<< . .

. 28
( : 33)

. . >>

which the maxim of action is to conform but the universality of a law

as such; and this conformity alone is what the imperative properly
represents as necessary.
There is, therefore, only a single categorical imperative, and it is this:
˜˜Act only in accordance with that maxim through which you can at the same time
will that it become a universal law.™™ (iv, 420“1)

If we rehearse what was said earlier about hypothetical imperatives, the
idea seems to be the following. A being endowed with a will is a being
endowed with a faculty of desire that acts not only in accordance with
laws, but in accordance with the representation of laws. Some of these
laws are prescriptive laws, or imperatives. In a hypothetical imperative,
the prescription (˜˜I ought to act™™ or ˜˜Act™™) is conditioned by an end to be
achieved (˜˜If I will X, I ought to Y™™). But a moral imperative is not of this
kind. It commands unconditionally (˜˜I ought to Y™™). Now the faculty of
desire, as the ˜˜subjective condition of life,™™ is always presented with
sensible motives and ends, which determine the antecedent (expressing
a goal to be achieved) and the consequent (expressing the means to
achieve that goal) of a hypothetical imperative. These are the default
conditions, as it were, in which the categorical imperative is called upon
to determine the faculty of desire, because there are, in fact, always ends
that determine what we think we ought to do. Therefore, if there is a
prescription proper to the categorical imperative, it can only be the
prescription that it be possible also to will unconditionally what is willed
under the condition of the particular ends expressed in hypothetical
imperatives. In other words, it can only be the prescription that we allow
ourselves only those ˜˜ought™™ clauses (prescriptions for action) that
remain in place once subjected to the test of the categorical imperative.
If I am correct in the interpretation I propose of Kant™s reasoning, the
interesting result is that the categorical imperative appears to intervene
only in second position as it were, or as a second-order principle: its role
is to evaluate the rules we already have, resulting from the hypothetical
premises expressing prudential and instrumental relations of ends and
means. Evaluating these rules under the categorical imperative can lead
either to prohibit, or to allow, the maxim governing one™s action (˜˜I
ought to Y™™) that follows from hypothetical imperatives. Understanding
the role of the categorical imperative in this way takes care of the objec-
tion of vacuity evoked above. The reason the categorical imperative
appears to be empty is that indeed on its own it provides neither specific
goals to achieve nor specific means to achieve them. These are provided
by the conditions under which human beings act and exercise their will

in determining their actions: the conditions of life. It is in relation to
these conditions that the categorical imperative acquires specific con-
tent, in determining which of the maxims resulting from the application
of hypothetical imperatives we ˜˜can™™ or ˜˜cannot™™ will to remain valid
once they are separated from what conditions them, namely from the
particular ends they serve.

˜˜Can also will™™ and moral judgment
So far I have freely used the terms ˜˜maxim™™ and ˜˜rule™™ to apply to any
self-prescribed rule of action. To better understand the role that the
categorical imperative plays in moral judgments (determining judg-
ments about what we ought to do), it will help to pause for a while and
examine Kant™s distinction between maxims, precepts, and (practical)
In Groundwork, Kant defines a maxim as ˜˜a subjective principle of
action,™™ that is, a rule that the rational agent assigns to herself based on
her particular empirical circumstances and inclinations. By contrast, a
practical law is an objective principle of action, that is, a rule that the
agent ascribes to herself by virtue of being a rational agent, indepen-
dently of any particular empirical circumstances or inclinations: a rule,
therefore, that applies to all rational beings strictly as such. Kant adds
that a maxim is a rule according to which a subject acts, but only a law is a
rule according to which he ought to act (handeln soll), that is, an impera-
tive (iv, 421n).
This addition is surprising. For surely the consequent of hypothetical
imperatives (the ˜˜then . . . ™™ clause in ˜˜If I will the end X, then I ought to
Y™™) expresses a manner in which I ought to act. True, the rule of action
thus expressed is binding only for the subject who wills the end X,
expressed in the antecedent of the hypothetical imperative. It thus
meets the first criterion for being ˜˜a maxim, not a law™™: it is ˜˜subjective,™™
not ˜˜objective,™™ in the sense explained above. But it does not meet the
second criterion (expressing only the way a subject acts, not the way she
ought to act), unless one wants to restrict the expression ˜˜ought™™ to the
moral ought. But this is not how Kant has been using the term, and this
would deprive the idea of hypothetical imperatives of all meaning.
Kant returns to the distinction between ˜˜maxim™™ and ˜˜law™™ in the very
first paragraph of the Critique of Practical Reason. There he begins by
defining practical principles as ˜˜propositions that contain a general
determination of the will on which depend several practical rules.™™

These rules are of two sorts. They are ˜˜subjective, or maxims, when the
condition is regarded by the subject as binding only for his will.™™ They
are ˜˜objective, or practical laws, when the condition is cognized as objec-
tive™™ (v, 19).15 Here the distinction between maxim and law is not a
distinction between ˜˜what we do™™ (maxims) and ˜˜what we ought to do™™
(law). It is a distinction, rather, between a rule that we assign to ourselves
(¼what we ought to do) under a condition particular to ourselves, and a
rule that is binding for all. In both cases, then, there is an ˜˜I ought,™™ a
prescription, but one is subjective (valid only for me), while the other is
objective (valid for all). Let us note, moreover, that strictly speaking the
maxims seem to be not the hypothetical imperatives themselves but
rather the consequent of these imperatives, which are binding only
under the condition that the antecedent be posited or asserted. A law,
by contrast, is valid ˜˜under an objective condition,™™ that is, a condition
valid for all.16 Both maxim and law thus express what we ought to do.
They differ merely in that the former is binding only under a particular
condition, and thus binding for some, not all subjects. The latter, in
contrast, is binding universally, for all. Only the first criterion of the
distinction (subjectivity of the maxim, versus objectivity of the law) seems
thus to be retained.
And yet, even in the Critique of Practical Reason, some of Kant™s indica-
tions seem to bring us back to the second criterion stated in Groundwork
for being a mere maxim (i.e. being an expression of ˜˜what we do™™ rather
than ˜˜what we should do™™). Thus, for example: ˜˜Someone can make it his
maxim to let no insult pass unavenged™™ (v, 19). There is no necessity for
such a maxim to present itself as an ˜˜I ought.™™ It may be only the
recognition of a rule I in fact follow, and will follow: ˜˜I will not tolerate
an insult without avenging myself.™™ Of course, to be followed, this rule
must be endorsed by the subject who follows it: it is relevant to a person,
in Harry Frankfurt™s sense of the term, an individual capable of

The text is ambiguous as to whether it is the ˜˜principles™™ or the ˜˜several practical rules™™
thought under them that are subjective or objective. This ambiguity does not create a
major difficulty, though. For the relevant contrast will be between (objective) law, which
will refer both to the categorical imperative and to the duties determined under it; and
(subjective) precepts, which will refer to (1) the general principle of hypothetical impera-
tives, (2) particular hypothetical imperatives, and even (3) the detached consequents of
those imperatives. More on this below.
This condition is the will itself, ˜˜the will as will™™ (v, 20). This point will be elucidated only in
section three of Groundwork, when Kant maintains that a free will and a will that is
determined under the representation of moral law are one and the same. Moral law is
the law of a free will. Cf. above, nn. 8, 12, 13.

second-order attitudes on its own desires.17 But it does not have the
strongly normative character of a formulation such as: ˜˜I should not
tolerate an insult without avenging myself.™™ Kant does seem to distin-
guish the maxim, understood in that weak sense, from both types of
imperative, hypothetical and categorical. He writes:
Imperatives, therefore, hold objectively and are quite distinct from
maxims, which are subjective principles. But the former either deter-
mine the conditions of the causality of a rational being as an efficient
cause merely with respect to the effect and its adequacy to it, or they
determine only the will, whether or not it is sufficient for the effect. The
first would be hypothetical imperatives and would contain mere pre-
cepts of skill; the second, on the contrary, would be categorical and
would alone be practical laws. Thus maxims are indeed principles but
not imperatives. But imperatives themselves, when they are conditional “
that is, when they do not determine the will simply as will, but only with
respect to a desired effect, that is, when they are hypothetical impera-
tives “ are indeed practical precepts but not laws. (v, 20)

What we now have, then, are not two terms (maxims and laws), but
four terms. The most general term is that of principle. There are three
kinds of principle: (1) laws (which are both objective and imperative),
(2) precepts (subjective and imperative), (3) maxims (subjective and non-
imperative). The precepts include what Kant had called, in Groundwork,
the rules of skill and the counsels of prudence, i.e. the hypothetical
imperatives, expressing an ˜˜I should™™ or ˜˜I ought to™™ (Ich soll) under
the condition of a previously determined end.18 The assertion of the
antecedent of a hypothetical imperative is itself a maxim, i.e. a rule that
expresses the manner in which our will is, in fact, determined, without

Cf. Harry Frankfurt, ˜˜Freedom of the will and the concept of a person,™™ in The Importance of
What We Care About (New York: Cambridge University Press, 1988), pp. 11“25.
Again, there is some uncertainty about what the precept is supposed to be: is it the whole
hypothetical imperative, or just the consequent that is asserted once the antecedent is
asserted? In a way, it would seem natural to say that the hypothetical imperative itself (˜˜if I
will, then I ought™™) holds universally, and thus objectively in Kant™s sense, precisely
because in it, the antecedent remains problematic: it would thus be a law, only the detached
consequent would be a precept. But this is not how Kant seems to use his distinction; he
does seem to include the hypothetical imperative itself among the precepts. This is
probably because what he has in mind is that the hypothetical imperative grounds the
assertion of a subjective ˜˜I ought™™. Note also that in the text cited, Kant mentions only rules
of skill as examples of precepts, which gives force to Wood™s claim that strictly speaking
there are no assertoric imperatives, only problematic ones (rules of skill). Against this view,
I would still maintain my own, for the reasons stated above (see n. 11).

this determination needing to take the form of an ˜˜I must™™ or ˜˜I should™™
or ˜˜I ought to™™ (Ich soll).19
These distinctions should now help us clarify the role of the cate-
gorical imperative in the transition from mere maxims and precepts to
practical laws. Consider Kant™s example, in the Critique of Practical
Reason, whether it is permissible to keep a deposit for myself:
I have, for example, made it my maxim to increase my wealth by every
safe means. Now I have a deposit in my hands, the owner of which has
died and left no record of it. This is, naturally, a case for my maxim. Now
I want only to know whether that maxim could also hold as a universal
practical law. I therefore apply the maxim to the present case and ask
whether it could indeed take the form of a law, and consequently
whether I could through my maxim at the same time give such a law as
this: that everyone may deny a deposit which no one can prove has been
made. I at once become aware that such a principle, as a law, would
annihilate itself since it would bring it about that there would be no
deposits at all. (v, 27)

The rule, ˜˜I will increase my wealth by every safe means,™™ is a maxim: a
rule that expresses the end that, in fact, I assign in general to my actions.
The case that presents itself (I have in my possession a deposit whose
existence is known to no one) falls clearly under the authority of this
maxim. Consequently, not only can I formulate the hypothetical impera-
tive: ˜˜If I want to increase my wealth by every safe means, then I ought to
deny the existence of a deposit which no one can prove has been made.™™
I can also posit its antecedent, and detach the consequent: ˜˜Now I want
to increase my wealth by every safe means. So, I ought to deny the
existence of the deposit, etc.™™ This conclusion is, in the vocabulary laid
out above, a precept: an imperative that holds only for me, under the
condition set by the end I have given myself: increase my wealth by all
available, safe means. There then arises the question: is it possible to will
this precept to be universally binding? Namely: suppose I take no
account of my particular end, can the precept be willed to bind (and
thus, to bind universally, unconditionally)? This would make it a prac-
tical law, in the sense stated above. We thus have here instances of the

Of course the antecedent might first itself have been derived as a precept, for instance: ˜˜If I
want to live comfortably, I ought to increase my wealth by all safe means.™™ Here, a rule that
could, as a mere maxim (˜˜I will increase my wealth by all safe means™™), function as the
antecedent in a hypothetical imperative (as in the example discussed below), is formulated
as a precept (˜˜I ought to™™), as the consequent of another hypothetical imperative.

three main types of principle distinguished above: the assertion of the
antecedent of the hypothetical imperative is a maxim; the assertion of
the consequent of the imperative, detached as the conclusion of the
inference, is a precept. A precept one took to be binding independently
of the condition formulated in the antecedent of a hypothetical impera-
tive, would thus bind universally and unconditionally: it would be a law.20
The test of the categorical imperative is applied to the precept (the
conclusion of the hypothetical syllogism). The test consists in asking
whether I could consent to the universalization of the precept. In the
case cited, the answer is negative: to universalize the precept ˜˜I must
deny the existence of the deposit which no one can prove has been
made,™™ would lead to canceling the very practice of deposits, and
thus my very action, which depends precisely on the existence of
that practice. Universalizing the detached consequent thus makes it
self-contradictory. By modus tollens, the antecedent itself must therefore
be denied. The modus tollens goes like this:
1 If I make it my maxim to (or, if I will) increase my wealth by all safe
means, then I ought to deny the existence of the deposit which no one
can prove has been made.
2 I ought not to deny the existence of the deposit which no one can
prove has been made (result of the test of universalization, about
which more below).
3 So, I do not make it my maxim to (or, I will not) increase my wealth by
all safe means.
We see, on this example, that the categorical imperative subjects to the
test of possible universalization first the particular rule of action, or
precept (˜˜I ought to deny the existence of the deposit which no one

It may seem strange to ask whether the precept can be universally binding, namely a law:
everyone should deny the deposit. After all, all we want to know is whether it is permissible
to deny the deposit, not whether one should deny it. But the test by universalization of the
initially ˜˜subjective™™ maxim is precisely a test of permissibility: the (subjective) maxim, or
precept, is permissible only if I could also will it as a universal, unconditional precept: a law.
This supposition is a mere fiction, but the fiction is what helps decide what I take to be
permissible. If one is surprised that such a strong test should be called upon to establish
something fairly mild “ only the permissibility, not even the requirement, of an action “
one should remember that the test is negative: if no contradiction emerges in the course of
the test, then the action is permissible, although of course not required of all. On the other
hand, if a contradiction does emerge, then the result is more forceful: the action is
forbidden, and its contradictory opposite morally required. On the test itself and the
problems it raises, see below.

can prove has been made™™). Then, by virtue of the first test, the cate-
gorical imperative also evaluates the maxim on which the precept
depended. The opportunity I receive, to fraudulently keep a deposit
which was entrusted to me, leads me to put into question not only the
precept resulting from the assertion of my maxim, but the maxim itself.
A rule which seemed uncontroversial (to increase my wealth by every
safe means) is challenged when the precept that it induces does not
withstand the test of the categorical imperative.
We also see that the test consists in separating the precept (here, ˜˜I
ought to deny the existence of the deposit which no one can prove has
been made™™) from the end expressed in the antecedent of the hypo-
thetical imperative (here, ˜˜increase my wealth by all safe means™™). The
precept, ˜˜I ought to keep the deposit™™ resulted analytically from willing
the end expressed in the antecedent. If the precept does not withstand
the test, then willing the end must be rejected as well.

Contradiction in conception, contradiction in the will
Kant writes that if we supposed the universalization of the precept that
recommends not returning the deposit I have the opportunity to keep
with impunity, ˜˜there would be no deposits at all.™™ This is one of the two
main kinds of impossibility outlined by the test of the categorical impera-
tive: universalizing the maxim (the precept) leads to contradicting the
very concept of the action prescribed by the detached consequent of the
hypothetical imperative. This first kind of contradiction is now com-
monly called ˜˜contradiction in conception™™ and is distinguished from a
second type, called ˜˜contradiction in the will.™™21 In the latter case, uni-
versalizing the maxim would make the relevant action impossible not by
virtue of some intrinsic contradiction, but because the will which would

Cf. iv, 424: ˜˜Some actions are so constituted that their maxim cannot even be thought
without contradiction as a universal law of nature, far less could one will that it should
become such. In the case of others that inner impossibility is indeed not to be found, but it
is still impossible to will that their maxim be raised to the universality of a law of nature
because such a will would contradict itself™™ (italics are Kant™s). The expressions ˜˜contra-
diction in conception™™ and ˜˜contradiction in the will™™ have been coined by Onora O™Neill:
cf. Constructions of Reason: Explorations of Kant™s Practical Philosophy (Cambridge: Cambridge
University Press, 1989), ch. 5. See also Christine Korsgaard, Creating the Kingdom of Ends
(Cambridge: Cambridge University Press, 1996), ch. 3. See also Christine Korsgaard™s
preface to Groundwork, p. xx.

want to universalize such a maxim would contradict itself. Such is the
case, Kant maintains, with the maxim: ˜˜I do not care to contribute to the
welfare of my fellow human beings™™ (iv, 423). Endorsing this maxim as a
universal law would put us in contradiction with a fundamental purpose
of all human beings, that of insuring their own welfare. Universalizing
the maxim that recommends not to make it one™s business to contribute
to the welfare of others does not lead to an intrinsic impossibility (contra-
diction). But it is impossible not to want one™s own welfare, which means
it is impossible to renounce in advance the help of others in case of
distress. Consequently, the maxim: ˜˜I do not care to contribute to the
welfare of others,™™ which it is impossible to will to be a universal law, is
rejected by the test of the categorical imperative.
Two interesting points emerge from this example. First, the test of
contradiction in the will takes into account the ends of empirical human
beings (here, well-being) and not only those of rational beings.22 But
second, the point of view on these ends required by the categorical
imperative is a point of view that is in some ways impersonal, or perhaps
more exactly, universally personal. When I consider adopting a maxim
or a precept, I am required to consider whether, regardless of my
particular goals, I could want this precept to become a universal law,
that is, to be adopted by all, as a categorical imperative. If I could, then
the precept is permissible (but not obligatory). If not, the precept is
forbidden, and the contrary precept (for example: I ought to contribute
to the well-being of others) is obligatory.
The first type of contradiction (contradiction in conception) deter-
mines ˜˜perfect duties,™™ i.e. duties that ˜˜allow no exception in favor of the
inclination.™™ The second type of contradiction determines ˜˜imperfect
duties,™™ i.e. duties for which a degree of latitude is left to the agent in
taking his inclinations into account in the determination of his actions.
This distinction is linked to another, that between ˜˜narrow™™ and ˜˜wide™™
or meritorious duties. Narrow duties are commands or interdictions
concerning specific actions (for example the interdiction to lie). Wide
or meritorious duties are those which prescribe an end without specify-
ing what actions must be undertaken to pursue this end. The duty of

The test by ˜˜contradiction in the will™™ also takes into account, of course, the ends of human
beings as rational beings. One of the examples Kant gives is that of testing the maxim ˜˜I will
not trouble myself with the cultivation of my talents.™™ This maxim is rejected because its
universalization puts it in contradiction with the proper end of a rational being, which is
that ˜˜all the capacities in him be developed, since they serve him and are given to him for all
sorts of possible purposes™™ (iv, 423).

benevolence toward others (caring for other people™s welfare) is a duty
of this type. This duty, determined under the test of contradiction in the
will, is thus a duty at once ˜˜imperfect™™ and ˜˜wide.™™23
One readily perceives the importance of the distinction between the
two kinds of contradiction, and the different kinds of duties they respect-
ively determine. The first kind of contradiction (contradiction in con-
ception, determining ˜˜perfect™™ and ˜˜narrow™™ duties), seems to bear the
whole weight of the so-called Kantian rigorism in matters of morality.
The second kind of contradiction (contradiction in the will, which deter-
mines the ˜˜imperfect™™ and ˜˜wide™™ duties), leaving room for inclination
and granting the agent a certain latitude in determining the action held
to be good, seems to reintroduce a prudential dimension (tied to the
search for happiness) into moral determination.
Now it is striking to note that the test by ˜˜contradiction in conception™™
seems especially relevant when applied to actions depending on human
convention.24 Thus for instance, it is clear that universalizing the maxim:
˜˜I will not return borrowed money™™ would put an end to (would be the
negation of) the existence of loan, that universalizing the maxim: ˜˜I will
make false promises™™ would be the negation of the practice of promise
itself, that universal appropriation of deposits would be the negation of
the very practice of deposit, and so on. All these actions are actions
depending on convention, and they are the object of juridical regulation.
Kant examines such actions in the first part of the Metaphysics of Morals
(the Doctrine of Right). It is clear that in this domain, all duties are both
perfect and narrow duties: they command obedience to an (external)
rule which determines with precision which actions are permissible or

For the distinction between ˜˜perfect™™ and ˜˜imperfect,™™ ˜˜narrow™™ and ˜˜wide™™ duties, see
Groundwork (iv, 422n) and Metaphysics of Morals (vi, 390). For the duty of benevolence as an
˜˜imperfect™™ and ˜˜wide™™ duty, see Groundwork (iv, 423“4); Metaphysics of Morals, x27 (vi,
450“1). The distinction between duties that are ˜˜narrow™™ and ˜˜perfect™™ on the one hand,
and duties that are ˜˜wide™™ and ˜˜imperfect™™ on the other hand, can also be characterized
under the second formulation of the categorical imperative: ˜˜So act that you use humanity,
whether in your own person or in the person of any other, always at the same time as an
end, never merely as a means™™ (iv, 429). The distinction is this time between the cases
where the adoption of a maxim or precept would contradict the idea that humanity is an
end in itself, and those where it would simply not contribute to the promotion of humanity
as an end in itself (see iv, 429“31). Kant maintains that the results from the tests under
these two formulas are the same. The scope of this chapter does not allow me to examine
this point, nor to consider the role played by the third formulation of the categorical
imperative, cited at the beginning of this chapter.
What Christine Korsgaard calls, after Rawls, ˜˜practices™™ (see Kingdom of Ends, p. 85; John
Rawls, ˜˜Two concepts of rules,™™ Philosophical Review, vol. 64 [1955], pp. 3“32).

<< . .

. 28
( : 33)

. . >>