A theory of Justice: Revised Edition

Doing and Deserving
(Princeton, Princeton University
Press, 1970), pp. 64f.
276
Distributive Shares

---

scheme of taxes and burdens designed to put a price on certain forms of
conduct and in this way to guide men’s conduct for mutual advantage. It
would be far better if the acts proscribed by penal statutes were never
done.
41
Thus a propensity to commit such acts is a mark of bad character,
and in a just society legal punishments will only fall upon those who
display these faults.
It is clear that the distribution of economic and social advantages is
entirely different. These arrangements are not the converse, so to speak,
of the criminal law, so that just as the one punishes certain offenses, the
other rewards moral worth.
42
The function of unequal distributive shares
is to cover the costs of training and education, to attract individuals to
places and associations where they are most needed from a social point of
view, and so on. Assuming that everyone accepts the propriety of self- or
group-interested motivation duly regulated by a sense of justice, each
decides to do those things that best accord with his aims. Variations in
wages and income and the perquisites of position are simply to influence
these choices so that the end result accords with efficiency and justice. In
a well-ordered society there would be no need for the penal law except
insofar as the assurance problem made it necessary. The question of
criminal justice belongs for the most part to partial compliance theory,
whereas the account of distributive shares belongs to strict compliance
theory and so to the consideration of the ideal scheme. To think of dis-
tributive and retributive justice as converses of one another is completely
misleading and suggests a different justification for distributive shares
than the one they in fact have.
49. COMPARISON WITH MIXED CONCEPTIONS
49. Comparison with Mixed Conceptions
While I have often compared the principles of justice with utilitarianism,
I have not yet said anything about the mixed conceptions. It will be
recalled that these are defined by substituting the standard of utility and
other criteria for the second principle of justice (§21). I must now con-
sider these alternatives, especially since many persons may find them
more reasonable than the principles of justice which seem at first anyway
to impose rather stringent requirements. But it needs to be emphasized
straightway that all the mixed conceptions accept the first principle, and
41. See H. L. A. Hart, 
The Concept of Law
(Oxford, The Clarendon Press, 1961), p. 39; and
Feinberg, 
Doing and Deserving,
ch. V.
42. On this point, see Feinberg, ibid., pp. 62, 69n.
277
49. Comparison with Mixed Conceptions

---

therefore recognize the primary place of the equal liberties. None of these
views is utilitarian, for even if the principle of utility is substituted for the
second principle, or for some part of it, say the difference principle, the
conception of utility still has a subordinate place. Thus insofar as one of
the chief aims of justice as fairness is to construct an alternative to the
classical utilitarian doctrine, this aim is achieved even if we finally accept
a mixed conception rather than the two principles of justice. Moreover,
given the importance of the first principle, it seems that the essential
feature of the contract theory is preserved in these alternatives.
Now it is evident from these remarks that mixed conceptions are much
more difficult to argue against than the principle of utility. Many writers
who seem to profess a variant of the utilitarian view, even if it is ex-
pressed vaguely as the balancing and harmonizing of social interests,
clearly presuppose a fixed constitutional system that guarantees the basic
freedoms to a certain minimum degree. Thus they actually hold some
mixed doctrine, and therefore the strong arguments from liberty cannot
be used as before. The main problem, then, is what can still be said in
favor of the second principle over that of utility when both are con-
strained by the principle of equal liberty. We need to examine the reasons
for rejecting the standard of utility even in this instance, although it is
clear that these reasons will not be as decisive as those for rejecting the
classical and average doctrines.
Consider first a mixed conception that is rather close to the principles
of justice: namely, the view arising when the principle of average utility
constrained by a certain social minimum is substituted for the difference
principle, everything else remaining unchanged. Now the difficulty here
is the same as that with intuitionist doctrines generally: how is the social
minimum to be selected and adjusted to changing circumstances? Anyone
using the two principles of justice might also appear to be striking a
balance between maximizing average utility and maintaining an appropri-
ate social minimum. If we attended only to his considered judgments and
not to his reason for these judgments, his appraisals might be indistin-
guishable from those of someone following this mixed conception. There
is, I assume, sufficient latitude in the determination of the level of the
social minimum under varying conditions to bring about this result. How
do we know, then, that a person who adopts this mixed view does not in
fact rely on the difference principle? To be sure, he is not conscious of
invoking it, and indeed he may even repudiate the suggestion that he does
so. But it turns out that the level assigned to the required minimum that
constrains the principle of average utility leads to precisely the same
278
Distributive Shares

---

consequences that would arise if he were in fact following this criterion.
Moreover, he is unable to explain why he chooses the minimum as he
does; the best he can say is that he makes the decision that seems most
reasonable to him. Now it is going too far to claim that such a person is
really using the difference principle, since his judgments may match
some other standard. Yet it is true that his conception of justice is still to
be identified. The leeway behind the scenes for the determination of the
proper minimum leaves the matter unsettled.
Similar things can be said concerning other mixed theories. Thus one
might decide to constrain the average principle by setting up some dis-
tributional requirement either by itself or in conjunction with some suit-
ably chosen minimum. For example, one might substitute for the differ-
ence principle the criterion to maximize the average utility less some
fraction (or multiple) of the standard deviation of the resulting distribu-
tion.
43
Since this deviation is smallest when everyone achieves the same
utility, this criterion indicates a greater concern for the less favored than
the average principle. Now the intuitionistic features of this view are also
clear, for we need to ask how the fraction (or multiple) of the standard
deviation is to be selected and how this parameter is to vary with the
average itself. Once again the difference principle may stand in the back-
ground. This sort of mixed view is on a par with other intuitionistic
conceptions that direct us to follow a plurality of ends. For it holds that
provided a certain floor is maintained, greater average well-being and a
more equal distribution are both desirable ends. One institution is unam-
biguously preferable to another if it is better on each count.
Different political views, however, balance these ends differently, and
we need criteria for determining their relative weights. The fact is that we
do not in general agree to very much when we acknowledge ends of this
kind. It must be recognized that a fairly detailed weighting of aims is
implicit in a reasonably complete conception of justice. In everyday life
we often content ourselves with enumerating common sense precepts and
objectives of policy, adding that on particular questions we have to bal-
ance them in the light of the general facts of the situation. While this is
sound practical advice, it does not express an articulated conception of
justice. One is being told in effect to exercise one’s judgment as best one
can within the framework of these ends as guidelines. Only policies
preferable on each score are clearly more desirable. By contrast, the
43. For a view of this kind, see Nicholas Rescher,