UNREFLECTIVE DISEQUILIBRIUM Race-Conscious Admissions After SFFA

For many years, there was a settled constitutional rule that universities could sometimes prefer certain applicants over others on the basis of race. In Students for Fair Admissions v. Harvard College (“SFFA”),¹ the Supreme Court unsettled that rule. For advocates of racial integration and equality of opportunity, the decision comes as a major blow—a ruling whose “devastating impact,” in Justice Sotomayor’s assessment, “cannot be overstated.”² But, in our view, SFFA can also be seen as a portrait of a body of law that is in transition. The ruling is ambivalent, unsure of its own import, and thus contains the seeds of more than one possible future.

Dobbs v. Jackson Women’s Health Organization,³ handed down a year before, makes for an instructive comparison. Both were blockbuster rulings in which the conservative majority finally triumphed on a central question that had divided the Court for decades. In other respects, though, the opinions could hardly have been more different. Dobbs tore up the Court’s abortion precedents root and branch, unequivocally “return[ing] the issue of abortion to the people’s elected representatives.”⁴ Its denunciation of abortion rights was sweeping and definitive. And the Court’s ruling therefore left hardly any room for a “next case” about the status of abortion as a specially protected facet of liberty.

SFFA differed markedly in its orientation both to the past and to the future. Looking backward, the Court purported not to overrule its affirmative-action precedents, but to faithfully apply them. “We have never permitted admissions programs to work [as Harvard’s and the University of North Carolina’s did],” Chief Justice Roberts claimed, “and we will not do so today.”⁵ Looking forward, Roberts made clear that SFFA had not purged all traces of race consciousness from the college admissions process—and he expressly contemplated that some manner of dialogue with regulated universities over the pivotal lines would therefore continue. Anticipating that next round of litigation, he remarked on where universities should turn for “legal advice on how to comply”—not to the dissenters—and he warned that whatever latitude universities enjoyed did not extend to “simply establish[ing] through … other means the regime we hold unlawful today.”⁶

All of this prompts a question: If “[e]liminating racial discrimination means eliminating all of it,”⁷ as the majority said, why couldn’t the Court just put the issue to rest? The short answer is that, despite its sometimes strident rhetoric, there were certain forms of race consciousness that even the majority did not want to condemn. “[N]othing in this opinion,” Roberts explained, “should be construed as prohibiting universities from considering an applicant’s discussion of how race affected his or her life, be it through discrimination, inspiration, or otherwise.”⁸ “A benefit to a student who overcame racial discrimination” would be permissible, for instance, as long as it was “tied to that student’s courage and determination.”⁹ Likewise, the Court would have no objection to “a benefit to a student whose heritage or culture motivated him or her to assume a leadership role or attain a particular goal,” if the “benefit” is “tied to that student’s unique ability to contribute to the university.”¹⁰ In cases such as these, the Court reasoned, the student would “be treated based on his or her experiences as an individual—not on the basis of race.”¹¹

To the dissenters, these “supposed” limits on the majority’s holding were “nothing but an attempt to put lipstick on a pig.”¹² The Court had merely “announce[d] a false promise to save face and appear attuned to reality,” Justice Sotomayor wrote, and “[n]o one is fooled.”¹³ Specifically, what made the promise false (in her view) was that “[t]he Court’s opinion circumscribes universities’ ability to consider race in any form by meticulously gutting respondents’ asserted diversity interests.”¹⁴

But we doubt that only a concern for appearances—as opposed to genuine ambivalence—was at work here. It is true that SFFA deemed the claimed benefits of race-based measures too “amorphous” to withstand strict scrutiny in this context.¹⁵ But the thrust of the Court’s later discussion must therefore have been that awarding the “benefits” it described would not amount to a racial classification triggering strict scrutiny at all.¹⁶ True, just what bounds that permissible practice (and distinguishes it from “the regime we hold unlawful today”¹⁷) is far from clear. But that does not necessarily make the safe harbor empty or a sham. (If the safe harbor were intended to encompass nothing, after all, including it—backed by a vague admonition to adhere to the spirit of the ruling—would seem a major blunder.) What the combination of the safe harbor and its uncertain bounds more plausibly reflects, we think, is that the Court is torn between conflicting impulses and uncertain, so far at least, about just how to reconcile them.

In fact, those conflicting impulses are evident on the face of the opinion. On the one hand, the Court is firmly committed to the proposition that race-based decisions are inherently demeaning and unjust. But on the other hand, the Court does not want to forbid universities from taking account of genuine differences between applicants—in terms of their experiences, aptitudes, and other relevant characteristics. Those two commitments would be easy to square in a factual context where nobody’s race had anything to do with any of their relevant characteristics. But as the dissenters emphasized, that is not our world.¹⁸ And the majority’s diffidence about its own holding—its felt need to authorize consideration of the effects or significance of race, if not race itself, for each applicant “as an individual,” if not for larger groups—strongly suggests that, at some level, it appreciated this fact about our social reality as well.

Our main aim here is to show how difficult it will be to draw and justify lines that reconcile these conflicting commitments. Broadly speaking, reasoned distinctions should be expected to arise from the contours of the principles that make conduct permissible or impermissible in the first place. But when two deeply different principles point toward very different prohibitions—and yet the Court is unwilling to relinquish either conviction—the tug of war that results will yield lines that are essentially arbitrary, if indeed they are discernible at all. The predicament that the Court began to confront in SFFA is of the latter kind, and the prospects for a workable doctrine to emerge are therefore slim.

Importantly, we do not mean that SFFA worked no change in the law. To the contrary, it seems clear that any policy that unambiguously employs “racial classifications” will now be forbidden. Thus, for example, an effort to admit at least some minimum overall number of Black students—with admissions thresholds explicitly or implicitly adjusted as needed to accomplish that aggregate aim—would be held unconstitutional. But when it comes to what the university may value in individual applicants, or how their race might figure as an input to judgments about their likely attributes and experiences, we doubt that principled lines can be drawn in anything like the terms that the Court suggested. The Court’s admonition to tackle those difficult issues by adhering to the spirit of its ruling in SFFA—that is, by remembering that “what cannot be done directly cannot be done indirectly”¹⁹—thus seems naïve at best. We are left with the distinct impression that the Court failed to appreciate the depth of the conflict between the principles to which it found itself drawn. Put another way, in contrast to the “reflective equilibrium” between attractive principles and intuitions about particular cases to which many legal thinkers aspire, SFFA reveals the Court to be in a state of unreflective disequilibrium—pulled in different directions, but lacking, so far, the willingness to adjust its starting points as necessary to reconcile that ambivalence.²⁰

Our argument will unfold as follows. In part I, we first present in a bit more detail the SFFA majority’s effort to distinguish between illegitimate consideration of a person’s race, on the one hand, and legitimate consideration of the ways that race figures in or has affected a person’s life, on the other. To the extent that this distinction gives the appearance of a clean solution—a way of accounting for the relevance of race without compromising the commitment to colorblindness—we show that this appearance is false. Often, understanding how a person’s race has affected them, or how it figures in their present self-conception, requires also appreciating that it is their race—or even their being Black, being white, or the like—that is playing this role. A thoroughgoing race blindness in admissions would thus make impossible the very practices that the Court was understandably loath to restrict.

Because a strict rule of colorblindness cannot reconcile the conflict that we have described, we turn in part II to other possibilities, looking to the values that the Court invoked in support of its holding in SFFA itself. Perhaps, for instance, the central issue is whether a policy uses “race for race’s sake,” as the Court said that Harvard and UNC’s practices did.²¹ Or perhaps it (also) matters whether a policy trades on racial generalizations—especially the “offensive and demeaning assumption that [students] of a particular race, because of their race, think alike.”²² Yet, as we will explain, criteria such as these also fail to do justice to the Court’s conflicting impulses. Rendered one way, they would again contradict the Court’s commonsense assurances that universities may attend to individuals’ actual experiences and identities. But if these criteria are relaxed in the ways required to accommodate those assurances, they cease to explain why the conduct that SFFA presumably sought to prohibit differs significantly from the conduct that it approved.

In part III, we try another tack. Perhaps the pivotal line must be drawn not only in terms of which properties the university values in an applicant, or the premises of the inferences that it makes about them, but also in terms of the manner in which the university comes to know the facts that set that reasoning in motion in the first place. In particular, perhaps all consideration of race in admissions really could be forbidden, but with a crucial proviso: “an applicant’s discussion”²³ of their race and its significance is fair game. We can see the allure of drawing such a distinction: the university itself is not assigning any significance to race, the thought would go, but it is open to hearing from applicants about whatever they deem important. As we will explain, however, that appeal also dissipates on closer inspection. Because communication is fundamentally a cooperative endeavor, it is impossible to specify in a principled way just what must come “from” the applicant—as opposed to figuring in the context from which the applicant can fairly assume that a well-socialized reader will draw when interpreting their own words.

Finally, we return in part IV to the theme with which we began: the inner conflict about the relevance of the social reality of race that SFFA lays bare, and the Justices’ failure—thus far, at least—to entertain the revisions to their own convictions that would be necessary to resolve it.

Because we are claiming that the SFFA Court was caught between conflicting impulses or principles, we should begin by entertaining a possible reading of the opinion that would seem to dispense with any such conflict. According to that reading, the new constitutional boundary in this area is simple: universities may not assign any significance whatsoever to an applicant’s race in making admissions decisions. We will first sketch out the analysis that might seem to support such a categorical or “colorblind” rule, and that might even have led the SFFA Court to think that its opinion supported (or at least could be squared with) such a rule. We will then explain why that analysis cannot be correct.

The linchpin of the colorblindness reading of SFFA is a distinction between a person’s race, on the one hand, and the miscellaneous other properties that they might possess in some manner connected to their race, on the other. Once one draws that distinction, SFFA might seem to bar universities only from considering the former—that is, from counting the bare fact of anyone’s race in their favor (or against them). Insofar as certain other attributes are of interest to a university, the fact that a student came to possess one of those properties on account of their race does not deprive that other attribute of its ordinary significance. Facts about “how race affected [an applicant’s] life, be it through discrimination, inspiration, or otherwise,”²⁴ are thus relevant just insofar as the same “[e]ffect[s] [on] [the applicant's] life” would also have mattered to the decision-maker if they had been effects of something else instead. What the university may not consider, in other words, is whether an applicant satisfies some race-specific description.

This account captures one highly intuitive understanding of what it is to make a decision about someone “on the basis of race,” and, at first blush, it does seem to cleanly partition such decisions from those based on the other properties that might properly matter to a university. Suppose, for example, that a university values grit, understood as a certain kind of resilience in the face of setbacks. And suppose that a particular applicant exhibited or developed grit through grappling with the disadvantages, such as discrimination, that they have faced on account of their race. Even if this person’s grit (or the evidence thereof) is thus causally traceable to their race, and even if the university happens to know this, valuing the grit qua grit is not valuing the person’s race. It is only if the university responds differently to the specific property of grit developed on account of being Black than to the more general property of grit that it employs a racial classification—subjecting the university’s practice to strict scrutiny that, after SFFA, it will almost surely fail.²⁵

Upon further inspection, though, the partition on which this line of thought relies is illusory. In many cases, one cannot engage in the kind of consideration of race-related properties that the Court blessed—indeed, that it said “nothing in [its] opinion should be construed as prohibiting”—without also considering, at some level, an applicant’s race or racial self-identification. That means that, whether the Justices in the majority appreciated it or not, the conflicting impulses to which SFFA gave expression cannot be reconciled by simply forbidding all consideration of an applicant’s race.

We will offer three different kinds of cases to demonstrate as much. This terrain could be divided in different ways, and nothing much turns on the precise delineation. But, for clarity of exposition, we will call these types race as decoding key, race as fungible ingredient, and race as quantifier.

Race as Decoding Key. In the first type of case, responding differently to applicants of different races is plainly required if a university is to identify and respond appropriately to the wholly race-independent properties that applicants possess. Race thus functions as a kind of decoding key; it allows the university to translate from the non-race information before it to the (also non-race) properties that it is actually interested in. In light of the Court’s recognition that race could figure in someone’s life as a source of “inspiration,” for example, consider this pair of essay fragments (which differ only in the second word of each):

We cannot think that, under SFFA, a university is obliged to award these two essays the same score. And when the university favors Applicant A’s essay over Applicant B’s (as it surely would), it would be obtuse to describe that differential treatment as being “on the basis of race” in any sense relevant to equal protection. After all, the different essays demonstrate many differences between the applicants that reasonable assessors (who attach no value to anyone’s race) would deem pluses or minuses in their files. In fact, the two essays differ in the very contents of what they are each saying. A typical reader of Applicant A’s essay might think of the “fight” for the Voting Rights Act, the “sacrifices” of Bloody Sunday, and so forth—and Applicant A is thus naturally understood as intentionally citing the people involved in those or similar events as inspirations. A typical reader of Applicant B’s essay would understand Applicant B to be saying something quite different about their motivations.²⁶ A rule of neutrality as to race does not plausibly require neutrality as to those very different worldviews.

We take it, therefore, that post-SFFA equal protection law licenses the disparate evaluations of these two applicants that seem intuitively correct. Were it otherwise, the SFFA Court’s invitation for universities to consider how a person’s race might figure in their life “through … inspiration” could not have been serious. Put another way, it seems inconceivable that, in blessing the consideration of the valuable properties that Applicant A demonstrates, the Court actually meant that they can be valued only insofar as the malignant properties that Applicant B demonstrates are too.

That means, though, that SFFA’s rule is not strictly colorblind: It does sometimes permit assessors to treat applicants differently in light of the racial identities that they avow. We will defer until part II the question of just when this might be permitted or forbidden. To a first approximation, however, the race as decoding key scenario might suggest that applicants’ racial identities (or avowals) are valid grounds for differential treatment to the extent that they are employed in decoding which relevant non-race properties an applicant does or does not possess (just as the avowed races of Applicant A and Applicant B are used in our examples).

Race as Fungible Ingredient. A second kind of case poses a similar problem. Here, however, race is relevant not simply as a decoder, but as an ingredient—interchangeable in principle with others—in a higher-order property that a university is free to value on its own terms. In these cases, the relationship between race and the higher-order property is not of the specifically causal variety that the Court mentioned expressly.²⁷ And yet it seems clear—both as a matter of common sense, and in light of the Court’s discussion—that these tighter relationships are no different from causal connections for purposes of SFFA’s distinction between permissible and impermissible considerations.²⁸

Suppose, for instance, that a university values the property of being thoughtful about one’s own inherited advantages and disadvantages. Any applicant who has that property at all will have it in some specific way. In other words, there will have to be one or more properties that:

• (1) 

  the applicant has;

• (2) 

  are (at least for them) inherited advantages or disadvantages; and

• (3) 

  the applicant is thoughtful about.²⁹

There will be some applicants (of all races) whose race is a property that satisfies this schema, and thus that indirectly qualifies them as possessing the higher-level property of interest. And presumably here, just as when an applicant’s race figures as a cause of some more general effect (such as “overcoming significant adversity”³⁰), nothing prohibits a university from recognizing that an applicant possesses the schematic, non-race-specific property that it values across the board. In other words, the law could not plausibly require that students who are thoughtful about an inherited advantage or disadvantage that falls along one particular axis (i.e., race) must be disfavored relative to others who are thoughtful about their positions on other such axes (such as family wealth or the like). Notwithstanding the majority’s causal language, its point seems clear: the bar on valuing an applicant’s race does not extend to evenhanded consideration of other properties just because some applicants’ possession of those independently valued properties might happen to involve or relate to their race (whether in terms of causal history or otherwise).

In fact, the same point is nicely illustrated by one of the Court’s own examples, drawn from the trial record, of what its holding permits. (These examples seem to have gone unnoticed, presumably because they were initially obscured in the Court’s opinion.³¹) To support its explanation of what, “as all parties agree,” remains permissible, the Court cited UNC’s discussion of how Andrew Brennen, a Black student, had answered a question about personal motivation on his college application. Brennen had described the stereotyping that he experienced as a Black man—such as expectations that his interests should be limited to “rap music” or “the hood”—and explained that he “do[es] what [he] do[es] because people do not expect it from [him], [and] because others who look like [him] are not able to do it.”³² Suppose, then, that an assessor rewarded Brennen’s essay because, among other things, it showed that he had the property of being motivated to defy pejorative, stereotypic expectations predicated on one’s socially salient identities.³³ Here again, what made it the case that the applicant had the valued property was, in part, his race.³⁴ After all, it is stereotypes about Black men that Brennen was committed to defying, and that would not amount to having the valued property if he were not a Black man himself. (One cannot really “defy” stereotypes about what other people are like.) So, if Brennen had not conveyed (either expressly or by implication) that he was Black, he would not have been credited with a property that the university values. Yet we take it that this case also clearly falls on the right side of SFFA, as the Court’s invocation of it as an illustration appears to confirm.

Here too, then, a sensible reading of SFFA forecloses a strict rule of colorblindness. What matters, it seems, is that the university is valuing not Brennen’s race, but rather a non-race property that Brennen simply possessed for race-linked reasons or in a race-linked way.

Race as Quantifier. Our third and final demonstration of SFFA’s non-colorblind commitments resembles the first, in that an applicant’s race is used to discern their non-race properties, but it differs in how race serves that role. Rather than serving simply to decode what an applicant is saying, the applicant’s race often enables a university to assess the magnitude of the non-race properties that it takes other aspects of the applicant’s file to demonstrate. Consider, for example, these three essay fragments:

SFFA leaves no doubt that an assessor can award points to Applicant C in light of what the first essay demonstrates about “that student’s courage and determination.”³⁵ And what the essay does tend to show about that of course depends on the experience that the student is describing and how difficult the assessor has reason to think that sort of experience really is. That is why Applicant C can surely be favored over the parking-challenged Applicant E; the experiences that the assessor has reason to think each underwent are very different in the respect that the university properly cares about.

Given that much, though, it seems almost equally plain that Applicant C can be favored over Applicant D, too. The experiences that each underwent simply differ, at least as far as the assessor’s evidence goes, in terms of how much courage and determination each evinces or is apt to engender. Indeed, it would be passing strange if SFFA’s admonition to treat each student “as an individual” based on “that student’s” experiences or characteristics required treating Applicants C and D as if they were the same person. (Even if one tried to implement that rule, moreover, it verges on incoherent.³⁶)

We will return to this and similar examples—but for now, our point is the same as with the last two.³⁷ Any plausible version of “considering an applicant’s discussion of how race affected his or her life, be it through discrimination, inspiration, or otherwise” will also involve considering the race of applicants, at least in the sense of awarding literal or figurative points in a manner that is sensitive to the identities that they convey or avow. Although such decisions are made on the basis of race in a certain sense, they evidently are not (always) made “on the basis of race” in the sense that would constitute a “racial classification” and trigger strict scrutiny. Put another way, under SFFA, some race-sensitive admissions practices are evidently more like stopping someone for questioning based on a suspect description that includes race: The decisions at issue are race sensitive, but courts have avoided scrutinizing them by deeming them not really based on race, or not based on race as such, or something to that effect.³⁸SFFA essentially promises the same treatment for some class of race-sensitive decisions—including, we take it, each of the three uses of race we have sketched so far—in the context of university admissions.

Even if the conflicting impulses reflected in SFFA cannot be reconciled by a simple rule of colorblindness, perhaps there is another principle that can do the job. After all, the SFFA majority certainly seemed to think that some such principle existed. Otherwise, it could not have thought that a university that read its opinion sympathetically—appreciating that “the prohibition against racial discrimination is ‘levelled at the thing, not the name’”—could avail itself of the leeway that the Court had acknowledged, on the one hand, without effectively “establish[ing] through application essays or other means the regime we hold unlawful today,” on the other.³⁹

Over the next two sections, we will consider whether the values that the Court invoked in SFFA can ground an account of “the thing” that distinguished Harvard’s and UNC’s admissions policies from the forms of race consciousness that even the Court deemed permissible. We conclude, as we’ve foreshadowed already, that there is no such principled distinction to be found here. While the Court did appeal to certain overarching normative concerns in condemning the Harvard and UNC programs—in particular, faulting them for valuing “race alone” and for engaging in “stereotyping”—neither idea can be reconstructed so as to explain both the permissibility of the forms of race consciousness that the Court approved and the impermissibility of the policies that we take it the Court sought to condemn.

When the Court described what was unacceptable, even galling, about the Harvard and UNC programs, it stressed not just that but how they considered race. “The point of respondents’ admissions programs,” according to Chief Justice Roberts, was “that there is an inherent benefit in race qua race—in race for race’s sake.”⁴⁰ Similarly, he faulted the universities for having “concluded, wrongly, that the touchstone of an individual’s identity is not challenges bested, skills built, or lessons learned but the color of their skin.”⁴¹ In what seems the same spirit, the majority repeatedly characterized the disfavored policies as affording preferences to some applicants “on the basis of race alone.”⁴² Indeed, it was the fact that “some students may obtain preferences on the basis of race alone” that, according to the majority, showed the Harvard and UNC programs to be at odds with Grutter v. Bollinger.⁴³ And, in responding to the dissenters, Chief Justice Roberts deemed it a damning fact that “race—and race alone—explain[ed] the admissions decisions for hundreds if not thousands of applicants to UNC each year.”⁴⁴

As we hear it, this rhetoric betrays a sense that what is so pernicious is not just any consideration of race, but Harvard’s and UNC’s (purported) practices of affording it sole, exaggerated, or “inherent” significance. The Court’s ire seems aimed at attention to race that is in some sense partitioned off from other attributes—at practices under which just being of one or another race, without more, triggers some difference in treatment. And, rendered charitably, this is indeed a recognizable normative commitment that could theoretically guide judgments in this area.

The clearest example of the forbidden sort of racial preference would be a regime that attached some kind of moral value to the sheer fact of the applicant’s skin color. After all, that sort of “[d]istinction[] between citizens solely because of their ancestry” really is “by [its] very nature odious to … the doctrine of equality.”⁴⁵ The paradigm of such a regime, of course, is not affirmative action programs but the systematic subordination of Black Americans under Jim Crow. Still, the same principle might also be transgressed by a regime that, perhaps in the name of some conception of cosmic justice between people of different phenotypes, deemed a dark-skinned applicant inherently more worthy or deserving of admission.⁴⁶ Whether such a preference were large or small, it could still be said to rest upon “race alone” in the sense evoked by the Court’s references to seeing an “inherent benefit” in race or valuing “race for race’s sake.”

If the principle condemning judgments based on “race alone” reached only that far, though, it would not explain the invalidity of many (perhaps any) actual affirmative-action policies. We doubt that many such programs have been predicated on the notion that some applicants are, simply by dint of “the color of their skin,”⁴⁷inherently better applicants or more entitled to admission. As Elizabeth Anderson puts much the same point, affirmative-action policies need not attach value to anyone’s “minimal race” (i.e., real or imagined bodily differences taken as indicators of shared ancestry or origin) or even to “character race” (i.e., a propensity for normatively significant traits taken to be intrinsic to a group defined by minimal race).⁴⁸ What matters for purposes of an affirmative action program, rather, is people’s membership in “racialized groups”—that is, their possession of traits that are socially represented as forming a character race and that make those people vulnerable to particular injustices as a result.⁴⁹

The important question about the Court’s condemnation of valuing “race alone” or “race for race’s sake,” therefore, is what it means for the practice of treating someone’s membership in a racialized group—specifically, their membership in one that is the subject of stigma and social exclusion—as a consideration in favor of admitting them. Under the Court’s pre-SFFA precedents, universities were permitted to value certain benefits that would flow from the admission of a particular student simply by dint of their membership in such a group. Suppose, for instance, that an elite university, whose graduates often go on to attain societal leadership positions, thinks it would be good for those positions to be more “visibly open” to young people who might otherwise discount themselves.⁵⁰ Admitting a Black student would then have greater expected value than admitting an otherwise-similar white student, because if the admitted student goes on to occupy a leadership role, there is an additional social benefit that will be captured only in the scenario where the student is Black.⁵¹

Now, if we had to wager on whether the SFFA majority would still permit a university to award a “plus” to a Black applicant on these grounds, we would bet against the university. But we do not think that SFFA gives a reason—or, really, that the Court could give a convincing reason—for such a prohibition, at least not one that can also be squared with the logic of what SFFA permits. For starters, it seems clear that the kind of “plus” that we have just described does not value “race for race’s sake”; to the contrary, an applicant’s race is valued, insofar as it is, purely for the sake of something else.⁵² And while awarding this “plus” certainly does transgress a norm of colorblindness—since it predicates decisions about people on their race—we have already shown that colorblindness, as such, cannot be “the thing”⁵³ that SFFA elevates to a guiding value for universities looking to comply with its demands. Once an applicant’s race is being considered—in the sense that admissions decisions may sometimes be sensitive to it—it is not obvious what line would be crossed by valuing the positive social effects that we have just described, even where it is an applicant’s race that links their admission to those effects.

What might distinguish this kind of race sensitivity from the kinds that SFFA recognized as benign? The distinction to which the Court would most likely be drawn, we think, is that here it really is “race alone”—meaning, for purposes of the “visibly open” rationale, the racialized group into which others are apt to sort a person—that serves as a merit or qualification, albeit on thoroughly instrumental grounds. As Anderson puts it in her defense of an integrative model of affirmative action, race “does not function fundamentally as a proxy for some other morally relevant property” in this model, but rather as the “direct object of … concern”: the institution is interested in precisely “[t]he causal power of a person’s race to break down race-based barriers to opportunity.”⁵⁴ This is what Justice Thomas was highlighting, we take it, in deriding the interest recognized in Grutter v. Bollinger⁵⁵ as resting on “racial aesthetics.”⁵⁶ And this same feature of the “visibly open” rationale—the seemingly unmediated relevance of race—makes it natural to see such an admissions practice as employing “racial classifications,” and hence triggering strict scrutiny, if anything does.

On closer inspection, however, all of this seems quite artificial. Note, first, that the property that the university actually values in our example—being such that one’s admission is apt to increase the apparent openness of leadership positions to all—does not have anything like a one-to-one relationship with race. Black applicants will not possess this property unless they also possess various others, such as the academic qualifications necessary to succeed at the university in the first place. (For this reason, the rationale does not actually recommend admitting anyone based on “race alone.”)⁵⁷ And conversely, many applicants who are not Black will possess the sought-after property nonetheless—because they are openly gay or transgender, or are visibly disabled, or have some other visible trait that many see, accurately or not, as counting against their prospects for holding positions of power in our society.

With these facts in view, the university’s consideration of race here does not seem fundamentally different than our earlier examples in which race figures as a fungible ingredient of another, independently valued property.⁵⁸ If a university values being motivated to defy pejorative, stereotypic expectations predicated on one’s socially salient identities, we noted, it will also end up rewarding some students for, among other things, being Black. The same is true of being such that one’s admission is apt to increase the apparent openness of leadership positions to all. In the former case, the fact that some possess the valued property in a way that is grounded in their race does not preclude crediting their possession of the valued property, period. The question remains, then, why the answer should be any different for the latter one.

If there is still an intuitive difference here, we suspect that it has something to do with notions of desert, merit, or fairness. After all, the physical appearance that makes a Black applicant satisfy the description at issue here does not in any sense reflect well on them. And there are many applicants for whom the same “plus” is, if not entirely unavailable, at least much more difficult to obtain.⁵⁹ In order to make these facts normatively relevant, however, one would have to endorse one of two broader principles.⁶⁰ The first possibility is that admissions decisions must be based on people’s deserts (which the “plus” we have described admittedly is not), rather than on the expected benefits of admitting them. But we can dispense with that broad principle quickly: Nothing in equal-protection law requires universities to approach admissions as an exercise in awarding people their just deserts rather than furthering an institutional mission.⁶¹ Any such rule would be transgressed by countless other considerations in university admissions, ranging from legacy status to whether an applicant’s academic interests (regardless of their intrinsic “merit”) align with the university’s strengths.

Second, one could posit that, in assessing the expected benefits of admitting different students, the social facts that undergird the extra value of a Black graduate’s success—including the fact that their physical appearance is stigmatized—must be disregarded. To motivate such a view, consider Justice O’Connor’s characterization of Batson v. Kentucky⁶²:

It is conceivable, at least, that the Court could think that, for like reasons, the social facts essential to the “plus” in our example must be disregarded. But this is where we most directly come up against the other impulse clearly reflected in SFFA. In eschewing colorblindness and approving the practice of considering how race has affected people’s lives, the Court does “permit the State to act on the premise that racial differences matter.” If the Court really thought that universities must disregard the social reality of race, in other words, it would have to require universities to simply feign confusion when a Black applicant describes their experience of stereotype threat, or when such an applicant touts their commitment to defying expectations that they will be interested only in “the hood,”⁶⁴ and so forth.

So, while we might predict that the Court would frown upon a university’s embrace of Grutter’s “visibly open” rationale—if only because it squarely overlaps the explicit rationale of past affirmative-action programs—we do not see a principled basis for treating the form of race consciousness at work in that rationale differently. At the least, the Court’s denunciations of valuing “race alone” or “race for race’s sake” do not suggest one.

In SFFA, the Court faulted not only the notion that there is an “inherent benefit in race qua race” (a matter, seemingly, of what is valued), but also the notion that “race in itself says something about who you are”⁶⁵ (a matter of what is believed or inferred). Even if there is no convincing reconstruction of the former line to be found, perhaps the Court’s conflicting impulses are reconcilable, at least in part, through a distinct rule against the latter sort of thinking. A rule against “stereotyping,” in other words, might explain in a principled way both what the Court sought to forbid and what it resolved to permit. Here too, however, the conflict runs deeper than the Court seems to have appreciated.

No doubt the Court thought that a commitment to eschewing stereotypes and treating people “as individuals” could make sense of its different prescriptions. The Court explained that concerns about “treat[ing] individuals as the product of their race” are a refrain of its modern race cases.⁶⁶ It noted that even Grutter had frowned upon any practice of assuming “that minority students always (or even consistently) express some characteristic minority viewpoint.”⁶⁷ And when the majority commented directly on the line between good-faith compliance and lawyerly circumvention, it presented a kind of individualization as the key:

The question, though, is whether this rhetoric can be translated into a principled rule that does not invalidate the very practices that the Court, recognizing the social reality of race, acknowledged as permissible. And although we can imagine a few different interpretations of the “anti-stereotyping” norm that the Court articulated, none can thread that needle.⁶⁹

The first and perhaps the most natural interpretation is a simple bar on any group-to-individual inferences based on an applicant’s race. Whether or not such a rule could be justified in principle, it cannot do the job here because it would effectively disallow the consideration of “how race affected [an applicant’s] life”⁷⁰ that the Court invited. Our earlier example of Applicant C’s essay about the experience of stereotype threat and classroom racial isolation, for instance, relies on just this kind of group-to-individual inference. The assessor must move from a belief about what the experience of being the only Black student in an AP class is generally like to an estimate of what Applicant C’s experience of being the only Black student in an AP class was (probably) like. Without that inferential step, the anecdote does not say anything useful about Applicant C’s “experiences as an individual.”⁷¹

A second interpretation of the anti-stereotyping norm would vindicate the obvious verdict about Applicant C by narrowing the rule’s scope. Specifically, perhaps only race-based inferences about people’s viewpoints—not about their experiences, or about the traits of character that tend to flow from experiences of the relevant kinds—are problematic. That would accord with much of the Court’s rhetoric in SFFA. “[W]hen a university admits students on the basis of race,” the Court explained, “it engages in the offensive and demeaning assumption that students of a particular race, because of their race, think alike.”⁷² Likewise, the Court characterized “the pernicious stereotype” on which “Harvard’s admissions process rests” as the belief that “a black student can usually bring something that a white person cannot offer,”⁷³ a belief grounded in the pedagogical value that Harvard saw in exposing students to a diversity of “background[s] and outlook[s].”⁷⁴

But this interpretation fails in the opposite way: by the Court’s lights, at least, it presumably permits far too much. Insofar as the case for affirmative action has traditionally involved group-to-individual inferences about racial groups, many of those inferences have been about applicants’ experiences, not their viewpoints. All else equal, for instance, Black and Latino students are more likely to have experienced discrimination, or to have been affected by growing up in a culture whose mass media portrays people who look like them in narrow or demeaning terms, and so on, than are white students. Those are, among other things, reasons to make different inferences about the personal characteristics (including “courage and determination”⁷⁵ but also including various ingredients of academic excellence) of students who present with “equal” records of academic achievement—regardless of what opinions or worldviews they might hold.⁷⁶ If the anti-stereotyping norm does not bar a university from taking account of this information in assessing individual applicants, it leaves a good deal of latitude for affirmative action to operate.⁷⁷

A third possible reconciliation might look not to what is inferred about someone, but to the property from which—or, put differently, the class about whom—it is inferred. That seems more promising in part because, at the end of the day, SFFA means that any line between what is permitted and what is prohibited will have to be recognizable as an account of what constitutes a racial classification in the first place.⁷⁸ In that spirit, the Court could accommodate the logical verdict about Applicant C by distinguishing an inference drawn on the basis of someone’s being the only Black student in their AP classes from an inference drawn from their being Black. The latter rests on a generalization or classification at the level of entire racial groups, but the former does not. Indeed, one could go a step further and insist that the fact of Applicant C’s race really does no work in our example at all; the only property of interest is having labored under stereotype threat where the stereotypes at issue are the stereotypes held about Black people. It comes as no surprise that practically everyone who has that property is Black, but still, the applicant’s being Black is not being afforded any significance, and thus the inference does not rest on any notion that anyone’s “race in itself says [something] about who [they] are.”⁷⁹

The problem for this kind of account is that, like the last one, it presumably rules out far too little (at least from the majority’s point of view). After all, there are very few generalizations about people’s experiences or characteristics for which the relevant reference class really is a racial category, full stop. Context would always matter to a conscientious assessor, whether the context is writ large (being Black and growing up in a society whose mass media portrays Black people in demeaning terms) or writ small (as in the AP classes example) or something in between (being Black and not especially wealthy).⁸⁰ The way would thus be open for a university to ascribe meaning to a person’s race so long as it did so in a sensible and relevant way—for instance, distinguishing the likely experiences of a Black applicant who grew up in the United States from those of one who grew up in another part of the world. Indeed, if this sort of principle were taken to capture the substance of SFFA, it would turn out that the Court essentially recapitulated Grutter’s “demand[] that race be used [only] in a flexible, nonmechanical way,” as part of “truly individualized consideration.”⁸¹SFFA would then prove to have preserved much of Grutter’s substance by relocating what had been a narrow-tailoring question to the threshold analysis of whether a racial classification is really at issue at all.

* * *

Where does this leave us? We first argued that the Court could not reconcile the commitments voiced in SFFA by simply mandating colorblindness (and that SFFA therefore cannot reasonably be read as prescribing such a rule). But once that maximalist interpretation is ruled out, we have now suggested, it is not clear how the Court’s disapproval of affirmative action could be reduced to any principled (or perhaps even discernible) rule. In particular, we have tried and failed to identify a principle, implicit in SFFA, that would vindicate the perception of the opinion as a fundamental departure from the Court’s precedents in this area, on the one hand, while also respecting the commonsense limits that the Court placed on its own holding, on the other. As a result, we are left genuinely uncertain how a university that wants to comply with the Court’s holding, and that takes to heart its admonition that “the prohibition against racial discrimination is ‘levelled at the thing, not the name,’”⁸² is supposed to distinguish permissible from impermissible consideration of race—or how a future Court that wants to build on SFFA’s foundation is supposed to do that either.

Still, there is another possible way of reconciling the Court’s conflicting commitments in this area that we have not yet considered. We offered in part I some examples of the forms of race consciousness that the SFFA Court seemed committed to permitting. But while all of our examples there do involve consideration of an applicant’s race, they also all involve essays authored by the applicants themselves. Of course, we framed the examples that way in keeping with the Court’s framing of its own opinion, which left universities free to “consider[] an applicant’s discussion of how race affected his or her life.”⁸³ Having adopted an essay format, though, we then focused on the content that is conveyed and what a university is free to make of it—and perhaps that misses the Court’s point. Again, in all of our examples, there is a sense in which it is “an applicant’s discussion” that the university is responding to, and not (at least directly) the facts themselves. So perhaps what the Court’s closing discussion conveys (and what our examples illustrate) is not that certain ways of considering race are permissible, but rather that considering the applicant’s discussion of race-related matters is, to the Court’s mind, categorically distinct from considering those matters themselves.

What would ground or motivate such a distinction? Intuitively, perhaps it is one thing for a university to offer applicants a menu of racial labels—so that it may then ascribe whatever significance it attaches to their avowed race—and another for the university simply to stand open to hearing from applicants whatever they have to say, including about how their race (meaning, really, their membership in a racialized group)⁸⁴ matters to them or should inform the university’s consideration.⁸⁵ At least, we can appreciate how the latter posture might strike one (or might have struck the Court) as open-minded, noncommittal, detached, even race neutral, in ways that the former is not. Here again, suspect descriptions employed in policing offer a useful analogy. Some courts have thought it significant that police officers “were not the source” of a suspect description that “included the fact that the suspect was African-American”: the officers “did not exercise their own judgment in the selection of the suspect,” the thought goes, but were merely passively responding to the facts put before them by private citizens.⁸⁶ We have our doubts about the aptness of that characterization of that situation. But the intuition underlying that take on the use of racial suspect descriptions strikes us as importantly related to one that might undergird a special allowance for universities to consider “[an] applicant’s discussion of” race-related matters that the university may not consider in an unmediated way.

The question, then, is whether this way of reconstructing the line that the Court has undertaken to draw in this area is more principled, or at least more workable, than the approaches that we found wanting above. To answer that question, we can return again to our example of an essay by a Black applicant (Applicant C) that discusses racial isolation and stereotype threat.⁸⁷ We have already shown that an assessor who comes away from that essay with a view of the student’s grit (or “courage and determination”)⁸⁸ is necessarily bringing to bear their understandings about the experience that the student described—understandings that, in the nature of things, cannot be individualized to “that student.”⁸⁹ The thought on the table now is that this might be permissible because, given the essay format, the individual applicant is really the moving force in the process—whereas, in the cases that the Court would denounce as “stereotyping,” that is not the case.

But that effort to distinguish the various cases does not withstand scrutiny. Granted, it is true that when the assessor makes an assumption about Applicant C’s experience—namely, that it was basically like those that Black students typically undergo in like circumstances—the assessor is doing only what the applicant has tacitly invited them to do. After all, Applicant C has described an experience at a certain level of generality—including their race and certain other details, but not including various others. It is natural to suppose that if the writer did not want the reader to envision an ordinary or typical instantiation of that scenario, the writer would have said either more or less. This sort of pragmatic inference is an important part of how communication works in general.⁹⁰ In our example, then, the reader forms a view about “that student’s courage and determination”⁹¹ based partly on race-related information (what the racially defined experience is usually like) that does not come “from” the student, but which the student has nonetheless invoked and invited the reader to bring to bear on their case. And perhaps that does allay any concern that the university has imposed its own conviction “that the touchstone of an individual’s identity” is “the color of their skin” and sorted Applicant C accordingly—rather than engaging with Applicant C’s perspective, “as an individual,”⁹² about the relevance of race to understanding their experiences or assessing their qualifications.

If this is what licenses the race-based inference about Applicant C, however, essays even more schematic than theirs ought to suffice too. Consider these possibilities (the first of which is unchanged from above), and suppose that each is offered as an answer to a question about experiencing adversity:

Although some of these may be better than others in the sense of being more effective as essays, we do not see any principled distinctions to be drawn in terms of whether each applicant has individually invited the race-related inferences that the university is apt to draw—or even in terms of whether the relevant information about their probable characteristics is coming “from” the applicant (as opposed to the store of knowledge that the applicant naturally presumes to be common ground). Applicant F, for instance, is making the same kind of communicative leap of faith as Applicant C: Rather than specifically invoking stereotype threat, Applicant F is supposing that the reader will already know why the sort of racial isolation described is ordinarily difficult—or, at least, will have an adequate grasp of how difficult it is, even if the reader (perhaps like Applicant F) lacks a rich understanding of why. Likewise, Applicant F leaves implicit just how this adversity contributed to their character, rather than singling out “courage and determination.” These losses of nuance may make the essay less impressive qua essay, and, depending on the reader, they may also make the essay less effective at conveying the degree of challenge that the writer underwent (or the positive effects that this had on them). But none of that suggests any difference in terms of the permissibility of the basic inference about the applicant’s personal qualities that the assessor is being invited to draw. (Recall that Applicant C, too, is relying on the reader to fill in significant and inevitable gaps in the usual way.)

The move from Applicant F to Applicant G is not different in kind. Some experiences are widely shared among young Black men who grow up in our racially stratified society at a certain time, and Applicant G is essentially telling the university that he is no exception—that what is true of most who share these qualities with him is also true of him—and inviting the university to afford that fact about him, “as an individual,”⁹³ the epistemic significance that it is actually due. Thus, a university that proceeds as Applicant G contemplates does not commit to the view “that race in itself ‘says [something] about who you are’”;⁹⁴ it simply permits an individual applicant to convey, using the shorthand made possible by common knowledge of certain facts about the social world, the sorts of experiences that he has had and how they have affected him.⁹⁵ The logical implication of the view that we are entertaining, then, would be that equal protection does not bar the university from “consider[ing] [Applicant G’s] discussion of how race affected [his] life” in this way either.⁹⁶ At that point, though, the distinction between an essay and a clearly optional checkbox—labeled so as to convey essentially the same content that Applicant G does in his essay—would come awfully close to the kind of form-over-substance rule that SFFA disavowed.⁹⁷

A slightly different thought experiment may help to underscore our point here. Suppose that a researcher canvassed the literature on racial hierarchy, stigma, and stereotyping in the contemporary United States and developed a list of the causal mechanisms through which being regarded as Black typically (though not necessarily always) operates to a person’s disadvantage. Like with any such list, the causal mechanisms could be individuated in a more or less fine-grained manner. Just to illustrate the point, we will posit a few nested descriptions of one such mechanism:

• I. 

  living with a stigmatized identity

  • I.A. 

    being judged negatively on grounds unrelated to one’s merits

    • I.A.1. 

      being judged negatively, by one’s teachers, on grounds unrelated to one’s merits

      • I.A.1.i. 

        being judged negatively, by one’s second-grade teacher, on grounds unrelated to one’s merits

Now suppose that, for each verbal formulation on our researcher’s hypothetical list (of which these represent just a tiny fraction), some Black college applicant asserts that they have had this experience and that it has affected them in the typical way. For some of the descriptions, we take it that SFFA clearly permits a university to credit that assertion and to take the experience as relevant to assessing the applicant’s qualities. But we cannot imagine how the Court would specify a point—a depth in traveling down the tree—at which this license kicks in. (How could it be that one has to specify that it was one’s second-grade teacher, perhaps among others, who operated as a vector of racial stigma in one’s childhood? But if one does not, why would one have to specify that it was teachers, as opposed to others, who did so?) Nor can we see what would justify such a requirement—in effect, a rule that, like allegations of fraud in a federal complaint, experiences of racism must be pled with particularity in a college essay—in the first place.⁹⁸ Again, the natural conclusion is that an applicant who simply affirms that their race has figured in their life in the usual way—or the way that is usual for Black Americans—has given the university relevant information that it is free to use, in the context of their entire application, in making inevitably rough judgments about “that student’s unique ability to contribute to the university.”⁹⁹

Stepping back from the examples, our basic point is simple. The allure of distinguishing “an applicant’s discussion” from the facts that the applicant might discuss seems to be a sense that the applicant, not the regulated actor, is then responsible for any significance attached to their race or their race-related experiences. But, in fact, the exercise of making sense of an applicant’s “discussion” is—like communication in general—fundamentally cooperative. Whatever the applicant says, they will be counting on the reader to fill in various details based on the common ground that the two share as inhabitants of the same social world. Unless information about race were somehow purged from that reservoir of context, applicants who identify themselves by race will be trading on the reader’s sense, not the applicant’s own, of the relevance of the facts that the applicant conveys. Yet we still lack any principle that could guide judgments about when, if ever, inviting the reader to bring this context to bear (or taking up such an invitation) is impermissible.¹⁰⁰

In theory, excising race from the communicative context altogether would obviate this intractable line-drawing problem—but, as will probably be clear by now, that is a nonstarter. Our main argument to that effect is simply that it would render unintelligible all of the examples with which we began (and which we take the Court to have effectively approved).¹⁰¹ But we cannot help also remarking on the artificial, absurd exercise that such a regime—in which applicants cannot take for granted any understandings about the usual significance, in some context, of being of the race that they are—would create for applicants.¹⁰² Imagine, for instance, a version of Applicant C’s essay that undertook to explain, for the benefit of a reader feigning ignorance, what “all of the worst stereotypes about people who look like me” actually are. Such an essay would then need to persuade that reader that, in light of the history and current reality of anti-Black racism in the United States (which it might also fall to the applicant to describe), the fear of behaving in ways that could seem to confirm those stereotypes really does tend to weigh more heavily on Applicant C than it might on others who do not share their racial identity and so are not subject to the same expectations.

The objections to such a regime practically make themselves. One glaring problem is that placing these artificial explanatory burdens on applicants would disable them from writing good essays qua essays—from demonstrating the skills, including perceptiveness about the reader’s frame of mind and the ability to tailor one’s expression to it, that good writing involves.¹⁰³ Even apart from that, the exercise would be a hollow and futile one: Admissions officers could not actually constitute themselves as blank slates when it came to race, with all of the naïveté (and, we suppose, gullibility) that would entail. Ultimately, simple charity tells us that this cannot have been what the SFFA Court contemplated: It cannot have meant that the university may “consider” an applicant’s discussion of how the social reality of race affected them, but only as a Martian observer unacquainted with any of the patterns and social meanings that make race meaningful in the first place would.

The difficulty of discerning principled lines from the Court’s discussion in SFFA, we have claimed, reveals more than a failure of craft. It manifests a deep tension—one that the majority may have believed it could finesse, or that it may simply have failed to appreciate, but that its ruling has nonetheless brought closer to the surface of equal-protection doctrine than ever before. At one pole, the Justices who formed the majority have long believed that governmental consideration of race is a profound evil—and SFFA was supposed to represent the long-delayed triumph of that conviction. Yet the very fact of that impending victory seems to have engendered a new awareness that, as Justice Kennedy had once warned, “postulate[s]” about colorblindness are sometimes “not sufficient” in “the real world.”¹⁰⁴ Among other things, the trial record included various examples of students’ essays (and testimony) explaining the role of race in their lives.¹⁰⁵ To bar those students of color from conveying the whole of their experiences and aspirations—in an admissions process geared to assessing the personal qualities of each student “as an individual”—must have struck some of the Justices as a bridge too far. So the majority said, in effect, that considering race is categorically forbidden, but nothing in its opinion should be taken to cast doubt on the plainly reasonable practice of genuinely attending to individual people’s relevant characteristics. What the Court seems to have failed to appreciate is that—thanks to features of our factual situation that the Justices are powerless to alter—these two commitments are flatly inconsistent.¹⁰⁶

“Many of our most serious conflicts,” John Rawls observed, “are conflicts within ourselves.”¹⁰⁷ Although our commitments to broad principles and our convictions about particular cases are both “capable of having for us … a certain intrinsic reasonableness,” it is not uncommon for them to fail to align.¹⁰⁸ And when we encounter such a conflict, reasonableness demands that “some of these judgments must eventually be revised, suspended, or withdrawn.”¹⁰⁹ For this reason, Rawls argued,

In a sense, SFFA revealed that the Court’s conservative majority could not live with its own abstract commitment to colorblindness. But the Court did not squarely or openly reckon with that predicament. As far as the opinion reflects, the Justices in the majority did not interrogate why the forms of race consciousness that seemed plainly reasonable to them struck them that way, or how their own pat denunciations of “stereotyping” or of valuing “race qua race” might need to be qualified in light of those new insights. “Those who suppose their judgments are always consistent,” Rawls noted, “are unreflective or dogmatic.”¹¹¹ They are also ill-positioned to craft principled, sound, and workable doctrines of constitutional law.

The dilemma that the Court faces in the wake of SFFA is apt to prove especially intractable because it is difficult to imagine the current Court making either of the revisions that coherence would demand. The Justices could, of course, more candidly and fully retreat from their commitment to colorblindness: for instance, they could openly permit universities to account for the obvious relevance of people’s memberships in racialized groups, even if not to value the very fact of anyone’s race as such.¹¹² Alternatively, they could abandon the idea that applicants should be treated “as individuals”—with each evaluated based on their own “unique ability to contribute to [a] university”—and fully embrace the injustice of excluding plainly relevant information about many applicants, especially those whose experiences and aptitudes are most shaped by the social reality of race. As the Court tries to build on SFFA’s foundation in the years to come, we expect that it will become only more clear that there is no principled middle ground.