Is the death penalty in its death throes?  A pair of Supreme Court cases this week suggests an interesting pattern is emerging. We see Justice Breyer, carrying a legacy from the late Justice Harry Blackmun, dramatically seizing particular cases to declaim that the Court needs to finally say that the death penalty as a whole can no longer survive constitutional scrutiny. We see the Court dodging such foundational questions but nevertheless dramatically issuing abrupt summary reversals of death sentences when state courts, by affirming  death sentences in the face of the Court’s minimal clear ground rules for capital punishment, engage in doctrinal insubordination. So consider this week’s duet.

The reversal case, Lynch v. Arizona, is both striking and somewhat mysterious. The issue should be clear. As background, here is what the Court said in 1994 in Simmons v. South Carolina: If the prosecutor argues for the death penalty on the ground that the defendant otherwise would pose a threat of violent danger to society, and if the only alternative to a death verdict is life without possibility of parole (LWOP in the vernacular), due process requires the trial judge to advise the jurors –in effect, reassure it—even if they vote against death, the defendant can never leave prison.

A fair reading of Lynch’s case suggests that the trial judge, with later approval by the Arizona Supreme Court, flatly contradicted this clear declaration. And the prosecutor’s chief argument before the Supreme Court in defense of the state court decision was hopeless from the start. The state argued that even if the official default sentence was LWOP, the jury still could reasonably fear the possibility of Lynch’s release from prison because the Governor could exercise clemency or the legislature could retroactively convert LWOP sentences to parole legibility. The Court merely had to restate what it had already made clear—that this argument would nullify the Simmons principle, since obviously clemency or legislative change are always theoretically possible. That should have been, or may have been, all that was necessary for this decision, but for another argument that seems incomprehensible. The trial judge in the case had apparently told the jury that the alternatives to death were LWOP or a 25-year minimum term followed by the possibility of “release.” If that last term meant parole, the instruction contradicted clear state law. If “release” meant clemency, then state was just rephrasing its losing argument. And that’s what the Court apparently inferred.

The Lynch decision was an unsigned per curiam—a format usually reserved for no-brainer cases, which are very often therefore unanimous. Not so here, because Justices Thomas and Alito dissented (but then of course we know that Chief Justice Roberts was in the majority). The dissent made a heroic but opaque effort to reconcile the judge’s weird instruction with state law, so as to preclude any conflict with Simmons. But mostly the dissenters argued that Simmons itself was simply wrong and needs to be overruled. That position is a loser.

While Justice Breyer could happily join the majority in Lynch, he was on a lonely ice floe in Tucker last week. There he deployed the rarely-used but venerable device of a dissent from denial of certiorari, because he had failed to persuade the necessary four justices to even hear the case. But when these dissents are published they afford the writer free space to make a broad declaration of jurisprudence and they often foretell later decisions.

Last year, in a case called Glossip v. Gross, Justice Breyer had diverged from the specific technical issue in the case to write a long solo opinion that was a virtual treatise on arguments why the death penalty is so riddled with arbitrary and irrational outcomes as to no longer merit a pretense of legality. In Tucker, he made a single, simpler point as a kind of supplement to his treatise. The county (parish) where Tucker was tried supplies half the death sentences in the state, even though it has only five percent of the state’s population and five percent of its homicides. Such a geographic lottery, so clearly a function of individual prosecutor preference, itself nullifies the death penalty law.

Justice Ginsburg did join this lonely dissent. Justices Sotomayor and Kagan may well have been sympathetic but were not ready to push for review right now. Might they join Justice Breyer’s position when and if the time and case seem riper?

That question leads us to consider where the death penalty stands now. As compared, say, to 20 years ago, far fewer prosecutors are filing capital charges, and fewer people are getting death-sentenced or executed. Moreover, here and there a state drops the death penalty by statute or judicial decision. And the combination of Lynch and Tucker may reinforce this trend. But what about next year?

If Drumpf is the next President, we may see this judicial pattern and this trend continue, or Justices Thomas and Alito may gain new allies who will help them overrule cases imposing tough ground rules on death sentences—cases like Simmons! If Clinton wins, the Court may proceed cautiously, but new allies might help muster a majority for Justice Breyer’s position. Judicial abolition of the death penalty may seem wildly improbable—almost as improbable as a decision that the Constitution protects a right of same-sex marriage.

Meanwhile, capital punishment in America will lurch and stumble along.

Robert Weisberg is the Edwin E. Huddleson, Jr. Professor of Law and Faculty Co-Director of the Stanford Criminal Justice Center.