Chip Somodevilla / Getty / The Atlantic

If it’s close, don’t forget Congress.

In the current anxiety over the possibility of a disputed election, attention has focused most on the battle that could rage in America’s courts to count the votes. But Al Gore’s acceptance of the Supreme Court’s judgment in 2000 has obscured a more likely venue for that fight: Congress.

The reason is a flawed statute for counting electoral votes—the Electoral Count Act of 1887 (ECA)—combined with a political reality that was unimaginable just five years ago.

We have entered a moment of irrational partisanship. If the election is close, yet marred with violence or credible claims—on either side—of fraud, that partisanship will trigger a self-righteous and irrational response, on both sides, and likely together. Each will have its reasons for the actions it takes. Each will believe itself justified by the wrongs charged to the other.

On the right, the election will be a story of fake ballots and illegal voting. On the left, the story will focus on votes suppressed, or mail-in ballots ignored. Those charges, if true, would be outrageous in any election. But long before any sober look could determine whether they are true, they will trigger action by the American people, and most importantly, by a political elite. Both sides will have a ready account for why their own actions are justified. Those accounts, supported by endless reporting and commentary by a media financially motivated to tribalize us, will steel each side to the extremes I describe below. Each will believe that it is saving the republic. That belief is dangerous at any time. Unmoderated by a balancing and disinterested force—what we once imagined broadcast journalism could be—it will inspire the worst in all of us.

I am not saying that both sides are equally untethered to reality. President Donald Trump’s preemptive discrediting of this election is baseless and inexcusable. Nothing the Democrats have done comes anywhere close to that wrong. No fair review of the president’s words and behavior could judge him as acting in good faith. He knows he has everything to lose; he will do whatever it takes to avoid defeat.

But I am saying that in a crucial way, to remix Frank Luntz, at a moment of irrational partisanship, what matters is not what’s true, but what people believe. And what we believe will come to us wrapped in radically different channels of media. In the frantic months between the election by the people (November 3) and the count of the electoral votes in Congress (January 6), irrational partisanship is likely to create two very different partisan realities. Each side will act based on its own reality, emboldened to defeat the heresy implicit in the other.

As we’ll see, it has before.

In an ordinary election, the votes of citizens select the presidential electors who then elect the president. Votes determine the winner in a state. In all but two states, that winner is awarded all of the state’s electoral votes.

Those electoral votes are then counted in Congress. The law governing that count was inspired by an election very much like the election we might be about to face. In 1876, like today, both sides believed they were fighting for the soul of a nation. Republicans were desperately trying to secure the equality that the Reconstruction Amendments promised, while Democrats were successfully turning much of the nation to a new version of perpetual caste for African Americans. Both sides believed that the other was engaging in massive fraud, or would if it could, to guarantee a win.

After the votes were counted, three states had more than one “slate” of electors purporting to represent their voters. That fact led Congress to establish an extraordinary commission to decide which slate would be counted. The commission ultimately, by one vote, chose the Republican slate in each contested state, a decision that allowed President Rutherford B. Hayes to win the presidency by a single electoral vote.

That mess led to the ECA. Though the act details many aspects of the process by which electoral votes select the president, its most important innovation was its method for resolving disputes when more than one slate of electors gets presented from any one state. When that happens, in its essence, the law first directs Congress to follow the resolution of state law; when state law fails, the law directs the two chambers of Congress to vote to decide which slate gets counted; and when the vote of the two chambers of Congress is split, the slate certified by the governor breaks the tie.

In the 130 years since Congress introduced this innovation, its mechanisms have never really been tested. Hawaii sent multiple slates of electors in 1960, but they didn’t matter to the result, and the presiding officer, Vice President Richard Nixon, allowed the slate that went against him, but expressly noted that his decision was not to set a precedent. In no other election has the tie-breaking feature of the ECA been necessary. Congress has not needed to insert its judgment between the judgment of the presidential electors in each state and the ultimate result announced by the joint session of Congress in early January.

But the ECA presumes that if Congress must step in, it will act in good faith. If states send multiple slates of electors to Congress—as they did in 1876—the ECA charges Congress to apply the law to the facts and, sorting through the proffered slates, choose the one that accords with federal law. In a word, the ECA demands that Congress make a legal judgment, not a political choice. It vests the body with the work of a court, not the discretion of 535 politicians.

Yet if the Senate’s trial of the impeachment of President Trump is any indication of Congress’s capacity to play this legal role, there is real reason for concern. As that trial showed, in a time of irrational partisanship, legal standards evaporate quickly. The law of impeachment, for example, required Mitch McConnell to swear an oath to be “impartial”—which he did while simultaneously telling anyone who asked that there was no way he would ever be “impartial.” In a court, such duplicity would be called perjury. The Senate calls it politics. And in our current moment, is there any reason to believe that the law would constrain Congress in selecting a president any more than in removing him? As one representative remarked in 1884, as Congress was considering what would become the ECA three years later,

It has been demonstrated time and again that the political conscience is a flexible and elastic rule of action that readily yields to the slightest pressure of party exigencies … When the great office of President is at stake … it would be expecting too much of human nature, under the tyranny of party, to omit any opportunity to accomplish its ends, more especially under that loose code of morals which teaches that all is fair in politics, as in war or in love.

The ultimate aim of the ECA is to ensure fidelity between the vote of the people and the declared results of the Electoral College. It does this by charging state officials with the task of certifying the results under existing state law, and by committing Congress to respect those determinations. Section 6, as currently codified, charges the governor with certifying which electors were appointed, along with the vote producing that result, and reporting to the archivist (and she or he then to Congress) the results of any dispute about that appointment. Section 5, the “safe harbor” provision made famous by Bush v. Gore, promises that Congress will respect the results of any contest, if resolved at least six days before the College is to vote, under rules set before the election. And Section 15 specifies the procedure that Congress is to follow in counting the votes cast, including the procedures for resolving contests when more than one slate gets presented.

These rules are cumbersome and incomplete, and again, historically untested by any real conflict. Yet if applied in good faith, they should secure accuracy in the declared results.

But the law leaves obvious openings for political mischief. The most obvious is a provision of law enacted 30 years before the ECA—Section 2 of the title governing the president, addressing “failed elections.”

Section 1 of that title sets the day on which electors are appointed—Election Day. Section 2 says that if the election has “failed to make a choice,” then the legislature gets to “direct” how the electors are appointed.

“Failed to make a choice” referred originally to a very specific case. Some states, such as New Hampshire, required that electors receive a majority (not a mere plurality) of the state’s votes. If no majority was achieved, there was a runoff. Section 2 covered that case, by allowing the states to specify a different way to select the electors if the initial conditions were not satisfied. To this day, many states rely upon Section 2 to resolve the case of a tie.

But “failed to make a choice” suggests something more than a technical rule about a required majority. And applying the plain meaning of those words, one could well suggest that they empower a legislature to pick a slate of electors whenever the election has, in a practical way, “failed.” Claims of fraud, or voter suppression, or the threat of a pandemic or illegal voting—all of these are conceivable reasons to claim that an election has failed. All of these give a plausible cover to any legislature choosing to appoint an alternative slate—at least if the election is close enough, in conditions that give reason to doubt.

And here is the proverbial camel’s nose under the tent. Nothing blocks politicians within the states from sending additional slates of electors to Congress. The ECA requires that the president of the Senate (now Vice President Mike Pence) open those “certificates and papers purporting to be certificates of the electoral votes.” What are “papers purporting to be certificates”? Who knows? Yet the law enters these certificates into the system that Congress must work through to decide who will become the next president.

For example, imagine that the vote in Florida is very close, but ultimately goes for former Vice President Joe Biden. The Republican legislature declares that the vote was fraudulent or incomplete, or illegally affected by the emergency of a pandemic, and that therefore the election has “failed to make a choice.” Imagine then that the legislature selects its own slate of electors, which it claims better represents the will of its people. And then imagine— and this turns out to be crucial—that the Republican Governor Ron DeSantis certifies that slate of electors, specifying precisely how many votes in the legislature each elector received. Those electors then meet on December 14, vote for President Donald Trump, and send their votes and certificates to Washington. Meanwhile, on the same day, the Democratic electors from Florida also meet, also vote, and also send their votes to Washington. On January 6, when the joint session of Congress meets to count the votes, Pence opens two slates from Florida — one for Biden, the other for Trump, and only one, Trump’s, certified by “the executive.”

Congress must then decide which slate it will count. Section 15 governs that procedure. When the joint session gathers, Pence moves alphabetically through the states. When he gets to Florida, he opens the two slates, and hands them to four “tellers”—two from each chamber, appointed to declare the votes as they are counted. Under the precedent of generations of practice, those tellers should report that at least one slate is not properly certified—the Democratic slate, missing the governor’s signature. The Republican slate is a closer call, since although Section 6 certainly meant for the votes of the people to be reported, the language speaks only of “votes cast” for any elector, which could refer to votes by the legislature, despite the earlier votes by the people. But let’s assume that both fail to be certified under Section 6, leading then to the next question under the ECA.

Defenders of each slate will then seek the protection of Section 5’s “safe harbor” provision. This was the provision at issue in Bush v. Gore, driving the Court to cut short the time for a recount to ensure that Florida could take advantage of that protection. But what the Court obscured is that there is no partisan-free procedure for determining whether the conditions of safe harbor actually apply. Specifically, what most seem to miss is that when any slate of electors gets presented to Congress, any member of Congress (if paired with a member from the other chamber) can object to that slate of electors having actually satisfied the requirements of safe-harbor status.

So if the Democratic slate from Florida is presented as having earned safe-harbor status, Representative Matt Gaetz and Senator Ted Cruz could sign an objection to that presumptive determination. Upon being presented with that objection, the two chambers separate and then vote on whether to sustain the objection or reject it. If both chambers vote to reject it, the Democratic slate would earn safe-harbor status, and be counted in the final result. But if the chambers split, then neither slate of electors would have safe-harbor status. As the DePaul University legal scholar Stephen Siegel describes in the most comprehensive review of the Electoral Count Act, “There is, in effect, a one-house veto over an electoral vote achieving [safe-harbor] status.”

If neither slate is presumptively valid against the other, then the joint session must pick which will be counted. If both chambers agree, that slate gets counted. But if they can’t agree, then the tie-breaking rule kicks in: If Congress is tied, but one slate is certified by the governor, that slate presumptively prevails. If we assume that Congress is divided, and that each house votes in a partisan manner, then under the rules of the ECA, the Republican slate from Florida would be counted—even if election officials in that state had certified the opposite.

My claim is not that such a game is always possible. Even in a time of irrational partisanship, the objections need a colorable basis: The vote must be close; the claimed flaw—whether fraud or pandemic-driven incapacity—must be substantial; the right of the legislature to intervene must be plausible, which it is because the ECA seems expressly to allow the legislature to act when the vote has “failed” (Section 2), and because the Supreme Court has (perhaps carelessly) suggested that the legislature can reclaim its power to appoint electors “at any time.” There may be no way that the Democratic governor of Louisiana could successfully conspire with Louisiana Democrats to generate an alternative slate—Louisiana is too solidly red. But swing states, such as Florida, are different.

And this is where the great unraveling could begin. When the parties recognize that this potential exists, and truly believe that they must do whatever they can to win, whatever the cost, both will have a strong incentive to send their own slates of electors to Washington wherever they can. And more important, where they can, each would have an extremely strong incentive to ensure that their slate is certified by the state’s governor.

If nothing but politics constrained either party, the Democrats would win this game—there are 288 electoral votes in states with Democratic governors (and three more in D.C.). But plainly, other constraints will guide party officials on whether to act against the vote of their own state’s citizens. No doubt, many will rightly worry that naked partisanship would poison their next election. But in close contests, marred by plausible claims of fraud, radical steps will be easier to defend.

There’s no predicting how this game ends. All we can say for certain is that at this moment, it is easier to imagine partisans stealing the election in Congress than it is to imagine changing the votes, or the counts of the votes, in enough states to change the ultimate results.  

So what could stop this great unraveling?

The simplest and most obvious solution is for the Democrats to take control of the Senate. If they do that, then they can decide as they like. Any objection to safe-harbor status could be rejected by both chambers. Any additional slate of electors could be ignored by both chambers. Thus, in a moment like ours, just as important as who presumptively wins the people’s vote for president is who controls both houses of Congress on January 6. Or put more directly: Chuck Schumer’s selection as majority leader could well be the necessary condition to Joe Biden becoming president.

But if the Democrats don’t take control of the Senate, the protections are less certain.

The best defense would be to inspire a good-faith application of the ECA by both houses of Congress. The safe-harbor protection of Section 5 enunciates a legal standard. So too do the certification procedures of Section 6, as well as the counting provisions of Section 15. If the politicians could resolve to apply the law, rather than a partisan view of the law, all of these problems would evaporate. Alternatively, following the suggestion of the Yale law professor Bruce Ackerman and Representative Ro Khanna, Congress could appoint another commission (as it did in 1876) to make that legal determination for it.

If that fails, the next line of defense will be the courts. State courts could intervene, and possibly enjoin state officials from wrongly sending an alternative slate. In theory, they could order the governor not to sign any unauthorized slate. That might well work. But if the politicians ignore those orders and certify the slates anyway, then the issue will shift to the federal courts.

And here there is a huge opportunity for the Supreme Court to settle a question that Bush v. Gore, unfortunately, left open: Once a state decides to hold an election, does the Constitution require that state to conform the selection of its electors to the results of that election?

You’d think the answer to that question would be clear. It is not. The Supreme Court has said, “When the state legislature vests the right to vote for President in its people, the right to vote as the legislature has prescribed is fundamental.” But immediately after this sentence, it wrote, “The State, of course, after granting the franchise in the special context of Article II, can take back the power to appoint electors.” The Court then quoted a case from a century before in which that take-back power was said to exist “at any time.” So can a legislature therefore ignore the results of the election? Is “after an election” within the scope of “at any time”?

The simplest fix would be for the Court to hold that while a state can cancel an election in advance, once the election has started, the people’s “fundamental” right (as the Court has called it) to have their vote matter cannot be recalled. That ruling would weaken the games that Congress could play, though not eliminate them. Remember, each legislature would be insisting that its slate actually represents the vote of its people better than the results of a flawed election.

Yet any federal judicial intervention in this process—especially once the counting by Congress has begun—is unlikely. This part of the process, at least, seems clearly committed to Congress; courts are therefore most likely to view the count as a political question, beyond the scope of judicial review. In principle, we could imagine a judge telling the politicians to obey the rules embedded in the ECA. In practice, counting on such an extraordinary judicial intervention would be extraordinarily risky.

Instead, again, the only certain path to preventing an irrational result at a moment of irrational partisanship is to remove the relevant partisan division. A Democratic Senate would protect against the games the ECA enables; nothing in the rhetoric of Democratic party leaders suggests they would play those games themselves, unless provoked.  

Our system for electing the president is old and fragile. It depends on both sides acting in good faith. But if we’ve entered an age when good faith is gone, and irrational partisanship has taken its place, then the system we have will not work with a divided Congress. In normal times, we can depend on politicians to do the right thing. In extraordinary times, the “right thing” becomes anything that will ensure a partisan result.

Nations don’t survive such moments well. In 1876, the deal that resolved the presidential-selection crisis condemned African Americans to almost another century of state-sanctioned oppression. If we’re to survive today, we must find a way to inspire something better than the poison that drives our politics, or make it irrelevant through an overwhelming result for the Democrats. We might not do either. That is reason enough to be afraid.

This story is part of the project “The Battle for the Constitution,” in partnership with the National Constitution Center.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.