A theory of Justice: Revised Edition

parties depends on weighing various considerations. The reasoning is
informal and not a proof, and there is an appeal to intuition as the basis of
the theory of justice. Yet, as I have remarked (§21), when everything is
tallied up, it may be clear where the balance of reasons lies. If so, then to
the extent that the original position embodies reasonable conditions used
in the justification of principles in everyday life, the claim that one would
agree to the principles of justice is perfectly credible. Thus they can serve
as a conception of justice in the public acceptance of which persons can
recognize one another’s good faith.
It may be helpful at this point to list some of the main grounds in favor
of the two principles of justice over the principle of average utility. That
32. Thus while Brandt holds that a society’s moral code is to be publicly recognized, and that the
best code from a philosophical standpoint is the one that maximizes average utility, he does not
maintain that the principle of utility must belong to the code itself. In fact, he denies that within the
public morality the final court of appeal need be to utility. Thus by the definition in the text, his view
is not utilitarian. See “Some Merits of One Form of Rule Utilitarianism,” 
University of Colorado
Studies
(Boulder, Colo., 1967), pp. 58f.
159
29. Main Grounds for the Two Principles

---

the conditions of generality of principle, universality of application, and
limited information are not sufficient by themselves to require these prin-
ciples is clear from the reasoning for the utility principle (§27). Further
assumptions must, therefore, be incorporated into the original position.
Thus, I have assumed that the parties regard themselves as having certain
fundamental interests that they must protect if they can; and that, as free
persons, they have a highest-order interest in maintaining their liberty to
revise and alter these ends (§26). The parties are, so to speak, persons
with determinate interests rather than bare potentialities for all possible
interests, even though the specific character of these interests is unknown
to them. They must try to secure favorable conditions for advancing these
definite ends, whatever they are (§28). The hierarchy of interests and its
relation to the priority of liberty is taken up later (§§39, 82), but the
general nature of the argument for the basic liberties is illustrated by the
case of liberty of conscience and freedom of thought (§§33–35).
In addition, the veil of ignorance (§24) is interpreted to mean not only
that the parties have no knowledge of their particular aims and ends
(except what is contained in the thin theory of the good), but also that the
historical record is closed to them. They do not know, and cannot enumer-
ate, the social circumstances in which they may find themselves, or the
array of techniques their society may have at its disposal. They have,
therefore, no objective grounds for relying on one probability distribution
rather than another, and the principle of insufficient reason cannot be
invoked as a way around this limitation. These considerations, together
with those derived from regarding the parties as having determinate fun-
damental interests, imply that the expectation constructed by the argu-
ment for the utility principle is unsound and lacks the necessary unity
(§28).
30. CLASSICAL UTILITARIANISM,
IMPARTIALITY, AND BENEVOLENCE
30. Classical Utilitarianism
I now want to compare classical utilitarianism with the two principles of
justice. As we have seen, the parties in the original position would reject
the classical principle in favor of that of maximizing average utility. Since
they are concerned to advance their own interests, they have no desire to
maximize the total (or the net balance) of satisfactions. For similar rea-
sons they would prefer the two principles of justice. From a contractarian
point of view, then, the classical principle ranks below both of these alter-
160
The Original Position

---

natives. It must, therefore, have an entirely different derivation, for it is
historically the most important form of utilitarianism. The great utilitari-
ans who espoused it were certainly under no misapprehension that it
would be chosen in what I have called the original position. Some of
them, particularly Sidgwick, clearly recognized the average principle as
an alternative and rejected it.
33
Since the classical view is closely related
to the concept of the impartial sympathetic spectator, I shall look at this
concept in order to clarify the intuitive basis of the traditional doctrine.
Consider the following definition reminiscent of Hume and Adam
Smith. Something is right, a social system say, when an ideally rational
and impartial spectator would approve of it from a general point of view
should he possess all the relevant knowledge of the circumstances. A
rightly ordered society is one meeting the approval of such an ideal
observer.
34
Now there may be several problems with this definition, for
example, whether the notions of approval and relevant knowledge can be
specified without circularity. But I shall leave these questions aside. The
essential point here is that there is no conflict so far between this defini-
tion and justice as fairness. For suppose we define the concept of right by
saying that something is right if and only if it satisfies the principles
which would be chosen in the original position to apply to things of its
kind. It may well be the case that an ideally rational and impartial specta-
tor would approve of a social system if and only if it satisfies the princi-
ples of justice which would be adopted in the contract scheme. The de-
finitions may both be true of the same things. This possibility is not ruled
out by the ideal observer definition. Since this definition makes no spe-
cific psychological assumptions about the impartial spectator, it yields no
principles to account for his approvals under ideal conditions. One who
accepts this definition is free to accept justice as fairness for this purpose:
one simply allows that an ideal observer would approve of social systems
to the extent that they satisfy the two principles of justice. There is an
essential difference, then, between these two definitions of right. The
impartial spectator definition makes no assumptions from which the prin-
33. 
Methods of Ethics,
pp. 415f.
34. See Roderick Firth, “Ethical Absolutism and the Ideal Observer,” 
Philosophy and Pheno-
menological Research,
vol. 12 (1952); and F. C. Sharp, 
Good and Ill Will
(Chicago, University of
Chicago Press, 1950), pp. 156–162. For Hume’s account, see 
Treatise of Human Nature,
ed. L. A.
Selby-Bigge (Oxford, 1888), bk. III, pt. III, sec. I, especially pp. 574–584; and for Adam Smith, 
The
Theory of Moral Sentiments,
in L. A. Selby-Bigge, 
British Moralists,
vol. I (Oxford, 1897), pp. 257–
277. A general discussion is found in C. D. Broad, “Some Reflections on Moral-Sense Theories in
Ethics,” 
Proceedings of the Aristotelian Society,
vol. 45 (1944–45). See also W. K. Kneale, “Objec-
tivity in Morals,” 
Philosophy,
vol. 25 (1950).
161
30. Classical Utilitarianism

---

ciples of right and justice may be derived.
35
It is designed instead to single
out certain central features characteristic of moral discussion, the fact that
we try to appeal to our considered judgments after conscientious reflec-
tion, and the like. The contractarian definition goes further: it attempts to
provide a deductive basis for the principles that account for these judg-
ments. The conditions of the initial situation and the motivation of the