There are 27 days left in the 2016 presidential election. The shrinking calendar is likely to add a little-needed infusion of mania into an already maniacal race, especially as it concerns the wildest of the four remaining candidates.

In the last 72 hours, Donald Trump has proven that he is willing to chart new territories of the paranoid and unpredictable: holding a pre-debate panel with women who have accused former president Bill Clinton of sexual indiscretion and assault, humiliating his running mate by breaking with him on foreign policy, and publicly sparring with Speaker Paul Ryan, the man who will be his most powerful ally in Congress—should Trump actually win the White House.

It is unclear which of these developments—the spectacle of Republicans trying to explain away or rationalize sexual assault, or the fracturing of the Republican Party itself—will dominate the remaining weeks of this campaign. But what may inflict the longest lasting damage on the American political system may be something else entirely. One sees suggestions of it in the recent and virulent strain of paranoia that has crept into Donald Trump’s campaign speech as of late: “You gotta watch your polling places!” he intoned last week. “I’m afraid the election’s going to be rigged.”

The mostly conservative narrative about election fraud is pervasive, though unfounded: voter fraud is at nearly zero percent in the United States. But it’s not exactly surprising that Trump’s increasing and alleged concern for electoral fairness comes at the precise moment that polling shows his support to be decreasing in several key states and his campaign finds itself in a tailspin. In fact, the two may be closely (and inversely) related: If the race doesn’t look like it’s going in the right direction, contest the results—a strategy adopted by spoilsports and despots alike.

Trump has thus far offered no evidence for his new concern beyond “you’ve been reading the same stories I’ve been reading,” (perhaps these stories include a debunked research paper on the Democratic primary) but he has nonetheless urged his supporters forward. On Trump’s website, those heeding the call can now sign up to be a “Trump Election Observer” with the explicit goal of helping the candidate “Stop Crooked Hillary From Rigging This Election!” (Never mind that the party trying hardest to tamper with American politics is Russia, not Rodham—and that thus far the Republican has been the beneficiary of said manipulations).

What, practically, does Trump mean? It’s not as if his supporters can follow unsuspecting voters into the booth, or conduct cyber-monitoring or check the hanging chads—but U.S. election law is remarkably permissive when it comes to so-called “election challengers.”

According to a study by the Brennan Center, 39 states currently permit private citizens—including, say, Trump supporters or Clinton supporters—to challenge prospective voters on Election Day. What happens to the voter varies, explained Wendy Weiser, director of the Brennan Center’s Democracy Program.

In some states, those challenging must provide election officials with some sort of valid reason for contesting a voter’s eligibility.

But in 24 states, private citizens can challenge a voter at the polls without offering any documentation that might call this eligibility into question.

All this means that a challenger intent on, for example, “stopping crooked Hillary from rigging the election” could feasibly single out any number of voters without any evidence to offer, and try to keep those voters from casting a ballot by calling into question their eligibility.

These challenges, furthermore, can “be based on race, ethnicity and language,” according to Weiser—in part owing to their inception during the Reconstruction Era. “A lot of these laws were passed as a package to keep newly enfranchised African Americans from going to the polls,” said Weiser.

Ultimately, much of the decision-making rests in the hands of election officials, many of whom are volunteers or retirees.

This presents a further set of problems.

“It’s intimidating to poll workers,” Weiser explained. “Many of them are older. Some are doing this for the first time.”

And the net result could be a “really volatile situation. It makes polling places the sites of confrontation and worse”—like harassment, “especially in a heated election. There are places where people can cross the line into impermissible conduct.”

It remains unclear how successful Trump’s mobilization efforts might be, but Weiser is concerned that election officials may not be prepared for this year’s influx of challengers. While other years have featured similarly-minded conservative efforts, including True the Vote, “This is the first in a long time where a presidential candidate with such a big platform has done so.”

Indeed, the very suggestion—preemptively—of fraudulent election behavior is largely unprecedented in the modern era (not, of course, the assertion of fraudulent election behavior after the fact). Such fearmongering is “quite common in less mature democracies,” said David Carroll, director of Democracy Programs at the Carter Center—which oversees a significant amount of international election monitoring:

“It’s seen in many of the places that we work in: places undergoing transition. In the context of weak and emerging democracies, it is not unusual. Quite often, it’s a party that fears it’s going to lose that is saying that.”

An official from the United Nations, which doesn’t monitor elections but does offer technical assistance, pointed to a practice that has been successful in “hotly contested” races, as this year in Lesotho and Nigeria, or Malawi in 2012. Ahead of the election itself, leaders of political parties sign a code of conduct, committing to accepting the results of the election. Should they have any complaints with its results, they vow not resort to violence, but to instead pursue legal means. The signing ceremonies of these pledges are sometimes televised, with presidents of neighboring countries in attendance. “Peer pressure,” the official explained to me.

The problem here is that Trump has shown no interest in being convinced that the election might be fair. His reluctance to sign a pledge to support his own party’s nominee—something he ultimately reneged on, anyway—makes it hard to imagine him signing anything that might bind him to anything, no matter how weak the language in the agreement.

And—as with birtherism—Trump’s party is already so animated by this conspiracy that, to a certain extent, it can live on without his active endorsement. According to Pew, only 38 percent of Trump supporters believe the vote will be counted fairly. That may not be surprising, but consider the widespread nature of this paranoia.

According to a 2012 Cooperative Congressional Election Study (emphasis mine):

31 percent of Romney voters… endorsed the idea that Obama was not the legitimate winner of the race.

Among those with “very positive” views of the tea party, 51 percent held that perspective.

But 22 percent of Romney voters with a less positive view of the tea party also questioned the legitimacy of the election, suggesting that this belief was not restricted to a fringe of the party.

Herein lies the most potentially devastating effect of Trump’s florid and public propulsion of a “rigged system.” His corps of challengers may very well sow chaos on election day, but the laws allowing them to do so are at least laws—able to be repealed or updated (or kept in place), but at the very least subject to formal legislative governance.

Public trust, however—the core of our social compact—is not a matter of legislation: It is a terrible thing to lose, and a difficult thing to regain. And Trump has lately determined that his survival may be contingent its erosion; the last three days, after all, have shown the country how much he’s prepared to risk. It’s a dangerous play for the candidate, but unquestionably, a far more hazardous one for democracy.