In a 4-3 decision, the Supreme Court has upheld the use of race in admissions at the University of Texas-Austin. Swing justice Anthony Kennedy unexpectedly joined the liberals and wrote the majority opinion in the case, Fisher v. University of Texas; with Antonin Scalia’s death and Elena Kagan’s recusal, there were only seven votes.

The majority opinion is incoherent, to the point of being downright comical in places. But it is weirdly of a piece with the Supreme Court’s previous decisions on this topic—and a good opportunity to look back at the road that got us here.

The case revolves around UT’s complicated admissions system. Most students are admitted through the “Top Ten Percent Plan,” which allows students graduating at the top of their high-school class to attend any state-funded university they want. (Thanks to a 2009 reform, the Austin campus is now allowed to cap automatic admissions at 75 percent.) The remaining slots are allocated according to scores on an “Academic Index” and a “Personal Achievement Index.” The latter takes race into account.

In 2008, Abigail Fisher was denied entry and sued. Fisher alleged that the system violates the 14th Amendment’s guarantee of “equal protection of the laws.” Not only does it discriminate on the basis of race, but it does so on top of two other measures that do a good job of promoting diversity by themselves, the Top Ten Percent Plan and the non-racial elements of the Personal Achievement Index (such as socioeconomic status).

Owing to UT’s unique circumstances, Fisher was able to make this case quite forcefully. For a period around the turn of the century, an appeals-court decision had forced UT to use race-blind admissions. The state passed the Top Ten Percent Plan in response, and in 2000, the university proudly announced that “enrollment levels for African American and Hispanic freshmen have returned” to their previous level. But when the Supreme Court okayed the use of race in 2003, the university injected it back into the system, seemingly just because it could. (Nonetheless, Fisher was perhaps not the ideal plaintiff: Even if she hadn’t been white, she probably wouldn’t have gotten in.)

Her case made it to the Supreme Court in 2013, but it was anticlimactically sent back to a lower court with some instructions. Last Thursday, the nation’s highest court finally put the litigation out of its misery.

The upshot is that the UT system is constitutional—but any legislator, university administrator, or lower-court judge looking for broader guidance as to what’s legal and what’s not will be sorely disappointed. The opinion does more to confuse than it does to clarify the rules, building on the tensions and contradictions of previous decisions.

For example, the Court saddles the university with “a continuing obligation to … periodically reassess[] the admission program’s constitutionality, and efficacy, in light of the school’s experience and the data it has gathered since adopting its admissions plan, and by tailoring its approach to ensure that race plays no greater role than is necessary to meet its compelling interests.” Yet those compelling interests are not “in enrolling a certain number of minority students, but … in obtaining ‘the educational benefits that flow from student body diversity.’”

So universities must track their “student body diversity” using “data” to make sure they don’t emphasize race any more than needed to achieve their goals, but they can’t have a “certain number of minority students” as a goal.

And to confuse matters even further, schools’ affirmative-action goals must be “sufficiently measurable to permit judicial scrutiny of the policies adopted to reach them.” In Kennedy’s estimation, the University of Texas achieved this with the “concrete and precise goals” of “ending stereotypes, promoting ‘cross-racial understanding,’ preparing students for ‘an increasingly diverse workforce and society,’ and cultivating leaders with ‘legitimacy in the eyes of the citizenry.’” Only a minority of Supreme Court justices, given voice in a lengthy dissent by Samuel Alito, thought something seemed off about this. (“For instance, how will a court ever be able to determine whether stereotypes have been adequately destroyed?”)

You can’t make this stuff up. Unless, of course, you build up to it slowly over the course of several decades, the way the Supreme Court did. Its first major affirmative-action case, Bakke, dates to 1978, and it was full of the same type of odd reasoning.

The Court took a very particular approach to “equal protection,” subjecting the use of race in admissions to “strict scrutiny”—a legal term meaning it must be narrowly tailored to serve a compelling state interest. This case is where the ban on “certain numbers”—outright quotas—came from. It also held that schools couldn’t try to remedy the broad historical disadvantages that certain racial groups have faced, which is probably the number-one argument that affirmative-action supporters employ when they’re not in court. What schools could do is use preferences to promote the supposed educational benefits of diversity.

Subsequent cases—notably two decided in 2003—were variations on this theme. Grutter revolved around the question of whether one school’s policy, which involved a nebulous “plus” factor for race designed to enroll a “critical mass” of minorities, was in reality a veiled quota system. (The Court said it was okay.) Gratz involved a system that blatantly gave an extra 20 points—on a 150-point scale—to applicants from underrepresented groups. (This was not okay.)

As Ramesh Ponnuru has pointed out, there was another option for the courts. The 14th Amendment didn’t even need to come into play—indeed it shouldn’t have, because it was not originally intended or understood to ban all government distinctions based on race. Some legal scholars would even do Ponnuru one better, arguing that originally, the 14th Amendment didn’t forbid outright school segregation, much less affirmative action. (While that may sound like the rant of a deranged racist, this point is often part of liberals’ case for a “living Constitution.”)

Courts should have looked not to the Constitution but to the Civil Rights Act of 1964, which states unequivocally: “No person in the United States shall, on the ground of race, color, or national origin, be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any program or activity receiving Federal financial assistance.”

If the Supreme Court had invalidated affirmative action on statutory grounds in 1978, Congress would have been forced to have an actual debate about the statute—specifically, about whether it should ban affirmative action along with other forms of racial discrimination. But in Bakke, the Supreme Court simply shunted the Civil Rights Act of 1964 aside, claiming the language did nothing but implement the 14th Amendment and therefore need not be considered separately.

Today, we have an odd kind of supervised federalism, where states and schools are free to use affirmative action, but where courts periodically swoop down and invalidate policies, doubling down on a vague, decades-old precedent that never made sense to begin with and certainly doesn’t reflect the laws Americans actually voted for.

That’s how we end up with a decision like Fisher—where justices fight over how much they should defer to universities’ opaque, shifting judgments of how much diversity they need; where schools are allowed to pursue a “critical mass” of minorities but not a “certain number” of them; and where institutions are instructed to go through the pretense of repeatedly “reassessing” the racial preferences they have used for years and have no intention whatsoever of discarding.

Robert VerBruggen is managing editor of The American Conservative.