This past presidential election was among the most unsavory in U.S. history, and it might not even be over yet. Unless we are very careful, this election could yet come close to crippling constitutional government.

The problem is straightforward enough: namely, that large sections of the losing side stubbornly refuse to admit defeat. That is bad in itself. Around the world, one of the commonly accepted criteria in judging a democratic society is whether the losing party agrees to stand aside after losing an election. Initially, Hillary Clinton conceded defeat, and did so with grace and maturity, partly (it seems) under White House pressure. Still, an alarming number of her followers remained diehards, and the Clintons have now joined the demand to recount votes in Wisconsin and other swing states.

That recount demand is rooted in some murky statistics, and some assertions that can now be shown to be bogus. Initially, New York magazine quoted an expert study purporting to show that votes in critical states had been manipulated by computer hacking. According to those early claims, a technical comparison of regions within particular states showed that places using electronic voting turned out surprisingly high votes for Trump relative to areas that relied on paper ballots. Hence, said the advocates, there was prima facie evidence of malfeasance, and votes in swing states should be audited carefully before any final decisions can be made about the election’s outcome—however long that process might take.

Actually, these suggestions were absurd, and were promptly recognized as such by many liberal media outlets. The allegations were, for instance, utterly rejected by quantitative guru Nate Silver, who is anything but a Trump supporter. As he and others noted, the non-urban areas that were vastly more inclined to vote for Trump were also the ones most likely to use electronic voting. Big cities, in contrast, commonly used paper ballots. Obviously, then, we would expect votes cast by computer to lean heavily toward Trump in comparison to those marked on paper. That result certainly did not mean that Russian techies in secret fortresses in the Urals were hacking computers in Wisconsin and Michigan to delete votes cast for Hillary Clinton. Not long after the New York story, the main computer expert cited made clear that he himself did not accept the hacking explanation, although he still felt that an electoral autopsy was called for. That process is now underway, and with the support of the Clinton camp.

The fact that such mischievous allegations have even been made bespeaks liberal desperation at the defeat of their candidate, and a bottom-feeding attempt to seek any explanation for the catastrophe those liberals feel they suffered on November 8. Sadly, though, these unfounded allegations will remain alive in popular folklore for decades to come, with the simple takeaway: Republicans stole the 2016 election.

Those electronic issues pale in comparison with Democratic Party resistance in the Electoral College, where delegates are scheduled to meet on December 19. Normally, those electors would simply be expected to confirm the results of the November ballot, but liberals have demanded that they do the opposite, and actively nullify the result. Some electors have already stated that they will refuse to accept the majority votes cast in their Trump-leaning states. Notionally, these “Hamilton electors” will take this course not from any partisan motivation but rather to draw attention to the perceived injustices of the Electoral College system, which in their view should be replaced by a national popular vote. Online petitions urging other electors to join the defection have garnered millions of signatures.

Donald Trump’s lead in the college was so substantial—probably 306 to 232—that a handful of “faithless electors” should not affect the overall result, which could be overturned only through the concerted efforts of dozens of pro-Hillary activists. That is extremely unlikely to happen, but all the credentialed experts dismissed as unthinkable so many other things that actually have happened in this turbulent year.

For the sake of argument, imagine that enough electors go rogue to flip the election. Think through the likely consequences of such an outcome—in which Hillary Clinton is inaugurated in January, rather than Donald Trump. It is inconceivable that a Republican Congress would accept this result. It would offer zero cooperation in any legislative efforts, and it would presumably stonewall any and all approval of Clinton-nominated officials or judges. The only way to operate the government in those circumstances would be for the president to make extensive use of executive orders, and to fill official posts through an unprecedented volume of recess appointments. Theoretically, that method might even be used to fill Supreme Court vacancies. Constitutional government would have broken down, and we would be facing something like a Latin American presidential dictatorship. For several years, Washington’s political debate would be reduced to something like a Hobbesian war of all against all.

Does anyone really want to see a Clinton presidency at such a cost?

Nor is it easy to see how such a cycle could ever be broken once set in place, and particularly how the precedent set in the Electoral College could ever be overcome. Would not Republican electors seek revenge in 2020 or 2024? In that event, the November elections would become merely an opening gambit in an interminable legal process.

It is also ironic to see Hillary’s supporters demanding action in the Electoral College on the grounds of her convincing win in the popular vote. As they argue, how could any administration seriously claim a “mandate” with just the 46 percent or so of that vote earned by Donald Trump? Older election aficionados might cast their minds back to 1992, when an incoming Clinton administration decided to go full steam ahead on a number of quite radical policies, including a bold attempt to establish a national health-care system. The president then was Bill Clinton, who owed his presidency to gaining just 43 percent of the popular vote. Mandates are strange and flexible beasts.

Through the years, we have witnessed a number of elections so catastrophic that they seemingly threaten the existence of one or the other party. In the mid-1970s, few serious observers believed the Republican Party would survive the Watergate crisis, and similar pessimism reigned on the right following Obama’s victory in 2008. Yet despite such disasters, political currents soon changed, and Republicans won historic victories in 1980 and 2010. The despairing Democratic Party of the late 1980s likewise managed to resurrect itself sufficiently to hold power through much of the following decade. The lesson is straightforward: complain all you like about defeat, but console yourself with the prospect of future recovery and victory, probably in as little as two years’ time. To that extent, the American political system is remarkably forgiving of even egregious failure.

But that system also depends on elections securing clear and commonly agreed outcomes, in accord with principles very clearly described in the Constitution. If those decisions are not accepted, and are subject to constant sniping and subversion, then that Constitutional settlement will simply run aground.

If people don’t learn to lose, the Constitution fails.

As for me, I will breathe again on December 20.

Philip Jenkins is the author of The Many Faces of Christ: The Thousand Year Story of the Survival and Influence of the Lost Gospels. He is distinguished professor of history at Baylor University and serves as co-director for the Program on Historical Studies of Religion in the Institute for Studies of Religion.