A number of correspondents have explained the paradox this way: Roberts and many others object to laws and policies (like affirmative action) that actually use racial terms in their language. A law that doesn’t do this is “neutral” and thus poses no problem.
This distinction is foreign to the Constitution. What the Equal Protection Clause as interpreted by the courts actually targets is “intentional discrimination.” Constitutionally, it doesn’t matter whether lawmakers write that purpose into the language or not. As anyone who has studied American history knows, it’s possible to further white supremacy without admitting anything publicly; that’s how Southern states kept the ballot all white for a century despite the 15th Amendment’s prohibition of race discrimination in voting. Jim Crow election laws didn’t say only whites could vote; instead, they set up a maze of restrictions—grandfather clauses, poll taxes, literacy tests—designed to winnow out non-whites and, to a lesser extent, poor people of any race.
A law that says, “whites can vote, non-whites not so much,” for example, is invalid on its face. But what if it says, “persons convicted of a crime of moral turpitude may not vote,” when that rule disenfranchises blacks far more often than whites? In a 1985 case called Hunter v. Underwood, the Supreme Court found that, neutral in language or not, the “moral turpitude” clause was invalid under the 15th Amendment. The historical evidence, wrote then-Justice William H. Rehnquist, “demonstrates conclusively that [the ‘turpitude’ provision] was enacted with the intent of disenfranchising blacks.” The sponsors of the measure had admitted in 1901 that they wanted, as one of them said, “to establish white supremacy in this State." The sponsors were all dead by 1985, but the passage of time didn’t matter, nor did the fact that it also disenfranchised some poor white voters.
The intent, not the language, is what matters under the case law. And it’s not necessary that those creating or enforcing the law actually hate minorities; all that must be shown is that they intended to treat minorities less favorably than the majority.
That’s what “discrimination” is. It needn’t speak its name; it needn’t wear a robe or burn a cross; it may speak in a soothing voice. Discriminatory lawmakers can have more than one motive. But if they intentionally target race, it is discrimination.
How do we determine “discriminatory intent”? Legislators, even those with bad intentions, tend to be more discreet today than they were in segregation-era Alabama. But the Supreme Court has developed ways of smoking out discriminatory intent. If you want a thorough introduction to how this is done, read Judge Nelva Gonzales Ramos’s 147-page opinion in the Texas-ID case, Veasey v. Perry. Ramos follows Supreme Court precedent carefully as she analyzes the history of SB 14. In a 1977 case called Village of Arlington Heights v. Metropolitan Housing Development Corporation, the Court instructed lower courts to ask certain questions when neutral-seeming laws are challenged. Do they affect minorities more than whites? Some civil-rights statutes forbid such a “disparate impact,” but under the Constitution impact by itself is never enough to prove discrimination. More evidence is needed. What is the history of discrimination in the city or state that enacted the law, and what was going on when the law was passed? Was the law enacted by unusual, hasty, or secretive procedures? What did the lawmakers themselves have to say about the law during the process?