Tải bản đầy đủ - 0trang
§2. Features of Ideal Moral Agents
sense. It’s a deep question (which I leave aside here) whether Kant’s conception of reason includes far more than reason.
It is useful, then, to use “reasonable” and “rational” as handy terms to
mark the distinction that Kant makes between the two forms of practical
reason, pure and empirical. Pure practical reason is expressed in the categorical imperative, empirical practical reason in the hypothetical imperative.
These forms of practical reason must also be distinguished from particular
categorical and hypothetical imperatives (as we shall see in a moment, the
particular maxims at step ) that satisfy the corresponding requirements
of practical reason in particular circumstances. The terms “reasonable” and
“rational” remind us of the fullness of Kant’s conception of practical reason
and of the two forms of reason it comprehends.
. Recall what we said last time in connection with Gr Pref: (–):
that Kant is concerned with the principles of a pure will, that is, with the
reasoning of fully reasonable and rational human agents. Such agents are
ideal: although affected by natural desires and inclinations, they are not
determined by them and always act as the principles of pure reason require.
They are also, let’s suppose, lucid and sincere in the sense that they know
(or can formulate) the reasons from which they act, and they can state these
reasons when appropriate.
Now, I take the CI-procedure to represent in procedural form all the
requirements of practical reason (both pure and empirical) as those requirements apply to our maxims. (This procedural representation is important
later when we discuss Kant’s moral constructivism.) In assessing the maxims
implicit in their actions, I suppose that ideal reasonable and rational agents,
who are also lucid and sincere, use these principles intuitively in their moral
thought and judgment.
Further, Kant takes for granted that the application of this procedure
presupposes a certain moral sensibility and a capacity for moral judgment
(MdS, Intro :–).2 Having a moral sensibility means, among other
things, knowing that wanting to make a deceitful promise raises a moral
question: Is such a promise permissible? Similarly, if great suffering tempts
us to want to take our life, that raises a moral question; so also do the
2. On this presupposition, see the instructive discussion by Barbara Herman, “The Practice of
Moral Judgment,” Journal of Philosophy (August ), –; reprinted in The Practice of Moral
Judgment (Cambridge, Mass.: Harvard University Press, ), pp. –.
needs and misery of others who seek my help. By representing the requirements of practical reason, the CI-procedure articulates a mode of reﬂection
that Kant thinks can help us to gain clarity about such matters, given our
moral sensibility and capacity for judgment as developed, as it must be, in
growing up and living in society.
It is a serious misconception to think of the CI-procedure as an algorithm intended to yield, more or less mechanically, a correct judgment.
There is no such algorithm, and Kant knows this. It is equally a misconception to think of this procedure as a set of debating rules that can trap liars
and cheats, scoundrels and cynics, into exposing their hand. There are no
. To clarify the remark above about a sincere agent, let’s distinguish
three kinds of reasons: explanatory reasons, agents’ reasons, and grounding
Explanatory reasons are part of an explanation of why someone acted as
he did. The explanation is partly psychological—a belief-desire explanation,
say—but it may refer to repressed or unconscious motives, and to other
psychological elements that agents are unaware of and would not count
among their reasons.
Agents’ reasons are the reasons that agents count as their reasons and
that truthful and sincere agents would, when appropriate, acknowledge as
their reasons. When people are lucid before themselves, that is, when they
know what moves them and do not act under the promptings of repressed
or unconscious motives, then agents’ reasons tend to coincide, or else to
overlap, with explanatory reasons. Kant’s ideal reasonable and rational and
sincere agents are also lucid.
Grounding reasons are of two kinds: reasons of rationality and justifying
reasons. Reasons of rationality are those that show a decision or an action
(for an agent in certain circumstances) to be rational, or sensible, as the
case may be; while justifying reasons show a decision or an action to be
reasonable, right, or just, or whatever is appropriate.
In the case of ideal agents—those who are reasonable, rational, and
lucid as well as truthful and sincere—the three kinds of reasons tend to
coincide, or else to overlap, when the actions in question are fully intentional, i.e., undertaken in the light of deliberation and judgment.
§. The Four-Step CI-Procedure
. It is important to recognize that the moral law, the categorical imperative,
and the CI-procedure are three different things.
The moral law is an idea of reason. It speciﬁes a principle that applies
to all reasonable and rational beings (or reasonable beings for short)
whether or not they are, like us, ﬁnite beings with needs. It holds for God
and the angels, and for reasonable beings elsewhere in the universe (should
there be such), as well as for us.
The categorical imperative, as an imperative, is directed only to those
reasonable beings who, because they are ﬁnite beings with needs, experience the moral law as a constraint. As such beings, we experience the moral
law in that way, and so the categorical imperative speciﬁes how that law
is to apply to us (Gr II:– [–]).
For the categorical imperative to be applied to our situation, it must
be adapted to our circumstances in the order of nature. This adaptation is
made by the CI-procedure as it takes into account the normal conditions
of human life by means of the law of nature formulation (Gr II: ).
. With these remarks as a preface, I now set out the CI-procedure in
four steps.3 At the ﬁrst step, we have the agent’s maxim, which is, let’s
suppose, rational from the agent’s point of view: that is, the maxim is rational given the agent’s situation and the available alternatives, together with
the agent’s desires, abilities, and beliefs (taken to be rational in the circumstances). Kant speaks of a maxim as a subjective principle: it is a principle
from which the subject acts (Gr II:n. ). When the agent’s maxim is
rational from the agent’s point of view, as supposed here, it may be said
to be subjectively valid.
The maxim is also assumed to be sincere: that is, it reﬂects the agent’s
actual reasons for the intended action as the agent, presumed to be lucid,
3. Modulo a few minor variations, my account of the CI-procedure follows that of Onora (Nell)
O’Neill in her Acting on Principle (New York: Columbia University Press, ). See also Paul Dietrichson, “When Is a Maxim Universalizable?” Kantstudien (); and Thomas Pogge, “The Categorical
Imperative,” in Grundelgung zur Metaphysick der Sitten: Ein Kooperativer Kommentar, ed. Ofried Hoăffe
(Frankfurt: Vittorio Klosterman, ). I have followed Barbara Herman’s supposition in a number
of her papers that when we apply the CI-procedure, we are to assume that the agent’s maxim is
would truthfully describe them. The CI-procedure applies, then, to maxims
that lucid and rational agents have arrived at in view of what they regard
as the relevant features of their circumstances. We should add that the
procedure applies equally well to maxims that rational and sincere agents
might arrive at (but have not), given the normal circumstances of human
To conclude: the agent’s maxim at the ﬁrst step is both sincere and
rational. It is a particular hypothetical imperative (to be distinguished from
the hypothetical imperative); and since it uses the ﬁrst-person pronoun, let’s
say that it expresses the agent’s personal intention to act from the maxim.
It has this standard form:
() I am to do X in circumstances C in order to bring about Y
unless Z. (Here X is an action and Y is an end, a state of
Note that the maxim has an “in order to” clause and so refers to an end.
For Kant, all actions have ends (MdS, Intro :f.). The nature of the clause
is important in distinguishing between duties of justice and other kinds of
duties, but I leave this aside here.
. The second step generalizes the maxim of the ﬁrst step; the result is
what we may call a universal precept (not Kant’s terminology) that applies
to everyone. When this precept passes the test of the CI-procedure, it is a
practical law, an objective principle valid for every rational being (Gr II:
n. ). So we have:
() Everyone is to do X in circumstances C in order to bring
about Y unless Z.
At the third step we are to transform the universal precept at () into
a law of nature to obtain:
() Everyone always does X in circumstances C in order to bring
about Y, as if by a law of nature (as if such a law was implanted in us by natural instinct) (II: [–]).
The fourth step is the most complicated; it raises questions which we
cannot thoroughly discuss here. The intuitive idea is this: