As an election that gave us a President with nearly three million fewer votes than his opponent should make clear, there are huge defects in the democratic quality of American institutions. One of the most important and nefarious anti-democratic features of American legislatures today—including Congress—is the partisan gerrymander, which can, and often does, significantly overweigh some votes while devaluing others. On Monday, the Supreme Court announced that it would be hearing a case concerning the constitutionality of the partisan gerrymander. This case, Gill v. Whitford, may well end up being remembered as a voting rights landmark. However, it might instead reveal the 2016 election as a tragic moment for American democracy that allowed opponents of voting rights to consolidate their hold on power for the foreseeable future. As is so often the case, which of these possibilities comes to fruition will depend on Justice Anthony Kennedy.

Gill involves a particularly egregious case of gerrymandering. In 2010, a unified Republican government hired consultants to use sophisticated computer software in order to redraw districts in a way that would maximize Republican seats in the legislature, by concentrating likely Democratic votes and carefully diffusing Republican ones. The results were remarkable. In 2008, 29 districts in the state Assembly were within three points of the state’s presidential vote; in 2012, the first election after the gerrymander, there were only seven. And these changes produced the intended skew: Wisconsin Republicans got less than 49 percent of the vote—but won 60 of the seats in the state’s 99-member Assembly. Since then, Republicans have continued to parlay a nearly evenly split electorate into huge legislative majorities.

The potential constitutional issues with partisan gerrymandering should be obvious. An electoral map that awards 60 percent of the seats of a legislature to a party that got less than 50 percent of the vote in a two-party race is at complete odds with the essential holding of the great Warren Court decisions Baker v. Carr and Reynolds v. Sims. These decisions held that when states deliberately refused to redraw districts to reflect population shifts, with the inevitable result of effectively disenfranchising urban voters, they violated the equal protection clause of the Fourteenth Amendment. Reynolds held that in all elections, except those for the United States Senate (which is permanently malapportioned by the Constitution), legislative districts had to be drawn to reflect a “one person, one vote” standard. It is hard to see how districting that intentionally overrepresents one group of voters and underrepresents another can be constitutional under these precedents.

And yet, the Supreme Court has allowed partisan gerrymandering to get worse and worse over the years. In the 2004 case Vieth v. Jubelirer, the Supreme Court declined to rule a partisan gerrymander of the Pennsylvania legislature unconstitutional. Four justices—Antonin Scalia, William Rehnquist, Clarence Thomas, and Sandra Day O’Connor—held that partisan gerrymanders were inherently “non justiciable.” That is, even if partisan gerrymanders are unconstitutional, they are inherently a “political question” that cannot be resolved by the courts. According to Scalia’s plurality opinion, the appropriate remedy to a partisan gerrymander is the power given to Congress in Article 1, §4 to alter legislative districts, not judicial review.

But like the pre-Baker argument that all districting claims were nonjusticiable, this argument is transparently specious. For example, Democratic Congresses in the first half of the twentieth century had no incentive to alter malapportioned districts that favored state Democrats (and, therefore, also favored House incumbents because state legislatures draw congressional districts.) Petitioning a Republican Congress and President to remedy constitutional gerrymandering that favors Republicans in 2017 will be equally futile. This is precisely the kind of failure of ordinary democratic processes where judicial review is most defensible.

And yet, Thomas remains on the Court, and Chief Justice Roberts and Justices Alito and Gorsuch are nearly certain to join him in holding that partisan gerrymanders are political questions. The four Democratic nominees on the Court are nearly certain to hold the Wisconsin gerrymander unconstitutional, although it is not clear whether they will be able to reach a consensus on how to adjudicate cases in which the facts are less extreme. The fate of this case, and of restoring democracy to the Wisconsin legislature will, therefore, be left in the hands Anthony Kennedy, who will have to be convinced to hold the gerrymander unconstitutional.

Some precedent might be relevant here in weighing the possible outcomes. As Rick Hasen of UC Irvine Law School explains, Kennedy’s opinion in Vieth was idiosyncratic. On the one hand, he did not join Scalia and the other conservative justices in finding that partisan gerrymanders were inherently nonjusticiable. On the other hand, he rejected the standards that the liberal justices proposed using to determine when gerrymandering was unconstitutional. While implying that Fourteenth Amendment challenges to partisan (as opposed to racial) gerrymandering should be rejected, Kennedy suggested that “[t]he First Amendment may be the more relevant constitutional provision in future cases that allege unconstitutional partisan gerrymandering.” If “an apportionment has the purpose and effect of burdening a group of voters’ representational rights,” Kennedy asserted, it might violate the First Amendment.

The litigants challenging the Wisconsin gerrymander, therefore, will rely heavily on the First Amendment with an eye toward persuading Kennedy. And, as they’ve done on same-sex marriage cases, the liberal justices are likely to go along with any rationale that Kennedy would support to rule the Wisconsin gerrymander unconstitutional. The question is: Will it work?

It’s hard to say. The most encouraging sign for liberals is Kennedy’s decision to join an opinion earlier in the term that struck down a race-based gerrymander in North Carolina. This would indicate that, at least in some districting cases, Kennedy is willing to act against his typical partisan interests. On the other hand, Kennedy may draw a sharp distinction between race-based and partisan gerrymanders, holding the former unconstitutional while making challenges to the latter nearly impossible. While one shouldn’t read too much into it, Kennedy joining the four other Republican nominees on Monday to stay a lower court decision holding a partisan gerrymander unconstitutional over a dissent by the four liberals on the bench is at least mildly discouraging.

Hopefully Kennedy can indeed be convinced to rule the Wisconsin gerrymander unconstitutional. If he can’t, the Electoral College malfunction that allowed Donald Trump to appoint Antonin Scalia’s replacement on the Court may provoke further antidemocratic reverberations for years to come.