The Body That Thinks

On consciousness, the world, and the question of AI

We say that we have a body.

As though the body were something we owned.

We have a house, a book, a pair of glasses — and a body. Perhaps our language betrays an old assumption: somewhere inside this body there is the real human being. A self. A consciousness. A thought. The body is the container in which all of this resides.

Thought always has a body.

But perhaps the problem begins right there.

In Nature on 27 August 2026, cognitive scientist Abeba Birhane reviewed Anthony Chemero’s new book Intertwined Creatures: The Embodied Cognitive Science of Self and Other. She begins with a question that has gradually moved from the margins of philosophy into public debate: Can artificial intelligence ever become conscious?

Yet her title also contains a warning. Perhaps the question itself rests on a misconception.

And the misconception concerns not the machine first of all.

It concerns the human being.

Birhane describes a long Western tradition in which the mind is conceived as something bounded and internal: a solitary, almost disembodied observer looking out upon the world. Cognition can then easily be understood as the processing of representations inside the head.

Chemero’s book challenges this picture, at least as Birhane presents it. But before going further, I need to make one qualification: I am reading Chemero here through two reviews, not through the two books themselves. One concerns his book from 2009, the other his new book from 2026. We should therefore not simply treat them as expressions of exactly the same project.

In fact, the difference between them is interesting.

Two different radicalizations?

In Radical Embodied Cognitive Science from 2009, Chemero’s project, as Erik Myin describes it, was explicitly directed against representationalism in cognitive science. He wanted to investigate whether perception, cognition and action could be explained as necessarily embodied phenomena without postulating mental representations as the decisive intermediary.

That is a fairly radical philosophy-of-science thesis.

It is not merely the claim that the body influences the brain, or that sensation and movement contribute to thought. It raises the question of whether we need, at all, the model in which the world must first be represented inside the organism before it can acquire meaning.

In Birhane’s review of Intertwined Creatures, seventeen years later, the emphasis is elsewhere. Here she highlights the embodied, dynamic and social human being, but also relationships, recognition, power and the social conditions in which human beings live. Among other things, she points to Chemero’s use of feminist political theory and to the way gender, skin colour, disability, class and other social positions affect the possibilities people encounter in the world.

This does not necessarily mean that Chemero has abandoned his earlier project. But the reviews do not give me sufficient grounds for claiming that the argument in 2026 is simply the same argument as in 2009.

Perhaps what we see instead is a movement.

From the question of how the organism relates to its environment, to the question of how the human being is also intertwined with other people and the structures of society.

From body and environment to body, environment and the other.

That makes the title Intertwined Creatures particularly suggestive.

Intertwined creatures.

Heidegger’s world

Here I find it difficult not to think of Heidegger.

His In-der-Welt-sein — being-in-the-world — was an attempt to break with the idea of the human being as an independent subject that first exists in itself and only afterwards has to establish a connection with the world.

We are always already in it.

When I use a hammer, I do not normally regard the hammer as an object with certain physical properties. It belongs to an activity together with the hand, the nail, the plank and what I am trying to accomplish.

Only when the hammer breaks does it distinctly emerge as an object.

Something similar applies to the body.

As long as it does what we expect of it, it can almost disappear from our attention. We walk without calculating our legs. We grasp the coffee cup without working out the distance between cup and hand.

But the body can make itself felt.

Through pain, dizziness, fatigue, illness or age.

Then we discover something that was true all along: we never stood outside it.

What the world affords

In Chemero, this takes a more cognitive-scientific form through the concept of affordance, taken from the ecological psychology of James J. Gibson.

An affordance can, in simplified terms, be understood as a possibility for action that exists in the relation between a living being and its surroundings.

A staircase is not merely height, width and material. For one person it is something easily climbed. For a small child it may be something to crawl up. For someone with severely reduced mobility it may be a barrier.

Myin emphasizes that affordances in Chemero are precisely relational: they arise between features of the environment and the capacities of the organism.

This changes something in our understanding of perception.

The world need not first be a collection of meaningless sensory impressions to which the brain subsequently adds meaning. It already appears as a world in which action is possible.

A cup can be grasped.

A door can be opened.

A railing can provide support.

A road can be walked.

This brings Chemero close to a thought I associate with Merleau-Ponty.

I never see from nowhere

For Merleau-Ponty, the body is not merely one object among other objects.

It is the way the world opens itself to me.

I always see from somewhere. I hear from somewhere. I move from somewhere. I have no access to the world from a neutral vantage point outside my own body.

There is no view from nowhere.

The body therefore becomes more than an instrument at the disposal of consciousness. It is a condition for there being a world for me at all.

That also makes the sentence “I have a body” increasingly difficult.

I do not first possess a completed self and then become equipped with a body.

I am bodily present.

But where does thought take place?

Here Chemero’s earlier work becomes genuinely challenging.

Myin describes how radical embodied cognitive science attempts to locate meaning in the relation between the organism and an information-rich world, rather than moving it inside the head. Toward the end of Chemero’s 2009 book, this leads to the idea of consciousness as a relation between organism and environment — in a certain sense located both “in the head and in the world.”

It is tempting to follow him a long way.

But resistance is needed here as well.

Myin provides some himself. He suggests that Chemero may be too quick to dismiss Andy Clark’s point about the significance of external symbols. Experiments with chimpanzees indicate that interaction with symbols can expand the kinds of cognitive tasks they are capable of performing. Myin’s point is that this does not necessarily force us back into a classical theory of internal representations. The symbols themselves may alter the organism’s future possibilities for interacting with the world.

That is an interesting objection.

Perhaps the choice is not between “everything takes place inside the head” and “nothing does.”

Language itself should make us cautious.

A written word exists outside the body. Yet it can change what I think. A map can alter where I go. A mathematical formula makes it possible to think thoughts I might hardly be able to sustain without the symbols before me.

The external world of human beings may itself be part of their cognitive life.

Thinking in the kitchen

In Birhane’s review of the new book, this idea becomes almost comically concrete.

You are on your way to work and looking for your keys. Then you see the laundry and remember that you need detergent. While doing something else, a book reminds you that you meant to take it with you. An object in your bag prompts yet another action — and in the end you leave without the very book you intended to bring.

Where did the thought take place?

Of course something happened in the brain.

But the room did something too.

The objects did something.

They reminded, interrupted, attracted and redirected attention.

We do not merely think about the world. In many ways, we think with it.

Perhaps that is why a notebook can be more than a storage place for thoughts I have already had. When I write, I see something I did not see before I wrote it. The sentence in front of me answers back. I cross out a word and discover that the thought itself has changed.

Where, then, does thought begin, and where does it end?

The intertwined human being

Here the new Chemero book seems to move the question further.

Birhane describes the human mind as embodied, dynamic and social, inseparably connected to its surroundings and to other people. She emphasizes that what matters is not only the relation between brain and body, but that this brain and this body live in a particular environment.

And this environment is not socially neutral.

How others see us, which identities are recognized, which abilities are attributed to us, and which possibilities society opens or closes can become part of the world in which we are actually able to act.

Affordance then begins to acquire something like a social dimension.

Not every door is equally open to everyone, even when it is physically ajar.

The human being is not only embodied.

The human being is dependent.

We are born dependent on others, learn language from others and become human in a world already inhabited, interpreted and shaped.

Even our most private thoughts bear traces of other people.

And then AI returns

We can now return to the question with which the Nature review began:

Can artificial intelligence become conscious?

Birhane answers rather firmly on the basis of her reading of Chemero. Today’s AI systems do not inhabit worlds, do not enter into relationships in the way human beings do, and do not have feelings. They are human-made artefacts shaped by human labour and culture.

This is an important corrective to the idea that linguistic fluency alone could establish consciousness.

But I am not certain that the question is therefore settled.

What happens if AI one day is no longer merely a language model running on a server?

Imagine a machine able to move through the world. It has vision, hearing and touch. It must navigate obstacles, learn what it can grasp, what it can cross and what it ought to avoid. It depends on energy, maintenance and other agents in order to continue functioning. Its actions have consequences for the possibilities available to it later.

Would it then begin to have something resembling affordances?

Would a staircase be something different for it than it is for us, yet still something it could encounter as possible or impossible to climb?

If Chemero’s radical thesis really says that cognition arises in the dynamic coupling between organism and environment, then the question must at least be asked.

But another question immediately follows.

Is any physical connection to the world enough for us to call something embodiment?

A temperature sensor is not cold.

A damage sensor is not pain.

A low battery is not necessarily hunger.

And being capable of being switched off is not automatically the same as knowing that one will die.

Here I sense a difference between having sensors and being vulnerable.

The living body is not merely an apparatus for gathering information. It needs the world in order to remain alive. It can be wounded. It ages. It cannot simply be replaced without raising the question of who, if anyone, continues.

Is this vulnerability itself a condition for the kind of consciousness we know?

I do not know.

Nor do the two reviews provide grounds for a definite answer.

But perhaps this is a better question than whether a language model “really thinks.”

The body we forgot

Philosophy has been extraordinarily preoccupied with thought. That is understandable. Philosophy works with concepts, sentences and arguments.

Yet every philosophical argument has had a bodily history.

Descartes doubted while sitting, walking, eating and sleeping. Kant wrote at a desk. Heidegger walked. Merleau-Ponty tried precisely to make visible the bodily character of our access to the world.

None of them thought from a point outside life.

Perhaps, then, the most important lesson from Chemero is not that “the body thinks.” That formulation, too, may be too simple.

The more radical claim may be this:

Thought does not belong to an isolated place.

It arises in the relation between a living human being and a world, the things within it, and other people.

Then the question of AI also becomes less simple than I first thought.

Before we ask whether the machine can become like the human being, we need to know what kind of creature the human being is.

Not merely a brain.

Not merely a body.

But a living, vulnerable and dependent being that is always already in a world.

And perhaps that is precisely why the body has remained so philosophically under-communicated for so long.

It has always been so close to us that we have barely seen it.