Colin S. Diver’s speeches, letters, and articles
University of Illinois, October 2, 2003
From equality to diversity: The detour from Brown to Grutter
Remarks of Reed College President Colin Diver
As interpreted by the Supreme Court, the Equal Protection Clause is a doctrine about the relation of ends and means. Whenever the government uses a classifying criterion to allocate burdens or distribute benefits, that criterion must fit reasonably tightly with the goal of the governmental program at issue. And whenever the government uses a so-called "suspect" classification-such as race-its goal must be particularly important ("compelling") and the means chosen must be particularly tightly fitted ("narrowly tailored") to achieve that compelling goal. Applying this test is excruciatingly difficult in practice because human beings and human institutions rarely define ends without referring to the means chosen to effectuate them. Not surprisingly, those whose practices are attacked for violating the equal protection standard try to couch the interests being pursued in a way that makes seem virtually inevitable the selection of the method they have chosen, to attain those interests. The job of the attacker, conversely, is to formulate the ends in a sufficiently general way as to suggest that the actor could just as well have chosen a variety of alternative methods.
A. Diversity as a goal
The problem is well illustrated in the context of trying to decide whether an interest labeled "diversity" can be used to justify racial selection in admissions. In her opinion for the Grutter majority, Justice O'Connor variably characterizes the state's interest as: "obtaining ï¿½the educational benefits that flow from a diverse student body'"; "attaining a diverse student body"; and "ï¿½assembling a class that is both exceptionally academically qualified and broadly diverse'." In two of these formulations, she quotes from the university's brief. It is, of course, in the university's interest to keep the word "diversity" in the definition of the goal or interest to be served. Such a formulation makes the use of race in the selection process more readily justifiable than if the interest to be achieved were characterized as the attainment of some set of general educational outcomes, such as transmitting some body of knowledge or improving some set of cognitive skills.
But what does it mean to have an interest in, or goal of, attaining "diversity," as in the expression "a diverse student body"? In purely descriptive terms, the unadorned term "diversity" conveys nothing other than a vague sense of variety. One might say that human beings are unique; so in that sense every group of people is equally and maximally diverse. One might equally say that human beings are all different from, say trees or ostriches; so in that sense all groups of people are equally nondiverse. Between those two tautological extremes, one cannot even begin to decide whether a given group of people is "diverse," without solving a number of problems.
First, one must specify the dimensions or characteristics that are relevant to the inquiry at hand. Imagine a group of three people with combined scores on the SAT of 400, 1000, and 1600, respectively. Is the group diverse? Arguably yes in the context of college admission; probably not in the context of a neighborhood touch football game. Second, one must specify what differences matter. Is a group of students with SAT scores of 1400, 1410, and 1420 diverse? Probably not. Third, for discontinuous variables like ethnicity or legacy status, one must specify which characteristics count and which don't. Imagine a group consisting of one Arab American, one Slavic American, and one Polynesian American. Is the group ethnically diverse? Yes, if one has a Western European baseline in mind; not if one is looking only for African Americans, Latinos, or Native Americans. Fourth, one cannot compare people across multiple dimensions without creating a composite scale that combines the multiple dimensions into a single metric. Assume a group consisting of one black, one flute player, and one farm boy. Is it a diverse group? The question is meaningless, because each person has been classified according to a different dimension. All one can say is that the members of the group are different (diverse) along each implicit dimension. But if they all have SAT scores of 1400, or are all neo-conservative, does that enhance or diminish their collective diversity?
Beyond the descriptive problems, there is the normative problem. One cannot say that one values diversity, or has a goal of seeking diversity, without having a reason why it is better to have greater, rather than lesser, variation along or among the relevant dimensions. With regard to many of the criteria conventionally used to rank college applicants, it is hardly obvious that diversity is a virtue. Consider academic ability, as traditionally measured by grade point averages and aptitude test scores, which are widely acknowledged to be the most important criteria for determining eligibility for admission to selective colleges and universities. Elite institutions of higher education seek to enroll an entering class whose academic credentials are not only high, but also uniformly high. This preference for homogeneity reflects a view, widely espoused by instructional faculty, that it is harder to teach a class with widely divergent academic skills than a class with uniform academic skills. For many other commonly applied admissions criteria, such as athletic prowess or legacy status, notions of diversity seem thoroughly beside the point. A school with a strong intercollegiate football program, for example, seeks to attract enough talented players to field a winning team. Beyond that, the school is unlikely to care about the football-playing abilities of its matriculants. Schools that give favorable weight to legacy status do so in the interest of keeping the alumni happy, not to achieve any particular mix of alumni children and nonalumni children.
So, an educational institution cannot coherently justify race-based discrimination (or, for that matter, discrimination based on any criterion) simply by invoking an interest in "diversity" per se. It must specify what kinds of criteria are important to it, how it measures or identifies variability within those criteria, and why greater variability along those dimensions is better than lesser variability. Michigan's solution to this challenge, and the one evidently embraced by the Court in Grutter, is to say, in effect: "We value those forms of diversity that produce better educational outcomes." This strategy has obvious appeal. First, it invokes the understood and accepted mission of the university, which is, after all, to educate (whatever that may mean), not simply to collect people of divergent characteristics, the way a zoo seeks to collect examples of many species. Second, it appeals to our intuition-indeed our conviction-that the quality of an educational experience is enhanced when one receives a wide variety of cognitive inputs into the learning process, and perhaps a wide variety of affective inputs into the developmental process. Finally, and for that reason, this strategy invokes a quasi-constitutional interest-academic freedom-that can receive special weight in the legal calculus.
B. Multidimensional diversity
So far so good. But the defender of academic diversity needs to answer at least two more questions: What are those forms of diversity that provide educational benefits, and in particular what does race-or more precisely, particular racial or ethnic groups-have to do with that diversity? The answer given by Powell in Bakke was to adopt a very catholic definition of educational diversity factors, and to claim that race was merely one among many-just another "plus" factor. "Ethnic diversity," he said, "is only one element in a range of factors a university may properly consider in attaining the goal of a heterogeneous student body." 
Several passages in Justice O'Connor's Grutter opinion suggest that she is invoking a similarly multidimensional definition of diversity. For example, she specifically approves much of Justice Powell's Bakke argument, including his approving citation of the multifactorial Harvard plan. Similarly, O'Connor applauds the Michigan law school plan for "adequately ensur[ing] that all factors that may contribute to student body diversity are meaningfully considered alongside race in admissions decisions," and then cites approvingly the list of nonracial "diversity" factors considered by the law school.
This catholic definition of educational diversity appeals to what Owen Fiss calls the "individualistic ideal" in equality jurisprudence. But, despite its surface appeal, the multidimensional definition of diversity simply cannot solve the conceptual problem. As indicated above, multidimensional diversity is conceptually meaningless without an overarching criterion for comparing people along different dimensions. To use the Michigan law school policy, suppose we have a group with one person each who lived abroad, was fluent in a non-English language, overcame adversity, did extensive community service, and had a successful pre-law career. Is the group meaningfully diverse? If one adds to the group a person who is African American, does that enhance the diversity of the group, in an educationally relevant sense? One can answer those questions only one criterion at a time. It may well be, for example, that having people who speak multiple languages in the class will enhance the education of all. Likewise, a group assembled from many countries could have that affect. And so on. But the mere multiplication of examples of different backgrounds tells us little, at least without some evidence that diversity along multiple dimensions produces even better educational outcomes than diversity along just one or a few.
There is a further problem with Powell's multidimensional view of diversity. If race is constitutionally toxic as a discriminant, why is that toxicity diminished when race is combined with other discriminants? Some poisons lose their toxicity when diluted with other ingredients.; others do not. Which kind of "poison" is racial discrimination? As a generic matter, it seems to me to be an example of the latter. Suppose a prosecutor uses a wide range of criteria, including race, in deciding whom to prosecute. Or a welfare agency uses a wide range of criteria, including race, in deciding whom to aid. Our intuition is that these behaviors violate the equal protection norm, and, I think, properly so. The "diminished toxicity" theory may seem more palatable in the context of higher education admissions, where the relationship between public purpose and some notion of heterogeneity has intuitive plausibility. But, in the end, it seems to me that race cannot be hidden in a halo of other factors even in this context. If race is constitutionally toxic, it has to be confronted on its own terms, as educationally justified or not.
C. Racial diversity
There is much in Justice O'Connor's opinion-unlike Justice Powell's opinion in Bakke-to suggest that she acknowledges this burden. In the section of her opinion identifying the "interest" being served by the law school's use of race in admissions, she is careful to avoid attaching the adjective "racial" to the word "diversity."  But it is plain from the surrounding context that racial diversity is what she is really talking about. In discussing the educational benefit of diversity, she refers almost solely to benefits flowing from racial forms of diversity. For example, she quotes the university's claim that diversity "promotes ï¿½cross-racial understanding,' helps to break down racial stereotypes, and ï¿½enables [students] to better understand persons of different races'." She cites sociological studies of the impact of racial diversity on various educational and vocational outcomes. She quotes a brief submitted by military leaders that addresses quite explicitly the need for a "racially diverse" officer corps. And, perhaps most tellingly, she states that the "legitimacy" of our institutions require that "the path to leadership be visibly open to talented and qualified individuals of every race and ethnicity."
1. Race as Proxy for Viewpoint: What, then, is the connection between racial diversity and educational outcomes? The "robust exchange of ideas" image invoked by Justice Powell suggests that persons of particular races are needed to contribute certain ideas to the exchange. As commentators on all sides of the affirmative action debate have noted, however, his comes dangerously close to the essentialist notion that all blacks think alike or all Latinos think alike or, for that matter, all whites think alike. Such a proposition is not only demonstrably false, but also highly offensive to our constitutional ideals. Justice O'Connor seeks to distance herself from that view by quoting the university's disclaimer of "any belief that minority students always (or even consistently) express some characteristic minority viewpoint on any issue." But she quickly asserts that the experience of being a minority in the country is "likely" to affect one's views and, later, that minority students are "likely" to have experiences of importance to the law school.
As Gary Orfield, a staunch defender of affirmative action, states: "The argument about diversity is only about probabilities." He goes on to recite pages of statistics showing the differential probabilities that members of various racial groups will live in poverty, be unemployed, rent rather than own their home, have little wealth, attend low-achieving schools, be victims of crime, and so forth-all for the purpose of establishing the proposition that whites, blacks, Latinos, and others are likely-but hardly certain-to have differing views on a wide variety of educationally relevant matters.
If race were not a suspect classification, this probabilistic strategy would be unexceptionable-just like the famous example of the farm boy from Idaho in the Harvard admissions blurb quoted by Justice Powell in Bakke. To the guardians at Harvard's gates, the category "farm boy from Idaho" conveyed enough information to suggest an educationally relevant distinction from the category of, say, "northeastern suburbanite." So the category "black," by itself, suggests an educationally relevant distinction from the category "white." The problem is that "black" and "white" are constitutionally suspect classifications, and thus bear a higher burden of justification than "farm boy" or "Idaho"-or, for that matter, "farm boy from Idaho."  If blacks are, say, more likely to live in poverty or live in one-parent families, why can't Michigan achieve its desired diversity by giving preferences for those characteristics? If blacks are more likely to be suspicious of the police, why can't Michigan test for that viewpoint directly, through personal statements, or indirectly by looking for persons who have certain kinds of experiences with the criminal justice system? Michigan might, of course, respond by invoking administrative convenience. At some point, it becomes easier, and more reliable, to use race as a proxy for these other-and educationally more relevant-forms of diversity. The problem with this answer, however, is that administrative convenience, even at its strongest, seems a weak reed with which to combat the moral and constitutional presumption against using racial classifications. Even if it were probabilistically true that blacks were less qualified that whites for certain classes of jobs, we would not permit employers to discriminate against blacks for that reason.
A further weakness in the "race-as-proxy-for-viewpoint" argument is the inconsistent and incomplete way in which Michigan-and its counterparts virtually everywhere-pursues diversity of viewpoint. If Michigan were genuinely interested in maximizing diversity of educationally relevant viewpoints, one would expect it to grant preferences to-or least keep systematic track of-many characteristics other than race. As Peter Schuck has pointed out, for example, one would surely expect diversity-seeking educators to pay attention to the religious beliefs or religious associations of its students in the admissions process. Likewise, one might expect admissions officers to look more actively for ways to assure a dispersion of political viewpoints, especially given the strong preponderance of liberal views at many elite educational institutions. Further, one would expect colleges and universities to give considerably greater weight to various measures of socioeconomic disadvantage or cultural isolation than they typically do. As if to answer this objection, Justice O'Connor lists nonracial "diversity" factors considered by the Michigan Law School, but never tells us the relative number of cases in which those factors actually influence admissions decisions or, more important, the relative weight that those factors receive, as compared to the extremely heavy weight given to race.
2. Race for Its Own Sake: An alternative strategy for justifying the educational benefit of racial diversity is to argue that race is an educationally relevant variable in its own right, rather than as a proxy for something else, such as viewpoint. The argument here is that, like it or not, race is a hugely important social fact of our contemporary society. Because of deeply embedded vestiges of our shameful racial history, our society is deeply divided and categorized by race. Members of minority races view themselves, and believe themselves to be viewed by others, first and foremost as members of their race. Whites, even if they do not see themselves as members of a distinct race, see minorities as different because of their race. Race permeates interpersonal interactions, social patterns, and institutional structures. Race is one of the principal categories by which government agencies, private researchers, and the media try to make sense of social phenomena.
All of this is true, if somewhat overstated. But what does it have to do with the legitimate goals of higher education? One possible argument is that the best way to learn to cross these racial divides in our society is by interacting with people of other races in school. Thus Justice O'Connor talks about the value of Michigan's program in promoting "cross-racial understanding" and helping to "break down racial stereotypes" and enabling students to "better understand persons of different races."
There are in our economy organizations that conduct workshops on promoting "cross-racial understanding." These organizations often use role-playing and other forms of experiential exercises to achieve their purposes. Such organizations can argue convincingly that they need to have a minimal number of participants drawn from each of the races on opposite sides of the divides they seek to bridge. A division of the University of Michigan that ran such programs could thus plausibly argue for race-conscious selection procedures for its workshops.
There are several problems, however, with using this analogy to justify affirmative action as it is conventionally practiced in the admissions programs of undergraduate and professional schools. First, these institutions are not exclusively or even primarily engaged in promoting cross-racial understanding. One need not adopt the derisive tone of Justice Scalia's opinion in Grutter to question how much importance the law school really attaches to this goal, given the lack of evidence that it attempts to mandate any instruction that fosters cross-racial understanding or measure the extent to which students develop such understanding. Ironically, it was at the undergraduate level-whose race-conscious admission program was struck down-that the University of Michigan appears to have made a greater curricular and extracurricular investment in promoting cross-racial understanding. The fact that the law school felt a need to promote cross-racial understanding suggests that efforts to do so at the undergraduate institutions from which it draws its students (including Michigan's undergraduate college) have not had a notably high rate of success. Indeed, there is reason to believe that many affirmative action programs do more to retard than to promote cross-racial understanding. Studies of race-conscious affirmative action programs have consistently found that they perpetuate or even exacerbate negative racial stereotypes among majority group members. This may explain the somewhat surprising finding of Stanley Rothman and his co-authors that an increase in a school's racial diversity correlated with a decrease in the quality of its educational program, as evaluated by that school's faculty and students.
A second problem with the analogy to a racial-understanding workshop is that it does not tell us which racial divides need to be bridged, and therefore which racial mixes need to be engineered. Why does Michigan explicitly favor the admission only of blacks, Latinos, and Native Americans, but not, say, Arabs or Slavs or Pacific Islanders? Why does it lump Latinos together, or Native Americans together, when in fact those labels embrace a very wide variety of disparate subgroups? And why does Michigan consistently admit such different numbers and percentages of blacks, Latinos, and Native Americans? An antiseptic-that is, remedially blind-diversity rationale provides no basis for an answer to these questions.
One basis for an answer suggested by the literature is that white students are much more likely to have grown up and gone to school in environments in which there are few, if any, blacks, Latinos, or Native Americans. Whites, more than members of these groups, live segregated lives before college and therefore are the primary beneficiaries of multiracial experiences in college. I have always found this argument to be deeply troubling. It says, in effect, that minorities are being recruited to these schools for the educational benefit of whites. That asymmetry would be troubling enough, but it seems to me especially disturbing when one considers that the minority students recruited for this purpose are typically admitted with academic credentials substantially lower, on average, than those of the white students. Not surprisingly, the academic performance of these minority students typically lags far behind that of their white peers. Worse yet, there is evidence that minority students at these schools "underperform" academically-that is, their academic performance is lower than the performance of whites with comparable entry credentials.
The fact that race-conscious affirmative action programs have produced a systematic "mismatch" between the levels of ability and preparation of minority students and the institutions that they attend is not necessarily a basis for scrapping such programs. But the mismatch is certainly a high a price to pay for whatever racial sensitivity training is being given to the white students. If minority students with lower academic credentials are to be recruited for the education of a group of academically better-prepared whites, the colleges ought at least to give full disclosure to the minority students, and arguably should pay them for their services.
3. Race and "Legitimacy": Justice O'Connor makes another argument in Grutter to explain why race might be seen as a primary educational variable, not simply a proxy for other educational benefits. This is her argument from legitimacy. She discusses with apparent endorsement the assertion made in the military leaders' amicus brief that a racially diverse officer corps is necessary to maintaining morale. A somewhat similar point is made in the corporate leaders' brief about the need for a diverse workforce to deal with a diverse customer and client base. These statements carry the implication that the legitimacy of institutions, in the eyes of those who work for them or who deal with them, is a function of the racial composition of their leadership or their membership. If this is not a legitimization of racism, it is surely a legitimization of racialism. Arguments of exactly this sort were made to defend the regime of Jim Crow torn down by Brown and its progeny. Whites, it was said repeatedly, will not work for blacks; whites will not fraternize establishments occupied by blacks; whites will not do business with blacks. These statements were, as sociological propositions, true enough. But as constitutional arguments, they were brusquely-and quite correctly-rejected by the courts. Can similar arguments now be used to defend the use of race in another context? Perhaps so, but not, it seems to me, on the grounds of diversity for diversity's sake.
O'Connor tries to explain her "legitimacy" argument by saying: "it is necessary that the path to leadership be visibly open to talented and qualified individuals of every race and ethnicity." The word "visibly" bears a great deal of weight in that sentence. Without it, the sentence would seem to require only that schools like Michigan refrain from discriminating against historically disfavored races, not that they affirmatively favor them. The addition of the word "visibly" implies that the credibility of the antidiscrimination policy depends on whether members of those races are actually-and, for that matter, "visibly"-hired. This claim comes perilously close to adopting the "role model" theory firmly rejected by the Court-and by Justice O'Connor-in Wygant. If a school board may not use racially selective criteria for hiring black teachers to serve as role models for black students, can law schools justify the use of racially selective admissions criteria to help law firms and corporations demonstrate the sincerity of their antidiscrimination policies? A further weakness in O'Connor's "visibility" argument is that the persons visibly hired must be "talented and qualified." One of the most damning arguments against affirmative action is that it casts a suspicion of lesser competence over all members of the benefited races. O'Connor never even acknowledges, much less responds to that common argument, nor the argument that minorities would perform better-even "visibly" better-if they attended schools at which their academic skills better matched that of their non-minority counterparts. There are answers to these arguments, indeed persuasive ones, but the "legitimacy" defense of race-conscious admissions surely cannot rest simply on O'Connor's ipse dixit.
D. Academic freedom
Finally, a word must be said about the burden of proof borne by the proponent of the proposition that educational diversity justifies race-conscious admissions programs. If that proposition is to be judged by the "rational relation" test-applied by the courts to resolve equal protection challenges to non-suspect classifications-it must surely be upheld. Both sociological studies and personal experience of most educators indicate that racial diversity can produce educational benefits. An educational institution would be free, under the rational relation test, to pursue this particular kind of diversity even if it does not systematically pursue other kinds of diversity. If, on the other hand, the test to be applied is the "strict scrutiny" standard-at least as it has been traditionally understood-- it seems to me that the proposition fails. The diversity interest is too vaguely formulated and too weakly supported in the educational programs of most schools to comprise a truly "compelling" state interest. And the particular criteria used in most race-conscious admissions programs are too indiscriminant, selective, and heavily weighted to satisfy the requirement of "narrow tailoring."
Seeming to recognize this fact, Justice O'Connor, like Justice Powell before her, invokes academic freedom as a basis for lowering the heavy burden of proof that race-based classifications customarily bear. The quasi-constitutional value of promoting the "robust exchange of ideas" and other claimed educational benefits of racial diversity confers on educational institutions a protective mantle of deference. If educators claim educational benefits, the Court is apparently disposed to believe them. This is not implausible. We educators like to think we know what we are doing. And we like to think that we are capable of surmounting tendencies toward self-interest, biased judgment, preoccupation with appearances, herd behavior, and the countless other characteristics that often distort human and organizational decisionmaking. Further, we take solace in the fact that race-conscious admissions policies have stood the test of time. They have been practiced widely for about 30 years, fine-tuned and adjusted along the way. They seems to work. They have met the market test.
But should the Court defer to educators on a matter like racial discrimination? In the four cases consolidated under Brown, the defendant school districts, invoking the opinions of "leading sociologists and educators," argued forcefully that segregation produced the best educational outcomes. The Supreme Court did not deign even to dignify those arguments with a rebuttal. And much more recently, when the Court declared unconstitutional the exclusion of women from Virginia Military Institute, it did not defer to VMI's invocation of the educational benefits of separate-sex military education.
One might argue, with Justice Ginsburg, that deference to academic judgment is more warranted when a school is using a suspect classification for inclusive, rather than exclusive, reasons. But hers is an explicitly remedial justification for race-based criteria-one in which the concepts of "inclusion" and "exclusion" have at least the possibility of deriving some coherence from history. Those concepts get virtually no traction in an argument from diversity. Diversity is all about heterogeneity, which implies engaging in both inclusion and exclusion simultaneously, so as to achieve a desired mix. Admitting more blacks and Hispanics necessarily means admitting fewer whites and Asian Americans. Whether such a policy deserves deference depends utterly on the context. The purported educational benefits of achieving a certain desired "mix" was the basis for the infamous Jewish quotas at Ivy League universities early in the past century,  and, allegedly, is the basis for concerns at some California schools that the number of Asian Americans is too large. Educators who attempted to argue today that quotas on Jews or Asian Americans are necessary to achieve desirable educational outcomes would receive very little, if any, deference-and deservedly so. Conversely, we seem quite willing to grant deference to other educators who offer educational justifications for concededly "exclusive" institutions, such as all-women's or predominantly black colleges. Yet an educator who tried to demonstrate the educational benefits of an all-white or, increasingly, even an all-male institution would probably be hooted off the stage. My point is neither to defend nor attack any of these arrangements, but simply to express skepticism that the concept of "inclusiveness" can help us pick out the contexts in which we ought to defer or not defer to the supposed expertise of educators.
 See Owen Fiss, "Groups and the Equal Protection Clause," 5 Phil. & Pub. Affairs 107 (1976).
 See Owen Fiss, "Groups and the Equal Protection Clause," 5 Phil. & Pub. Affairs 107 (1976).
 539 U.S. at (slip op. at 15), quoting the Brief for Respondents Bollinger et al., at i.
 Id. at 16.
 Id. at 17, quoting Brief for Respondents Bollinger et at., at 13.
 See Jim Chen, "Diversity in a Different Dimension: Evolutionary Theory and Affirmative Action's Destiny," 59 Ohio St. L.J. 811 (1998): "everybody talks about diversity, but no one knows what it means." For suggested taxonomies of "diversity," see Patricia Gurin at al., "Diversity and Higher Education: Theory and Impact on Educational Outcomes," 72 Harv. Educational Rev. 330 (2002); Levinson, supra n. 19.
 See Michael T. Nettles et al., "Race and Testing in College Admissions," in Gary Orfield & Edward Miller, eds., Chilling Admissions: The Affirmative Action Crisis and the Search for Alternatives (1998), at 97.
 438 U.S. at 314. Another illustration of the catholicity of Powell's conception is his quotation from William Bowen, then president of Princeton, on the wide range of student characteristics that contribute to "informal learning":
[A] great deal of learning occurs informally. It occurs through interactions among students of both sexes; of different races, religions, and backgrounds; who come from cities and rural areas, from various states and countries; who have a wide variety of interests, talents, and perspectives . . .
Id. at 313 n. 48, See also Justice Powell's quotation from the Harvard plan, describing the wide range of factors that count toward student "diversity" at that university, id. at 316.
 Id. at 25. 26.
 539 U.S. at (slip op. at 25).
 Id. at 26: [A]dmittees who have lived or traveled widely abroad, are fluent in several languages, have overcome personal adversity and family hardship, have exceptional records of extensive community service, and have had successful careers in other fields.
 Fiss describes the individualistic ideal as "the idea of treating people as ï¿½individuals'-recognizing each person's unique position in time and space, his unique combination of talent, ability and character, and his particular conduct." Fiss, supra n. 33, at 126-27. The attractiveness of this ideal may help to explain the Court's hostility in Gratz to an admission system that reduces applicants to an arithmetic formula.
 Interestingly, Justice Kennedy, in his dissenting opinion, quite explicitly recognizes a state interest in "racial" diversity: "In the context of university admissions the objective of racial diversity can be accepted based on the empirical data known to us, but deference is not to be given with respect to the methods by which it is pursued. 539 U.S. at (Kennedy slip op. at 2).
 Id. at 17-18.
 Id. at 20.
 Id. at 20, quoting from Brief for Respondent Bollinger et al., at 30.
 Id. at 21.
 Id. at 26.
 Gary Orfield, "Introduction," in Gary Orfield & Michael Kurlaender, eds., Diversity Challenged 1, 16 (2001).
 See Anthony T. Kronman, "Is Diversity a Value in American Higher Education?" 52 Fla. L. Rev. 861, 879 (2000) (arguing that different races have different experiences that "shape attitudes, producing characteristically different beliefs and judgments about society as a whole, and contrasting impressions of the relations among races."
 438 U.S. at 316.
 I've never been sure whether it's the "farm" or the "Idaho" that is doing the work in that example, or whether there is something special about farms in Idaho.
 There is an extensive literature speculating on whether racial discrimination is economically efficient as a form of "statistical discrimination." See, e.g., Stewart Schwab, "Is Statistical Discrimination Efficient?" 76 Am. Econ. Rev. 228 (1986). Even if it is, the moral intuition of most people is that it should be prohibited.
 Schuck, supra n. 19, at 37-38.
 O'Connor tells us that: "The Law School frequently accepts nonminority applicants with grades and test scores lower than underrepresented minority applicants . . ." 539 U.S. at 26. But this tells us very little about the true impact of these factors. The reality at most schools is that they receive very little weight as compared to the weight given to race. The primary exceptions, at undergraduate institutions, are athletic ability and legacy status, neither of which are intended to promote viewpoint diversity.
 539 U.S. at (slip op. at 18).
 "This [promotion of cross-racial understanding] is not, of course, an ï¿½educational benefit' on which students will be graded on their Law School transcript (Works and Plays Well with Others: B+) or tested by bar examiners (Q: Describe in 500 words or less your cross-racial understanding)." 539 U.S. at (Scalia, J., concurring in part and dissenting in part) (Scalia slip op. at 2).
 The programs are described in Gurin et al., supra n. ,at 343-46.
 See Linda Hamilton Krieger, "Civil Rights Perestroika: Intergroup Relations after Affirmative Action," 86 Calif. L. Rev. 1251, 1331-32 (1998).
 Stanley Rothman, Seymour Martine Lipset & Neil Nevitte, "Racial Diversity Reconsidered," The Public Interest 25 (Spring 2003).
 Numerous commentators have noted the arbitrariness and selectivity of racial and ethnic distinctions drawn in most affirmative action programs. E.g., Christopher A. Ford, "Administering Identity: The Determination of 'Race' in Race-Conscious Law," 82 Calif. L. Rev. 1231 (1994); Levinson, supra n. 19, at 597; Schuck, supra n. 19, at 14-17, 39.
 See Tables 1-3 in Chief Justice Rehnquist's opinion. 539 U.S. at (slip op. 7).
 See, e.g., Orfield , supra n. 50, at 15: "Whites are by far the most segregated group of students in American pulbic schools and thus are likely to have the least knowledge of effective intergroup skills." See also Gurin et al., supra n. 37, at 352, claiming that in the University of Michigan cohort the "most consistent" educational effects of diversity were experienced by the white students.
 See K. Edward Renner, "Racial Equity and Higher Education," Academe, http://www.aaup.org/publications/Academe/03jf/03jfrenart.htm (accessed 2/6/03) ("the main value of ï¿½diversity' is for white students to learn from minorities. This uses blacks and Hispanics for white ends").
 The literature contains numerous reported estimates of the gaps in SAT scores and high school GPAs among racial groups at most elite colleges and universities. Most estimates of the SAT gap range from 100 to 300 points in the combined SAT scale. Most estimates of the GPA gap range up to a full grade-point average (1.0 point on a 4-point scale). See, e.g., Gail Heriot, "Strict Scrutiny, Public Opinion, and Affirmative Action on Campus: Should the Courts Find a Narrowly Tailored Solution to a Compelling Need in a Policy Most Americans Oppose?" 40 Harv. J. on Legis. 217, 229 (2003); Thomas J. Kane, "Misconceptions in the Debate Over Affirmative Action in College Admissions," in Gary Orfield & Edward Miller, eds., supra n. 38, at 18; Schuck, supra n. 19, at 18-19.
 See Stephen Cole & Elinor Barber, Increasing Faculty Diversity: The Occupational Choices of High-Achieving Minority Students 101-04 (2003); Heriot, supra n. 66, at 230. Cf. Beth Dawson et al., "Performance on the National Board of Medical Examiners Part I Examination by Men and Women of Different Race and Ethnicity," 272 J.A.M.A. 674 (1994) (discussing racial differential in rate of failure on NBME Part I examination).
 See Bowen & Bok, supra n. 27, at 72-77; Cole & Barber, supra n. 67, at 121-138.
 The distinction is discussed in Stephen L. Carter, "When Victims Happen to Be Black," 97 Yale L.J. 420 (1988).
 See Gewirtz, supra n. 19, at 127; Levinson, supra n. 19, at 588-90.
 539 U.S. at (slip op. at 20).
 476 U.S. 267, 274-75 (1986) (plurality opinion).; id. at 288 (O'Connor, J., concurring). In Wygant, Justice O'Connor did distinguish between the "role model" theory and the goal of "promoting racial diversity among the faculty," an interest she said had not been pressed upon the Court in that case. Id. at 288 n.*. A majority of Justices appeared to have approved the rejection of "role model" theory in City of Richmond v. J.A. Croson Co., 488 U.S. 469, 498 (1989).
 Justice Thomas is the most vocal spokesman for this view on the current Court. See, e.g., Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 241 (1995) (Thomas, J., concurring). Justice Powell made a similar point in Bakke. 438 U.S. at 298. The point has been made repeatedly in the secondary literature. See, e.g., Stephen L. Carter, Reflections of an Affirmative Action Baby 47-69 (1991); Shelby Steele, The Content of Our Character: The Second Betrayal of Black Freedom in America (1998); Kreiger, supra n. 60, at 1259-63.
 There is evidence that this has in fact happened in the University of California system following adoption of Proposition 209. See Heriot, supra n. 66, at 230-31; Schuck, supra n. 19, at 75-78; Stephan Thernstrom & Abigail Thernstrom, "The Consequences of Color-Blindness," Wall St. J., April 7, 1998, at A18.
 For examples of the sociological evidence, see the essays collected in Orfield & Kurlaender, eds., supra n. 50 (especially Mitchell J. Chang, "The Positive Educational Effects of Racial Diversity on Campus," at 175; and Sylvia Hurtado, "Linking Diversity and Educational Purpose: How Diversity Affects the Classroom Environment and Student Development", at 187). See also Gurin et al., supra n. 37. But see Rothman et al, supra n. 61 (reporting that student and faculty satisfaction with the quality of a college's educational program is inversely correlated with the degree of its racial diversity).
 See, e.g., Gerald Gunther, "Foreword: In Search of Evolving Doctrine on a Changing Court: A Model for a Newer Equal Protection," 86 Harv. L. Rev. 1 (1972). But see Jed Rubinfeld, "Affirmative Action," 107 Yale L.J. 427 (1997), who argues that the Supreme Court has distorted the strict scrutiny test, from its original (and proper) function to "smoke out" illegitimate motives for using race as a discriminant, to a more recent (and improper) cost-benefit test analysis of the consequences of using race as a discriminant. Most affirmative action programs, Rubinfeld argues, would easily satisfy the "smoking out" test.
For speculations on some of the less-than-noble motivations that might sustain race-conscious admissions programs, see Schuck, supra n. 19, at 36, 58, 71; Shelby Steele, "Affirmative Action Must Go," N.Y. Times, Mar. 1, 1995, at A19. In his opinion in Grutter, Justice Thomas variously ascribes affirmative action programs to elitism, faddishness, aesthetics, and image-consciousness. See Thomas slip op. at 2, 6, 22. He even hints that, because race-conscious admission practices disadvantage minorities academically, the might be motivated by a kind of subconscious racism. Id. at 23-27.
 Briggs v. Elliott, Brief for Appellee; see also Davis v. County School Bd., Brief for Appellee.
 United States v. Virginia, 518 U.S. 515 (1996).
 See Gratz, 539 U.S. at (Ginsburg, J., dissenting) (Ginsburg slip op. at 4).
 The Court has explicitly instructed district courts to grant deference, at the remedial stage, to administrators of higher educational institutions undergoing court-ordered desegregation. E.g., United States v. Fordice, 505 U.S. 717 (1992). In this context, academic "judgment" is confined by a clear directive to remove previous racial barriers.
 See, e.g., H. Broun & G. Britt, Christians Only: A Study in Prejudice (1931).
 There is an extensive literature touting the educational advantages of attending predominantly black institutions of higher education. See, e.g., Walter R. Allen, "African-American College Student Outcomes at Predominantly White and Historically Black Public Colleges and Universities," 62 Harv. Educ. Rev. 26 (a992); Lamont Flowers & Ernest T. Passcarella, "Cognitive Effects of College Racial Composition on African American Students After 3 Years of College," 40 J. of College Student Development 671 (1999).
 Cf. Justice Stevens's concession that the "faculty-diversity" rationale used to justify race-conscious hiring in the Wygant case could be turned on its head by those who would argue that segregated classes lead to better achievement. 476 U.S. at 316.