Two stories that appeared within days of each other in the American press this fall might seem, on their face, unrelated. In one, the White House defended its decision to bar CNN, MSNBC, and Politico from routine access, with the Department of Justice arguing in court filings that the outlets' reporting "threatened national security and disseminated false information." In the other, sources describe growing concern within government that President Trump may deploy the National Guard near polling places in the coming elections, notwithstanding state and federal laws that explicitly forbid a military presence at voting sites.

Treated in isolation, each story invites a narrow legal analysis: does the First Amendment protect a press outlet's access to the White House press pool? Does the Insurrection Act, or some novel reading of it, permit troops near ballot boxes? But history rarely lets us examine these things in isolation, and it shouldn't now. The pattern, not the individual episode, is the story. And the pattern is a familiar one to anyone who has studied how democracies actually die: not in a single dramatic seizure of power, but in the quiet, sequential disabling of the institutions that referee political contests.

The Press as the First Casualty

Governments attacking journalists is not new to American history. What is unusual is the explicit legal framing now being offered: that unfavorable coverage itself constitutes a security threat justifying exclusion. CNN, MSNBC, and Politico have sued, arguing the ban amounts to retaliation for editorial decisions rather than any legitimate security concern. The DOJ's counter-argument, that the outlets spread "false information," should give any student of history pause, because it is nearly a textbook definition of the pretext regimes use to narrow the space for public contestation.

Political scientists studying democratic backsliding describe it as "a process of regime change towards autocracy that makes the exercise of political power more arbitrary and repressive, and that restricts the space for public contestation and political participation." That definition was not written with Trump in mind; it describes a general phenomenon observed across dozens of countries and, notably, applied to periods of American history itself, including the Jim Crow-era South, where the machinery of democracy was hollowed out from within while its outward forms—elections, courts, newspapers—remained nominally intact.

The mechanism then was not a formal repeal of Black suffrage. It was the Red Shirts, paramilitary terror groups that emerged in Mississippi in 1875 and spread through the Carolinas, using organized intimidation rather than legal abolition to strip the right to vote of its meaning. The genius, if one can call it that, of that strategy was its deniability. No law on the books said Black Southerners could not vote. Violence and threat simply made the exercise of that right too dangerous to attempt. The forms of democracy survived; its substance did not.

Troops at the Ballot Box: An Old Fear With New Urgency

This is why the National Guard story deserves more attention than a single news cycle. The laws against military presence at polling places exist precisely because American history, and world history, offers ample warning about what happens when armed state power positions itself near the machinery of an election. The concern is not that soldiers would forcibly stop anyone from voting. The concern is the same one that animated the Red Shirts: that visible force near polling places changes turnout not through law but through fear, particularly in the communities most likely to oppose the party deploying it.

Consider Russia in 1917. The Bolsheviks did not abolish elections outright. They allowed a genuinely free vote for the Constituent Assembly, then dissolved it the moment the Socialist Revolutionaries won the majority the Bolsheviks hadn't anticipated. The lesson historians draw from that episode is not that Lenin hated elections; it is that he was willing to tolerate them only so long as they produced the correct outcome, and abandoned them the instant they did not. The infrastructure of democratic choice—ballots, assemblies, voting rolls—can be preserved as theater long after its actual power to constrain rulers has been hollowed out.

A similar logic played out half a world away and decades later. The 1954 Geneva Accords called for nationwide Vietnamese elections in 1956 to reunify the country. Ngo Dinh Diem, with American backing, simply refused to hold them, calculating (almost certainly correctly) that Ho Chi Minh would win. The elections were not rigged. They were cancelled, because cancellation was easier than losing.

Angola offers a still darker variant. The Lisbon Peace Accords of May 1991 produced elections in September 1992, which the MPLA won. Jonas Savimbi, having lost, simply rejected the results and returned to civil war. Within roughly eighteen months, 3.3 million people were refugees and a thousand were dying daily. The mechanism of an election, faithfully administered, offered no protection at all once one side decided the result was unacceptable and had the guns to say so.

None of this is to suggest the United States in 2025 is Bolshevik Russia, Diem's Vietnam, or Savimbi's Angola. It is to suggest that the protective assumption many Americans carry—that democratic collapse would look obviously different from ordinary political conflict, that we would surely notice—is not well supported by the historical record. Backsliding rarely announces itself. It accumulates.

What Makes This Moment Different, and What Doesn't

There is a countervailing historical lesson worth taking seriously too: authoritarian ambition and authoritarian capacity are not the same thing. Mussolini, for all his rhetoric of totalitarian control, never actually eliminated the independent power of the Italian king, who dismissed him from office in 1943, nor of the Catholic Church, which retained enormous social authority throughout the fascist period. Historians now generally describe Mussolini's Italy as authoritarian rather than truly totalitarian, precisely because rival institutions retained enough independent standing to survive and eventually help end the regime.

America in 2025 has real rival institutions still functioning. CNN, MSNBC, and Politico did not simply accept exclusion; they sued, and a federal court is actively weighing their claim. That is not nothing. It is, in fact, the system working exactly as designed: a free press using an independent judiciary to check executive retaliation. The DOJ has to argue its case in open court rather than simply decree the outlets banned by fiat. Reports of concerns about ICE detention facilities, however unsettling, were themselves surfaced through the kind of document review that a free press exists to conduct.

And crucially, the public does not appear to be sleepwalking through this. Recent polling shows Trump's unpopularity is a dominant factor influencing voters, with his standing weighing heavily on the electorate's mind. Whatever else is true, an electorate that dislikes a president and says so in polls, and that will get the chance to express that dislike at the ballot box, is not yet an electorate that has lost its voice.

That is the genuine, if fragile, distinction between this moment and the darker precedents above. The Constituent Assembly in Russia was dissolved after it voted the wrong way. The Vietnamese elections were simply never held. The Angolan election's loser had an army of his own. In the American case, the courts are open, the polls are being conducted and published without apparent interference, and the election itself has not been cancelled or delayed. The guardrails are being tested, not yet removed.

Why Vigilance Now Matters More Than Panic

The historical record on democratic erosion suggests a specific kind of vigilance is warranted, one focused less on any single action and more on the accumulation and the trajectory. A press ban defended in court is a legal dispute. A press ban defended in court while National Guard deployment near polling places is being seriously discussed internally, while concerns about detention infrastructure emerge, while a president openly treats unfavorable coverage as an existential threat rather than an irritant, is a pattern. Patterns are what history teaches us to watch for, because individually deniable actions can add up to something that, in aggregate, none of the individual actors would ever admit to intending.

The Jim Crow South did not require a single law abolishing Black suffrage. It required a few decades of accumulated intimidation, disenfranchisement statutes justified on other grounds, and the exhaustion of anyone who tried to fight each individual instance in isolation. That is the trap of incrementalism: each step is defensible on its own terms, and by the time the cumulative effect is undeniable, the institutions that might have stopped it have already been weakened by the steps that preceded it.

The remedy history offers is not despair but specificity. Courts still hearing press access cases need to rule, and rule quickly, because delay itself is a form of erosion. Congress and state officials who have jurisdiction over Guard deployment and polling place security need to state clearly, publicly, and in advance that troops will not appear at voting sites, removing ambiguity before it can be exploited. And voters, who according to the polling remain very much engaged and very much aware of what is at stake, need to keep doing the one thing that has, in every historical case examined here, actually mattered: showing up, and making their preference for the rule of law over the discretion of any single man unmistakable.

Democracies that survive stress tests are not the ones that never faced them. They are the ones where enough institutions—courts, a free press, an attentive electorate—held their ground long enough for the moment to pass. That is still possible here. But it is not guaranteed by the mere existence of a Constitution, any more than it was guaranteed by the existence of a Constituent Assembly in 1917 or an election commission in Angola in 1992. It is guaranteed, if at all, by people who notice the pattern and act while there is still time to matter.