The strongest argument for throwing Trump off the ballot is the weakness of the counterarguments.

When I review divided appellate-court decisions, I almost always read the dissenting opinions first. The habit formed back when I was a young law student and lawyer—and Federalist Society member—in the late 1980s, when I would pore (and, I confess, usually coo) over Justice Antonin Scalia’s latest dissents.

I came to adopt the practice not just for newsworthy rulings that I disagreed with, but for decisions I agreed with, including even obscure cases in the areas of business law I practiced. Dissents are generally shorter, and almost always more fun to read, than majority opinions; judges usually feel freer to express themselves when writing separately. But dissents are also intellectually useful: If there’s a weakness in the majority’s argument, an able judge will expose it, sometimes brutally, and she may make you change your mind, or at least be less dismissive of her position, even when you disagree. Give me a pile of Justice Elena Kagan’s dissents to read anytime—I love them even when she’s wrong, as I think she often is. You can learn a lot from dissents.

Last night, I reviewed the three separate dissents in Anderson v. Griswold, the landmark 4–3 Colorado Supreme Court case holding that Section 3 of the Fourteenth Amendment prohibits Donald Trump from ever serving again as president of the United States. I had been skeptical of the argument, but not for any concrete legal reason. To the contrary, I believed the masterful article written by the law professors (and Federalist Society members) William Baude and Michael Stokes Paulsen had put the argument into play. And I had read (not to mention heard, at length, on the phone) and took quite seriously what my friends Judge J. Michael Luttig and Professor Laurence H. Tribe had to say about it here in The Atlantic—that the Fourteenth Amendment clearly commands, in plain language, that Trump never hold federal office again.

Their points were strong. But much as I never want to see Trump near the White House again, I wasn’t quite buying them. The argument seemed somehow too good to be true. And frankly, from a political standpoint, it would be better for the country if Trump were thrashed at the polls, as I think he ultimately would be. There had to be a wrinkle. I just knew it.

But last night changed my mind. Not because of anything the Colorado Supreme Court majority said. The three dissents were what convinced me the majority was right.

The dissents were gobsmacking—for their weakness. They did not want for legal craftsmanship, but they did lack any semblance of a convincing argument.

For starters, none of the dissents challenged the district court’s factual finding that Trump had engaged in an insurrection. None of the dissents seriously questioned that, under Section 3 of the Fourteenth Amendment, Trump is barred from office if he did so. Nor could they. The constitutional language is plain. You can’t be president if you previously took an oath “as an officer of the United States … to support the Constitution of the United States” but “engaged in insurrection or rebellion” against, or have “given aid or comfort to the enemies of,” that Constitution or the nation it charters.

Nor did the dissents challenge the evidence—adduced during a five-day bench trial, and which, three years ago, we saw for ourselves in real time—that Trump had engaged in an insurrection by any reasonable understanding of the term. And the dissenters didn’t even bother with the district court’s bizarre position that even though Trump is an insurrectionist, Section 3 doesn’t apply to him because the person holding what the Constitution itself calls the “Office of the President” is, somehow, not an “officer of the United States.”

Instead, the three dissenters mostly confined themselves to saying that state law doesn’t provide the plaintiffs with a remedy. But that won’t help Trump. This case seems headed for the Supreme Court of the United States, which has no authority to make definitive pronouncements about state law. In Colorado, the Supreme Court of Colorado has the last word on that. And it now has spoken.

Yet even the dissenters’ contentions about state law made little sense. Chief Justice Brian Boatright argued that, while Colorado law requires its secretary of state to examine the constitutional qualifications of presidential candidates, it doesn’t allow her to consider whether they are constitutionally disqualified.

Nothing in the state statute suggests that’s the case, and it’s plainly illogical. Every qualification necessarily establishes a disqualification. If the Constitution says, as it does, that you have to be 35 years of age to serve as president, you’re out of luck—disqualified—if you’re 34 and a half. By the same token, if you’ve engaged in an insurrection against that Constitution in violation of your oath to it, you’ve failed to meet the ironclad (and rather undemanding) requirement that you not have done that.

Boatright’s suggestion that the insurrection issue presents something too complex for Colorado’s election-dispute-resolution procedures is equally unconvincing. Reviewing the tabulation of statewide votes can be complicated—remember those Florida “chads” in 2000?—but the courts have to get it done, and quickly. It’s hard to imagine that assessing the undisputed record of Trump’s miscreance presents any more complexity than that.

And no stronger is Justice Carlos Samour’s suggestion that Trump was somehow deprived of due process by the proceedings in the district court. This was a full-blown, five-day trial, with sworn witnesses and lots of documentary exhibits, all admitted under the traditional rules of evidence before a judicial officer, who then made extensive written findings of fact under a stringent standard of proof. Every day in this country, people go to prison—for years—with a lot less process than Trump got here. As for the expeditiousness of the proceedings, that’s in the very nature of election disputes: Recall, once again, Florida in 2000. And Samour’s suggestion that Trump was denied a fair trial because he didn’t have a jury is almost embarrassing: Any first-year law student who has taken civil procedure could tell you that election cases are not even close to the sort of litigation to which a Seventh Amendment jury-trial right would attach.

The closest the dissents come to presenting a federal-law issue that ought to give someone pause comes in Samour’s argument that Section 3 is not self-executing—that it can’t be enforced unless Congress passes a law detailing how it can be enforced. The majority opinion, though, along with Paulsen and Baude and Luttig and Tribe, have disposed of that argument many times over. All you need to do is to look, as any good Scalia-like textualist would, to the words and structure of the Fourteenth Amendment.

True, Section 5 of the amendment gives Congress the power to enact enforcement legislation. But nothing in the amendment suggests that such legislation is required—that Section 3 (or any other prohibition in the amendment) has no teeth unless Congress implants them. To hold otherwise would mean that Section 1 of the Fourteenth Amendment—which contains the more familiar prohibitions against state deprivations of equal protection and due process—would likewise have been born toothless. Which would mean that, if every federal civil-rights statute were repealed tomorrow, states could immediately start racially resegregating their schools. That’s not the law, and thankfully so.

So the dissents showed one thing clearly: The Colorado majority was right. I dare not predict what will happen next. But if Trump’s lawyers or any members of the United States Supreme Court want to overturn the decision, they’d better come up with something much, much stronger. And fast.