Why shouldn’t vote weight matter in drawing district lines, “especially when voter-based apportionment…is being substantially disregarded?,” Justice Anthony Kennedy asked Texas Solicitor General Scott Keller at oral argument this morning in Evenwel v. Abbott.

Justice Kennedy’s question hit the nail on the head. The case is a challenge to Texas’s practice, shared by the other states, of drawing district lines based on raw population, rather than voter population, with the result that the value of an individual vote can vary wildly based on the number of nonvoters included in different districts. Plaintiffs Sue Evenwel and Edward Pfenninger live in districts with relatively few noncitizens. When voting in state senate elections, their votes are worth 50 percent less than a vote cast in Brownsville, right up against the Mexican border. All of the districts have about the same raw population. But the Brownsville district contains more non-voters, including undocumented immigrants, thereby boosting the voting power of district residents who are able to vote.

Ms. Evenwel and Mr. Pfenninger’s complaint is straightforward enough: how on earth is that disparity consistent with the Court’s “one-person, one-vote” doctrine, which holds that “an individual’s right to vote for state legislators is unconstitutionally impaired when its weight is in a substantial fashion diluted when compared with votes of citizens living in other parts of the State.”

Many observers expected today’s argument to be a rout for the plaintiffs, but it didn’t quite turn out that way. The Court’s liberals peppered plaintiffs’ counsel Will Consovoy with questions regarding the Court’s precedents (which generally apply raw population data), the apportionment of House seats among states by raw population, the interest in “equal representation,” and the availability of data to apportion based on data correlated with voter population.

Consovoy parried them with ease. As he explained, while the Court’s cases had generally assumed that raw population would work well enough, they never directly considered what to do when voter population and raw population diverge to the extent seen here. By contrast, those decisions are crystal clear that the underlying rational of “one-person, one-vote” is enforcement of the right to vote by ensuring that all votes are equal and none diluted. The Court, Consovoy explained, has long rejected the analogy to House apportionment, on the basis that dividing House seats among the states implicates very different interests, including tax apportionment, state control of voter qualification, and the relative balance of power among the states. And as for data, Texas actually used citizen voting age data from the Census’s American Community Survey when it drew its districts—as basically every state does, in order to comply with the Voting Rights Act.

It was apparent that Justices Ginsburg, Kagan, and Sotomayor had come prepared with questions meant to trip up the plaintiffs’ argument. But none of them gained traction.

However, Keller, representing the State of Texas, faced more skeptical questioning from the Chief Justice and Justice Kennedy, who were clearly troubled by the sheer magnitude of the vote disparity on display, far greater than any the Court had previously upheld. Keller’s primary point was that, while there might be discrimination, it wasn’t “invidious discrimination,” which is what the Fourteenth Amendment’s Equal Protection Clause protects against. States, he argued, have the power to choose among different views of the nature of representation as a “core sovereign function.” But his argument that states could legitimately choose to take into account vote-equality interests was met with the inevitable question, by Justice Kennedy: if those interests have merit, why would states be licensed to ignore them?

And, as the Chief Justice noted, the Court cannot ignore that the doctrine it is applying is called “one-person, one-vote.”

Finally, Ian Gershengorn appeared at the lectern for the United States. While nominally supporting Texas, the U.S. resisted its view that states have a choice and would instead limit their options to raw population, with the possibility of adjusting it in unusual circumstances. That view drew even greater skepticism than the idea that states could ignore vote weight, with Justice Alito questioning, “It’s your argument that voters are irrelevant?” Gershengorn attempted to highlight, with mixed success, the potential disruption that would result from a decision requiring consideration of voter populations.

Apparently troubled by the large vote-weight disparities in Texas and other states with large undocumented populations, the Chief Justice several times pitched a possible compromise position that states would be required to consider vote weight only when necessary to avoid disparities of 20 percent or more. That would, in effect, exempt many or most states and localities from any additional burden, while still preserving the core “one-person, one-vote” doctrine.

The outcome is impossible to predict. The Court’s liberals appeared set against the plaintiffs’ vote-equality argument, while Justice Alito seemed to be on the other side. Justice Thomas and (uncharacteristically) Scalia were silent, although Thomas has previously expressed sympathy for the plaintiffs’ position. And in the middle are Justice Kennedy and Chief Justice Roberts, whose compromise view has a chance, and perhaps even more than that, of prevailing.