In 1945 a black man named Sweatt applied to the University of Texas Law School, but was refused admission because state law provided that only whites could attend. The Supreme Court declared that this law violated Sweatt’s rights under the Fourteenth Amendment to the United States Constitution, which provides that no state shall deny any man the equal protection of its laws.1 In 1970 a Jew named DeFunis applied to the University of Washington Law School; he was rejected although his test scores and college grades were such that he would have been admitted if he were black or Filipino or a Chicano or an American Indian. DeFunis asked the Supreme Court to declare that the Washington practice, which offered less exacting standards to minority groups, violated his rights under the Fourteenth Amendment as well.2
The Washington admissions procedures were complex. Applications were divided into two groups. The majority—those not from the designated minority groups—were first screened so as to eliminate all applicants whose predicted average, which is a function of college grades and aptitude test scores, fell below a certain level. Majority applicants who survived this initial cut were then placed in categories that received progressively more careful consideration. Minority group applications, on the other hand, were not screened, but all received the most careful consideration by a special committee consisting of a black law professor and a white professor who had taught in programs to aid black law students. Most of the minority applicants who were accepted in the year in which DeFunis was rejected had predicted averages below the cut-off level, and the law school conceded that any minority applicant with his average would certainly have been accepted.
The DeFunis case split those political action groups that have traditionally supported liberal causes, as Ann Fagan Ginger’s useful collection shows. The B’nai B’rith Anti-Defamation League and the AFL-CIO, for example, filed briefs as amici curiae in support of DeFunis’s claim, while the American Hebrew Women’s Council, the UAW, and the UMWA filed briefs against it.
These splits among old allies demonstrate both the practical and the philosophical importance of the case. In the past liberals held, within one set of attitudes, three propositions: that racial classification is an evil in itself; that every person has a right to an educational opportunity commensurate with his abilities; and that affirmative state action is proper to remedy the serious inequalities of American society. In the last decade, however, the opinion has grown that these three liberal propositions are in fact not compatible, because the most effective programs of state action are those that give a competitive advantage to minority racial groups.
That opinion has, of course, been challenged. Some educators argue that benign quotas are ineffective, even self-defeating, because preferential treatment will reinforce the sense of inferiority that many blacks already have. Others make a more general objection. They argue that any racial discrimination, even for the purpose of benefiting minorities, will in fact harm those minorities, because prejudice is fostered whenever racial distinctions are tolerated for any purpose whatever. But these are complex and controversial empirical judgments, and it is far too early, as wise critics concede, to decide whether preferential treatment does more harm or good. Nor is it the business of judges, particularly in constitutional cases, to overthrow decisions of other officials because the judges disagree about the efficiency of social policies. This empirical criticism is therefore reinforced by the moral argument that even if reverse discrimination does benefit minorities, and does reduce prejudice in the long run, it is nevertheless wrong because distinctions of race are inherently unjust. They are unjust because they violate the rights of individual members of groups not so favored, who may thereby lose a place as DeFunis did.
DeFunis presented this moral argument, in the form of a constitutional claim, to the courts. The Supreme Court did not, in the end, decide whether the argument was good or bad. DeFunis had in fact been admitted to the law school after one lower court had decided in his favor, and the law school said that he would be allowed to graduate however the case was finally decided. The Court therefore held that the case was moot and dismissed the appeal on that ground. But Mr. Justice Douglas disagreed with this neutral disposition of the case: he wrote a dissenting opinion in which he argued that the Court should have upheld DeFunis’s claim on the merits.
Many universities and colleges have taken Justice Douglas’s opinion as handwriting on the wall, and have changed their practices, in anticipation of a later Court decision in which his opinion prevails. In fact, his opinion pointed out that law schools might achieve much the same result as the University of Washington admissions procedures by using a more sophisticated policy. A school might stipulate, for example, that applicants from all races and groups would be considered together, but that certain minority applicants’ aptitude tests would be graded differently, or given less weight in overall predicted average, because experience had shown that standard examinations were for different reasons a poorer test of the actual ability of these applicants. But if this technique is used deliberately to achieve the same result, it is devious, and it remains to ask why the candid program the University of Washington used was either unjust or unconstitutional.
DeFunis plainly has no constitutional right that the state provide him a legal education of a certain quality. His rights would not be violated if his state did not have a law school at all, or if it had a law school with so few places that he could not win one on intellectual merit. Nor does he have a right to insist that intelligence be the exclusive test of admission. Law schools do rely heavily on intellectual tests for admission. That seems proper, however, not because applicants have a right to be judged in that way, but because it is reasonable to think that the community as a whole is better off if its lawyers are more intelligent. Intellectual standards are justified, that is, not because they reward the clever, but because they seem to serve a useful social policy.
Law schools sometimes serve such policy better, moreover, by supplementing intelligence tests with other sorts of standards: they sometimes prefer industrious applicants, for example, to those who are brighter but lazier. They also serve special policies for which intelligence is not relevant. The University of Washington law school, for example, gave special preference not only to the minority applicants but to veterans who had been at the school before going to war; neither DeFunis, nor any of the briefs submitted in his behalf, complained of that preference.
DeFunis does not have an absolute right to a law school place, or a right that only intelligence be used as a standard for admission. He says he nevertheless has a right that race not be used as a standard, no matter how well a racial classification might work to promote the general welfare or to reduce social and economic inequality. He does not claim, however, that he has this right as a distinct and independent political right that is protected by the Constitution specifically, like his right to freedom of speech and religion. The Constitution does not condemn racial classification directly, as it does condemn censorship or the establishment of a religion. DeFunis claims that his right that race not be used as a criterion of admission follows from the more abstract right of equality that is protected by the Fourteenth Amendment, which provides that no state shall deny to any person the equal protection of the law.
But the legal arguments made on both sides show that neither the text of the Constitution nor the prior decisions of the Supreme Court decisively settle the question whether, as a matter of law, the equal protection clause makes all racial classifications unconstitutional. The clause makes the concept of equality a test of legislation, but it does not stipulate any particular conception of that concept.3 Those who wrote the clause intended to attack certain consequences of slavery and racial prejudice, but it is unlikely that they intended to outlaw all racial classifications, or that they expected that this would be the result of what they wrote. They outlawed whatever policies would violate equality, but left it to others to decide, from time to time, what that means. There cannot be a good legal argument in favor of DeFunis, therefore, unless there is a good moral argument that all racial classifications, even those that make society as a whole more equal, are inherently offensive to an individual’s right to equal protection for himself.
There is nothing paradoxical, of course, in the idea that an individual’s right to equal protection may sometimes conflict with an otherwise desirable social policy, including the policy of making the community more nearly equal overall. Suppose a law school were to charge a few middle-class students, selected by lot, double tuition in order to increase the scholarship fund for poor students. It would be serving a desirable policy—equality of opportunity—by means that violated the right of the students selected by lot to be treated equally with other students who could also afford the increased fees. It is, in fact, part of the importance of DeFunis’s case that it forces us to acknowledge the distinction between equality as a policy and equality as a right, a distinction that political theory has virtually ignored. He argues that the University of Washington violated his individual right to equality for the sake of a policy of greater equality overall, in the same way that double tuition for arbitrarily chosen students would violate their rights for the same purpose.
We must therefore concentrate our attention on that claim. We must try to define the central concept on which it turns, which is the concept of an individual right to equality made a constitutional right by the equal protection clause. What rights to equality do citizens have as individuals which might defeat programs aimed at important economic and social policies, including the social policy of improving equality overall?
There are two different sorts of rights they may be said to have. The first is the right to equal treatment, which is the right to an equal distribution of some opportunity or resource or burden. Every person, for example, has a right to an equal vote in a democracy; that is the nerve of the Supreme Court’s decision that one man must have one vote even if a different and more complex arrangement would better secure the collective welfare. The second is the right to treatment as an equal, which is the right, not to receive the same distribution of some burden or benefit, but to be treated with the same respect and concern as anyone else. If I have two children, and one is dying from a disease that is making the other uncomfortable, I do not show equal concern if I flip a coin to decide which should have the remaining dose of a drug. This example shows that the right to treatment as an equal is fundamental, and the right to equal treatment derivative. In some circumstances the right to treatment as an equal will entail a right to equal treatment, but not, by any means, in all circumstances.
DeFunis does not have a right to equal treatment in the assignment of law school places; he does not have a right to a place just because others are given places. Individuals may have a right to equal treatment in elementary education, because someone who is denied elementary education is unlikely to lead a useful life. But legal education is not so vital that everyone has an equal right to be admitted.
DeFunis does have the second sort of right, that is, a right to treatment as an equal in the decision over which admissions standards should be used. He has a right that his interests be treated as fully and sympathetically as the interests of any others when the law school decides whether to count race as a pertinent criterion for admission. But we must be careful not to overstate what that means.
Suppose an applicant complains that his right to be treated as an equal is violated by tests that place the less intelligent candidates at a disadvantage against the more intelligent. A law school might properly reply in the following way. Any standard will place certain candidates at a disadvantage as against others; but an admissions policy may nevertheless be justified if it seems reasonable to expect that the overall gain to the community exceeds the overall loss, and if no other policy that does not provide a comparable disadvantage would produce even roughly the same gain. An individual’s right to be treated as an equal means that his potential loss must be treated as a matter of concern; but that loss may nevertheless be outweighed by the gain to the community as a whole. If it is, then the less intelligent applicant cannot claim that he is cheated of his right to be treated as an equal just because he suffers a disadvantage that others do not.
The University of Washington may make the same reply to DeFunis. Any admissions policy must put some applicants at a disadvantage, and a policy of preference for minority applicants can reasonably be supposed to benefit the community as a whole even when the loss to candidates like DeFunis is taken into account. If there are more black lawyers they will help to provide better legal services to the black community, and so reduce social tensions. It might well improve the quality of legal education for all students, moreover, to have a greater number of blacks taking part in classroom discussions of social problems. If blacks are seen as successful law students, further, this might well encourage other blacks to apply who do meet the ordinary intellectual standards, and that, in turn, would raise the intellectual quality of the bar. In any case, preferential admissions of blacks should decrease the difference in wealth and power that now exists between different racial groups, and so make the community more equal overall.
It is, as I said, controversial whether a preferential admissions program will in fact promote these various policies, but it cannot be said to be implausible that it will. The disadvantage to applicants like DeFunis is, on that hypothesis, a cost that must be paid for a greater gain; it is in that way like the disadvantage to less intelligent students that is the cost of ordinary admissions policies. 4
We now see the difference between DeFunis’s case and the case we imagined, in which a law school charged students selected at random higher fees. The special disadvantage to these students was not necessary to achieve the gain in scholarship funds, because the same gain would have been achieved by a more equal distribution of the cost among all the students who could afford it. That is not true of DeFunis. He did suffer from the Washington policy more than those majority applicants who were accepted. But that discrimination was not arbitrary; it was a consequence of the meritocratic standards he approves. DeFunis’s argument therefore fails. The equal protection clause gives constitutional standing to the right to be treated as an equal, but he cannot find, in that right, any support for his claim that the clause makes all racial classifications illegal.
If we dismiss DeFunis’s claim in this straightforward way, however, we are left with this puzzle. How can so many able lawyers, who supported his claim both in morality and law, have made that mistake? These lawyers all agree that intelligence is a proper criterion for admission to law schools. They do not suppose that anyone’s constitutional right to be treated as an equal is compromised by that criterion. Why do they deny that race, in the circumstances of this decade, may also be a proper criterion?
They fear, perhaps, that racial criteria will be misused; that such criteria will serve as an excuse for prejudice against the minorities that are not favored, like Jews. But that cannot explain their opposition. Any criteria may be misused, and in any case they think that racial criteria are wrong in principle, and not simply open to abuse.
Why? The answer lies in their belief that, in theory as well as in practice, DeFunis and Sweatt must stand or fall together. They believe that it is illogical for liberals to condemn Texas for raising a color barrier against Sweatt, and then applaud Washington for raising a color barrier against DeFunis. The difference between these two cases, they suppose, must be only the subjective preference of liberals for certain minorities now in fashion. If there is something wrong with racial classifications, then it must be something that is wrong with racial classifications as such, not just classifications that work against those groups currently in favor. That is the unarticulated premise behind the slogan, relied on by defendants of DeFunis, that the Constitution is color blind. That slogan means, of course, just the opposite of what it says: it means that the Constitution is so sensitive to color that it makes any institutional racial classification invalid as a matter of law.
It is of the greatest importance, therefore, to test the assumption that Sweatt and DeFunis must stand or fall together. If that assumption is sound, then the straightforward argument against DeFunis must be fallacious after all, for no argument could convince us that segregation of the sort practiced against Sweatt is justifiable or constitutional.5 Superficially, moreover, the arguments against DeFunis do indeed seem available against Sweatt, because we can construct an argument that Texas might have used to show that segregation benefits the collective welfare, so that the special disadvantage to blacks is a cost that must be paid to achieve an overall gain.
Suppose the University of Texas admissions committee, though composed of men and women who themselves held no prejudice, decided that the Texas economy demanded more white lawyers than they could educate, but could find no use for black lawyers at all. That might have been, after all, a realistic assessment of the commercial market for lawyers in Texas just after the war. Corporate law firms needed lawyers to serve booming business, but could not afford to hire black lawyers, however skillful, because the firms’ practice would be destroyed if they did. It was no doubt true that the black community in Texas had great need of skillful lawyers, and would have preferred to use black lawyers if these were available. But the committee might well have thought that the commercial needs of the state as a whole outweighed that special need.
Or suppose the committee judged, no doubt accurately, that alumni gifts to the law school would fall off drastically if it admitted a black student. The committee might deplore that fact but nevertheless believe that the consequent collective damage would be greater than the damage to black candidates excluded by the racial restriction.
It may be said that these hypothetical arguments are disingenuous, because any policy of excluding blacks would in fact be supported by a prejudice against blacks as such, and arguments of the sort just described would be rationalization only. But if these arguments are in fact sound, then they might be accepted by men who do not have the prejudices the objection assumes. It therefore does not follow from the fact that the admissions officers were prejudiced, if they were, that they would have rejected these arguments if they had not been.
In any case, arguments like those I describe were in fact used by officials who might have been free from prejudice against those they excluded. Many decades ago, as the late Professor Bickel reminds us in his brief for the B’nai B’rith, President Lowell of Harvard University argued in favor of a quota limiting the number of Jews who might be accepted by his university. He said that if Jews were accepted in numbers larger than their proportion of the population, as they certainly would have been if intelligence were the only test, then Harvard would no longer be able to provide to the world men of the qualities and temperament it aimed to produce, men, that is, who were more well rounded and less exclusively intellectual than Jews tended to be, and who, therefore, were better and more likely leaders of other men, both in and out of government.
It was no doubt true, when Lowell spoke, that Jews were less likely to occupy important places in government or at the heads of large public companies. If Harvard wished to serve the general welfare by improving the intellectual qualities of the nation’s leaders, it was rational not to allow its classes to be filled up with Jews, even though the men who reached that conclusion might well prefer the company of Jews to that of the Wasps who were more likely to become senators. (Lowell suggested he did, though perhaps the responsibilities of his office prevented him from frequently indulging his preference.)
It might now be said, however, that discrimination against blacks, even when it does serve some plausible policy, is nevertheless unjustified because it is invidious and insulting. The briefs opposing DeFunis make just that argument to distinguish his claim from Sweatt’s. Because blacks were the victims of slavery and legal segregation, they say, any discrimination that excludes blacks will be taken as insulting by them, whatever arguments of general welfare might be made in its support. But it is not true, as a general matter, that any social policy is unjust if those whom it puts at a disadvantage feel insulted. Admission to law school by intelligence is not unjust because those who are less intelligent feel insulted by their exclusion. Everything depends upon whether the feeling of insult is produced by some more objective feature that would disqualify the policy even if the insult were not felt. If segregation does improve the general welfare, even when the disadvantage to blacks is fully taken into account, and if no other reason can be found why segregation is nevertheless unjustified, then the insult blacks feel, while understandable, must be based on misperception.
It would be wrong, in any case, to assume that people in the position of DeFunis will not take their exclusion to be insulting. They are very likely to think of themselves as members of some other minority, like Jews or Poles or Italians, whom comfortable and successful liberals are willing to sacrifice in order to delay more violent social change. If we wish to distinguish DeFunis from Sweatt on some argument that uses the concept of an insult, we must show that the treatment of the one, but not of the other, is in fact unjust.
So these familiar arguments that might distinguish the two cases are unconvincing. That seems to confirm the view that Sweatt and DeFunis must be treated alike, and therefore that racial classification must be outlawed altogether. But fortunately a more successful ground of distinction can be found to support our initial sense that the cases are in fact very different. This distinction does not rely, as these unconvincing arguments do, on features peculiar to issues of race or segregation, or even on features peculiar to issues of educational opportunity. It relies instead on further analysis of the idea, which was central to my argument against DeFunis, that in certain circumstances a policy that puts many individuals at a disadvantage is nevertheless justified because it makes the community as a whole better off.
Any institution that uses that idea to justify a discriminatory policy faces a series of theoretical and practical difficulties. There are, in the first place, two distinct senses in which a community may be said to be better off as a whole in spite of the fact that certain of its members are worse off, and any justification must specify which sense is meant. It may be better off in a utilitarian sense, that is, because the average or collective level of welfare in the community is improved even though the welfare of some individuals falls. Or it may be better off in an ideal sense, that is, because it is more just, or in some other way closer to an ideal society, whether or not average welfare is improved. The University of Washington might use either utilitarian or ideal arguments to justify its racial classification. It might argue, for example, that increasing the number of black lawyers reduces racial tensions, which improves the welfare of almost everyone in the community. That is a utilitarian argument. Or it might argue that, whatever effect minority preference will have on average welfare, it will make the community more equal and therefore more just. That is an ideal, not a utilitarian, argument.
The University of Texas, on the other hand, cannot make an ideal argument for segregation. It cannot claim that segregation makes the community more just whether it improves the average welfare or not. The arguments it makes to defend segregation must therefore all be utilitarian arguments. The arguments I invented, like the argument that white lawyers could do more than black lawyers to improve commercial efficiency in Texas, are utilitarian, since commercial efficiency makes the community better off only if it improves average welfare.
Utilitarian arguments encounter a special difficulty that ideal arguments do not. What is meant by average or collective welfare? How can the welfare of an individual be measured, even in principle, and how can gains in the welfare of different individuals be added and then compared with losses, so as to justify the claim that gains outweigh losses overall? The utilitarian argument that segregation improves average welfare presupposes that such calculations can be made. But how?
Jeremy Bentham, who believed that only utilitarian arguments could justify political decisions, gave the following answer. He said that the effect of a policy on an individual’s welfare could be determined by discovering the amount of pleasure or pain the policy brought him, and that the effect of the policy on the collective welfare could be calculated by adding together all the pleasure and subtracting all of the pain it brought to everyone. But, as Bentham’s critics insisted, it is doubtful whether there exists a simple psychological state of pleasure common to all those who benefit from a policy, or of pain common to all those who lose by it; in any case it would be impossible to identify, measure, and add the different pleasures and pains felt by vast numbers of people.
Philosophers and economists who find utilitarian arguments attractive, but who reject Bentham’s psychological utilitarianism, propose a different concept of individual and overall welfare. They suppose that whenever an institution or an official must decide upon a policy, the members of the community will each prefer the consequences of one decision to the consequences of others. DeFunis, for example, prefers the consequences of the standard admissions policy to the policy of minority preference the University of Washington used, while the blacks in some urban ghetto might each prefer the consequences of the latter policy to the former. If it can be discovered what each individual prefers, and how intensely, then it might be shown that a particular policy would satisfy on balance more preferences, taking into account their intensity, than alternative policies. On this concept of welfare, a policy makes the community better off in a utilitarian sense if it satisfies the total collection of preferences better than alternative policies would, even though it does not satisfy the preferences of some.6
Of course, a law school does not have available any means of making accurate judgments about the preferences of all those whom its admissions policies will affect. It may nevertheless make judgments which, though speculative, cannot be dismissed as implausible. It is, for example, plausible to think that in postwar Texas the preferences of the people were mostly in favor of the consequences of segregation in law schools, even if the intensity of the competing preference for integration, and not simply the number of those holding that preference, is taken into account. The officials of the Texas law school might have relied upon voting behavior, newspaper editorials, and simply their own sense of their community in reaching that decision and, though they might have been wrong, we cannot now say, even with the benefit of hindsight, that they were.
So even if Bentham’s psychological utilitarianism is rejected, law schools may—and in fact do—appeal to preference utilitarianism to provide at least a rough and speculative justification for admissions policies that put some classes of applicants at a disadvantage. But once it is made clear that these utilitarian arguments are based on judgments about the actual preferences of members of the community, a fresh and much more serious difficulty emerges.
The utilitarian argument, that a policy is justified if it satisfies more preferences overall, seems at first sight to be an egalitarian argument. It seems to observe strict impartiality. If the community has only enough medicine to treat some of those who are sick, the argument seems to recommend that those who are sickest be treated first. If the community can afford a swimming pool or a new theater, but not both, and more people want the pool, then it recommends that the community build the pool, unless those who want the theater can show that their preferences are so much more intense that they have more weight in spite of the numbers. One sick man is not to be preferred to another because he is worthier of official concern; the tastes of the theater audience are not to be preferred because they are more admirable. In Bentham’s phrase, “each to count for one and none for more than one.”
These simple examples suggest that the utilitarian argument not only respects, but embodies, the right of each citizen to be treated as the equal of any other. The chance each individual’s preferences have to succeed, in the competition for different social policies, will depend upon how important his preference is to him, and how many others share it, compared to the intensity and number of competing preferences. His chance will not be affected by the esteem or contempt of either officials or fellow citizens, and he will therefore not be subservient or beholden to them.
But if we examine the range of preferences that individuals in fact have, we shall see that the apparent egalitarian character of a utilitarian argument is often deceptive. Preference utilitarianism asks officials to attempt to satisfy people’s preferences so far as this is possible. But the preferences of an individual for the consequences of a particular policy may be seen to reflect, on further analysis, either a personal preference for his own enjoyment of some goods or opportunities, or an external preference for the assignment of goods and opportunities to others, or both. A white law school candidate might have a personal preference for the consequences of segregation, for example, because the policy improves his own chances of success. But he may also have an external preference for those consequences because he has contempt for blacks, disapproves of social situations in which the races mix.
The distinction between personal and external preferences is of great importance for this reason. If a utilitarian argument counts external preferences along with personal preferences, then the egalitarian character of that argument is corrupted. The chance that anyone’s preferences have to succeed will then depend not only on the demands made by the personal preferences of others on scarce resources but on the respect or affection they have for him or for his way of life. If external preferences tip the balance, then the fact that a policy makes the community better off in a utilitarian sense would not provide a justification compatible with the right of those it disadvantages to be treated as equals.
This corruption of utilitarianism is plain when some people have external preferences because they hold political theories that are themselves contrary to utilitarianism. Suppose many citizens, who are not themselves sick, are racists in political theory, and therefore prefer that scarce medicine be given to a white man who needs it rather than a black man who needs it more. If officials following a utilitarian theory count these political preferences at face value, then the theory will be, from the standpoint of personal preferences, self-defeating, because the distribution of medicine will not be, from that standpoint, utilitarian at all. It would be utilitarian, from that standpoint, only if the preferences of people concerning their own treatment when sick were counted. In any case, self-defeating or not, the distribution will not be egalitarian in the sense defined. Blacks will suffer, to a degree that depends upon the strength of the racist preference, from the fact that others think them less worthy of respect and concern.
There is a similar corruption when the external preferences that are counted are altruistic or moralistic. Suppose many citizens, who themselves do not swim, prefer a new pool to a new theater because they approve of sports and admire athletes, or because they think that the theater is often immoral and ought to be repressed. If such altruistic preferences are counted, so as to reinforce the personal preferences of swimmers, the result will be a form of double counting: each person who wants to swim in a new pool will have the benefit not only of his own preference but also of the preference of someone else who takes pleasure in his being able to do so. If the moralistic preferences are counted, the effect will be the same: actors and audiences will suffer because their preferences are held in lower respect by citizens whose personal preferences are not themselves engaged.
In these examples, external preferences are independent of personal preferences. But of course political, altruistic, and moralistic preferences are often not independent, but grafted on to the personal preferences they reinforce. If I am white and sick, I may also hold a racist political theory. If I want a swimming pool for my own enjoyment I may also be altruistic in favor of my fellow athlete, or I may also think that the theater is immoral. The consequences of counting these external preferences will be as grave for equality as if they were independent of personal preference, because those against whom the external preferences run might be unable or unwilling to develop reciprocal external preferences that would right the balance.
External preferences therefore present a great difficulty for utilitarianism. That theory owes much of its popularity to the assumption that it embodies the right of citizens to be treated as equals. But if external preferences are counted in overall preferences, then this assumption is jeopardized. That is in itself an important and neglected point in political theory; it bears, for example, on the liberal thesis, first made prominent by Mill, that the government has no right to enforce popular morality at law. It is often said that this liberal thesis is inconsistent with utilitarianism, because if the preferences of the majority that homosexuality should be repressed, for example, are sufficiently strong, utilitarianism must give way to their wishes. But the preference against homosexuality is an external preference, and the present argument provides a general reason why utilitarians should not count external preferences of any form. If utilitarianism is suitably reconstituted so as to count only personal preferences, then the liberal thesis is a consequence, not an enemy, of that theory.
It is not, however, always possible to reconstitute a utilitarian argument so as to count only personal preferences. Sometimes personal and external preferences are so inextricably tied together, and so mutually dependent, that no practical test for measuring preferences will be able to discriminate among the personal and external elements in any individual’s overall preference. That is especially true when preferences are affected by prejudice. Consider, for example, the preference of a white law student to associate only with white classmates. This may be said to be a personal preference for an association with one kind of colleague rather than another. But it is a personal preference that tends to be parasitic upon external preferences: a white student often prefers the company of other whites because he has racist social and political convictions, or because he has contempt for blacks as a group.
If such associational preferences are counted in a utilitarian argument used to justify segregation, then the egalitarian character of the argument is destroyed just as if the underlying external preferences were counted directly. Blacks would be denied their right to be treated as equals because the chance that their preferences would prevail in the design of admissions policy would be crippled by the low esteem in which others hold them. In any community in which prejudice against a particular minority is strong, then the personal preferences upon which a utilitarian argument must fix will be saturated with that prejudice; it follows that in such a community no utilitarian argument purporting to justify a disadvantage to that minority can be fair.7
This final difficulty is therefore fatal to Texas’s utilitarian arguments in favor of segregation. The preferences that might support any such argument are either distinctly external, like the preferences of the community at large for racial separation, or are inextricably combined with and dependent upon external preferences, like the preferences of white students for associating with white classmates and of white lawyers for associating with white colleagues. These external preferences are so widespread that they must corrupt any such argument. Texas’s claim that segregation makes the community better off in a utilitarian sense is therefore incompatible with Sweatt’s right to treatment as an equal guaranteed by the equal protection clause.
It does not matter to this conclusion whether external preferences figure in the justification of a fundamental policy, or in the justification of derivative policies designed to advance a more fundamental policy. Suppose Texas justifies segregation by pointing to the apparently neutral economic policy of increasing community wealth, which could be said to satisfy the personal preferences of everyone for better homes, food, and recreation. If the argument that segregation will improve community wealth depends upon the fact of external preference; if the argument notices, for example, that because of prejudice industry will run more efficiently if factories are segregated; then the argument has the consequence that the black man’s personal preferences are defeated by what others think of him.
Utilitarian arguments that justify a disadvantage to members of a race against whom prejudice runs will always be unfair arguments, unless it can be shown that the same disadvantage would have been justified in the absence of the prejudice. If the prejudice is widespread and pervasive, as in fact it is in the case of blacks, that can never be shown. The preferences on which any economic argument justifying segregation must be based will be so intertwined with prejudice that they cannot be disentangled to the degree necessary to make any such contrary-to-fact hypothesis plausible.
We now have an explanation that shows why any form of segregation that disadvantages blacks is, in the United States, an automatic insult to them, and why such segregation offends their right to be treated as equals. The argument confirms our sense that utilitarian arguments purporting to justify segregation are not simply wrong in detail but misplaced in principle. This objection to utilitarian arguments is not, however, limited to race or even prejudice. There are other cases in which counting external preferences would offend the rights of citizens to be treated as equals and it is worth briefly noticing these, if only to protect the argument against the charge that it is constructed ad hoc for the racial case. I might have a moralistic prejudice against professional women, or an altruistic preference for what I take to be “virtuous” men. It would be unfair for any law school to count preferences like these in deciding whom to admit to law schools; unfair because these preferences, like racial prejudices, make the success of the personal preferences of an applicant depend on the esteem and approval, rather than on the competing personal preferences, of others.
The same objection does not hold, however, against a utilitarian argument used to justify admission based on intelligence. That policy need not rely, directly or indirectly, on any community sense that intelligent lawyers are intrinsically more worthy of respect. It relies instead upon the law school’s own judgment, right or wrong, that intelligent lawyers are more effective in satisfying personal preferences of others, like the preference for increasing wealth or winning law suits.
We therefore have the distinctions in hand necessary to distinguish DeFunis from Sweatt. The arguments for an admissions program that discriminates against blacks are all utilitarian arguments, and they are all utilitarian arguments that rely upon external preferences in such a way as to offend the constitutional right of blacks to be treated as equals. The arguments for an admissions program that discriminates in favor of blacks are both utilitarian and ideal. Some of the utilitarian arguments do rely, at least indirectly, on external preferences, like the preference of certain blacks for lawyers of their own race; but the utilitarian arguments that do not rely on such preferences are strong and may be sufficient. The ideal arguments do not rely upon preferences at all, but on the independent argument that a more equal society is a better society even if its citizens prefer inequality. That argument does not deny anyone’s right himself to be treated as an equal.
We are therefore left, in DeFunis, with the simple and straightforward argument with which we began. Racial criteria are not necessarily the right standards for deciding which applicants should be accepted by law schools. But neither are intellectual criteria, or indeed, any other set of criteria. The fairness—and constitutionality—of any admissions program must be tested in the same way. It is justified if it serves a plausible policy that respects the right of all members of the community to be treated as equals, but not otherwise. The criteria used by schools that refused to consider blacks fail that test, but the criteria used by Washington do not.
We are all rightly suspicious of racial classifications. They have been used to deny rather than to respect the right of equality, and we are all conscious of the consequent injustice. But if we misunderstand the nature of that injustice because we do not make the simple distinctions that are necessary to understand it, then we are in danger of more injustice still. It may be that preferential admissions programs will not in fact make a more equal society. They may not have the effects their advocates believe they will. That strategic question should be at the center of debate about these programs. But we must not corrupt the debate by supposing that these programs are unfair even if they do work. We must take care not to use the equal protection clause to cheat ourselves of equality.
February 5, 1976
Sweatt v. Painter, 339 US 629, 70 S.Ct. 848. ↩
DeFunis v. Odegaard, 94 S.Ct. 1704 (1974). ↩
See my “The Jurisprudence of Richard Nixon,” The New York Review of Books, May 4, 1972. ↩
I shall argue later in this essay that there are circumstances in which a policy violates someone’s right to be treated as an equal in spite of the fact that the social gains from that policy may be said to outweigh the losses. These circumstances arise when the gains that outweigh the losses include the satisfaction of prejudices and other sorts of preferences that it is improper for officials or institutions to take into account at all. But the hypothetical social gains described in this paragraph do not include gains of that character. Of course, if DeFunis had some other right, beyond the right to be treated as an equal, which the University of Washington’s policy violated, then the fact that the policy might achieve an overall social gain would not justify the violation. (See my “Taking Rights Seriously,” The New York Review of Books, December 17, 1970.) If the Washington admissions procedures included a religious test that violated his right to religious freedom, for example, it would offer no excuse for using such a test that it might make the community more cohesive. But DeFunis does not rely on any distinct right beyond his right to equality protected by the equal protection clause. ↩
In the actual Sweatt decision, the Supreme Court applied the old rule which held that segregation was constitutionally permitted if facilities were provided for blacks that were “separate but equal.” Texas had provided a separate law school for blacks, but the Court held that that school was by no means the equal of the white school. Sweatt was decided before the famous Brown case in which the Court finally rejected the “separate but equal” rule, and there is no doubt that an all-white law school would be unconstitutional today even if a separate black school provided facilities that were in a material sense the equal of those provided for whites. ↩
Many economists and philosophers challenge the intelligibility of preference utilitarianism as well as psychological utilitarianism. They argue that there is no way, even in principle, to calculate and compare the intensity of individual preferences. Since I wish to establish a different failing in certain utilitarian arguments, I assume, for purposes of this essay, that at least rough and speculative calculations about overall community preferences can be made. ↩
The argument of this paragraph is powerful but it is not, in itself, sufficient to disqualify all utilitarian arguments that produce substantial disadvantages to minorities who suffer from prejudice. Suppose the government decides, on a utilitarian argument, to allow unemployment to increase because the loss to those who lose their jobs is outweighed by the gain to those who would otherwise suffer from inflation. The burden of this policy will fall disproportionately on blacks, who will be fired first because prejudice runs against them. But though prejudice in this way affects the consequences of the policy of unemployment, it does not figure, even indirectly, in the utilitarian argument that supports that policy. (It figures, if at all, as a utilitarian argument against it.) We cannot say, therefore, that the special damage blacks suffer from a high unemployment policy is unjust for the reasons described in this essay. It may well be unjust for other reasons; if John Rawls is right, for example, it is unjust because the policy improves the condition of the majority at the expense of those already worse off. ↩