Political Polarization Feels Unprecedented. The Record Says Otherwise — Mostly.

Poll Americans about the state of the country and a consistent result comes back: most people believe political divisions are at or near a historic peak. A majority describe the current moment as among the most dangerous for democracy in their lifetimes. Cable news hosts say it. Political scientists say it. Your uncle at Thanksgiving says it with particular emphasis.

Here's the problem: Americans said exactly the same thing in 1968, when cities were burning and a presidential candidate was assassinated during the primaries. They said it in 1919, when labor violence and the first Red Scare produced more domestic bombings in a single year than most people remember. They said it in 1864, when a sitting president wasn't sure he'd win re-election because the war he was prosecuting had become politically catastrophic. They said it in 1800, when the peaceful transfer of power between opposing parties was genuinely uncertain and both sides were printing pamphlets accusing the other of planning to destroy the republic.

And before America existed to have opinions about itself, the Athenians were saying it during the Peloponnesian War, the Romans were saying it during the late Republic, and the Tang Dynasty Chinese were saying it during the An Lushan Rebellion, which killed somewhere between thirteen and thirty-six million people and nearly ended the dynasty entirely.

The feeling of living through uniquely dangerous political times is not a reliable indicator of uniquely dangerous political times. But — and this matters — it's not nothing either. Some of those moments really did precede collapse. The trick is figuring out which variables actually predicted the outcome.

Building a Rough Map of Fracture

Comparing political crises across civilizations is not a precise science. The data is incomplete, the contexts are radically different, and anyone who gives you a specific "polarization score" for the Roman Senate versus the US Congress is doing something closer to numerology than history. That said, historians and political scientists have identified a set of conditions that appear with notable consistency in the periods that preceded genuine societal fracture — as opposed to the periods that felt catastrophic but ultimately resolved.

The conditions that tend to show up before recovery look something like this: intense partisan conflict exists, but institutional trust, while damaged, hasn't completely collapsed. Rival factions still accept the legitimacy of shared processes — courts, elections, legislative bodies — even while disputing specific outcomes. Economic grievances are real but distributed broadly enough that no single group has concluded that the system is structurally rigged against them specifically. And critically, the conflict has identifiable off-ramps: elections, negotiations, reforms that could plausibly address the underlying complaints.

The conditions that tend to show up before actual collapse are different in character, not just in degree. Institutional legitimacy doesn't just weaken — it fragments, with different factions recognizing entirely different authorities as valid. Economic distress becomes concentrated enough that specific groups have a rational case that they cannot win within the existing system. Violence shifts from episodic to systematic. And perhaps most importantly, elite defection accelerates: the people who have the most to lose from instability start hedging their bets on the stability itself.

The late Roman Republic hit all of these markers before it fell. The Weimar Republic hit them. The Tang Dynasty hit them during the An Lushan Rebellion. The American Republic in 1861 hit them — and then had a Civil War, which is the honest answer to "has America experienced genuine political fracture before."

Where the US Actually Sits

Here's the part that's genuinely complicated, and where honest analysis has to resist the pull toward either reassurance or alarm.

On the indicators that historically precede recovery, the US has some things going for it. Elections continue to occur and produce transfers of power, including ones that the losing side contested. Courts remain functional and their rulings, however disputed, are largely complied with. The military has not fractured along political lines, which is a variable that matters enormously — most historical collapses involved military defection or division. And public opinion data, for all its messiness, shows that large majorities of Americans across party lines still express preference for democratic governance over authoritarian alternatives, even as trust in specific institutions has eroded.

On the indicators that historically precede fracture, the picture is less comfortable. Economic inequality in the US has reached levels not seen since the 1920s, and the geographic and demographic sorting of that inequality means that the people at the bottom of it are increasingly concentrated in specific communities with specific political identities — which is precisely the condition that historical examples suggest makes economic grievance politically explosive rather than diffuse. Institutional trust has declined not just in government but in media, medicine, and law — the secondary institutions that historically buffer political conflict. And elite behavior, that canary-in-the-coal-mine variable, shows measurable signs of hedging: wealthy Americans have increased foreign asset holdings, dual citizenship applications have risen, and political donors who once funded both parties have increasingly concentrated on one side or the other.

None of this is determinative. The historical record is full of societies that hit multiple warning indicators and pulled back. The US in 1968 looked, on several of these dimensions, quite alarming — and it didn't collapse. The institutional resilience that Americans tend to take for granted is itself a historical variable: the US has a longer continuous constitutional history than almost any other nation on Earth, and that continuity has real inertia.

The Part Everyone Gets Wrong

The most common mistake in discussions of political polarization — from pundits, from politicians, from well-meaning op-ed writers — is treating it as primarily a cultural or psychological problem. The argument goes: if people would just talk to each other more, consume less partisan media, and approach the other side with empathy, the polarization would ease.

The historical record is not very supportive of this framing. In virtually every case where polarization resolved without catastrophe, the mechanism wasn't a change in how people felt about each other — it was a change in the underlying material conditions that had made conflict rational. The New Deal didn't reduce American political tension in the 1930s because it made conservatives and liberals like each other more. It reduced tension because it addressed, imperfectly but meaningfully, the economic conditions that were making political radicalization rational for large numbers of Americans.

Conversely, the periods that collapsed into genuine violence were almost never caused by insufficient empathy. The Roman Republic's late-period fractures weren't a communication problem. The conditions that produced the Civil War weren't resolvable by more civil dialogue. Real fracture tends to happen when the material stakes are high enough that factions have rational reasons to fight rather than negotiate.

This suggests that the most useful question isn't "how do we get people to be nicer to each other" — it's "what are the specific material grievances that are making conflict feel rational to large numbers of people, and what would actually address them."

Five thousand years of data doesn't give you a clean answer to that. But it does tell you which question to ask. And that, given how often we ask the wrong one, is a decent start.