Why Uncertainty Feels Like Anxiety: The Entropy Model

Most people describe anxiety as a feeling that arrives without a reason. It shows up before the meeting, or at three in the morning, or in the middle of an ordinary afternoon, and it doesn't seem attached to anything.

There's a framework in psychology that suggests it is attached to something, and that the something is more specific than we usually assume. Not danger. Uncertainty.

The paper is Psychological Entropy: A Framework for Understanding Uncertainty-Related Anxiety, published in Psychological Review in 2012 by Jacob Hirsh, Raymond Mar, and Jordan Peterson. They call their framework the entropy model of uncertainty, or EMU. It has been cited several hundred times and sits underneath a good deal of the predictive-processing work I've written about elsewhere.

The core claim is compact enough to state in a sentence. Uncertainty is experienced subjectively as anxiety.

Not caused by. Experienced as. That distinction is the whole argument.

What Entropy Means Here

In physics and information theory, entropy measures disorder and unpredictability in a system. The more possible states a system could be in, the higher its entropy.

Applied to a mind, entropy is the number of live possibilities you're holding at once. How many things this situation could mean. How many ways it could go. How many responses are available and how uncertain you are about which one works.

Hirsh and his colleagues propose that uncertainty arises specifically from conflict between competing affordances. An affordance is a possible action or interpretation the environment seems to offer. When your available readings of a situation, or your available responses to it, are pulling against each other, entropy rises. When one clear reading and one clear action are available, it falls.

This is why an ambiguous email from your boss can generate more distress than a difficult but clear one. The difficult message is unpleasant and low-entropy. The ambiguous one contains four possible meanings, all live, none resolvable, and the system has to hold them all.

Anxiety as a Signal, Not a Malfunction

The claim that anxiety is the felt quality of unresolved uncertainty reframes what it's doing.

Anxiety, on this account, is not a system failure. It's a readout. It tells you that your internal model isn't currently able to predict what's happening or what to do about it. That's information. In an organism that has to act in an uncertain world, a signal that says your model is failing here is enormously useful.

This matches what I see clinically more often than the alternative. The clients who come in most anxious are frequently not the ones facing the worst circumstances. They're the ones facing the least resolved ones. A person with a bad diagnosis and a treatment plan is often steadier than a person waiting three weeks for test results. The waiting is higher entropy. Nothing bad has happened yet, and that's precisely the problem.

It also fits a pattern I've written about separately, where anxiety in the present turns out to be about something unresolved from before. An old situation that never got a clear reading stays open. The model never closed. The entropy never came down.

Why We Build Beliefs

The third piece of the framework explains something about human behavior that can otherwise look irrational.

Hirsh and colleagues argue that clear goals and belief structures reduce the felt experience of uncertainty by narrowing the spread of competing possibilities. A belief system is, functionally, a filter. It tells you what things mean and what to do. It cuts the number of live options.

This is genuinely useful. It's also the mechanism behind some of our worst tendencies. A rigid ideology is extremely effective entropy management. So is a conspiracy theory. So is a fixed and punishing story about who you are, which is often what I'm actually working with in the room. I'm the kind of person who ruins things is a painful belief and a stable one. It reduces uncertainty. It answers the question in advance.

Which is part of why insight alone rarely dislodges these stories, and why beliefs that hurt us can be so much harder to give up than they should be. They're doing a job.

Why Change Feels Like Getting Worse

Here's the part I find most clinically useful.

If beliefs reduce entropy, then changing a belief must raise it. Temporarily, and sometimes sharply.

The moment you begin to question a long-held story about yourself, you don't immediately get a better story. You get an absence. The old model is loosening and no new one has consolidated. Possibilities that had been closed reopen. Entropy climbs. And since entropy is experienced as anxiety, the felt experience of beginning to change is increased anxiety.

This is why people so often quit therapy right when it starts working. It's why the week after a real insight is frequently worse than the week before. The client isn't regressing. They're in the high-entropy gap between an old model that's failing and a new one that hasn't formed. It's the same territory as why the brain prefers familiar suffering to unfamiliar healing.

Naming this in advance changes how people move through it. When someone knows that the discomfort is the sound of a model reorganizing rather than evidence that things are getting worse, they can tolerate considerably more of it.

The Balance Point

The framework implies a range rather than a target.

Too much entropy and you can't function. Nothing is stable enough to act on. Too little and you're rigid: certain, protected, and unable to update when the world changes. Neither end is healthy, and the work is usually about moving toward the middle from whichever side someone is stuck on.

I want to be careful here, because there's a related idea that gets conflated with this one. The entropic brain hypothesis, which comes out of Robin Carhart-Harris's work rather than this paper, maps psychiatric conditions onto entropy levels directly, putting depression and rigid rumination at the low end and psychosis at the high end. It's a productive idea and I've written about where healing lives on that spectrum. But it's a different framework from a different literature, and the 2012 entropy model doesn't propose that taxonomy. Worth keeping them separate.

Where This Connects

The entropy model sits inside a larger body of work. Hirsh and colleagues build explicitly on Karl Friston, whose free energy principle describes the brain as a system that minimizes surprise by continually updating its model of the world. The entropy model is, in a sense, the affective side of that story. Friston tells you what the system is doing. Hirsh tells you what it feels like from the inside when it isn't working.

The same framework explains why meaning matters in a way that isn't merely philosophical. Meaning organizes perception and guides action. It tells you what things are and what to do about them. A person without any framework for making sense of their life is holding an enormous amount of unresolved possibility, which is one reason the meaning-making dimension of suffering shows up as anxiety so reliably.

And entropy isn't only something we flee. Rising uncertainty also drives attention, curiosity, and exploration. The same signal that makes you anxious makes you look. Whether it becomes dread or interest depends a great deal on whether you have the resources to meet it.

What This Changes

If anxiety is the felt quality of an unresolved model, then the treatment question shifts.

The goal isn't to eliminate anxiety. A person with no anxiety would be a person whose model was never challenged, which would mean a person who had stopped learning. The goal is to build a system that can tolerate more uncertainty without collapsing, and that can update without falling apart while it does.

That's a different project from symptom reduction. It's slower. It looks less like removing something and more like widening a range. But it's closer to what actually seems to help.

If you'd like to explore what that kind of work could look like for you, you're welcome to book a consultation.

Reference Hirsh, J. B., Mar, R. A., & Peterson, J. B. (2012). Psychological entropy: A framework for understanding uncertainty-related anxiety. Psychological Review, 119(2), 304–320.