I’ve been collecting this pattern for years. It shows up in every field where someone careful is trying to change someone else’s mind with evidence — and failing. Scientists, educators, clinicians, policy researchers. Different domains. Same wall.

You present the data. Irrefutable, in some cases — photographs, measurements, peer-reviewed consensus. The person doesn’t update. Doesn’t pause. In some cases becomes more certain than before you started.

What I kept noticing was that the experts asking “why doesn’t this work?” were asking a physics question about a problem that isn’t, at its root, a physics problem. They were using the right tool for the wrong system.

Here is what the system actually is.

1. Why doesn’t evidence work?

The short answer: because the person you’re talking to is not, in that moment, running an evidence-processing system.

Human cognition has more than one operating mode. There is a mode oriented toward genuine inquiry — open to updating, tolerant of uncertainty, capable of holding “I don’t know” without distress. And there is a mode oriented toward stability — one whose primary function is not truth-seeking but threat-management. Its job is to keep the internal system coherent, predictable, and safe.

When a belief is functioning as a regulatory tool — when it is the thing that makes a chaotic or frightening world feel ordered and comprehensible — that belief becomes structurally protected from revision. Not because the person is incapable of reason, but because dismantling the belief would destabilize the internal regulation it is providing.

Evidence, in that context, does not land as information. It lands as threat.

And when a system perceives threat, it does not open up and examine the incoming data with curiosity. It defends. This is not a character flaw. It is basic threat-response architecture, operating exactly as designed.

The astrophysicist presenting photographs of the curved Earth from space is, from inside the nervous system of a flat-Earth believer, not a colleague offering data. She is a destabilizing force attacking something that is keeping that person internally afloat. The sophistication of her evidence is irrelevant to that process. In fact, the more unambiguous her evidence, the more threatening it registers — and the more forcefully the defensive system responds.

This is why evidence doesn’t work. It is aimed at the wrong system.

2. Are conspiracy believers less intelligent or less educated?

No. And the research is fairly consistent on this.

Intelligence does not protect against motivated belief. In some measurable ways, higher verbal intelligence correlates with a greater capacity to construct sophisticated defenses for beliefs one already holds — a phenomenon researchers sometimes call “myside bias.” The smarter you are, the better you are at finding reasons why you are right.

Education is similarly unreliable as a shield. What matters is not the volume of information a person has access to, but the function a particular belief is serving for them. If a belief is providing regulation — a sense of coherence, identity, or safety — then more information does not dislodge it. It gets filtered, reinterpreted, or absorbed into the existing structure.

It is uncomfortable to sit with this, because it removes the explanation that feels most available: they just don’t understand. If they just don’t understand, the solution is simple — explain better, explain more clearly, explain until they do understand. That is a solvable problem, and it keeps the expert in a position of competence.

The actual picture is more uncomfortable: some people understand perfectly well and still do not change their position, because understanding was never the issue.

The question worth asking instead is not “what do they know?” but “what is this belief doing for them?” That is the question that leads somewhere.

3. Are they doing this consciously — do they know, on some level, that it isn’t true?

Mostly no, and the distinction matters.

This is not a performance. It is not, for the vast majority of people, a conscious choice to disbelieve evidence they privately accept. The belief is genuinely held — not despite the evidence, but through a filtering process so automatic and so complete that contrary evidence rarely arrives in its unmodified form.

Every information system — human or otherwise — has filters. The question is whether those filters are oriented primarily toward accuracy or toward coherence. In a system under chronic stress, or one that learned early that certain sources of authority could not be trusted, the filters tend to prioritize coherence. Incoming information gets processed through the question “does this fit what I already know to be true?” before it gets processed through the question “is this actually true?”

This happens below the level of deliberate thought. By the time the person is consciously aware of the evidence presented, it has already been evaluated by a faster, older system — and in many cases, quietly reclassified.

This is why conspiracy thinkers often describe the experience of encountering mainstream evidence as itself suspicious. It doesn’t feel like information to them. It feels like the very thing they’ve been warned about: the official narrative, the cover story, the thing powerful people want them to believe. The filter doesn’t just reject the evidence. It incorporates it as further proof.

That architecture is internally consistent. And it is almost entirely invisible to the person running it.

4. Why do they get more entrenched when I push harder?

Because pressure on a regulatory system produces more regulation, not less.

This is perhaps the most reliably observable pattern in these conversations, and it is almost perfectly predictable once you understand what is structurally happening. It has been documented well enough that it has a name — the backfire effect, though the precise conditions under which it reliably occurs are still debated.

The mechanism is straightforward: when a belief is functioning as a stabilizing structure, and an external force threatens that structure, the system does not respond with openness. It responds by reinforcing. The defenses thicken. The alternative explanations multiply. The emotional investment increases.

There is also a social dimension. By the time a person has argued their position in front of others — whether online or in person — backing down is no longer just an epistemic act. It is a social one. It means having been wrong publicly. It means, in many cases, having misled people they care about. The cost of updating has compounded far beyond the original question of whether the Earth is round.

Pushing harder treats the problem as if it were a debate to be won. But the problem is not a debate. It is a structure that is serving a function. Winning a debate does not eliminate the function. It just removes the current vehicle for serving it, which typically gets replaced by a stronger one.

The frustrating implication of this is that the correct response to entrenchment is often less pressure, not more. This does not mean agreement. It means that confrontation, in this context, tends to produce the opposite of its intended effect.

5. Why do some people believe one conspiracy but not others — is there a pattern?

Yes. The pattern follows function, not content.

If you look carefully at which conspiracy theories attach to which people, the content often looks random — flat Earth, moon landing denial, vaccine skepticism, political cover-ups, historical revisionism. But the structure is consistent: the belief positions the person as someone who can see what others cannot, operates outside a corrupt or incompetent mainstream, and belongs to a community of others who also see clearly.

The specific content of the belief is almost secondary. What matters is the combination of functions it serves:

Coherence: It explains why the world feels threatening, arbitrary, or unfair — not by chance, but by design. Someone is responsible. That is, paradoxically, more bearable than chaos.

Worth elevation: “I have done my own research. I see through the propaganda. I am not one of the sheep.” This is not a trivial function. For people whose sense of value and competence has been consistently undermined in other domains, the identity of “someone who knows the truth” can be profoundly stabilizing.

Group belonging: Conspiracy communities are often genuinely warm, engaged, and mutually supportive. The belief is the price of admission, but what members receive in return is real social connection and shared identity.

Institutional distrust: This one is worth sitting with carefully, because it is often the most legitimate component. Many conspiracy believers have direct personal experience of institutions — medical, governmental, judicial, educational — behaving badly, lying, or failing them. Their distrust is not irrational. It was learned. The problem is that generalized institutional distrust, once established, becomes a filter that cannot distinguish between genuine corruption and accurate science. Everything coming from official sources is pre-dismissed.

The pattern that emerges: people tend to adopt conspiracy beliefs in the domains where their unmet needs are highest. Someone who has experienced profound social exclusion finds belonging. Someone whose competence has been chronically dismissed finds expertise. Someone whose world has felt uncontrollable finds an explanation that restores the sense of a legible, if sinister, order.

This is not weakness. It is an adaptive response to real conditions. Understanding that does not require agreeing with the belief. But it changes what kind of response might actually be useful.

6. Is there anything that actually works? Is there a way in?

Yes. But it requires starting somewhere other than the evidence.

Everything that research and careful observation suggests is effective has one thing in common: it reduces the threat before introducing the challenge. Which is another way of saying: it changes the operating mode of the conversation before it tries to change the content of the belief.

This looks like a number of things in practice.

Genuine curiosity before correction. Not performed curiosity — the kind that is actually a setup for the correction you were planning to deliver anyway — but real interest in understanding how the person arrived at their position. This does more than create rapport. It shifts the dynamic from threat to something closer to inquiry. In a non-threatening context, the filters relax. Not always. Not quickly. But the direction of change reverses.

Affirming before questioning. This does not mean agreeing with the belief. It means acknowledging the legitimate core beneath it — the real distrust, the real experience of institutional failure, the real desire for a coherent explanation of a difficult world. When that is genuinely seen and named, the defensive function of the belief becomes temporarily less necessary.

Questions rather than statements. A question cannot be directly argued against. It creates a different kind of cognitive engagement — one that invites the person to examine their own reasoning rather than defend against yours. Not Socratic gotcha-questions, but genuine ones: What would it take to change your mind? Has anything ever made you question this? What was the moment you started to see it this way?

Time. Almost no one changes their mind in the moment. The most reliable evidence is of slow, gradual shifts — often traceable to a relationship, rather than an argument. The person who eventually updates rarely points to the decisive piece of evidence. They point to a person who made them feel safe enough to question.

None of this is a reliable formula. People in certain operating modes are not reachable by any method available to an outsider. But “not always” is genuinely different from “never,” and the conditions for the possible cases look remarkably consistent.

7. Should I keep trying — and if so, why?

This is not a question about conspiracy theories anymore. It is a question about what you are actually doing when you engage with this.

If the goal is to win — to produce, in the other person, a visible and immediate acknowledgment that you were right and they were wrong — the honest answer is that this outcome is rare, and pursuing it will exhaust you. The evidence-versus-belief dynamic is almost perfectly designed to produce the opposite of that outcome.

But if the goal is something else — to model a different way of engaging with uncertainty, to stay in contact with someone you care about, to chip at something slowly and without expectation — then the answer changes. Those things are possible. They happen. They require a different measurement of success.

There is also a third option, which is knowing when not to try. Not every conversation is a therapeutic opportunity. Not every interaction requires you to be the bridge. Protecting your own cognitive and emotional resources is not a failure of commitment to truth. It is a precondition for doing any of this sustainably.

The astrophysicist watching the moon-landing denier is not failing at science when her evidence doesn’t land. She is using the right tool for the wrong system.

What the problem actually is — a question about what human beings need to feel safe, coherent, and valued in a world that often fails to provide those things — is a different kind of inquiry altogether. I’ve been building a map for it. It started with a different question, in a different field, and kept expanding because the same shape kept appearing.

That map is what the questions above are drawn from. If something here finally made that behavior look like a system with a logic — that’s where the rest of it lives.

The framework behind these answers

The explanations in this piece derive from TEG-Blue — The Emotional Gradient Blueprint — a research framework that integrates neuroscience, developmental psychology, trauma research, and systems thinking into a unified map of how human beings regulate, and what happens when regulation fails.

TEG-Blue does not treat conspiracy thinking as stupidity, pathology, or moral failure. It treats it as a predictable output of specific structural conditions: systems that learned, at some point, that the world was not safe enough to be faced without protection — and that built what they needed to face it anyway.

Understanding that architecture doesn’t require agreeing with its products. But it makes those products legible. And legibility, for people trained to understand complex systems, is where useful work can begin.

The full framework is available at teg-blue.org — built for researchers, scientists, and those who think in systems.

TEG-Blue is an independent research framework. The models described in this piece represent a synthesis of over fifty established theories across neuroscience, trauma psychology, and systems science — integrated into a navigable architecture for understanding human emotional behaviour.

Series: Why Humans Are So Frustrating · No. 01

Last updated: 2026-03