What happens when a splinter group breaks off from the fundamental American consensus that we can trust an election? The aftermath of the Bush-Kerry race offers one potential answer. And for anyone hoping that Trump’s followers will quietly fold themselves back into the system, the 2004 experience suggests otherwise. The splinter is still out there.

Over the 16 years that followed the 2004 election, candidates have won and conceded; presidents have been inaugurated. But the loosely defined movement that launched back then has lived on. Most of its members are left-wing, though not all of them identify as Democrats. They’ve come to define their cause not around John Kerry’s rightful presidency, but around the idea of election integrity. Some are fixated on voter suppression; some subscribe to deep-state conspiracies about the manipulation of voting machines. What they share is a conviction that the 2004 election was a sham, and that it exposed a sweeping, anti-democratic cabal. Jonathan Simon, a onetime pollster-turned-lawyer-turned-chiropractor who worked with Freeman on his early analysis, summed up the prevailing view at a congressional hearing after the 2004 vote: “What we’re dealing with here, although the formality is all in place, is a stuffed animal, not a real animal—a taxidermic model of democracy.”

And many of them still believe that. Their continued commitment to the idea even today reveals that, once sown, doubt in the democratic process is difficult to dispel. Rather than recede with age, in many cases these 2004 skeptics’ concerns only deepened. And today, many of these 2004 figures have found a new cause in the 2020 election, embracing Trump’s claims about the results even if they are on the opposite end of the ideological spectrum. The movement is starting to split, as others refuse to align themselves with the president and his supporters, and even think it’s dangerous to do so.

But despite this current divide, both sides are still united in one conviction: They still think the democratic process is deeply flawed—enough to throw an election in 2004, and enough to do it again. Simon, who still questions the 2004 election but doesn’t buy Trump’s claims, says his own reaction to the president has been edifying. “I recognize the irony: Now I’m the one asking for more evidence,” he says. Still, he identifies with something at the core of even Trump’s wildest rants. “The general proposition that you can’t trust the system,” he says, “that is valid.”

The 2004 election was supposed to prove that America had learned its lessons from 2000, when the Bush-Gore race came down to a disastrously convoluted vote count in Florida, followed by weeks of uncertainty and a final Supreme Court ruling. With a political will fueled by images of hanging chads and squinting Florida poll workers, Congress in 2002 passed the bipartisan Help America Vote Act, or HAVA. The act required states to set up voter registration databases, voter identification procedures and provisional ballots for people whose names didn’t show up on the rolls at their home precincts. It also helped local districts purchase the latest in voting technology: optical scan ballot-readers and touch-screen machines that would, in theory, eliminate the human error and potential fraud inherent in paper counting. Even so, the Kerry campaign was prepared for a fight: It had assembled a nationwide network of lawyers to fight post-election battles, if necessary.

Even after Kerry conceded the race, accusations about shenanigans in Ohio emerged on several fronts. The first is what might be called garden-variety voter suppression—perpetrated, accusers say, by Ohio Secretary of State Ken Blackwell, who was also the co-chair of Bush’s Ohio campaign. A 2005 report about the election by U.S. Rep. John Conyers, then a Democratic Congressman from Michigan, outlines voter suppression tactics reminiscent of the Jim Crow era: flyers directing voters to the wrong polling places; voting machine shortages in college towns and heavily-minority neighborhoods; people forced to stand outside on Election Day for hours, in pelting rain.

But people like Freeman, who dug into the exit poll numbers, also circulated a darker theory, centered on fraud that occurred after the vote—via those high-tech voting machines, whose results couldn’t be verified against an independent paper trail. Questions about why the exit polls were so inaccurate—which the mainstream media and even the pollsters attributed to flawed polling methodology—led others to spin out theories about who might have been driven to change the votes themselves. Depending on whom in the movement you talk to, it might have been Karl Rove, working with the private companies that manufactured the machines; it might have been the military-industrial complex; it might have been a shady “they” of extra-governmental elites. It was all impossible to disprove, because it was impossible to prove: If the numbers defied explanation, then any idea, however uncomfortable or wild, could theoretically be true.

For a short time, the breakdown in Ohio gained some purchase in official channels. On December 8, Conyers convened a hearing of an ad hoc committee on voting irregularities in Ohio. It was a toothless event, without a single Republican in attendance, but Democrats made passionate statements in favor of election reform. “The system of voting broke down Nov. 2, 2004,” said Rep. Sheila Jackson Lee of Texas, who called for an independent audit of discarded ballots; a national Election Day holiday; and a rule that barred officials who oversaw elections from working for a partisan campaign. When the hearing turned to questions from the audience, people made suggestions that would sound familiar to anyone who has listened to Trump this year: encouraging electors to change their votes; urging Congress not to certify the election.

Many in the audience seemed open to the deeper, darker theories. When Simon stood up at the citizens’ portion of the hearing and said, “They won’t count our votes as long as they own the machines,” the room broke out in applause. But the politicians would only go so far. “Paper ballots,” U.S. Rep. Jerrold Nadler of New York declared at the hearing, “are extremely susceptible to fraud.” And those who believed that the election had been stolen got no help from the mainstream press, where even left-leaning outlets wouldn’t take up the idea of a vast web of fraud. In The Nation, Alexander Cockburn was caustically dismissive: “As usual, the conspiracy nuts think plans of inconceivable complexity worked at 100 percent efficiency, that Murphy’s law was once again in suspense and that 10,000 co-conspirators are all going to keep their mouths shut.”

Before long, the mainstream apparatus had moved on. In August, 2005, Conyers issued a report about Ohio’s voting discrepancies—you can still buy it on Amazon—but it didn’t lead to any change. In 2005, Mother Jones published an article refuting the stolen-election claims. (Blackwell himself would eventually go on to serve on Trump’s transition team; in 2017, he was appointed to a Trump administration commission on voter fraud.)

Still, the believers carried on their lonely work, taking their fight well beyond Ohio. Among them was Paul Lehto, who, at the time of the 2004 election, was an attorney volunteering at a polling station near his home in Everett, Washington. That night, he noticed a strange discrepancy: A one-vote difference between the number of ballots certified by human poll workers and the number that was recorded by the touch-screen voting machines. No explanation of human error or technical glitches seemed to make sense, so he started to investigate. At first, “it was fun,” he says. “It wasn’t my habit to back down from a challenge. My dad was an IRS auditor for 32 years and a very good one. Maybe there was a gene for being a hunting dog and sniffing them out.”

Long after Washington state and the rest of the country had moved on from the election, Lehto and his brother-in-law, an engineering professor with training in statistics, continued their solo investigation. “Sometimes I only half-jokingly say ‘I was part of John Kerry’s legal army but I never got the order to stand down,’” Lehto told me. They filed Freedom of Information Act requests, asking for the logs from every voting machine in the county. They filed a lawsuit against the county and the voting machine company, which an appellate court eventually tossed out because by then, the election machines had been replaced.

Since then, Lehto says, he has spoken about election fraud to Republican, Democratic, and socialist gatherings, Kiwanis and Lions Clubs, librarians and branches of the NAACP. He says he gets broad agreement on sweeping statements: That politicians can be crooked; that crooks cheat; that crooked incumbents make election law. “Knowing what everybody knows about computers, do you think we can stop the crooks from altering the computers to make sure that they always get reelected?’” he tells his audiences. “Then the lightbulbs come on,” he told me.

The feeling Lehto taps into is easy to harness, says Joseph Uscinski, a political science professor at the University of Miami who specializes in conspiracy theories. Only a few hardy souls will commit the time to mapping out dark plots, he says: “It’s not like Grandma is sitting at home running regression models.” But plenty of Americans—at every point on the political spectrum—are open to theories that cast unseen power brokers in dark shadows. “A lot of people just have antagonism to the system writ large,” he says. They divide the country between “us, the good people, and then the corrupt elite.”

As Uscinski points out, Trump and his supporters weren’t the only ones who thought fraud was likely in 2020. When Uscinski did polling in October to see how much Americans trusted the upcoming election, he found that 70 percent of Republicans believed there was a conspiracy to rig the election with mail-in ballots. But nearly the same number of Democrats believed the election would be rigged because the post office wouldn’t deliver the ballots that had been cast. In August, Hillary Clinton said Biden shouldn’t concede if the election was close.

For years, mistrust in the process fueled the core of people who doubted the results of the 2004 election. Simon self-published a book, Code Red, which he continued to update with information from subsequent elections. Freeman shared his analysis with Robert F. Kennedy Jr., who wrote a high-profile Rolling Stone article in 2006, titled, “Was the 2004 Election Stolen?” Freeman co-wrote his own book with the same title, believing his pursuit would make a difference.

“I wrote the entire first draft of this book in four months, working from midnight to late in the morning, because I thought, ‘You know, I’m saving democracy here. And I’m also going to write a bestseller,’” Freeman told me. “Obviously, I was delusional, but nevertheless it really powers you.”

He was, indeed, wrong about whether the book would take off: The book got scant attention, as did another book published around the same time by Mark Crispin Miller, a New York University professor, called Fooled Again: How the Right Stole the 2004 Election, & Why They’ll Steal the Next One, Too (Unless We Stop Them). The fact that election-fraud arguments were roundly ignored by politicians and the press became a kind of evidence in itself. “My book came from a major publisher and was very meticulously sourced and was nonpartisan, inasmuch as I was not then and am not now a Democrat,” Miller told me. “And yet when the book came out, it was like hitting a brick wall. It got a total of two newspaper reviews in the whole country. One was a hit piece.” His election fraud work, he says, marked a shift in his reputation from a mainstream intellectual to a fringe figure. (Miller is known for challenging other consensus narratives, suggesting that the government knew in advance about the 9/11 attacks and questioning whether masks can stop the spread of Covid.)

As the idea of election fraud became more and more marginalized, some of the movement’s urgency was lost. “Obama’s victory took the wind out of the sails of this thing,” says Freeman, who also concedes that he got weary of the fight. “Realistically from my perspective, what do I get out of this?” he told me. “If I’m unsuccessful, I waste a lot of time and effort … if I’m marginally successful, I get crucified. If I’m really successful, I get killed.” When I asked him if he was serious, he said he was—he got spooked when, in the thick of his activism, his tire nearly fell off as he drove home from work in a rainstorm, the result of four lug nuts that had mysteriously gone loose.

Still, in 2016, when Donald Trump the presidential candidate started making noise about a rigged election, Freeman tried, through various channels, to reach out to him and offer help. He never managed to make contact.

Now, as Trump makes his own broad-ranging claims of election fraud, the stalwarts of the movement born in the wake of Kerry’s loss are processing different emotions. Freeman and Simon, for instance, old comrades-in-arms, now find themselves in a kind of civil war, split over whether Trump’s rantings are a golden opportunity or a toxic diversion.

On December 2, after Trump gave a 45-minute speech about election fraud that most news networks refused to cover, Freeman called up the “Election Integrity” Google group he had started years earlier and posted: “From an EI perspective, this is the most important speech anyone has ever made.” Simon responded with horror. “I personally regard Donald Trump as more of a threat against our nation’s wellbeing than I do rigged elections,” he later told me. “Rigged elections is a chronic problem—it’s more akin to rheumatoid arthritis—and Donald Trump is more like brain cancer.”

Miller, on the other hand, thinks blind hatred of Trump is distracting other diehards from their own principles. He’s irked “to see them now either laughing off these signs of theft, or even saying ‘We have to get some sane people in the White House,’ as if it’s their call to determine who won. It’s not up to them. It wasn’t up to us in 2004.”

Now, he and others in the movement are watching with a sense of personal concern as the mainstream apparatus rejects Trump’s ideas outright. When YouTube announced it was removing videos that claimed widespread 2020 election fraud, Miller and Freeman wondered if old footage of themselves, talking about the 2004 race, would be swept up in the purge.

Meanwhile, some advocates want to focus on election reforms that distinguish their aims from Trump’s—and don’t rely on shadowy theories of fraud. Greg Palast, a former journalist for the Guardian and the BBC who has made election reform his signature cause, wants to concentrate on fighting “duller, less Hollywood, less exciting means” of suppressing the vote: provisional ballots that are rejected; signatures that are challenged; voters lists that are unfairly purged. This fall, he conducted his own investigation into vote suppression in Georgia, then joined the ACLU in a lawsuit to restore 198,000 voters to the state’s rolls.

Even many of Miller’s ideas for reform are, if not politically viable, at least straightforward: tightening the chain of custody for ballots (which would essentially eliminate mail-in ballots); making Election Day a federal holiday; making voter registration automatic on a person’s 18^th birthday. “If there were any kind of political will to improve American elections, it could be done simply,” he says.

But Miller and others also agree with Simon’s warnings, from long ago, about the hackability of voting machines. Lehto thinks elections will never be secure until electronic machines disappear, the black box of uncertainty is gone and people can look at hard paper records of every single vote. “Let’s just reverse the burden of proof. Say, ‘OK, prove Biden won,’” he told me. “There is no proof Biden won. It’s just numbers that popped out of computers.”

I asked Lehto what, then, he thinks should happen next. In a country where election results are unknowable, who should get inaugurated on January 20? He demurred. “At that level I would say: ‘My fellow Americans, none of us really know what the election results are,’” he said. “We all have our opinions, but those are really based on faith. It’s like a religious belief.”

That’s an idea Freeman raised to me, too. He sent me the results of an online poll that asked Americans who they legitimately thought won the 2020 election; he was struck, not by the numbers for Biden or Trump, but by the fact that only two percent said they weren’t sure. For Freeman, uncertainty about election results is a constant, a given. And yet, he concedes, life goes on. Inauguration Day will come. The president will still be the president—something he has begrudgingly accepted for the last 16 years.

“I mean I don’t have a choice, right? I have to pay my taxes,” Freeman says. “I accept them purely as most people accept them, I think, which is: They have the power.”