Embodied Cognition

In the 2003 movie Freaky Friday, Lindsay Lohan and Jamie Lee Curtis swap bodies but keep their personalities. It's a catchy premise that encodes one of the deepest assumptions of Western philosophy: the self is a thing that rides around inside a body, independent of its vehicle. The assumption is wrong.

The Body Shapes the Self

Pawel Tacikowski's body-swapping experiments at the Karolinska Institute used head-mounted cameras and synchronised touch to make pairs of friends feel they'd switched bodies. The illusion was powerful enough that when the friend's body was threatened with a fake knife, participants flinched as though their own body were in danger. But the really interesting finding was cognitive: while embodying their friend, participants rated their own personalities differently — shifting toward traits they'd previously attributed to the friend. The brain, confronted with a mismatch between what it sees (the friend's body) and what it knows (my personality), resolves the conflict by updating the personality model rather than rejecting the illusion.¹

This isn't an isolated curiosity. The experimental literature on body illusions has been piling up for two decades, and the findings are remarkably consistent: change a person's perceived body, and you change how they think. People who experience an out-of-body illusion through virtual reality report reduced fear of death. People made to feel invisible show less social anxiety in front of a crowd. People embodied in a four-year-old's body through VR start unconsciously associating themselves with child-like traits and perceive the world as larger. White people who inhabit a Black virtual body show reduced implicit bias scores. Domestic violence offenders who embody a female victim get better at recognizing female emotions.¹

The rubber hand illusion — where synchronised touch on a hidden real hand and a visible fake hand causes people to experience the fake hand as their own — is the simplest version. But what makes it philosophically important is that the illusion triggers real physiological changes: the skin temperature of the real hand drops, and the magnitude of the temperature change correlates with the intensity of the illusion. If the fake hand is threatened, the person shows a genuine startle response. The brain doesn't just believe the fake hand is real — it treats it as real in a way that alters the body's homeostatic regulation.¹

Self-Concept as Ongoing Negotiation

The deeper lesson, as Jane Aspell puts it, is that "you can't just study the brain in a vat. The brain exists in the body and in the environment. If you look at it separated from those things, you're only getting half the picture."¹

This is the core claim of embodied cognition: the body is not a peripheral device that executes commands from the brain. It's a constitutive part of cognition itself. The way we think is shaped — not just influenced, but structured — by the body we inhabit. This connects to the extended mind thesis but goes further: Andy Clark argues that cognitive processes can extend beyond the body into tools and environment; embodied cognition argues that even within the organism, cognition is not localised in the brain but distributed across brain-body systems.

Consider what the body-swapping experiments reveal about memory. Tacikowski found that some participants did worse on memory tests during the illusion — but not all of them. The people who were better at embracing their friend's body and personality did better on memory, suggesting that memory is impaired when a person's body clashes with their concept of self. If your representation of self is disturbed, memory encoding degrades. This makes sense if selfhood isn't a static identity but an active, ongoing process of self-modeling that the brain requires in order to organise experience into storable form.

The Historical Neglect

It's remarkable how recently this has entered neuroscience. As Olaf Blanke notes, "science and philosophy have traditionally mostly associated the self with reflective cognitive aspects of the mind." The body was treated as a container — something the self lives in — rather than as part of the self. The R. Baumeister observation that "everywhere in the world, self starts with body" dates to 1999, and it still reads as a provocation in some philosophical circles.¹

The neglect has philosophical roots. Descartes' division of mind and body, the computational metaphor of the brain as a computer running software, the functionalist position that mental states are defined by their causal roles regardless of physical implementation — all of these push toward a disembodied picture of cognition. Even predictive processing, which has done more than any other framework to reconnect brain and body through interoception, started as a theory about information processing in cortical hierarchies.

But the body keeps reasserting itself. Anil Seth's finding that people feel greater ownership of a virtual hand that pulses in sync with their heartbeat. Barrett's constructed emotion theory, in which emotions aren't brain states but whole-body predictions involving interoception, proprioception, and autonomic regulation. The discovery that gut microbiome composition correlates with mood and anxiety. The simple, stubborn fact that exercise is as effective as antidepressants for mild-to-moderate depression. None of this is intelligible if the brain is running a program in a vat. All of it makes sense if the mind is what a body does.

Sensory Substitution: Unused Ports

If the body shapes cognition, what happens when you add entirely new sensory channels? Sunny Bains's survey of sensory substitution research suggests the answer is: the brain just plugs them in.² Udo Wächter wore a belt lined with vibrating pads for six weeks — whichever pad pointed north buzzed. After a few days he had an internal map of his city. Peter König, the belt's inventor, became "intuitively aware of the direction of my home" even while 100 miles away. Wächter felt the vibration moving in his dreams.

The deeper story is Paul Bach-y-Rita's work on "tactile vision." Starting in the 1960s, his team showed that blind people could identify images "seen" through a grid of stimulators on their backs, and later through a mouthpiece with 144 electrodes on the tongue. The tongue turns out to be ideal for sensory input — nerve-dense, touch-receptor-packed, and bathed in saliva as a conductor. When Bach-y-Rita's team connected an accelerometer to the tongue display for people with damaged inner ears, it not only restored their balance but the effects persisted after removing the mouthpiece, sometimes for hours or days. The brain had recalibrated using the new channel.²

The philosophical import is this: the brain doesn't care where sensory data comes from. It cares about statistical structure. Kohler's 1950s experiments showed that people wearing goggles that flipped the world upside down adapted completely after several weeks — their brains rewired to process the inverted images as normal. The brain has what Bains calls "unused sensory ports just waiting for the right plug-ins." This is neuroplasticity applied to the extended mind thesis in its most literal form: extend the body's sensory apparatus, and the brain extends its cognitive map to match.

When the Body Is Missing: Congenital Analgesia

The inverse case is equally revealing. Steven Pete was born with congenital analgesia — the inability to feel pain.³ He chewed off a quarter of his tongue while teething. He broke his leg at a roller-skating rink and didn't notice until people pointed at the blood soaking through his pants. He was taken from his parents by child protective services because the accumulation of injuries looked like abuse. Other children would recruit new students to fight him, saying "if you can't feel pain, you will once I'm done with you."

What makes this relevant to embodied cognition isn't the medical condition but what Pete reports about his relationship to his body: he's not reckless but hypervigilant, because without pain signals he has no interoceptive alarm system. Internal injuries terrify him — appendicitis, especially, because he'd have no warning. The absence of a single sensory channel doesn't just remove a feeling; it restructures the entire self-model. If Seth is right that the bodily self is built from interoceptive predictions, congenital analgesia is a natural experiment in what happens when a major prediction channel goes dark. You don't lose your self, but the self you build is organized around what you can't feel — a constant cognitive compensation for missing data.

Virtual Embodiment and the Self-Model

Thomas Metzinger takes embodied cognition into virtual reality and discovers that the self-model is far more hackable than anyone suspected.⁴ Metzinger had his first out-of-body experience during a meditation retreat at nineteen — floating above himself, hearing his own breathing. As a materialist philosopher of mind, this was deeply unsettling. Years later, he and the neuroscientist Olaf Blanke figured out how to induce similar experiences in a lab. Blanke had found that electrically stimulating specific brain regions could create the sensation of floating above one's body, perceiving a doppelganger, or feeling an invisible presence nearby.

The virtual-embodiment experiments that followed are the body-swapping studies from Tacikowski's work scaled up. Metzinger put on a VR headset and saw his own body, filmed from behind, being stroked. He could feel the stroking, but the body it was happening to seemed to be in front of him. He felt "marooned outside of himself." Mel Slater and Mavi Sanchez-Vives extended this: using tracked VR with a virtual mirror, they could convince subjects that a virtual body was their own in just a few minutes of guided movement. "We have the illusion that our body model is very stable," Sanchez-Vives said, "but that's only because we've never encountered anything else."

What Metzinger concluded from all this is that we don't just live inside a model of the external world — we live inside models of our own bodies, minds, and selves. These self-models can be adjusted in ways that seem logically impossible, like portraying a self that exists outside the body. The therapeutic implications are already being explored: virtual embodiment for chronic pain, body dysmorphia, and trauma. But the deeper philosophical lesson is the one that connects to the rubber hand illusion: the self is not a thing inside the brain. It's a model the brain runs — and like any model, it can be updated, extended, or relocated.⁴

The Proteus Effect

Perhaps the most practically significant implication is the "Proteus effect": people's behaviours and thinking patterns are influenced by the physical appearance of avatars they inhabit in virtual environments. If this generalises — and the body-swapping literature suggests it does — then the coming wave of VR and AR technologies won't just change what we see. They'll change who we think we are. The therapeutic applications are already in development: body illusions for chronic pain (reduced by almost 40% in one study), anorexia, phantom limb, and trauma-related depersonalisation.¹

The deeper implication, though, is for everyday self-understanding. The personality you take to be yours isn't stored in your brain like software on a hard drive. It's a running inference that your brain performs by integrating sensory signals from your body, contextual information from the environment, and social information from other minds — continuously, moment to moment. Change the body and you change the inference. The self is not the ghost in the machine. The self is the machine's best guess about what kind of machine it is.