I am grateful to those who contributed such thoughtful responses to my proposal. Although I cannot discuss all their points, I will consider some of the most challenging here.
The Meaning of Equal Opportunity
Scanlon points out that there is another conception of equality of opportunity besides mine, which in some instances is more appropriate, and I agree. I do not wish to apply my proposal for equalizing opportunity to such positions as opera star or mathematics professor. In what might be called “contests,” our purpose is to discover the person who can best carry out a socially valuable task, and it is quite acceptable that extraordinary native talents will prevail. Although I did not properly restrict the scope of my inquiry, I am concerned with equating opportunity for kinds of advantage whose individual enjoyment does not impose costs on others. For example, my being healthy or educated does not impose costs on others, the way my becoming an opera singer would, were I, someone lacking vocal talent, given a sufficiently large handicap to enable me to win the role of Papageno in The Magic Flute.
Now some would argue against this distinction, claiming that if Joe, of low native ability, is deemed entitled to the relatively large share of society’s educational resources necessary to educate him, then those resources cannot be spent on others whose education would eventually deliver more socially valuable output. I do not deny there is a potentially difficult issue here, nor can I immediately propose the exact parameters of its solution. But I think the distinction is defensible, and the defense will be found in elaborating the view that it is reasonable for a person to expect to enjoy the kind of advantage that a modicum of education brings (where the size of the modicum is an increasing function of society’s per capita wealth), but not reasonable for a person to expect to enjoy the advantage that being an opera star brings.
My proposal divides the problem of defining equality of opportunity into two parts. The first, which I had hoped would be relatively uncontroversial, is the procedure of distributing income or health services or whatever resource is relevant to the acquisition of the advantage in question among different types so as to equalize advantage across types for all persons at a given centile of their type distribution of “effort.” The second part is the partitioning of the relevant population into types, a procedure which would inevitably be fraught with controversy. One virtue of the proposal, I think, is that it concentrates the political debate on this second part, and solves at least a substantial part of the problem (namely, the first part).
Samuel Scheffler asks why society should not debate the method proposed in the first part as well. Of course, it may — but my hope is that consensus can be reached at least on a definition of equality of opportunity for a certain kind of advantage, even if the definition leaves open the partition of society into types. I wish to argue that my definition is the right one — or, more precisely, is a statistically feasible approximation to the right one — even if we do not have a solution to what the “correct” type partition is.
Eric Maskin disagrees with the “principle that the set of choices available to a type is accurately reflected by the empirical distribution of choices actually made.” Thus, for Maskin, even if all members of a type performed exactly the same harmful action (e.g., smoking for 30 years), it would not be correct to conclude that no other action was accessible to them. I agree that it would be preferable to measure less obliquely when an action is beyond the control of a person, but, again pragmatically, I think such measurement is rarely possible. We tend to believe that people can control their behavior when we have no apparent reason to think otherwise. One could say that I subscribe to an assumption of charity: if no member of a type avoids an action that has deleterious consequences, then I say that, by virtue of their type, these individuals had no other action accessible to them. For instance, in one important case that inaccessibility may be due to the inability to perceive correctly the (high) probability of harmful consequences of that action. Such misperception (take smoking, for example) might well be psychologically very difficult to avoid: if all those who are similarly situated to me (e.g., Black male steelworkers) perform some action, it may be difficult for me to believe that that action is stupid or harmful, for if I did, might I not be compelled to call into serious question the way of life of those with whom I identify?
Arthur Ripstein believes that responsibility should be decoupled from control. He advocates that we hold persons responsible for behavior if it imposes severe costs on others. I agree that the costs of Smith’s asocial behavior should not be borne entirely by those upon whom they happen to fall, but it does not follow that Smith should bear those costs, as we would deem appropriate were we to hold her responsible for the behavior. If Smith were clearly not in control of the behavior, then I say it should be viewed as a kind of natural disaster, which no one is responsible for, and whose costs everyone should share in bearing. It surely does matter, to me, how a person’s expensive tastes were acquired. If, to take an extreme example, a person were hypnotized to regurgitate every time he was fed anything but filet mignon, and such hypnosis were irreversible, I would support his receiving a state subsidy, if without such a subsidy he would have to live on the streets to support his bizarre eating habit.
Susan Hurley says that my approach is unworkable if the kind of action the person is supposed to take to defend herself from disadvantage is multidimensional. I completely agree: my proposal is only well-defined, let alone justifiable, if there is a unidimensional kind of “effort” a person can take to lower the probability of suffering the disadvantage in question. In the examples I gave, that effort was “avoiding smoking” or “getting educated,” both of which can be measured in (unidimensional) years. My method for equalizing opportunity is only intended for a situation in which individuals must expend some costly, unidimensional effort to achieve the advantage in question, and in those cases, I believe my procedure is more attractive than Hurley admits. In such cases, I think equality-of-opportunity requires equalizing advantage at given centiles of effort across types, because, but for the traits that define type and for which they are not responsible, they have behaved similarly. The kind of agnosticism Hurley expresses concerning this judgment would require her to challenge almost any counterfactual claim, and therefore much of the foundations for the theory of statistical inference. Short of having the kind of data that Eric Maskin wants, and that I think are much less available than the kind I use, I see no more practical solution to the problem of assigning responsibility.
Although Robert Solow and Richard Epstein occupy very different positions in the political spectrum, their criticisms of my proposal are quite similar, in that they both think it would require unacceptable government intrusion or would induce intolerable levels of political intrigue in its implementation. There is a simple, but unfortunately technical, answer to these charges.
Suppose the problem is to come up with a tax code that implements equality of opportunity for some kind of advantage. And suppose that a “tax code” consists in a specification of the rule by which any person must compute the tax she must pay. Now if the tax code specifies a different rule for every type, then the charges of abuse that Solow and Epstein level against the proposal are real. But suppose the tax code specifies the same rule for every type, perhaps a fairly complicated rule (e.g., one for which marginal rates of taxation change at many different income levels). As long as the rule is the same for all types, no person will have any incentive to take actions to change his type — indeed, the Internal Revenue Service will never have to identify the type of any individual. That (single) rule can be chosen as the optimal rule, among all possible single rules, for the objective of reaching equality of opportunity. Now restricting the planner’s problem to the choice of a single rule for all types will not enable us to implement the degree of equality of opportunity that could be implemented were the planner able to vary the rule across types, assuming that no taxpayer behaved dishonestly or “strategically.” In general, the planner’s problem can allow a degree of tax-rule differentiation across types that will not provide incentives to dissemble. Indeed, we already have such differentiation in our present tax code: for instance, taxpayers are subject to different tax rules according to the number of dependents they have, and whether they are disabled in certain ways. In other words, accepting the real danger of opportunistic behavior of the population constrains the degree of equality of opportunity that can be implemented, but it does not reduce that degree to zero.
Elizabeth Fox-Genovese and Nancy Rosenblum both criticize my approach from rather nihilistic viewpoints. Fox-Genovese doubts that we can ever “calculate the conditions and attributes that invalidate formal equality of opportunity,” and asserts that my proposal presupposes an “extraordinary illusion of mastery” of the complexity of individual motivations and behavior. I simply do not share her agnosticism. Suppose we limit ourselves to three characteristics of persons, arguably each outside of individual control: race, performance on IQ or similar tests, and socio-economic position of parents. (I need take no position on whether IQ is heritable, environmentally determined, measures intelligence, or whether tests which measure it are culturally biased. I assert only that a person’s IQ is largely beyond her control, and that the kind of talent IQ measures is important for the acquisition of many kinds of advantage in our society.) The enjoyment of many kinds of advantage is, I conjecture, quite sharply differentiated across the types defined by these three characteristics. Can Fox-Genovese really claim that an attempt to reduce the differences in these degrees of advantage is ill-conceived? How can she further assert that my proposal “moves the politics of equality of opportunity a giant step closer toward a self-defeating politics of an unrealizable equality of outcome?” Is it not obvious that, if any contemporary society moves towards equality of opportunity, then outcome inequalities will be reduced? Reducing outcome inequalities generated by inequality of opportunity is not the same thing as eliminating all inequality of outcome, which I am not advocating. Fox-Genovese’s analogy between the Kentucky Derby and an economy is mistaken. The Kentucky Derby is (aside from the business of betting) a contest to discover the best horse, or the best horse-jockey team, but the purpose of an economy is not, in my view, to discover the best economic competitors, but rather to provide a decent life for all.
Nancy Rosenblum’s nihilism is evidenced in her claim that I wish away the “ambiguity and incoherence that characterize American political ideology;” my project is “wildly off the mark of anything psychologically and politically possible for us” because of the “massive confusion” in popular culture about personal responsibility. Even if my proposal is philosophically justifiable, it is hopelessly naive to believe it could be implemented in the confused American context, one where “intuitions about equality are ambiguous and contradictory.” For Rosenblum, more “modest moves” must be taken towards equality of opportunity.
Perhaps Rosenblum is right in her diagnosis of our national consciousness; but her concern is really quite different from mine. My purpose is to clarify what commonly held perceptions about personal responsibility (in the sense of having control over acts) and equality of opportunity (in the sense of leveling the playing field) imply for social policy. I do not contend that these perceptions are universal, but they are held by many, indeed many of polar political persuasions. If I am right that these views imply the kind of policy I have described, then there is a basis for productive public debate. I would not expect to win over large numbers of voters to my view in the short run, even should I be correct. But I do strongly believe that a commitment to democracy, which Rosenblum and I share, is built upon a view that people, even if their current views are “ambiguous and incoherent,” are capable of reasoned and objective deliberation. Otherwise, we should be looking for philosopher kings.