They Aren’t Arguing With You
On predictive processing, the consolation of misreading, and what a reader’s objection actually tells you
There is a particular frustration reserved for careful writers. Not the frustration of being ignored — that’s clean, if lonely — but the frustration of being responded to. Of having spent weeks honing an argument, stripping it of opinion, qualifying every claim that needs qualifying, and then receiving a reply that addresses something you never said. A position you didn’t take. A motive you don’t hold. A conclusion you didn’t reach.
The writer quoted in the exchange that prompted this post put it with admirable restraint: “The response to my work is data about the reader, not the argument.” He called it either consoling or alarming, depending on the day.
We think it’s something more specific than either. We think it’s a window — straight into the architecture of how minds actually work. And once you understand the mechanism, the misreading doesn’t just become tolerable. It becomes, in the truest sense, evidence of learning.
The Brain Is Not a Receiver
The standard model of reading goes something like this: words carry meaning, meaning travels from the page to the mind, the mind decodes it. The writer encodes; the reader decodes. Communication is transmission.
This model is wrong. Not approximately wrong — structurally wrong.
The brain is not a passive receiver of information. It is a prediction machine. At every moment, it is generating a model of what it expects the world to be — including what it expects the next sentence to say, what it expects the author to believe, what it expects this argument to imply. When new information arrives, the brain doesn’t process the information itself. It processes the gap between the information and its prediction.
This is the core claim of predictive processing, the most influential framework in contemporary cognitive neuroscience, developed most rigorously by Karl Friston and elaborated by Andy Clark, Jakob Hohwy, and others. The brain is a hierarchical generative model — it builds layered predictions about incoming signals at every level of perception and cognition, and propagates only what doesn’t fit: the prediction error.
What this means for reading is profound. A reader doesn’t encounter your text in any neutral sense. They encounter the delta — the divergence between what you wrote and what their existing model expected you to write. Everything that confirms the model passes through nearly frictionless. Everything that violates it generates a signal. That signal demands resolution.
Precision Makes It Worse. That’s the Point.
Here is the counterintuitive insight that the exchange above surfaces: a precise argument generates more heat than a vague one.
Why? Because vagueness is accommodating. A loose claim leaves enough room for the reader’s model to project agreement onto it. The reader fills the gaps with their own prior beliefs and finds you, pleasingly, correct. No friction. No signal.
A precise argument is a different kind of object entirely. It closes those gaps. It specifies exactly what it means and doesn’t mean. It forces the reader’s model to confront a definite shape — and either that shape fits the existing architecture of their beliefs, or it doesn’t.
When it doesn’t fit, the brain registers prediction error. The more precise the argument, the more exactly it maps the contours of where the reader’s model breaks. The objection that follows isn’t a response to your argument. It’s the sound of a model attempting to defend itself.
This is why rigorous writing generates more controversy than sloppy writing. Rigorous writing is specific enough to actually collide with something.
The Argument They’re Having With Themselves
Now we arrive at the heart of it — the insight the original author glimpsed but left unnamed.
When a reader objects to your work, reconstructs a position you never took, or argues with conviction against a case they built themselves, they are not, in any meaningful sense, engaging with you. They are engaged in something more intimate and more urgent: they are experiencing cognitive dissonance.
Cognitive dissonance — Leon Festinger’s foundational concept, still one of the most robust findings in social psychology — is the discomfort of holding two conflicting cognitions simultaneously. In the reading context: the reader’s existing model on one side, your argument pressing against it on the other. The discomfort is real and physical. It demands resolution.
The brain has two options. It can update the model — accept the prediction error as valid, revise beliefs accordingly. Or it can reject the signal — declare the incoming information invalid, defend the prior, and discharge the tension by externalizing it.
Externalization requires a target. You are the nearest available one.
The objection the reader voices is not about your argument. It is the brain’s attempt to resolve an internal conflict by attributing the source of discomfort to something outside itself. The argument they’re actually having is with themselves. You are just the justification for it.
This is not bad faith. It is the ordinary operation of a mind under pressure.
Why Strong Priors Are Hard to Move
The Bayesian brain framework gives us even more precision here. In Bayesian terms, the brain doesn’t weight all prediction errors equally. It weights them against the precision of the prior — the confidence the brain assigns to its existing model.
A high-precision prior is a strongly held belief. It has been reinforced repeatedly, confirmed by experience, woven into a larger structure of related beliefs that would also need to shift if this one shifted. The brain, quite rationally, requires a much larger prediction error to update a high-precision prior than a low-confidence one.
This is why a reader who has spent years building a professional identity around a particular framework will push back harder against a challenge to that framework than a reader encountering the domain for the first time. It’s not stubbornness in any moral sense. It’s a correctly calibrated resistance to high-cost belief revision.
What looks like refusing to engage with the argument is actually the brain running its update algorithm and deciding, based on available evidence and the cost of revision, that the prior wins this round.
The writer experiences this as rejection. It isn’t. It’s the system working as designed.
The Objection Is a Map
Here is where the reframe becomes practically useful — especially for those of us who think for a living, who write to develop ideas rather than merely to broadcast them.
A reader’s objection is not noise. It is diagnostic information. It reveals, with unusual precision, exactly where their prior model diverges from yours — and how strongly that divergence is defended.
If you learn to read objections this way, they become something far more valuable than feedback. They become a map of the belief architecture you’re trying to engage with. The places where readers push back hardest are the places where your thinking most directly confronts a strongly held prior. Those are, almost by definition, the places where the intellectual stakes are highest.
The writer who treats reader responses as data about the reader — not about the argument — isn’t being dismissive. They’re being epistemically accurate. The response tells you something real about the structure of the receiving mind. It just doesn’t tell you much about whether your argument is right.
What This Means for How We Think
This is the point at which the mechanism connects to something larger — the question of why structural thinking is so difficult, and why tools that support it matter.
If the brain processes only prediction error — only the delta between new input and existing model — then the quality of any thinker’s engagement with new ideas is bounded by the quality of their existing model. A rich, flexible, well-articulated model generates finer-grained prediction errors. It notices more. It can update more precisely. A coarse, rigid, or unexamined model generates blunt responses: confirmation or rejection, agreement or outrage, with very little in between.
This is why the serious thinker’s real task isn’t to consume more information. It’s to maintain and develop the model — to keep it explicit enough to be revised, granular enough to register subtle divergences, and structurally coherent enough that an update in one area propagates correctly to related areas.
That is extraordinarily hard to do inside your head alone. Memory is lossy. Connections between ideas degrade. The model you think you hold and the model you actually hold drift apart over time, invisibly.
This is the problem NotesCanvas is built to address — not as a place to store information, but as a place to make your model visible. When your thinking is externalized into a canvas, you can see where your arguments connect and where they don’t. You can see the gaps. You can see what you’ve been assuming without examining. You can watch a new idea arrive and trace, in structural terms, exactly where it fits and where it creates tension.
In predictive processing terms: you can do the update consciously, deliberately, in full view of what’s changing — rather than letting the brain handle it in the dark, where the easiest resolution is often just to reject the signal.
The Consolation, Properly Named
The writer concluded: “The response to my work is data about the reader, not the argument. That’s either consoling or alarming, depending on the day.”
We’d put it differently. It is neither consoling nor alarming. It is clarifying.
The misreading is not a failure of communication. It is communication revealing its true nature. Meaning never transfers intact from one mind to another. It is always reconstructed — assembled from the incoming signal and whatever the reader already carries. The writer can refine the signal to an extraordinary degree, and still cannot control the reconstruction. The text, as the author says, belongs to whoever picks it up.
But here is the consolation that science provides: the reader who objects most loudly is the reader whose model was actually disturbed. They engaged. Their brain registered the prediction error. The discomfort is real, which means the encounter was real.
The reader who agrees too easily may have simply found enough vagueness to project their existing beliefs onto. The reader who pushes back has met something that didn’t fit their world.
That’s not failure. That’s the friction of actual thinking — two minds, two models, and the irreducible gap between them briefly, violently, productively visible.
The argument they’re having is with themselves. But you gave them something worth arguing about.
NotesCanvas is a thinking environment for serious thinkers — a place to build the kind of explicit, navigable model that makes real engagement with hard ideas possible. If you found this piece useful, it was built with exactly that purpose in mind.