In these last days before the Iowa caucuses, Bernie Sanders has enjoyed a surge in the polls that has thrown moderate Democrats into a panic. A poll released on Wednesday by Crooked Media/Change Research puts Sanders at the top of the field with 27 percent, trailed by Pete Buttigieg (19 percent), Joe Biden (18 per cent), Elizabeth Warren (15 percent), and Amy Klobuchar (10 percent). In national polls, the gap between the Biden and Sanders continues to narrow. In the aggregation of polls by Real Clear Politics, Biden stands at 28.4 percent to Sanders’s 23 percent. It’s easy now to see a path for victory for Sanders: He could win the first three states (Iowa, New Hampshire, Nevada), which would give him enough momentum to overtake Joe Biden (whose supporters are much less committed than Sanders’s loyal base).

The prospect of Sanders being the nominee has unnerved moderate Democrats as well as their allies among Never Trump Republicans. This panic has led a super PAC (Democratic Majority for Israel) to start pushing anti-Sanders ads. It’s also produced a spate of nervous columns warning of the risks of a Sanders candidacy in New York magazine, The New York Times, The Atlantic, and Slate.

Jonathan Chait’s New York piece is the most valuable example of this genre because it contains a revealing contradiction. Chait rehearsed the familiar litany of concerns about Sanders: the unpopularity of the socialist label that Sanders has embraced, Sanders’s history of associations with radical groups like the Socialist Workers Party, the indiscreet sexual writing Sanders did as a young man, and the success that Democrats have had in fielding moderate candidates in the 2018 midterms compared to the failure of more progressive candidates in the same election cycle.

All of these are serious concerns, although many of them can convincingly be shown to be less worrying than Chait thinks. For example, Chait argues that in 2018, “while progressives managed to nominate several candidates in red districts—Kara Eastman in Nebraska, Richard Ojeda in West Virginia, and many others—any one of whose victory they would have cited as proof that left-wing candidates can win Trump districts, not a single one of them prevailed in November.” But as Princeton historian Matt Karp notes, even though Ojeda lost, he did 25 percent better than Hillary Clinton did in 2016.

Karp observes, “When pundits try to bring up Richard Ojeda as evidence of failure, remember that this guy had the single strongest performance of any non-incumbent Democrat in the entire 2018 election. If Bernie Sanders runs 25 points ahead of his partisan lean in a national election, he’ll win a landslide so big that even Jon Chait’s most garishly stupid takes will be buried for a thousand years.”

But the best refutation of Chait comes from Chait himself. At the beginning of his article, Chait notes, “Polarization has given any major party nominee a high enough floor of support that the term ‘unelectable’ has no real place in the discussion.” This sentence is a self-destruct button: Once he pushes it, Chait’s whole argument is blown to smithereens.

Donald Trump’s 2016 victory illustrates how polarization undermines previously self-evident ideas about electability. After all, who could be less electable than Trump was when he ran against Hillary Clinton? He had no political experience and a checkered past that included multiple bankruptcies, ongoing feuds with party elders like John McCain and Mitt Romney, a history of making misogynistic comments, and multiple accusations of sexual misconduct (highlighted in the weeks before the election by the release of an audio where he boasted about grabbing women by their genitalia).

If anyone was ever unelectable, it was Donald Trump. That’s what many analysts, myself included, thought. Chait was a particularly strident advocate of this idea. In an infamous 2016 column, which he later retracted, Chait argued that liberals should want Trump to be the Republican nominee because he would be easy to trounce.

We all make mistakes. The test is whether we learn from them. The mistake of underestimating Trump in 2016 was based on discounting the power of polarization—a lesson directly pertinent to the question of Sanders’s electability.

Trump won because many Republicans who despised him put party loyalty above all. They thought Hillary Clinton was the devil incarnate, while Trump was a lesser sinner who would at least give them conservative judges. The Never Trump faction never consisted of much more than a small, if vocal, scattering. Even some Never Trump conservatives, like New York Times columnist Bret Stephens, who didn’t vote for Trump in 2016 are now thinking about returning to the fold because they are horrified by the possibility of Bernie Sanders or Elizabeth Warren winning.

The lesson of 2016, then, is that a progressive candidate like Sanders or Warren can win so long as Democrats remain animated by partisan fervor. To win, they would only need a unified Democratic base plus a small number of those who in 2016 voted third party or didn’t vote at all.

In previous elections, progressives were told to hold their noses and vote for unexciting candidates like Bill Clinton, Al Gore, or Hillary Clinton in order to defeat the Republicans. If Sanders or Warren is the nominee, then moderate Democrats will be the ones who have to hold their noses. The alternative is reelecting the worst Republican president in modern times.

The problem with writers like Chait is that they discuss electability as if it were a fixed and objective fact of nature. But electability depends on the choice of the electorate, very much including voters like Chait himself. If moderate Democrats of the Chait variety put aside their ideological disagreements with progressives and pledge to support the nominee no matter who it is, then the electability problem becomes slight.

So far, it doesn’t seem like moderates are prepared to make that sacrifice. Indeed, there are signs some are preparing to fight to the end. Democratic National Committee Chair Tom Perez has used his position to fill up committees at the next convention with centrists, many of whom are corporate lobbyists hostile to economic populism. If there is a floor fight at the convention, Perez has staffed the committee with those who want to thwart progressives.

The electability issue will be settled by choices made by this faction of the party: If they truly oppose Trump and all he stands for, they won’t use their positions to undermine a progressive nominee. But if they fight tooth and nail against the winner, then the possibility of Trump’s reelection is much higher. For the good of the party, moderate Democrats need to prepare to become junior partners in the coalition. They are not used to making this sacrifice. But the electability problem will be solved if they swallow their pride.