not merely because we cannot intuit the perfection of the divine will, having rather
to derive it only from our own concepts of which morality itself is foremost, but also
because if we do not so derive it (and to do so would involve a most flagrant circle in
explanation), the only remaining concept of the divine will is made up of the attrib-
utes of desire for glory and dominion combined with the awful conceptions of might
and vengeance, and any system of ethics based on them would be directly opposed to
morality.
But if I had to choose between the concept of the moral sense and that of perfection
in general (neither of which at any rate weakens morality, though neither is capable of
serving as its foundation), I would decide for the latter, because it preserves the indefinite
Idea of a will good in itself free from corruption until it can be more narrowly defined. It
at least withdraws the decision on the question from the realm of sensibility and brings it
to the court of pure reason, although it does not even there decide the question.
For the rest, I think I may be excused from a lengthy refutation of all these doc-
trines. It is so easy, and presumably so well understood even by those whose office
requires them to decide for one of these theories (since their students would not toler-
ate suspension of judgment), that such a refutation would be superfluous. What inter-
ests us more, however, is to know that all these principles set up nothing other than
heteronomy of the will as the first ground of morality, and thus they necessarily miss
their goal.
In every case in which the object of the will must be assumed as prescribing the
rule which is to determine the will, the rule is nothing else than heteronomy. The
imperative in this case is conditional, stating that if or because one wills such and such
an object, one ought to act thus or so. Therefore the imperative can never command
morally, that is, categorically. The object may determine the will by means of inclina-
tion, as in the principle of one’s own happiness, or by means of reason directed to
objects of our possible volition in general, as in the principle of perfection; but the will
in these cases never determines itself directly by the conception of the action itself but
only by the incentive which the foreseen result of the action incites in the will—that is:
I ought to do something because I will something else. And here still another law must
be assumed in me as the basis for this imperative; it would be a law by which I would
necessarily will that other thing; but this law would in its turn require an imperative to
restrict this maxim. Since the conception of a result to be obtained by one’s own
powers incites in the will an impulse which depends upon the natural characteristic of
the subject, either of his sensibility (inclination and taste) or understanding and reason;
and since these faculties according to the particular constitution of their nature find
satisfaction in exercising themselves on the result of the voluntary action, it follows
that it would really be nature which would give the law [to the action]. This law, as a
law of nature, would have to be known and proved by experience, and as in itself
contingent it would be unfit to be an apodictical practical rule such as the moral rule
must be. Such a law always represents heteronomy of the will: the will does not give
itself the law, but an external impulse gives the law to the will according to nature of
the subject which is susceptible to receive it.
The absolutely good will, the principle of which must be a categorical imperative,
is thus undetermined with reference to any object. It contains only the form of volition
in general, and this form is autonomy. That is, the capability of the maxims of every
good will to make themselves universal laws is itself the sole law which the will of
every rational being imposes on himself, and it does not need to support this by any
incentive or interest.
FOUNDATIONS OF THEMETAPHYSICS OFMORALS 883
444