The Loop
Three researchers in three different fields discovered the same thing. Nobody noticed they were saying the same thing.
In 1992, John Zaller published a book about how public opinion actually works. His finding was uncomfortable: most people don’t have opinions. They have impressions. When a pollster calls, they don’t retrieve a stable belief from storage. They construct one on the spot from whatever they heard most recently. The elite speaks. The public absorbs. The absorbed frame becomes “my opinion.” Nobody notices the ventriloquism.
In 2017, Dan Kahan at Yale published something worse. He showed that intelligence doesn’t fix this. It accelerates it. Smart people don’t evaluate evidence more clearly. They filter evidence more skillfully — in the direction of whatever they already believe. The more scientifically literate someone is, the further they diverge from people on the other side of a political question. Knowledge becomes a weapon for defending identity, not a tool for discovering truth.
In 2019, Robert Shiller published Narrative Economics and applied the same insight to markets. His argument: economic events don’t create narratives. Narratives create economic events. “Recession is coming” isn’t a prediction. It’s a cause. Enough people believe it, they cut spending, demand drops, and the recession arrives. The story writes itself into reality. Then the reality confirms the story. Then nobody questions where the story came from.
Three researchers. Three domains. The same loop.
The Architecture
Strip away the academic language and here is what all three found:
A belief forms before the evidence arrives. Once formed, the belief filters all incoming evidence. Confirming evidence is absorbed. Disconfirming evidence is dismissed, reframed, or never encountered. The belief strengthens. The strengthened belief filters more aggressively. The loop closes.
Zaller found it in the public. A news cycle frames an issue. People who haven’t thought about the issue absorb the frame. The frame becomes their position. When new information arrives, they accept it if it matches the frame and reject it if it doesn’t. The opinion hardens into something that feels self-generated. It was never self-generated. It was received, accepted, and defended.
Kahan found it in the individual. Identity-protective cognition. You don’t hold a belief because the evidence supports it. You hold it because it protects your membership in a group, your story about who you are, your model of the world. The evidence that threatens the model isn’t engaged with. It’s neutralized. And the smarter you are, the more sophisticated your neutralization.
Shiller found it in markets. A narrative spreads — “housing never goes down,” “AI will change everything,” “inflation is back.” The narrative changes behavior. Changed behavior changes prices. Changed prices confirm the narrative. Confirming prices recruit more believers. The narrative becomes self-fulfilling. Not because it was true. Because enough people acted as if it were.
The mechanism is identical every time:
Belief forms. Evidence filters. Reality conforms. Loop closes.
Where They Diverge
The researchers diverge on where the loop starts.
For Zaller, it starts with elites. Politicians, media, thought leaders set the frame. The public doesn’t originate beliefs. It receives them. The loop begins at the top and distributes downward.
For Kahan, it starts with identity. The group you belong to — political, professional, cultural — determines which beliefs are safe to hold. The loop begins with the tribe and radiates inward.
For Shiller, it starts with the story itself. Narratives have their own epidemiology. Some are contagious. Some aren’t. The ones that are contagious — emotionally resonant, simple, repeatable — spread through populations the way viruses spread through bodies. The loop begins with the narrative’s fitness, independent of any single person or institution.
These aren’t contradictions. They’re different entry points into the same system. Elites launch narratives. Identity determines which ones stick. Contagion determines which ones scale. All three are operating simultaneously, on the same people, all the time.
The Cost
The cost in public opinion is manufactured consensus. People walk around holding positions they believe are their own, arrived at through their own reasoning, reflecting their own values. Zaller showed that most of this is received, not reasoned. The feeling of having an opinion is not the same as having formed one.
The cost in individual cognition is sealed doors. Kahan showed that the smartest people are the most trapped — they have the strongest tools for defending a position, which means they have the strongest tools for never updating one. Intelligence without intellectual honesty is a lock, not a key.
The cost in markets is bubbles and crashes. Shiller showed that narratives can sustain prices at levels no fundamental justifies — and that when the narrative breaks, the reversal is violent precisely because the price was held up by story, not structure. The dot-com crash, the 2008 crisis, every crypto cycle — same loop, same collapse.
And the cost across all three domains is the same: people living inside models of reality that haven’t been tested against reality. Confident in beliefs that were never formed from evidence. Defending positions that were never chosen.
The Exit
Each researcher also, in their own way, identified the exit.
Zaller’s exit is direct experience. When people have personal, concrete experience with an issue, elite framing loses its grip. The frame can’t overwrite what you lived through. This is why the 99-point gap in AI attitudes exists — one afternoon with the tool does what no amount of media framing can undo.
Kahan’s exit is identity decoupling. When the belief stops being about “who I am” and starts being about “what works,” the filtering weakens. This is rare. It usually requires a situation where the cost of the wrong belief is personal and immediate. Abstract debates stay captured by identity. Concrete consequences break through.
Shiller’s exit is skin in the game. Prediction markets, short sellers, anyone who pays a price for being wrong — they update faster because reality charges admission. A narrative can survive in a survey indefinitely. It cannot survive in a market where someone is betting against it with real money. The narrative of “housing never goes down” lasted years in public opinion. It lasted exactly until traders started shorting subprime mortgage bonds.
Three exits. Same principle: contact with reality breaks the loop. Experience over framing. Consequences over identity. Cost over narrative.
The loop is strong. It runs almost everyone, almost all the time. But it has one weakness.
It cannot survive the thing it claims to describe.
“The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge.”
The loop is the illusion. Reality is the cure. The cure only works if you take it.