A theory of Justice: Revised Edition

The Analysis of Knowledge and Valuation
(La Salle, Ill., Open Court Publishing Co., 1946). The
whole of sec. 13 of ch. 18 is relevant here. Lewis says: “Value to more than one person is to be
assessed as if their several experiences of value were included in that of a single person.” Page 550.
However, Lewis uses this idea to give an empirical account of social value; his theory of right is
neither utilitarian nor empirical. J. J. C. Smart, in reply to the idea that fairness is a constraint on
maximizing happiness, puts the point neatly when he asks: “if it is rational for me to choose the pain
of a visit to the dentist in order to prevent the pain of toothache, why is it not rational of me to choose
a pain for Jones, similar to that of my visit to the dentist, if that is the only way in which I can prevent
a pain, equal to that of my toothache, for Robinson?” 
An Outline of a System of Utilitarian Ethics,
p. 26. Another brief statement is in R. M. Hare, 
Freedom and Reason
(Oxford, The Clarendon Press,
1963), p. 123.
Among the classical writers the conflation of all desires into one system is not to my knowledge
clearly asserted. But it seems implicit in Edgeworth’s comparison between “mécanique celeste” and
“mécanique sociale” and in his idea that someday the latter may take its place with the former, both
being founded upon one maximum principle, “the supreme pinnacle of moral as of physical science.”
He says: “As the movements of each particle, constrained or loose, in a material cosmos are continu-
ally subordinated to one maximum sum-total of accumulated energy, so the movements of each soul,
whether selfishly isolated or linked sympathetically, may continually be realising the maximum
energy of pleasure, the Divine love of the universe.” 
Mathematical Psychics,
p. 12. Sidgwick is
always more restrained and there are only hints of the doctrine in 
The Methods of Ethics.
Thus at one
point he may be read to say that the notion of universal good is constructed from the goods of
different individuals in the same way as the good (on the whole) of a single individual is constructed
from the different goods that succeed one another in the temporal series of his conscious states
(p. 382). This interpretation is confirmed by his saying later: “If, then, when any one hypothetically
concentrates his attention on himself, Good is naturally and almost inevitably conceived to be
pleasure, we may reasonably conclude that the Good of any number of similar beings, whatever their
mutual relations may be, cannot be essentially different in quality.” Page 405. Sidgwick also believed
that the axiom of rational prudence is no less problematical than that of rational benevolence. We can
equally well ask why we should concern ourselves about our own future feelings as about the feelings
of other persons. Pages 418f. Presumably he thought the answer identical in each case: it is necessary
to achieve the greatest sum of satisfaction. These remarks seem to suggest the conflation view.
164
The Original Position

---

individual (with no aversion to risk) who tries to maximize his own
prospects, the classical doctrine is the ethic of perfect altruists. A surpris-
ing contrast indeed! By looking at these principles from the standpoint of
the original position, we see that a different complex of ideas underlies
them. Not only are they based upon contrary motivational assumptions,
but the notion of taking chances has a part in one view yet none in the
other. In the classical conception one chooses as if one will for certain
live through the experiences of each individual, seriatim as Lewis says,
and then sum up the result.
38
The idea of taking a chance on which person
one will turn out to be does not arise. Thus even if the concept of the
original position served no other purpose, it would be a useful analytic
device. Although the various principles of utility may often have similar
practical consequences, we can see that these conceptions derive from
markedly distinct assumptions.
There is, however, a peculiar feature of perfect altruism that deserves
mention. A perfect altruist can fulfill his desire only if someone else has
independent, or first-order, desires. To illustrate this fact, suppose that in
deciding what to do all vote to do what everyone else wants to do.
Obviously nothing gets settled; in fact, there is nothing to decide. For a
problem of justice to arise at least two persons must want to do something
other than whatever everyone else wants to do. It is impossible, then, to
assume that the parties are simply perfect altruists. They must have some
separate interests which may conflict. Justice as fairness models this
conflict by the assumption of mutual disinterest in the original position.
While this may prove to be an oversimplification, one can develop a
reasonably comprehensive conception of justice on this basis.
Some philosophers have accepted the utilitarian principle because they
believed that the idea of an impartial sympathetic spectator is the correct
interpretation of impartiality. Indeed, Hume thought that it offered the
only perspective from which moral judgments could be made coherent
and brought into line. Now moral judgments should be impartial; but
there is another way to achieve this. An impartial judgment, we can say, is
one rendered in accordance with the principles which would be chosen in
the original position. An impartial person is one whose situation and
character enable him to judge in accordance with these principles without
bias or prejudice. Instead of defining impartiality from the standpoint of a
sympathetic observer, we define impartiality from the standpoint of the
litigants themselves. It is they who must choose their conception of jus-
38. See