What do you think?
Rate this book
The Brain, In Theory
Why engineering and computational analogies are poorly suited to the study of biological cognition
Mainstream theories of the brain are often expressed through engineering concepts—computation, code, control, reverse-engineering, optimization. These theories cast the living organism as a machine and the brain as a computer. The fact that cognition is a biological phenomenon seems merely anecdotal; biology is considered just “implementation.” In The Brain, In Theory, Romain Brette argues that the brain is not a “biological computer” because living organisms are not engineered. Engineering is the use of knowledge to solve technical problems, to build an artifact with a plan. But, Brette reminds us, Darwin’s insight is precisely that evolution is not a case of engineering. Unlike engineering, evolution has no predetermined goals, plans, or knowledge.
Brette reviews the main theoretical frameworks for thinking about the brain, including computation, neural representations, information, and prediction, and finds them poorly suited to the study of biological cognition. He proposes understanding the brain as a self-organized, developing community of living entities rather than an optimized assembly of machine components. With this new perspective, Brette brings life back to the study of the brain and cognition.
Mainstream theories of the brain are often expressed through engineering concepts—computation, code, control, reverse-engineering, optimization. These theories cast the living organism as a machine and the brain as a computer. The fact that cognition is a biological phenomenon seems merely anecdotal; biology is considered just “implementation.” In The Brain, In Theory, Romain Brette argues that the brain is not a “biological computer” because living organisms are not engineered. Engineering is the use of knowledge to solve technical problems, to build an artifact with a plan. But, Brette reminds us, Darwin’s insight is precisely that evolution is not a case of engineering. Unlike engineering, evolution has no predetermined goals, plans, or knowledge.
Brette reviews the main theoretical frameworks for thinking about the brain, including computation, neural representations, information, and prediction, and finds them poorly suited to the study of biological cognition. He proposes understanding the brain as a self-organized, developing community of living entities rather than an optimized assembly of machine components. With this new perspective, Brette brings life back to the study of the brain and cognition.
304 pages, Hardcover
Published April 7, 2026
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 4 of 4 reviews
July 28, 2026
I can't claim to understand all the arguments Romain makes, as some are quite technical -- related to or involving concepts in information theory, computer science, engineering, and neuroscience, but the ones I do understand are compelling, and I got the sense of someone familiar in all of these areas, widely read, and critical.
It's important to have been 'steeped' in a particular view of problems in order to then understand where the (often implicit) assumptions are made and Romain had, as I think he himself says, never questioned whether neurons compute, which shows just how deep these assumptions in the fields of neuroscience run.
As a reviewer on the back of the book says, Romain is like the Wittgenstein of neuroscience, and he is, because he's basically questioning the assumptions many competing neuroscience views all share. Brains compute, but how? Brain waves? Neurons? Brain regions? This book is like applied philosophy of science to this field.
Finally, I can tell that this book is very much needed in this area, as the online discourse around it from academics working in the area is heated and the resistance tells me that this work is at the very least making those who oppose it articulate why and make explicit the assumptions they make in their thinking.
It's important to have been 'steeped' in a particular view of problems in order to then understand where the (often implicit) assumptions are made and Romain had, as I think he himself says, never questioned whether neurons compute, which shows just how deep these assumptions in the fields of neuroscience run.
As a reviewer on the back of the book says, Romain is like the Wittgenstein of neuroscience, and he is, because he's basically questioning the assumptions many competing neuroscience views all share. Brains compute, but how? Brain waves? Neurons? Brain regions? This book is like applied philosophy of science to this field.
Finally, I can tell that this book is very much needed in this area, as the online discourse around it from academics working in the area is heated and the resistance tells me that this work is at the very least making those who oppose it articulate why and make explicit the assumptions they make in their thinking.
August 11, 2026
Are humans machines? Is the brain a computer? How is the mind programmed?
In our age of AI these questions are commonplace and seem reasonable enough. This owes to the fact that neuroscience and cognitive are built on the assumption that, yes indeed, brains are computers.
Romain Brette attempts here to debunk this assumption. He does so posited from a vantage point that is within the science. Being a physics trained neuroscientist, he demonstrates, like a wolf in sheeps clothing, so to speak, how computational neuroscientists are wrong in using notions such as encoding, computation, information, etc. to explain the brain. Brette sweeps over these concepts and leaves no survivors, neither literal nor metaphorical interpretations of the computational terminology are left standing
Brette makes it clear that the confusion stems from attempting to reverse-engineer the brain, when it wasnt engineered in the first place. He looks at the biology of various living systems and points out that they are nothing like machines, and the analogies to artifactual machines that are used in the cognitive neurosciences are false.
Given this amount of disanalogies, why is the computational perspective still so prevalent? There are certainly many reasons, but one is the transparent utility in thinking of brains as computers. If the brain is a computer and the mind is the software, then the silicon valley techbros have some hope of realizing their dream of "ethical" slavery. If however, we choose to see brains not as computers, but as collectives, as Brette would have it, then what utility does that bring? I wish he shared his thoughts and dreams regarding this.
Now, while Brette does provide an alternative way of understanding cognitive agents, namely through their anticipating, he does indeed not elaborate on the utility of such a view. While elevating your intellect through an accurate understanding of nature is a goal common to all scientists, I am afraid that Brette will hit a wall regarding adoption of his ideas, especially from the computer science adjacent world. Then again, maybe that is Brette's goal here, to make the line in the sand between the study of the living and the study of machines discernible. From the computer scientist point of view this does sound nice. To be able to study models of computation genuinely without having to refer to the messiness of the biology. Maybe deep learning lecturers finally can delete their one slide referring to the biological plausibility of neural networks. And for the biology side, ditching computation can finally allow for concepts such as action, goal-directedness, and subjectivity to enter the scene of cognition, topics which are serious challenges.
I wish this was a laymans book, because it is certainly important in the age of AI, but it is not, it is too difficult. As a cognitive scientist you have to bend your mind against everything youre taught to know, but even if youre untrained in science, the metaphors of engineering are so commonplace in society that seeing through them is a challenge.
Overall, this is a great book! If you arent convinced of "embodied" perspectives, this book will challenge you. And if Brette doesnt convince you in the end, you will certainly pick up facts and curiosities about life that you didnt consider before. Regardless which way you lean, science is better off when conceptual unclarities are brought to the front.
In our age of AI these questions are commonplace and seem reasonable enough. This owes to the fact that neuroscience and cognitive are built on the assumption that, yes indeed, brains are computers.
Romain Brette attempts here to debunk this assumption. He does so posited from a vantage point that is within the science. Being a physics trained neuroscientist, he demonstrates, like a wolf in sheeps clothing, so to speak, how computational neuroscientists are wrong in using notions such as encoding, computation, information, etc. to explain the brain. Brette sweeps over these concepts and leaves no survivors, neither literal nor metaphorical interpretations of the computational terminology are left standing
Brette makes it clear that the confusion stems from attempting to reverse-engineer the brain, when it wasnt engineered in the first place. He looks at the biology of various living systems and points out that they are nothing like machines, and the analogies to artifactual machines that are used in the cognitive neurosciences are false.
Given this amount of disanalogies, why is the computational perspective still so prevalent? There are certainly many reasons, but one is the transparent utility in thinking of brains as computers. If the brain is a computer and the mind is the software, then the silicon valley techbros have some hope of realizing their dream of "ethical" slavery. If however, we choose to see brains not as computers, but as collectives, as Brette would have it, then what utility does that bring? I wish he shared his thoughts and dreams regarding this.
Now, while Brette does provide an alternative way of understanding cognitive agents, namely through their anticipating, he does indeed not elaborate on the utility of such a view. While elevating your intellect through an accurate understanding of nature is a goal common to all scientists, I am afraid that Brette will hit a wall regarding adoption of his ideas, especially from the computer science adjacent world. Then again, maybe that is Brette's goal here, to make the line in the sand between the study of the living and the study of machines discernible. From the computer scientist point of view this does sound nice. To be able to study models of computation genuinely without having to refer to the messiness of the biology. Maybe deep learning lecturers finally can delete their one slide referring to the biological plausibility of neural networks. And for the biology side, ditching computation can finally allow for concepts such as action, goal-directedness, and subjectivity to enter the scene of cognition, topics which are serious challenges.
I wish this was a laymans book, because it is certainly important in the age of AI, but it is not, it is too difficult. As a cognitive scientist you have to bend your mind against everything youre taught to know, but even if youre untrained in science, the metaphors of engineering are so commonplace in society that seeing through them is a challenge.
Overall, this is a great book! If you arent convinced of "embodied" perspectives, this book will challenge you. And if Brette doesnt convince you in the end, you will certainly pick up facts and curiosities about life that you didnt consider before. Regardless which way you lean, science is better off when conceptual unclarities are brought to the front.
August 30, 2026
A book about how we understand the brain. It is mostly a critique, based on careful argumentation, of the view that the brain executes computations and the conceptualizations that follow from that, such as that the brain represents things and that it processes information. The author puts forward an alternative view that is based on embodiment and interaction with the environment, although the development of this alternative view is not the main focus. This book was highly insightful for me on several levels. It made me see how the computer metaphor is extremely pervasive - it permeates our language on how we talk about the brain to a much greater extent than I was aware of. It also made me see that many of the ideas we have are essentially homuncular, and with that kind of dualist/magical. It is amazing to me how this crops up again and again, and that it is so hard to realize that the main ways we talk about this subject are highly problematic. I will probably have to sit with these ideas longer and turn them over to appreciate what it changes exactly for my own views (I'm not a neuroscientist but I do work in a somewhat related field), but it sure made me see things differently. As I said, the development of an alternative view is not the focus of this book - it only gives the start of the beginning of how to see things differently. On the one hand I had hoped for more in that respect but on the other it is exciting that so much remains to be done from a theory point of view. The book could probably have been a bit more concise, there is quite some repetition, but in some ways this is also useful to get the problems with the current framework across as clearly as possible (defusing counterarguments from several perspectives etc). It is not necessarily easy to read - the subject just requires careful writing and the discussing of deep subject matter. A radical book that I suspect will be on my mind for a while.
September 13, 2026
Fantastic!
Displaying 1 - 4 of 4 reviews