Into the Forge of Both Sides-ism
Give me grace.
Both sides have radicalized. Both sides are extreme. Both sides are the problem.
This is the responsible position, the one you take when you want to sound thoughtful without getting into a fight. I held it for years. It felt fair, measured, the kind of thing a serious person says at dinner without anyone needing to push back. You critique the right’s populism and the left’s institutional capture in the same breath, everybody nods, nobody feels accused or attacked, necessarily.
The data doesn’t cooperate. But the more interesting question turned out to be why the comfortable frame made it so hard to see.
I’ve rewritten this eight times, attempting to thread this needle.
Timur Kuran is an economist who noticed something strange about revolutions: they seem to come from nowhere. The Soviet system looked stable, then collapsed. The Shah seemed secure, then fell in weeks. His explanation was preference falsification. People misrepresent their private beliefs under social pressure, publicly affirming what they privately doubt. When everyone falsifies, no one knows what anyone actually believes. Public opinion becomes a hall of mirrors.
Kuran’s deeper discovery was what happens over time. If you spend long enough publicly affirming positions you privately question, you start to forget you ever questioned them. The performance overwrites the private belief. Preference falsification doesn’t just hide what people think. Given enough time, it destroys what people think.
I bring this up before the data, before the politics, because the “both sides” frame is itself a species of preference falsification. We say “both sides are the problem” not because we’ve examined the evidence and found symmetry, but because symmetric blame is socially safe. It doesn’t require choosing. It doesn’t invite accusation. And we’ve been performing it so long that many of us have forgotten there was ever a question to examine.
When you look past the frame, the picture isn’t symmetric.
In 2017, Pew Research published a study tracking ten core political values from 1994 to 2017. The median Democrat shifted substantially leftward, with the sharpest acceleration after 2011. The median Republican shifted rightward, but less dramatically and more gradually. The partisan gap more than doubled.
A 2026 study from Cambridge, published in Royal Society Open Science, put a finer point on it. Analyzing American National Election Studies data from 1988 to 2024, they found the left became 31.5 percent more socially liberal. The right moved 2.8 percent more conservative. On actual policy positions across three and a half decades, one side’s center of gravity shifted dramatically. The other barely budged.
These aren’t polls about one hot-button issue. These are composite measures across the full range of policy positions: immigration, race, criminal justice, the role of government, social values.
But notice what I just did. I presented a scoreboard. And the temptation, mine included, is to read it as vindication: see, they started it. That impulse is the frame talking. The numbers matter as evidence that the “both sides” story is hiding a causal chain. They don’t matter as a final score. The mechanism underneath is what we need to see.
The engine underneath
Two visions of human nature sit beneath every political argument. Thomas Sowell mapped them in A Conflict of Visions: the constrained vision sees people as fixed and flawed, so you design systems that work despite our limits. The unconstrained vision sees people as perfectible, so you get the right arrangement and solve problems the old systems only managed.
You don’t need the labels to see which one found institutional fuel after 2008. Universities, professional credentialing, media, corporate HR: they provided the infrastructure for one vision to accelerate while the other held roughly in place. Each new norm made the previous position look retrograde, which pushed the next norm further, which made the thing that was progressive five years ago look conservative now.
But beneath the institutional machinery, something deeper was at work. Karl Popper, writing in wartime exile while watching the most educated civilization in Europe devour itself, identified what he considered the most dangerous idea in Western political thought: historicism. The belief that history moves toward a knowable destination. Once you believe you know where history is going, three things follow. Opposition becomes irrational. Present costs become acceptable. And coercion becomes a moral obligation. You’re not forcing anyone. You’re helping them arrive where they were going anyway.
“The right side of history” is the tell. It sounds like moral clarity. It functions as prophecy. And prophets don’t need consent.
Popper’s insight explains why the left’s shift was institutional rather than merely ideological. Historicism changes what institutions do. If progress has a direction, then institutions that advance that direction are doing their job, and institutions that resist it are failing. The credentialing bodies, the editorial boards, the HR departments didn’t conspire. They each, independently, aligned with what felt like the direction of history. The result was a ratchet: each turn generated pressure for the next, because standing still now meant standing on the wrong side of something inevitable.
The most striking version happened on race. Zach Goldberg, analyzing ANES data, documented something counterintuitive: between 2012 and 2020, white liberals shifted so far leftward on racial attitude measures that on several survey questions they scored more “liberal” than Black and Hispanic respondents. The people most affected by racial discrimination shifted their attitudes moderately. The people most removed from it shifted dramatically.
That inversion is hard to explain as awakening. You’d expect genuine awakening to start closest to the experience and radiate outward. This started in seminar rooms, editorial meetings, social media and propagated out.
Immigration tells the same story from a policy angle. In 2006, twenty-six Senate Democrats voted for the Secure Fence Act, including Obama and Schumer. Border security was a bipartisan position. By 2018, enforcement was cruelty. People who didn’t change their mind on immigration found the categories changing around them.
Eric Hoffer’s observation about mass movements illuminates why these shifts felt so total. The doctrine matters less than what the movement provides psychologically: belonging, identity, purpose, someone to blame. If you’re twenty-eight in a major city with a degree that isn’t producing the life you expected, a movement that offers a moral framework, a community of the righteous, and a clear enemy answers questions that career and income are failing to answer. The psychological pull and the institutional enforcement arrived together. One side’s shift had both engine and fuel. The other side’s shift hadn’t happened yet.
This doesn’t mean racism isn’t real or that border enforcement is simple. It means the fuel for these particular shifts included more than new information about race or immigration. And when the advocates move further than the people they’re advocating for, it’s worth asking who the movement is serving.
What happened next
Then the right shifted. Differently.
The Cambridge researchers who documented the 2.8 percent policy shift offered an interpretation: the right’s recent energy reflects “outgroup animosity for a perceived ‘woke’ left,” as one of the study’s authors put it, more than firm belief in extreme policy positions. The right developed new tolerances for the previously unacceptable. It didn’t develop new ideas of its own.
By 2025, only 30 percent of Republicans strongly disapproved of January 6th, down from 51 percent in 2021. By 2023, 63 percent still believed the 2020 election was stolen. Between 2022 and 2024, the share who say “health of democracy” is a critical issue fell from 51 to 32 percent. On immigration, the share who say undocumented immigrants should not be allowed to stay legally jumped from 42 to 66 percent between 2020 and 2024.
Moderate Republicans fell below 20 percent for the first time in Gallup’s tracking: 18 percent in 2024, down from the low 30s at the turn of the century. But this looks less like existing conservatives radicalizing than like moderates leaving the party. The coalition didn’t move right. It shed everyone who wasn’t already there.
I need to say something directly here, because the structure of this essay could be read as softening the right’s pathology: reactive does not mean less dangerous. Election denial is real. Institutional nihilism is dangerous. The personality cult is dangerous. A car crash caused by one driver drifting into the other lane is still a car crash. The passengers in both vehicles are equally injured. The right’s norm erosion (anti-democratic proceduralism, January 6th, the delegitimization of institutions themselves) is its own crisis, and its consequences are playing out right now.
But the right’s shift didn’t materialize from nothing. The data shows a delay, not a parallel: the left shifted through roughly 2020, and the right’s hardening accelerated after. The left’s shift was programmatic: build through institutions, capture credentialing, make the new norms load-bearing. The right’s shift was reactive: tear it down, burn it out, replace deliberation with grievance. Same psychological mechanism, the demand for certainty in the face of exhaustion, wearing different clothes.
Popper would recognize both. The historicist plans the future against the will of the open process: once you know the destination, correction becomes coercion. The authoritarian offers relief from the chaos the planning created: the exhaustion of uncertainty produces a demand for someone who will just decide. Both are enemies of the open society. But one is cause and the other is consequence, and flattening the sequence into “both sides” makes both problems harder to fix.
Why we can’t say this
Here is where Kuran’s insight becomes structural.
The Cato Institute and YouGov found in 2020 that 62 percent of Americans say the political climate prevents them from sharing their views. Self-censorship was highest among Republicans (77 percent) and independents (59 percent), but a majority of Democrats (52 percent) reported it too. More than half of Democrats don’t feel free to say what they think.
The More in Common project, in their Hidden Tribes study, identified what they called the “Exhausted Majority”: roughly two-thirds of Americans caught between the politically active wings. Progressive Activists (about 8 percent of the population) and Devoted Conservatives (about 6 percent) drive the discourse. The other 86 percent watches.
This is preference falsification at scale. And the “both sides” frame is its signature product. We default to symmetric blame because naming the sequence carries social cost. Say the left shifted first and further, and from the left you’re a right-wing apologist. Say the right’s response is genuinely dangerous, and from the right you are some type of traitor. Symmetric blame is the position that avoids both accusations.
Kuran showed that revolutions surprise everyone because the gap between public expression and private belief was invisible until it collapsed. The “both sides” frame works the same way. It feels like analysis or fairness. It functions as camouflage. Underneath it, a large majority privately suspects the symmetry story is wrong, but publicly performs it because the cost of saying so is too high.
And the deepest damage operates exactly the way Kuran predicted. Perform the frame long enough and the original question disappears. You can no longer access the observation because it’s been buried under years of diplomatic equivalence. The frame doesn’t just hide the asymmetry. It erodes your capacity to see it. Knowledge corruption, Kuran called it: falsification so prolonged that the private belief itself degrades.
In the mid-1990s, journalists at major outlets roughly matched the general population on many political attitude measures. By the 2020s, that match was gone. Newsroom surveys show dramatic leftward clustering, particularly on social and cultural issues. This isn’t conspiracy. It’s Kuran’s mechanism operating inside institutions: when everyone in the room shares the same assumptions, those assumptions stop being assumptions. They become “how things obviously are.” The person who disagrees doesn’t get silenced. They just feel increasingly off, and eventually they leave or stop talking, and the room gets more uniform, and the frame survives because the room producing it no longer contains the perspective that would challenge it.
I notice something about myself as I write this. I keep reaching for balance. Every time I name the asymmetry, the instinct fires to immediately offset it with a matching critique of the other side. The pull toward symmetric blame is strong precisely because asymmetric diagnosis invites the very social cost I described three paragraphs ago. I am performing the preference falsification I’m writing about. The frame has its hooks in me too.
You don’t fix institutional capture by asking voters to calm down. You don’t fix reactive nihilism by asking institutions to loosen up. You fix each by naming each, which requires tolerating the discomfort of saying something that will be misread by someone on every side.
The 86 percent caught between the active wings, the ones performing compliance or calling themselves independent because the alternatives require pretending: they are the open society trying to hold. And the first step in holding is refusing the frame that makes the problem invisible: seeing the sequence clearly enough to diagnose each pathology for what it is, instead of flattening both into a symmetry that comforts everyone and helps no one.
Accurate diagnosis is not partisanship. Misdiagnosis for the sake of balance is.
Sources and Inspiration
A Conflict of Visions - Thomas Sowell
The True Believer - Eric Hoffer
Private Truths, Public Lies - Timur Kuran