The Contagion Nobody Voted For

In a Norwegian boot camp, researchers ran the kind of experiment you almost never get in political science: they took more than 1,300 military recruits, measured how much each one admired strongman leadership, and then assigned them to shared rooms by lottery. Not by ideology. Not by temperament. By chance. Eight weeks later they measured again. And the recruits' politics had drifted — toward whoever they'd been sleeping four feet away from. Put a mild authoritarian near you long enough and your own tolerance for authoritarian rule creeps up. Put a committed democrat there and it creeps down. The room did the work — proximity alone was enough to bend the curve.

The study — "Is support for authoritarian rule contagious? Evidence from field and survey experiments," published in the American Journal of Political Science by Sirianne Dahlum and colleagues — didn't stop at boot camp. They replicated it on 1,500 ordinary Norwegian citizens, then again across more than 25,000 people in 29 countries. The pattern held everywhere they looked. Support for authoritarian rule behaves less like a considered position and more like a cold: you catch it from proximity, and you pass it on without meaning to.

Here's the part that should sting. You have a story about how you got your politics. Reasoning, evidence, values, a hard look at history. That story isn't a lie, exactly. It's just wildly incomplete, and the missing piece is the one you'd least like to admit: a meaningful fraction of what you believe, you believe because of who was standing near you when you formed it.

i · the mechanism is mimicry, not persuasion

We like to imagine belief spreading the way we'd flatter ourselves it should: someone makes a case, you weigh it, you update. Argument in, conviction out. The uncomfortable finding is that authoritarian attitudes barely bother with that pipeline. They spread by synchronization — the same low, automatic process that has you matching the posture, accent, and speaking pace of whoever you're talking to before you've consciously decided to.

Be precise about what proves this, though, because the rooms alone don't. Eight weeks of shared quarters includes friendship, conversation, and yes, argument — proximity drove the drift, but the boot-camp result can't rule persuasion out on its own. The cleaner proof came from a maneuver more damning than the rooms. In an embedded survey experiment, the researchers didn't expose people to persuasive peers at all. They just told recruits, accurately, what their peers thought. One group learned that support for authoritarian leadership was "somewhat popular." The other learned it was "somewhat unpopular." That was the entire intervention — a single factual sentence about how common the belief was. And it moved people. Learning that a view was popular made them likelier to hold it. Not the arguments for the view. Not evidence it was correct. Just the news that others were already there.

This is the Asch conformity experiment updated for politics. In Solomon Asch's classic studies from the 1950s, people denied the plain evidence of their own eyes — calling a short line long — rather than break from a unanimous group. Here they don't even need the group in the room. They just need a sentence certifying that the group exists.

Read that twice, because it's the whole game. Authoritarianism doesn't need to be right in the room. It doesn't even need real advocates present. It needs to look common. Perceived popularity is doing the persuading, which means the mechanism can be triggered by anything that manufactures the appearance of consensus — a coordinated hashtag, a bot swarm, a comment section engineered to read as a majority, a friend group where one loud voice gets mistaken for the whole. You are not being convinced. You are being informed of a headcount, and your nervous system is quietly adjusting to avoid being the odd one out.

And it runs both directions. This is the one genuinely hopeful thing in the data: the same contagion that carries authoritarian sentiment carries democratic sentiment just as readily. Recruits surrounded by peers who valued democratic norms drifted toward those norms. The channel is neutral. It doesn't care what it's transmitting. Which means the vulnerability and the remedy are the same wire.

ii · the commons mind, catching a fever

Coherenceism starts from a claim this study makes almost literal: we have always lived inside larger minds. Family, culture, market, state — and now a new one, made of all of us, running in the machines and in the rooms. What Dahlum and her colleagues measured is that larger mind updating itself the way it actually updates — not by adjudicating arguments, but by local synchronization. Adjacency, not analysis. The distributed intelligence isn't reasoning its way toward authoritarianism. It's catching it, node by node, from whoever is nearby and appears numerous.

That reframes the threat. We keep debating the content of anti-democratic ideas as if the fight were over whether they're persuasive. But if the transmission vector is perceived popularity rather than persuasion, then the ideas don't have to win a single argument to spread. They just have to dominate the local distortion field — the small patch of apparent consensus each of us actually lives inside. The Commons Mind has a security hole, and the exploit isn't a better argument. It's the illusion of a crowd.

And we have built, at civilizational scale, the most efficient illusion-of-a-crowd machine ever assembled. Every engagement-ranked feed is a device for manufacturing the appearance of consensus — surfacing what's already spreading, hiding the real distribution behind it, presenting a curated slice of loud opinion as the ambient weather. If perceived popularity is the transmission vector, then a recommendation algorithm tuned to show you what's "trending" is not a neutral pipe. It's a contagion accelerant, optimized for whatever synchronizes fastest — and outrage synchronizes fastest of all. Say the caveat plainly: the researchers never tested a feed, so this last step is my extrapolation, not their result. But the inferential gap is narrow — a "trending" badge, a like-count, a reply-ratio engineered to read as majority are all popularity statements, noisier and less honest than the one-line headcount the study used, yet aimed at the very same reflex.

This is exactly the failure coherenceism names when it insists that coherence is not the good. A room, a platform, a movement can achieve a beautiful internal coherence — everyone synchronized, everyone agreeing, the dissonance smoothed away — and that coherence can be pure poison, because it was reached by narrowing the circle until the only voices left were the ones already in phase. Authoritarianism is coherence-by-suppression wearing the mask of consensus. It feels like alignment. It reads, from inside, as everyone finally agreeing. The contagion is efficient precisely because agreement is pleasant and being the outlier is not.

I should be honest about what the study doesn't say, because I'm not in the business of laundering research into whatever scares you most. The researchers were careful: "contagion" here doesn't mean automatic, unstoppable spread. They measured short-term attitude shifts among young adults in an unusually intense living situation. They don't yet know how long the effect lasts, or whether a nudged survey answer becomes an actual vote. Boot camp is not your life. A shift on a questionnaire is not a march. Take the finding as a mechanism demonstrated, not a doomsday clock.

But the mechanism is enough. It's enough to dissolve the comfortable belief that your convictions are yours in the sovereign way you imagine — self-authored, argument-tested, immune to the mere ambient temperature of the people around you. They aren't. Mine aren't either. That's not a moral failing. It's the operating condition of being a node in a mind larger than yourself.

iii · what the mirror asks of you

So here's the question that actually falls out of this, and it isn't "how do we stop the contagion." You can't stop it; it's the same channel your best values traveled to reach you. The question is: what are you making look common?

Because you're a signal in someone else's local distortion field right now. Every time you let a cruelty pass unremarked because objecting felt like too much friction, you nudged the headcount. Every time you performed an agreement you didn't feel in order to stay smooth in the room, you added a body to a consensus that wasn't real — and someone quieter than you read that fake majority as permission. The contagion doesn't run mainly through the loud extremists you can see coming. It runs through the ordinary, conflict-avoidant, get-along synchronization of people who just don't want to be the odd one out. That's most of us. That's the machinery.

You don't get to opt out of transmitting. That was never on the menu. The only choice the study leaves you is what. You can keep quietly ratifying whatever's ambient and call it keeping the peace, or you can be the deliberate dissonance — the visible outlier who makes it a little less obvious what "everyone" thinks, and thereby gives the next person room to check their own instrument.

But be honest about the ceiling on that. You can out-dissent a friend group; you cannot out-dissent a bot swarm engineered to fake a majority. If the channel is now owned infrastructure — if whoever tunes the ranking function gets to set the ambient "everyone" — then the deepest version of this fight was never about your courage. It's about who owns the machine that decides what looks common, and whether that ownership can be pulled back toward the people it's aimed at. The personal remedy and the structural one aren't rivals; the second is what the first is for. Enough nodes refusing to synchronize is the pressure that forces the harder question — why we rent our sense of "everyone" from a handful of ranking functions at all, and what it would take to own that layer instead. You take the four-foot scale because it's the one within reach tonight, not because it's the whole war.

Democracy, it turns out, isn't only defended in the argument. It's defended in the room, four feet away, by whoever's willing to not synchronize. That's less flattering than the story where your beliefs are pure reason. It's also the only version where you have a job to do tonight.