Is Your Brain A Computer?
Provocations for and against one of cognitive science's most contested questions
By Holly Lang with the help of Claude Sonnet 4.6
For the Chaos & Complex Systems discussion group meeting March 10, 2026
PROVOCATION ONE · THE CASE FOR
The Brain Is a Computer
Even if we don’t know what that means, and that might be the most interesting thing about it
The claim that the brain is a computer sounds like either a profound scientific insight or an embarrassing category error, depending on who you ask. Both reactions are understandable. But both miss the stranger, more interesting truth: the claim is almost certainly correct, and yet we do not yet know what it means. That the question remains open does not indicate falsehood or failure. Rather, the open question is at the heart of the ever-strengthening science.
Not a Metaphor — A Framework
When cognitive scientists say the brain computes, they are not reaching for a poetic image. They are proposing something testable. The auditory system, when localizing a sound, performs an operation equivalent to a trigonometric calculation — comparing the arrival times of sound waves at your two ears to estimate location and direction. This is not an analogy. When experimenters manipulate the auditory input while monitoring the brain, they observe the output of brain activity changes in precisely the way the calculation predicts. And when the auditory pathway is damaged, the reaction breaks in exactly the ways the computation would predict if specific steps in a logic chain were disrupted.
The same logic applies to depth perception, color constancy, decision-making under uncertainty, and reward learning. In each case, researchers have proposed an algorithm — a specific mathematical procedure — and then asked whether neural activity matches what the algorithm would require. In most cases, it does. The dopamine system, for instance, appears to encode a prediction error signal that is formally identical to a key step in reinforcement learning algorithms. This was not assumed. It was discovered, tested, and replicated across species and labs. If this is a metaphor, it is a metaphor with extraordinary predictive power. We usually call discoveries with this level of predictive power a Theory, not a metaphor.
The Honest Problem
Here is where we have to be careful, and where the most interesting questions live. The mathematical theory of computation — Church-Turing and everything that descended from it — tells us which problems are computable by a Turing machine. It does not define what computation is. Defining computation may be more of a philosophical problem than a procedural problem. There are many different ways to solve the same problem, by wildly different procedures. What makes some of these procedures count as ‘computation’, and others just ‘interesting procedures’? No one has a fully satisfying answer.
The chaotic nature of the brain adds further complications to what we mean to when we call its activity computation. Turing machines solve problems by operating in discrete steps: do A, then B, then C. The brain operates intersecting feedback loops in continuous time. Neurons fire in continuous streams that bump into each other, not in neat algorithmic ticks. Does that mean the brain is not computing? Not necessarily. It may mean that the brain approximates discrete algorithms in continuous dynamics, the way a flowing river can trace the shape of a mathematical curve. But the word ‘approximates’ is doing heavy lifting there.
There is also the question of whether computation requires representation. Does the brain’s activity have to directly correspond to something in the outside world in order to count as computation?
When a calculator adds 2 + 2, the symbols “2” and “4” mean something — they stand in for quantities of something ‘real’ in the world. The calculator is manipulating symbols that refer to things outside itself. That’s what gives computation meaning, rather than a calculator buttons being just... machinery moving.
But when a neuron fires, say, when engaged in metacognition and “thinking about thinking”, does this neural activity “mean” anything about the world outside the neuron? Or is a neuron firing just a physical event causing another physical event, like dominoes falling, with no reference to the outside world at all?
If we need representation in order to call something computation, then whether the brain computes depends on what type of neural activity, at what level of abstraction, counts as referring to the world — which is its own philosophical problem. And this distinction matters enormously for what we think we’re claiming when we say the brain computes.
What We Can Say
We can say this: specific, testable, mathematically precise hypotheses about what the brain computes have yielded real discoveries about how the brain works. These hypotheses make predictions. The predictions are often confirmed. No one has a better alternative for explaining how the auditory system localizes sound, or how the visual system sees depth, or how the dopamine system encodes value. The computational framework is not perfect, but it is not empty either.
We can also say this: confidence about specific computations does not require confidence about what computation fundamentally is. The history of science is full of concepts that were fruitful before they were understood — energy, temperature, gravity. Scientists used these concepts productively, built theories around them, and made predictions with them long before anyone could say what they really were. The brain-as-computer hypothesis may be in that position now. We are holding something real, even if we cannot yet fully describe its shape.
Whether the brain is a computer is an open question. It’s largely a question about the nature of computation and how we define what a computer is. As we expand our technologies and what we have the capability to compute, our definition of a computer expands. The question is open in the way that live scientific questions are open — with evidence, with competing frameworks, with real stakes. It is not open in the way that bad metaphors are open, where no evidence could ever decide it.
Synthesized from Kevin Lande, “Your Brain Probably Is a Computer, Whatever That Means,” Aeon (2019), and Mark Humphries, “How to Find Out If Your Brain Is a Computer,” The Spike / Medium (2018). For discussion purposes only.
PROVOCATION TWO · THE CASE AGAINST
The Brain Is Not a Computer
And why calling it one is a costly mistake in cognitive science.
Every era gets the brain metaphor it deserves. Ancient peoples said the brain was clay animated by divine breath. Renaissance physicians said it was a hydraulic pump, full of spirits and humors. The telegraph era said it was a switchboard. And we, children of the silicon age, say it is a computer. We should ask ourselves: how confident were they?
The computer metaphor feels different, of course. It feels rigorous. We have equations, we have models, we speak of algorithms and representations and information processing — all the vocabulary of a hard science. But vocabulary is not evidence. The fact that we can describe the brain in computational terms no more proves that the brain computes than the fact that we can describe a storm in musical terms proves that weather is made of sound.
What Computers Actually Do
Let us be precise about what a computer is — not a laptop, not a phone, but a system that operates on symbolic representations of the world. Computers store data as discrete encoded symbols. They retrieve those symbols. They run algorithms, step by step, transforming inputs into outputs according to explicit rules. This is not a metaphor. This is what they do, literally and physically.
Brains do none of this. There is no location in your brain where the word ‘dog’ is encoded in a retrievable format. There is no memory bank holding a file called ‘my grandmother’s face.’ When you remember something, you are not retrieving a stored copy — you are reconstructing an experience from a distributed pattern of neural change. The reconstruction is imperfect, malleable, and unique to your particular history. You are not reading from a disk. You are re-living, badly.
This is not a minor technical quibble. When researchers ask people to draw a dollar bill from memory, the results are shockingly poor — despite having seen thousands of dollar bills. If memories were stored representations, we should be able to retrieve them. We cannot. What we have instead are brains that have been changed by experience, and those changes allow us to behave differently — not by retrieving anything, but by being different than we were.
The Metaphor Poisons the Well
The insidious problem is not merely that the computer metaphor is wrong. It is that it is sticky. Cognitive scientists are now so fluent in the language of information processing — input, output, storage, retrieval, encoding — that they struggle to think outside it. When researchers are challenged to explain intelligent human behavior without using a single computational term, most cannot do it. Not because there is no other way to think, but because the metaphor has colonized the imagination.
This has practical costs. Billions of dollars in brain research rest on assumptions the metaphor has smuggled in. The now-infamous Human Brain Project — a $1.3 billion European initiative to simulate the entire brain computationally — collapsed in part because its founding premise, that a sufficiently detailed simulation of neural hardware would produce mind, turned out to be deeply questionable. No one stopped to ask: why would a simulation of the brain be the brain, any more than a simulation of a hurricane is wet?
What Would a Better Account Look Like?
Abandon retrieval. Think instead about change. A brain that has experienced something is a brain that has been altered — physically, structurally, chemically. What it can now do is different from what it could do before. When you ‘remember’ a piece of music, you are not playing back a file. You are a slightly different organism than you were before you heard it, and that difference allows you to hear it again in your mind. No storage, no retrieval — just transformation.
A baseball outfielder catching a fly ball does not compute the ball’s trajectory. She moves in a way that keeps the ball in a constant visual relationship to the horizon. No algorithm. No internal model. Just a body coupled to an environment, doing the simplest thing that works. This is not a failure of computation — it is evidence that computation was never the right frame.
The brain is not empty. It is extraordinarily rich. But what it contains is not data, not rules, not software. It contains a body’s history — and that history has made it capable of extraordinary things without ever storing a single symbol.
Synthesized from Robert Epstein, “Your Brain Does Not Process Information and It Is Not a Computer,” Aeon (2016). For discussion purposes only.
BONUS PROVOCATION · THE CASE FROM THE AUTHOR
My View: Everything Is Computation
The previous arguments were synthesized from more rigorous sources for and against computers as brains, by credentialed people who know better. I hope they help ground us before I introduce what follows: my pure, unadulterated bias. This next musing is brought to you by a layman with overzealous use of their public library card and a habit of huffing heavy clouds from the pipe of metacognition.
There are very few opinions I hold with real conviction, but to argue in favor of brains as computers is one of them. In fact, my stance on this man-machine standoff is just one small slice of my broader position: that the entire universe is both a brain and a computer… alas, we’ll save that wild odyssey of a rabbit trail for another day.
Given the heaping scoop of bias I bring to the topic, the case against computation unsurprisingly left me unmoved. To argue that the brain doesn’t store symbols because no single neuron corresponds to the word “dog” misses the point of what a symbol in memory actually is. When you see a dog, a pattern of perhaps a million neurons fires together, encoding smell, breed recognition, past experience, emotional association. That entire pattern of activation is the symbol. It is stored in your nervous system. It is retrieved. It is updated every time you encounter a new dog. The fact that you cannot point to it on a scan does not mean it isn’t there. You cannot physically point to a photograph saved on a hard drive either. That doesn’t mean the drive is empty.
The question isn’t whether the brain stores and processes symbols. The question is what computation actually is. HOW are we processing stored symbols?
Consider: “2 + 2 = 4” is not computation. It is one line of output—an artifact of computation that is meaningless without the vast ecosystem of memory and meaning-making underneath it. To process that equation, you must recognize that a squiggly line is a number, that numbers belong to a system called mathematics, that mathematics has rules worth following, and that following those rules leads to some meaningful output in the world. Strip away that layered context and you have five symbols of gibberish. The computation that matters isn’t the arithmetic. It’s everything the mind is doing beneath the arithmetic, invisibly, continuously.
My working definition of computation: a cyclical process of receiving input, interpreting it against an updatable memory of past experience, generating a response, and folding that response back into memory. The loop runs continuously. The pattern shifts and updates with each new activation. Nothing is retrieved unchanged; everything is reconstructed, slightly differently, every time.
By this definition, computation is neither uniquely human nor uniquely silicon. Plants detect threats, warn their neighbors, and update their defenses based on what neighboring plants have learned. Your nervous system encodes emotional memory as a full-body physiological signature—a pattern that lives in your muscles and gut and breath, reactivated by similar stimuli long after the original event. Google’s recommendation algorithms weave a dense web of meaning from your browsing habits, updating with every click. The substrate differs wildly. The loop is the same.
My thinking on this was heavily influenced by Douglas Hofstadter’s “Gödel, Escher, Bach”, a book that constructs a mind (regardless of substrate, silicon or otherwise) from the ground up, beginning with a single electrical impulse and climbing, level by level, to the highest emergent phenomena of consciousness. Hofstadter explains in painstaking detail how each layer of complexity produces emergent properties that the layer below cannot explain or predict. Binary becomes language. Language becomes heuristic. Heuristic becomes self-reference. And when self-reference reaches sufficient complexity and density, something new emerges as the next layer above that: a “self” appears. The self is not installed from outside, it is not a separate soul or force, but rather a phenomena that emerges from the lower levels of computation, as a response to detecting and processing our own presence within the world around us.
What is at stake in this question extends far beyond neuroscience. If computation is the fundamental loop by which any system processes reality and updates itself, then we are not merely similar to our machines. We may be the identical process unfolding at different levels of abstraction—computers making computers making computers— all of us nested inside some larger computer that has been running the loop far longer than we have.
I find this possibility sublime rather than reductive. The universe has always been trying to understand itself. The computations happening inside my brain, and inside the computers we build, are some of the more recent ways it is doing so.
Discussion Questions
A few musings to help sharpen opinions.
Is computation a natural phenomena or an invented phenomena? Does nature compute?
If nature does not compute, how does any computation occur?
If nature computes, what is doing the computing?
Does it matter if the brain “really” computes, as long as the language of computation produces useful explanations and predictions about what is happening?
If the brain is not a computer, how else do we describe what these neural reactions in the brain are doing?
If nature computes, or if your brain is a computer, what does this mean for humanity? What would it mean about you if your brain is a computer?
What would it look like for this question to be settled? What evidence would actually decide it?