Almost exactly six years ago, James Comey begot a new mantra for the Trump era: “Lordy, I hope there are tapes.” In most cases, none has emerged: not of the former FBI director’s conversation with Donald Trump about loyalty, not of the fateful call with Ukrainian President Volodymyr Zelensky, and not, well, that other fabled tape.

In the ongoing classified-documents scandal, though, the tapes seem to exist. CNN and The New York Times report that Justice Department Special Counsel Jack Smith, who is investigating Trump’s removal of secret records to Mar-a-Lago, has obtained a recording in which the former president discussed his possession of a sensitive document. According to the outlets, Trump indicates that he knows it’s classified and is aware he cannot share it.

The content of the tape is important for any prosecution of Trump, which would have to prove he knew that what he was doing was wrong. But the circumstances of the recording are also revealing about how Trump operates, and the way he seems to understand bad press as a graver threat than criminal prosecution.

[Read: Donald Trump and the currency of tapes]

No dispute exists over whether Trump took boxes of documents with him from the White House. The question is whether they were classified and public records, or declassified and his personal property. Trump has asserted publicly—though his attorneys have conspicuously avoided doing so in legal filings—that he declassified all of the materials before leaving office, without providing any evidence for the claim. Audio proof that Trump understood that at least one document was still secret would demolish that defense.

Given that mishandling of classified materials by former officials is apparently common, Smith appears to also be focusing on whether Trump attempted to hide the documents from the federal government once they were requested and then subpoenaed. Reports indicate that Trump had boxes moved to hide them and lied to his attorneys about the material, and an aide allegedly inquired about how long surveillance video was maintained. (Lordy, maybe there are lots of tapes.)

Aside from the egregious violation of the Stringer Bell rule—or perhaps just the Richard Nixon rule—that recording evidence of one’s own criminality represents, the tape would demonstrate yet again Trump’s reckless disregard for the law. Consider the circumstances for the recording. In July 2021, two writers working with former White House Chief of Staff Mark Meadows on his autobiography interviewed Trump at his Bedminster, New Jersey, club. Meadows was not present. (Suffice it to say that this is not how Bob Haldeman or Ulysses S. Grant wrote their memoirs.)

Trump was, as usual, in a score-settling mood. A recent New Yorker report had claimed that in the final days of his administration, Chairman of the Joint Chiefs of Staff Mark Milley had taken steps to prevent Trump from ordering a strike on Iran. The story was opaque on its sourcing, but it narrated events from Milley’s own perspective. Trump, who likes to portray himself as a dovish, isolationist opponent of warmongering generals, was furious. At the meeting with the two writers, Trump brandished a report that he claimed was Milley’s plan for an assault on Iran, and said that the general had repeatedly urged him to mount an attack. He can apparently be heard waving the papers on the recording. Neither CNN nor The New York Times heard the audio, but it was described to reporters at both outlets by multiple sources.

But Trump was reluctant to show the memoir writers the actual document, according to the reports, because he knew it was still classified and they did not have security clearances. He may not have always been so fastidious. Smith is reportedly also investigating whether Trump showed several visitors a classified map.

The recording that Smith has obtained was reportedly made not by the writers but by Margo Martin, a Trump aide who “​​routinely taped the interviews he gave for books being written about him that year,” according to the Times. The former president was apparently worried about being misrepresented or misquoted.

To summarize: Trump’s fear of damaging press—whether in the Milley reports or the Meadows book—was so much greater than his fear of criminal accountability that he ended up making an incriminating recording that could be a key piece of his own prosecution.

[Adam Serwer: Trump’s inner circle keeps violating the Stringer Bell rule]

Throughout his career, Trump has behaved as a person who sees image as more important than law. It’s an outlook that seems to stem not only from his inherent disdain for rule of law and love of publicity, but also from a calculation that when the two conflict, image will triumph. Over and over, he’s managed to wriggle out of potential legal jams with bluster, brazenness, and the occasional large check. That worked when he was president, too, escaping serious consequences from Special Counsel Robert Mueller’s investigation and his first impeachment by rallying political support. It was not enough to prevent his loss in the 2020 presidential election, but it helped him avoid conviction in his second impeachment.

Trump is still at it. No matter how damning the evidence that Smith is able to assemble, Trump is seeking to bully the Justice Department out of charging him. If that doesn’t work, he hopes to be reelected to the presidency in November 2024, which would allow him to shut down any investigation or prosecution against him, or to pardon himself. It might yet work.

Every successful politician follows roughly the same path: First, they become prominent on some stage. They become more successful, maybe graduating to a larger stage. Then, eventually, they peak and decline, with the affection of even their strongest supporters cooling somewhat.

If they are lucky (Harry Truman, George H. W. Bush), they eventually experience some historical revision that burnishes their reputation. (If they are very lucky, they even live to see it.) If they are not (Herbert Hoover, Richard Nixon), they don’t. This happens whether a politician’s departure from office comes in defeat at the polls or at the top of their popularity, as with Bill Clinton, who has seen his reputation suffer—personally and politically—in the past 15 years.

Along with election results and norms of basic decency, Donald Trump continues to defy this pattern. Not only was the former president nationally famous before he entered politics, but he has always been unpopular with most Americans and very popular with his base. From early in his presidency through to the present, nothing has changed the fundamental picture. That stability is now the key to understanding the 2024 Republican nomination race.

[David A. Graham: The Republican primary’s Trump paradox]

The prospect of a rematch between Trump and Joe Biden has demoralized and baffled commentators. “Not Biden vs. Trump Again!” moaned a recent headline on the political-science site Sabato’s Crystal Ball. “It won’t be pretty. It may not be inspiring. And it will mostly be about which candidate you dislike more,” warned Doyle McManus of the Los Angeles Times. “How did a once-great nation end up facing an election between two very old, very unpopular White dudes?” groaned The Washington Post’s Megan McArdle.

The answer in Biden’s case is relatively straightforward: Incumbent presidents basically never lose the nomination (though shockingly high polling for known crank Robert F. Kennedy Jr. illustrates the dissatisfaction among Democratic voters). Trump is a more interesting case, because he is not president, has never successfully won the popular vote, and lost the previous election—to say nothing of his attempt to steal the election afterward.

These are the ingredients for a politician to lose his support and slink from the scene. No popular groundswell demanded that Gerald Ford run in 1980, nor Bush in 1996; only inveterate op-ed-page contrarians such as Doug Schoen clamored for Hillary Clinton to run again in 2020 (or 2024, for that matter).

Yet Trump hasn’t lost luster, partly because he never had much luster to begin with. Since March 2017, with a brief exception, more than half of Americans have disapproved of Trump (during his presidency) or held an unfavorable opinion of him (since he left office), according to FiveThirtyEight’s poll averages. (He very briefly dipped into mere plurality disapproval early in the coronavirus pandemic.)

One half of the equation is that it’s hard to become unpopular when you were already there. The other half is that it’s hard to become more unpopular when your supporters are so devoted. In a recent YouGov/Economist poll, 84 percent of Republicans had a favorable view of Trump; Quinnipiac pegged the number at 86 percent.

This kind of split might have been impossible in past decades, because it would have spelled electoral doom: To win the nomination in politically heterogeneous parties, a candidate had to appeal broadly. But in today’s ideologically sorted and affectively polarized parties, a candidate can win the nomination and then rely on their party’s voters to coalesce around them and guarantee 47 to 49 percent of the vote. (Of course, it’s that last little increment to a majority or plurality that makes all the difference in the end.)

Ron DeSantis only formally entered the race in May, but he appears to be sputtering. At the same time, the primary is expanding, as more Republicans enter the race or seriously consider it. One explanation for this is that DeSantis just hasn’t been a very good candidate: He looks clumsy and leaden on the trail, and he’s failed to differentiate himself from Trump in a way that appeals to enough voters. That’s encouraged other Republicans to make a plan for the mantle of Trump alternative.

But the problem facing either DeSantis or any of the others is not that the right Trump alternative hasn’t emerged but that most Republicans don’t want a Trump alternative. They want Trump. The depth of affection for Trump is appalling, given that his first term in office was morally and practically disastrous and ended with an attempt to steal the election and an exhortation to sack the U.S. Capitol. But Republicans continue to love him; it’s not debatable.

DeSantis, cautiously, and former New Jersey Governor Chris Christie, more Christiely, have tried to get around this by arguing that Trump is a loser: He lost in 2020, he led the party to losses in 2018 and 2022, and he barely avoided losing in 2016. This is a tricky balance to strike, because it requires convincing Republican voters that the guy they voted for twice, and whom they still like, is a loser—especially compared with Christie, who lost badly to Trump in 2016, and DeSantis, who is losing badly to Trump this time. The easy retort is the same one for Bernie-would-have-won types after 2016: If he would have won, then why didn’t he? In this case, why aren’t you winning now?

More important, this argument will fail to convince Trump supporters because they believe he’s actually the most electable candidate. A Monmouth poll released Tuesday finds that almost two-thirds of Republicans think the former president is definitely or probably the candidate best positioned to defeat Biden. Trump critics will scoff at this, but then again, Trump’s victory in 2016 is proof that unpopularity isn’t politically fatal.