¹ Rawls, John, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971)Google Scholar.

² Ibid., p. 16.

³ Ibid., p. 11.

⁴ Ibid., p. 15.

⁵ Ibid., p. 137.

⁶ Ibid., p. 15.

⁷ Ibid., p. 85.

⁸ Ibid., p. 303.

⁹ Ibid., p. 302.

¹⁰ Ibid.

¹¹ Ibid.

¹² For further discussions of the problem of justify-ing Rawls's proposed original position or of choosing from competing alternatives, see especially: Nagel, Thomas, “Rawls On Justice,” The Philosophical Review, 82 (04, 1973), 220–234CrossRefGoogle Scholar; Hare, R. M., “Rawls' Theory of Justice—I,” Philosophical Quarterly, 23 (04, 1973), 144–155CrossRefGoogle Scholar; and Feinberg, Joel, “Justice, Fairness and Rationality,” Yale Law Journal, 81 (04, 1972), 1004–1031CrossRefGoogle Scholar.

¹³ Rawls, p. 154.

¹⁴ Ibid., pp. 150–6.

¹⁵ This can be seen from the following:

“His expectation is defined … by a weighted sum of utilities of representative individuals, that is, by the expression Σ P_i u_i, where p_i is the likelihood of his achieving the ith position, and u _i the utility of the corresponding representative man” (pp. 164–5).

Now, if an agent in the original position reasons according to the principle of insufficient reason that he has an equal likelihood of turning out to be any particular individual, then he may compute p_i in each case by determining the fraction which the ith person represents of the total population (in the case of one individual, 1/n). His total prospect is thus

1/n u _i + 1/n u ₂ + … + 1/n u _n or Σ 1/n u_i/n

The above expression, however, is identical to the average utility, for since the total utility must be Σ u_i and the total number of persons is n, the average utility for a given society must represent the total utility divided by the number of its persons, or Σ u_i/n. Hence, the society which maximizes the average utility also maximizes the expectation of total utility for any individual in the original position-provided that one assumes he has an equal likelihood of turning out to be anyone. See especially pp. 164–5.

¹⁶ Rawls, p. 154.

¹⁷ Ibid., p. 173.

¹⁸ Ibid., p. 157.

¹⁹ Ibid., p. 154.

²⁰ According to what Rawls calls the “four-stage sequence” (pp. 195–201), particular facts will be revealed in due course, but only after the principles of justice are chosen. At that point enough information about the particular society is to be provided to allow for the intelligent design of a constitution and other policies for the just society. But since the principles of justice are supposed to have general applicability regardless of the particulars which may later be revealed, no factual information about the society in question is to be introduced at that stage of the deliberations. There is thus no basis for estimations of probability apart from the principle of insufficient reason which Rawls claims to be inadequate.

²¹ Ibid., p. 155.

²² Ibid., p. 397.

²³ Rawls introduces an additional premise which he calls the “Aristotelian Principle.” He grants, however, that this “principle does not assert that any particular kind of activity will be preferred. It says only that we prefer, other things equal, activities that depend upon a larger repertoire of realized capacities and that are more complex” (429–30). It should be noted that this principle would in no way affect the indeterminacies we shall explore in the doctrine of rational plans as it is meant to support the relevant premises about primary goods. The addition of the Aristotelian Principle, in other words, would not provide the basis for Rawls's assumptions concerning the relative worth of liberty, equality of opportunity and income and wealth and the amounts required of each. Rational plans might be preferred, according to the Aristotelian Principle, because they are more complex or more inclusive even though those plans may require different primary goods or the same goods in a different order of priority.

Furthermore, we shall set aside the Aristotelian Principle in the discussion which follows because Rawls indicates that the argument is not meant to depend on it:

“The correctness of the definition of a person's good in terms of the rational plan for him does not require the truth of the Aristotelian Principle. The definition is satisfactory, I believe, even if this principle should prove inaccurate, or fail altogether” (433). For further criticism of the Aristotelian Principle, see Barry, Brian, The Liberal Theory of Justice (Oxford: The Clarendon Press, 1973), pp. 27–30Google Scholar.

²⁴ Rawls, p. 421.

²⁵ Ibid., p. 408.

²⁶ Ibid., p. 413.

²⁷ Ibid., pp. 408, 416.

²⁸ Ibid., pp. 409, 421.

²⁹ Ibid., pp. 154–55.

³⁰ The term “threshold” to describe these assumptions is also used by Barry, “The Liberal Theory of Justice,” p. 97Google Scholar.

³¹ Rawls, p. 154.

³² Ibid., pp. 449, 564.

³³ Ibid., p. 408.

³⁴ Ibid., p. 412.

³⁵ Ibid., p. 409.

³⁶ Ibid., note 11, p. 408.

³⁷ Ibid., p. 416.

³⁸ Ibid., p. 423.

³⁹ Ibid., p. 432.

⁴⁰ Ibid., p. 143.

⁴¹ Ibid., p. 125.

⁴² Teitelman, Michael, “The Limits of Individualism,” The Journal of Philosophy, 69 (10 5, 1972), 545–556, p. 550CrossRefGoogle Scholar.

⁴³ Rawls, p. 396.

⁴⁴ Ibid., p. 42.

⁴⁵ Ibid., p. 156.

⁴⁶ Ibid., p. 42.

⁴⁷ Ibid., p. 43.

⁴⁸ Ibid.

⁴⁹ Ibid.

⁵⁰ Ibid.

⁵¹ For certain complications in the application of this threshold notion to the difference principle see the discussion of that principle in 1.2*.

⁵² Rawls, p. 302.

⁵³ Ibid., p. 207.

⁵⁴ Ibid.

⁵⁵ Ibid., p. 61.

⁵⁶ Berlin, Isaiah, Four Essays On Liberty (New York: Oxford University Press, 1969), p. 129Google Scholar.

⁵⁷ Ibid., p. 130.

⁵⁸ Rawls, p. 206.

⁵⁹ Ibid., p. 221.

⁶⁰ Ibid., p. 224.

⁶¹ Ibid.

⁶² Ibid., p. 196.

⁶³ Ibid., p. 197.

⁶⁴ Ibid., p. 542.

⁶⁵ Ibid., p. 543.

⁶⁶ Ibid., p. 72.

⁶⁷ Ibid., p. 87.

⁶⁸ Ibid., p. 74.

⁶⁹ Ibid.

⁷⁰ Ibid., p. 301.

⁷¹ Ibid., p. 74.

⁷² Ibid., p. 512.

⁷³ Ibid., p. 157.

⁷⁴ Ibid., p. 538.

⁷⁵ Ibid., p. 534.

⁷⁶ Ibid., p. 440.

⁷⁷ Ibid., p. 158.

⁷⁸ Ibid.

⁷⁹ Ibid.

⁸⁰ Ibid.

⁸¹ Ibid., p. 80.

⁸² Ibid., p. 64.

⁸³ Ibid., p. 81. The distribution to X₁; the most favored, is marked off along the horizontal axis and the other two positions are plotted on the vertical axis.

⁸⁴ Ibid.

⁸⁵ Ibid., pp. 81–82.

⁸⁶ Ibid., p. 80.

⁸⁷ Ibid., pp. 150–151.

⁸⁸ Rawls uses this term on page 153.

⁸⁹ See the argument for the principle of average utility described in Footnote 15.

⁹⁰ The guaranteed minimum principle would clearly be the proper conclusion if conditions sufficiently favorable to fully implement it are assumed. Such an assumption would conform to what Rawls calls “ideal theory” (see pp. 245–7). The problem of a suitable alternative to maximin for the “less than favorable conditions” of “non-ideal theory” (p. 245) has been discussed by Douglas Rae and James Fishkin, “A Weak Theory of Justice,” prepared for the Public Choice Society, Meetings of March 21–3, 1974, New Haven, Connecticut. See also Hare's comparison of the actual consequences of Rawls's argument to an “insurance strategy,” Hare, R. M., “Rawls' Theory of Justice-II” The Philosophical Quarterly, 23 (07, 1973), p. 249Google Scholar.

⁹¹ Given conditions which allow everyone to at least reach the threshold, all maximum (or equality) distributions would also be guaranteed minimum distributions, but all guaranteed minimum distributions would not be maximin (or equality) distributions. Only the latter, weaker principle, is warranted by the proposed premises.

⁹² Of course, the first of the three features would still remain to be weighed against the gambling argument. Rawls has never claimed, however, that the first feature could stand alone as a case against the gambling argument. He claims that his case against the principle of insufficient reason is “plausible” only in light of a certain estimation of the risks and of the probable gains. See, for example, p. 169.

⁹³ This argument is meant to apply to both the general and special conceptions. Even though it is to be found in a chapter entitled “The Reasoning Leading to the Two Principles of Justice,” it begins, on page 150, in terms of the “general conception of justice as fairness” and the problem of distributing “all primary social goods.”

⁹⁴ Rawls, p. 85

⁹⁵ The veil of ignorance might be raised sufficiently for them to judge the relative impact of the principle chosen on the various social positions so as to allow them to make their choice according to a unified theory of primary goods. Such a procedure would also require that the number of persons in the original position be imagined as corresponding to the number of positions in the society for which the choice is to be made. Rawls's version of the original position included no such requirement.

⁹⁶ Rawls, p. 151.

⁹⁷ For other games of fair division, particularly those more suited to the n-person case, see Luce, R. D. and Raiffa, Howard, Games and Decision (New York: John Wiley and Sons, 1957), pp. 363–368Google Scholar.

⁹⁸ Our discussion of maximin has focused almost entirely on the question of whether it would be chosen in the original position. We have not fully explored its substantive policy disadvantages although some of these were mentioned in the discussion of the difference principle and inequality in 1.2*. For a fuller discussion, see the paper by Douglas Rae in this symposium.