When the Supreme Court in late April decided the case Schuette v. Coalition to Defend Affirmative Action, it again directed attention to what is surely one of the most peculiar issues in American politics. The Court’s own opinions on the use of affirmative action in higher education have been frequently confusing and unsatisfactory. Yet this muddle accurately reflects Americans’ broader ambivalence about a policy that our own political ideals might seem both to endorse and to condemn.

The Court’s confusion has been evident since 1978, when it decided Regents of the University of California v. Bakke, issuing its first major opinion on the use of affirmative action in higher education. Bakke examined a policy of the University of California Medical School at Davis, which reserved sixteen seats out of one hundred in its total entering class for qualified minority applicants. In deciding the case, the Court, on the one hand, upheld the use of affirmative action by a university as a means of assembling a diverse student body, but, on the other hand, struck down the particular system in use at the UC-Davis Medical School. Although a university might fairly treat race as one among various “plus” factors taken into account when considering a particular application, the Court argued, each student was nevertheless entitled to individualized consideration. A rigid numerical quota such as that at Davis did not grant such consideration to every student on an equal basis, because while minority applicants could compete for all one hundred of the available seats, white applicants could compete for only eighty-six of them.

Ironically, however, only one justice on the Court, Lewis Powell, Jr., actually held this view. Four justices would have simply rejected the use of race in admissions decisions altogether. Four others thought not only that the use of racial preferences was constitutional, but also that the particular program at UC-Davis passed muster. Powell joined the former group in striking down the medical school’s program, but he joined the latter in agreeing that race could be validly taken into account in order to achieve the goal of diversity. Powell’s middle-of-the-road position—that universities may consider race as long as they do not employ a numerical quota system—thus determined the outcome of the case, even though not one other justice agreed with him.

In 2003, Bakke was replayed in almost uncanny fashion when the Court decided a pair of companion cases, Gratz v. Bollinger and Grutter v. Bollinger. Both involved challenges to affirmative action policies at the University of Michigan. Gratz dealt with the university’s undergraduate admissions program, which employed a “points system,” in which applicants received up to 150 points for a variety of factors. Applicants receiving a total of at least one hundred points were admitted, and students from underrepresented minority groups were automatically awarded twenty points, one fifth of the total required for admission. Grutter, the second case, took up the corresponding policy at Michigan’s law school. Instead of using a points system, the law school gave each applicant “individualized, holistic review... giving serious consideration to all the ways”—including race along with other qualities—“an applicant might contribute to a diverse educational environment.”

With Justice Sandra Day O’Connor this time playing the role of Lewis Powell, the Court again managed to split the difference. It struck down the undergraduate admissions policy at issue in Gratz, holding that the automatic assignment of twenty points to every minority applicant was akin to Bakke’s quota system in denying each student equal and individualized consideration. But it upheld the law school’s policy in Grutter, arguing that the holistic review employed there satisfied the requirements of equal protection. As in Bakke, however, the Court fractured badly; only one other justice, Stephen Breyer, joined O’Connor in striking down the one policy and upholding the other. Four justices would have struck down both programs; three justices would have upheld both. The victorious middle-ground position thus had less support than either ­alternative.

Schuette in turn arose directly from the controversy in Gratz and Grutter. Following those decisions, Michigan voters approved a 2006 ballot proposition that amended the state constitution in order to forbid the use of affirmative action in public education, including the state’s public universities. Opponents challenged that outcome, arguing that the amendment to Michigan’s state constitution was invalid because it violated the Equal Protection Clause of the United States Constitution. The Supreme Court in Schuette rejected their position, holding that Michigan voters were indeed entitled to prohibit the use of affirmative action if they so desired.

The issue in Schuette is thus somewhat different from that in Bakke or Gratz/Grutter: not whether affirmative action is constitutional, but whether voters may amend their state constitution in such a way as to deny public universities the option of employing affirmative action. As Justice Kennedy put it in his opinion for a plurality of the Court, “This case is not about how the debate over racial preferences should be resolved. It is about who may resolve it.” Six justices agreed that the voters were entitled to resolve it through a popular referendum and constitutional amendment. Yet these six produced four separate opinions among them, not one of which received the support of more than three justices. And the longest opinion of all—longer than all of the others combined—was an impassioned dissent from Justice Sonia Sotomayor, who complained bitterly that the Court was being blind to the reality of ­continued racial inequality.

The Court’s efforts to deal with affirmative action have thus been noteworthy for their consistent inconsistency. They are consistent in hewing to what I have called a middle-of-the-road position. In somewhat simplified but not inaccurate form, we might summarize it thus: some consideration of race is acceptable, but numerical quotas are not. Because the Court has been so deeply fractured, however, with no approach able to win the support of a majority of justices, it has failed to delineate any consistent rationale for its decisions. From a legal point of view, this is unsatisfactory. For the sake not only of voters and politicians, but also of lower courts seeking to apply precedents in subsequent cases, it would be preferable for the Court to settle on a clear rationale that could predictably indicate what is permitted and what is not. From the standpoint of moral or political philosophy, however, there may be something to be said for the resolution the Court has—almost despite itself—reached, even if few individual justices have endorsed it. We can see this by considering the arguments about affirmative action that are standard in our political debates. For, much like the Court’s decision, these arguments are also frequently muddled and unsatisfactory.

Conservatives typically oppose affirmative action or any use of race-based preferences. They like to cite Justice John Marshall Harlan’s ringing dissent in Plessy v. Ferguson: “Our Constitution is color-blind, and neither knows nor tolerates classes among citizens.” To affirm with the Declaration of Independence that all men are created equal means that every citizen is entitled to be judged on his or her own merits, and not simply lumped together with other members of a race, as if one’s personal talents, interests, and experiences were all outweighed by the single fact of skin color. To be sure, the United States long failed to live up to this fundamental principle of equality. But we did not fight a civil war and amend the Constitution merely to replace slavery’s brutal caste system with a morass of ethnic preferences that perpetuate racial balkanization instead of overcoming it.

This is a familiar argument, but it is in some ways a strange one for conservatives to make. Conservatives emphasize the value of tradition, of inheriting from our forebears a way of life and its corresponding values. They stress the importance of patriotism and of national pride. They affirm Edmund Burke’s famous description of society as a contract stretching across generations: “a partnership not only between those who are living, but between those who are living, those who are dead, and those who are to be born.” Yet this perspective arguably lends support to affirmative action. For if we are part of an ancestral community reaching back into the past and forward into the future, then we bear some continuing responsibility for its faults. We cannot take pride in our country’s achievements without also accepting our share of guilt for the injustices it has done. And if we have benefited from those past injustices, then we have some obligation to address their consequences in the present.

Liberals, by contrast, typically defend affirmative action as an appropriate way of leveling the racial playing field by giving a hand up to minorities who have suffered various forms of oppression, especially the centuries of slavery and segregation inflicted upon African-Americans. Yet the obstacles that such policies create for less advantaged white Americans—the most advantaged, of course, are not the ones who bear affirmative action’s costs—rest uneasily with core liberal values. Liberals are deeply reluctant to see social costs imposed upon people without clear individual responsibility. We should not assume, for example, that criminals “deserve” severe punishments, or welfare recipients a place at the bottom of the economic totem pole, without first considering whether factors beyond their control—poor education, broken homes, the loss of low-skill jobs due to globalization—have left them unfairly disadvantaged. Yet a white student who has met the qualifications for college admission but because of her race is denied a place given instead to an equally or less qualified minority applicant—the situation of Jennifer Gratz in the case bearing her name—is unquestionably being burdened on the basis of something for which she bears no ­responsibility.

Indeed, the insistence that we not distribute social benefits or burdens on the basis of undeserved factors is at the heart of the most creative and influential contemporary statement of liberal political philosophy. In A Theory of Justice (1971), John Rawls described how principles of justice should be deduced from the “original position,” a hypothetical situation in which we know nothing about ourselves that might unfairly influence our choice of principles: not our sex, or our economic class, or our particular skills, or, of course, our race. These factors are all, Rawls says, “arbitrary from a moral point of view” and thus should not be allowed to bias our conception of justice. It is difficult to see, however, how such a formula could produce anything but color-blind principles of justice. The white college applicant clearly bears no responsibility for his race, and to disadvantage him because of it is, as Rawls might say, to treat him not as an end, but rather as a means to the achievement of a social goal. Justice cannot mean that we must treat blacks as ends but may treat whites as means. We might expect liberals, rather, to borrow a line from Justice John Roberts, who, in a decision dealing with segregation in Seattle schools, said, “The way to stop discrimination on the basis of race is to stop discrimination on the basis of race.”

One could, of course, give different accounts of both the conservative and liberal positions than I have given here. But these arguments suffice to illustrate the problematic nature of affirmative action, for they show that one could have good reasons for expecting conservatives to support and liberals to oppose it. That the opposite is the case merely reveals how difficult it is to resolve the controversy over ­affirmative action on the basis of political principles familiar to us from our ordinary political discourse. Perhaps, indeed, those principles do not determine a particular, “correct” resolution of the issue at all. Perhaps, instead, we are simply thrown back upon the need for prudential judgments about what will best serve the nation and heal its racial wounds over the long run.

In that spirit, we might look to one last feature of Justice O’Connor’s opinion in Grutter for a token of wisdom. While upholding the University of Michigan’s use of race in law school admission, she also indicated that the justification for such racial preferences grows weaker over time, so that at some point they will no longer be compatible with the Constitution’s demand for equal protection. She even sought to set a deadline: “It has been 25 years since Justice Powell [in Bakke] first approved the use of race to further an interest in student body diversity in the context of public higher education.... We expect that 25 years from now, the use of racial preferences will no longer be necessary....”

This sounds more like a legislative than a judicial judgment, and as a piece of constitutional interpretation it is perhaps rather unsatisfactory. Yet it may contain a great deal of political prudence. If society is indeed a partnership “between those who are living, those who are dead, and those who are to be born,” then we do indeed have a duty to redress our ancestors’ wrongs. At some point, however, the claims of those “yet to be born” become compelling. As Justice O’Connor suggests, there is a kind of statute of limitations—perhaps lacking a clear cut-off date, but existing nonetheless—on the demands of intergenerational justice. It is one thing for the sins—the political sins, at any rate—of the fathers to be visited upon their sons, another for them to be visited upon their grandsons and great­grandsons, as if the debt could never be paid. If those who are living accept responsibility for the sins of those who are dead, we may look forward to the day when those who are yet to be born, white and black alike, will finally find their past debts paid off. Or at least forgiven.

Peter Meilaender is Professor of Political Science at Houghton College.