¹ Rawls, John, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971), p. 3.Google Scholar

² Such formulations can be found not only in Rawls and in utilitarian authors, but also on the Continent—for example, in Habermas, Jürgen, Moralbewuβtsein und kommunikatives Handeln (Frankfurt: Suhrkamp, 1983), pp. 75f.Google Scholar, where Habermas proposes that social norms should be assessed by reference to the interests of all those affected by them. (For an English translation, see Habermas, Jürgen, Moral Consciousness and Communicative Action, trans. Lenhardt, C. and Nicholsen, S. W. (Cambridge, MA: MIT Press, 1990].)Google Scholar

³ We generally do not believe, for example, that neglecting to save someone is morally on a par with killing her. Nor do we think that one ought to kill whenever doing so will save a larger number from being killed. We admit that it might be better for people in general if we accepted these principles. The disposition to kill one for the sake of saving several may be desirable from the standpoint of others at large. But we do not let the recipients' standpoint alone settle the matter.

⁴ I use “contractarian” in reference to the idea of a hypothetical social contract (as opposed to, e.g., a historical one). Though my primary example is Rawls, I mean to include more thinly veiled social contracts as exemplified by Gauthier, David, Morals by Agreement (Cambridge: Cambridge University Press, 1987).CrossRefGoogle Scholar

⁵ Pure egalitarianisms, though they accept individualism and humanism, also fall outside the CC mold. It makes sense to stipulate that prudent prospective participants assess their Positions in relative terms and care greatly not to fall below the mean. However, the stipulation that this is all they care about — that they do not care at all about their absolute positions (that they are indifferent, for example, between higher and lower levels of equality)—is at odds with the assumption of rational prudence. Though outside the CC mold, pure egalitarianisms may be subject to some of the same problems as CC conceptions of justice—for example, to the first problem discussed below (Section III).

⁶ “Whatever a people cannot impose upon itself, cannot be imposed upon it by the legislator either” (Immanuel Kant, “On the Common Saying: ‘This May Be True in Theory, but It Does Not Apply in Practice,’” in Kant's Political Writings, ed. Reiss, Hans [Cambridge: Cambridge University Press, 1970], p. 85Google Scholar [Akademieausgabe, vol. VIII. p. 304f.]).Google Scholar I believe that “cannot” is to be understood here in the sense of moral possibility. See my “Kant's Theory of Justice,” Kant-Studien, vol. 79, no. 4 (Winter 1988), p. 423.Google Scholar

⁷ Catharine MacKinnon, for example, would presumably not recognize the lesser availability of pornography as a loss. See, for example, her Feminism Unmodified (Cambridge, MA: Harvard University Press, 1987).Google Scholar

⁸ Whether it can reasonably be confined to this role is another question, which has haunted utilitarianism: Why should a judge follow the best rule even in those few exceptional cases where she knows that violating this rule will lead to a better outcome? See, for example, Lyons, David, “Utility and Rights,” in Theories of Rights, ed. Waldron, Jeremy (Oxford: Oxford University Press, 1984).Google Scholar

⁹ I think the analogous claim could be made about contractarianism and consequentialism understood as guides to the assessment of conduct and character, but this is not my topic here.

¹⁰ Amartya K. Sen is the most prominent of these critics. See “Equality of What?” in his Choice, Welfare, and Measurement (Cambridge: Cambridge University Press, 1982), esp. p. 366Google Scholar; and his Inequality Reexamined (Cambridge, MA: Harvard University Press, 1992), esp. pp. 79–87.Google Scholar

¹¹ Someone using the maximin rule of decision making strives to maximize her minimum Pay-off. In the case at hand, prospective participants using the maximin rule would rank alternative institutional schemes by the worst social position each would produce: if one institutional scheme produces a superior bottom position than another, then the former is Preferred to the latter because it secures a higher minimum.

¹² We can specify this notion in various ways, for example, as physical attractiveness, as being disposed toward a cheerful temper, or as some combination of the two.

¹³ Variants of this example work for CC conceptions that use an aggregation function other than maximin. If a CC conception specifies f as averaging (conceives prospective participants as maximizing probability-weighted expectations), for instance, the example could stipulate (as seems plausible empirically) that the attractive and the unattractive differ in the extent to which education and income differentials affect their quality of life: a given advantage in education or income means more to the unattractive, because it contributes more to their chances of achieving rewarding personal lives. As in the example in the text, institutions should then be designed so as to favor the unattractive in the distribution of such advantages. For the sake of simplicity, the examples are somewhat crude. The same point could be made through illustrations involving more fine-grained or even scalar dimensions, more dimensions than our three (education, class, genetic attractiveness), or correlations between class and talents or class and motivation.

¹⁴ See, for example, Arneson, Richard, “A Defense of Equal Opportunity for Welfare,” Philosophical Studies, vol. 62, no. 2 (05 1991)CrossRefGoogle Scholar; Cohen, G. A., “On the Currency of Egalitarian Justice,” Ethics, vol. 99, no. 4 (07 1989)CrossRefGoogle Scholar; and Roemer, John, A Future for Socialism (Cambridge, MA: Harvard University Press, 1994), especially the section on what socialists want. Arnartya Sen (supra note 10) may not belong on this list. To me, at least, his examples suggest that he envisions a conception of justice that is not sensitive to all natural inequalities that affect quality of life (as colloquially understood).CrossRefGoogle Scholar

¹⁵ See Rawls, John, Political Liberalism (New York: Columbia University Press, 1993), pp. 74f., 106.Google Scholar

¹⁶ Ibid., p. 74. The other two higher-order interests are the interests in developing and exercising the two moral powers (ibid.), namely,

a capacity for a sense of justice and a capacity for a conception of the good. A sense of justice is the capacity to understand, to apply, and to act from the public conception of justice which characterizes the fair terms of social cooperation. Given the nature of the political conception as specifying a public basis of justification, a sense of justice also expresses a willingness, if not the desire, to act in relation to others on terms that they also can publicly endorse…. The capacity for a conception of the good is the capacity to form, to revise, and rationally to pursue a conception of one's rational advantage or good. (Ibid., p. 19)

¹⁷ In Political Literalism, Rawls formulates his two principles of justice as follows. First Principle: “Each person has an equal claim to a fully adequate scheme of equal basic rights and liberties, which scheme is compatible with the same scheme for all; and in this scheme the equal political liberties, and only those liberties, are to be guaranteed their fair value” (Ibid., p. 5). Second principle: “Social and economic inequalities are to satisfy two conditions: first, they an? to be attached to positions and offices open to all under conditions of fair equality of opportunity [the opportunity principle]; and second, they are to be to the greatest benefit of the least advantaged members of society (the difference principle]” (ibid., p. 6). The first principle is to have lexical priority over the second and, within the second principle, opportunity principle is to have lexical priority over the difference principle (Rawls, , A Theory of Justice, p. 302f.).Google Scholar The word “lexical” is short for “lexicographical” (ibid., p. 42f.). The paradigm case of a lexical priority is this: when arranging words alphabetically, as in a lexicon, one gives priority to the first letter over the second, to the second over the third, and so on. For example, one lists all words beginning with an “a” ahead of all words beginning with a “b” regardless of the second and subsequent letters—as when “azure” is listed ahead of “baal.” Rawls conceives the priority of one principle over another along the same lines: in the comparison of basic structures, an advantage in terms of a lexically prior principle overrides any disadvantage in terms of lexically subordinate ones.

¹⁸ See Pogge, Thomas W., Realizing Rawls (Ithaca: Cornell University Press, 1989), pp. 44–47, 114–16, 183–89.Google Scholar I also show there that CC conceptions meeting this constraint (I called them “semiconsequentialist”) can accommodate our conviction that the distribution of health care should favor those with greater health problems by defining the relevant social good as “access to health care when needed.”

¹⁹ This demand would generally lead to institutions favoring the naturally handicapped, even if the parties do not employ the maximin rule. But there could be exceptions. It might be argued, for instance, that lack of access to higher education generally reduces the qualilty of life of gifted persons more dramatically than, and (other things equal) below, that of the less gifted. Contracting parties presumed to know this would then have reason to prefer institutions under which the naturally better endowed are favored in the distribution of the social good education.

²⁰ The principle requires that “those who are at the same level of talent and ability, and have the same willingness to use them, should have the same prospects of success regardless of their initial place in the social system, that is, irrespective of the income class into which they were born” (Rawls, , A Theory of Justice, p. 73, cf. pp. 275, 301).Google Scholar Does fair equaliy of opportunity rule out group inequities involving race and gender as well? Rawls does not incorporate mis constraint—perhaps because he believes that formal equality of opportunity (together with the first principle of justice) suffices to wash out race-induced and gender-induced inequalities over time.

²¹ This illustration uses class-induced inequalities in order to stay as close as possible to Rawls's formulation of fair equality of opportunity.

²² I have designed the example so that the inequality in average annual incomes is very modest indeed (compared to present realities), in order to reassure the reader that the fair value of the political liberties is really being maintained and that the difference principle is really satisfied.

²³ It may be said that I have not told a plausible story showing that the maintenance of fair equality of opportunity can have net costs; if so, another such story can be substituted. One might further object that the difficulty I indicate cannot arise at all, because maintaining fair equality of opportunity has important benefits that will always outweigh any costs. This strikes me as unlikely. Even if it were a fact, however, it would not support the inherent moral importance that Rawls (and most of us) attach to the avoidance of class inequities. Fair equality of opportunity would not be demanded in its own right by justice (by prudent prospective participants), but would come about as a merely contingent by-product of the realization of economic efficiency or the difference principle. This scenario would confirm, not defeat, my claim that CC conceptions cannot explain why fair equality of opportunity should have any moral standing in and of itself. I am indebted to Ling Tong for this objection and subsequent discussion.

²⁴ To ensure comparability between this and the previous table, let us assume that any 0xpenditures for education are counted not as income of those adults who pay them (parents under S₁, and taxpayers under S₂), but as income of the youngsters who are receiving the educational benefit.

²⁵ Rawls explicitly repudiates this idea. See Rawls, John, “Social Unity and Primary Goods,” in Utilitarianism and Beyond, ed. Sen, Amartya K. and Williams, Bernard (Cambridge: Cambridge University Press, 1982), p. 164.Google Scholar

²⁶ Nozick, Robert, Anarchy, State, and Utopia (New York: Basic Books, 1974), p. 157n.Google Scholar

²⁷ The answer may well depend on whether they have offspring or not. Those who do will probably care about their children's and grandchildren's chances of occupying privileged positions. Interestingly, this offspring argument could work only for race and class, but not for gender: in the first two cases, fair equality of opportunity benefits the offspring of the worst off—those from the lowest class or of the currently most disfavored race. Fair equality of opportunity for women, by contrast, benefits one's offspring only if one's offspring are predominantly female. Assuming zero correlation between social disadvantage and predominantly female offspring, and assuming also that disadvantaged mothers do not care more for their daughters than for their sons, we cannot then argue that fair equality of opportunity for women is in the interest of the worst off because it benefits their offspring.

²⁸ He envisions a second stage, after ideal theory, in which the preferred conception of justice and the ground rules it selects are tested for stability. This test is conceived as an either-or affair and is never integrated with the results of the first stage. Rawls never aims for an overall appraisal of alternative institutional schemes, an appraisal that would take account of how each would actually work in practice (degree of stability, incidence of noncompliance, etc.). See, for instance, his Political Literalism, pp. 64Google Scholar, 133f., 140f.

²⁹ I have in mind constitutions whose words are highly protective of civil rights and democratic procedures but bear little or no relation to the realities in the relevant country. The successive constitutions of the USSR are a clearcut example, as are many constitutions in present and former Third World dictatorships and one-party regimes.

³⁰ Rawls, , “Social Unity and Primary Goods,” p. 163.Google Scholar

³¹ Such a security threshold would presumably be defined in terms of frequency-based probabilities. Here one may need to take account of various personal characteristics. Thus, it is quite possible that a certain basic freedom is reasonably secure for some citucns (defined by wealth, geography, race, gender, or age, for example) but not for others.

³² The discontinuity could be somewhat mitigated: justice demands, one might say, that the basic liberties be made more secure than the first principle mandates, and this demand on a par with, and to be balanced against, the demands of either the opportunity principlex or the difference principle.

³³ This is said to include that one be free “from psychological oppression and physical assault.” See Rawls, John “Reply to Alexander and Musgrave,” Quarterly Journal of Economics, vol. 88, no. 4 (11 1974), p. 640.CrossRefGoogle Scholar

³⁴ Rawls, , A Theory of Justice, p. 242.Google Scholar The quote shows nicely how his commitment to the prospective-participant perspective attenuates the uplifting idea (chosen for the cover of the paperback) that “[e]ach person possesses an inviolability founded on justice that even the welfare of society as a whole cannot override” (ibid., p. 3).

³⁵ Ibid., p. 242.

³⁶ One may also be attracted to strict-liability laws that criminalize harmings through rnsiness, mental confusion, and the like, where the main benefit is not deterrence but prevention of like harms caused by the same offender.

³⁷ There is a similar categorical commitment to various other procedural safeguards – for example, to the prohibition of double jeopardy (Fifth Amendment to the U.S. Constitution) to the prohibitions on informing the jury about the accused's prior convictions.

³⁸ In nonideal theory, we must settle what kinds of abridgments of basic liberties are called for when it is impossible to safeguard all basic liberties for all. What sorts of abridgments of the basic rights of the accused would the contracting parties countenance for the sake of reducing to a certain extent the excessive incidence at a certain level of crimes of a certain kind? It may be said that we are not (yet) in a nonideal situation, that current crime rates are not excessive, do not undermine the claim that the basic right to physical integrity is fulfilled for all; and it may then be concluded that the first principle prohibits the lower standards of evidence, assuming it mandates evidence beyond a reasonable doubt in ideal theory. The issue cannot be so easily avoided, however. For we must also justify how, in ideal theory, the basic liberties are carved out of a much wider domain of candidate freedoms, how one particular design of a universalizable scheme of basic liberties is chosen. This involves snowing that the parties would indeed consider the existing insecurity of the right to physical integrity consistent with the first principle being fully satisfied, thereby protecting the basic rights of the accused from abridgment. At some level, the parties must trade off between citizens' freedom from being innocently convicted and citizens' freedom from being innocently victimized by crime. For some reflections about this trade-off, see Reiman, Jeffrey and van den Haag, Ernest, “On the Common Saying that It Is Better that Ten Guilty Persons Escape than that One Innocent Suffer: Pro and Con,” Social Philosophy and Policy, vol. 7, no. 2 (Spring 1990).CrossRefGoogle Scholar

³⁹ A punishment is more or less severe (as I use this term) the more or less it hurts the person punished. In many jurisdictions, rich persons receive larger fines than poor ones for the same offense. This does not mean that the severity of punishment varies with wealth. To the contrary, those differentials accommodate the concern that severity should not vary with wealth. Only a larger fine can punish a rich offender as severely as a smaller fine punishes a poor one.

⁴⁰ The word “expected” is to serve as a reminder that costs and benefits have to be assessed from the perspective specified by the relevant CC conception. If the hypothetical contracting parties use the maximin rule, for example, they define the expected net benefit of a punishment scheme in terms of its impact on the quality of life of the worst off (in comparison with the quality of life that the worst off under alternative schemes would enjoy).

⁴¹ I assume that, for CC conceptions, the imposition of punishments will involve expected costs in terms of quality of life. (Otherwise prospective participants would want equal, namely maximal, punishments for all crimes, and the prospective-participant perspective would be in even greater trouble than I allege.) This assumption is plausible, because Prospective participants run some risk of being (correctly or falsely) convicted and subjected to any punishment they authorize, and may have to pay a share of the cost of inflicting punishments. It does not follow, however, that more severe punishment for a particular crime will entail greater cost. When elasticity is high, the opposite may be true: though each person convicted of the crime receives more jail time (say), the crime is less frequent and fewer Persons are punished for it. Since the increase in the first factor may be overwhelmed by the decrease in the second, total jail time imposed for the crime may actually decline.

⁴² I assume, as seems inevitable, that any reasonable CC conception counts involuntary Premature death as a major evil.

⁴³ The figures provided by the American Automobile Association are as follows: The annual death toll on U.S. highways hovers around forty-three thousand. About half these deaths are caused by accidents in which at least one intoxicated person is relevantly involved (i.e., involved not merely as passenger, but as driver or pedestrian). Of course, in some of these cases there is no drunk driver, and in others the intoxication of one or more drivers did not contribute to the accident. There is some dispute over what the death toll on U.S. highways would be if there were no drunk driving, but the lowest estimate seems to be that ten thousand premature deaths (plus one hundred thousand injuries and one billion dollars property damage) would be avoided every year. For these figures, see Ross, H. Laurence, Confronting Drunk Driving (New Haven: Yale University Press, 1992), pp. 35–38.Google Scholar Over a period of seventy years, then, a U.S. resident's chance of death due to drunk driving works out to about one in four hundred.

⁴⁴ We must consider that the projected “reform” would increase the killing of police officers: notorious drunk drivers, when flagged down, may believe that they have little left to lose. I owe this point to Michael Levin.

⁴⁵ Note that the argument for using the death penalty here extends, for purposes of assessing social institutions, the same concern and respect to all citizens–criminals and innocents alike. It counts the execution of a guilty person as no less bad than the vehicular death of an innocent. A higher execution rate would be called for, if the lives of innocents were weighted more heavily.

⁴⁶ For the record, let me say that – though I am, unconvinced by the draconian practices that (I have argued) are entailed by the CC approach – I do believe that punishments for grave and repeated drunk-driving offenses ought to be significantly more severe than they currently are and than they ought to be according to Husak, Douglas, “Is Drunk Driving a Serious Offense?” Philosophy anJ Public Affairs, vol. 23, no. 1 (Winter 1994), pp. 52–73.CrossRefGoogle Scholar But then this essay is not about appropriate punishments for the various drunk-driving offenses.

⁴⁷ A good many victims of drunk driving are, of course, victims of their own drunk driving.

⁴⁸ In the long passage I quoted from Rawls at the beginning of Section VA, he writes that the representative citizen can accept the strict-liability statute “at least if the penalties imposed are not too severe.” But whence this limitation? The harms at stake (“civil strife”) are very serious, and, if only the death penalty can avert them, then it is possible that the death penalty is called for to protect optimally the representative citizen's right to life.

⁴⁹ For the United States, witness, for example, punishments of so-called “accessories after the fact” and punishments under the Racketeer Influenced and Corrupt Organizations (RICO) statute. In both cases the punishment for given criminal conduct can become much more severe on account of circumstances beyond the offender's knowledge and control (an element of strict liability) and can become grossly disproportionate to the offender's culpability.

⁵⁰ See Scanlon, Thomas M., “Contractualism and Utilitarianism,”Google Scholar in Sen, and Williams, , eds., Utilitarianism and Beyond, esp. pp. 110–28.Google Scholar