even if it is only a "Cease and desist" order
from a court). What would he want then? What we are
being asked is to imagine someonewho could say,
honestly and truly, "The only thing I want is not to
be helped by anyone else. 11 Helped to do what? Taken
strictly this is not a desire or a possible end at all,
and this is shown by the fact that such a person may
want and desire and need the ,ssistance of someone
else in order to achieve it. 1
This argument leads naturally to another reading of the second
contradiction test. The claim is that !!.Q.matter what your ends are,
it is possible that you might need the assistance of others to achieve
them. Presumably the same sort of thing might be said about talents.
On this basis, we might read the second contradiction test as saying
that everyone will have some purposes that cannot be satisfied in
 119

these worlds of the universalized maxims. l~e do not need to identify

the purpose specifically, for everyone wi11 have some purpose or other
for which they might conceivably need help, and so cannot will a world
in which no help is available. Even if you do not have such a purpose
right now, one might say, one can imagine either situations in which
you would almost certainly adopt such a purpose, or situations in which
one of your enduring purposes would be threatened in a way that would
make it necessary for you to will the availability of some assistance.
Singer's point is that we could construct an example of such a situation
even with regard to the purpose of avoiding the assistance of others.
One cannot be certain that one will never need the assistance of others,
no matter what one's purposes are.
On this reading, mutual assistance and developed talents and
powers are what Rawls calls primary goods - they are things that we
want whatever else we want. 12 For this reason everyone must will their
availability, regardless of her or his personal plans and projects.
Talent and assistance are the most general resources of action, used
in the achievement of almost every purpose, needed by every person
regardless of what he or she is up to. It is therefore not necessary
to specify the end for whose sake each person wills the availability
of these means. You need only have some ends. Presumably every being
who is practically rational has some ends. This reading has the virtue
of fitting nicely with Kant's remarks: we need our talents for 11
all
sorts of possible purposes"; instances in which we need the aid of
others "can often arise." I suppose that it is still possible to
object that the "primary goodness" of these pervasively useful
resources of action is a mere empirical fact about humanbeings.
 120

Duties are to apply to all rational beings, and perhaps we can imagine
rational beings for whomthese are not resources - beings who have no
ends for which they need assistance, beings who are born into whatever
world they inhabit with all the talents and powers they \•lill ever need
or indeed could ever have. But it is hard to see what this objection
comes to. Certainly, if rational beings had no use for one another's
assistance, there would be no duties of mutual aid - but there is no
reason to suppose anyone would \'lant to deny that. One could say that
the test establishes that if there are any primary goods or all-pervasive
resources of action, everyone must will their availability - and in the
case of humanbeings, cooperation and talents are at least amongthem.
In any case this would be rather like claiming that it is just an
empirical fact about us that we are not gods. Philosophical readers
will no doubt be divided in their opinions about that.
So far, I have described one reading based on the idea that there
is an implicit purpose associated with the maxims rejected by the
second test, which, after the manner of the first test, would be
thwarted in the world of the universalized maxim; and one reading based
on the idea that talent and cooperation are primary goods whose
availability must be willed no matter what one's ends and purposes
are or might be. A third strategy for dealing with the second
contradiction test is to identify as the purpose thwarted in the
world of the universalized maximthe one actual end that Kant says
everyone can be presumed to have - happiness. This is, of course,
especially plausible as a way of dealing with the second example,
the mutual assistance case, but it could apply to both.
One can move naturally from the claim that mutual assistance and
 121

talents are a sort of primary goods needed for whatever ends we might
have to the claim that they are needed for happiness, for happiness
itself is sometimes viewed by Kant as a sort of second-order end -
the dream of "the sum of satisfaction of all inclinations." (G 15/399) 13
The desire to be happy is a desire to satisfy as many of your
inclinations as possible, and that in turn involves a desire that the
means to do so, in general, be available. This has the advantage of
locating the contradiction in the thwarting of a specific purpose,
but a purpose that everyone surely has. The contradiction is nowmore
definite than it is if we just speak of ends in general - it applies to
everyone, and it is the same contradiction for everyone (which on the
previous reading might have been in doubt). In addition, there is a
possible architectonic point to it - for if we read the second test in
this fashion, we can say that the first contradiction test is based on
the technical imperative, while the second contradiction test is
based on the imperative of prudence. The reading also finds support in
some of the remarks Kant makes in the Doctrine of Virtue, where the
duty of assisting others is subsumedunder the obligatory end of
taking the happiness of others as one's end, and is traced to the fact
that "our self-love cannot be divorced from our need of being loved
by others ... so that we make ourselves an end for others." (DV53/393)
Furthermore, the uncertainties associated with the pursuit of happiness,
which Kant so strongly emphasizes, can be used to support the claim
that every one of us has a reason to contribute to the development of
humantalents and a cooperative and helpful community, whether it
seems so in light of our present purposes or not. Kant could reply
to Sidgwick's objection that no one can knowwhether he can do without
 122

the assistance of others because "the concept of happiness is so

indefinite that, although each person wishes to attain it, he can
never definitely and self-consistently state what it is he really
wishes and wills." (G 35/418) The uncertainty of happiness has the
effect of locking us pennanently into something like Rawls' original
position - we may knowwhat our private conception of the good is now,.
but the possibility that it will undergo a radical change has the same
sort of effect as if we did not knowat all. In light of that, we can
see by Rawlsian arguments that it behooves us to be fair, and to make
a fair contribution to the commongood.14
The idea that what is at stake at least in the duty of helpfulness
is not willing a world in which one's happiness might be endangered is
fairly commonin the literature. There are two problems with it,
however. One difficulty is that it makes one wonder why Kant did
not just posit happiness as the purpose of adopting this maximand
assimilate the example to the first contradiction test (with the
difference that the end and so the duty belong to everyone and that it
is the violation of the prudential imperative that provides the
contradiction. Certainly, one can easily imagine how this argument
would go: we simply add to the maximthat the man supposes he can
better achieve his happiness by not troubling about others. (This is
different from Singer's move of supposing that the man actually wants
to benefit from the helpfulness of others, because nowwe get that
part of the argument from the uncertainty of happiness.) On
universalization he discovers that, if no one troubled about anyone,
his already uncertain chances of happiness will be greatly reduced,
and so we have a contradiction of the first type. Whydidn't Kant
 123

make the argument like this?

A passage from Ebbinghaus1 s 11Interpretation and Misinterpretation
of the Categorical Imperative 11 provides a suggestion concerning the
answer to this question:
In deciding the unlawfulness of the maxim11Never help
anyone in need11 there is no question of consequences
so far as these concern the possibility of attaining
.!!!t.own happiness as!.!!. end if this maximis followed
by myself and others. But in coming to this decision
there is indeed a question of consequences so far as
these concern the possibility of taking !!!Y.own happiness,
and consequently my own rescue in distress, as an end,
if I will the maximof hardheartedness as a la\i-:-:-.-
The reason why I cannot will the maximof indifference
as a law is not that by so doing I should stand in the
way of my own happiness. Whether my chance of happiness
will be improved or impaired by freeing everyone, and
consequently myself, from all obligation to help others
depends on circumstances about which there can never
be any! priori decision. But I can indeed decide
a priori, as in the case of every analytic truth, that
Tf my will includes a will to be left without help in
need, I cannot possibly take happiness as my end and
consequently must be able to will my own unhappiness. 15
It might be tempting (though this is not Ebbinghaus s point) to say
1

that this marks a difference between the two tests. In the first test
it is clear that you would not be able to attain the end that is
thwarted. In the second test it is not clear, because, as Ebbinghaus
says, it is a matter of unpredictable fact who might benefit and who
might lose from the universal practice of personal isolationism. So
this problem can perhaps be overcome. It is a question of the
conditions under which one can rationally adopt an end, not under which
one can rationally expect to achieve it - it is a case in which
expectations are wholly unreliable.
The second and more important difficulty with this reading is that
it seems to work much less well for the duty of cultivation. Kant's
example of a world in which there is no cultivation is the South Sea
 124

Islands, and there is no reason to believe he thought of the people

there as miserable - indeed he describes them as given up to pleasure.
Of course, one might \'lant to make something of the fact that not
everyone has the advantages of a tropical island paradise. But whatever
others might be inclined to think, Kant was enchanted with Rousseau,
and he had a low opinion of the happiness brought into the world by
cultivation. This emerges often in the teleological history writings.
On the other hand, the ends-in-general reading finds a clear basis in
Kant's discussion of this duty in the Doctrine of Virtue, where he
says that the duty of natural perfection is:
"Cultivate your powers of mind and body so that they
are fit to realize any end you can come upon, for it
11

cannot be said which of these ends could, at some time,

becomeyours. ( DV 52/392)

This brings us to the last reading of the second test contradiction

that I will discuss, and the one that I think fits most naturally
with Kant's other arguments. On this reading, the end that is
contradicted in the worlds of the universalized maxims in the case of
the second test is freedom - in the specific sense of the capacity and
opportunity to adopt and pursue ends. The argument is that in the
world of the universalized maximsone's freedom is cramped, restricted,
and limited in ways that are unacceptable to a being that wills its
own freedom. Here, freedom is assigned to every rational being as an
end that is a specifically rational end. Wemust will our own
freedom and therefore we cannot will a state of affairs in which it is
thwarted. A world in which there is no development of humantalents
and powers is certainly such a world; a world in which there is no
cooperation can be argued to be so as well.
Kempfavors this reading of the second test, and spells it out in
 125

the case of the duty of cultivation as follO\oJS:

What chiefly distinguishes man from the rest of creation,
according to Kant, is his possession of freedom; this in
turn depends on his possession of reason, not in the sense
that he is capable of theoretical activity, but in the sense
that he can set ends or purposes before himself (whereas the
rest of creation can merely fulfill passively the purposes
of nature). And this gives its point to the expression,
which Harrison finds obscure, "For as a rational being he
necessarily wills .... " Whatever a man's private aim or
purpose in life may be, the fact that he has such a purpose
is a sign of his rationality, even though all men, being
imperfectly rational, have some purposes which they would
not have if they were perfectly rational, and fail to
have some which they would then have. Nowany human
purpose requires the exercise of some talent or capacity
for its fulfillment; for a talent or capacity just is the
ability to take appropriate means to given ends. Man's
ability to conceive of purposes would be of no value,
and his freedom would be incomplete, if he were not also
endowedwith the capacity for discovering and adopting
the best means for the attainment of those purposes.16
As the passage brings out, the argument for these duties combines the
rationally necessary willing of freedom, with the fact that talent and
cooperation are a sort of primary good for its exercise. In this
connection it is interesting to note that in the DeweyLectures,
Kantian Constructivism in Moral Theory", Rawls uses this sort of
11

account of the primary goods. (In Theory of Justice, Ra\'1lsI account

of the primary goods is closer to the ends-in-general reading of the
second contradiction test.) Rawls says:
... we take moral persons to be characterized by two moral
powers and by two corresponding highest-order interests in
realizing and exercising these powers. The first power is
the capacity for an effective sense of justice, that is,
the capacity to understand, to apply, and to act from (and
not merely in accordance with) the principles of justice.
The second moral power is the capacity to form, to revise,
and rationally to pursue a conception of the good.
Corresponding to the moral powers, moral persons are said
to be moved by two highest-order interests to realize and
exercise these powers... Since the parties represent
moral persons, they are likewise moved by these interests
to secure the development and exercise of the moral powers.17
 126

But at the basis of the desire for primary goods are the
highest order interests of moral personality and the need
to secure one's conception of the good (whatever it is).
Thus the parties are simply trying to guarantee and to
advance the requisite conditions for exercising the powers
that characterize them as moral persons. Certainly this
motivation is neither heteronomous nor self-centered:
we expect and indeed want people to care about their
liberties and opportunities in order to realize these
powers, and we think they show a lack of self-respect
and weakness of character in not doing so.18
Here again the guiding idea is that it is not just any (as Kant would
say, arbitrary, conditioned, or contingent) purpose for which we need
these resources but this further thing that is realized in the course
of 11setting 11
and pursuing our other ends, namely "moral personality"
or "autonomy" or freedom. It is noteworthy that Kempappeals to the
teleological argument in his passage - humanbeings differ from all
the rest of creation in our power to set ends, in our freedom from the
domination of natural purposes themselves - our natural purpose,
signalled by our freedom from instinctual bonds, is precisely to work
out our own ends - to be free. Thus this reading of the second test
is supported by the argument made in the previous chapter on the basis
of "Conjectural Beginning of HumanHistory": even before Kant
establishes the obligatory ends that are supposed to follm·, from the
second contradiction test, Kant allows reason a role in the setting of
ends and so allows for a certain kind of freedom. This deserves
emphasis, for one might have thought that it was inappropriate to
bring freedom in at this stage of the argument. If the only ends that
could be freely adopted were the obligatory ends themselves, any
.
claims about our capacity for freely adopting ends might seem to have
to wait upon the success of the second contradiction test itself; we
could not appeal to that capacity in making the very arguments that
 127

would be intended to establish it. But in fact this is not Kant's

view. On Kant's view, the objects of rational deliberation, the
objects upon which the will exercises its deliberative powers, are
first and foremost maxims- and both ends and means are adopted as
parts of or within maxims. As we can, as rational beings, accept and
reject maxims, so we can accept and reject the ends those maxims
contain. And quite apart from that, the descriptipn in "Conjectural
Beginning" shows how rational processes can interest us in new ends
as well as leading us to the discovery of various means.
The teleological arguments are especially relevant to the duty of
self-cultivation because Kant describes the historical process as one
in which humankindas a species is forced to develop all of its
talents and powers, in spite of the initial unhappiness this causes.
In those writings Kant explicitly regards the development of talents
and powers as a develoi:xnentof rationality and freedom. For instance,
in "Idea for a Universal History" Kant says:
Manaccordingly was not to be guided by instinct, not
nurtured and instructed with ready made knowledge; rather,
he should bring forth everything out of his own resources.
Securing his own food, shelter, safety and defense ... all
amusementwhich can make life pleasant, insight and
intelligence, finally even goodness of heart - all this
should be wholly his ownwork. In all this Nature seems
to have moved... just as if she had willed that, if man
ever did advance from the lowest barbarity to the highest
skill and mental perfection and thereby worked himself up
to happiness (so far as it is possible on earth), he alone
should have the credit and have only himself to thank -
exactly as if she aimed more at his rational self-esteem
than at his well-being. (OH: Idea for a Universal History,
14/19-20)
Later, talking about howwe are spurred to self-development by the
"unamiable characteristics of unsociability from whence opposition
springs" (OH: Idea for a Universal History 15/21), Kant declares
 128

Thanks be to Nature, then, for the incompatibility, for

heartless competitive vanity, for the insatiable desire
to possess and to rule! Without them, all the excellent
natural capacities of humanity would forever sleep,
undeveloped. (OH: Idea for a Universal History 16/21)
Were it not for these unamiable characteristics:
Men, good-natured as the sheep they herd, would hardly
reach a higher worth than their beasts; they would not
fill the empty place in creation by achieving their end,
which is rational nature. (OH: Idea for a Universal His-
tory 15/16/21)
In Religion Within the Limits of Reason Alone, these characteristics
of vanity and competitiveness are classed amongthe less happy results
of what Kant calls the 11Predisposition to Humanity11 , which is part of
the original predisposition of humannature. The 11natural purpose11
of this predisposition and even of its less amiable results is there
identified as 11a spur to culture 11
• (R22) Throughout these writings
then, the point of the development of talents is not that they will or
would bring us happiness, and not merely that they are useful for many
ends, but that they develop something that is sometimes called our
rational nature, sometimes called our humanity, and is identified with
the capacity to set an end. Keeping in mind that according to the
Doctrine of Virtue, The power to set an end - any end whatsoever -
11

is the characteristic of humanity11 (DV51/392) and in turn that an

11

end is an object of free choice 11 (DV43/384), it seems to me that the

identification of freedom(= humanity) as the end relevant to the
second contradiction test is well supported, at least in the case of
the duty of self-cultivation. This also fits well with the sort of
argument that is based on the Fonnula of the End in Itself, for there
of course the explicitly identified end which we are to further is
humanity11 or "rational nature"; which is freedom in the sense of the
11

capacity to set an end.

 129

In order to give a practical contradiction reading of the second

contradiction test, we must locate some purpose which is thwarted in
a world where talents are not developed and in a world where no one
helps anyone else. Since no purpose is stated in the maximsthat Kant
thinks fail the second contradiction test, this purpose must either
be one that is implicitly associated with the maximor one that can
be assigned to all rational beings without specific empirical
information. The difficulty with the first strategy - finding a
purpose implicitly associated with these maxims- is that it
assimilates the second contradiction test too closely to the first;
and it tends to beg the question about why the duties arising out of
the second test are in a special sense incumbent upon everyone,
regardless of her or his situation. The second strategy seems more
promising: we can combine the idea that talents and cooperation are
primary goods with some universal purpose that is therefore likely to
be thwarted if they are unavailable. We have looked at three possible
accounts, which are really very similar. On one account, we say
simply that these primary goods are needed for ends in general, and
every rational being is sure to have some end or other which requires
them. This is a sort of predictive reading - we will find a
contradiction in each case, though it may be different in each case.
On the second account, we say that these primary goods are needed for
happiness, especially given its uncertainty: this has the advantage of
making it an actual rather than a predicted end which is in question:
we knowfrom the text of the Foundations that according to Kant all
dependent rational beings do have happiness as an end. (G 33/415}
Because happiness is a second-order and uncertain end, this reading
 130

encompasses the idea of "ends in general". Finally, on the third

account, we focus on another idea that encompasses "ends in general",
namely freedom in the sense of the capacity to adopt an end, which
Kant calls humanity11 •
11
If it can be argued that a rational being must
will its own freedom, then we can argue that rational beings must
will the availability of the primary goods as means to the exercise and
genuine effectiveness of this freedom. Unless talents and powers are
ready and cooperation is available, our power to adopt and pursue ends
is severely restricted; freedom cannot be realized in a world that is
not, as it were, arranged to welcome its exercise.
Perhaps the most important thing is the fact that all of these
readings work fairly well. We can find arguments that are plausible
if not decisive for assigning ends in general, happiness, and freedom
to rational beings as such as things they will necessarily will. This
is enough to establish that a practical contradiction reading can be
given of the second contradiction test, even if there is disagreement
as to which one. As far as textual support goes, it looks as if the
happiness reading is strongest in the case of the duty of
cooperativeness and the freedom reading is strongest in the case of the
duty of cultivation. And I can see no argument against resting with
that as a final conclusion. The freedom reading has one advantage
over the other two which is of interest, especially to those who are
inclined to be convinced by Sidgwick's objection. It really does not
depend at all on empirical considerations (except those needed to
establish that talents and cooperation are indeed a sort of primary
good), for here there is no question of predicting that you will have
an end requiring these means or of arguing that you cannot be certain
 131

whether you will have such an end as part of your happiness. On the
freedom reading it is enough to say that you must will that these
generally useful means be available if you should decide on an end that
requires them, and that as one who wills her own freedom you want that
decision to be open to you. It is not a question of predicting what
ends you will or might have, and deciding whether you could or could
not achieve them without talent and cooperation. It is a question of
protecting your capacity to choose ends and your opportunity to pursue
the ends you do choose. This means that the freedom reading is in a
certain way more powerful than the other two, and immuneto a sort of
objection that is always raised to them.
In the last chapter of The Varieties of Goodness, G. H. VonWright
puts forward a description of just action which he says is similar to
Kant's view and to the Golden Rule: it involves appeals to fairness,
and arguments against giving oneself a special privilege, and it
establishes duties on the basis of arguments essentially similar to
those used by Kant. (I won't go into details since it is with Kant
that I am concerned here.} But VonWright believes that moral duties
so derived apply only within what he calls a moral community, and
that "the moral communityunder consideration consists of those who
have more to gain from never being harmed by anybody else than from
sometimes hanning somebodyelse. 1119 NowVonWright thinks that it can
be argued, as plausible fact, that anyone who is not a God will be in
this moral community. What is in question is not what we gain in
individual cases from hanning others, versus what we stand to lose in
such cases. What we "have to gain" must be thought of as what we would
gain from a policy of freely hanning others subtracting what we would
 132

lose from everyone else's having a policy of freely harming us. 20 And
in making this calculation we must take into account the fact that if
we thus placed ourselves outside of the moral communityothers who
remained part of the cooperative unit would band together against us.
Thus to be outside of the moral communityone must be stronger than all
other humanbeings combined. It is not likely that we will find
anyone that strong; and yet
Yet it is hardly an a hriori necessity that an individual
man should be weaker-tan all other men jointly. (I say
'hardly,' since we may wish to deny the name of 'man' to
a being of superhumanpowers.) Suppose a man invents
some fearful weapon, to which he alone has access and by
means of which he can wipe out any number of men who
withstand his wishes or encroach upon his privacy. He
could be a kindly man and never want to do harm to anybody.
Could he be just and moral? I shall not attempt to answer
the question. As long as he keeps his secret weapon and
is aware of the superiority, which it confers upon him in
relation to the rest of humanity, he, in any case, does
not belong to the moral community, which is determined
by the basic inequality in our example.
The fiction of the superman, although logically
possible, is yet highly unrealistic, someonemay say.
Is it highly unrealistic? I would ask. Substitute,
in our example, a team of men for an individual - and
the example is less unrealistic. Substitute a state -
and we shall recognize it as thoroughly realistic ...
But what about the imagined superman, who is not in
the community? Is it not his moral duty, too, 2fver
to do evil to his neighbor? I would answer No.
Nowalthough VonWright's case is a little different than the one
under consideration in Kant's fourth example, it is easy to see how
some such argument could be used as a sort of extended version of
Sidgwick's argument against Kant's analysis of the duty of helpfulness.
Perhaps someonewould want to argue that a rational being after the
manner of VonWright's superman, living amongus, could be certain
that whatever its happiness turned out to consist in, it could secure
that happiness without anybody's help. Hence this being need not will
 133

a world in which mutual cooperation is available.

Even if that were true, which is doubtful, however, it would only
be true because despite the uncertainties about the nature of
happiness, the superman perceived its happiness as consisting of
certain sorts of ends: those that can be had without the willing
cooperation of free and equal others. That is already a curious view
of happiness, but never mind. The point I want to make is that the
freedom reading, which emphasizes the humancapacity to set 11 ends, is
11

more resistant to this sort of objection than the others. It certainly

cannot be denied that there are ends which involve genuine community
relations amonghumanbeings either as means or as necessary conditions.
Indeed such relations may be ends, quite apart from moral
considerations. The superman may want to have a friend or a lover
or an equal partner in enterprise. Friendship and romantic love
are good examples, being amongthe ends that Kant supposes involve
moral relations between those involved, and amongthe best of those
ends which liberation from the control of instincts has enabled human
beings to invent. There are others like them. Nowwhatever one
supposes about the character of one's happiness, it seems to me that
it is easy to see that a being that wills its own freedom cannot will
to cut itself off from the whole category of such ends. Howcould a
free being claim that it would never under any circumstances choose an
end in this category? Anyonewho wills her own freedom must will a
state of affairs in which such ends can be chosen at wi11 as well as
others. There is no reason for excluding this category from the scope
of choice. And I think it must be urged that in fact, more of our
specifically humanends are of this nature than are not. This, then,
 134

is one example of the somewhatgreater power of the freedom

interpretation of the second contradiction test. 22
In any case, the second contradiction test does admit of
practical contradiction arguments, and some of them are quite strong.
What the arguments share is the idea that talents and cooperation are
a sort of primary good or general resource facilitating the adoption
and pursuit of all sorts of ends; and that any rational being, as a
pursuer of ends, has a reason for willing the sort of world in which
these primary goods or general resources are available. If the
universalization of your maximwould render them unavailable, it
cannot be willed. One whose vocation is the free adoption and pursuit
of ends cannot will a world in which the major resources for free
action in general are absent, any more than she can will in a
particular case that her action be ineffective and her own purpose
thwarted. A practically rational being must will a world for action.
 135

NOTES

1. The proper fonnulation of maxims is discussed more extensively in

Chapter 4, Section I.
2. See Chapter 4, Section III.
3. The correct specification of the purpose is very important in a
case like this. In the example, I am supposing that the thief's
purpose is to make something his property, to own it. The argument
would not apply in cases where the purpose was different, and
stealing can have a different purpose. Other cases would have to
be tested on their ownmerits. Stealing food for the purposes of
survival when, for some reason, there is no other way to get it,
for example, would not fail the test by the argument that I have
given, since there the purpose is not to own something or have it
for your own use, but to stay alive. A different purpose makes it
an entirely different kind of case.
4. John Stuart Mill, Utilitarianism, p. 310, in Warnock.
5. Marcus Singer, Generalization in Ethics, p. 259.
6. Onora Nell, Acting on Principle, p. 63.
7. See Barbara Hennan, "Morality as Rationality", pp. 104ff., for a
discussion of the basis of the hypothetical imperative. Most
other commentators on Kant that I have read make surprisingly
little of the important argument for the analyticity of the
hypothetical imperative.
8. Allen Wood, "Kant on False Promises", pp. 615-616.
9. Kant himself explicates all of the duties of virtue in terms of
the idea of treating humanity as an end in itself in the Doctrine
of Virtue. On the other hand, the arguments in the Metaphysical
Elements of Justice are not in any obvious way a series of
applications of the first contradiction test, so perhaps it will
not be thought useful to appeal to the methods of presentation
used by Kant in the Metaphysics of Morals to detennine what the
best method of presenting a given duty is. I discuss the method
of presentation used in the Metaphysical Elements of Justice in
Section III of Chapter 4.
10. Marcus Singer, Generalization in Ethics, pp. 276ff. Harrison's
idea is in "Kant's Examples of the First Fonnulation of the
Categorical Imperative", pp. 208ff. in Wolff. WhenHarrison
attributes a practical contradiction view to Kant it is by way of
criticism: Harrison takes it that Kant wants to get a theoretical
logical contradiction, but is only able to show that everyone could
not act on a certain maxim, which Harrison is sure is no
 136

contradiction at all. Singer, who is a more sympathetic reader

of Kant than Harrison, takes the suggestion more seriously.
11. Marcus Singer, Generalization in Ethics, pp. 277-278; pp. 272-273
12. John Rawls, A Theory of Justice, Section 15, pp. 92ff., Sections
60-63, pp. 395-416.
13. Sometimes, especially in the Critique of Practical Reason, Kant
seems to espouse a sort of hedonistic the.ory of happiness which
seems to me to be somewhatat odds with the 11 Ideal of the
I magination" view of the Foundations. In the Second Critique
passages, Kant sounds like Bentham: "The only thing he considers
in making a choice is how great, how long-lasting, how easily
obtained, and how often repeated this agreeableness is. 11 (C2 22/23)
This covers Bentham's intensity, duration, certainty (stretching a
point), and fecundity - and of the two remaining (for the
individual) on Bentham's list, one, propinquity, is mistaken. So
Kant has left out only purity, and he discusses that a few lines
later, as a feature of certain 11more refined joys and delights."
But this view does not seem to me to sustain what Kant says about
the uncertainty of the prudential imperatives. Pleasure is not
obviously less predictable than any other end if it is viewed
this way. In any case there are tensions among Kant's various
remarks about happiness, so it is perhaps important to note that
in saying happiness is a second order end I have the 11 ideal of the
imagination" view in mind. On the other view, everything else
might be thought of as means to the end of happiness, and then it
is not a second order end.
14. Another possible Rawlsian interpretation of the second
contradiction might be based on the idea of fairness or
reciprocity. The person who considers the maximof never helping
others has probably already accepted the help of others on general
cooperative tenns, one might say. He has already accepted benefit
from a helpful world and so ought to do his share. He has already
willed a helpful world and cannot will it away after making use of
it. In "Justice as Fairness 11 , this sort of thing is identified as
a duty of fair play. The difficulty in using this style of
argument in the context of Kant's interpretation would lie in the
dependence upon empirical fact - that a person has already
willingly accepted the benefits of cooperation. But except for
the reference to a temporally located empirical fact the idea
would be the same as in the argument now under discussion.
15. Julius Ebbinghaus, 11Interpretation and Misinterpretation of the
Categorical Imperative 11 , translated by H. J. Paton. In Wolff,
pp. 114-115. Ebbinghaus may intend the remark as a general point
about the way all of the contradiction arguments work; he does not
make the use of it that I suggest here.
16. J. Kemp,"Kant's Examples of the Categorical Imperative", pp.
240-241 in Wolff.
 137

17. John Rawls, 11Kantian Contructivism in Moral Theory: The Dewey

Lectures 1980.11 Lecture One, 11 Rationa1 and Ful1 Autonomy
11
,

p. 525.

18. John Rawls, 11Kantian Constructivism in Moral Theory: The Dewey

Lectures 1980.11 Lecture One, "Rational and Full Autonony",
p. 527.

19. G.H. VonWright, Varieties of Goodness, p. 211.

20. G.H. VonWright, Varieties of Goodness, pp. 212-213.
21. G.H. VonWright, Varieties of Goodness, pp. 213-214.
22. In this regard it is interesting to note that Mill appeals to
this idea of our capacity to 11set 11 new ends in his defense of
political and civil liberties in On Liberty. All of the
specifically humanends from which the higher quality pleasures
flow must have been discovered sometime, Mill argues, and are the
products of "originality". Because these pleasures are the best
source of happiness, the utilitarian is in favor of conditions
under which yet more of them are likely to be discovered, opening
up new sources of happiness to the humanrace. He then argues
than liberty and the encouragement of "individuality" provide the
conditions under which these important discoveries are made. Mill
also uses the appeal to the humancapacity of setting new ends in
his argument for self-cultivation: cultivating one's abilities
makes one capable of higher quality pleasure, gives one the
experience that makes one a good judge of pleasure, and may lead
to the discovery of new pleasures. Thus both Mill and Kant see
the humancapacity to take new kinds of ends and so to extend
the range of things and activities that can be valued as at
least an important source of moral duties. (On Liberty, Chapter
III)
 Chapter 4
The Leibnizian Interpretation

Immediately after the argument leading to the Fonnula of

Universal Law, Kant recasts the categorical imperative into the
Formula of the Lawof Nature:
The universality of law according to which effects
are produced constitutes what is property called nature
in the most general sense (as to form}, i.e., the
existence of things so far as it is determined by
universal laws. [By analogy], then, the universal
imperative of duty can be expressed as follows: Act
as though the maximof your action were by your will
to becomea universal law of nature. (G 39/421)
In this chapter, I discuss the significance of the Formula of the
La\',of Nature for the establishment of the claim that the categorical
imperative is a law of reason. I will argue that, in taking the law
of nature as the "Typic of pure practical judgment" (C2 70/67}, Kant
was describing the conditions in which a purely rational choice can
be made.
In the Foundations, the initial set of examples follows the
Formula of the Lawof Nature, and the universalization arguments that
I have been discussing are given in terms of it. It is in one s 1

maximconceived as a law of nature that the first contradiction is

found; in willing one 1 s maximas a law of nature that the second
contradiction is found. It is, in fact, the language associated with
the law of nature that has made Kant1 s readers suppose that either
theoretical or teleological contradictions must be what Kant has in
mind: for it is systems of nature" and "laws of nature" that "could
11

not subsist 11
, "would annihilate themselves" (C2 27/27} and so forth.

138
Such language, taken in conjunction with Kant's remarks about the use
of the law of nature as the Typic of practical judgment, leads very
naturally to the theoretical or teleological interpretation. For one
might want to say that, while the problem in one's will is ultimately
a practical contradiction, this practical contradiction can only be
discerned through the discovery of a theoretical or teleological
contradiction in the system or law of nature that corresponds to the
universalization of one's maxim. The idea behind such a reading would
be that practical contradictions are somehowvague or hard to spot,
and that theoretical or teleological contradictions in laws of nature,
which are supposedly more readily discernible, can serve as indicators
of practical contradictions in universalized maxims. In that case, it
would be a theoretical or teleological contradiction we would look
for in actually working the two contradiction tests {or at least the
first one), while it would be a practical contradiction - and so
something genuinely a part of a theory of practical reason - that would
be thus revealed.
This seems at first sight like an extremely plausible reading of
the use of the law of nature as a Typic. But I have already argued
at length that such a reading of the two contradiction tests is neither
necessary {since practical contradictions are not vague) nor especially
plausible {since the tests can be workedmore easily and with fewer
problems on a practical contradiction reading). What I will therefore
do in this chapter is offer a different reading of the point and
significance of the use of the law of nature as Typic. Since I want
to deny that the point of the Typic is the one I have described in
the preceding paragraph, somesuch reading is necessary to my case for
 140

the practical contradiction interpretation of the two contradiction

tests. Indeed, giving a reading of the purpose of the Typic is
perhaps especially important for the practical contradiction view
for just this reason. For admittedly, it might be thought that the
practical contradiction view does not makeenough of the difference
between the Formula of Universal Lawand the Formulaof the Lawof
Nature. Since I have described the contradiction tests in terms of,
for instance, the universal adoption of a policy or the standardization
of a procedure, my reading is open to the objection that the idea of
a law of nature" is not really doing any work in it.
11
This is not
true of the theoretical or the teleological interpretations. And it
is certainly clear from the section "Of the Typic of Pure Practical
Judgment11 in the Critique of Practical Reason that Kant takes the idea of
the law of nature to be doing very important work; he suggests that
moral judgment is never made and indeed could not be madewithout it:
If the maximof action is not so constituted as to stand
the test of being made the form of a natural law in
general, it is morally impossible [though it may still
be possible in nature]. Even conmansense judges in this
way, for its most ordinary judgments, even those of
experience, are always based on natural law. Thus it
is always at hand, but in cases where the causality from
freedom is to be judged, natural law serves only as the
~ of a law of freedom, for if common sense did not have
something to use in actual experience as an example, it
could make no use of the law of pure practical reason
in applying it to that experience. (C2 72/69-70)
I therefore need to say something about why, on my reading, Kant might
say that we cannot makemoral judgments without the device of
conceiving our maxims, not merely as everybody's maxims, but as laws
of nature. And in fact, I believe that I can give a much stronger
reading of the point and purpose of the Typic than the one I have
 141

described.

I. Fonnal Implications of the Formula of the Lawof Nature

So far the reading I have given has made two explicit uses of the
idea of a law of nature. One is that imagining your maximas a law of
nature is imagining that everyone with a similar purpose does act on
it - not just that they might or could, but that they do, as if by a
reflex or instinct. This makes the law of nature into a way of
insisting on the idea of universalization - which is a weak use of the
Typic, but certainly one of the things that it is meant to do. The
idea of a standard procedure is meant to capture this force of "law of
nature" - a standard procedure is not just an often-used one or a
commonone, but the obvious, automatic one, the one used unless there
is some special reason why not. To this extent a standard procedure is
like an instinct - it charts a certain course for the agent. And this
must be insisted on if we are to get the sort of contradiction the
tests require. After all, many sorts of actions that clearly should
not pass the contradiction tests!!:! very widely practiced; think for
instance of bribery in some parts of the business world, or some kinds
of academic cheating. 1 It is only when one considers these procedures
as absolutely universal and standardized that one can get the required
contradiction. The law of nature idea is, then, a way of insisting on
this. The second use I have made of the idea of the law of nature is
more significant: it can be used to answer the objection that the
practical contradiction view requires a time lapse before the
contradiction will appear, and describes a situation that could not
last rather than a situation that could not be. As I said in Chapter
 142

3, the law of nature that could not last also could not be. Laws of
nature are eternal, so even if a time lapse were needed before the
contradiction could occur, it would at any given momentalready have
passed.
It is also clear that Kant intended the idea of a law of nature
to provide a model for a maxim: not just in the sense that the maxim
is tested for its morality by its fitness to be a law of nature, but
in the sense that we can see what is and is not to be included in the
maximthat is to be tested by treating it as a potential law of
nature. Maximsmust include both the end that is to be achieved and
the means by which it is to be achieved, in order to exhibit the
purposive structure of the action. In taking a law of nature as a
model for a maxim, the effect would be the end and the cause would be
the means; the model implies that both must be included. Kant has
already made this connection between means and causes in the proof
of the hypothetical imperative: one who wills the end must will the
means precisely because willing involves regarding oneself as a cause.
Taking the law of nature as a model for the maximthat is to be tested
has implications, therefore, about the exact specification of the
maxim. These implications can, I think, be used to help resolve some
of the standard difficulties concerning the correct formulation of
maxims: for example, that maximscan be so specified that they
resist universalization and therefore pass the test regardless of
their content, or so stated that they cannot be universalized even
when laudable or innocent.
Since on Kant's view a law of nature must itself be fully
universal, the law of nature model for a maximspecifies that the
 143

maximmust describe the action in such a way that it is the real,

full, and sufficient cause of the purpose. This will in turn
guarantee that the adoption of the maximitself is the real, full, and
sufficient cause of the action: this is because a maximis our
"conception of a law". (G 29/412} Since the connection between means
and end is a connection of sufficient cause to effect, one must
include in the statement of the maximall and only what is genuinely
relevant to the production of the effect - the purpose - in the given
circumstances. A maximis a proposed justification of the action; and
it is the connection between the action and the purpose that is
supposed to justify the action. This connection is what Kant calls the
11
form11 of the maxim.
This is what Kant has in mind in the First Section of the
Foundations when he first introduces the idea of a maxim (G 16/400}.
As the examples are meant to show, two people might perform similar
actions for a different purpose - for instance, one person might knock
you over to take away your wallet, and another, in extraordinary
circumstances, might knock you over to save you from an oncoming
bullet. In the one case knocking you over is an unjustifiable assault;
in the other it is not only justified but laudable. Similarly, two
people with the same purpose might perform different actions - if the
purpose is, say, getting the best candidate elected, one might go from
door to door trying to persuade people, while another might stuff
ballot boxes. Again one is justified and the other not. This shows
that neither the type of action alone nor the purpose alone can justify
the action. The justification lies in the connection or relation
between them, and it is this that the maximis supposed to express.
 144

The connection succeeds in justifying the action only if it is

universalizable; if it is not, either the action is unjustified or
something has been left out. In order for the action to be justified,
the connection between the action and the purpose must be a sufficient
reason for the action; this is what the universalization test tests
for. A sufficient reason must be universalizable; if it is not, there
are obviously cases in which it is not sufficient and we h~ve not
yet found out whether we are in such a case or not. A basic reason
for an action (a candidate for a sufficient reason) is that the action
is the cause of an end one wants to achieve: the relevant connection
between action and purpose is causal. If the connection is one that
can always exist, both cause and reason are sufficient. A cause is
therefore our model for a reason: hence the law of nature test. But
this means that what is included in the maximmust stand in a
genuinely causal relation to the achievement of the purpose in the
circumstances, and that the cause must be fully stated. Thus, it is
because our maximis our conception of a law that the causal
connections implied or asserted by the maximmust be both ones that we
genuinely believe in and that are genuinely motivating us - they must
be the real reasons for our actions.
This provides the basis for argument against the claim that
maxims can be "tailored" so that they pass the universalization test.
One can tailor a maximby making it so specific that it is essentially
resistant to universalization: for instance, one can claim that among
the circumstances influencing one to perform a certain action is the
fact one has the social security number: 359-40-1082. Then when one
universalizes, the action is to be performed only by those who have
 145

that number, and since only one person has that number, the situation
remains the same. But if the maximis modelled on a causal law, the
possession of just this numbermust be shown to influence the causal
features of the situation; for example, it might be shown that your
having just this number structures the causal situation so that the
proposed action is the only way to achieve the intended purpose.
But of course having that social security number is really only causally
revelant to a quite limited and highly specific set of activities -
writing certain digits in the blank following the words "social
security number" on a form, for instance, where it is one's purpose to
fill out the form correctly. There is only a very well-defined set of
effects of which the possession of just those digits in one's social
security numbercan be said to be the cause, and only in certain
kinds of circumstances do they influence the ordinary causal relations
between actions and purposes, and so find a place amongthe relevant
circumstances included in maxims. One's social security number is a
reason to write certain digits on fonns; but it is unlikely to be a
reason for any commonlywrong actions.
Barbara Hermanpoints out that we can test for the relevance of
what is included in maxims by means of a counterfactual question test:
The tailored maximis, again, "WheneverI am in
financial difficulties, I will make a deceitful promise,
in order to borrow money, if the promisee is red-haired,
28 years old, and it is a Thursday" - and it is proposed
as a possible 'correct' maximof the act of deceiving
a red-haired 28 year old (Arthur) next Thursday, in
order that George may extricate himself from his financial
embarrassments. The procedure for blocking this variety
of tailoring will be to construct a set of counterfactual
questions to test the relevance of the various features
in the descriptive portion of the maxim- that is, those
parts of the maximwhich exhibit departures from a wholly
general characterization, e.g., deceiving for personal
 146

gain, etc. For example, we will want to knowhow

George would answer, 11lf Arthur did not have red hair,
would you still make the deceitful promise?" and "If
the occasion for making the promise turned up on Tuesday
instead of Thursday, would you still make the deceitful
promise?" If the answers to these questions were
affirmative, then we would have excellent grounds for
declaring 1 red-headedness 1 and 'being on a Thursday•
irrelevant to a descr~ption of what it is the agent
(George) wills to do.
Willing is regarding oneself as a cause: the test reveals whether
the agent really believes that what is included in his maximis
causally relevant to the achievement of his purpose, and so whether
the inclusions are really motivationally relevant to what he wills.
A similar problem arises when the maxim's specificity seems to
render perfectly innocuous actions wrong. A standard classroom
objection is that maximssuch as, say, the maximof taking steps to
becomea doctor cannot be universalized because everyone cannot be a
doctor. Here, the intuitive problem is that "being a doctor" is an
insufficiently motivational description of the purpose of adopting
the maximof taking steps to becomea doctor (which is not to say that
being a doctor cannot be desired for its own sake). The motivational
description must include something about why being a doctor is valued
for its own sake, and it will make a difference why. Wanting to be a
doctor is one reason for adopting the maximof taking steps to become
a doctor, but the objection that there would still be a
universalization problem if everyone, by some fluke, wanted to be a
doctor, is correct. On a Kantian account a mere inclination is never
a sufficient reason for the adoption of an end (and whenever you adopt
a maximyou adopt its end). The agent is expected to take into
account, and to include in her maxim, factors such as whether she can
 147

do good and exercise her talents by this career choice. If everyone

wants to be a doctor and this agent has no special vocation for the
business then this agent will not do good and exercise her talents, and
will be wrong to be a doctor merely because she wants to. In a case
where taking steps to becomea doctor is right, the motivational
description wi11 include these factors: and certainly we can
universalize a maximof adopting a career which one enjoys and in
which one can do good and exercise one's talents. 3
So even in the case where the end is valued for its own sake the
motivational grounds of its being so valued must be included and must
be of the right sort. The entire description in the maximmust be one
that is motivationally relevant, and only some descriptions are. If
we view the contradiction tests as ones in which a basic reason is
tested for its sufficiency, then it is clear that what we start with
must be a basic reason. This will only be the case if both the end
and the means are correctly given: the end must be motivationally
specified, and the relation of the action to the end must be causal.
In many cases of apparently "tailored" maxims it will be unclear which
of these two things has gone wrong. Thus, Hermanpoints out with
respect to her example that:
It is not impossible to imagine a case where Arthur's
having red hair would be part of the circumstances that
gave someonea motive for acting: suppose George has
an irrational grudge against all red-haired people
(having been injured by a red-haired bully as a child),
and took every opportunity to cause them injury. But in
such a case we would do well to rethink the structure of
the maxim, as it is no 1onger clear what kind of action
is being contemplated. 4 --
In this case, where George would only cheat a red-haired person, we
cannot make out any causal connection between George's purpose
 148

{obtaining moneyto get out of a jam) and the red-headedness of

George's victim. Red-haired people are not better sources of money.
So in this case, we must move the factor of red-headedness into the
purpose of the action; then we must, as Hermandoes, seek some
motivational description of this purpose (cheating the red-headed)
which by itself makes no obvious sense. 5
The idea that a maximis our conception of a law of nature is,
I think, supposed to bring out both of these points about the
specification of maxims. The relation of action to purpose must be
genuinely causal, and the purpose must be the real and fully specified
effect aimed at by the will, the thing of which the will conceives
itself as a cause.
Combiningthe idea that the maximmust include both means and
purpose (cause and effect) with the idea that what we are seeking is a
practical contradiction also helps to block another form of the
objection that good or innocent actions cannot pass the universalization
tests. One of the favorite idle games of Kant's more hostile critics
is to combine the theoretical contradiction interpretation with a
misstated maximto show that highly commendableactions will turn out
to be wrong. Bradley, for example, comes up with this:
The essence of morality was a similar contradiction.
'Negate the sensuous self'. But if the sensuous self
is negated, possibility of morality disappears.
Morality is thus as inconsistent as theft. 'Succor
the poor' both negates and presupposes (hence posits)
poverty: as Blake comicafly says:
Pity would be no more,
If we did not make somebodypoor.
If you are to love your enemies, you must never be
without them; and yet you try to get rid of them.
Is that consistent? In short, every duty which
presupposes something to be negated is no duty; it
is an immoral rule, because self-contradictory. 6
 149

Singer, who goes to the trouble of refuting cases of this structure

one by one, mentions two other similar cases. One comes from
Brentano: we must not refuse bribes, for the maximof never accepting
a bribe would destroy the practice of bribery, and in a world where
there were no bribes we could not refuse them, and so could not act on
the-maxim. Another, traced to Frank ChapmanSharp and Lucius Garvin,
says that we must not be vaccinated against a communicabledisease if
the vaccine must be obtained from people who have the disease, for if
everyone is vaccinated the disease will be stamped out, no more
vaccine available, and so vaccination impossible - it cannot be
universalized. 7 As Singer so patiently shows, the purpose of these
maximsis anything but defeated by their adoption, so if we remember
to include the purpose, and to look for a practical contradiction with
respect to this purpose, we will certainly find no contradiction.
Amongthe formal implications of the Formula of the Lawof Nature,
then, are all of these things. One of the things that I have mentioned
is of special importance for the practical contradiction view - that
the fact that laws of nature are eternal helps to block the objection
that a time lapse is needed for a practical contradiction to emerge.
It is important on any interpretation to emphasize that the univer-
salization of a maximis really to be seen as a universalization,
not merely a generalization, of the procedure. The maximis to be
imagined as implanted in us by a natural instinct. In addition, the
idea that the maximis one's conception of a law and is to be regarded
as a candidate for a law of nature provides guidance in the correct
fonnulation of maxims. The purpose must be an effect aimed at by the
will, and the action and circumstances must be so described that in
 150

those circumstances the action is the cause of the purpose. It is

this that shows that the purpose is at least a basic or possible
reason for the action, and so a candidate for a sufficient reason for
the action. For it is the sufficiency of our reasons for actions for
which the categorical imperative tests.

II. The Leibnizian Interpretation

The significance of the idea of selecting a maximas if it were
to be a law of nature goes beyond these formal features. The point
of the Typic is not just to provide us with a formal model for our
maxims, but also to provide us with a perspective - a point of view
from which our maxims can and should be assessed. It is in terms of
this perspective that we are to understand why it is rational to act
according to the categorical imperative.
The author of a law of nature or a system of nature is in a
special position when assessing candidates for laws of nature and
judging their possi:bility and their value as such. The position of
one who is creating the laws of nature is unrestricted in a particular
way, for the author of nature is not limited by any considerations of
what is naturally possible. Since what is naturally possible is not
settled until after the laws of nature are chosen, the only constraints
on the choice of these laws are purely rational constraints: the
author of a law of nature need only consider whether the law is a
reasonable one. Thus it is a position from which one's choice can be
said to be guided by pure reason. I will call this the world-creating
position; for the creator of laws and systems of nature is indeed a
creator of the world.
 151

The significance of the world-creating position is very great

in Kant's moral philosophy. The world-creating position is a model
of the perspective from which a purely rational choice is made. It
is, in fact, one of a numberof such models that appear in Kant's
ethical writings. Besides the Typic, two other models are of great
importance: the model of the legislating subject of the Kingdomof
Ends in the Foundations; ~nd the model of the free being choosing his
own character in Religion Within the Limits of Reason Alone. What
these models or images have in commonis that they represent attempts
to describe a perspective from which choice is constrained only by
reason. All represent positions in which the laws and principles of
choice themselves are being chosen. They are different ways of
representing what I will call the Standpoint of Practical Reason.
And I will say that Kant's strategy for determining whether pure reason
can be practical and what its dictates are if it is practical is to
take up the Standpoint of Practical Reason and to see what guidance
from reason can be found there.
The world-creating position has a special place amongthe various
representations of the standpoint of practical reason because it is
evident that Kant thinks it is the one from which we can get the most
detailed and determinate knowledgeof the dictates of pure reason. This
is why it is selected for use as the Typic. The position of
legislating subject in the Kingdomof Ends and the character-choosing
position have different roles. The former is used to account for the
specific sense of obligation and the idea of humandignity that are
associated with moral decision: our dignity comes from the fact that
we are bound to the law only because it is our own law and we are its
legislators; while we are obligated by it because we must also be
regarded as subjects in the Kingdomof Ends. (G 57-58/438-439) The
character-choosing position, as I will show later (Chapter 7), is used
to illuminate the idea of autonomyand the sort of motivation we have
for acting according to the dictates of pure reason. But it is from
the world-creating position that we can best see what those dictates
are. In the rest of this section, I will elaborate further on the
significance of this position and explain its special usefulness as a
Typic for the moral law.
In his discussion of the Typic in the Critique of Practical
Reason Kant insists both on the agent's position as the supposed
creator of an order of things and on the independence of the agent's
decision from any considerations of the physical possibility of the
object willed. The actual use of the Typic is introduced in this
passage:
The rule of judgment under laws of pure practical
reason is: Ask yourself whether, if the action which
you propose should take place by a law of nature of
which you yourself were a part, you could regard it as
possible through your will. Everyone does, in fact,
decide by this rule whether actions are morally good
or bad. Thus people ask: if one belonged to such an
order of things that anyone would allow himself to
deceive when he thought it to his advantage, or felt
justified in shortening his own life as soon as he was
thoroughly weary of it, or looked with complete indifference
on the need of others, would he assent of his own will to
being a memberof such an order of things? Noweveryone
knows very well that if he secretly permits himself to
deceive, it does not follow that everyone else will do
so, or that if, unnoticed by others, he is lacking in
compassion, it does not mean that everyone else will
immediately take the same attitude toward him. The
comparison of the maximsof his actions with a universal
natural law, therefore, is not the determining ground
of his will. But such a law is still a type for the
estimation of maximsaccording to moral principles. If
the maximof action is not so constituted as to stand
 153

the test of being made the form of a natural law in

general, it is morally impossible [though it still may
be possible in nature]. {C2 72/69-70)
Here the emphasis is simply on whether the agent could agree to be
part of a certain order of things. A parallel passage in Religion
Within the Limits of Reason Alone brings out the idea of world-creation
more strongly:
Take a man who, honoring the moral law, allows the
thought to occur to him {he can scarcely avoid doing
so) of what sort of world he would create under the
guidance of practical reason, were such a thing in his
power, a world into which, moreover, he would place
himself as a member. He would not merely make the
very choice which is determined by that moral idea of the
highest good, were he vouchsafed solely the right to
choose; he would also will that [such] a world should
by all means come into existence... {RS)
Because this passage from the Religion speaks of a "choice which is
determined by that moral idea of the highest good" (RS), it is worth
emphasizing that the 11Typic11 section in the Critique of Practical
Reason occurs as a part of the chapter on the concepts of good and
evil: "The Concept of an Object of Pure Practical Reason11 • The Typic
is explicitly declared to be a Typic for the "law of freedom" and
consequently for the concept of the absolutely good'' (C2 71/68):
11

the highest good, we learn later, is the "entire object of a pure

practical reason." (C2 113/109) Thus the Religion passage is about
the same thing as the passage in the Typic section of the Critique of
Practical Reason; by the time he reaches both remarks he has already
established that the good must be determined in accord with the law,
must refer to what can be chosen in accord with the law.
Whenwe take up the perspective described by the Typic, then,
the point is that we are to think of what sort of a world we would
 154

create under the guidance of practical reason - where a "sort of

world" is specified in terms of the system of laws characterizing
a world. This is what one would expect in Kant, for he thinks that
without laws the various things that we might experience (or rather,
since this would not quite be experience, intuit) would not be knit
together into a single, self-identical, unified world. A consistent
and ordered and complete set of laws form the criterion for the
identity of a world, just as, we will see later, a consistent and
ordered and complete set of maxims forms the criterion for the identity
of a character. In this, Kant works within the tradition of Leibniz,
who imagines God choosing the best of all possible worlds by choosing
a certain set of laws or principles:
Thus we may say that in whatever manner Godmight have
created the world, it would always have been regular
and in a certain order. God, however, has chosen the
most perfect, that is to say the one which is at the
same time the simplest in hypotheses and the richest in
phenomena,as might be the case with a geometric line,
whose construction was easy, but whose properties and
effects were extremely remarkable and of great
significance.a
Furthermore, it is an absolutely central feature of the
Leibnizian philosophy that when God is envisioned as choosing the
world, God is envisioned as doing so, not arbitrarily, but in
accordance with reason. This is something Leibniz, arguing against
those who believe that the truths of reason are themselves productions
of the will of God, was prone to insist on: 9
In saying, therefore, that things are not good according
to any standard of goodness, but simply by the will
of God, it seems to me that one destroys, without
realizing it, all the love of Godand all his glory;
for why praise him for what he has done, if he would
be equally praiseworthy in doing the contrary? Where
will be his justice and his wisdomif he has only a
 155

certain despotic power, if arbitrary will takes the

place of reasonableness, and if in accord with the
definition of tyrants, justice consists in that which
is pleasing to the most powerful? Besides it seems that
every act of willing supposes some reason for the willing
and this reason, of course, must precede the act. This
is why, accordingly, I find so strange the expressions of
certain philosophers who say that the eternal truths of
metaphysics and Geometry, and consequently the principles
of goodness, of justice and of perfection, are effects
only of the will of God. To me it seems that all these
follow from his understanding, which does not depend upon
his will any more than does his essence.10
For to think that God acts in anything without having any
reason for his willing, even if we overlook the fact
that such action seems impossible, is an opinion which
conforms little to God's glory. For example, let us
suppose that God chooses between A and B, and that he
takes A without any reason for preferring it to B. I
say that this action on the part of God is at least
not praiseworthy, for all praise ought to be founded
upon reason which~ hypothesi is not present here. My
opinion is that God does nothing for which he does not
deserve to be glorified.11
Leibniz's God, unrestricted of course by any limitations of physical
possibility, chooses a world by choosing a set of laws, and does so for
good and sufficient reason. It was for Leibniz's readers one of the
central points of controversy whether, in necessarily always choosing
for a sufficient reason, God chose freely. 12 Leibniz, and those who
followed in his path, thought that God certainly did.
This picture of a free and rational choice was taken up by Kant
in his ethical theory. So the point of the Typic might be described
this way: it puts us in the position of Leibniz's God, and therefore
in the position of making a free and rational choice. Since from this
position we can make a choice that is unconstrained by any but rational
principles, it is from this position that we can make a choice that is
at once purely rational, and, arguably, perfectly free: an autonomous
choice. Thoughwe make the choice with an end in view - the end in
the maxim, in the case of the first test; and ends-in-general, or
happiness, or freedom, in the case of the second test - we are free
to choose that principle of action on which we would act in a world of
our own rational choosing, rather than merely the one dictated by the
circumstances in which we find ourselves. Weassess the principle for
its rationality as a practical principle, not just for its efficacy
in the achievement of this particular end. The choice we make is a
rational choice, since it is for its rationality that we choose the
principle; and a free choice, since we do not let the circumstances
dictate what our maximwill be. While this understanding of the
Typic helps us to get a grasp on both the rationality and the freedom
of choices made from the perspective provided, it is primarily on the
idea of rationality of such choices that I want to focus here. I
will discuss the connection of this sort of choice to freedom further
in Chapter 7.
The idea that the position described is one in which we are
unconstrained by physical necessities (or limits in the physical
possibilities) is one that Kant comes back to a number of times in the
Critique of Practical Reason, and is an important feature in his
argument for the pure rationality of choices made from this position.
For example, early in the chapter on Goodand Evil Kant says:
If the object is taken as the determining ground of our
faculty of desire, its physical possibility through the
free use of our strength must precede the decision as to
whether it is or is not an object of the practical reason.
But if, on the other hand, the a priori law can be regarded
as the determining ground of action, which is consequently
seen as determined by the pure practical reason, then the
judgment as to whether or not something is an object of
the pure practical reason is wholly independent of any
question of our physical ability; the only question is
whether we should will an action directed to the existence
 157

of an object if it were within our power.

Consequently, the moral possibility of the action
takes precedence, for in this case it is not the
object but the law of the will which is its ground
of determination. (C2 59-60/57-58)
And again, later:
Since in all precepts of the pure will it is only a
question of the determination of the will and not of
the natural conditions (of practical ability) for
achieving its purpose, it thereby happens that the
practical concepts a priori in relation to the supreme
principle of freedom immediately becomecognitions,
not needing to wait upon intuitions in order to
acquire a meaning. This occurs for the noteworthy
reason that they themselves produce the reality of that
to which they refer (the intention of the will) - an
achievement which is in no way the business of
theoretical concepts. (C2 68/66)
These passages are both from the "Concept of an Object of Pure
Practical Reason." In this chapter, it is Kant's aim to answer the
criticism (of the Foundations), made by Pistorius, that - in Kant's
words - "the concept of the good was not established before the
moral principle, as in his opinion was necessary." (C2 9/8-9) They
form part of a general address to this problem whose outline is
familiar to readers of the Foundations: if it were settled first
what the good is, and only then what the moral principle is, we would
have only hypothetical imperatives and so no moral principle at all.
But in these chapters Kant is not merely talking about the best way to
locate or define a principle of right or obligation - he is interested
in the good. So here the emphasis and point of the argument is
slightly different from that of the Foundations: we want the good to
be a rational and universal concept (C2 60/58), but it will not be so
unless we define the moral principle first and thereby get ourselves a
principle in terms of which we can identify a rational and universal
 158

conception of the morally good - namely, that the morally good is what
is chosen by, or is the choice of, the good-willed person who is
governed by the moral law. And Kant thinks it is an important part of
his argument here - the argument about the good - and this means that
the decision as to what is good will not be preceded by any decision
about what is physically possible. SometimesKant makes it sound as if
he thinks this is because what is good ..:i!the intention of the will
(see the second of the two quotations just given) and this good thing
can be produced directly by the good-willed person. But this is only
true of what is unconditionally good. In other passages the argument
is that what is good is what the good-willed person would will were
it within her power. (As in the first of the two quotations just given.)
Thus the point is that the judgment whether a thing is good need not
be preceded by a judgment whether the thing is physically possble:
goodness is not some (theoretically identified but very interesting)
property of the objects around us, but a purely rational characteri-
zation of something as part of the world that ought to be. The
decision as to a thing's goodness is a rational and practical decision,
completely independent of whether the thing can be brought about.
It is one made on the basis of rational principles alone. Such deci-
sions are made by means of the Typic as I have described it: you ask
yourself whether it belongs to the sort of world you would rationally
will if you were a rational creator in the position of Leibniz's
God.
I have already said that the main point of the argument about
being unconstrained by considerations of physical possibility is that
it leaves us, as it were, with no basis for choice except pure
 159

rationality itself. I have also said that such a position is

represented by any situation in which what you are doing is choosing
the laws of choice themselves. Nowthis is the position of someone
who is considering the adoption of a maxim- that person is choosing
the law that will govern her actions and choices. From the standpoint
of the desire for happiness, in the real world, you think primarily
about whether your maximwill work - whether it will get you both the
particular object that you nowseek and your happiness as well. That
decision, though made with reason, is not one of pure reason since your
choice is limited by the physical possibilities. But when you take up
the standpoint of practical reason, only rational constraints govern
the choice of your principle of choice. These rational constraints are
supposed to be represented by the two contradiction tests; and they
represent the only constraints reason places on a principle of choice
that is to be a law; that it not contradict itself, and that it not
contradict a rational will.
The idea that what the Typic tells us to do is to put ourselves in
the position of Leibniz's God in order to make a purely rational choice
can be extended to explain why there are two tests and what the two
tests are supposed to do. Leibniz's Godmakes the choice of the best
of all possible worlds on the basis of two principles: the principle of
non-contradiction and the principle of sufficient reason. The principle
of non-contradiction determines what is a possible world, and so
determines the set God is to choose from; the principle of sufficient
reason dictates that one choose the best of these; and as a fully
rational being Godcan do no less. Nowthese two principles of pure
rational choice correspond to the two contradiction tests. The first
 160

determines whether a maximis even a possible one in a world that you

would will for yourself; the second, taking the world of the
universalized maximas certainly possible, determines whether it is the
one which, as rational, you would will. In Leibniz's metaphysics, this
will be the most metaphysically perfect one: simplest in hypotheses
but richest in phenomena. But Kant says that there are two notions of
perfection;
... perfection, can be taken in either a theoretical or
a practical sense. In the former, it means nothing
more than the perfection of anything in its own kind
(transcendental perfection), or the perfection of a
thing merely as a thing generally (metaphysical
perfection); and we need not discuss these here. The
concept of perfection in its practical meaning, however,
is the fitness of or sufficiency of a thing to any kind
of ends. This perfection, as a characteristic of man
and thus as internal, is nothing else than talent, or
skill, which strengthens and completes talent. The
supreme perfection in substance, i.e., God (hence
external), when regarded practically, is the sufficiency
of this Being to all ends in general. (C2 41-42/41)
Fromthe practical point of view, the perfection either of your own
will or of the sort of world you would will were it within your power
can also be understood in terms of sufficiency for any sort of ends.
If your talents are developed and you have a policy of helpful and
cooperative behavior, your will approaches this sufficiency. If you
were willing a world, choosing laws for it, from a practical point of
view, you would give that world the laws that would make it sufficient
for any sort of ends - the more sufficient, the more perfect. So you
would will that everyone would develop their talents rather than not,
that everyone would be helpful rather than not, by way of willing the
most practically perfect "order of things."
This enables us to read the two contradiction tests in terms of
 161

Leibniz's two principles of reason. In the first test, you are in the
position of Leibniz's God, looking for a world that is not
contradictory, amongthose in which a certain end - the one in your
proposed maxim, is to be brought about. In the second test, you are
in the position of Leibniz's God, choosing amongnon-contradictory
worlds the most practically perfect. Fromthe standpoint of practical
reason, the first of these two judgments (whether your maximis
contradictory when universalized) is analytic. The second (whether
your maximcan be willed as universal law) is synthetic. Leibniz,
who believes all truth to be in some sense necessary(!_ priori),
nevertheless distinguishes the two sorts of truth arising from his two
principles of reason: what is true by virtue of non-contradiction is
necessary, whereas what follows from God's choice of the actual world
as the best, as opposed to some other possible world, is contingent,
though! priori and certain. 13 Similarly, Kant will say in the
Metaphysics of Morals that, from the point of view of pure practical
reason, the first principle of the doctrine of law is analytic,
whereas the first principle of the doctrine of virtue is synthetic.
The concept with respect to which the first principle of Law is
14
analytic and that of virtue is synthetic is freedom (DV56/396).
This is what makes it possible for there to be a pure rational choice
from the standpoint of practical reason, and what makes it possible to
represent this choice in terms of the position of an author of nature.
Weneed only rememberthat a world without practical contradiction is a
world for action, a world for freedom.
In Theory of Justice, Rawls makes the following remarks about this
approach to ethical theory:
 162

Somephilosophers have thought that ethical first

principles should be independent of all contingent
assumptions, that they should take for granted no
truths except those of logic and others that follow
from these by an analysis of concepts. Moral
conceptions should hold for all possible worlds.
Nowthis view makes moral philosophy the study of
the ethics of creation: an examination of the
reflections an omnipotent deity might entertain
in detennining which is the best of all possible
worlds. Even the general facts of nature are to be
chosen. Certainly we have a natural religious
interest in the ethics of creation. But it would
appear to outrun humancomprehension. From the
point of view of the contract theory it amounts to
supposing that the persons in the original position
knownothing at all about themselves or their world.
How, then, can they possibly make a decision? A
problem of choice is well defined only if the
alternatives are suitably restricted by natural laws
and other constraints, and those deciding already
have certain inclinations to choose amongthem.
Without a definite struT;ure of this kind the question
posed is indeterminate.
Nowsince I have claimed that Kant's theory is in a very important way
a theory of this type (in the sense that it is a theory that appeals to
the ethics of creation of the best of all possible worlds, and one in
which the constraints of natural law are in a certain way ignored - but
not, of course, in the sense that it presumes no truths but those of
logic and analysis), I would like to say something about how Kant's
view escapes the difficulties Rawls mentions.
As I have described the actual workings of the two contradiction
tests, neither of them is indetenninate
11 11
in the way Rawls describes.
This is especially obvious in the case of the first contradiction
test, where both criteria Rawls mentions for a well-defined problem of
choice are obviously met. The situation that the agent is to consider -
the world of the universalized maxim- is given in terms of the laws
of nature, plus this one other law, the universalized maximitself.
 163

Whenwe try to think what this situation would be like, our thinking
is governed by our knowledgeof the laws of nature in manyways. For
example, in the promising case, we take into account the psychological
law that people are not likely to go on accepting promises they know
will not be kept. People learn from experience, and they are not
boundlessly generous: these facts are, if you will, contingent.
Indeed, this is sometimes offered as a criticism of Kant, a proof that
his test is not pure of empirical considerations in the required way -
although I hope that what I am about to say will provide an answer to
that objection. Again, I have already described how Kant's readers
are concerned over the empirical 11 character of the claim that we need
11

one another's help for the pursuit of ends in general. Any consideration
of whether a universalized procedure will work must take into account
empirical considerations of how things work. In all the cases where
the maximfails because the action proposed involves the abuse of a
practice, the 11
empirical 11 fact that practices cannot exist unless they
have the support of proper use by a sufficient number of people and of
an open social acknowledgementplays a role in showing what the
contradiction is. Wherewe try to find a contradiction in the
universalization of the use of some natural act for a certain purpose,
psychological facts carry an even heavier weight. In short, in working
the contradiction tests, the proposed new law is "plugged in" to the
world as it is, and general facts about the world as it is are taken
into account. This is true of both tests. The agent is in the
position of Leibniz's God, but only with respect to a world consisting
of this one law, a certain proposed psychological law of humannature,
plus other laws as they actually are.
 164

11
There is also, in each case, an inclination
11
that detennines
whether or not the world of the universalized maximis choiceworthy.
In the first case, the inclination
11 11
is the one involved in the
maximitself - the choiceworthiness of the world of the universalized
maximis measured directly in tennsof the satisfiability of one's
purpose by one's proposed method in that world. Yet because you
choose, not merely a principle that will get you your end in this
world, but the principle upon which you would act to achieve your end in
a world of your ownchoosing, we could say that the touchstone for
your choice is the free pursuit of this end. In the case of the second
contradiction test, I argued, there are a variety of ways to characterize
the purpose that serves as a touchstone for the decision:
11 11
ends-in-
general, happiness, freedom. One reason for favoring the reading
according to which the purpose in question is freedom is the fact that
the concept with respect to which the first test is analytic and the
second synthetic is freedom. In either case we could say that the
touchstone for the will's choice is its ownfreedom: in the one case
it chooses not to act against freedom, in the other case it chooses
the best world for freedom. There is a touchstone for the choice, so
that it is not impossible to make. Andyet neither the presence of
this touchstone nor the employmentof our knowledgeof the laws of
nature is a violation of the formality and purity of the test. The
choice is still one in which one is constrained only by the rational
necessities. It is practical contradiction that stops you from willing
the universalization of the rejected maxim. Wemust appeal to
theoretical facts to determine whether there is a practical
contradiction in willing the world of the universalized maxim, but
 165

these theoretical facts do not govern our decision. They only

describe what object is to be decided about. And yet what is being
decided about is a world of a certain sort. You are not detennining
whether something is physically possible in this world, but whether
something is morally possible - possible in a world for action. In the
case of the first contradiction test, one of the things in the world
you are considering is you, acting on the maximin question, having the
purpose in question: and in the case where the maximfails, you,
frustrated in your attempt to achieve your purpose. In the case of the
second contradiction test, one of the things in the world is you,
trying to act in general, trying to achieve ends. In either case, the
choice for or against the world of the universalized maximis made by
an absolutely pure practical reason. And the only question that pure
practical reason need ask itself in order to settle the issue is: is
the world of the universalized maximin this case a world for action?
Is it the best of all possible worlds for action?
Even for Leibniz's God, a pure rational choice is a choice of
what is perfect. In the realm of the practical, for Kant, perfection
is the sufficiency of a thing to any sort of end. Thus in detennining
the perfection of a world in the practical sphere, we ask about its
sufficiency to any sort of end. In the first contradiction test there
is a given end, and we ask about that. But a world in which everyone
constrained their actions by the first contradiction test would be
sufficient to all of the ends in question. And the second contradiction
test determines the world's sufficiency for ends in general - it is
better to say, for the pursuit of ends in general. What the two
tests have in commonis the idea of choosing a best possible world for
 166

action. In either case, we look to the idea of a world in which we

are to act, to set and to pursue ends. The result of everyone acting
morally is a world perfect for acting in. It is a world for the free
rational pursuit of ends. The Kingdomof Ends - the whole of rational
beings as ends in themselves as well as of the particular ends which
each may set for himself {G 51/433), is Kant's model of this perfect
world for action. Everyone acts only on universalizable maxims. Where
there are practices, they are not abused, and since the purpose of a
practice is to smooth the path of the humanpursuit of ends, this is a
central feature of a perfect world for actton. Indeed, the simplest
and most effective practices may be used - whereas one of the most
severe practical limitations we face in the real world is the need to
design practices that cannot be abused {and so the need, often, to
settle for ones that are less effective and efficient than more easily
abused ones are). Since talents are developed and cooperativeness a
policy, the resources of action are present. In the Kingdomof Ends
you can set any end that you like (within the limits of moral
acceptability) and pursue it in the most efficient and effective
manner. You can decide for yourself what your happiness will consist
in and you are free to pursue it. It is the will 's ideal.
Nowthe explanation why the choice of a world for free rational
action is a pure rational choice is this: free rational action is the
work of the will: it is the function of practical reason. Thus in
making the perfection of the world for action the touchstone of its
choice the pure practical reason chooses only what is essential to the
will's - its own, that is - functioning. The function of practical
reason is to bring reason into action - to act rationally, on maxims-
 167

and what I have described is simply reason choosing according to what

the optimumconditions for its own functioning are. The laws that
reason chooses are the laws of its own optimal functioning, or the laws
for the will 's optimal functioning - the laws of the world in which the
will can set and pursue ends.
A parallel with theoretical reason will make the point here
clearer. While I think this point could be made in terms of the
specific doctrines and arguments of the Critique of Pure Reason, I do
not need to go into extensive detail here, and it will serve my
purposes better to make the argument in a simple way. Even the few
passages I have already quoted from Leibniz's Discourse of Metaphysics
ring with Leibniz's conviction that the sort of world created by
rational Godmust be rational throughout. Everything has a reason,
happens for a reason, exists for a reason. Wemay not always be able
to see it but it is so. The world makes sense, all the way down. Were
it not for our own limitations, it could be understood, explained, and
justified in every detail. This is the principle of sufficient
reason, and the very essence of metaphysics.
Nowin the Critique of Pure Reason this compelling vision is a
major source of Kant's worry. Perhaps we knowwhat it would be like
for the world to be reasonable all the way down, but have we any way
to ascertain whether it is? Howdo we knowthat the world as it is
in itself confonns to the demandsof reason? And of course Kant's
answer will be that we only knowit of the world of experience, and
only knowit of that world because we cannot have experience except
insofar as our intuitions are ordered in a certain way. Beyond that,
reason can make only regulative demands, teaching us to seek in the
 168

world the sort of systematic order that would make the world make sense
to us and seem a reasonable place. The principles of reason,
governing the understanding, aim at making the world an understandable
place, an explainable place, a comprehensible place. In place of
metaphysics, in Leibniz s sense, we have regulative principles
1

governing scientific investigation. Instead of knowledge that the

world is rationally ordered, we have rules governing our quest for
rational order in the world. These rules are aimed at making the
world a world for understanding.
The point of this simpleminded statement of Kant s view is to
1

emphasize the parallelism of function of rational principles in the

two cases. Reason has two employments- theoretical and practical -
which are given by the regulative principles governing the understanding
on the one hand and the moral law on the other. The understanding-
under-regulative-principles and the will are, as it were, the
representatives of reason in its two domains. In both cases, the
function of rational principles .:ii to enable these faculties of reason
to do what they do. Regulative principles governing the understanding
are aimed at making the world a comprehensible place - so that we can
indeed understand it, insofar as that is possible to us. Moral
principles are aimed at making the world a world for the free rational
pursuit of ends - so that we can freely and rationally pursue ends in
it. Understanding and action are rational activities. They are the
characteristic activities of humanbeings as the only knownrational
beings. The function of the principles of reason is to make these
activities possible. Indeed, principles of reason can be seen as
descriptive of these sorts of activities: we are not really
 169

understanding and not really acting unless we are understanding and

acting as these principles prescribe. Kant says:
Pure reason is in fact occupied with nothing but itself
It can have no other vocation. (Cl 556; A680/B708)16
Whenwe consider a rational principle from the point of view of its
function, it becomes clear, then, that there is no violation of
formality in making its fitness for a world for free rational action
the touchstone of a universalized maxim's reasonableness. This simply
reflects what practical reason~-
This analogy between the regulative principles of the theoretical
reason and the moral laws extends to an actual connection when, in the
Critique of Pure Reason and the Critique of Judgment, we learn that
reason in its two employmentshas the same aim and ideal - that of
purposive unity, a complete teleological system organized around a
final purpose.
This highest fonnal unity, which rests solely on
concepts of reason, is the purposive unity of things.
The speculative interest of reason makes it necessary
to regard all order in the world as if it had
originated in the purpose of a supreme reason. Such
a principle opens out to our reason, as applied in the
field of experience, altogether new views as to how
the things of the world may be connected according
to teleological laws, and so enables it to arrive
at their greatest systematic unity. (Cl 560; A686-687/
B714-715)
A systematic and purposive unity of things is the ideal of both
employmentsof reason because it is what it would take to make the
world both fully comprehensible to us and fully justified - because
these two things are, in the end, one and the same. We could fully
explain things only if we knewwhy they must be so, why they could
not be other than they are - but such an explanation would have to
 170

appeal, not to a mere cause, but to a purpose for whose sake things
are organized as they are. Of a mechanistic law, at however high a
level of generality, one can always ask why it is so and not
otherwise; it is only a final purpose that can give a final answer to
the demandfor the reason why of things. But a final purpose would
provide not only a complete explanation of things for the satisfaction
of the theoretical understanding, but a complete justification of
things from a practical and moral point of view. In the Critique of
Judgment, we learn that the final purpose for both the theoretical
and the practical purposive systems is the same: it is humanbeings
under moral laws, who are happy in proportion as they are virtuous.
I will give part of that argument from the Critique of Judgment later
on (see Chapter 5), but there is one consequence of it that must be
mentioned here.
I have claimed that the point of the Typic is to place us in the
situation of Leibniz's God choosing the best of all possible worlds, so
that we can make a choice that is based upon pure reason. In trying
to explain why in this case this approach does not give rise to the
difficulties about the constraints of a determinate problem of choice
mentioned by Rawls, I pointed out that in working the two contradiction
tests the laws of nature as they actually are do play an informing
role - not in the sense of constraining our choice to what is
physically possible, but in the sense of enabling us to envision the
world we would be choosing if we chose the world of the universalized
maxim- of providing us with a detenninate object of choice. Insofar
as what we are choosing from the standpoint of practical reason is a
world for action this does not seem to me to violate the formality of
 171

the tests. The question is not whether the world we envision and
decide about is physically possible, or desirable from the point of
view of an inclination, but whether it has a practical contradiction
in it.
But there is something more satisfactory than this to be said in
the end. For it turns out from the standpoint of practical reason
there is something to say about what all the laws of nature ought to
be, even the laws of non-humannature. Specifically what the general
facts of nature ought to be does, as Rawls puts it, outrun human
comprehension, yet we can say that there is one feature they will have
in the best of all possible worlds for action. This feature is that
they will be so arranged that every person's happiness is in proportion
to her moral virtue. Exactly what arrangement would achieve this result
we cannot know, but that the best of all possible worlds for action
must have this arrangement is a dictate of pure practical reason. 17
To this extent we can see from the standpoint of practical reason even
what the general facts of nature must be - at least what they must be
1i ke.

This is the doctrine of the Dialectic of the Critique of Practical

Reason, on which Kant bases his claims about the moral basis of
religious faith. Granting that virtue and happiness in proportion to
virtue are features of the entire object of a practical reason (see
Chapter 5, Section I), we can say that the point is this: in the best
of all possible worlds for action, any good end - every good end - can
be achieved. It is for humanreason to set the end and find the
means to achieve it, and to do this we must have knowledgeof and
make use of the laws of nature. Rememberthe proof of the principle
 172

behind hypothetical imperatives - the humanagent regards herself as

a cause, and so uses the available causal connections to bring about
the end. The laws of nature make humanrational agency possible.
Without a knowledge of them, we could not bring about or even aim at
any ends. Yet the laws of nature, not being chosen by us, might also
seem to make some purposes impossible of achievement. The laws of
nature might be such that no effort of ours would be sufficient, or
even such that every reasonable effort we make might go entirely
wrong. By "the niggardly provision of a stepmotherly nature" a will
might be "wholly lacking in power to accomplish its purpose" and "the
greatest effort" might not "avail it to achieve anything of its end."
(G 10/394) Then we would be in a condition of alienation, unable to
foresee or control the results of our efforts. Our intentions might
be good, perfect - and yet this would not be a world for action.
Recalcitrant nature would have disarmed humanagency, making the
achievement of our ends impossible. In order to act, we need sufficient
cooperation from the laws of nature .
... his effort is bounded; and from nature, although
he may expect here and there a contingent accordance,
he can never expect a regular harmonyagreeing
according to constant rules (such as his maxims are
and must be, internally) with the purpose that he yet
feels himself obliged and impelled to accomplish.
Deceit, violence, and envy will always surround him,
although he himself be honest, peaceable, and kindly;
and the righteous menwith whomhe meets will,
notwithstanding all their worthiness of happiness,
be yet subjected by nature, which regards not this,
to all the evils of want, disease, and untimely
death, just like the beasts of the earth. So it
will be until one wide grave engulfs them together
(honest or not, it makes no difference) and throws
them back - who were able to believe themselves the
final purpose of creation - into the abyss of the
purposeless chaos of matter from \'lhich they were drawn.
The purpose, then, which this well-intentioned person
 173

had and ought to have before him in his pursuit of moral

laws, he must certainly give up as impossible. {C3 303}
In the best of all possible worlds for free rational action, this
cooperation will be a feature of the laws of nature; for without
enough cooperation to make humanagency real and effective, this world
cannot be a world for action at all.
Fromthe standpoint of pure practical reason, then, we choose a
world for action: the best of all possible worlds for the free rational
pursuit of ends. This is what pure practical reason, without any
other specific purpose, aims at and demands. The point of the Typic
is to place us in the standpoint of pure reason, so that the choice
of our maxims can be made in a way that is governed only by rational
principles. By taking up the position of the author of an order of
nature in which we are to act, we put ourselves in a position to make
a decision about our maximthat is based on reason alone.

III. Practical Reason and Justice

I have claimed that when we take up the standpoint of practical
reason, and attempt from that standpoint to choose our maximson the
basis of reason alone, the two rational bases of our choice correspond
to Leibniz's two basic rational principles. Exercising the principle
of non-contradiction, we look for maximsthat do not contradict
themselves practically when universalized; maximswhose purposes can
still be realized in the world of the universalized maxim. Exercising
the principle of sufficient reason, we seek a world that is as perfect
as possible a place for humanaction - for the free rational adoption
and pursuit of ends. Weadopt such policies and procedures as we
 174

would choose as laws for a world for action, were it in our power to
choose. By taking up a standpoint in which our choice is based purely
on these rational principles rather than on any of our private
interests or inclinations, we are enabled to make a choice that is
based on pure reason, to adopt maximsand perform actions on a purely
rational basis. By acting the way we would act in a world of our own
rational choosing, we exercise our autonomy. Instead of allowing
ourselves to be constrained by the actual circumstances in which we
must function, instead of being channelled along whatever path among
the circumstances leads to the satisfaction of our desires, we act
just as we would were we able to choose the very world in which we
are to act. The consequences may sometimes not be what we would have
them, but our actions are our very own, the ones we would perform in a
world for action. Wedo this by acting according to the laws of a
world for action. These laws, modelled on the laws of nature, say
that our use of means to ends must be universalizable, like the
relation of a sufficient cause to an effect, in order that there be a
s uffi ci ent reason for our act ions ..
The Typic is a Typic for the objectively good. One could say
that when we take up a practical perspective towards the world, we
view things in terms of good and evil. It is because these are
intelligible, rather than perceptible or scientifically ascertainable,
features of things that we must use the Typic to descry them. Since
good and evil have to do with action and purpose, means and ends, we
take cause and effect as our models for these. The will is a
causality belonging to rational beings; a means is a cause, a purpose
an effect of the will, a maxim, giving the connection of means and end,
 175

is the agent's conception of a causal law, a law for a world for

action. The practical perspective is one from which the \·1orld appears
as the arena for the free pursuit of ends, and objects and practices
and procedures in the world are available tools for the achievement
or realization of good ends. Fromthe standpoint of practical reason,
the world is viewed as a place in which every individual is to invent
and pursue a happiness, and things are seen in relation to this
invention and pursuit.
In the Dialectic of the Critique of Practical Reason, as I have
mentioned, Kant speaks of what this means as far as our assessment of
nature is concerned: what it would be like for nature to provide us
with a perfect world for action, one in which we were not alienated
from the results of what we do. Since we have no power over the laws
of nature, this is a subject for hope and for faith. In the Introduction
and first part of the Metaphysical Elements of Justice, Kant gives an
account of the practical necessity of the state that is also based on
the idea of what is necessary to make the world into a world for
action. Kant in a sense presents the state as a kind of moral or
practical entity by means of which humanbeings transfonn the physical
world of land and what land produces into the means of action, and
justice as a condition of a world rendered practical. In this
section I will describe these ideas, for I think that they lend
support to the argument of this chapter. I also think that I can
thus give a reading of why the unexpected emphasis on property in
this book turns up.
The argument of the Metaphysical Elements of Justice starts out
to be about external freedom. A condition of justice is a condition
 176

in which a rational external freedom is realized amonghumanbeings.

Kant introduces what he calls a "universal principle of justice" that
runs:
Every action is just (right) that in itself or in
its maximis such that the freedom of the will of
each can coexist together with the freedom of
everyone in accordance with a universal law.
( DJ 35/230)
or:
act externally in such a way that the free use of
your wil 1 is compatible with the freedom of everyone
according to a universal law. (DJ 35/231)
According to Kant the only innate and original right is the right to
freedom, and any other rights that humanbeings have must be derived
from this. From the right to freedom, however, Kant immediately
develops a set of "proprietary rights" divided according to their
objects. One can have a proprietary right in a thing, which is one's
possession; in the will of another with respect to a particular
11
act",
as whenyou have been promised something; and in the status of another
11

in relation to oneself" - as when you enter into legally governed

relations such as marriage, parenthood, employment, etc. (DJ 54-55/
247-248) Nowthe argument of the first part of the Metaphysical
Elements of Justice follows this pattern: Kant first undertakes to
establish the idea of proprietary rights on the basis of the categorical
imperative and the innate right to freedom; then, having established
that these are all natural rights, he tries to show that we are both
authorized and obligated to leave the "state of nature" and enter into
a juridical condition in which these rights will be fully realized and
protected. The point of the existence of the state, and the basis of
the requirement of allegiance to it, is the protection of the right to
 177

freedom and the associated proprietary rights. Since it can be

argued that the point of proprietary rights is to facilitate action,
it will follow that the point of the state is to protect a condition
which is designed for action.
For my purposes the interesting part of this argument is the
connection between freedom and the idea of a proprietary right, for it
is here that the idea of rendering the world practical comes in. Kant
establishes the idea of proprietary rights on the basis of what he
calls "the juridical postulate of practical reason" which runs:
it is possible to have any and every external object of
my will as my property. (OJ 52/246)
Kant explains the juridical postulate as follows:
In other words, a maximaccording to which, if it
were made into a law, an object of will would have
to be in itself (objectively) ownerless ... conflicts
with Lawand Justice.
[The reason for this postulate is as follows.]
An object of my will is a thing that I have the
physical power to use. Let us suppose that it were
absolutely not within my power de jure to make use of
this thing, that is, that such power would not be
consistent with the freedom of everyone in accordance
with a universal law. In that case, freedom would be
robbing itself of the use of its will in relation to
an object of the same will inasmuch as it would be
placing usable objects outside all possibility of
being used. In other words, it would reduce these
objects to nought from a practical point of view and
make them into~ nullius, although formally the will
involved in the use of these things is still consistent
with the freedom of everyone in accordance with
universal 1aws. ( OJ 52/246)
Consequently, it is an a priori assumption of practical
reason that any and every object of my will be viewed as
and treated as something that has the objective possibility
of being yours or mine. (DJ 53/246}
The juridical postulate is meant to establish the concept of property
or proprietary right. The argument here is negative: if a thing
 178

cannot be someone's property, it is reduced to nothing from a practical

point of view, and there is no reason to restrict our freedom by
robbing ourselves of the available means of action arbitrarily. The
point about property is this: from the practical point of view the
things in the world appear to us as means of action, and property is
necessary to rendering an object into such a means. That is why if a
thing cannot be property it is rendered res nullius from a practical
point of view: it is not a means. The reason that something must be
property in order to be, effectively, a means, is that it is not really
possible to act, to carry out plans and projects, without some
authority over the means of action that goes beyond the mere power of
what Kant calls physical detention. Kant explains:
Thus, for example, I do not call an apple mine simply
because I hold it in my hand (possess it physically), but
only if I can say: 1 possess it even when I let it out
11

of the hand that is holding it. 11

Similarly, I cannot say
of the land on which I am camping that it is mine just be-
cause I am camping on it; I can say that it is mine only if
I can assert that it is in my possession even if I leave the
place in question. The reason for this is that, in the
first case (of empirical possession), if someonewere to
wrench the apple out of my hand or to carry me off from the
place where I was camping, he would not injure me with respect
to my external property, although, of course, he would injure
what is internally mine (my freedom). But he would not in-
jure me as far as my external property is concerned unless I
could also claim to have possession of the object even with-
out detention of it; therefore, in the present case, I can-
not call these objects (the apple and the camp) mine. (DJ
54/247-248)
If internal freedom is to have any external realization at all, it
must be possible to use physical objects as means. For this to be so,
you must not merely be able to detain them physically, but to have
some sort of authority over them. The apple and the camp are not
resources of action for the person from whomthey can be wrested away
 179

without violation of right or freedom. Intuitively the idea is very

simple: in order to effectively carry out your plans, you must be
able to eat the apple you have picked, or to save it for later; to
plant crops in your 11camp11 , and leave them without losing them; to
have some mastery over the physical world beyond what you can obtain
by brute strength. The argument does not establish full-fledged
rights to private property, but only that things (and other people's
actions) be at some times and in some places guaranteed to an
individual 1 s use. Such a guarantee is a proprietary right, rendering
the thing or action a property
11 11
and thereby rendering it a means of
action. Property, like other moral or practical notions, is an
intelligible concept, and its derivation from the juridical postulate
of practical reason is actually a Kantian Deduction:
The possibility of this [nonempirical] kind of
possession and the Deduction of the concept of
nonempirical possession are founded on the juridical
postulate of practical reason: It is a duty of
11

justice to act towards others so that external objects

(usable objects) can also becomesomeone's [property]".
The possibility and the Deduction [of nonempirical
possession] are at the same time bound up with the
exposition of the latter concept [of property], which
grounds external property on nonphysical possession
alone. The possibility of nonphysical possession cannot
in any way be intuited as true (simply because it is a
concept of reason for which no corresponding intuition
can be given). Instead, its possibility follows
directly from the aforementioned postulate, for, if it
is necessary to act according to this basic principle
of right and justice, then the intelligible condition
(of a mere de jure possession) must also be possible.
It should surprise no one that the theoretical
principles of external property become lost in the
intelligible world and represent no advance in knowledge,
for the possibility of the concept of freedom, on which
they rest, is not susceptible of theoretical Deducation
and can only be deduced from the practical law of
reason (the categorical imperative) as a fact [Faktum]
of practical reason. (DJ 60/252)
 180

The point of proprietary rights, then, is to render the physical

world into an arena for action by rendering the land and the other
objects in the world into means of action. The point of the state is
to protect and realize these proprietary rights. The state is a
morally-practical entity whose purpose is to establish and protect a
world practically conceived - a world of means and ends. The state is
a device by which humanbeings transform the world of land and objects -
the world of nature - into a world of means and resources - a world
for action.
 181

NOTES

1. For this point and the examples used to illustrate it, I am

indebted to the students in my seminar on Kant's ethics at
Yale in the Spring of 1980.
2. Barbara Herman, "Morality as Rationality", p. 67. I am guided
by Herman's discussion throughout this section.
3. In Chapter 6, I discuss the question of the extent to which there
is room in the Kantian theory for action that is permissible and
governed or motivated by one's inclinations rather than moral
considerations. This example and what I say about it suggests
that the obligatory ends must always play a role in decisions
not involving considerations of justice. Hhat I say later comes
downless hard on this point. I am not certain what to think
about this case.
4. Herman, "Morality as Rationality", p. 68.
5. Perhaps it will be thought, however, that there are no such
limits on what counts as a motivational description of the
purpose. It is possible to ascertain whether there is a genuinely
causal relation or an imagined causal relation between action and
purpose in many cases: so it will be possible to tell whether the
agent really thinks the action is a cause of the purpose. But are
there any constraints on what purposes an agent might have? Maybe
it will be thought that the purpose could be just anything - that
one could want to cheat the red-haired or hurt people on Thursdays.
Given the freedom to set an end, any end whatever, it might seem
that any purpose is possible and might be in a genuine maxim. It
would be odd if my rock-bottom reason for wanting to hurt someone
was that his social security number had certain digits or his
hair were a certain color, but if I am free to adopt any end
cannot it be that I adopt the end of hurting someonewith a
particular numberor a particular hair color? Here there are a
number of things to say. One is that just as the action must bear
some intelligible relation to the purpose, so the purpose must be
one that is intelligible for a humanbeing to have: it must be one
that in Anscombe's terms has a point. (Intention, ss. 37ff.)
Another might be that the end is one whose adoption is motivated
by a possible incentive of the will. To apply this criterion we
must have some sort of psychological or anthropological theory about
what the possible incentives of the will are. I think it is
arguable that Kant provides such a theory, though a very spe~ulative
one, in his description of our "predisposition" in Religion Within
the Limits of Reason Alone. The three parts of the predisposition
are the three sources of incentives of the will, and Kant even says
that each by itself would motivate the will. Motivation by any of
these incentives is psychologically (theoretically) intelligible,
even if the action is wrong. But all of the possible incentives
 182

are supposed to be covered by the various parts of the

predisposition. A humanbeing can no more be motivated by just
anything that she can believe just anything: even mistakes are of
particular kinds. Perhaps the most important thing to say,
however, is the thing that Hermaninsists on - the maximmust
after all be true, the real one. ("Morality as Rationality",
pp. 62-65) Regardless of what might be conceivably possible, we
do not in fact usually want to hurt people because of the color
of their hair or the digits in their social security number.
6. F. H. Bradley, "Duty for Duty's Sake", p. 155.
7. Marcus Singer, Generalization in Ethics, pp. 279-292.
8. Leibniz, Discourse on Metaphysics, Section II, pp. 4-5.
9. See Monadology,sections 46 (p. 261), 53, 54 (pp. 262-263) for
parallel passages. In 46, Leibniz mentions Descartes and Poiret
as those who believe that eternal truths are dependent on God.
10. Leibniz, Discourse on Metaphysics, Section II, pp. 4-5.
11. Leibniz, Discourse on Metaphysics, Section III, pp. 6-7.
12. See Leibniz's Correspondence with Arnauld.
13. Leibniz, Discourse on Metaphysics, Section XIII, p. 22.
14. A more complete discussion of the two principles and the claim
that the one is analytic and the other synthetic is in Chapter 6.
15. John Rawls, A Theory of Justice, Section 26, pp. 159-160.
16. See Aristotle, The NicomacheanEthics, Book IX, Chapter 8:
" ..• for reason in each of its possessors chooses what is best
for itself, and the good man obeys his reason."
17. Wecannot knowwhat arrangements would bring about this result,
because of the intelligible nature of virtue. If nature were so
set up that happiness were always a causal consequence of virtuous
action, those who never acted from duty but always, as it happened,
according to it, would be happy even if they were thoroughly evil,
and so the result would not be achieved that way. Such a person
is possible, according to Kant,
... when reason employs the unity of the maxims in
general, a unity which is inherent in the moral law,
merely to bestow upon the incentives of inclination,
under the nameof happiness, a unity of maximswhich
otherwise they cannot have... The empirical character
is then good, but the intelligible character is still
evi 1 . ( R 32)
 Chapter 5

Humanityas an End in Itself

In this chapter I will be concerned with the Formula of the End in

Itself, which reads:
Act so that you treat humanity, whether in your own person or
in that of another, always as an end and never as a means only.
(G 47/429)
In the summaryof the three formulae, Kant explains that this fonnulation
gives the material of the moral law. (G 54-55/436)
The argument leading up the Fonnula of the End in Itself, like the
others, is an argument with the structure: if there is a categorical
imperative or practical law of reason, this is what it must be like.
This means that a careful examination of this argument and others that
are related to it ought to show us why a law of practical reason must
have just this material or end - it ought to show us why humanity is the
objectively rational end. Mypurpose in this chapter is to explicate
this argument, establish humanity as the objective end, and to show that
this is the sort of material appropriate to a rational principle.
In the case of this fonnulation, however, there are special chal-
lenges. To say that humanity is the right sort of material for a ration-
al principle and is an objectively rational end will already be start-
ling to those who suppose that Kant's claim to have put forth a principle
of pure practical reason rests mainly on the formality of the categorical
imperative. Kant himself incessantly insists on the contrast between
"material" and fonnal throughout the ethical writings, and yet he
11 11

183
 184

unquestionably says that humanity is an objective end and the material

of the moral law. I suspect it is usually believed that the claim that
the categorical imperative is a principle of reason rests squarely on
the Formulaof Universal Law, (which is "fonnal 11 ) and the claims of the
other formulae to be rational principles are based on their presumed
equivalence with the Universal Lawformulation. The Formula of the End
in Itself is intuitively more attractive and (despite what Kant says)
may be found easier to work with in directing concrete moral thinking.
So I suppose that doubters are inclined to take the following view:
that the material formulation is attractive and captures manyof our
moral intuitions, while the Universal Lawformulations sustain claims
about rationality but do not give clear results, and Kant allows too
muchto rest on the presumption of equivalence.
I think it can be shownthat such suspicions are misguided, and that
the claim that humanity is an end in itself can be supported on rational
grounds independent of the presumedequivalence of the different formu-
lations. Most importantly, I want to argue that the notion of humanity
as an end in itself adds something to Kant's case for the rationality
of the categorical imperative that will be missed if the Formula of the
End in Itself is treated merely as the more intuitively attractive ver-
sion of the principle. Wecannot understand what Kant means by acting
from duty or being motivated solely by reason until we have understood
what is involved in making humanity one's end. It is in terms of the
idea of objective ends that Kant describes humanvirtue and explains
what it is to act from duty rather than merely in accordance with it.
And it is in terms of the idea of objective ends that we can answer the
Humeanchallenge, and show in detail howpure reason can set an end and
motivate the will.
 185

Accordingly, in this chapter I will discuss the idea of humanity as

an end in itself, and in the following chapter I will deal with the role
of this idea in explaining action from duty - perfectly rational action.
· Once these arguments are in place, someof the objections to Kantian
ethics that are based on the emphasis on ·formali'smwill' be dealt with.
In particular, at the end of the next chapter, I will have something to
say about two extremely troublesome issues: the way in.which Kant deals
with the role of emotions, passions, and affections in moral motivation
and the virtuous life; and his rigorism
.
about
. exceptionless
. . moral
. rules,
such as the duty not to lie. I do not intend to challenge the usual
interpretations of what Kant thinks about these subjects, but only to
. .
place his reasons for taking these positions in a new light. Once we
understand what Kant thought was involved in the act of making humanity
one's end, his claims about the nature of virtue in the· Doctrine of
Virtue and other later writings will be seen to be important parts of
the theory of pure rational action. This, in any case,. is what I hope
to show.

I. TwoDistinctions in Value
Since the argument leading to the establishment of the Formula of
the E~d in Itself depends upon the notion of unconditioned goodness, it
is necessary first to take a look at that notion. The Foundations opens
with the claim that:
Nothing in the world - indeed nothing even beyond the world
- can possibly be conceived which could be called good
without qualification except a good will. {G.9/392-393)
As Kant presents the teleological argument, it becomesclear that what
he means is that the good will is the only unconditionally good thing
and 11the supreme condition to which the private purposes of menmust for
 186

the most part defer. 11

(G 12/396} Again:
This will must indeed not be the sole and complete good but
the highest good and the condition of all others, even of the
desire for happiness. (G 12/396}
Happiness, by contrast to the good will, is referred to as a "conditional
purpose. 11 {G 12/396}
The unconditioned/conditioned distinction in value is to be under-
stood (i) by reference to other uses of the distinction, in the theore-
tical realm, and {ii) in contrast with the means/end distinction. I will
start with the latter. The means/end distinction can be said to be a
distinction in the way we value things {states of affairs, activities,
circumstances, and conditions in the more ordinary sense - as for ins-
tance happiness, luxury, adventure, status). That is, we can value
things as means or!! ends, depending upon whether we value them for the
sake of something else or for their own sake. Exampleswould be of the
usual sort: tools, instruments, appliances, money, chores, would usually
be valued as means; happiness, beautiful objects, enjoyable and rewarding
activities, as ends. By contrast, the conditioned/unconditioned distinc-
tion is a distinction not in the way we value things but in the circum-
stances (conditions) in which they are objectively good. A thing is
unconditionally good if it is good under any and all conditions, if it
is good no matter what the context. In order to be unconditionally good,
a thing must obviously carry its ownvalue with it - have its value in
itself (be an end in itself). "Intrinsic" value would be a natural term
for this, except that in our tradition "intrinsic" has so often been
opposed to "instrumental" that it would be misleading, since a thing that
is not unconditionally good can be valued as an end: I will therefore
simply avoid the use of the term. Nowa thing is conditionally valuable
if it is good only whencertain conditions are met: if it is good
 187

sometimesand not others. Thus, to elaborate on Kant's own examples,

"the coolness of a villain makes him not only far more dangerous but also
more directly abominable in our eyes than he would have seemedwithout
it", {G 10/344) whi1e coo1ness in a genera1 or a surgeon is usually an
excellent thing. Power, riches, and health are good or not depending
upon what use is made of them. To say that a thing is conditionally
valuable is to say that it is good when and only when the conditions of
its goodness are met. Wecan say that a thing is good absolutely
{really} either if it is unconditionally good or if it is a thing of
conditional value and the conditions of its goodness are met. Here it
is important to notice that "good absolutely" is a judgment applying to
real particulars: this woman'sknowledge, this man's happiness, and so
on. To say of a thing that it is good absolutely is not merely to say
that it is the type of thing that is usually good {knowledgeand happi-
ness are usually good) but that it contributes to the actual goodness of
the world: here and now the world is a better place for this. This we
do not say about the coolness of the villain or the knowledgeof the
traitor or the happiness of the evil man. WhenKant says that the only
thing good without qualification is a good will, he means that the good
will is the one thing for which the world is always a better place, no
matter 11what it effects or accomplishes." {G 10/394)
The two distinctions interact in the following ways. Whena thing
is valued as a means {or is the sort of thing valued as a means}, it will
always be a conditionally valuable thing, and the goodness of the end
will be a condition of the goodness of the thing valued as a means. In-
struments therefore can only be conditionally valuable. If the condi-
tions of their goodness are met, however, they can be good absolutely.
The more important point is about things valued as ends. These are, with
 188

the exception of the objective ends, also conditionally good. In parti-

cular, happiness, under which Kant thinks our other private purposes are
subsumed, is only conditionally good, for:
It need hardly be mentioned that the sight of a being
adorned with no feature of a pure and good will, yet enjoying
uninterrupted prosperity, can never give pleasure to a rational
impartial observer. Thus the good will seems to constitute
the indispensable condition even of worthiness to be happy.
(G 9/393)
Nowthis is where we must appeal to the other uses of the unconditioned/
conditioned distinction in Kant to understand the full force of what he
is saying. If anything is conditioned in any way, reason seeks its con-
dition, continually regressing upon the conditions until it reaches
something unconditional. It is this characteristic activity of reason
that generates the antinomies of theoretical speculative reason described
in the Critique of Pure Reason. The most accessible example is causal
explanation - if we explain a thing in terms of its cause, we then go on
to explain the cause itself in terms of its cause, and this process con-
tinues. Reason does not want to rest until it reaches something that
needs no explanation (although this turns out not to be available): say,
something that is a first cause or its owncause. A causal explanation
truly satisfying to reason would go all the way back to this evident
first cause, thus fully explaining why the thing to be explained is so.
These are familiar sorts of movementsin philosophy, so there is no need
to belabor the point. To apply it here, it is only necessary to point
out that just as to really explain a thing we would have to find its
unconditioned first cause, so to really justify a thing (where justify
is "show that it is good") we would have to show that all the conditions
of its goodness were met, regressing upon the conditions until we came
to what is unconditioned. Since the good will is the only uncondition-
ally good thing, this means that it must be the source and condition of
 189

all the goodness in the world; goodness, as it were, flows into the world
from the good will, and there would be none without it. If a person has
a good will, then that person's happiness (to the extent of his or her
virtue) is good; if the person's ends are good then the means to those
ends are good. This is why the highest good, later identified as the
whole object of practical reason, is virtue and happiness in proportion
to virtue: together these comprise all that is good - the unconditional
good and the private ends that are rendered good by its presence. Hence
also the Kingdomof Ends, as "a whole of rational beings as ends in
themselves as well as of the particular ends which each may set for him-
self" (G 51/433), is a kingdomin which the absolute good is fully
realized. 1 The Good is a system of ends, synthesized by the categorical
imperative, including the totality of all that is absolutely good under
the unconditioned good - a practically rational systematic whole.
As pure practical reason it likewise seeks the unconditioned
for the practically conditioned (which rests on inclinations
and natural need); and this unconditioned is not only sought
as the determining ground of the will but, even when this is
given (in the moral law), is also sought as the unconditioned
totality of the object of pure practical reason, under the
nameof the highest good. (C2 112/108)
It is this purely rational conception of goodness (the unconditionally
good and the conditionally good of which the unconditioned condition is
met) that Kant works with in the argument leading to the Fonnula of the
End in Itself.

II. The Argumentfor the Formulaof the End in Itself

After discussing the Formulaof Universal Law, Kant begins to raise
the important question of the rational motivation for applying these
standards to our maxims:
The question then is: Is it a necessary law for all
 190

rational beings that they should always judge their actions

by such maximsas they themselves could will to serve as
universal laws? If it is such a law, it must be connected
(wholly a priori) with the concept of the will of a rational
being as such. (G 44/426)
But here it is a question of objectively practical laws and
thus of the relation of a will to itself so far as it deter-
mines itself only by reason; for everything which has a rela-
tion to the empirical automatically falls away, because if
reason of itself alone determines conduct it must neces-
sarily do so a priori. The possibility of reason thus deter-
mining conduct must now be investigated. (G 45/427)
This investigation begins with some considerations about ends:
The will is thought of as a faculty of determining itself
to action in accordance with the conception of certain laws.
Such a faculty can be found only in rational beings. That
which serves the will as the objective ground of its self-
determination is an end, and, if it is given by reason alone,
it must hold alike for all rational beings .... The subjective
ground of desire is the incentive, while the objective ground
of volition is the motive. Thus arises the distinction be-
tween subjective ends, which rest on incentives, and objective
ends, which depend upon motives valid for every rational being.
Practical principles are formal when they disregard all sub-
jective ends.•.. (G 45/427)
Kant then goes on to point out that if we could find an objective end,
we would have the basis for a categorical imperative, applying to all
rational beings. Reciprocally, if the "objective ground of the will 1s
self-determination 11
is an end, then it looks as if we need some objec-
tive and rational end in order to have a categorical imperative that can
motivate the will. This objective end must be something that has its
value in itself. Kant then proceeds to argue that it must be "man, and
in general every rational being11 11humaniti' or "rational nature. 11

(G 46-47/427-429)
The argument by which Kant reaches this conclusion depends upon the
distinction of unconditioned and conditioned value described in the pre-
ceding section. In fact it takes in part the form of a regress upon
conditions:
 191

All objects of inclinations have only a conditional worth,

for if the inclinations and the needs founded on them did not
exist, their objects would be without worth. The inclinations
themselves as sources of needs, however, are so lacking in
absolute worth that the universal wish of every rational being
must be indeed to free himself completely from them. There-
fore, the worth of any objects to be obtained by our actions
is at all times conditional. Beings whose existence does not
depend on our will but on nature, if they are not rational
beings, have only a relative worth as means and are therefore
called "things"; on the other hand, rational beings are desig-
nated "persons" because their nature indicates that they are
ends in themselves, i.e. things which may not be used merely
as means. Such a being is thus an object of respect and, so
far, restricts all [arbitrary] choice ..•.
For, without them, nothing of absolute worth could be found,
and if all worth is conditional and thus contingent, no supreme
practical principle for reason could be found anywhere.
{G46-47/428)
In this passage Kant appears to be looking around at a world viewed
practically to see what in it might be objectively and necessarily valu-
able. It is the first part of this search that reads like a regress
upon the conditions: movingfrom the objects of our inclinations, to
the inclinations themselves, finally {later) back to ourselves, our
rational nature. The final step, that rational nature is itself the
objective end, is reenforced by this consideration:
The ground of this principle is: rational nature exists as
an end in itself. Mannecessarily thinks of his own existence
in this way; thus far it is a subjective principle of human
actions. Also every other rational being thinks of his exis-
tence by means of the same rational ground which holds also
for myself; 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. {G47/429)
I have quoted these somewhatobscure passages so extensively because I
would like to attempt a reconstruction of this argument. The reconstruc-
tion does not precisely follow the text but I hope will capture Kant's
meaning. What follows is this somewhatspeculative reconstruction.
In the discussion of good and evil in the Critique of Practical
Reason (C2 59-74/57-71), Kant discusses what he refers to as an old
 192

fonnula of the schools: Nihil appetimus, nisi sub ratione bani; nihil
aversamur, nisi sub ratione mali. If this is taken to mean that "we
desire nothing except with a view to our weal or woe" it is "at least
very doubtful. 11
But if it is read as saying "we desire nothing, under
the direction of reason, except in so far as we hold it to be good or
bad" it is "indubitably certain." (C2 61-62/59-60)2 Similarly, in the
Foundations Kant says that "the wi11 is a faculty of choosing only that
which reason, independently of inclination, recognizes as practically
necessary, i.e., as good." (G 29/412) Insofar as we are rational agents
we will choose what is good - or take what we choose to be chosen as good.
Suppose then that you make a choice, and you believe what you have
opted for is a good thing. Howcan you justify it or account for its
goodness? In an ordinary case it will be something for which you have
an inclination, something that you like or want. Yet it looks as if the
things that you want, if they are good at all, are good because you want
them - rather than your wanting them because they are good. For "all
objects of inclinations have only a conditional worth, for if the inclina-
tions and the needs founded on them did not exist, their objects would be
without worth." (G 46/428) The objects of inclination are in themselves
neutral: we are not attracted to them by their goodness; rather their
goodness consists in their being the objects of humaninclinations. (I
would say: goodness in objects is not a theoretical property we perceive
and respond to, but a practical characterization we bring to the world
of objects.)
This, however, makes it sound as if it were our inclinations that
made things good. This cannot be right, for "the inclinations themselves,
as sources of needs, however, are so lacking in absolute worth that the
universal wish of every rational being must be indeed to free himself
 193

completely from them." (G 46/428) Nowwithout necessarily endorsing

what Kant says here, we can easily agree that there are some inclinations
of which we want to be free: namely, those whose existence is disruptive
to our happiness. Take the case of a bad habit associated with an habi-
tual craving - it would not be right to say that the object craved was
good simply because of the existence of the craving when the craving
itself is one that you would rather be rid of. So it will not be just
any inclination, but one that we choose to act on, that renders its
object good.
Even consistency with our own happiness does not make the objects
of inclination good, however. (NowI am, admittedly, disconnected
from the passage I am interpreting - for Kant leaves it at the undesir-
ability of having inclinations at all.) This is partly because we are
not certain what our happiness consists in, but more because of a claim
that has already been made in the opening lines of the Foundations: we
do not believe that happiness is good in the possession of one who does
not have a good will. This is, of course, our great tanptation - to
believe that our own happiness is unconditionally good. Andyet it is
not really a tenable attitude, and the one whoattempts to hold it is
certainly guilty of making an exception of himself. For either one must
have the attitude that just one's own happiness is unconditionally good,
which is rather a remarkable feat of egocentricism, 3 or one has to have
the attitude that each person's happiness is unconditionally good. But
since "good" is a rational concept and "what we call good must be, in
the judgment of every reasonable man, an object of the faculty of desi re 11
(C2 62-63/60-61), we cannot rest with the position that everyone's happi-
ness, whatever it might be, is absolutely good. For:
 194

Thoughelsewhere natural laws make everything harmonious, if

one here attributed the universality of law to this maxim,
there would be the extreme opposite of harmony, the most
arrant conflict, and the complete annihilation of the maxim
itself and its purpose. For the wills of all do not have one
and the same object, but each person has his own {his own
welfare), which, to be sure, can accidentally agree with the
purposes of others who are pursuing their own, though this
agreement is far from sufficing for a law because the occa-
sional exceptions which one is permitted to make are endless
and· cannot be definitely comprehendedin a universal rule. In
this way a harmonymay result resembling that depicted in a
certain satirical poemas existing between a married couple
bent on going to ruin, "Oh, marvelous harmony, what he wants
is what she wants"; or like the pledge which is said to have
been given by Francis I to the EmperorCharles V, "Whatmy
brother wants (Milan), that I want too." (C2 27-28/28)
Given that the good must be a consistent, harmoniousobject of rational
desire and an object of the faculty of desire for every rational being,
one can take neither everyone1 s complete happiness nor one's own happi-
ness to be good without qualification: the former does not form a con-
sistent harmoniousobject; and the latter cannot plausibly be taken to
be the object of every rational will if the former is not. Thus happi-
ness cannot in either form be the "unconditioned condition" of the good-
ness of the object of your inclination, and the regress upon the condi-
tions cannot rest here. Wehave not yet discovered what if anything
makes the object of your choice good and so your choice rational.
Nowcomes the crucial step. The answer, I think, is that what makes
the object of your rational choice good is that it.:!.! the object of a
rational choice. That is, since we still do make choices and have the
attitude that what we choose is good in spite of our incapacity to find
the unconditioned condition of the object's goodness in this (empirical)
regress upon the conditions, it must be that we are supposing that ra-
tional choice itself makes its object good. I am going to give this idea
a tag - I will refer to it as the idea that rational choice has a value-
conferring status. WhenKant says: "rational nature exists as an end in
 195

itself. Mannecessarily thinks of his own existence in this way; thus

far it is a subjective principle of humanactions" (G 47/429), I read him
as claiming that in our rational choices and actions we view ourselves
as having a value-conferring status in virtue of our rational nature.
Weact as if our own choice were the sufficient condition of the goodness
of its object: this attitude is built into rational action. WhenKant
goes on to say: "Also every other rational being thinks of his existence
by means of the same rational ground which holds also for myself; 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"
(G 47/429), I read him as making the following argument. If you view
yourself as having a value-conferring status in virtue of your power of
rational choice, you must view anyone who has the power of rational
choice as having in virtue of that power a value-conferring status. This
will mean that what you make good by means of your rational choice must
be hannonious with what another can make good by means of her rational
choice - for the good is a consistent, harmonious object shared by all
rational beings. Thus it must always be possible for others "to contain
in themselves the end of the very same action" (G 48/430: Paton has it
more happily: "share in the end of the very same action"). 4 The way to
guarantee this possibility is to act according to universalizable princi-
ples.
Finally, the moral egoist is a man who limits all ends to
himself, sees no use in anything except what is useful to him
and, as a eudaemonist, locates the supreme detennining ground
of his will merely in utility and his own happiness, not in
the thought of duty. For, since every other man also fonns
his own different concept of what he considers happiness, it
is precisely egoism that results in [the eudaemonist's] having
no touchstone of the genuine concept of duty, which absolutely
must be a universally valid principle. - So all eudaemonists
are practical egoists.
The opposite of egoism can be only pluralism. that is the
 196

attitude of not being occupied with oneself as the whole

world, but regarding and conducting oneself as a citizen of
the world. (ANTH 11-12/130)
Thus, regressing upon the conditions, we find that the uncondi-
tioned condition of the goodness of anything is rational nature, or the
power of rational choice. To play this role, however, rational nature
must itself be something of unconditional value - an end in itself. This
means, however, that you must treat rational nature wherever you find it
(in your own person or in that of another) as an end. This in turn means
that no choice is rational which violates the status of rational nature
as an end: rational nature becomesa limiting condition of the ration-
ality of choice and action. It is an objective end, so you can never act
against it without contradiction. If you overturn the source of the
goodness of your end, neither your end nor the action which aims at it
can possibly be good, and your action will not be fully rational.
In order to knowwhat is meant by "treating rational nature as an
end", we need only consider this argument, and see howrational nature
got to be an end in itself. Whatwas in question was the source of the
goodness of an end - the goodness, say, of some ordinary object of incli-
nation. This source was traced to the power of rationally choosing ends,
exercised in this case on this end. So when Kant says rational nature
or humanity is an end in itself, it is the power of rational choice that
he is referring to, and in particular, the power to set an end (to make
something an end by conferring the status of goodness on it) and pursue
it by rational means. This is made very clear later in the Foundations
in Kant's summaryof the argument:
Rational nature is distinguished from others in that it
proposes an end to itself. This end would be the material of
every good will. Since, however, in the idea of an absolutely
good will without any limiting condition of the attainment of
this or that end, every end to be effected must be completely
 197

abstracted (as any particular end would make each will only
relatively good), the end here is not conceived as one to be
effected but as an independent end, and thus merely negatively.
It is that which must never be acted against, and which must
consequently never be valued merely as a means but in every
volition also as an end. Nowthis end can never be other than
the subject of all possible ends themselves,[*] because this
is at the same time the subject of a possible will which is
absolutely good; for the latter cannot be made secondary to
any other object without contradiction. The principle: Act
with reference to every rational being (whether yourself or
another) so that it is an end in itself in your maxim, is thus
basically identical with the principle: Act by a maximwhich
involves its ownuniversal validity for every rational being.
That in the use of means to every end I should restrict my
maximto the condition of its universal validity as a law for
every subject is tantamount to saying that the subject of ends,
i.e., the rational being itself, must be made the basis of all
maximsof actions and must thus be treated never as a mere
means but as the supreme limiting condition in the use of all
means, i.e., as an end at the same time. (G 56/437-438)
*[Paton has "subject of all possible ends himself" (not them-
selves) and that seems correct.JS
It is the power to take something as your end and pursue it by rational
means that is identified with "humanity", and it is that power or capa-
city that we are adjured to treat as an end in itself. The test that
parallels the first contradiction test is concerned with cases in which
we are using another merely as a means and not at the same time as an
end, and the test that parallels the second contradiction test is con-
cerned with cases in which we are failing to promote humanity as an end.
In either case, what is involved is a failure to properly acknowledgein
your conduct the value-conferring status either of another or of yourself.
Wecan make this plausible, and also see why Kant takes the two formulae
to be identical, by considering the examples and the way Kant explains
them.
In the suicide case, Kant says that i f, in order to escape from
11

burdensomecircumstances, he destroys himself, he uses a person merely

as a means to maintain a tolerable condition up to the end of life. 11

(G 47/429) In the passage quoted above, Kant pointed out that it was a
 198

consequence of his argument that "the subject of a possible will which

is absolutely good... cannot be madesecondary to any other object with-
out contradiction. 11
(G 56/437} This is what happens in the case of
suicide: the end, in the example, is a tolerable condition" and the
11

means is the destruction of a rational being - hence a rational being is

being used as a mere means to a relative end. The reason why this is
said to be a contradiction rather than merely a case of misordered values
is that the relative end must get its value from the thing that is being
destroyed for its sake. Howeverobvious it may seem that a tolerable
condition is a good thing, it is good only because of the value conferred
upon it by the choice of a rational being. Destroy the rational being,
and you cut off the source of the goodness of this end - it is no longer
really an end at all, and it is no longer rational to pursue it.
In the promising case, Kant uses two expressions which are central
to what he wants to say. One is that the other person 11cannot possibly
assent to mymodeof acting against him and the second, already men-
11

tioned, that the other person cannot share in the end of the action. In
both of these expressions, the apparent element of subjectivity can be
eliminated by a strong reading of the "cannot". It is not that others
don't happen to agree to the way you propose to act, or that they don't
happen to have the same end that you seek by this action - it is that
they cannot. Suppose, for instance, you cometo me and ask to borrow
some money, falsely promising to pay it back next week and, by some
chance, I knowthat your promise is false. Suppose also that I have the
same end that you do, in the sense that I want you to have this money,
and so I turn the moneyover to you anyway. Here I have the same end as
you, and tolerate your attempts to deceive me to the extent that they do
 199

not prevent my turning over the moneyto you. Yet even in this case it
is not true, and cannot be true, that I assent to your way of behaving
towards me and share in the end of your action. If I call you on it
openly, and say, don1 t give me that, just take this money"then what
11

I am doing is not accepting a false promise, but giving you a handout,

scorning your promise. If I don't call you on it, and keep these
thoughts in my heart, it is all the same: I am still not accepting your
false promise. In this case I am pretending to accept your false pro-
mise: "you' 11 pay me next week, I say, "sha11 we say Wednesday?
11 11
,

wondering what sort of fool you take me for, but prepared to live up to
myownstandards of courtesy and decent humanintercourse all the same.
But there is all the difference in the world between actually doing some-
thing and pretending to do it. Knownfalse promises cannot be accepted;
knownlies cannot be believed; in all cases a tolerated manipulation is
not an agreement to be manipulated, but something else. If I call you
on the false promise, I turn the case into a different sort of action:
a handout on my part, not a promise on yours. If I don1 t, I engage in a
sort of counter-manipulation, by pretending to engage in a promise with
you when I am not really doing so. (In this case, by the way, I am on a
Kantian account guilty of not treating you as an end, and of deceiving
myself into the idea that your bad behavior gave me a license for that.)
In neither case can l be described as agreeing to your way of behaving
towards me, for in both cases I fix it so that something else is hap-
pening. Even if I do this covertly there is no assent. If I say to you
later: 1 knewperfectly well that you never intended to pay me back,
11 11

you can hardly say 11well, that s all right then, for you assented to it.
1 11

So there is a criterion here for whether another can assent to your way
 200

of behaving: if it is a case in which, so long as the other person

knowswhat you are trying to do and has the power to stop you, then what
you are trying to do is not really what is happening at all, then it is
a case in which the other person cannot possibly agree to your way of
behaving. Knowledgeof what is going on and some power over what is go-
ing on are the conditions of assent: if under these conditions your
action as you envision it cannot even really take place, then it cannot
be assented to. This criterion, like the first test generally, works
more clearly for practices than for natural acts, since practices are by
their nature joint activities. But even in the case of natural acts it
is not impossible to apply. Youcannot wrest from me what I freely give
to you; and if I have the power to stop you from wresting something from
me and do not use it, I am in a sense freely giving it to you.
What this shows is that it is not possible for me to allow myself
to be manipulated (treated as a mere means) - at least not without consi-
derable self-deception. This is because I must in all my actions regard
myself as an end: I cannot rationally allow myself to be treated as a
means. If I have the relevant knowledgeand power to give me control
over what is happening, then I can only allow what I take to be good to
happen: for when I choose, I regard myself as.! cause of what is to
occur. I must myself determine whether I will confer goodness upon these
results, and in doing so I cannot rationally treat my own power of con-
ferring value as something that is of less value than the results them-
selves. Hence too the point about th~ other person being able to "con-
tain" or share the end of the very same action (G 48/430): another can-
not share your end if it is the nature of your action to place that end
above or higher than his or her rational nature or power of rational
 201

choice.
In every case in which the first contradiction test is violated,
something of this nature is going on - there is some manipulation, some
taking advantage, which involves a failure to treat the other as an end
in itself and a possible fellow first-cause of the results. Although
Kant focusses on the acceptability of your treatment of the other person
directly involved in the action, a parallel point can be made about your
treatment of others who use the same sort of procedure you propose to use
but do so only properly. Whenyou tell a lie for a certain purpose, the
lie works to achieve the purpose only because most people tell the truth.
In such a case it is not just the person to whomyou lie that you treat
as a means (.manipulate} but all of those who tell the truth. This is
because you allow their actions to fuel your method, and that is expli-
citly treating their rational nature as a mere means: indeed it is
making a tool of other peoples• good wills. Wheneveryou use a method
that works only because others do not use it - which is what the first
contradiction test always shows - you make an instrument of the rational
nature of others, and treat them as mere means. Wheneveryou act towards
an individual in a way that she could not assent to, you manipulate her,
and treat her rational nature as a mere means. I will argue later that
this concept of manipulation is especially important in this regard - for
in manipulation you "work" the reason of the other, treating the other's
power of rational choice as a set of levers which, pushed with due skill,
give you the results that you desire. 6
The third and fourth examples, of the duty of self-perfection and
the duty to promote the happiness of others, admit of very clear accounts
in terms of the idea of acknowledgingthe value-conferring power of
 202

rational beings as ends in themselves. In the case of the duty of self-

perfection, it is a question of developing and realizing the capacities
which enable you to exercise your power of rational choice - the talents
and powers that make it possible for you to set and pursue ends. It is
your powers as an agent that are to be promoted. In the case of the duty
to promote the happiness of others, Kant says·:
Humanitymight indeed exist if no one contributed to the
happiness of others, provided he did not intentionally
detract from it; but this harmonywith humanity as an end
in itself is only negative rather than positive if everyone
does not also endeavor, so far as he can, to further the
ends of others. For the ends of any person, who is an end
in himself, must as far as possible be my end[*], if that
conception of an end in itself is to have its full effect
on me. (G 48-49/430} *[this should be "as far as possible
be my end~"]7
This is because the full realization and acknowledgementof the fact
that another is an end in itself involves viewing the ends upon which
this person confers value as good - and when one acknowledges that some-
thing is good, one acknowledgesit to be "in the judgment of every
reasonable man, an object of the faculty of desire" (.C262-63/60-61).
To treat another as an end in itself is to treat his ends as good, just
as regarding yourself as an end in itself involves taking your own ends
to be good.
In general, one might say that the two formulae are equivalent
because the notions of a world for action and a value-conferring agent
are correlative. It is because of one's capacity to set and pursue ends
using reason that one chooses a world for action: it is because of this
same capacity that one regards oneself as an end in itself. A world for
action is a world well-designed as the setting in which ends in them-
selves are to engage in their characteristic activity. A world for
action is a world for ends in themselves, since action, as the creation
 203

and realization of value, is what we as specifically humanCth.eposses-

sors of humanity
11 11
} primarily choose. Choosing as if you were choosing
a world for action, and making humanity the limiting condition of your
choice, come to the same thing. In both cases the idea is that we make
our capacity to set and pursue ends our primary value. One might say
that, above all, it is our capacity to value ends as good that we value.
Humanbeings bring goodness into the world through the capacity to value
things. Whenwe view the world from the standpoint of practical reason,
in terms of good and evil and our capacity to be the cause of good or
evil in it, it is the very capacity to take up and act from this point
of view that we value most of all.
Humanityis an objective end, because it is the unconditioned
condition of all value. As such, it is the ground of a categorical
imperative, for Kant has argued that if there were such an end it would
serve as the basis for a practical law, applying to all rational beings
because an objective end is an end for all rational beings. But an end
also serves the will as the objective ground of its self-determination
11 11

CG45/427) and although "Objectively the ground of all practical legisla-

tion lies (according to the first principle} in the rule and in the form
of universality, which makes it capable of being a law (at most a natural
law}; subjectively, it lies in the end. 11 (G 49/431} Paton renders this
last sentence 11the ground for every enactment of practical law lies
objectively in the rule and in the form of universality which (according
to our first principle) makes the rule capable of being a law (and indeed
a law of nature); subjectively, however, it lies in the end.... 118 The
motive for the individual to choose as if choosing a world for action
lies in taking humanity as an end in itself: it is by making this
 204

objective end one's own subjective end that one can be movedto act as
the moral law commands. This is the basis of Kant's theory of virtue,
and it is in termsof this idea that we are to understand the notion
of an action that is motivated by reason alone.

III. Conferring Value

In the next chapter I will turn to the discussion of the role of
humanity as an end in itself in the motivation of the perfectly rational
action - the action that is done for the sake of duty - as this role is
described by Kant in the Metaphysics of Morals. Before I do that, how-
ever, I want to say something more about the idea of conferring value
in relation to the notion of an end in itself. In particular, I want
to point out two other contexts in which this notion operates: the
discussion of the incentives of pure practical reason in the Critique
of Practical Reason; and the teleological arguments about the role of
humanity in the Critique of Judgment. Seeing how the notion operates
in these contexts will, I think, lend support to the reading of the
argument leading to the Formula of the End in Itself which I have given.
The discussion of the moral incentive in the Critique of Practical
Reason lends support to the idea that as rational beings we must regard
what we choose as good and that we are tempted to take the fact that
something contributes to our own happiness as the unconditioned condi-
tion of its value. A description of the moral incentive will bring this
out. The moral law functions as an incentive, according to Kant, by
checking the inclinations that may oppose it, thus producing an effect
on feeling. At first, it looks as if the effect on feeling would be
negative - the sense of thwarted inclinations. Kant says:
 205

The essential point in all determination of the will

through the moral law is this: as a free will, and thus not
only without co-operating with sensuous impulses but even
rejecting all of them and checking all inclinations so far
as they could be antagonistic to the law, it is determined
merely by the law. Thus far, the effect of the moral law
as an incentive is only negative, and as such this incentive
can be knowna priori. For all inclination and every sen-
suous impulse is based on feeling, and the negative effect
on feeling (through the check on the inclinations) is itself
feeling. Consequently, we can see a priori that the moral
law as a ground of detennination of the will, by thwarting
all our inclinations, must produce a. feeling which can be
called pain. (C2 75/72)
But Kant wants to show howthis functioning of the moral law can pro-
duce a positive effect on feeling as well. In seeking this result, Kant
develops a rather complex view of the effect of the law on feeling, in
which we can discern four different but closely interrelated "feelings"
caused by the law, which together contribute to the effect of the moral
incentive. The discussion is introduced as follows:
All inclinations taken together (which can be brought into
a fairly tolerable system, whereupontheir satisfaction is
ca11ed happiness} constitute se 1f-rega rd (so 1i psi smus}. This
consists either of self-love, which is a predominant benevo-
lence towards one's self (philautia) or of self-satisfaction
(arrogantia}. The former is called, more particularly,
selfishness; the latter, self-conceit. Pure practical reason
merely checks selfishness, for selfishness, as natural and
active in us even prior to the moral law, is restricted by
the moral law to agreement with the law; when this is done,
selfishness is called rational self-love. But it strikes
self-conceit down, since all claims of self-esteem which
precede conformity to the moral law are null and void. For
the certainty of a disposition which agrees with this law is
the first condition of any worth of the person (as will soon
be made clear), and any presumption [to worth] prior to this
is false and opposed to the law. Nowthe propensity to self-
esteem, so long as it rests only on the sensibility, is one
of the inclinations which the moral law checks. Therefore,
the moral law strikes downself-conceit. (C2 75-75/73)
A dual effect on each of the two sides of self-regard makes up the four-
fold response which constitutes the moral incentive. The checking of
the "selfish" inclinations is negatively pain, but positively a sense
 206

of freedom or release. The checking of the self-esteem that accom-

panies them is negatively, humiliation, but positively, respect. It is
this that is, of course, identified as the moral incentive, though the
other effects are present as well, and seem to work together in pro-
ducing the effect of the incentive:
Since the idea of the moral law deprives self-love of its
influence and self-conceit of its delusion, it lessens the
obstacle to pure practical reason and produces the idea of
the superiority of its objective law to the impulses of
sensibility; it increases the weight of the moral law by
removing, in the judgment of reason, the counterweight to
the moral law which bears on a will affected by sensibility.
(C2 78/75-76)
In the case of self-love, the effect is painful because of the thwart-
ing of those inclinations that oppose the moral law (and the potential
thwarting of any inclination). The positive side of this feeling is
not muchemphasized in the incentive section but plans an important
role in the discussion of moral education in the 11Methodology11 :
The heart is freed from a burden which has secretly pressed
upon it; it is lightened when in instances of pure moral
resolutions there is revealed to man, who previously had not
correctly knownit, a faculty of inner freedom to release
himself from the impetuous importunity of the inclinations,
to such an extent that not even the dearest of them has an
influence on a resolution for which he nowmakes use of his
reason •.. there is a consciousness of an independence from
inclinations and circumstances and of the possibility of
being sufficient to myself, which is salutary for me in yet
other respects. (C2 165/160-161)
In the case of self-conceit, the negative side is humiliation; but the
positive side is respect for the very law that humiliates us:
If anything checks our self-conceit in our ownjudgment, it
humiliates. Therefore, the moral law inevitably humbles
every manwhen he compares the sensuous propensity of his
nature with the law. Nowif the idea of something as the
determining ground of the will humiliates us in our self-
consciousness, it awakens respect for itself so far as it
is positive and the ground of detennination. The moral law
therefore, is even subjectively a cause of respect. (C2
77/74)
 207

Nowalth.ough Kant speaks of self-esteem as "one of the inclinations

which the moral law checks" lC2 76/731, it is clear that th.e influence
of self-esteem or self-conceit is more pervasive than that of other
inclinations and th~t the way in which it is checked is different from
that of the other inclinations. Self-conceit is not merely thwarted -
it is humiliated; and the result is respect, identified by Kant as the
moral incentive itself. I suggest, therefore, that self-conceit is
something that accompaniesor pervades the influence of the inclina-
tions generally, something that is present and active wheneverwe are
tempted to act on any inclination whatever. Self-conceit .i!, the ten-
dency to suppose that the objects of our inclinations are good just
because they contribute to our happiness - the tendency, that is, to
suppose ourselves fit to confer value even whenwe do not makethe
moral law a limiting condition of our choice •
••• we find our pathologically determined self, although
by its maximsit is wholly incapable of giving universal
laws, striving to give its pretensions priority and to make
them acceptable as first and original claims, just as if it
were our entire self. This propensity to makethe subjec-
tive determining grounds of one's choice into an objective
determing ground of the will in general can be called self-
love; when it makes itself legislative and an unconditional
practical pdnciple, it can be called self-conceit. The
moral law, which alone is truly, i.e., in every respect,
objective, completely excludes the influence of self-love
from the highest practical principle and forever checks
self-conceit, which decrees the subjective conditions of
self-love as laws. (C2 77/74)
In this passage it is quite clear that self-conceit is not merely one
amongmanyinclinations, but that side of self-love that tends to re-
.
gard the dictates of self-love as legislative.
Thus we conceive how it is possible to understand a priori
that the moral law can exercise an effect on feeling, since
it blocks the inclinations and the propensity to make them
the supreme practical condition (i.e. self-love) from all
 208

participation in supreme legislation. ( C2 77/74).

This 11propensity 11 is the propensi_tyto evil described in Religion Within
the Limits of ReasonAlone as follows:
The law, rather, forces itself upon him irresistibly by
virtue of his moral predisposition; and were no other incen-
tive working in opposition, he would adopt the law into his
supreme maximas the sufficient determining ground of his
Willkur; that is, he would be morally good. But by virtue
of an equally innocent natural predisposition he depends
upon the incentives of his sensuous nature and adopts them
also (.in accordance with the subjective principle of self-
love) into his maxim.•..
... Hence the distinction between a good man and one who is
evil ••. must depend upon subordination (the form of the maxim),
i.e. which of the two incentives he makes the condition of
the other. Consequently man (eventhe bes~is evil onlyin
that he reverses the moral order of the incentives when he
adopts them into his maxim. He adopts, indeed, the moral law
along with the law of self-love; yet when he becomesaware
that they cannot remain on a par with each other but that one
must be subordinated to the other as its supreme condition,
he makes the incentive of self-love and its inclinations the
condition of obedience to the moral law; whereas, on the con-
trary, the latter, as the supreme condition of the satisfac-
tion of the former, ought to have been adopted into the
universal maximof the Willkur as the sole incentive.
(.R 31-32).

The conclusion that I would like to draw from this discussion is

this. As rational agents we require that the conditions of the goodness
of our choices be met - we require that what we choose be good. Further-
more, we are inclined, due to the evil propensity of our nature, to be-
lieve ourselves capable of conferring value on the objects of our choice
without regard for the moral law. Weare inclined to believe that what
makes us happy is thereby good, and behind this attitude is a secret
presumption of our ownworth: self-conceit. The moral law itself,
demandinguniversalization and consistency, alerts us to the falseness
of this attitude: this subjective principle of humanactions must be
universalized, objectified. If my choice can confer value upon its
object, so can another's, and all rational beings must be regarded as
 209

ends in themselves; and th.is provides a limiting condition on choice.

This limiting condition i's itself the moral law, in the Formula of the
End in I tse 1f.
The second set of passages that I want to discuss in this connec-
tion comesfrom the Critique of Judgment. These passages seem to me
to support in a very forceful way the idea that it is our power to con-
fer objective value that Kant thinks of as having unconditional worth.
In the "Methodologyof Teleological Judgment", Kant is concerned
with the question of what might appropriately play the role of final
purpose of creation. He has established the idea of a natural purpose,
which provides the basis for a teleological interpretation of nature;
but in order to view nature as a teleological system, we must discover
its final purpose: we must discover, that is, a reason for the exis-
tence of nature itself. This will not be a purpose internal to nature,
but one outside or independent of it: "an objective supreme purpose,
such. as the highest reason would require for creation." (C3 286n} In
carrying out his inquiry, Kant undertakes a familiar sort of regress
argument. In this case, the condition of a given thing is its purpose,
and the regress must end with something which is in itself a final pur-
pose, "that purpose which needs no other as the condition of its possi-
bility"; something about which "it can no longer be asked why" it exists.
(C3 284-285) Starting with the idea of a natural purpose, the argument
proceeds to what Kant calls an ultimate purpose: this will be that
which we judge to be the purpose within nature towards which all nature
is organized. This ultimate purpose, being as it were nature's contri-
bution to the final purpose, will give us an idea of what this final
purpose is.
 210

Beginning with consideration of vegetable nature~ we can reason

back to the final purpose as follows:
•.. a more intimate knowledgeof its indescribably wise
organization does not permit us to hold to this thought
(that it is a mere mechanism), but prompts the question:
Whatare these thi'ngs created for? If it is answered:
For the animal kingdom,which is thereby nourished and has
thus been able to spread over the earth in genera so various,
then the further question comes: What are these plant-
devouri'ng animals for? The answer would Be something like
this: For beasts of prey, which can only be nourished by
that which has life. Finally we have the question: What
are these last, as well as the first-mentioned natural
kingdoms, good for? For man in reference to the manifold
use which his understanding teaches him to makeof all these
creatures. He is the ultimate purpose of creation here on
earth, because he is the only being who can form a concept
of purposes and who can, by his reason, make out of an
aggregate of purposively formed things a system of purposes.
{_C3 276)
Kant then goes on to inquire for the more specific feature of human
life that is the purpose of nature.
If nowthat must be found in manhimself which is to be
furthered as a purpose by meansof his connection with
nature, this purpose must either be of a kind that can be
satisfied by nature in its beneficence, or it is the
aptitude and skill for all kinds of purposes for which
nature (external and internal) can be used by him. The
first purpose of nature would b_eman's happiness, the
second his culture. (C3 279)
Here the argument joins up with the arguments of the teleological
historical writings wh.ichI described in Chapter 2, as well as the
teleological argument in the beginning of the Foundations. Happiness
does not seem to be something that nature can achieve or aims at
achieving: the evidence favors culture. But also, just as in the
argument for the formula of the end in itself:
Happiness, on the contrary, as has been shownin the pre-
ceding paragraphs by the testimony of experience, is not
even a purpose of nature in respect of man in preference
to other creatures, muchless a final purpose of creation.
Menmay of course make it their ultimate subjective purpose.
 211

But if I ask, in reference to the final purpose of creation,

Whymust men exist? then we are speaking of an objective
supreme purpose, such as the highest reason would require
for creatton. lf we answer: These beings exist to afford
objects for the benevolence of that supreme cause, then we
contradict the condition to which the reason of man subjects
even his own inmost wish for happiness {viz. the harmony
with his own internal moral legislation).. (.C3286n)
The answer then will be that the ultimate purpose of nature is culture,
in a specific sense:
There remains therefore of all his purposes in nature only
the formal subjective condition, viz. the aptitude of set-
ting purposes in general before himself and (independent of
nature in his purposive determination) of using nature, con-
formably to the maximsof his free purposes in general, as
a means. This nature can do in regard to the final purpose
that lies outside it, and it therefore may be regarded as
its ultimate purpose. The production of the aptitude of a
rational being for arbitrary purposes in general {conse-
quently in his freedom) is culture. (C3 281)
The purpose of nature is the culture of humankind, and culture is that
which furthers our aptitude for setting purposes in general before our-
selves and using nature as a means to realize them.
The final purpose of nature, Kant argues, is morality itself. It
is in morality that the aptitude of setting purposes before ourselves
finds its completion, for an end or purpose must be unconditionally
good, and only in morality do we find the unconditioned condition of
its goodness:
Nowwe have in the world only one kind of beings whose
causality is teleological, i.e. is directed to purposes,
and is at the same time so constituted that the law accord-
ing to which they have to determine purposes for themselves
is represented as unconditioned and independent of any
natural conditions, and yet in itself necessary. The being
of this kind is man, but man considered as a noumenon,the
only natural being in which we can recognize, on the side of
its peculiar constitution, a supersensible faculty (freedom)
and also the law of causality, together with the object,
which this faculty may propose to itself as highest purpose
(the highest good in the world).
Only in man, and only in him as subject of morality, do
we meet with unconditioned legislation in respect of purposes,
 212

which therefore alone renders him capable of being a final

purpose, to which the whole of nature is teleologically
subordinated. (C3 285-286)
It is our capacity to set ends - to freely choose what shall be an end
by means of reason, that not only makes us ends in ourselves, but which
forms the final purpose of nature teleologically conceived. It is only
this capacity that has its value completely in itself; so that this not
only forms the basis of a possible categorical imperative, but also the
only possible basis for a complete teleological view of creation.
Without men the whole creation would be a mere waste, in
vain, and without final purpose. But it is not in reference
to man's cognitive faculty {theoretical reason) that the
being of everything else in the world gets its worth; he is
not there merely that there may be someoneto contemplate
the world. For if the contemplation of the world only affor-
ded a representation of things without any final purpose, no
worth could accrue to its being from the mere fact that it is
known;we must presuppose for it a final purpose, in reference
to which its contemplation itself has worth. Again it is not
in reference to the feeling of pleasure or to the sum of
pleasures that we think a final purpose of creation as given;
i.e. we do not estimate that absolute worth by well-being or
by enjoyment (whether bodily or mental), or, in a word, by
happiness. For the fact that man, if he exists, takes this
for his final design gives us no concept as to why in general
he should exist and as to what worth he has in himself to
make his existence pleasant. He must, therefore, be supposed
to be the final purpose of creation, in order to have a
rational ground for holding that nature must harmonize with
his happiness if it is considered as an absolute whole
according to principles of purposes. Hence [that which we
seek] is only the faculty of desire, not however, that which
makes man dependent (through sensuous impulses) upon nature,
nor that in respect of which the worth of his being depends
upon what he receives and enjoys. But it is that worth which
he alone can give to himself and which consists in what he
does, howand according to what principles he acts, and that
not as a link in nature's chain but in the freedom of his
faculty of desire. That is, a good will is that whereby alone
his being can have an absolute wort~ and in reference to
which the being of the world can have a final purpose.
{C3293)
Or as Kant puts it in an earlier footnote:
There remains, then, nothing but the value which we ourselves
give our life, through what we cannot only do but do
 213

purposively in such independenceof nature that the existence

of nature itself can only be a purpose under this condition.
(.C3 284n)

It is we, with our power of valuing things, that bring to the world
such value as it has - and even the redemption of nature is up to us.
 214

NOTES

1. It is because we must will whatever is good - all of wh.at is good -

that Kant later says we must make the Highest Goodour end. The
Highest Goodappears to be impossible, and it is because the moral
law therefore seems to ask of us the adoption of an impossible end
that a dialectic of practical reason is generated. It is this
11 11

that creates the need for a practical religious faith. People are
sometimes puzzled over the fact that acting as if the Highest Good
were one's end is no different from acting according to the moral
law generally; they cannot see what the demandto take the Highest
Goodas one's end amounts to. The answer is that the adoption of
this end is itself a sort of action, and one that is practically
equivalent to taking up a virtuous attitude of will - resolving
to act not merely according to duty but from duty.
2. Despite an awkwardnessof translation here, it is obvious that this
means our rational desires are only influenced by the conception of
things as good or bad, and not that we can have a rational desire
for what is bad.
3. Somediscussion of egoism and its oddities might be useful in sup-
porting this point. For examples, Nagel on the egoist as a sort
of solipsist in The Possibility of Altruism, Chapter XI; G.E. Moore
in Principia Ethica, Sections 59-61; and Brian Medlin, "Ultimate
Principles and Ethical Egoism.11

4. Groundworkof the Metaphysic of Morals, translated by H.J. Paton,

p. 97. In Gennanthe passage is:
... dass sie als verntlnftige Wesenjederzeit zugleich als
Zwecke, d.i. nur als solche, die von ebenderselben Handlung
auch in sich den ZweckmUssenenthalten K6nnen, gesch8tzt
werden sollen. (Vorlander, p. 53)
5. Groundwork,Paton, p. 105. The Germanis:
Dieser kann nur nichts anderes als das Subjekt aller
m6glichen Zweckeselbst sein, weil dieses zugleich das Subjekt
eines rnt,glichen schlechterdings guten Willens ist; ...
(Vorlander, p. 61)
6. See Chapter 6, pp. 269-275.
7. Paton has it:
For the ends of a subject who is an end in himself must, if
this conception is to have its full effect in me, be also,
as far as possible, mx_ends. (Groundwork,p. 98}.
In German:
Denn das Subjekt, welches Zweckan sich selbst ist, dessen
ZweckemOssen,wenn jene Vorstellung bei mir alle Wirkung
 215

tun soll, auch soviel m6glich meine Zweckesein. (Vorlander,

p. 53)
8. Groundwork,Paton, p. 98. In German:
Es liegt n8mlich der Grund aller praktischen Gesetzgebung
objektiv in der Regel und der Form der Allgemeinheit, die
sie ein Gesetz {allenfalls Naturgesetz} zu sein f8hig macht
(nach dem ersten Prinzip}, subjektiv aber im Zwecke,....
(Vorlander, p. 54}
 Chapter 6

Acting FromDuty

In order to show that pure reason can be practical, or, to say the
same thing, that there is a practical employmentof pure reason, it is
not enough to show that there is a law of reason that applies to actions.
It must also be shownthat it is possible for us to be motivated to
act on that law. For Kant, the purely rational action is not merely
one that is in accord with the law, but one that is done from the law.
The truly worthy action is done from respect for the law, and motivated
entirely by reason. Whatwe need is at least a picture of howthis is
possible, of what it is like. Andthis is especially important because
of the power of the Humeanpicture on which the usual argument against
rational motivation is based: every action receives its motivational
impetus from desire for an end, reason cannot dictate an end, therefore,
reason cannot motivate an action.
In considering the Fonnula of Universal Law, it is clear that Kant
moves the discussion onto different grounds by introducing the idea of
a maxim,and of the rational standards that apply to maxims. Maxims
are the basic objects of the will, and ends are adopted in the context
of maxims: hence it seems no longer to be a question of whether there
is a specially rational end or not: the influence of reason on action
is accounted for by the existence of rational standards applying to
maxims. Wetest our maximsto see whether or not these standards are
met, and accept or reject them accordingly - this is the influence of

216
 217

reason on action. The existence of a categorical imperative as well as

hypothetical ones shows that this influence goes beyond the detennina-
tion of fitness of means to ends which Humesaw as the only possible
practical role of reason.
But a careful consideration of the process of deliberation as
here envisioned shows that this is not enough to account for the pos-
sibility of an action motivated entirely by reason. The person who acts
solely from duty is supposed to be motivated by the law alone:
Nowas an act from duty wholly excludes the influence of
inclinations and therewith every object of the will, nothing
remains which can determine the will objectively except the
law, and nothing subjectively except pure respect for this
practical law. This subjective element is the maximthat I
ought to follow such a law even if it thwarts all my in-
clinations. (G 17/400)
The difficulty arises precisely because of the formality
11 11
of the
categorical imperative in the universality formulations. As stated,
the categorical imperative in the Universal Lawformulation is a test
on maxims; in fact, it is a negative test, which accepts maximsas
permissible or rejects them as forbidden. But the categorical imperative
does not by itself seem to direct us to any particular actions, or to
the adoption of any particular maxims. Someother maximwhich is to be
accepted or rejected must always be present for the test to work.
This in any case seems to be the scenario in the Foundations,
where Kant envisages someonedetennining what his duty is as fonnulating
an unacceptable maximunder the influence of self-love, and then rejec-
ting the maximbecause it does not pass one of the two contradiction
tests. This procedure leads most obviously to duties of omission, or to
restraints - to the restriction of choice rather than the detennination
of it. Here it looks as if in the case where someoneactually takes
action from duty, as opposed to merely refraining, the person must first
 218

formulate some sort of negative maximof not acting - of neglecting the

required action - and then discover that this policy of neglect is not
permissible. In the first set of examples in the Foundations, Kant
does make it sound as if all duties are arrived at in this negative and
indirect way: the duties of helping others and developing one's talents
are arrived at by people whoactually stop to deliberately consider the
policies of never helping anyone and not developing talents, and then
discover by the contradiction tests that these policies are inmoral.
Even in the Critique of Practical Reason, when he introduces the idea of
the Fact of Reason, Kant sometimesmakes it sound as if it is whenwe
are formulating other maximsthat duty impresses itself upon us:
It is therefore the moral law, of which we becomeimmediately
conscious as soon as we construct maximsfor the will, which
first presents itself to us; and, since reason exhibits it
as a ground of determination which is completely independent
of and not to be outweighed by any sensuous condition, it is
the moral law which leads directly to the concept of freedom.
(C2 29/29-30)
Intuitively, this is an unsatisfactory picture. One wants to protest
that the world suffers morally not primarily or not only from people
whoget the wrong answers when they ask themselves such questions as
"shall I help others? shall I keep myword?" but from people who never
think to ask themselves such questions at all. The negative and indirect
approach does not tell us muchabout what role reason might play in
bringing the thoughtless or inconsiderate person to a sense of positive
duty. The duties of virtue, the obligatory ends, are positive duties;
so are some duties of justice, such as keeping your promise once you
have made it. So, under certain circumstances, are such things as
advising someone, calling the police, or giving encouragementto a
discouraged friend. Neglecting such duties out of squeamishness or an
 219

attention to one's owncomfort and convenience is one sort of evil;

but simply not thinking to do them is another, just as real. Certainly,
it is part of our everyday view of what a good person is like that that
person not only does her duty when she thinks of it, but is alert to
what we might call the occasions of duty. If we wish to think of the
good person as an eminently reasonable person, we need a view of the
role reason might play in alerting us to these occasions. Nowfor this,
we need a view of howthe categorical imperative might actually prompt
us to adopt certain maxims- not just to reject bad ones but to formulate
and adopt good ones.
It is important that the question raised here not be confused with
another motivational question, closely related to this, but, in Kant's
philosophy, explicitly declared to be unanswerable. That is the ques-
tion of howwe can be motivated to adopt the categorical imperative it-
self as an incentive of the will - of howwe can adopt the maxim, men-
tioned in the quotation from the Foundations above, "that I ought to
follow such a law even if it thwarts all my inclinations". (G 17/400)
That question cannot be answered because it lies at "the supreme limit
of all moral inquiry" {G82/462).
Yet howpure reason, without any other incentives, wherever
they may be derived, can by itself be practical, i.e., how
the mere principle of the universal validity of all its
maximsas laws {whichwould certainly be the form of a pure
practical reason), without any material {object) of the will
in which we might in advance take some interest, can itself
furnish an incentive and produce an interest which would
be called purely moral; or, in other words, how pure reason
can be practical - to explain this all humanreason is wholly
incompetent, and all the pains and work of seeking an explana-
tion of it are wasted. (G 81/461)
That pure reason can be practical is made knownto us by the Fact of
Reason, and it is this that, in the argument of the Critique of Practical
 220

Reason, reveals to us our freedom and our intellectual nature. The

question that I am raising here is not this unanswerable question of
howwe can be motivated by the categorical imperative. The question
that I am raising is that of how, once we have taken a good resolution
and made the moral incentive the supreme condition of all others in our
maxims, that incentive functions in guiding action. Suppose you do
make the moral law your law, howdoes it teach you what your duty is
and motivate you to do that duty? Or as Kant puts it at the beginning
of the incentive discussion in the Critique of Practical Reason:
For howa law in itself can be the direct determining ground
of the will (which is the essence of morality) is an insoluble
problem for the humanreason. It is identical with the pro-
blem of howa free will is possible. Therefore, we shall not
have to showa priori why the moral law supplies an incentive
but rather what it effects (or better, must effect) in the
mind, so far as it is an incentive. (C2 75/72)
It is this question that Kant associates with the question of the nature
of humanvirtue: the question of howrespect for the law, as an incen-
tive, overcomesthe obstacles provided by the inclinations and motivates
us to do what is right. The major discussions of the issue are in the
incentive discussion of the Critique of Practical Reason, along with
the Methodologyof that work, and in the Introductions to the Metaphysics
of Morals and the Doctrine of Virtue.
In the discussion in the Doctrine of Virtue Kant explicitly addres-
ses the problem that I have been describing, and undertakes to solve it
in terms of the notion of objective ends:
•.• the formal principle of duty is contained in the
categorical imperative: S0 act that the maximof your action
11

could becomea universal law. Ethics adds only that this

11

principle is to be conceived as the law of your ownwill and

not of will in general, which could also be the will~
another. In the latter case the law would prescribe a juri-
dical duty, which lies outside the sphere of ethics. - Maxims
are here regarded as subjective principles which merely
 221

qualify for giving universal law, and the requirement that

they so qualify is only a negative principle: not to come
into conflict with a law as such. - Howthen can there be,
beyond this principle, a law for the maximsof actions?
Only the concept of an obligatory end, a concept that be-
longs exclusively to ethics, establishes a law for the maxims
of actions by subordinating the subjective end (which every-
one has) to the objective end (which everyone ought to adopt
as his own)...
•.. it is the obligatory end that can make it a law to have
such a maxim, since for the maximitself the mere possibility
of harmonizin~·with a giving of universal law is already
sufficient. {DV48-49/389)
Kant's theory of virtue is given in terms of the doctrine of the obliga-
tory ends, and since it is the virtuous person who acts not merely ac-
cording to duty but from duty, we must look to the doctrine of obliga-
tory ends in order to explain that action more fully.

I. The Division of Duties

In this chapter I am going to try to spell out the view of rational
action presented in the Metaphysics of Morals and the role of ends in
rational action. In order to do this, it will first be necessary to
set up the division of duties in tenns of which Kant explains the nature
of virtuous action. The two kinds of duties are duties of justice and
duties of virtue. They can be distinguished in a surprising variety of
ways, which I will examine in this section.
Kant begins by distinguishing two ways in which duties can be
legislated: ethical and juridical legislation, distinguished in terms
of the motive for obeying the law.
Every legislation contains two elements (whether it pre-
scribes external or inner actions, and whether it prescribes
these a priori by mere reason or by the choice of another
personT: first a law which sets forth as objectively neces-
sary the action that ought to take place, i.e. which makes
the action a duty; secondly a motive, which joins a ground
detennining choice to this action subjectively with the
thought of the law. Hence the second element consists in
 222

this: that the law makes the duty into the motive. The
first presents an action as a duty, and this is a merely
theoretical recognition of a possible determination of the
power of choice [Willkur], i.e. of a practical rule. The
second connects the obligation to act in this way with a
ground for determining the subject's power of choice
[Willkur] as such.
•.. The legislation that makes an action a duty and also
makes the duty the motive is ethical. But the legislation
that does not include the motive 1n the law and so per-
mits a motive other than the Idea of duty itself is juridi-
cal ...
-The mere conformity or non-conformity of an action with
the law, without reference to the motive of the action, is
called its legalita (lawfulness). But that conformity in
which the Idea of uty contained in the law is also the
motive of the action is called its morality. (MM16-17/218-
219)
This distinction is used this way: duties of virtue are those that ad-
mit only of an ethical legislation, whereas duties of justice admit of
either ethical or juridical legislation. So Kant says that all duties
belong to ethics, at least indirectly; and when a juridical duty is done
from duty 11it is still a virtuous action (a proof of virtue)." (MM18/
220)
Kant says that we can be compelled to do a duty of justice (e.g.
MM18/220), but we cannot be compelled to do it from duty. Ethical legis-
lation is internal and autonomous. This connects the idea of ethical
legislation to the notion of an end, for:
That ethics is a doctrine of virtue (doctrine officiorum
virtutis - [doctrine of the offices of virtue]) follows from
the above definition of virtue whenwe connect it with the
kind of obligation proper to ethics. - Determination to an
end is the only determination of choice which in its very
concept excludes the possibility of compulsion through natural
means by another's act of choice. Another can indeed compel
me to do something tfiat,s not my end (but only a means to
his end), but he cannot compel me to make it .!!!lend. To have
an end which I have not myself madean endwouTdbe a self-
contradiction - an act of freedom which is still not free.
- But it is no contradiction that I myself set an end which
is also a duty, since I constrain myself to it and this is
altogether consistent with freedom. (DV39/381-382)
 223

It is worth noting that the sense in which we can be compelled to

do a duty of justice is a double one. Kant partly means that we 11can11
be compelled in a theoretical sense - it is always physically possible
to make someonedo a duty of justice, since they are all outer actions.
But he also sometimesmeans "can" in a moral sense. Others are morally
authorized to use coercion to enforce their rights, with which the
duties of justice are associated. (See the Introduction to the Metaphy-
sical Elements of Justice). Kant points out (MM18/220) that some
duties of virtue include outer acts - for example, one could be forced
to do something beneficent. But (i) one cannot be forced to do the thing
for a benevolent reason, because one has made the happiness of another
one's end,for this is physically impossible; and (ii) one cannot
rightly be forced to do the beneficent thing. It is consistent with
one's freedom that one be forced to do an action demandedby justice,
but it is not consistent with one's freedom to be forced to do an action
that is in accord with an obligatory end.
It is with this in mind that Kant makes another sort of distinction
between the two sorts of duties, and associates duties of justice with
our external freedom, and duties of virtue with our internal freedom.
(DV38/380) It is easy to see that the duties of justice would be
by and large associated with the first contradiction test and the duties
of virtue with the second. {Adifficulty with this point is Kant's use
of suicide as an example of the workings of the first contradiction
test, for suicide is a violation of a duty of virtue - a perfect duty
of virtue. In the Foundations, Kant associates the first test with per-
fect duty and the second with imperfect, but there are many perfect du-
ties of virtue discussed in the Doctrine of Virtue. This difficulty
 224

can be accomodated, however; I will comeback to it.) It is also easy

to see that the duties of justice are by and large either negative,
like the duties not to commitvarious crimes, or duties one acquires by
entering into various practices, like the duties associated with con-
tracts, promises, family life, and so forth.
Three more modesof distinction are important. Kant gives them in
section titles: "VI. Ethics does not give laws for actions {Ius does
that), but only for the maximsof actions." (DV48/388) "VII. Ethical
duties are of WideObligation, whereas Juridical Duties are of Narrow
Obligation." (DV49/390) "X. The First Principle of the Doctrine of
Lawwas Analytic: That of the Doctrine of Virtue is Synthetic." (DV56/
396) Each of these must be discussed.
Ethics gives laws for maximsbecause it is concerned with ends, and,
as Kant points out, "it is the obligatory end that can make it a law
to have such an end." (DV48/389) So long as we stay within the limits
of testing maxims, the categorical imperative only accepts or rejects
preexisting maxims, and so accepts or rejects the actions those maxims
describe. It is only when there is an obligatory end that we are obliged
to have a certain maxim- namely, a maximof doing something for the
sake of that end. For instance, the adoption of the end of promoting
the happiness of others makes it necessary to have certain maxims, to
do this or that for the sake of another's happiness. These maximsin
turn must be tested for legality - which is why justice always has a
certain priority over the duties of virtue in the Kantian system - no
end can justify an unjust action. Obviously, this is not a complete
picture of what ethical legislation is like, for we need to knowtwo
more things: howthe categorical imperative dictates ends; and whether
 225

and howthis idea of adopting an obligatory end and adopting a maximin

light of it applies to the case of doing a duty of justice from the
motive of duty.
Kant claims that the fact that ethical duties are of wide obliga-
tion follows from the fact that ethics gives laws for maximsrather
than directly for actions:
•.. for if the law can prescribe only the maximof actions,
not actions themselves, this indicates that it leaves a play-
room (latitudo) for free choice in following (observing) the
law, i.e. that the law cannot specify precisely what and how
muchone's actions should do toward the obligatory end. (DV
49/390)
In the case of the two obligatory ends of the Foundations (and here),
it is easy to see what Kant has in mind. Although everyone must make
the happiness of others and the cultivation of their natural perfections
an end, the moral law does not specify exactly what or howmuchwe will
do towards these ends. Whatdepends on our circumstances, and howmuch
on both our circumstances and our own true needs. (DV51-53/391-393)
Whichof your talents you will cultivate depends upon your profession,
your opportunities, andevenyour preferences; though neither your
body nor your mind can be permitted to go entirely to seed, for you do
have to be prepared for all sorts of possible ends. Andyou are not to
devote so much time and energy to the pursuit of the happiness of others
that you entirely neglect your own, for that would contradict itself
11 11
•

(DV53-54/393)
It is often thought that Kant used the pairs of tenns "narrow and
wide" "perfect and imperfect" interchangably. Gregor, for example,
finds it a problem that perfect duties appear in the Doctrine of Virtue,
which is explicitly about wide duties. 1 Certainly there is evidence for
this view: to my knowledgeKant makes no statement about the use of the
 226

two sets of tenns, and there are places where he opposes "narrow and
perfect" to "wide and imperfect". (e.g. DV113/446) In the Foundations
discussions of the examples, the two distinctions are lined up. I do
not want to enter at length into this confused issue, except to point
out that whatever one makes of the tenninology, perfect duties do have
a natural place in a doctrine of virtue. Even in the Foundations, Kant
remarks that since a perfect duty is one that admits of no exceptions
in the interests of inclination, he recognizes both inner and outer per-
fect duties. (G 39n/421n) The place for perfect duties in a doctrine of
virtue would be this: once you have adopted an end, you have perfect
duties of omission with respect to actions that ar~ contrary to that
end. If you make humanity in your ownperson your end, there may be
latitude about what you will do to further and promote it, but you have
a strict duty not to do the thing that would absolutely eliminate it:
to commitsuicide, say, or to use your own rational nature as a manipu-
lated means by engaging in lying or servility. These duties of omission
are certainly strict, for they admit of no exception at all; yet they
are associated with adopted, obligatory ends. Although the duty in
general is wide and does not give rise to specific maximsof actions, it
does give rise to specific prohibitions, and these duties are perfect.
Finally, Kant distinguishes the two "first principles" upon which
the doctrine of law and the doctrine of virtue are based, and claims that
the first principle of law is analytic and that of virtue is synthetic.
He explains as follows:
Weneed only the principle of contradiction to see that, if
external compulsion checks the hindering of hannonious outer
freedom in accordance with universal laws (and is thus an
obstacle to the obstacles to freedom), it could harmonize with
ends· as such. I need not go beyond the concept of freedom to
see that anyone may take whatever he pleases as his end.
 227

The first principle of Lawis therefore an analytic proposi-

tion.
But the principle of the doctrine of virtue goes beyond
the concept of outer freedom and joins with it, in accordance
with universal laws, an end which it makes a duty. The
principle is therefore synthetic. - Its possibility is con-
tained in the deduction (IX). (DV56/396)
The passage tells us that it is with respect to the concept of freedom -
of outer freedom - that the first principle of law is analytic. That
principle is:
Every action is just [right] that in itself or in its maxim
is such that the freedom of the will of each can coexist to-
gether with the freedom of everyone in accordance with a
universal law. (DJ 35/230)
Outer freedom is defined as 11independence from the constraint of an-
other's will"; so here the point is that if your action constrains the
will of another, it is consistent with freedom that both the laws of
freedom and other people proscribe your action. The law, and those who
would stop you on the basis of the law, are only "obstacles to the ob-
stacles" to external freedom. The law of outer freedom, however, can-
not demandeither an end or any particular maximof you:
It also follows that I cannot be required to adopt as one
of my maximsthis principle of all maxims, that is, to make
this principle a maximof my action. For anyone can still be
free, even though I am quite indifferent to his freedom or
even though I might in my heart wish to infringe on his free-
dom. That I adopt as a maximthe maximof acting justly is
a requirement that ethics [rather than jurisprudence] imposes
on me. (DJ 35/231)
This is because, as Kant tells us in the Doctrine of Virtue, it is only
an end that can make it a law to have a certain maxim. (DV48/389)
Accordingly, the passage about the analytic and synthetic character of
the two first principles continues:
It is by laying downobligatory ends that pure practical
reason widens the concept of duty beyond the concept of outer
freedom and its limitation by the merely fonnal condition of
its thoroughgoing consistency. In place of external com-
 228

pulsion it brings forth inner freedom - the power of self-con-

straint, not through the mediumof other inclinations but by
pure practical reason (which scorns all these intermediaries)
[that is, the inclinations through which others are able to
compel us to obey the law - my insertion]. (DV56-57/396)
The first principle of the doctrine of virtue, as a synthetic prin-
ciple, requires a deduction. This deduction is extremely important,
for here at last we learn howthe categorical imperative dictates
obligatory ends, ana thereby, which ends it dictates. In the following
passage Kant states the first principle, explains it, and then, with
rather alarming brevity, deduces it:
The first principle of the doctrine of virtue is: act ac-
cording to a maximof ends which it can be a universal law
for everyone to have. - According to this principle man is
an end, to himself as well as to others. And it is not enough
that he has no title to use either himself or others merely
as means (since according to this he can still be indifferent
to them): it is in itself his duty to makeman as such his
end.
This first principle of the doctrine of virtue, as a cate-
gorical imperative, admits of no proof, but it does admit a
deduction from pure practical reason. - What, in the relation
of man to himself and others, can be an end, that is an end
for pure practical reason. For pure practical reason is a
power of ends as such, and for it to be indifferent to ends
or to take no interest in them would be a contradiction, be-
cause then it would not detennine the maximsfor actions either
(since every maximcontains an end) and so would not be prac-
tical reason. But pure reason can prescribe an end a priori
only in so far as it declares it to be also a duty. -And this
duty is then called a duty of virtue. (DV55-56/395)
This passage is certainly obscure, but I will try to give a reading of
it. First, I will take "determine the maximsfor actions", in line with
the important quotation on p. 221 {from DV48-49/389), to mean 11make it
a law to have a certain maximof actions. 11
This phrase can be read
another way: it could mean, 11set a rational standard for the maxims
of actions." I will also consider howthe deduction would go on this
reading.
Since the deduction is a deduction from pure practical reason (not
 229

from pure reason in general, but taking the existence of pure practical
reason as a Fact), we can suppose it to start from the assumption that
pure practical reason is possible (because actual) and to consider the
conditions under which it is possible. Then the argument goes: take
it that pure practical reason is possible; then reason by itself must
be able to detennine action. To do this, it must be able to make it a
law to have a certain maximof action. Only an obligatory end can make
it a law to have a certain maxim, so if pure practical reason is possible
there must be an obligatory end. But pure practical reason can pre-
scribe an end only if the end is a duty, and this means that the end
must be one which it can be a universal law for everyone to have. Hence
the first principle of the doctrine of virtue is: act according to a
maximof ends which it can be a universal law for everyone to have. For
what in the relation of man to himself and others can be an end, that
is an end for pure practical reason.
In his remarks about this first principle, it is clear that the end
to which the principle directs us is humanity (man as such), and, I
think, whatever ends are consistent with the value of humanity. At this
point, the argument must be like the argument leading to the Formula
of the End in Itself, for Kant moves inmediately to the idea that one
must both refrain from treating humanity as a mere means and do something
positive to express the fact that humanity is one's end. The particular
ends to which this first principle leads us are of course the ends as-
sociated with the second contradiction test: the happiness of others
and one's own perfection. So I suppose that when he speaks of what can
be an end, he means what has the right rational status to be an end:
whatever end can be universalized. In effect, this argument, like the
 230

argument for the Formulaof the End in Itself, says that pure practical
reason is only possible if something is an objective end; but for
reasons we have seen before, only if humanity is valued as an end in
itself can there be any objective ends. Henceyou must make humanity
your end, and that will involve an inner commitmentboth to not treating
humanityas a mere means and to treating humanity, positively, as an
end. That, as we have already said, involves making the ends of others
your ends; whatever can be an end, is an end for pure practical reason -
and whatever can be an end is whatever can be made an end consistent with
the value of humanity. The argument, put intuitively, moves us from a
commitmentto pursuing only what is good to a conmitmentto pursuing
whatever is good, on the grounds that pure reason can be practical only
if we can be movedto pursue something solely by the fact that it is
good.
Supposing instead that by 11determine the maximsfor action
(since every maximcontains an end)11 Kant means 11set a rational standard
for maxims"the deduction might proceed as follows. If pure practical
reason is possible, it sets a standard for maxims(that they must be
able to be willed universally). If it sets a standard for maxims, how-
ever, it sets a standard for ends, for every maximcontains an end,
and so if somemaximsare accepted and some rejected, so are some ends.
Furthermore, if somemaximsmust be accepted, so must someends: the
ends associated with the maximsthat must be accepted are objective ends.
Since humanity is the material of the moral law, one could presumably
movefrom the idea that somemaximsmust be accepted to the idea that
humanitymust be made an end. This reading finds some support in Kant's
emphasis on the fact that every maximcontains an end.
 231

Howeverthat may be, it is clear that the end to which Kant thinks
the principle of virtue directs us is humanity, and the duties of virtue
are those that arise from the fact that we ought to make humanity our
end. Pursuing one's own perfection and the happiness of others are
ways of acting that express the fact humanity is one's end; indeed, the
whole of the Doctrine of Virtue consists of a careful, detailed, expli-
cation of howa person will act when she has made humanity in her own
person and that of others an end in itself. The duties of justice,
by contrast, involve ways of acting consistently with external freedom;
they describe a way of acting, but say nothing about the reasons we are
to act in that way. This means that it is whenyour action is governed
by the fact that you have made humanity an end in itself that it is a
truly reasonable action.

II. Acting FromDuty I: Obligatory Ends

The essence of a juridical duty is that it is demandedof us in the
nameof universal outer freedom. Consistency with external freedom is
the basis of the obligation to perform the required act or omit the for-
bidden one. But it does not matter, in the doctrine of law, whyyou do
so. It is possible for others to coerce you to perfonn such a duty: in-
deed, they are authorized to do so, to protect their own rights. It is
ethics that makes it a duty to do your duty from the motive of duty, and
so to bring moral worth to your action. It is the mark of a perfectly
virtuous person - a person of good character - not only to adopt and to
do something for the sake of the obligatory ends, but to do his or her
duty from the motive of duty.
The discussions in the Introductions to the Metaphysics of Morals
 232

and the Doctrine of Virtue provide us with a picture of how it is pos-

sible to be motivated to an action by reason alone, without aid or
impetus from any incentive of inclination. In order to make this pic-
ture as complete as possible, it is necessary to turn briefly to the
moral psychology presented in Religion Within the Limits of Reason
Alone. Kant himself introduces the Metaphysics of Morals with a state-
ment of his later views about the will, but the presentation of these
views in the Religion is more complete. In what follows I will give a
summaryof these views, and then describe the action from duty in the
case of the obligatory ends. In the following section I will take up
the case of the juridical action from duty.
In the Religion and in the Introduction to the Metaphysics of
Morals, Kant is interested in the distinction between two aspects of
the will (of the appetitive power of faculty of desire); in German, the
Willkur and the Wille. 2 In the Metaphysics of Morals, Kant describes
them as two aspects of the appetitive power. Willkur, often translated
as 11choice 11 , "power of choice", or "arbitrary will", is the appetitive
power in action - making choices, adopting maxims. The Wille, identi-
fied with practical reason,is the appetitive power as giving laws.
The appetitive power which acts in accordance with concepts,
in so far as the ground determining it to action lies in it-
self and not in the object, is called a power to act or to
refrain from acting at one's discretion. It is called choice
[Willkur] when it is joined with consciousness that its action
can produce the object; otherwise its act is called a wish.
The appetitive powerwhose inner determining ground, and so
the decision itself, lies in the subject s reason is called
1

the will [Wille]. The will [Wille] is therefore the appetitive

power viewed in relation to the ground that determines the
power of choice [Willkur] to the action, while the power of
choice [Willkur] is the appetitive power viewed in relation
to the action. The will [Wille] itself has, properly speaking,
no detennining ground: in so far as reason can determine the
power of choice [Willkur], the will [Wille] is, rather, prac-
ti.ca 1 reason i tse 1f. (_MM
9-10/213)
 233

It is perhaps helpful, and not too misleading, to think of the Willkur

as the representative of the will s negative freedom and the Wille as
1

the source of its positive freedom.3

In Religion Within the Limits of Reason Alone, Kant uses the dis-
tinction in describing what it is to have a good or an evil character. 4
One's character is the basis of all one1 s choices, actions, and deci-
sions; yet, as good or evil, it is something for which one must be re-
garded as responsible. 8ut nothing is morally evil (i.e.
11
capable of
being imputed) but that which is our ownact. 11
(R 26) Kant therefore
envisions us as, in effect, choosing our own character. This choice is
an intelligible act, outside of or preceding our temporal existence,
11 11

which shows up in all of our decisions and choices - it may be, during
the whole course of our life. In the Critique of Practical Reason,
Kant says:
There are cases in which men, even with an education which was
profitable to others, have shownfrom childhood such depravity,
which continues to increase during their adult years, that
they are held to be born villains and incapable of any im-
provement of character; yet they are judged by their acts,
they are reproached as guilty of their crimes; and indeed,
they themselves find these reproaches as well grounded as if
they, regardless of the hopeless quality ascribed to their
minds, were just as responsible as any other men. This could
not happen if we did not suppose that whatever arises from
man's choice (as every intentional act undoubtedly does) has
a free causality as its ground which from early youth expresses
its character in its appearances (its actions). These actions,
by the unifonnity of conduct, exhibit a natural connection.
But the latter does not render the vicious quality of the will
necessary, for this quality is rather the consequence of the
freely assumed evil and unchangeable principles. (C2 103/
99-100)

Since all choices and decisions are madeon the basis of maxims, and
since our character governs our choices and decisions, Kant pictures our
character as consisting in a sort of highest or most fundamental maxim.
This highest maximis most fundamental both in the sense that it
 234

governs all other decisions and choices and the adoption of maxims
generally, and in the sense that nothing (no further maxim)governs it:
it is an original choice. Since we are responsible for our character,
Kant sees the fundamental maximas being itself a sort of act. It is
the Willkur's choice of a law for its ownoperations - it is the choice
of a principle of choice .
... the judgment that the agent is an evil man cannot be
made with certainty if grounded on experience. In order,
then, to call a man evil, it would have to be possible a
priori to infer from several evil acts done with consciousness
of their evil, or from one such act, an underlying evil maxim;
and further, from this maximto infer the presence in the a-
gent of an underlying ground, itself a maxim, of all particu-
lar morally-evil maxims. (R 16)
... the source of evil cannot lie in an object determining
ll,can lie oty
the Willkur through inclination, nor yet in a natural impulse;
in! rule made~ the Willkur for the use of
1ts freedom my emphasls] that is, in a maxim. But now it
iirust not be considered pennissable to inquire into the sub-
jective ground in man of the adoption of this maximrather
than its opposite. If this ground were not itself a maxim
but a mere natural impulse, it would be possible to trace the
use of our freedom wholly to determination by natural causes;
this,however, is contradictory to the whole notion of freedom.
Whenwe say, then, Manis by nature good, or, Manis by nature
evil, this means only that there is in him an ultimate ground
(inscrutable to us) of the adoption of good maximsor of evil
maxims(i.e. those contrary to law), and this he has, being
a man; and hence he thereby expresses the character of his
species. (R 17)
The tenn "act" can apply in general to that exercise of free-
domwhereby the suprememaxim(in harmonywith the law or con-
trary to it} is adopted by the Wi11kur, but a 1so to the exer-
cise of freedom whereby the actions themselves {considered
materially, i.e. with reference to the objects of the Willkur}
are perfonned in accordance with that maxim. The propens1ty
to evil, then, is an act in the first sense (aeccatum origin-
arium), and at the same time the formal groun of all unlawful
conduct in the second sense, {R 26}
The original act of the Willkur, then, is to adopt a maximwhich will
govern its subsequent operations: a rule for the use of its freedom;
it is clear from these passages and the first book of the Religion
 235

generally, that this maximwill be for or against the moral law. The
moral law, as the law of practical reason, is dictated by the Wille,
and ought to be the maximof the Willkur. The nature of the fundamental
maximis this: it detennines which of the incentives associated with
self-love or the incentive of respect for the moral law will be the
unconditioned condition of the other. Both sorts of incentives inevita-
bly influence us, for it is part of our predisposition that they should.
(R 21) But it is for the Willkur to detennine whether it will satisfy
the demandsof self-love only when they are consistent with morality,
or, on the contrary, satisfy the demandsof morality only when they are
consistent with self-love. It does this by the way in which these in-
centives are incorporated into the fundamental maxim:
... freedom of the Willkur is of a wholly unique nature in
that an incentive can determine the Willkur to action only
so far as the individual has incorporated ,t into his maxim
(has made it the general rule in accordance with which he will
conduct himself); only thus can an incentive, whatever it
may be, co-exist with the absolute spontaneity of the Willkur
(i.e. freedom). But the moral law, in the judgment of reason,
is in itself an incentive, and whoever makes it his maximis
morally good. (R 19-20)
The law, rather, forces itself upon him irresistably by virtue
of his moral predisposition; and were no other incentive work-
ing in opposition, he would adopt the law into his supreme
maximas the sufficient determining ground of his Willkur;
that is, he would be morally good. But by virtue of an equally
innocent natural predisposition he depends upon the incentives
of his sensuous nature and adopts them also {in accordance
with the subjective principle of self-love) into his maxim.
If he took the latter into his maximas in themselves wholl{
adequate to the determination of the WilTkur, without troub ing
himself about the moral law (which, after all, he does have
in him), he would be morally evil. Nowsince he naturally
adopts both into his maxim, and since, further, he would find
either,7fit were alone, adequate in itself for the deter-
mination of the Willkur, ...
... the distinction between a good man and one who is evil
cannot lie in the difference between the incentives which they
adopt into their maxim(not in the content of the maxim), but
rather must depend upon the subordination (the form of the
maxim), i.e. which of the two incentives he makes the condition
 236

of the other. Consequentlyman (even the best) is evil only

in that he reverses the moral order of the incentives when
he adopts them into his maxim. He adopts, indeed, the moral
law along with the law of self-love; yet when he becomesaware
that they cannot remain on a par with each other but that one
must be subordinated to the other as its supreme condition,
he makes the incentive of self-love and its inclinations the
condition of obedience to the moral law; whereas, on the
contrary, the latter, as the supreme condition of the satis-
faction of the fonner, ought to have been adopted into the
universal maximof the Willkur as the sole incentive. (R 31-
32)
I have quoted this last long passage once already, in showing howclosely
this "propensity to evil" is related to the self-conceit which accom-
panies inclinations, and which is described as humiliated and broken
downby respect for the moral law according to the account in the
· Critique of Practical Reason.5 Kant believes that the propensity to
evil is a universal (though contingent) feature of humanbeings. It
must be overcomeby a moral revolution, whereby the ranking of incentives
in the fundamental maximis reversed, and they are put in their proper
order.
But if a man is to becomenot merely legally, but morallh,
a good man (pleasing to God), that is, a man endowedw;t
virtue in its intelligible character (virtus noumenon)and
one who, knowingsomething to be his duty, requires no in-
centive other than this representation of duty itself, this
cannot be brought about through gradual reformation so long
as the basis of the maximsremains impure, but must be effected
through a revolution in the man's disposition (a going over
to the maximof holiness of the disposition}. He can become
a newman only by a kind of rebirth, as it were a new creation
(John III; comparealso Genesis I, 2), and a change of heart .•..
... man is under the necessity of, and is therefore capable of,
a revolution in his cast of mind, but only of a gradual re-
form in his sensuous nature (which places obstacles in the
way of the former). That is, if a man reverses, by a single
unchangeable decision, that highest ground of his maximswhere-
by he was an evil man (and thus puts on the newman}, he is,
so far as his principle and cast of mind are concerned, a
subject susceptible of goodness, but only in continuous labor
and growth is he a good man. (R 42-43}
To have made this resolution is to have set oneself upon the path of
 237

virtue.
It is here that the objective ends come in; for in order to carry
out the virtuous resolution, one must freely adopt the ends of practical
reason:
For since the sensuous inclinations tempt us to ends (as the
matter of choice) which.may be contrary to duty, legislative
reason can check their influence only by another end, a moral
end set up against the ends of inclination, which must there-
fore be given a priori independently of the inclinations.
(DV38/379-380T
This passage, which appears almost at the beginning of the Doctrine of
Virtue, can be read as follows. Suppose you attempt this moral revolu-
tion, establfshing the moral incentive as supreme and sufficient for
your will. You resolve to do this. and in doing so you resolve to defeat
the universal humanpropensity to evil - the tendency to self-conceit,
and the temptation to allow self-love to be legislative. Nowwhat this
amounts to is adopting humanity, as an end in itself, as your own
unconditioned end. This can be seen both from the identity of the
fonnulae of the categorical imperative and from the nature of the pro-
pensity of evil which is to be defeated: it is a propensity to rank
your own happiness higher than humanity. Nowhumanity is not an end to
be achieved but an end not to be acted against; and since our humanity
just is our capacity to confer value through rational choice, the way
not to act against it is to act according to the categorical imperative,
in the Formula of Universal Law, which is the one that gives the right
way of acting in detail. To make the categorical imperative as far as
possible your law is not different from making humanity your end. Virtue
is made possible by adopting the categorical imperative materially.
This act, the adoption of humanity as one1 s highest end,makes it possible
to divest the sensuous inclinations of the self-conceit that plays their
 238

advocate, and to makemorality the condition of their satisfaction.

In order to effect your good resolution, therefore, you must make
humanity your unconditioned end, as the categorical imperative indeed
directs. Whenyou really make something your end, however, that involves
more than merely not treating it as a means, and it is this that leads
to the adoption of the obligatory ends, as we have seen. These, in
turn, are ends that can be promoted. So, in the case of action under-
taken in light of an obligatory end we have a clear picture of an action
that is motivated solely by reason and done entirely from duty. Having
made the good resolution that the categorical imperative should be your
law, you have taken humanity as your unconditional end. Since humanity
is the power to choose rationally and confer value, making it an end
involves making an end of your own talents and powers and of the things
that are valued as part of the happiness of others. Makingthese your
end is resolving to do something to promote them. It is an end that
can make it a law to have a certain maxim. Because you have these ends,
you adopt certain maximsof actions; say, to do this or that for the
sake of another's happiness. This maximis, in turn, tested by the
categorical imperative's first contradiction test, so that you may be
sure that what you do is consistent with justice and external freedom.
Supposing that it is, and that you perform the action, you have done an
acti.on that is moti·vated entirely by reason. No inclination has entered
at any point. The adoption of the ends and the maximsinvolved all
spring from the categorical imperative, the maximwhose fonn and matter
are one.
There is, however, a little more to it than that. In order to see
this, we must thi"nk about what it means to make something one's end. It
 239

is not as if, having adopted the obligatory ends as objective and per-
manent objects of the Willkur, you at that momentwork out the meansyou
will use to promote them, for all the rest of your mortal life. Nor
can you really be said to have made something your end if all that
amounts to is that you acknowledgeit to be an end, and refrain from
acting against it. Instead, a person who has an end must both take
action to promote it and also be alert to its occasions. Youcan only
be said to really have something as your end if you are inclined to
notice whenyour circumstances involve opportunities to promote or rea-
lize that end or threats to it and ff you are inclined to act according-
ly. Consider an example. Suppose one man says of another: "The
happiness of my friend is my end; it is very important to me; it is
one of the things that I care most about in the world." One thing we
may then expect of this man is that he will do something - take action -
to secure the happiness of his friend. But we will also expect that
the manwill be alert, in manyways, to the occasions relevant to the
happiness of his friend: we expect the happiness of his friend to occur
to him under such circumstances. For example, when he sees something
that his friend would like or care about, he remembersto tell his
friend about it, or if it is an object perhaps even buys it for his
friend; or when something happens that will be relevant to his friend's
interests and concerns, this relevance occurs to him. This is, to
varying extents, the way our ownprivate ends effect us; they color our
perceptions of the world. A- psychological theory of what it really
means to have something as one's end would tell us howthis occurs;
but I do not think that anyone as yet has such a theory. 6 Yet it is
unquestionable that something like this goes on; it is perception, not
 240

calculation, that alerts us to the bearing that our experiences might

have on the things that we care for. Weare generally alert to the
occasions of our own happiness, to circumstances and events in the world
that might influence our own plans and projects, and this alertness is
part of what it is to really have those plans and projects. If the man
says: my friend's happiness is my end, but in fact his friend and his
friend's happiness never so muchas cross his mind, we do not believe
him, even though he never does anything contrary to his friend's hap-
piness. Even if the man took action - say, visited this person on a
regular basis or performed some habitual favor for him - but still never
thought of his friend except when he was performing these habitual ac-
tions, we might feel inclined to doubt that his friend's happiness was
his end. Weshould perhaps think that he was seeking some other gratifi-
cation - perhaps the gratification of telling himself that his friend's
happiness is his end, when really, it is not.
I think that this has something to do with the imperfection of du-
ties that involve ends. For although we can make a point of adopting
an end, as an act of will, we do not have full and immediate control
over the extent to which we perceive and respond to the world around us
in terms of this end. It is only gradually, only by working at it, only
by practicing, and perhaps only by divine grace, that we can make some-
thing our end in the full sense - the sense in which its occasions are
luminous to us - by a resolution of the will. It is partly because this
is a gradual process that virtue, according to Kant, admits of degrees,
even though the resolution that is the basis of virtue is to be regarded
as an all-or-nothing action of the Willkur that either has been taken or
has not. It might even be possible to say that the width of duties has
 241

to do with the leeway we have in satisfying them, (exactly what we will

do to satisfy them); whereas their imperfection has to do with the fact
that they are satisfiable to varying extents, depending upon the extent
to which we have internalized the ends involved. I would not want to
claim that Kant's terminology always fits this reading, but I think it
might, roughly, in the following important passage:
But a wide duty is not to be taken as a permission to make
exceptions to the maximof actions, but only as a permission
to limit one maximof duty by another (e.g. love of one's
neighbor in general by love of one's parents) - a permission
that actually widens the field for the practice of virtue.
- As the duty is wider so man's obligation to action is more
imperfect; but the closer to narrow duty (Law) he brings the
maximof observing this duty (in his attitude of the will),
so much the more perfect is his virtuous action.
Imperfect duties, accordingly, are only duties of virtue.
To fulfill them is merit (meritum = +a); but to transgress
them is not so much{uiij (demeritus = -a) as rather mere
lack of moral worth = , unless the agent makes it his
prfiic,ple not to submit to these duties. The strength of one's
resolution, in the first case, is properly called only virtue
(virtus); one's weakness, in the second case, is not so much
vice {vitium) as rather mere want of virtue, lack of moral
strength (defectus moralis). As the word Tugend [virtue] comes
from tauqen [to be fit for], so Untugend [lack of virtue] comes
from E! nichts taugen [to be worthless].) Every action con-
trary to duty is called a transgression (peccatum). It is
when an intentional transgression has been adopted as a
bas·ic principle that it is properly called vice (vitium).
(DV49-50/390) -
Although the law does not specify exactly what and howmuchwe are
to do in promotion of the obligatory ends, we gain in merit and moral
perfection as we do more. Since moral perfection is itself one of the
obligatory ends, this is required of us, although this too is only a
wide duty. In our endeavor to becomemore perfect, it is the self-
conceit accompanyingthe sensuous inclinations as well as those inclina-
tions themselves that stand in our way. For one's own happiness is a
natural end, and we are naturally inclined to see the world in terms of
the occasions and opportunities of our ownhappiness - we think of things
 242

and events as they bear upon our private ends. In order to attain
virtue, we must learn to look beyond our private happiness to the in-
terests of humanity. Weneed not sacrifice our happiness or leave it
out of account, but we must learn at least to make it secondary to
this higher end. And there is no question that the more thoroughly
and perfectly humanity as an end in itself becomesthe end we have in
view in our actions, the more we are likely to devote outselves to
the obligatory ends. To do so is rational in a double sense. It is
rational because the moral law commandsthat humanity as an end in it-
self be the limiting condition of all our demandsfor happiness; and it
is rational because in the pursuit or promotion of an obligatory end we
engage in a kind of action that is motivated entirely and purely by
practical reason, with no impetus from the sensuous incentives at all.

III. Acting FromDuty II: Juridical Action

The case for a kind of action which is motivated solely by reason
is not difficult to understand in the case of the obligatory ends,
precisely because in this case we are·acting with a definite end. It
is only necessary to showthat reason can dictate this end; then we have
reason dictating the end, constructing the maxim, testing the maxim,
and an action that is rational throughout. The role of ends in the
other case - the juridical duty that is done from the motive of duty -
is a little more complicated.
SometimesKant speaks as if ends played no role in the case of the
juridical duty that is done from the motive of duty. For instance:
In the same way, every ethical obligation implies the concept
of virtue, but not all ethical duties are thereby duties of
virtue. Those duties which concern, not so mucha certain
end (matter, object of choice), but what is merely formal in
 243

the moral determination of the will (e.g. that the due action
should also be done from the motive of duty), are not duties
of virtue. Only an end winch is alsoa duty can be ca~led a
duty of virtue. For this reason there can be manyduties
of virtue (and also manydifferent virtues), while there is
only one fonnal element of moral choice (one virtuous atti-
tude of will), which is, however, valid for all actions.
(DV41/383)
But what it is virtuous to do is not necessarily a duty
of virtue in the proper sense. The practice of virtue can
have to do merely with the formal aspect of our maxims, while
a duty of virtue is concerned with their matter - that is with
an end which is also conceived as a duty. (DV55/394-395)
The third thing concerns the division of the material from
the formal (purposefulness from lawfulness) in the principle
of duty. Regarding this, it should be noted that not every
obligation of virtue (obligatio ethica) is a duty of virtue
(officium etfiicum s. virtutis): reverence for law as such
does not yet estabTish an end as a duty, and only an obliga-
tory end is a duty of virtue. - Hence there is only one
obligation of virtue, whereas there are Tan~ duties of virtue;
for there are indeed manyobjects which it 1s our duty to have
as our ends, but only one virtuous attitude of will, as the
subjective ground detennining us to fulfill our duty. This
ethical obligation extends over juridical duties too, but
it does not entitle them to be called duties of virtue. (DV
73/410)
These passages seem explicitly to divide off the specific duties of vir-
tue, which involve ends, from juridical duties done ethically, which do
not. There are, however, several difficulties with taking this as Kant's
whole or final view of the matter. First, if there are any positive
duties of justice, then in these cases it looks as if not only an omis-
sion but an action and its maximare required of us. But it is only an
end that can make it a law to have a certain maxim. Second, Kant some-
times says things, in the introduction to the Doctrine of Virtue itself,
which suggest that we are either acting under the influence of a freely
adopted end or we are acting under complusion. Third, the duty to do
one's duty from the motive of duty appears here as an obligatory end -
the end of cultivating one's moral perfection - and Kant is very con-
 244

cerned to explain why, in spite of what one might think, this is a

wide duty. Fourth, Kant sometimesspeaks as if a law-abiding action has
no merit unless, in performing that action, you have made the law your
motive in the sense that you have made the rights of humanity your
end. Thus it seems as if on the one hand we are told that this sort of
rational action involves no end, on the other hand it seems bound up in
someway both with our duty to promote humanity in our own person and
our duty to respect humanity in the persons of others. These claims
must be untangled if we are to understand the juridical action that
is done from duty.
I am going to begin by making some remarks about each of the four
difficulties that I have just mentioned. About the first, the associa-
tion of obligatory ends with the idea that it is our duty to have a
certain maxim,I havealready said something. Granting Kant's assertion
that the categorical imperative makes itself knownto us whenever we
formulate a maxim(the Fact of Reason), it is easy to see (always assum-
ing a good resolution has been taken) howwe could be motivated by pure
reason to omit an action, howeverdesirable, if reason thus alerts us
that it is wrong. Humesupposes that reason can only oppose an action
if it can set an end:
Since reason alone can never produce any action or give
rise to volition, I infer, that the same faculty is as incapa-
ble of preventing volition, or of disputing the preference
with any passion or emotion. This consequence is necessary.
'Tis impossible that reason cou'd have the latter effect of
preventing volition, but by giving an impulse in a contrary
direction to our passion; and that impulse, had it operated
alone, wou'd have been able to produce volition. Nothing can
oppose or retard the impulse of passion, but a contrary im-
pulse; and if this contrary impulse ever arises from reason,
that latter faculty must have an original impulse on the will,
and must be able to cause, as well as hinder, any act of vo-
lition. But if reason has no original influence, 'Tis im-
possible it can withstand any principle, which has such as
 245

efficacy, or ever keep the mind in suspense a moment.7

Nowalthough Kant does want to say that reason can give rise to a
"volition" on its own, and so in somesense does not want to deny that
part of Hume's point here, it nevertheless seems possible on the Kantian
account to see how reason can oppose an action without first setting
up an end. Wecan be motivated by practical reason (though we cannot
say how); it makes its standards knownwhenwe formulate a maxim; hence
if the maximis unacceptable, the action may be omitted under the in-
fluence of reason. Wecan even see on this basis howwe can perform an
action under the pure influence of reason, so long as the maximof its
opposite occurs to us and is rejected, as in the first set of examples
in the Foundations. In a certain way Kant's argument seems to go in a
reverse direction from that which Humethought the argument would have
to take, movingfrom the more obvious opposing influences of reason to
its positive capacity to set an end, as in the deduction of the first
principle of virtue. Yet the statement (DV38/379-380, quoted above on
p. 237), that legislative reason can check the influence of the sensuous
inclinations only by setting up a moral end, suggests that even in this
case the objective end may be operative. And in the case of the positive
juridical duty, where it seems that reason must direct us to the adoption
of the maximand so to the action, it looks as if we need an end that
will "makeit a law to have such a maxim." (DV48/389; quoted above
on p. 222)
The second difficulty I mentioned arises from various remarks in
which Kant seems to associate autonomousactions and freely adopted ends
in a way that makes it hard to imagine the juridical duty being done
autonomouslyin the absence of an end. I have already quoted, on p. 222,
 246

one of the passages where Kant seems to oppose free action, which is
based upon a freely adopted end, with compelled action {where another
coerces you by way of some end you already have}. There Kant says that
detennination to an end is the only determination of choice which in
11

its very concept excludes the possibility of complusion through natural

means by another's act of choice." (DV39/381-382} Shortly before this
remark Kant says:
An end is an object of the power of choice [Willkur] (of
a rational being}, through the thought of which choice is
determined to an action to produce this object. - Now
another person can indeed compel me to perfonn actions which
are means to his end, but he cannot compel me to have an end;
only I myself can make something my end. - Butthe notfon
that I am under obligation to take as my end something that
lies in the concepts of practical reason, and so to have a
material ground of choice beyond the formal one that Law
contains, would be the concept of an end which is in itself
a duty. The doctrine of this end would not belong to the
doctrine of Lawbut rather to ethics, since the concept of
self-constraint in accordance with moral laws belongs only
to ethics.
For this reason ethics can also be defined as the system
of the ends of practical reason - Ends and duties [to which
we can be compelled by others] differentiate the two divisions
of moral philosophy in general. (DV38-39/381}
These kinds of remarks strongly suggest that the autonomousaction always
involves a freely adopted end.
The third, and in someways, most forceful difficulty with the view
that the juridical duty done from the motive of duty involves no end is
that in the Doctrine of Virtue this sort of action shows up~ an end.
Part of the duty of cultivating our perfections (and so of treating
humanity in our own person as an end) is the duty of cultivating our
moral perfection, where moral perfection consists in the disposition to
do one's duty from the motive of duty. Nowof course in one sense this
looks like it is no problem at all: if the juridical duty done from the
motive of duty does involve acting without an end, then this is only to
 247

say that we must make an end of our capacity to act without any end;
that is, of our capacity to act from pure respect for the law. Yet
if this end were always involved in our motivation whenwe act, or try
to act, from duty, then for~' rational action would always involve
an obligatory end. For us, that is, the pure juridical act would always
be mediated by the attempt to realize a certain end - the end of living
up to humanity in our ownperson.
A notorious difficulty with this view, and one recognized by Kant,
is that moral perfection is a wide duty. As Kant himself points out,
it looks as if it is our rigorous duty always to do our duty from the
motive of duty. Kant explains why it is not in terms of our incapacity
to know(for certain) our ownmaxims: A thing cannot be absolutely
cormnanded
of us whenwe have no way of telling whether we are fulfilling
the conmandor not. He explains as follows:
Man's greatest moral perfection is to do his duty and this
from a motive of duty (to makethe law not merely the rule
butaTso the motive of his actions). - Nowat first sight
this looks like a narrow obligation, and the principle of
duty seems to prescr1be with the precision and strictness of
a law not merely the legality but also the morality of every
action, the attitude of will. But in fact the law, here
again, prescribes only the maximof the action: a maximof
seeking the ground of obligat1on solely in the law and not in
sensuous inclination {advantage or prejudice). It does not
prescribe the action itself. - For man cannot so scrutinize
the depths of his own heart as to be quite certain, in even
a single action, of the purity of his moral purpose and the
sincerity of his attitude, even if he has no doubt at all
about the legality of the action .
•.• The law does not prescribe this inner action in the human
mind itself but only the maximof the action: the maximof
striving with all one's might to make the thought of duty for
its ownsake the sufficient motive of every dutiful action.
(DV52-53/392-393)
The above passage is from the Introduction. In the body of the text,
Kant says:
In the first place, this duty consists, subjectively, in
 248

the purity {puritas moralis} of one's attitude in duty: that

even without any admixture of purposes derived from the sen-
suous inclinations, the law is of itself the motive and ac-
tions are done not merely in confonnity with duty but also
from the motive of duty. - Here the commandis "be holi'.
Secondly this duty is concerned, objectively, with the
whole moral end of perfection - that is, of performing all
one's duties and achieving the complete moral end in relation
to oneself. Here the commandis "be perfect." {DV113/446)
This duty to oneself is a narrow and perfect one in its
quality; but it is wide and imperfect in its degree, and this
because of the frailty {fragilitas) of humannature.
It is man's duty to strive for this perfection, but not
to achieve it (in this life), and his pursuit of perfection,
accordingly, is only a continual progress. Hence while this
duty is indeed a narrow and perfect one with regard to the
object {the Idea which one should make it one's end to
realize), in relation to the subject it is only a wide and
imperfect duty to oneseTf. (DV113/446)
Hhat Kant says about the objective part of this duty is of special in-
terest, because it gives us one possible answer to the question of how
we are motivated to the positive juridical duty. If you make an end
of your ownmoral perfection, then you have a reason to be alert to the
occasions of duty. As with any end, the extent of this alertness is
a matter of degree; and so your perfection, even in the sense of whether
you always rememberto do what you ought, is also a matter of degree.
For although you have madean end of moral perfection, you may in a
given case not do what is required of you: this is what Kant calls
frailty
11 11
• (R 25) Andeven whenyou do what is required of you, you
cannot be sure that duty was the sole motive, and that inclination and
vanity have played no role: this is what Kant calls "impurity". (R 25)
The fact that this duty involves an end explains how, with the best will
in the world, we can still fail to be complete in our observance of
what duty enjoins.
The fourth difficulty with the view that the pure juridical act
involves no end arises from the following passage:
 249

Although confonnity of actions with Law(being a law-

abiding man) is not meritorious, the confonnity with Law
of the maximof such actions regarded as duties, i.e.
reverence for Law, is meritorious. For by this we make
the right of humanity, or also the rights of men, our end
and widen our concept of duty beyond the notion of whaTTs
due (officium debiti), since another can demandby r1ghf
that my act1ons conform with the law, but not that the law
be also the motive for my actions. {DV50/390-391)
This passage expresses the view of the nature of the pure juridical act
that Kant's remarks about the role of ends in the determination of max-
ims and the argument for the first principle of virtue lead one to
expect. It is because you have made humanityyour end that the pure
juridical act has merit. And, if you have acquired this virtuous at-
titude of will, you will be led to do the positive duty of justice,
when it occurs, because the right of humanity is the end you have in
view. Your concern for the right of humanity can motivate you, not
just to limit your conduct in accordance with the moral law, but ac-
tually to adopt maximsthat are called for by the rights of others.
The important thing is to see whyall of these different accounts
of the action that is done from duty are not in conflict with one
another. At first appearance, of course, it looks as if they are:
there is so muchemphasis throughout Kant's writings on the idea that
whenyou act from duty you are responding to and motivated by the form
of the maxim,doing what you are called to do without regard to con-
sequences, and so forth, that it seems contrary to all that Kant says
to bring in even such ends as one's ownmoral perfection or the rights
of others. But it is not, for two important and closely related reasons.
One has to do with the kind of end that humanity is, i.e. that it is the
material of the moral law itself. The second emerges whenwe consider
what is involved in the adoption of a maxim.
 250

It must always be rememberedthat Kant thinks that every maximcon-

tains an end. This is certainly one of the things that he meanswhen
he says that practical reason is a faculty of ends. Consider again the
proof of the hypothetical imperative: wheneveryou will something,
you regard yourself as a cause. No act of will takes place without an
end in view. Every maximthat is adopted has an end in it, in some
sense, whether it is adopted because of an inclination for its end,
because of a rational interest in the end, or because of the form of
the maxim. The action nevertheless has an end, and when the maximis
freely adopted, the end is freely adopted also. It is this fact that
enables Kant to make the close association between internal freedom and
the adoption of ends that he makes at the beginning of the Doctrine of
Virtue, even while he says that not every virtuous action is a duty
of virtue.
Nowthis leads Kant to say that we always act with an end in view,
regardless of what motivates us to adopt a certain maxim. For without
an end in view, we are not willing; there is nothing of which we are
regarding ourselves as a cause. Even in the case where the juridical
action is demandedof us and we act out of pure respect for the law,
there must be some kind of end involved. It is easy to see why the
preservation of the rights of humanity is a good candidate for this
end. But the preservation of the rights of humanity is not an end that
can be promotedor achieved by means that are ascertained theoretically;
only the categorical imperative can teach us howto act with this end
in view. So Kant's view is that whenyou do a juridical duty from
duty, you adopt the maximbecause of its form and its necessity, and
the end you have in view in doing so may be either the right of humanity,
 251

or your ownmoral perfection, or perhaps both. Perhaps in the first

case we should say that you have the virtue of justice, and in these-
cond case we should say you have the virtue of honorableness, although
this is not Kant's language now. But in either case the fonn of the
maximis your motive; only it movesyou under the influence of a certain
attitude of will. In either case we could even say that it is the end
that makes it a duty to have the moral maxin of always acting as far
as possible out of pure respect for the moral law.
Other ideal ends of practical reason can play this role as well.
Sometimeswhen Kant refers to the end we have in view in doing our duty,
he identifies the Highest Goodas the end in question. For instance:
For these laws command absolutely, be the consequence what
it will; indeed, they even require that the consideration of
such consequence be completely waived when a particular act
is concerned; and thereby they makeduty an object of high-
est respect without offering or proposing to us an end (or
a final end) such as would have to constitute duty's recommen-
dation and the incentive to fulfillment of our duty. All
mencould have sufficient incentive if (as they should) they
adhered solely to the dictation of pure reason in the law.
Whatneed have they to knowthe outcomeof their moral actions
and abstentions, an outcomewhich the world's course will
bring about? It suffices for them that they do their duty;
even though all things end with earthly life and though, in
this life, happiness and desert maynever meet. Andyet it
is one of the inescapable limitations of manand of his
faculty of practical reason (a limitation, perhaps, of all
other worldly beings as well) to have regard, in every action,
to the consequence thereof, in order to discover therein what
could serve him as an end and also prove the purity of his in-
tention - which consequence, though last in practice (nexu
effective) is yet first in representation and intentionlnexu
finali). In this end, if directly presented by reason alone,
man seeks something that he can love; therefore the law, which
merely arouses his respect, even though it does not acknow-
ledge this object of love as a necessity does yet extend it-
self on its behalf by including the moral goal of reason among
its detennining grounds. That is, the proposition: Makethe
highest good possible in the world your own final end! is a
synthetic proposition a priori, which is introduced by the
moral law itself; although practical reason does, indeed, ex-
tend itself therein beyond the law. This extension is possible
 252

because of the moral law's being taken in relation to the

natural characteristic of man, that for all his actions he
must conceive of an end over and above the law. (R 6-7n)
Perhaps even more striking is this closely related passage, also from
a footnote in "On the Common
Saying: 1
This Maybe True in Theory
but it does not Apply in Practice. 111

The necessity of assuming as the ultimate end of all things

a highest good on earth, which it is possible to achieve with
our collaboration, is not a necessity created by a lack of
moral incentives, but by a lack of external circumstances
within which an object appropriate to these incentives can
alone be produced as an end in itself, as an ultimate moral
end. For there can be no will without an end in view, al-
though we must abstract from this end whenever the question
of straightforward legal compulsion of our deeds arises,
in which case the law alone becomesits determinant .•.. And
the necessity of an ultimate end posited by pure reason and
comprehendingthe totality of all ends within single principle
(i.e. a world in which the highest possible good can be
realized with our collaboration) is a necessity experienced
by the unselfish will as it rises beyondmere obedience to
formal laws and creates as its ownobject the highest good.
This idea of the totality of all ends is a peculiar kind of
determinant for the will. For it basically implies that if
we stand in a moral relationship to things in the world -
around us, we must everywhere obey the moral law; and to this
is added the further duty of working with all our power to
ensure that the state of affairs described (i.e. a world
conforming to the highest moral ends) will actually exist.
In all this, manmay see himself as analogous to the
divinity. For while the divinity has no subjective need of
any external object, it cannot be conceived of as closed up
within itself but only as compelled by the very awareness of
its own self-sufficiency to produce the highest good outside
itself. (Theory and Practice 65)
There is, besides these, another end which we can have in view whenwe
do our duty from duty, which is especially associated with the duties
of justice. Wecan see ourselves as working towards Perpetual Peace,
which according to Kant is the highest political good and the only con-
dition under which a universal juridical condition of things - that is,
a universal external freedom with rights guaranteed - can be realized.
{DJ 128-129/354-355)
 253

The person whoacts from duty is motivated only by respect for

the law, without regard to ends. Andyet it is nevertheless true that
in being so motivated the person who truly acts from duty and not just
in accordance with it will have an end in view, an end that this person
associates with actions that are done in accordance with the law. I
think that Kant's view should be understood in the following way.
The ends of practical reason form, as he says, a system: a totali-
ty composedof an unconditioned good and the conditional goods that
fall under it. The unconditioned good is humanity as an end in itself -
our capacity for rational choice, which whenfully realized is our
capacity for a good will. The conditional goods that fall under it
are the private ends that makeup the legitimate happiness of rational
creatures. These, together, makeup the Kingdomof Ends - the whole
of all rational beings as ends in thenselves as well as of the particu-
lar ends which each may set for himself. (G 51/433) The Kingdomof
Ends is a world for action, for the citizens of the Kingdomof Ends have
made humanity in their ownperson and that of others their end, and this
means that they have adopted the obligatory ends: they cultivate their
natural and moral perfections and, because they makean end of one
another's happiness, they jointly pursue a commongood. In the Kingdom
of Ends, our capacities for action are realized. All that is sought
is good, and all that is objectively good is sought. The Kingdomof
Ends is a practical ideal of reason. Whenwe join to it the idea of
the cooperation of nature in bringing about the correspondence of virtue
and happiness, we have the Highest Good- the shared ideal of practical
reason and a teleologically judging theoretical reason. Furthermore,
within the system that is the Kingdomof Ends, we can discern a basic
 254

framework- a form - that must be realized if the system is to exist

at all, as well as a more material componentthat can be realized to
a greater or lesser extent. The form is a juridical condition of the
world, only possible under the condition of perpetual peace. If the
rights of humanity are violated, the Kingdomof Ends does not exist.
The material component- the good that is realized in the Kingdomof
Ends - is greater or less according to the extent to which the obliga-
tory ends are adopted and realized by individuals and nature is cooper-
ative in securing happiness for the virtuous.
Nowwhenwe take up the standpoint of practical reason, and under-
take to choose our maximsas if we were choosing universal laws, we
take up a certain perspective. It is a perspective from which we see
the world in terms of the interests of humanity. Wecan describe these
interests in tenns of rational action, the setting and seeking of ends,
or even the free pursuit of happiness. Andfrom this perspective, it
is the system of the ideal ends of practical reason that emerges as the
end in view, the ideal final good. Even though this end is not the
motive of the rational action, it is a necessary feature of the perspec-
tive from which rational motivation is possible. And so a person who
has a truly virtuous attitude of will - a virtuous orientation of the
will - has these ideal ends of practical reason in view. The will
always sees itself as a cause - a cause of sanething - and the moral will
sees itself as a contributing cause of the Kingdomof Ends or Highest
Good. Thus, a person whodoes a juridical duty from duty does it with
an ideal end of practical reason in view.
The most important feature of the virtuous attitude is that the
person who has it has made humanity his or her end. In the juridical
 255

duty it is the rights of humanity that serve as the primary end. Be-
yond this, it seems as if Kant allows for a certain flexibility with
regard to the attitude of the will. Perhaps the end in view is the
entire final end - the Highest Good; perhaps one sees oneself as aiming
at the rights of humanity in this case, or as aiming at the universal
juridical condition of perpetual peace; or perhaps one is straining
for one's ownpurity and perfection of wi.11. Kant says that there are
various ends, and so various virtues:
Like anything formal, virtue considered as the will 1 s
firm resolution to conform with every duty is always one.
But in relation to the obligatory end of action or what one
ought to make one's end (the material element in the maxim),
there can be manyvirtues. And since obligation to the maxim
of such an end is called a duty of virtue, there are many
duties of virtue. (DV55/395)
The most important thing about this picture of acting from duty is
that the person whose action is from duty and so is purely rational is
not therefore seen as acting without an end. This person is not motiva-
ted by an inclination for this end, but by reason - and the end is an
end of reason. But the person has the end and acts with the end in
view. The Kantian ideal of good character is not a vision of a person
who is somehow
motivated by fonnal considerations of practical logic
instead of thinking about humanity and its rights and its happiness.
It is rather a vision of a person who, in taking up the standpoint of
practical reason, has made humanityand the interests of humanity, his
or her highest end.

IV. Moral Perfection and Happiness

A familiar worry about Kant's theory of rational and moral action
concerns the kind of place it leaves for the individual 1 s pursuit of his
 256

or her ownprivate ends, and his or her own private happiness. At

first it might seem as if this worry is completely misplaced, and
based on a misunderstanding. Kant's view is that the denands of
morality serve as conditions of the value of the pursuit of happiness,
not that happiness has no value. The Kingdomof Ends consists not only
of rational beings as ends in themselves but also of the private ends
that each sets for himself or herself. So long as your ends are pur-
sued without violating the rights of others, and so are pennissible,
they are good. Others even have a duty to help you pursue then, under
these conditions. Happiness is not excluded.
Nevertheless, readers of Kant have often supposed his view to be
that the virtuous person does not pursue her own happiness at all.
Christian Garve seems to have made this charge, for in "On the Common
Saying: 'This Maybe True in Theory, but it does not Apply in Prac-
tice"', Kant answered it with some impatience. The virtuous person
must, in doing her duty, not be motivated by considerations of her own
happiness, and must completely abstract from then. But she is not
expected to renounce the aim of happiness, and "where the question of
duty does not arise, and where there is no conflict with duty" the
pursuit of happiness is perfectly permissible. (Theory and Practice
64-66) This attitude shows up again in the Doctrine of Virtue, and in
Religion Within the Limits of Reason Alone. Since happiness under the
condition of virtue is a good thing, it is of course all right to
pursue one's own:
But since all other menwith the exception of myself would
not be all men, and the maximwould then not have the univer-
sality of a law, as it must have in order to be obligatory,
the law prescribing the duty of benevolence will include my-
self, as the object of benevolence, in the commandof practical
 257

reason. - Not that I am thereby obligated to love myself (for

this happens inevitably, apart from any command,and so there
is no obligation to it); it is rather that legislative reason,
which includes the whole species (and so myself with it) in
its Idea of humanity as such (not of men), includes me, when
it gives universal law, in the duty of mutual benevolence,
according to the principle that I am equal with all others
besides me, and permits you to be benevolent to yourself
under the condition of your being benevolent to every other
manas well. (DV118/451)
This is obvious enough - the duty of benevolence is based on the idea
that every humanbeing's ends are good, and there is no reason to ex-
clude your own. Indeed:
.•. a maximof promoting another 1 s happiness at the sacrifice
of my own happiness, my true needs, would contradict itself
were it madea universal law.• (DV53-54/393)
There can even be an indirect duty to promote one's ownhappiness, for:
Adversity, pain, and want are great temptations to trans-
gress one's duty. So it might seem that prosperity, strength,
health, and well-being in general, which check the influence
of these, could also be considered obligatory ends which make
up the duty of promoting one1 s own happiness, and not merely
the happiness of others. - But then the end is not the
agent's happiness but his morality, and happiness is merely
a means for removingobstacles to his morality - a pennissible
means, since no one has a right to demandthat I sacrifice
my ownends if these are not inmoral. To seek prosperity for
its own sake is no direct duty, but it can well be an indirect
duty: the duty of warding off poverty as a great temptation
to vice. But then it is not my happiness but the preserva-
tion of mymoral integrity that is my end and also my duty.
(DV47/388)
Indeed, this indirect duty forms the basis of one of the examples in
the First Section of the Foundations (G 15/399). In the Religion, Kant
says that all of the predispositions that together form the predisposi-
tion of the humanwill in general are toward good, because they all
enjoin the observance of the law. Our animal instincts, presumably, en-
courage us to care for ourselves and our offspring, and so to do our
duties to ourselves as animal and at the same time moral beings (DV83/
420); the predisposition to humanity spurs us on to the development of
 258

our natural perfections. (R 22-23) It is only because the predisposi-

tions to animality and humanity can be used 11contrary to their ends"
when self-love is made the condition of morality that we are even capa-
b1e of evi 1. 8
There is, however, still roomfor concern over this issue. The
difficulty remains because of something Kant says about the obligatory
ends, namely:
As the duty is wider, so man's obligation to action is more
imperfect; but the closer to narrow duty (Law) he brings the
maximof observing this duty (in his attitude of will), so
much the more perfect is his virtuous action. (DV49/390)
Onemight think, on the basis of this, that it is our duty to act from
duty as muchas possible. Weare enjoined to pursue our ownmoral per-
fection, and although Kant is usually careful to say, not that we must
do every action so far as possible from the motive of duty, but that we
must do every dutiful action so far as possible from the motive of duty,
it might still be thought that the perfection of our will is increased
as the general interests of humanity cometo take the place of our pri-
vate interests as the ends of our actions. Since it would be a contra-
diction for us to pursue the happiness of others to the exclusion of
our true needs {DV53-43/393), this does not lead to the problem(often
associated with utilitarianism) that moral philosophy makes it seem as
if there is no relief from relentless obligation, and no room for pri-
vate life. But it might be thought that it does lead to the view that
a person with a thoroughly virtuous character will devote time to her
own happiness only to the extent of maintaining the satisfaction of her
true needs, and give the rest of her life up to the pursuit of the
obligatory ends. So there is someof the same doubt whether it is not
a consequence of Kant's moral philosophy that most of us, innocently
 259

going about our daily pursuits and leading our private lives, are doing
something which, if not wrong, is certainly far less than right. In
pursuing our own happiness we may be doing something permissible; but
in pursuing the happiness of others we would be doi.ng something worthy.
In pursuing our own happiness we are doing something that reason allows,
· but it is not fully motivated by reason - it is motivated by inclina-
tion - and if we gave the time to the obligatory ends we could be doing
something rational instead. In the Foundations, Kant even speaks of
inclinations as something of which a rational being should prefer to
be altogether free. (G 46/428) It is true that later, in the Religion,
he goes back on this view:
Nowthe ground of this evil ... cannot be placed, as is so
commonlydone, in man s sensuous nature and the natural incli-
1

nations arising therefrom. For not only are these not directly
related to evil .•• we must not even be considered responsible
for their existence (we cannot be, for since they are implan-
ted in us we are not their authors). (R 30)
Natural inclinations, considered in themselves, are go?1'
that is, not a matter of reproach, arid 1t ,snot only fut, e
to want to extirpate them but to do so would also be harmful
and blameworthy. Rather, let them be tamed and instead of
clashing with one another they can be brought into a harmony
in a wholeness which is called happiness. (R 51)
But although inclinations and the happiness that is founded on them need
not be a matter of reproach, it is also not a matter of credit, as is
the pursuit of the obligatory ends.
NowI think it is correct that, on Kant1 s view, the virtuous person
would devote as muchtime as possible to the promotion of obligatory
ends, and attempt to limit the satisfaction of her wholly private incli-
nations to her true needs. But I do not think that this means we are to
picture the virtuous life as one in which private pursuits are constantly
being set aside or abstained from altogether to make room for the doing
 260

of 11good deeds", nor do I think that Kant intends to portray the vir-
tuous life as being necessarily, in an emotional sense, self-sacrifi-
cing. Weneed only remind ourselves of two things in order to dispel
the vision of an ascetic life that might seem to result from Kant's
views about virtue. One is the great scope of activities and projects
that can be plausibly construed to fall under the rubric of things
done for the sake of the obligatory ends. The other is _the important
fact that a person can find satisfaction and pleasure in the achieve-
ment of an end, so long as it has truly becomehis end, regardless of
the motives that led him to adopt the end and make it his own.
In regard to the first point, it need only be pointed out what a
wide range of everyday activities serve to promote the happiness of
someoneelse or to increase one's natural perfections. Speaking of the
width of duties, Kant says:
... a wide duty is not to be taken as a permission to make ex-
ceptions to the maximsof actions, but only as a pennission
to limit one maximof duty by another {e.g. love of one's
neighbor in general by love of one's parents) - a pennission
that actually widens the field for the practice of virtue.
{DV49/390)
One satisfies the duty of contributing to the happiness of others in
part simply in the course of the ordinary activities of friendship and
family life. So long as one's aim is genuinely to increase the happiness
of parent, child, spouse, or friend this is action taken towards the
obligatory end. Similarly, Kant speaks of the pursuit of natural per-
fections (a heading under which a greater portion of those of our
activities that are reasonably mentally or physically healthy can be
taken to fall) as being directed by the sort of life we choose to lead:
As to which of these natural perfections should take
precedence amongour ends and in what proportion to one
another we should make them our ends in keeping with our duty
 261

to ourselves, it remains for us to choose, according to our

rational deliberation about what sort of life we should like
to lead and whether we have the powers necessary for that
way of life (e.g. whether it should be manual labor, com-
merce, or scholarship). For, quite apart from the need to
support himself, which in itself cannot be the ground of
a duty, man_hasa duty to himself to be a useful memberof the
world, since this also belongs to the worth of humanity in
his own person, which he ought not to degrade. (DV112/
445-446)
So the ideal is not a person who gives up private life, but one who
allows the obligatory ends to detennine what form private life and every-
day activities will take. It is not a sacrifice of private life to re-
solve to do not merely what your inclination would lead you to, but some-
thing that is also useful, or helps to develop your mental and physical
powers, or makes someonebesides yourself happy, as often as possible.
Nor need it involve giving up your enjoyments. It does involve
giving up, as far as possible, the influence of your own expected enjoy-
ments on your decisions. Whenyou do your duty from duty, it will be
the thought of duty - in this case, the thought of the obligatory
nature of the ends - that movesyou. But there is no reason to suppose
that you can only enjoy the achievement of an end that is adopted for
the sake of the enjoyment it promises. This is the way we are inclined
to choose, since enjoyment is a thing that we seek. But it is not only
whenwe seek enjoyment and choose our ends for the sake of enjoyment
that we enjoy their achievement. Whensomething is really your end, you
will enjoy its realization regardless of whyyou adopted it as an end.
Whenyou allow your choice to be detennined by reason, you choose your
ends, not for their potential pleasantness, but for their goodness. But
this does not mean that the successful pursuit of these ends will not be
pleasant to you. It does mean that you may have to work at making these
good things very thoroughly and genuinely your ends before that pleasure
 262

will come:
Helping others to achieve their ends is a duty. If a man
practices it often and succeeds in realizing his purpose, he
eventually comes to feel love for those he has helped. Hence
the saying: you ou ~t to love your neighbor as yourself, does
1
not mean: you shou immediately (first) love him and (after-
wards) through the mediumof this love do good to him. It
means, rather: do good to your fellow-man, and this will
give rise to love of man in you (as an aptitude of the in-
clination to beneficence in general). (DV62-63/402)
By an 11aptitude of the inclination 11
, Kant does not mean that you come
to choose to do beneficent acts out of inclination, for then your choice
would not be worthy. He means rather that you may, with help from na-
ture and fortune, develop a structure of inclinations which allows you
to take pleasure in the achievement of this freely adopted end. The
end can continue to be your end because of its goodness, not because
of its pleasantness, even after it has becomepleasant to you. A range
of feelings, amongthem pleasure, follow from the adoption of ends, as
well as sometimespreceding them and motivating their adoption. And
appropriate feelings are, except in special circumstances of temperament
or fortune, the natural result of the adoption of ends. So although it
cannot be a direct duty to have a certain feeling, as Kant so often in-
sists, it can almost be an indirect one. For it can be a duty to have
an end, and the possession of the end will ordinarily bring the appropri-
ate feelings with it. It is this fact, I think, that enables Kant, in
spite of all that he says about howwe cannot be held responsible for
the temperamentand feelings that fonn the passive part of our nature,
to put forth for gratitude and sympathetic feeling as duties to others
in the Doctrine of Virtue. Sidgwick was surprised by this, and I am
certain other readers must be so as well:
In the case of Gratitude even the rigidity of Kant seems
to relax and to admit of an element of emotion as indispen-
 263

sable to the virtue. 9

But Kant's point is only that the emotional element will not be the
detenninant of choice, since emotion is something passive for which
we cannot be held responsible. Howeverinevitably the emotion followed
from the real adoption of the end, it would be the rational determination
of the will by the goodness of the end, not the feelings consequent on
the adoption of the end, for which we would be held responsible.
Here we find also the answer to make to those whoare put off by
Kant's remarks in the first set of examples - the examples of actions
that are clearly done from duty - in the Foundations. The point of
those examples is to emphasize that even in cases where the love of the
end cannot be present, a person can choose the end on the basis of rea-
son and pursue it with a real moral worth. This is especially obvious
in the case of the sorrowing manwho "tears himself out of a dead in-
sensibility" {G 14/398) to help others, though due to his temper and
situation he no longer can sympathize with or care for them. But this
is not to say that in the ordinary course of events the person who
chooses to help others on rational grounds will be, as such, one who
does sympathize with and care for them. The sympathywill follow from
the adoption of their happiness as an end, rather than the adoption of
the end following from the sympathy; but it will probably be there.
Kant's ideally virtuous person is likely to have the full complementof
emotions, concerns, and reactions that we expect of a virtuous person.
But he will have them because of his ends, rather than having his.ends
because of the emotions.
Nowif one asks, what is the aesthetic character, the
temperament, so to speak, of virtue, whether courageous and
hence joyous or fear-riddenand dejected, an answer is hardly
necessary. This latter slavish frame of mind can never occur
 264

without a hidden hatred of the law. Anda heart which is

happy in the performance of its duty {not merely complacent
in the recognition thereof) is a mark of genuineness in the
virtuous disposition - of genuineness even in piety, which
does not consist in the self-inflicted torment of a repentent
sinner ••• but rather in the firm resolve to do better in the
future. This resolve, then, encouraged by a good progress,
must needs beget a joyous frame of mind, without which man
is never certain of having really attained a love for the
good, i.e. of having incorporated 1t 1nto hTs maxim. (R 19n)
It seems from this passage that the joyous character Kant associates
with virtue combines two elements. One is the love of the good that
the virtuous person has achieved, which makes her happy in the perfor-
manceof her duty {and not, as Kant remarks, merely complacent about it).
The other is that sense of independence (firm resolve) that I mentioned
earlier 10 as part of the positive side of the moral incentive - the
sense that one's sensuous inclinations need not be a burden on one.
In a section of the Anthropologyentitled "Apologyfor Sensibility"
Kant remarks:
The ineradicable passive element in sensibility is really
the source of all the evil things said about it. Man's inner
perfection consists in his having control over the exercise
of all his powers, so that he can use them as he freely
chooses. (ANTH 24/144) - -
Indeed this can be seen as the main difference between the virtuous life
and the life lived for the sake of private happiness. The closer we
approach to perfect virtue, the more we choose our ends on the basis
of reason and for the sake of their objective goodness, the more truly
free we are. The seeker after happiness may find it by a shorter route,
but at the cost of letting nature assign to him his ends and his way of
life. The virtuous person, by contrast, exercises freedom in the choice
of his ends. 11 Having chosen them freely and for the sake of their
goodness by no means eliminates his chances of finding happiness in the
successful pursuit of his ends.
 265

V. Humanityas an End: The Duty to tell the Truth

In this last section, I want to discuss an issue notorious in
the literature on Kant's moral philosophy: his views on the strictness
of the duty to tell the truth. The two main expressions of these views
are someremarks found in the Doctrine of Virtue, and the article "On
a SupposedRight to Lie FromAltruistic Motives." In both places,
Kant puts forward the view that we must never under any circumstances
and for any result - howevergood and howeverurgent - tell a lie. He
also says that if we do tell a lie, no matter howgood our intention,
we are responsibile for any consequences that ensue from it, no matter
howunforeseeable. Nowwhile I do not think that what I am going to
say can entirely dispel the worrisome effect of Kant's views about this
issue, I do think it will put them in a somewhatdifferent light. I
also think that consideration of this issue can help us to understand
what Kant really means by "treating humanity as an end in itself".
Up until now I have focussed in this chapter on the Introduction
to the Doctrine of Virtue; it is nowimportant to look at the views
Kant puts forward in the main part of that work. There Kant spells out,
in detail, not only what actions will be taken by the person whomakes
humanity his or her end but also what attitudes will be adopted and even,
sometimes, what feelings will be experienced. Besides the Doctrine of
Method, which is concerned with howvirtue is to be taught and acquired,
the text is divided into accounts of duties to oneself and duties to
others. The duties to oneself are divided as follows: one has perfect
duties to oneself considered as an animal and at the same time as a moral
being; perfect duties to oneself considered simply as a moral being; and
imperfect duties to oneself which involve cultivating one's natural and
 266

moral perfections. Kant says that the perfect duties are concerned
with man's moral health and preservation, and the imperfect duties
.•• belong to his moral wealth... which consists in having
the power to realize all his ends, in so far as this can be
acquired; they belong to his cultivation of himself (as
active perfecting). (DV82/419)
The duties to others are divided differently: they are divided into
duties of love and duties of respect.
I have identified our "humanity"quite specifically with the power
to confer value upon an object through rational choice: to set an end,
and pursue it by rational means. All of the duties of virtue are
based upon the idea of placing a proper value on humanity in this sense,
and can be understood in terms of it. The duties to oneself as an
animal and a moral being, for instance, involve treating one's natural
person in such a way that its contribution to our capacity to set and
pursue ends is not diminished. It is the correllative of the imperfect
duty to cultivate natural perfections. Thus, for example, gluttony and
drunkenness are forbidden because
•.. as a result of overindulgence in food he is, for a certain
time, incapacitated for such actions as would require adroit-
ness (skill) and deliberation in the use of his powers. (DV
90/427)
With great delicacy Kant even rates the viciousness of these evils ac-
cordingly: gluttony is worse than drunkenness, since the latter is
at least stimulating to the imagination, whereas the former makes all
our faculties passive. Of the forms of drunkenness, spirits are worse
than wine, as wine is enlivening but spirits tend to make their users
withdrawnand silent. Indeed, wine at a proper-sized dinner party is
even a good thing, because it encourages a moral end: the exchange of
ideas. (DV91-92/428) That last is noteworthy: the exchange of ideas
 267

is a moral end because it contributes to the cultivation of all involved,

and because it places the diners in a proper relation to one another's
reason.
The duties to oneself as a moral being are duties not to lie, not
to be avaricious, not to be servile; to try to knowyourself clearly,
and judge yourself cleanly. While Kant makes somegeneral and rather
intuitive appeals to the idea that we must behave with the dignity
proper to a rational being and legislator in the Kingdomof Ends, there
is more than intuition at work in ascertaining what sort of behavior is
proper here. The guiding idea is that of making a proper use of one's
rational capacities; Kant thinks that in lying, avarice, servility,
and self-deception one is treating one's own rational capacities as
a meansand not at the same time as an end. A similar idea runs through
the account of the duties of respect to others: they involve taking
up a proper attitude towards the reason and (rational) character of
others. Weare to avoid arrogance, calumny, and mockery: thinking
ourselves better and wiser than others or thinking others worse and
sillier than we need to. The same goes for the duties of love to others,
which involve not only the practical beneficence of making the ends of
others your own, but attitudes of gratitude and sympathywhich show that
you fully understand the ways in which the ends of others are their ends.
Sympathyis a duty to 11partici pate actively in the fate of others" ( DV
126/457) - it shows that the fate of another is real to you, that the
ends of another are real ends for you. Gratitude is almost more a duty
of respect than of love, Kant tells us (DV127/458) for the ungrateful
person withholds from the benefactor a moral honor that the benefactor
has earned. Kant could have said more than this, for typically the un-
 268

grateful person tries to take a degrading view of the benefactor's

motive; and this is a clear case of a violation of the respect owing
to another, along the same lines as calumnyand mockery.
What holds all of these various duties together is the idea that
to treat humanity as an end is to treat each person's reason as reason.
While the most obvious consequence of this is that what ~nother chooses
as an end must be taken to be an end (the duty of pursuing the happiness
of others), the consequences are more extensive. To treat another with
respect in the Kantian sense is to treat her so far as possible under
the assumption that she is using her reason and using it well. This
is whycalumnyand mockerytake a special place as vices here: we owe
to others not only a practically generous attitude toward their plans
and projects, but a theoretically generous attitude towards their
thoughts and motives. And generous here does not mean charitable: it
means that we regard the best interpretation of another's words and
conduct as her due. If what she does seems bad or what she says seems
silly, we assume, as her right, that there was reason behind it, even
if misguided. Rather than calumnating her evil or mockingher foolish-
ness, we are to take it that either we or she has been misguided, and
honestly so. This is what it is to respect another as a rational equal:
it is always to suppose as far as possible that she is using her reason
and using it well. Everyoneknowshowenraging it is to have one's
views dismissed, say, as the product of prejudice or passion, one's
motives dismissed as mere guises for selfishness, and so on. It is this
treatment of others and this attitude towards another's use of reason
that Kant's duties of respect strictly forbid:
On this there is based a duty of respect for man even in
the logical use of his reason: a duty not to censure his error
 269

by calling it absurdity, poor judgment, and so forth, but

rather to suppose that his error must yet contain sane truth
and to seek this out, uncovering, at the same time, the
deceptive appearance (the subjective detennining ground of
judgment which, by an oversight, he took for the objective),
and so, by explaining to him the possibility of his having
erred, to preserve his respect for his ownreason. For if,
by such expressions, we deny all understanding to the manwho
opposes us in a certain opinion, howcan we make him under-
stand that he has erred in it? - The same thing applies to
the censure of vice, which must never break out into com-
plete contempt and denial of all moral worth to the immoral
man; for on this supposition he could never be improved, and
this is not consistent with the Idea of man, whoas such (as
a moral being} can never lose all his disposition to the good.
(DV133-134/463-464}
In dealing with our fellow rational beings, we are in this sense to
treat them as equals in the Kingdomof Ends, and not to forget that
Reasondepends upon this freedom for its very existence. For
reason has no dictatorial authority; its verdict is always
simply the agreement of free citizens, of whomeach one must
be pennitted to express, without let or hindrance, his
objections or even his veto. (Cl p. 593, A 739/8 767)
Whenwe are mockingor dismissive towards another's judgments or motives,
we in effect exclude him from full membershipin the Kingdomof Ends,
and thus fail to treat his humanity as an end in itself. To treat
another's humanity as an end in itself is therefore to treat the other's
reason as reason. An important part of what it means to treat someone's
reason as reason, however, is to address it as reason. This, it seems
to me, is the key to Kant's strictness about the duties of justice, and
especially, about lying.
WhenI speak of "addressing another's reason as reason" as a way
of treating that person as an end rather than a mere means, I meangiving
the person a fair chance to reason out, for himself, what he is to think
or to choose or to do. This idea already shows up in Kant's examples
of the Fonnula of the End in Itself, whenhe says that the other must
be able to share in the end of the very same action. Whenyou deceive
 270

someone,as in a false promise, the person you deceive does not get
to decide whether he will contribute to this end {your having the
money)or not. If you respect the person, in the sense of respecting
his reason, you will perhaps put the facts before him and just ask for
the money. Youtrust to his reason to reach a good conclusion, and of
course you take a risk, for he maydecide not to give it to you. But
you owe it to him as a rational being and fellow first cause to let him
determine whether this is a good end which he cares to promote or not.
Whenyou make the false promise, on the contrary, you are manipulating
him, and in particular you are manipulating his reason. Youmay even
say to yourself, "if he thinks he is going to get his moneyback, he
will surely give it to me, so I will get him to think that". His reason
to you is a sort of machine, and you knowwhat levers to pull to get
the desired results. Youtreat him as a machine and not as a rational
being with the right to decide for himself what to think and what to
do, and the right to his share in determining the destiny of things.
You exclude him from value-conferring and rational legislation.
This is why lying is for Kant a paradigm case of treating someone
as a mere means and not at the same time as an end. The use of language
is the case in which we seemmost clearly to be addressing one another
as rational beings, and addressing one another's reason. Languageis
what enables rational beings to form a communityin which each can play
a legislative role. Unless one is in some special context (singing a
song, telling a joke or a story), to address someoneis to imply that
you appeal to that person as a fellow rational being and a fellow citizen
in the communityof Reason. If you then say, not what you think is true,
not what is on your mind or in your heart, but what you think will pro-
 271

voke the right response, you degrade that person's very reason into
a tool for the achievement of your ownends. And this is exactly what,
according to the material fonnulation of the categorical imperative,
immorality is. So lying is a basic fonn of immoralconduct. Andof
course a pervasive form, too. For there may be no more everyday temp-
tation than the temptation ( even in our dealings with those cl o·sest to
us) to withhold something, or to embellish a story, or to place an
emphasis, so as to try to ensure that we get the reaction we want.
Wheneverwe do this, we manipulate. Andwheneverwe manipulate, unwill-
ing to take the risk of trusting to the other's reason, we fail to treat
the other as an end in himself or herself. Wehave, in however small
a way, violated the conditions of rational community,and failed to
live up to the ideal of the Kingdomof Ends. And we have, in however
small a way, been guilty of something that is the basic form of inmoral
conduct towards others - we have refused to treat another person as an
end, but made him or her into a means to our ownprivate ends.
Kant is so aware of our tendency to this sort of thing that he
identifies the skill of a man in having an influence on others so as
11

to use them for his ownpurposes11 as one of the primary meanings of the
w9rd prudent."
11
(see G 33n/416 and ANTH
183/322) The other meaning is
the more usual one: the ability to unite one's purposes into a system
for happiness. It should be obvious nowthat the moral law curbs
prudence in both of these senses. Respect for others demandsnot only
a generous view of another's reason and her right and capacity to use
it, but, as a consequence, an unflinching straightforwardness and
honesty.
Youact according to a maximadopted by the will, and whenyou will
 272

something you regard yourself as its cause. NowKant's view is that as

long as you do what is morally required of you, you are not responsible
for any unintended consequences. If doing what reason and morality
demanddo not have good results in this world, this is not your fault,
but rather an (apparent) flaw in nature. Youhave done your part to-
wards the achievement of the ideal, and it is up to other humanbeings
and to divine arrangements to do the rest. Youcannot do any good by
violating the rights of others, for this is to depart from the juridical
conditions outside of which no good can exist. If you do more or less
than is simply due, or required of you, then you are responsible. If
there are good consequences, you can claim somecredit for them; but
it there are bad consequences, you are answerable. Hence on Kant's
view we have a well marked sphere of responsibility; 12 it is possible
to ascertain in a practical way of what you are and are not to regard
yourself as the cause.
In connection with the discussions of lying, Kant startles his
readers by saying that if you tell a lie, however innocent, you are the
cause of whatever results from your having done so, no matter howinci-
dental to your intention that result is. This appears in both discus-
sions:
For example, a householder has ordered his servant to say
1
not at home' if a certain manasks for him. The servant
does this and, as a result, the caller slips away and commits
a serious crime, which would otherwise have been prevented by
the guard sent to arrest him. On whom(according to ethical
principles} does the blame fall in this case? On the servant,
surely, who violated a duty to himself by his lie, the results
of which his ownconscience imputes to him. (DV95-96/431)
This benevolent lie, however, can becomepunishable under
civil law through an accident (casus), and that which escapes
liability to punishmentonly by accident can also be condemned
as wrong even by external laws. For instance, if by telling
a lie you have prevented murder, you have madeyourself legally
 273

responsible for all the consequences; but if you have held

rigorously to the truth, public justice can lay no hand on
you, whatever the unforeseen consequences may be. After you
have honestly answered the murderer's question as to whether
his intended victim is at home, it may be that he has slipped
out so that he does not comein the way of the murderer,
and thus that the murder maynot be committed. But if you
had lied and said he was not at homewhen he had really gone
out without your knowingit, and if the murderer had then
met him as he went awayand murdered him, you might justly
be accused as the cause of his death. For if you had told
the truth as far as you knewit, perhaps the murderer might
have been apprehended by the neighbors while he searched the
house and thus the deed might have been prevented. There-
fore, whoever tells a lie, howeverwell-intentioned he might
be, must answer for the consequences, howeverunforeseeable
they were, and pay the penalty for them even in a civil
tribunal. (SRTL347-348/426-427)
It was only an accident (casus) that the truth of the state-
ment harmedthe occupant of the house; it was not a free act
(in a juristic sense). For to demandof another that he
should lie to one's ownadvantage would be a claim opposed
to all lawfulness. Each man has not only a right but even
the strict duty to be truthful in statements he cannot avoid
making, whether they hann himself or others. In so doing,
he does not do harm to him who suffers as a consequence;
accident causes this harm. For one is not at all free to
choose in such a case, since truthfulness (if he must speak)
is an unconditional duty. (SRTL349/428)
A possible way to understand Kant's view about this, based on what
I have said about the basis of the strictness of the duty to tell the
truth, would be this: Whenyou tell a lie, and thereby make a tool of
the reason of another, you also at the same time render yourself, from
your ownpoint of view, the one in control of the situation. By making
a machine of the other's reason, and pulling the levers to get the
result that you want, you make the consequencesyours. For you have,
from your ownpoint of view, madeyourself the 2.!!ll_ autonomousrational
will in the case, and so it is your will that must be regarded as the
cause of the results, howeverunexpectedly out of hand they may get. You
are, at least from your ownpoint of view, the only first cause in the
case. So long as you merely give another the information he demandsof
 274

you (assuming you have no choice but either to give it or to lie) you
have only done what is due, and it remains his action. It is partly
in the fact that our just actions may have, through these accidents,
bad results, that Kant finds the need of reason on which he bases his
doctrine of practical religious faith. Youmay hope that the laws of
nature are so arranged that these wrongs will be righted in the end.
But if you take control of the situation, and try to control it through
the instrumentality of the murderer's reason, you makeyourself at
least part cause of whatever happens.
I do not mean to suggest that this reading of these passages will
make them any more palatable to those who find them objectionable.
Somewill still find the doctrine intolerable; others may think that
although it may be true that whenwe make tools of others we must take
the consequences to ourselves, this is something that we ought to take
the risk of doing in cases of the sort Kant describes. I do not want
to settle that. What the reading does help to show is what sort of
thing Kant thinks is involved in the requirement of treating humanity
as an end in itself. Since the very basis and meaning of immorality is
that we treat others as mere means to our own private ends, instead of
giving them the respect due to free and equal rational beings, Kant
thinks this is to be avoided at all costs. Weowe it to the humanity in
ourselves and each and every other humanbeing that everyone be treated
as an end in itself;that we appeal to each person's reason freely and
frankly, granting them at any cost their inviolate right to legislation
in the Kingdomof Ends. Kant's rigorism is not to be understood as a
product of his trying to cling to logical principles in the face of all
moral insight and perception, but rather as the result of the high ideal
 275

of humanrelations that emerges from his moral philosophy. In order

to choose rationally, one must value humanity as an end in itself.
Valuing humanity as an end in itself commitsone always to treating
the reason of another as reason. To address another's reason straight-
forwardly; to grant to each as his or her due the right to participate
in detennining the destiny of things; to do this without angling for
the small and covert means of control that tempt us and without flinch-
ing at the risk we thereby certainly do take; to do all this with grace
and confidence, and, if we need it, faith in God: this is the ideal of
conduct we must try to live up to, if we are to be what Kant calls
"citizens of the world."
 276

NOTES

l. Mary J. Gregor, Lawsof Freedom, Chapter VIII. The distinction

between perfect and imperfect and narrow and wide is discussed in
general in the preceding chapter VII. There is also a useful
discussion of this issue in Onora Nell's Acting on Principle,
Chapter Four.
2. See the Note on Citations for an account of my treatment of the
use of these terms in quotations from Religion Within the Limits
of Reason Alone.
3. I will continue to use the English 11wil1 11 rather than "faculty of
desire" or "appetitive power" to describe the single thing of
which the Wille and the Willkur are two aspects.
4. This use of the term "character" is in 1i ne with the language of
the Critique of Practical Reason and the Religion, although some-
times, in other works, Kant will say that a person has either a
good character or no character at all. The reason for this, how-
ever, is easy to understand and settles any difficulty about the
terminology. One's character consists of the principles upon which
one acts, primarily the fundamental maxim. If one's fundamental
maximis bad, in a sense one has an evil character. But in another
sense only a rational principle can really be accepted as a maxim
by a rational being; a bad maxim,since it is not really rational,
is not really a principle. So a person who acts from a bad maxim
has no real principles, and therefore can be said to have no charac-
ter at all. This view is expressed in the Anthropology, where
Kant says,
For character requires maximsthat proceed from reason and
morally practical principles. So it is not correct to say
that the evil i'n a certain man is a quality of his character;
for in that case it would be diabolical. But man never
sanctions the evil in himself, and so there is really no evil
from princi"ples; it comesonly from abandoning principles.
(ANTH 158-159/293-294}
5. See Chapter Five, pp. 204-209.
6. Aristotle, I believe, comes the closest to trying to have a theory
about this: in the Ethics, the De Anima, and On the Movementof
Animals.
7. DavidHume,Treatise of HumanNature, p. 415.
 277

8. Kant's view in the Religion is startlingly close to that of Butler

in the first three Sermons. Butler also argues that although not
all of our natural motives and impulses have the good as their
object, they are so designed that they will all lead to the good
if governed by conscience. As in Kant, it is only because of the
possibility of misusing the impulses that are not directed toward
the good that we can go wrong.
9. Henry Sidgwick, The Methodsof Ethics, p. 223.
10. See Chapter Five, pp. 205-206.
11. I have followed one tendency in Kant in allowing it to sound here
as if there were two ways of choosing an end - according to predic-
ted pleasure or freely. This is, of course, oversimplified, as
even the discussion in "Conjectural Beginning of HumanHistory"
shows. I think that the nature of the adoption and possession of
an end is virtually unexplored territory. It seems to me that
psychology has not yet succeeded even in raising the question of
howwe get our ends and howwe are movedto act by them.
12. It is hard to identify specific passages in Kant where these ideas
about the sphere of responsibility are stated, other than the ones
I have given here. The clearest statement is perhaps that in the
Lectures on Ethics, pp. 59ff. It is a view of responsibility that
naturally accompanies a moral philosophy centered on the guidance
of rational ideals rather than the production of results. Living
up to the ideal marks out exact limits to your responsibility, a
way of doing just your share. Less is not allowed and more is not
required. On a utilitarian theory, by contrast, responsibility
can seem to extend as indefinitely as producible consequences do.
See the discussion by Bernard Williams in Utilitarianism For and
Against, especially pp. 93-118.
 Chapter 7
Conclusion: Autonomy

The third formulation of the categorical imperative might be called

the Formula of Autonomyin the Kingdomof Ends. No particular wording of
this formula seems to be singled out in the text as the official version,
but nere are a few:
By this principle all maximsare rejected which are not
consistent with the universal lawgiving of will. (G 49/431)
.•. the principle of every humanwill as a will giving
universal laws in all its maxims. (G 50/432)
Kant attaches an apology for not providing another set of discussions of
examples to the above formulation, so it has one claim to being a major
statement. Kant also uses the following formulations:

..• a categorical imperative .•• can only commandthat every-

thing be done from the maximof its will as one which could
have as its object only itself considered as giving univer-
sal laws. (G 50/432)
Act according to maximswhich can at the same time have them-
selves as universal laws of nature as their object. (G 55-56/
437)
These versions employ the notion of autonomyalone. I have called
this formula the Formula of Autonomyin the Kingdomof Ends for two rea-
sons. One is that, although some formulations employ the notion of auto-
nomywithout mention of the Kingdomof Ends, no formulation refers to the
Kingdomof Ends without mention of 11legislating 11
or use of the notion of
autonomy. It is important to rememberthat there are three arguments

278
 279
in the Second Section of the Foundations that have the form "if there
were a categorical imperative ... , 11 and that in his review of the formulae
Kant lists three. 1 The third is a Formulaof Autonomyin the Kingdomof
Ends, namely:
.•. all maximswhich stem from autonomouslegislation ought
to harmonize with a possible realm of ends as with a realm
of nature. (G 55/436)
Kant tells us that this formula provides 11a complete determination of all
maxims"and the "all-comprehensiveness or totality of the system of ends."
(G 55/436) Earlier he has explained that the Formula of Autonomyleads to
the concept of the Kingdomof Ends. (G 51/433) Other versions employing
the notion of the Kingdomof Ends are:
The rational being must regard himself always as legislative
in a realm of ends possible through the freedom of the will,
whether he belongs to it as memberor as sovereign. (G 52/434)
..• every rational being must act as if he, by his maxims,
were at all times a legislative memberin the universal realm
of ends. (G 57/438)
Act according to the maximsof a universally legislative
memberof a merely potential realm of ends.... (G 57/439)
The Formula of Autonomytells us that we are to act as if we were
legislating membersin the Kingdomof Ends. As legislators, we must
choose universal laws, and so the only difference between this and the
Formulaof Universal Lawis the emphasis on the fact that it is we who
make the laws, who give ourselves the laws. Kant explains that he
arrives at the Formulaof Autonomyby combining the other two formulae:
Objectively the ground of all practical legislation lies
(according to the first principle) in the rule and in the form
of universality, which makes it capable of being a law (at
most a natural law); subjectively, it lies in the end. But
the subject of all ends is every rational being as an end in
itself (by the second principle); from this there follows the
third practical principle of the will as the supreme condition
of its harmonywith pure practical reason, viz., the idea of
the will of every rational being as making universal law.
 280

By this principle all maximsare rejected which are not

consistent with the universal lawgiving of will. The will is
thus not only subject to the law but subject in such a way
that it must be regarded also as self-legislative and only for
this reason as being subject to the law (of which it can re-
gard itself as the author). (G 49/431)
The Fonnula of Autonomycombines the other two because it combines the
second key idea, that humanbeings as ends in themselves confer value on
ends through the choice of their maxims, with the first key idea, that
they must choose such maximsas admit of being universal laws. Arguments
for the identity of the first two formulae have already shown that if you
make humanity your end you must act only on such maximsas admit of being
universal laws. For if you act on a maximwhich cannot be universalized,
you claim for yourself and for your end a value higher than the value
that others can confer on their similar ends. That means that you are
failing to treat others as ends in themselves, for you are failing to
take their power of conferring value as a limiting condition of your own.
Whenyou do take other people's power of conferring value as a limiting
condition of your own, you choose only ends that could be the objects of
every rational will - objects that can be shared. You choose what can
be chosen within the Kingdomof Ends. The Formula of Autonomyin the
Kingdomof Ends makes these connections explicit: by making humanity
your end, you make it your _Q!'.!!llaw to act according to the Formula of
Universal Law. It is important to rememberthat the Formula of the End
in Itself was introduced as part of an investigation into the possibility
of reason determining conduct!. priori. (G 45/427)2 The introduction of
the notion of autonomy, which will be the key to the deduction of the
moral law, takes this investigation as far as it can go short of an actual
critical deduction.
So, after he has arrived at this formulation by the combination of
 281

the other two, Kant again gives us an argument of the form: if there
were a categorical imperative, this is what it would have to be like.
And this time it is a motivational argument, for its point is that if
there were a categorical imperative, it would have to be a law that we
give to ourselves, an autonomouslaw.
Thus the principle of every humanwill as a will giving
universal laws in all its maximsis very well adapted to being
a categorical imperative, provided it is otherwise correct.
Because of the idea of universal lawgiving, it is based on no
interest, and, thus of all possible imperatives, it alone can
be unconditional. Or, better, converting the proposition:
if there is a categorical imperative (a law for the will of
every rational being), it can only commandthat everything be
done from the maximof its will as one which could have as its
object only itself considered as giving universal laws. For
only in this case are the practical principle and the impera-
tive which the will obeys unconditional, because the will can
have no interest as its foundation. {G 50-51/432)
The argument for this appropriateness is a familiar one by this
point in the Foundations. Since obligation is a kind of rational neces-
sity, we must regard ourselves as subject to the law - as bound in some
way to follow it. But there are only two ways to be subject or bound to
a law: either your will is autonomousor it is heteronomous. If your
will is autonomous, the law on which it acts is your own law, and you
are subject to the law only because it is the law of your ownwill. If
your will is heteronomouswith respect to a law, you are bound to that
law by some external interest: that is, you are motivated to obey the
law by your interest in some promised reward or threatened punishment
that you associate with the law. In that case the law itself does not
interest you, for if you could achieve your end some other way, that
would be fine with you. This already shows that moral motivation cannot
be heteronomous, for in moral motivation we are directly interested in
the law itself. But in any case it is clear that a heteronomousrela-
tion of the will to a law implies that the will is acting on a
 282
hypothetical imperative: because of your interest in such and such an
end, you ought to do this. Hence if your will's relation to the law is
heteronomous, you are not acting on a categorical imperative. Since it
is a contradiction that we should be bound to the categorical imperative
by an interest (that would make it hypothetical), the only way we can be
bound to a categorical imperative is autonomously. The moral law must
be an autonomousprinciple of the will if we are bound to it at all,
which is to say if it exists at all.
If we now look back upon all previous attempts which have
ever been undertaken to discover the principle of morality,
it is not to be wonderedat that they all had to fail. Man
was seen to be bound to laws by his duty, but it was not seen
that he is only bound to act in accordance with his own, yet
universal, legislation, and he is only bound to act in accor-
dance with his ownwill, ••.. For if one thought of him as
subject only to a law (whatever it may be), this necessarily
implied some interest as a stimulus or compulsion to obedience
because the law did not arise from his ownwill. Rather, his
will was constrained by something else according to a law to
act in a certain way. By this strictly necessary consequence,
however, all the labor of finding a supreme principle for duty
was irrevocably lost, and one never arrived at duty but only
at the necessity of action from a certain interest. This might
be his own interest or that of another, but in either case the
imperative always had to be conditional and could not at all
serve as a moral command. (G 51/432-433)
Kant's deduction or validation of the moral law, both in the
Foundations and in the Critique of Practical Reason, depends essentially
on the claim that the moral law, as an autonomousprinciple which the
will makes and gives to itself, is the law of a free will: that "a free
will and a will under moral laws are identical." {G 65/447) This identi-
fication is the first step in the argument of the Third Section of the
Foundations, and it is obvious from the dispatch with which Kant makes
it that he thinks that the argument of the Second Section has thoroughly
prepared the way for it. Nowany deduction of the moral law will have
two parts: Kant must first show that a free will and a will under moral
 283

laws are identical, and he must then show us what reason we have to
believe that we have a free will and so really are under moral laws. In
the Third Section of the Foundations and in the section on the Deduction
in the Critique of Practical Reason, it is the second part of this task
that concerns Kant: the establishment (for practical purposes) of the
freedom of the will. Nowin one sense the moral law has not been fully
proved to be a law of practical reason, binding on every rational being,
until the entire deduction or validation is complete. Yet in another
sense the argument for the status of the categorical imperative as a law
of reason is completed once the identification of a free will and a will
under moral laws has been made. For the will is practical reason, and
"reason must regard itself as the author of its own principles 11
(G 67/
448): if a free will and a will under moral laws are really one, then
moral laws have been shownto be the laws of a fully practical reason.
At this point we do knowhowa fully rational being would behave: it is
only our own capacity for this kind of behavior that is still in ques-
tion. Since my concern has been with the establishment of the categori-
cal imperative as a law of pure practical reason, my task will be com-
pleted with an examination of this first step of the deduction. In what
follows, therefore, I am going to examine Kant's identification of a free
will and a will under moral laws, and to defend this identification
against an important objection.
Kant introduces his discussion this way:
As will is a kind of causality of living beings so far as
they are rational, freedom would be that property of this
causality by which it can be effective independently of
foreign causes determining it, just as natural necessity is
the property of the causality of all irrational beings by
which they are determined in their activity by the influence
of foreign causes. (G 64/445-446)
 284
This, however, is merely negative freedom - the independence of the will
from detennination by alien causes. The will in its activity - its choice
of maxims- is not directed or determined by any laws of nature. Wecan
easily imagine what it would be like for us to be determined in our choice
of maximsby some psychological law of nature. Suppose, to keep it sim-
ple, that there is such an ascertainable quality as the "strength of
desire", and it is a non-trivial psychological law of nature that we al-
ways seek the end which at any given momentwe most desire. {It isn't
really easy to imagine our behaving in this way, which would be rather
extraordinary, but it is easy to follow the example.) Nowif this were
so, nature would always be ready to dictate a certain maximof ends, and
we could calculate the means to get a (detennined) maximof action about
what to do to seek that end. But if the will is free there is no such
law of nature: all of our maxims, at every level, would be freely chosen.
Indeed, to have a free will is just to really have a will, for a will is
the capacity of acting according to one's conception of laws. (G 29/412)
But although nature does not determine the activity of our will, and does
not give a law to it, the will must have some principle and some law, for:
Since the concept of a causality entails that of laws accord-
ing to which something, i.e., the effect, must be established
through something else which we call cause, it follows that
freedom is by no means lawless even though it is not a pro-
perty of the will according to laws of nature. Rather, it
must be a causality according to immutable laws, but of a
peculiar kind. Otherwise a free will would be an absurdity.
{G65/446)
Either a free will has some principle in accordance with which it chooses
its maxims, or it chooses its maximsat random. But the second option is
absurd. The absurdity can be brought out in a numberof related ways.
Kant says the will is a causality, and a causality must have some law.
A thing that functions at randomis the very opposite of a causality.
 285

The will aims at an effect (every maximhas an end), but a thing that
operates randomly aims at nothing, by definition. One can also use a
Humeanargument here. If freedom of the will means that actions are
taken at randomor by chance, we could hardly appeal to it as a basis
for responsibility: indeed it is hard to see how the possessor of such
a will himself could knowvery far in advance what he was likely to do
next. 3 Finally - the point is always the same - one can make the same
sort of argument about the will that Leibniz made about God. The will
is supposed to be practical reason, but a will that acted at random
would have no reason for what it does. This is contrary to our views of
what a reason is and does.
What else, then, can the freedom of the will be but autonomy,
i.e., the property of the will to be a law to itself? The
proposition that the will is a law to itself in all its
actions, however, only expresses the principle that we should
act according to no other maximthan that which can also have
itself as a universal law for its object. And this is just
the formula of the categorical imperative and the principle
of morality. Therefore a free will and a will under moral
laws are identical. (G 65/446-447)
If a free will must have a law, but in order to be free must be seen as
choosing its own law or maximat every level, then one might say that
the only law it ever has any reason to choose is its own law. So if the
moral law is the law of autonomy, it will be the law of a free will. And
Kant thinks he has established that the moral law is the law of autonomy,
so he readily depends upon that here. That is all there is to this part
of the argument.
But it is possible, at a second glance, to feel deeply puzzled over
the readiness with which Kant takes himself to have established this
point. After all, the main point that the argument establishes is that
the free will must be an autonomouswill: it must give itself its own
law; that is howwe must think of it. But howcan it matter which law
 286

it gives itself? Even if we grant that a will acting under the autono-
mousmoral law is free in a way that a will determined by the laws of
nature is not, there is still another case: namely that of a will which
freely gives itself a law, but not the moral law. Such a will is autono-
mous in the sense that it directs its own activities, and makes its own
law, even though autonomyis not as it were the content of its law. Per-
haps a will could decide to give itself a law just like the imaginary
law of nature for an unfree will that I considered before: the law of
acting on its strongest desire. Since it is negatively free there is
nothing to prevent the free will from making this choice, or any other.
And then we can raise the question: in what sense would a will that
chose that law to govern its subsequent activities be any less free, or
any less autonomous, than a will that chose the moral law to govern its
activities. Both would be independent of the laws of nature, and each
would be autonomousin the bare sense of having chosen its own principle
of choice. This question becomeseven more important when we consider
that, in order that it be held responsible, an evil will must be regarded
as having chosen its ownlaw. So where does the difference lie?
This worry seems to have been on Sidgwick's mind when he criticized
Kant for depending on a conflation of two different notions of freedom.4
Sidgwick distinguished what he called "GoodFreedom"or "Rational Free-
dom"from what he called Neutral or Moral Freedom." Sometimes, we say
11

that a person acts more freely the more thoroughly he acts under the
guidance of reason - this is Goodor Rational Freedom. Sidgwick supposes
that we talk and think of the matter this way when we are thinking of the
conflict between reason and passion: when reason wins we are free and
in control, when passion wins we are unfree and mastered. Neutral or
Moral Freedom, by contrast, is the freedom to choose between good and
 287

evil. Sidgwick argues that the two notions are quite different:
... it is clear that if we say a man is a free agent in pro-
portion as he acts rationally, we cannot also say, in the
same sense of the term, that it is by his free choice that
he acts irrationally when he does so act. The two notions
of freedom must be admitted to be fundamentally different
in the two statements ...• 5
Sidgwick believes that Kant intends the notion of Neutral or Moral Free-
domwhen he is thinking about responsibility, and Goodor Moral Freedom
when he is urging the rationality of moral action. But he also thinks
that Kant simply conflates the two notions, and he cites the very pas-
sages I have just been discussing in evidence of that.
Similarly, in an earlier work, he explains that "since the
conception of causality involves that of laws... though free-
domis not a property of the will depending on physical laws,
yet it is not for that reason lawless; on the contrary, it
must be a causality according to inmutable laws, but of a
peculiar kind; otherwise, a free will would be a chimera
(Unding). And this immutable law of the "free" or autono-
11 11

mousu wi11 is, as he goes on to say, the fundamental pri nci-

ple of morality, "so that a free will and a will subject to
moral laws are one and the same.11
I have quoted this last phrase, not because it clearly
exhibits the notion of Rational Freedom- on the contrary,
it rather shows howeasily this notion may be confounded with
the other. A will subject to its ownmoral laws maymean a
will that, so far as free, conforms to these laws; but it
also may be conceived as capable of freely disobeying these
laws - exercising Neutral Freedom.6
Sidgwick rightly sees this as more than a verbal problem, because of
the motivational point of the identification of freedom and morality.
And I am afraid that most readers of Kant will feel the loss
to be serious; since nothing in Kant's ethical writing is more
fascinating than the idea - which he expresses repeatedly in
various forms - that a man realizes the aim of his true self
when he obeys the moral law, whereas, when he wrongly allows
his action to be determined by empirical or sensible stimuli,
he becomessubject to physical causation, to laws of a brute
outer world. But if we dismiss the identification of Freedom
and Rationality, and accept definitely and singly Kant's other
notion of Freedomas expressing the relation of the human
thing-in-itself to its phenomenon,I am afraid that this
spirit-stirring appeal to the sentiment of Liberty must be
dismissed as idle rhetoric. For the life of the saint must
be as muchsubject - in any particular portion of it - to the
 288

necessary laws of physical causation as the life of the scoun-

drel: and the scoundrel must exhibit and express his charac-
teristic self-hood in his transcendental choice of a bad life,
as mucha; the saint does in his transcendental choice of a
good one.
Since the transcendental choice of a character is the free will's
choice of a rule for the use of its ownfreedom, and the point of en-
visioning this choice as 11transcendental 11
is that it should be conceived
as a free choice, the problem here and the problem I raised about the
Foundations argument are obviously the same. In the Foundations argument,
what is in question is indeed what sort of law is taken by the free will
for its own principle, and it is the decision made about this that deter-
mines whether the character is good or evil. And it is true that Kant
seems to want to say both that only a will that chooses the moral law as
its ownmaximis free (in the sense of being autonomous)and that, since
character must be imputable whether it is good or evil, the maximmust
be regarded as chosen by the free will in either case. But if even an
evil will chooses its ownprinciple, how is it any less autonomousthan
a good one?
The issue can be recast in terms of Kant's later views about the
nature of the will, in The Doctrine of Virtue and Religion Within the
Limits of Reason Alone. I have already described these, but a reminder
will be helpful here. Kant distinguished the Wille - which is the will
as giving laws - from the Willkur, which is the will as active and
adopting maxims.
Lawsproceed from the will [Wille] - maximsfrom the power
of choice [Willkur]. In man the power of choice [Willkur] is
a power of free choice. The will [Wille], which does not look
to anything beyond the law itself, cannot be called either free
or unfree, since it does not look to actions but rather, in an
immediate way, to legislating for the maximsof actions (and
so to practical reason itself). Thus the will [Wille] func-
tions with absolute necessity and itself admits of no
 289
necessitation. It is therefore only the power of choice
[Willkur] that can be called free. {DV25/244; Wille and
Willkur my additions) --
The Willkur, being the free will described by Kant in the first part of
Section Three of the Foundations, must be envisioned as choosing its own
maxim. This maxim, the basis of the character, governs all of the deci-
sions and actions made by the person. Kant describes this maxim, in the
Religion, as a rule made by the Willkur for the use of its freedom {R 17).
In making this rule, the Willkur works within certain constraints: con-
straints which might, I think, be taken to define humannature. Human
beings are motivationally responsive to certain kinds of incentives: our
freedom, Kant tells us, consists in the fact that an incentive can de-
11

termine the Willkur to action .Q!!}l~ far~ the individual has incorpora-
ted it into his maxim(has made it the genera~ rule in accordance with
which he will conduct himself)" (R 19). Hence the Willkur's choice of a
highest maximis a choice of which amongthe available incentives will be
the basis of this maxim. For purposes of this argument the responsive-
ness of the humanwill in general to moral motives is taken for granted.
Part of the predisposition of the humanwill is a predisposition to per-
sonality which consists in the capacity for respect for the moral law as
in itself a sufficient incentive of the will. (R 21-23) Ultimately, as
Kantsees it, 8 the Willkur's choice comesdownto a choice between a moral
maxim(the categorical imperative itself as highest maxim)and a maximof
self-love. Nowthis is the choice that is in question in the issue that
has been raised. Kant's view, which Sidgwick has challenged, is that the
Willkur, in choosing a moral maxim, somehowaffirms, exercises, or
achieves its autonomy, whereas if it chooses a maximof self-love, it
does not. The difficulty, as I have described it, is that in either case
the Willkur has chosen its ownmaxim,and is to that extent autonomous.
 290

In order to resolve the difficulty, it is necessary to attend to some

important features of this original and originating choice.
First of all, it is important to see howthe two issues of the free-
domof action and the pure rationality of action come together in this
idea of the originating choice. I will take the idea of a purely ration-
al action to be this: it is action that can be fully justified, in the
sense that one has a satisfactory answer to the demandfor a reason no
matter howfar the question is pushed. Suppose that you wanted to fully
justify some everyday action. Whensomeonedemandsa reason for an
action, the appropriate answer is given by the maximof the action - the
maximembodies the reason. So you might say you wanted a certain end,
and you believed a certain means would bring it about. Behind your use
of this as a justification is your commitmentto the principle of hypo-
thetical imperatives: that whoeverwills an end wills the means. No-
tice that if what you were doing was attempting to explain the action
theoretically, you might also give the answer in tenns of the maxim, and
behind your use of the maximas an explanation would be a principle of
psychology closely.parallelling the principle of hypothetical imperatives,
something perhaps about the transfer of motivational impetus from an end
to the perceived means. If you are then asked to justify the adoption
of this maxim, it would not be enough to appeal to, say, your desire for
the end. For justification, like explanation, involves subsumption under
laws or principles. If your desire for the end were offered as a cause
of your acting on the maxim, this explanation would involve sane law of
nature concerning the motivating effect of desires on humanbeings (say,
that other things being equal we try to attain their objects, or some
such thing). If your desire is offered as a reason for your adoption of
the maxim, the justification commitsyou to a principle of some sort,
 291
to the effect that you take your desire for an object to be a reason for
trying to attain that object, in certain kinds of cases anyway. Nowif
this principle is not freely adopted, if it is just a fact about you that
you act to attain the objects of your desires and you could not do other-
wise, then the desire is not a reason, but only a cause, of your action.
Onecan imagine the persistent questioner demandinga justification for
even this principle: why try to attain the objects of your desires? Why
not just ignore them? Whynot seek sometreatment that would make them
go away? If it is not open to you to do this {ignore your desires, or
whatever) then you are not free, for your will is determined, and your
action is not fully rational, for you cannot justify or give a reason for
the ultimate principle on which you act. It is only a brute fact about
you that you act that way. Of course, eventually both justifications and
explanations must cometo an end, a point where there is no more to be
said. An explanation may end with a sort of brute fact, at howeverhigh
a level. ("That is just howelectrons behave.") A justification ends
with an ultimate reason, beyondwhich there is no further reason. But
even this ultimate reason, if it is a reason and not a mere determining
cause, must be freely adopted as a reason. Rational justification and
freedom go together: somethingcan only be a reason for your action if
you take it to be a reason by freely adopting a certain principle. In-
centives offer themselves as reasons, but if we are free and rational we
are motivated by them only in so far as we adopt them into our maxims,
or at least determine the order in which they are adopted. Nowthis is
what the highest maximdoes: it expresses what you ultimately take to
be a reason. Thus the problem about freedom of the will and the problem
about the full justification of action cometogether in the idea of the
choice of the highest maxim,and the problem of whether we can make sense
 292

of this choice. If the choice of the highest maximis a completely ran-

domand arbitrary choice, then action cannot be fully justified: there
is no reason for the principle on which it is ultimately based. Assum-
ing that we are not detennined, self-love and morality are at bottom both
arbitrary principles of choice, for they are themselves chosen arbitrar-
ily. At the same time, freedom of the will is, by the Foundations argu-
ment, an absurdity: it is ultimately lawless.
Nowat first sight it would seem as if the choice of the highest
maximmust be lawless. For it is a choice for which there can be neither
a reason nor a cause (no detennining ground). There cannot be a cause,
because it must be a free choice. There cannot be a reason, because
this choice is the ground of all reasons. Yet it must be chosen, for if
the highest maximwere just a brute fact about one's character, the
ground of all decisions and choices, one could not be said to be res-
ponsible for one's character at all. In that case the highest maxim
would be one's personal law of nature, a destiny, a complete individual
concept of the will. Wewould have what Kant calls "the freedom of a
turnspit, which when once woundup also carries out its motions of
itself." (C2 101/97) So because it is essential to Kant's view that we
be fully free and fully responsible, and not destined to act and choose
as we do, the highest maximmust be an object of a choice. Yet the
choice must be conceived to be madewith the purest spontaneity, for it
cannot be referred back to any still more fundamental maxim. Since this
maximwill determine what is to count as a reason for us, there can be -
it seems - no reason for choosing one way or another about this maxim.
Consequently, this spontaneous choice is inscrutable
11 11
: it cannot be
explained or understood.
Hence the source of evil cannot lie in an object determining
 293

the Willkur through inclination, nor yet in a natural impulse;

it can lie only in a rule madeby the Willkur for the use of
its freedom, that is, in a maxim. But now ,t must not be con-
sidered permissible to inquire into the subjective ground in
man of the adoption of this maximrather than its opposite.
If this ground itself were not ultimately a maxim, but a mere
natural impulse, it would be possible to trace the use of our
freedom to determination by natural causes; this, however, is
contrary to freedom. Whenwe say, then, Manis by nature good,
or Manis by nature evil, this means only that there is an
ultimate ground (inscrutable to us)* of the adoption of good
maximsor of evil maxims...• (R 17)
In the footnote to the passage Kant adds:
*That the ultimate subjective ground of the adoption of
moral maximsis inscrutable is indeed already evident from this,
that since this adoption is free, its ground (why, for example,
I have chosen an evil and not a good maxim)must not be sought
in any natural impulse, but always again in a maxim. Nowthis
maximalso must have its ground, and since apart from maxims
no determining ground of free Willkur can or ought to be ad-
duced, we are referred back endlessly in the series of subjec-
tive determining grounds, without ever being able to reach the
ultimate ground. (R 17-18)
As a free act, this inscrutable choice of the maximof character
cannot take place in time, for every event in time is causally deter-
mined. It is therefore conceived as an intelligible act, and as such is
placed squarely in the noumenalrealm:
In this existence nothing is antecedent to the determination
of his will; every action and, in general, every changing
determination of his existence according to the inner sense,
even the entire history of his existence as a sensuous being,
is seen in the consciousness of his intelligible existence
only as a consequence, not as a determining ground of his
causality as a noumenon. Fromthis point of view, a rational
being can rightly say of an unlawful action which he has done
that he could have left it undone, even if as an appearance it
was sufficiently determined in the past and thus far was ines-
capably necessary. For this action and everything in the past
which determined it belongs to a single phenomenonof his
character, which he himself creates, and according to which he
imputes to himself as a cause independent of all sensibility
the causality of that appearance. (C2 101/98-99)
To seek the temporal origin of free acts as such (as though
they were natural effects) is thus a contradiction. Hence it
is also a contradiction to seek a temporal origin of man's
moral character, so far as it is considered as contingent,
since character signifies the ground of the exercise of
 294

freedom; this ground {like the determining ground of free

Willkur generally), must be sought in purely rational repre-
sentations. {R 35)
Nowthis is important because the Willkur, as intelligible or noumenal
agency, is pure activity: there is nothing passive or receptive about
it, passivity or receptivity being characteristic of sensibility. This
means that the choice of the highest maximcannot be explained in terms
of the effect of strength of feelings, either moral or natural; and in
particular, the choice of self-love as one's highest principle cannot
ultimately be explained in terms of temptation. If we are susceptible
to temptation, even this must be imputable to us, and the grounds of the
susceptibility must lie in this choice. This, in fact, is what Religion
Within the Limits of Reason Alone is about.
For the fact is that in Religion Within the Limits of Reason Alone,
as in the Foundations argument, Kant takes it to be the choice of evil
that is in need of an explanation. That is, in the Religion as in the
Foundations Kant takes the moral law to be the law of a free will, and
so in a certain way the natural or obvious candidate to be the maximof
the free Willkur. Thoughit is emphasized that the choice is madewith
the most perfect freedom and spontaneity, that its ground is utterly
inscrutable, that neither cause nor reason can be offered as an account
of it, it is nevertheless clear that it is the choice of an evil maxim
that is the mystery for Kant. And this is because he thinks of the moral
law as being in a certain way already the law of a free Willkur; whereas
the adoption of the evil maximmaking self-love the condition of moral-
ity is an action that must be taken by the Willkur, a condition into
which a humanbeing must fall.
The trivial explanation of this would refer to the fact that our
predisposition is to good. That is, the incentives have a natural
 295

ordering, because the predisposition to personality is not merely a

susceptibility to be motivated by respect for the moral law as an incen-
tive, but a susceptibility to be motivated by the moral law as a suffi-
cient incentive of the will. The account of the moral incentive in the
Critique of Practical Reason shows us howthe moral incentive declares
its ownfitness to be the condition of all others: it functions by
striking downour self-conceit. Hence Kant says that:
Howit is possible for a naturally evil man to make himself
a good manwholly surpasses our comprehension; for howcan a
bad tree bring forth good fruit? But since, by our previous
acknowledgement,an originally good tree (good in predisposi-
tion) did bring forth evil fruit, and since the lapse from
good into evil (whenone remembersthat this originates in
freedom) is no more comprehensible than the re-ascent from
evil to good, the possibility of this last cannot be impugned.
For despite the fall, the injunction that we ought to become
better men resounds unabatedly in our souls .... (R 40)
The manwho has already made self-love his maximhas no reason to turn
away from self-love to morality. Yet the predisposition to personality
is such that moral incentives continue to make their demand- to be the
condition of self-love - perceived by us. But this explanation suggests
that it is a mere feature about our predisposition, a sort of fact about
transcendental psychology, that the moral law is the law of a free will.
The real explanation has to do with the nature of original choice.
Whenthe Willkur makes a law for the use of its ownfreedom, chooses
the law according to which it will makeall further choice, it is in a
special position. It is to make a choice, but because of its intelligi-
ble nature and purity of agency, there is nothing to determine that
choice, nothing even to tempt it one way or another. It is in a state
of absolute freedom. This is always true of a situation in which the
practical decision involves choosing the principles of choice themselves
rather than choosing objects in terms of already existing principles of
 296

choice. This is a position that I have talked about before. 9 It is the

standpoint of practical reason.
Nowthe explanation of the fact that a free will and a will under
moral laws is the same can be given in tenns of the content of the moral
law. The categorical imperative tells us we are always to select our
maximsas if they were to be universal laws of nature. To make the cate-
gorical imperative your highest maximis to resolve always to choose
your maximsas if you were makinguniversal laws of nature, and to make
their fitness to be universal laws of nature your motive for choosing
them. If you were choosing universal laws of nature, and you were not
choosing them for the sake of someend to which you were attracted, your
choice would be perfectly free and spontaneous: for there would be
nothing to determine your choice, and nothing even to tempt it one way
or the other. Thus, whenyou make the categorical imperative your maxim,
you resolve to choose each and every maxim~ if~ were making ~
original choice. The categorical imperative is the law of a free Willkur
because it adjures us always to choose our maximsas if we were in the
position of a free Willkur. The categorical imperative is the law of
practical reason because it is the maximthat the free Willkur chooses
from the standpoint of practical reason. And it is the maximthat the
free Willkur chooses from the standpoint of practical reason because its
content is the conunandalways to choose from the standpoint of practical
reason. A free will and a will under moral laws are the same. The cate-
gorical imperative is the principle of autonomybecause its entire con-
tent could be given in the formulation: always choose autonomously.
Hence the maximthat a free Willkur could be expected to take for the
11

use of its freedom" is the moral law, which in effect tells it to remain
a free Willkur. The moral law describes the situation of a free will and
 297

says to choose as if we were in that situation, with perfect freedom and

spontaneity. The law of self-love, by contrast, binds the Willkur over
to incentives and their objects. In making the law of self-love its
maxim, the Willkur gives up its position of perfect freedom for nothing,
for no reason. This is why, from a standpoint of practical reason, evil
is a state into which the Willkur must fall, and a complete and utter
mystery, a perversion of the humanwill.
In the Leibnizian Interpretation, I suggested that the point of the
Typic of Natural Lawis to place us in the position of Leibniz's God, a
position of world-choice. 1 °
Choosing from such a position makes your
choice rational, because there is nothing to guide your choice except
the principles of pure reason. It makesyour choice free, for your
choice is in no way constrained by the external circumstances. You
choose just as you would if you were choosing the world for yourself,
as if it were yours to create. Leibniz's God is at the standpoint of
practical reason, for a God has no inclinations to sway His choice:
In all this, manmay see himself as analogous to the divinity.
For while the divinity has no subjective need of any external
object, it cannot be conceived as closed-up within itself, but
only as compelled by the very awareness of its ownall-suffi-
ciency to produce the highest good outside itself. In the
case of the supreme being, this necessity (which corresponds
to duty in man) can be envisaged BY.us only as a moral need.
(Theory and Practice, 65n)
In the Formula of Autonomy,the freedom and the rationality of the choice
are seen to be one. The rationality of the choice lies in its freedom,
for the law of reason dictates that choices be made freely - with perfect
spontaneity. To put the same thing another way, spontaneity is choice
based on the principles of pure practical reason, and the principles of
pure practical reason are themselves concerned with the conditions for
spontaneous choice. The law that is dictated by the Wille is the law
 298

of the perfect functioning of the Willkur. In the Leibnizian Interpre-

tation, I argued that the function of a rational principle is to enable
a rational faculty to do what it does. 11 The will is rational agency,
and moral laws are aimed at rendering the world a perfect world for
rational action, a world in which the best conditions obtain for deter-
mining freely what your ends will be and pursuing them rationally. Now
we can say that the Willkur is the power of spontaneous choice, and the
law of the Wille is the law of the perfect functioning of ever-spon-
taneous choice.
Earlier in this discussion, I said that the problem of the freedom
of the choice of the highest maximand the problem of the full justifica-
tion and ultimate rationality of action would be solved together if it
could be shownthat the choice of morality as the highest maximwas not
arbitrary. As far as freedom is concerned, the choice is not arbitrary
because the choice is for freedom, for the preservation of the Willkur's
autonomy. The Willkur is already free and in the position into which the
categorical imperative places us; the choice of evil is from this pers-
pective an unintelligible downfall.
But the rational origin of this perversion of our Willkur
whereby it makes lower incentives supreme amongits maxims,
that is, of the propensity to evil, remains inscrutable to
us, because this propensity itself must be set downto our
account and because, as a result, that ultimate ground of
all maximswould in turn involve the adoption of an evil maxim
[as its basis]. Evil could have sprung only from the morally-
evil (not from mere limitations in our nature); and yet the
original predisposition is a predisposition to good; there is
then for us no conceivable ground from which the moral evil
in us could originally have come. {R 38)
So Kant, contrary to Sidgwick's impression, ends up by denying that the
notion of Moral or Neutral Freedomis a real notion of freedom:
Somehave tried to define freedom of choice as the power
to choose between the alternatives of acting with or against
 299

the law. But freedom of choice as phenomenongives frequent

examples of this in experience. For we knowfreedom (as it is
first made knowableto us through the moral law) only as a
neaative property in us: the property of not being necessita-
te to act by any sensuous determining ground. But we cannot
explain theoretically freedom in its positive aspect, as it
exercises necessitation on the sensuous power of choice - that
is, freedom as noumenon,as the power of man viewed merely as
an intelligent being. Wecan see only this: that while our
experience of man as a sensible being shows that he can choose
to act cwainst the law as well as in conformit¥ with it, his
freedom as an intelligible biiryg cannot be def,neal)y this,
since appearances cannot exp a,n a supersensible object (like
free choice); and that freedom cannot be located in the
rational agent's ability to choose what is opposed to his
(legislating) reason, even if experience proves often enough
that this happens (though we still cannot conceive the possi-
bility of it} •... Only freedom with regard to the inner
legislation of reason is really a power: the possibility of
deviating from le9islative reason is a lack of power.
(DV25-26/226-227)
Perhaps nothing brings out the difference between the theoretical
employmentof reason and the practical employmentof reason so vividly
as this fact: that while from the standpoint of theoretical reason
nothing is easier to understand than the fact that a humanbeing might
evade duty when it is in manifest conflict with her happiness, from the
standpoint of practical reason the sacrifice of her freedom for the
attainment of some object of inclination is wholly unintelligible. This
is one way to bring out why the attempt to give morality a theoretical
foundation cannot succeed. The theoretical and the practical standpoints
are different, and can seem to be at odds; the moral reality that is
obvious from the standpoint of practical reason, can be opaque to theory.
Perhaps it is not yet obvious what the solution to the difficulty
about the full justification and pure rationality of action is. Intro-
ducing this issue, I imagined a persistent questioner, who, in the search
for the unconditioned, demandedto knowwhy at every step in the justifi-
cation of an action, until the highest maximat last was reached. For a
 300
fully justified action there would always be a reason, a reason for the
adoption of every maximinvolved in the decision. Yet I said that there
was in one sense no reason for the choice of the highest maxim, since it
is by means of your legislation of the highest maximthat you determine
what is to count as a reason for you, what you will take for a reason.
Nowthis leads to the obvious question - howis this dilenma now resolved?
What is the justification of the highest maxim?
The answer comes from the Formulaof the End in Itself, and the argu-
ment for that formula. The justification of the choice of the moral law
itself as highest maxim, and of every choice that follows from that, is
this: that legislation is the prerogative of a rational being, and the
moral law is the law that a rational being as such legislates. The capa-
city for free legislation, for choosing maximsand so for determining
what will count as a reason for action and what will be deemedan objec-
tively good end, is its ownjustification. To value the humanity in your
ownperson and the humanity of others is to take that capacity as its own
justification; and it is to acknowledgethat a free will, wherever it is
found, has the prerogative of legislation.
Andwhat is it that justifies the morally good disposition
or virtue in making such lofty claims? It is nothing less than
the participation it affords the rational being in giving uni-
versal laws. He is thus fitted to be a memberin a possible
realm of ends to which his ownnature already destined him.
For, as an end in himself, he is destined to be legislative in
the realm of ends, free from all laws of nature and obedient
only to those which he himself gives. Accordingly his maxims
can belong to a universal legislation to which he is at the
same time also subject. A thing has no worth other than that
determined for it by the law. The legislation which determines
all worth must therefore have a dignity, i.e. unconditional and
incomparable worth. For the esteem, which a rational being
must have for it, only the word "respect'' supplies a suitable
expression. Autonomyis thus the basis of the dignity of both
humannature and every rational nature. (G 54/435-436)
Kant's view then is this: that there is only the value that we
 301

ourselves give our lives. (C3 284n) If there is anything at all of

value in the world, it is because there exists in the world a being who
is capable of valuing: a creator of values. The humanbeing, released
from the control of instinct, finds in practical reason and the freedom
that belongs to it a capacity that brings goodness and purpose into the
world: the capacity of making something one's purpose, of freely taking
it for one's good. This is humanity, the capacity to set an end. When
we make humanity itself our end - whenwe undertake to cherish this
capacity for valuing and always to acknowledgeits absolute right to a
legislative voice, we thereby render our actions and our choices objec-
tively good. Goodnessis not a property already belonging to certain
kinds of things in the world, which is to be ascertained and then sought
out. A thing is good because a free rational being who brings value into
the world chooses it as good. Goodand evil, right and wrong, are fea-
tures of the world that we bring into it through the manner of our choi-
ces and actions. Whenwe look at the world theoretically, we can find
no basis for objective judgments of moral value. The moral truth is
perspicuous to us only from the standpoint of practical reason.
 302

NOTES

1. See Chapter 1, pp. 25-26.

2. See Chapter 5, p. 190.
3. David Hume,Treatise of HumanNature, p. 411.
4. Henry Sidgwick, The Methodsof Ethics. Appendix: "The Kantian
Conception of Free Will."
5. Sidgwick, p. 511.
6. Sidgwick, pp. 514-515.
7. Sidgwick, p. 516.
8. See Chapter 6, pp. 234-236.
9. See Chapter 4, pp. 151-152.
10. See Chapter 4, p. 155ff.
11. See Chapter 4, pp. 167-169.
 303

BIBLIOGRAPHY

I. Worksby ImmanuelKant
AnthropologyFroma Pragmatic Point of View. (1798) Translated by
MaryJ. Gregor. The Hague: Martinus Nijhoff, 1974.
"Conjectural Beginning of HumanHistory" (1786) Translated by Emil
L. Fackenheim. In On History. Edited by Lewis White Beck.
Indianapolis: Bobbs-Merrill Library of Liberal Arts, 1963.
Critique of Judgment. (1790) Translated by J. H. Bernard. NewYork:
Hafner Library of Classics, 1951 rpt. 1974.
Critique of Practical Reason. (1788) Translated by Lewis White Beck.
Indianapolis: Bobbs-Merrill Library of Liberal Arts, 1956.
ImmanuelKant's Critique of Pure Reason. (1st edition 1781, 2nd
edition 1787) Translated by NormanKempSmith. NewYork: Macmillan,
St. Martin's Press, 1965.
The Doctrine of Virtue: Part II of the Metaphysics of Morals. (1797)
Translated by MaryJ. Gregor. NewYork: Harper Torchbooks, 1964.
Also translated as The Metaphysical Principles of Virtue by James
Ellington. Indianapolis: Bobbs-Merrill Library of Liberal Arts,
1964.
"The End of All Things" (1794) Translated by Robert E. Anchor. In On
History. Edited by Lewis White Beck. Indianapolis: Bobbs-Merrill -
Library of Liberal Arts, 1963.
Foundations of the Metaphysics of Morals. (1785) Translated by
Lewis White Beck. Indianapolis: Bobbs-Merrill Library of Liberal
Arts, 1959. Also translated as Groundworkof the Meta~hysic of
Morals by H.J. Paton. NewYork: Harper Torchbooks, 964.
Grundlegungzur Metaphysik der Sitten. Edited by Karl Vorlander.
Hamburg: Felix Meiner, Der Philosophischen Bibliothek Band 41, 1965.
"Idea for a Universal History Froma CosmopolitanPoint of View"
(1784) Translated by Lewis White Beck. In On History. Edited by
Lewis White Beck. Indianapolis: Bobbs-Merrill Library of Liberal
Arts, 1963.
Lectures on Ethics. (Lectures from 1775-1780, first published by
Paul Menzer in 1924) Translated by Louis Infield. London: Methuen
& Co., 1930. rpt. NewYork: Harper Torchbooks, 1963.
The Meta h sical Elements of Justice: Part I of the Meta h sics of
Morals. 797 Translated by John Ladd. Ind,anapo is: Bobs-
Merrill Library of Liberal Arts, 1965.
 304

"On a Supposed Right to Lie FromAltruistic Motives" {1797) in

ImmanuelKant: Critique of Practical Reason and Other Writings in
Moral Philosophy. Translated and edited by Lewis White Beck.
Chicago: University of Chicago Press, 1949. rpt. NewYork: Garland,
1976.
"On the ConmanSaying: 1 This Maybe True in Theory but it does not
Apply in Practice 111 {1793) translated by H. B. Nisbet in Kant's Poli-
tical Writin s. Edited by Hans Reiss. Cambridge: Cambridge Un1ver-
1
sity Press, 970.
"Perpetual Peace" {1795) Translated by Lewis l~hite Beck. In On
History. Edited by Lewis White Beck. Indianapolis: Bobbs-Merrill
Library of Liberal Arts, 1963.
Reliton Within the Limits of Reason Alone. (1793) Translated by
Theo ore M. Greene and Hoyt H. Hudson. La Salle, Illinois: Open
Court, 1934. rpt. NewYork: Harper Torchbooks, 1960.
"What is Enlightenment?" {1784) Translated by Lewis ~Jhite Beck.
In On History. Edited by Lewis White Beck. Indianapolis: Bobbs-
Merrill Library of Liberal Arts, 1963.

II. Other Works

Acton, H. B. Kant's Moral Philosophy. London: Macmillan, St.
Martin 1 s Press, 1970.
Anscombe,G.E.M. Intention. Basil Blackwell, 1957. 2nd edition:
Ithaca, NewYork: Cornell University Press, 1963.
Aristotle, The NicomacheanEthics. Translated by Sir David Ross.
Oxford: Oxford University Press, 1925. Revised by J. L. Ackrill and
J. O. Urmsonand reissued as a World's Classic Paperback, 1980.
Aune, Bruce. Kant1 s Theory of Morals. Princeton: Princeton
University Press, 1979.
Beck, Lewis White. A Conmentaryon Kant's Critique of Practical
Reason. Chicago: The University of Chicago Press, 1960; Phoenix
Books, 1963.
_______ . "Apodictic Imperatives" Kant-Studien Band 49,
1957-1958. rpt. in Kant: Foundations of the Metaphysics of Morals:
Text and Critical Essays. Edited by Robert Paul Wolff. Indianapolis:
Bobbs-Merrill Text and Co11D11entary
Series, 1969.
Bentham, Jeremy. An Introduction to the Principles of Morals and
Legislation. {1789) NewYork: Hafner Library of Classics, 1948
rpt. 1973.
 305

Bradley, F. H. "Duty for Duty's Sake Essay IV of Ethical Studies,

11
•

1876; 2nd edition, 1927. Oxford: Oxford University Press Paperbacks,

1962.
Broad, C. D. Five Types of Ethical Theory. NewYork: Routledge and
Kegan Paul, 1951.
Butler, Joseph. Five Sermons Preached at the Rolls Chatel. From
Fifteen Sermons Preached at the Rolls Chapel. London, 726.
Indianapolis: Bobbs-Merrill Library of Liberal Arts, 1956.
Dietrichson, Paul. "What Does Kant Meanby 'Acting from Duty?111
Kant-Studien Band 53, 1962. rpt. in Kant: A Collection of Critical
Essals. Edited by Robert Paul Wolff. Garden City, NewYork:
Doubeday Anchor Books, 1967. rpt. Notre Dame: University of Notre
DamePress.
. "Kant's Criteria of Universalizability" in Kant:
""'Fo_u_n_d_a_t,-·o_n_s_of-t-he
Metaphysics of Morals: Text and Critical Essays.
Edited by Robert Paul Wolff. Indianapolis: Bobbs-Merrill Text and
CommentarySeri es, 1969. This essay is based on Dietri chson' s "When
Is a MaximFully Universalizable?" Kant-Studien, Band 55, 1964.
Duncan, A.R.C. Practical Reason and Morality. London: ThomasNelson
and Sons, 1957.
Ebbinghaus, Julius. "Interpretation and Misinterpretation of the
Categorical Imperative" Translated by H. J. Paton. The Philosophical
Quarterl~, VolumeIV No. 15, 1954. rpt. in Kant: Foundations of the
Metaphysics of Morals: Text and Critical Essays. Edited by Robert
Paul Wolff. Indianapolis: Bobbs-Merrill Text and CommentarySeries,
1969.
Frankena, William K. "Obligation and Motivation in Recent Moral
Philosophy" in Essays in Moral Philosophy. Edited by A. I. Melden.
Seattle: University of Washington Press, 1958. rpt. in Perspectives
on Morality: Essays of William K. Frankena. Edited by Kenneth E.
Goodpaster. Notre Dame: University of Notre DamePress, 1976.
. "Sidgwick and the Dualism of Practical Reason"
=rh,....e____,..M-on.....,i,....s
.....
t-,...,,v-=-o
.....
1-um-e-58,No. 3, 1974. rpt. in Perspectives on Morality:
Essays of William K. Frankena. Edited by Kenneth E. Goodpaster.
Notre Dame: University of Notre DamePress, 1976.
Freud, Sigmund. Civilization and Its Discontents. (1930) Translated
by James Strachey. NewYork: W.W.Norton and Co. The Standard
Edition, 1961.
. Totem and Taboo. (1913) Translated by James
=s-=-tr_a_c..-he_y
___ N,,_e_w
York: W.W.Norton and Co. The Standard Edition, 1950.
Gregor, Mary J. Lawsof Freedom. Basil Blackwell, 1963. NewYork:
Barnes and Noble.
 306

Harrison, Jonathan. "Kant's Examples of the First Formulation of the

Categorical Imperative." The Philosophical Quarterly, Vol. VII No. 26,
1957. rpt. in Kant: Foundations of the Metaphysics of Morals: Text
and Critical Essays. Edited by Robert Paul Wolff. Indianapolis:
Bobbs-Merrill Text and CommentarySeries, 1969.
Hegel, G.W.F. Hegel's Phenomenologyof Spirit. (1807) Translated
by A. V. Miller. Oxford: Oxford University Press, 1977.
.,,,.......,.,..__,.,..--·Hegel's Philosophy of Right. (1821) Translated by
T. M. Knox. Oxford: Oxford University Press, 1952; Oxford University
Press Paperback, 1967.
Herman, Barbara. "Morality as Rationality: A Study of Kant's Ethics."
PhDThesis, Harvard University, 1976.
Hume, David. An Enguir¥ Concerning the Principles of Morals. (1751)
In Hume's Moral and Political Philosophy. Edited by Henry D. Aiken.
NewYork: Hafner Library of Classics, 1948 rpt. 1972.
.
.,,....,...,.__..,,,,..,,--- A Treatise of HumanNature. {1739) Edited by L.A .
Selby-Bigge. Oxford: Oxford University Press, 1888 rpt. 1975.
Kemp,J. "Kant's Examples of the Categorical Imperative" The
Philosophical Quarterly, Vol. 8 No. 30, 1958. rpt. in Kant:
Foundations of the Metaphysics of Morals: Text and Critical Essays.
Edited by Robert Paul Wolff. Indianapolis: Bobbs-Merrill Text and
CommentarySeries, 1969.
Leibniz, GJLF. Discourse on Metaphysics {1686) and Monadology.
(1714) In Leibniz: Discourse on Metaphysics/ Correspondence with
Arnauld/ Monadology. Translated by George R. Montgomery. La Salle:
Illinois: Open Court rpt. edition 1973.
Medlin, Brian. "Ultimate Principles and Ethical Egoism." Australasian
Journal of Philosophy XXXV,1957. rpt. in Morality and Rational Self-
Interest. Edited by David P. Gauthier. EnglewoodCliffs, NewJersey:
Prentice Hall, 1970.
Mill, John Stuart. On Liberty (1859) and Utilitarianism (1861).
In John Stuart Mill: Utilitarianism; On Liberty; Essay on Bentham;
Together with Selected Writings of Jeremy Bentham and John Austin.
Edited by MaryWarnock. NewYork: Meridian Books, 1962.
Moore, George Edward. Principia Ethica. Cambridge: Cambridge
University Press. 1st edition l903, 1st paper edition 1959 rpt. 1971.
Nagel, Thomas. The Possibility of Altruism. Oxford: Oxford
University Press, 1970.
Nell, Onora. Acting on Principle: An Essay on Kantian Ethics. New
York: Columbia University Press, 1975.
 307

Paton, H.J. The Categorical Imperative. London: Hutchinson &

Company,1947; rpt. Philadelphia: University of Pennsylvania Press,
1971.
Plato. Republic. Translated by Paul Shorey. In Plato: The Collected.
Dialogues. Edited by Edith Hamilton and Huntington Cairns. Princeton:
Princeton University Press, Bollingen Series, 1963.
Prichard, H. A. "Does Moral Philosophy Rest On a Mistake?" Mind 21,
1912. rpt. in Mill: Utilitarianism: Text and Critical Essays:- Edited
by Samuel Gorovitz. Indianapolis: Bobbs-Merrill Text and Commentary
Series, 1971.
Duty and Interest" from his Inaugural Lecture as
,..,.,.,_.,_..-,---,----.,..-•
11

White's Professor of Moral Philosophy, 1928. rpt. in Morality and

Rational Self-Interest. Edited by David P. Gauthier. Englewood
Cliffs, NewJersey: Prentice Hall, 1970.
Rawls, John. A Theory of Justice. Cambridge: Harvard University
Press, 1971.
_____ . "Justice as Fairness" Philosophical Review 67, 1958.
. "Kanthn Constructivism in Moral Theory: The Dewey
------,,.-,,..
Lectures 1980 Journal of Philosophy LXXVII,No. 9, 1980.
11

. "TwoConcepts of Rules" Philosophical Review 64,

.,...1"""95 -r-p...,..t-.in
...5.....Utilitarianism: Text and Critical Essays. Edited by
Samuel Gorovitz. Indianapolis: Bobbs-Merrill Text and Commentary
Seri es, 1971.
Ross, Sir David. Kant's Ethical Theory. Oxford: Oxford University
Press, 1954.
--- .........
--· The Right and The Good. Oxford: Oxford University
Press, 1930.
Sidgwick, Henry. The Methods of Ethics. 7th edition, 1907. rpt. New
York: Dover, 1966.
Singer, Marcus George. Generalization in Ethics. NewYork: Alfred A.
Knopf, 1961.
Smart, J.J.C. and Williams, Bernard. Utilitarianism For and Against.
Cambridge: CambridgeUniversity Press, 1973.
VonWright, Georg Henrik. The Varieties of Goodness. NewYork:
Routledge and Kegan Paul, 1963.
Wood,Allen. "Kant on False Promises" In Proceedings of the Third
International Kant Con§ress. Edited by Lewis White Beck. Dordrecht-
Holland: D. Reidel, 1 72.