A theory of Justice: Revised Edition

particular our aims are. This simply reflects the fact that no such contin-
gencies appear as premises in their derivation.
We may note also that the motivational assumption of mutual disinter-
est parallels Kant’s notion of autonomy, and gives another reason for this
condition. So far this assumption has been used to characterize the cir-
cumstances of justice and to provide a clear conception to guide the
reasoning of the parties. We have also seen that the concept of benevo-
lence, being a second-order notion, would not work out well. Now we can
add that the assumption of mutual disinterest is to allow for freedom in
the choice of a system of final ends.
30
Liberty in adopting a conception of
the good is limited only by principles that are deduced from a doctrine
which imposes no prior constraints on these conceptions. Presuming mu-
tual disinterest in the original position carries out this idea. We postulate
that the parties have opposing claims in a suitably general sense. If their
ends were restricted in some specific way, this would appear at the outset
as an arbitrary restriction on freedom. Moreover, if the parties were
conceived as altruists, or as pursuing certain kinds of pleasures, then the
principles chosen would apply, as far as the argument would have shown,
only to persons whose freedom was restricted to choices compatible with
altruism or hedonism. As the argument now runs, the principles of justice
cover all persons with rational plans of life, whatever their content, and
these principles represent the appropriate restrictions on freedom. Thus it
30. For this point I am indebted to Charles Fried.
223
40. The Kantian Interpretation

---

is possible to say that the constraints on conceptions of the good are the
result of an interpretation of the contractual situation that puts no prior
limitations on what men may desire. There are a variety of reasons, then,
for the motivational premise of mutual disinterest. This premise is not
only a matter of realism about the circumstances of justice or a way to
make the theory manageable. It also connects up with the Kantian idea of
autonomy.
There is, however, a difficulty that should be clarified. It is well ex-
pressed by Sidgwick.
31
He remarks that nothing in Kant’s ethics is more
striking than the idea that a man realizes his true self when he acts from
the moral law, whereas if he permits his actions to be determined by
sensuous desires or contingent aims, he becomes subject to the law of
nature. Yet in Sidgwick’s opinion this idea comes to naught. It seems to
him that on Kant’s view the lives of the saint and the scoundrel are
equally the outcome of a free choice (on the part of the noumenal self)
and equally the subject of causal laws (as a phenomenal self). Kant never
explains why the scoundrel does not express in a bad life his charac-
teristic and freely chosen selfhood in the same way that a saint expresses
his characteristic and freely chosen selfhood in a good one. Sidgwick’s
objection is decisive, I think, as long as one assumes, as Kant’s exposition
may seem to allow, both that the noumenal self can choose any consistent
set of principles and that acting from such principles, whatever they are,
is sufficient to express one’s choice as that of a free and equal rational
being. Kant’s reply must be that though acting on any consistent set of
principles could be the outcome of a decision on the part of the noumenal
self, not all such action by the phenomenal self expresses this decision as
that of a free and equal rational being. Thus if a person realizes his true
self by expressing it in his actions, and if he desires above all else to
realize this self, then he will choose to act from principles that manifest
his nature as a free and equal rational being. The missing part of the
argument concerns the concept of expression. Kant did not show that
acting from the moral law expresses our nature in identifiable ways that
acting from contrary principles does not.
This defect is made good, I believe, by the conception of the original
position. The essential point is that we need an argument showing which
principles, if any, free and equal rational persons would choose and these
principles must be applicable in practice. A definite answer to this ques-
31. See