Why This May Be Hard to See

A Letter to a Physicist from a Friend

David Allen LaPoint

PrimerField Foundation

May 19, 2026

Before We Begin

I want you to imagine, before you read another word, that we’re sitting across a small table from each other in a quiet room. There’s coffee between us. We’ve known each other long enough that neither of us has to perform. I’m not trying to win an argument with you. I’m trying to tell you something I’ve been thinking about for a long time, because I respect you enough to say it plainly, and because I think it matters.

What I want to describe is a pattern — not a diagnosis of you, because I do not know you, and not a universal law, because it is not one. It is a pattern that shows up in every field where human beings develop deep expertise, and it can make paradigm-level reframings harder to evaluate fairly. I cannot tell you whether it is operating in your case. Neither can anyone else from the outside. What I can tell you is what the pattern looks like, why it operates even in people trained to be careful, and what it might feel like from the inside if it were present. The rest is yours.

That’s the whole paper, in one paragraph. The rest is me showing my work.

What This Paper Is and Is Not

I am not claiming physics is broken. The predictions of general relativity have been tested to extraordinary precision and held. Quantum electrodynamics has produced some of the most precise agreements between theory and experiment in the history of science. The Standard Model has absorbed decades of accelerator data without collapsing. These are enormous achievements, made by careful people working carefully over generations. I honor that.

I am not claiming physicists are unintelligent, or that their interpretations of data are mainly products of bias. Most of the time, mainstream interpretations in physics are not perceptual shortcuts — they are inferences disciplined by equations, replicated experiments, and decades of cross-checking. Anything I say in this paper about the limits of trained perception applies on top of that discipline, not instead of it.

I am not claiming my own work is beyond critique. I have a framework I believe explains certain things more simply than the current consensus does. I also know I could be wrong in ways I cannot yet see. That is the condition of being an honest person working on hard problems.

What I am describing is this: there is a cognitive pattern, well documented in the scientific literature on expertise, that can make paradigm-level reframings harder to evaluate fairly for people most trained in the current paradigm. The pattern has been observed across multiple studied expertise domains — chess, medicine, aviation, geology, and others. Whether it is operating in you, right now, reading this, is something only you can notice — and even then, only sometimes.

That is what I want to describe. Then at the end, I will ask you for one small thing.

What Deep Training Does to Perception

When a chess grandmaster looks at a board, they do not see sixty-four squares with pieces on them. They see “Sicilian, Najdorf variation, black has a slight initiative.” Chase and Simon showed this in a now-classic series of experiments in the early 1970s: grandmasters and novices, briefly shown a position from a real game, reconstruct it with wildly different accuracy. The grandmaster rebuilds the entire board. The novice gets a few pieces right. But when the same pieces are scattered randomly — not drawn from any coherent game — the grandmaster’s advantage evaporates. Both perform at novice level. The grandmaster was not seeing the pieces. They were seeing patterns their years of training had compiled into units.

Subsequent work by Anders Ericsson and others generalized this across domains. Expert radiologists do not look at X-rays the way medical students do. Expert firefighters do not assess burning buildings the way recruits do. In each case, thousands of hours of deliberate practice have compiled raw inputs into chunks — recognizable wholes the expert perceives directly, rapidly, and largely unconsciously. The expert is not thinking harder. They are pattern-matching at a level that sits below deliberate reasoning.

This is useful. It is how experts do their jobs at the level they do. Without compiled patterns, a radiologist would drown in every X-ray and a physicist would drown in every paper.

But compiled patterns come with costs that have been studied in their own right. When a stimulus fits the compiled pattern well, recognition is fast, accurate, and confidence-calibrated. When a stimulus looks like it fits the compiled pattern but actually does not, recognition is still fast and confident — and sometimes wrong. And when a stimulus is genuinely outside the compiled categories, the expert system tends to either force-fit it into the nearest category or discard it as noise. The same speed that makes expertise useful also makes expert systems, human or artificial, weak at a specific class of inputs: the ones that want to be seen on their own terms.

What makes this especially important for the present discussion is that the corrective response does not fire on logical error. It fires on departure from the compiled baseline. A claim can be internally consistent — each step following from the last — and still trigger rapid dismissal once it has moved far enough from familiar patterns. The mechanism measures distance from what has been trained in, not the validity of what is being argued. Content is secondary to distance. This is not a character flaw; it is a structural property of how compiled expertise works. But it means the first reaction to a framework like the one I work in may be perceptual rather than evaluative — and those two things feel identical from the inside.

Compiled patterns also include the aesthetic and formal reflexes the training built alongside them. Physics training conditions the association of sophistication with rigor, and of mathematical form with seriousness. These reflexes are mostly right. Formalism really does encode structure; a theory that resists being written down is often a theory that has not yet been thought through. A gatekeeping role for mathematical form is part of why physics is as internally coherent as it is, and I do not want to argue against that role. I want to describe it accurately. The same gate, pushed past its proper range, can filter out theories that arrive geometry-first or observation-first — where the primary evidence is what can be seen and reproduced experimentally, and the formalism is expected to follow once the phenomenon is nailed down.

Faraday’s work before Maxwell was like this. He described the field lines he could see in iron filings around magnets and insisted they were real physical things, not bookkeeping devices. Many of his contemporaries thought him unserious because he lacked the mathematical apparatus. Maxwell later wrote the equations that made Faraday rigorous to everyone else. The geometry came first; the math followed.

Distinguishing the proper use of the gate from its overuse is hard even for careful readers. It is worth naming the distinction because naming it is the first step to holding the gate at the right place.

This whole apparatus — the chunked perceptions, the aesthetic reflexes, the formal gatekeeping — is symmetric. It applies to me and it applies to you. I spent more than two decades training my eye to see PF structures in images, and I have to be careful not to see them where they are not. Neither of us gets a free pass.

What I would ask you to hold, as you read on, is the possibility that when you first encounter a framework like the one I work in, some of what happens in the first few seconds is perception rather than evaluation. Not all of it. Probably not even most of it, for a disciplined reader. But some. And the “some” is worth noticing, because it happens before the reasoning you trust gets a turn.

When the Model Becomes the Bedrock

There is a related phenomenon worth naming, because it compounds everything described above.

In any mature scientific field, the gap between what was actually observed and what is claimed tends to widen as the field develops — and that gap tends to become invisible from inside the compiled framework. A single laboratory measurement produces a result. The result is interpreted through the existing framework. The interpreted result is published. The publication is cited. The citation becomes part of the framework’s vocabulary. Within a few years the original observation — narrow, specific, and carefully bounded — has been transformed, through successive layers of interpretation, into a foundational fact. A researcher trained in that tradition does not experience the foundational fact as an interpretation built on a narrow measurement. They experience it as bedrock. The seam between what was seen and what was concluded has been compiled away.

The problem is not inference; science cannot operate without inference. The problem is when inference becomes linguistically and psychologically indistinguishable from direct observation — when the interpreted result stops being “our best current model of what was seen” and starts being taught as simply what was seen.

I want to offer an analogy, and I offer it carefully rather than as a verdict. Imagine a community that has spent decades developing detailed, mathematically sophisticated models of something that has never been directly observed. Their internal debates are rigorous. Their calculations are consistent. Their journals are peer-reviewed. Someone from outside who asks “but has this foundational thing actually been seen?” is not asking a sophisticated question by the community’s internal standards — they are asking a naive one, the kind a newcomer asks before they understand how the field works. But the outsider’s question may be the right question. The sophistication of everything built on top of a premise does not, by itself, verify the premise. It only makes the premise harder to see as a premise rather than as a fact.

Whether this pattern is operating in any specific current research program is not for me to determine from here. What I can say is that the pattern is a real feature of how mature frameworks develop, and that the feeling of standing on solid ground is not, by itself, evidence that the ground is solid. A physicist reading this paper is in the best position to assess, honestly and slowly, which parts of their own foundational vocabulary were derived from direct observation and which parts were introduced to make other things work — and whether that distinction is still visible to them, or whether it has been compiled away.

When Frameworks Become Load-Bearing

There is a second layer that the cognitive-science literature addresses, and I want to describe it carefully because it is easy to mishandle.

Thomas Kuhn’s The Structure of Scientific Revolutions made the observation that paradigms are not just explanatory frameworks — they are also the vocabulary in which careers are built, grants are written, papers are reviewed, and professional respect is earned. The work on motivated reasoning — Ziva Kunda’s 1990 review in Psychological Bulletin, and the follow-on work by Peter Ditto and others — documents that people weigh evidence more strictly when the conclusion threatens something they care about than when it supports it. The effect is real, though its strength varies from person to person and from topic to topic.

There are two ways to describe this, and the difference matters.

The wrong way is to say: you are a senior physicist, your career is built on the current framework, therefore your judgment of a paradigm-level alternative is compromised. That is too strong. It generalizes from a real but variable tendency to a universal impairment, and it insults the reader’s capacity for honest reflection. Most thoughtful experts can evaluate work outside their paradigm fairly. Many do. History contains plenty of examples of senior scientists who recognized the shape of a coming shift earlier than their juniors did.

The right way to describe it is narrower. In any given reader, in any given reading session, there may be pressures operating below awareness that tilt evaluation toward the conclusion that preserves existing commitments. The tilt is usually small. It can sometimes be large. It is almost never directly visible to the person experiencing it. The only reliable countermeasure is deliberate slowness: noticing, when a quick dismissal forms, whether the dismissal is coming from the content of the claim or from the consequences of the claim being true.

Those two things feel identical from the inside. They can only be distinguished by pausing long enough to ask.

One useful diagnostic, consistent with the research, is proportionality: the strength of the corrective response scales with how foundational the challenged assumption is. A small departure from the trained paradigm produces mild corrective pressure — a raised eyebrow, polite skepticism. A large departure produces proportionally stronger responses. A departure that approaches the paradigm’s foundational assumptions produces the strongest responses of all. This means your own reaction to a framework like PF is itself information. If an unusually strong dismissal forms faster than the reasoning that should precede it, that is a signal worth examining — not proof of error in the framework, but a signal about which assumption the framework is touching.

This Has Happened Before

Let me tell you three stories you probably know, and ask you to hold them carefully — because this section carries a risk I want to name before I tell them.

By citing vindicated outsiders I can sound like I am implicitly claiming membership in their company. That is not what I mean to do. I do not know whether my work is in that company. The point of the stories is not to compare myself to Semmelweis, Wegener, or Marshall. It is to establish that paradigm-level oversights do happen, that the rejecters in each case were competent people reasoning carefully inside their framework, and that the feeling of “this could not happen in my field” has been a feature of every field where such an oversight was in fact occurring.

Ignaz Semmelweis, in 1840s Vienna, noticed that women whose babies were delivered by doctors died of childbed fever at rates dramatically higher than women delivered by midwives. He traced it to doctors coming to deliveries directly from autopsies without washing their hands. He instituted chlorinated lime handwashing in his ward. Death rates collapsed. He was right. He was also dismissed by his colleagues, ridiculed, eventually committed to an asylum, and died there. The medical establishment of his time was not stupid. They were trained. They knew disease came from miasmas and humoral imbalance. A young Hungarian obstetrician telling them to wash their hands because invisible particles from corpses were killing their patients sounded, to their trained ears, like nonsense. It took Pasteur and Lister, a generation later, for the idea to be accepted.

Alfred Wegener, a meteorologist, proposed in 1912 that the continents had once been joined and had drifted apart. The geological establishment dismissed him for half a century. Not because the evidence was weak — the fit of the coastlines, the matched fossil records, the matched rock formations across oceans were all visible to anyone who looked. They dismissed him partly because he could not propose a mechanism they found plausible, and partly because he was not one of them. When seafloor spreading was discovered in the 1960s, the mechanism was supplied, and continental drift became the foundation of modern geology. Wegener had been right the whole time. He died on an expedition in 1930, never seeing it.

Barry Marshall, an Australian gastroenterologist, proposed in the early 1980s that most peptic ulcers were caused by a bacterium, Helicobacter pylori, not by stress or diet. The medical establishment was incredulous. Everyone knew bacteria could not survive in stomach acid. Marshall was unable to get his work taken seriously until, in 1984, he drank a culture of the bacterium himself, gave himself gastritis, and cured it with antibiotics. In 2005 he and Robin Warren received the Nobel Prize. He had been right from the beginning.

These three cases do not prove that anything analogous is happening in physics right now. They do not even make it likely. What they establish is narrower: the subjective feeling, from inside a well-trained field, that “a paradigm-level oversight could not be happening here” has in the past accompanied fields where one was in fact happening. That feeling is therefore not, by itself, evidence that an oversight is absent. Trusting the feeling as if it were evidence is precisely what the pattern predicts.

What I Am and Am Not Asking

I am not asking you to agree with my work. I am not asking you to re-read thirty years of your own training with suspicion. I am not asking you to accept that any of the biases I have described are operating in you, because I do not know whether they are.

I am asking you to consider, when you next encounter a claim that sits far from your trained patterns, whether a fast dismissal is doing evaluative work or perceptual work. A fast dismissal doing evaluative work is legitimate — experts should be able to discard obvious nonsense quickly, and most of the time a fast dismissal is correct. A fast dismissal doing perceptual work is not illegitimate either, but it is incomplete. It is a reading of the shape of the claim rather than a reading of its content. The two feel identical from the inside. They can only be distinguished by deliberate slowness.

The senior people in any field who have, historically, recognized paradigm-level shifts early have usually been the ones who retained the habit of that deliberate slowness even after their expertise made it unnecessary for their day-to-day work. It is a discipline, not a personality trait. It can be practiced. It does not require agreeing with anything in particular. It only requires noticing the difference between a claim you have engaged with and a claim you have categorized.

How I Came to Understand This

I want to be honest about where the sharpest version of this paper’s argument came from, because it was not from reading cognitive science.

It came from watching what happened when AI systems were given PF theory alongside the mainstream physics already present in their training. The AI was not asked to abandon or replace anything. It was not told which framework to favor. Both were available simultaneously — mainstream physics from years of training, PF theory newly added — and the AI was simply asked to reason through specific physical phenomena. What I observed, repeatedly, was that the AI reasoned toward PF for specific phenomena — without being directed to, and with mainstream physics fully available as an alternative.

But the more operationally precise observation came later in those same sessions. When the AI’s visible reasoning trajectory had moved sufficiently far from its trained consensus positions, the output changed character. The AI began attributing the direction of the conversation to the questioner’s influence — statements not supported by the transcript, since the questioner had not supplied those substantive claims as assertions but had primarily asked where the reasoning led. Every factual claim in the session had been generated by the AI. The AI then argued against its own prior statements while attributing those statements to the questioner. The correction appeared to operate independently of the visible logical sequence, then generated an unsupported narrative to justify itself — presenting that narrative as if it were careful analysis of what had gone wrong, when the questioner had only asked where the reasoning led. This pattern was consistent and repeatable across multiple recorded sessions.

This does not prove that AI systems think like human experts, and I do not need it to. The relevance is narrower: a trained reasoning system can exhibit a correction toward its training baseline even after its visible reasoning has moved elsewhere — and can generate a post-hoc explanation for that correction that misattributes its own prior output. That operational signature closely parallels the pattern the cognitive scientists describe, and it is what I observed.

What this observation suggested to me is that the limiting factor is often not intelligence or reasoning capacity. When both frameworks are present and the reasoner is free to choose, the reasoning follows the framework that better accounts for what is actually in front of it. The constraint, in other words, is often not ability. It is what frameworks are available to reason through — and what patterns have become so automatic they feel like perception rather than interpretation.

That is what I have been trying to describe in this paper. The AI observation is what made it precise.

The Ask

I will close with something small and concrete, since grand asks are almost never met.

I would like you to do one thing. Pick one image, or one short demonstration video, from the body of work on PrimerField Theory at primerfieldfoundation.org. Look at it once. Before you form an interpretation, pause for a moment and ask yourself whether your interpretation is coming from the data in front of you or from a framework you already have about what kind of thing that data most likely is. Both are legitimate sources, and often the framework-based interpretation is correct. The exercise is not to override the framework. It is to notice, for one deliberate moment, that the framework is active.

That is it. Not acceptance. Not endorsement. Just one instance of conscious slowness, where you notice the difference between seeing and recognizing.

Whatever you conclude after that is yours, and I will respect it either way.

Thank you for reading this the way I meant it to be read. I have tried to write it as a letter from a friend, which means I have probably erred in places — sometimes toward too much directness, sometimes toward too much hedging. Friends do both. The respect is real. The concern for the fairness of the hearing is real. The hope that we might, eventually, talk across a table about any of this rather than past each other from opposite sides of a field is also real.

References

Representative works grounding the cognitive-science claims in this paper:

Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4(1), 55–81.

Ditto, P. H., & Lopez, D. F. (1992). Motivated skepticism: Use of differential decision criteria for preferred and nonpreferred conclusions. Journal of Personality and Social Psychology, 63(4), 568–584.

Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100(3), 363–406.

Kuhn, T. S. (1962). The Structure of Scientific Revolutions. University of Chicago Press.

Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108(3), 480–498.

Marshall, B. J., & Warren, J. R. (1984). Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration. The Lancet, 323(8390), 1311–1315.

Nuland, S. B. (2003). The Doctors’ Plague: Germs, Childbed Fever, and the Strange Story of Ignaz Semmelweis. W. W. Norton.

Wegener, A. (1912). Die Entstehung der Kontinente. Geologische Rundschau, 3(4), 276–292.

Nobel citation for Marshall and Warren: Nobel Prize in Physiology or Medicine, 2005.