Selfhood

Two extraordinary first-person accounts — one from a woman who lost her self, one from a woman who never had one — converge on the same unsettling conclusion: the self isn't a thing you have, it's a process you do, and it can start, stop, and break in ways that reveal just how constructed the most intimate-seeming part of your experience really is.

Jane: Having a Self and Losing It

Jane Charlton works as head of European Human Rights at the UK Ministry of Justice. She holds conversations, makes jokes, talks about cats and Portuguese wines. From the outside, she seems completely normal. From the inside, she doesn't fully exist.¹

Jane has depersonalization disorder (DPD). She describes it as "grieving for my own death, even though I seem to be around to witness it." The constituent parts are all still working — she can articulate her experience, she knows who she is, she passes every test of intellectual self-knowledge. But an essential element is missing: the felt sense of being a subject rather than an object. Her friend Sarah describes it as looking at a snowglobe — the world is inside the glass, but you're on the outside, unable to connect the "I" looking in with the world in the globe.¹

This maps onto a distinction from Husserl's phenomenology that suddenly stops feeling academic when you meet someone living it. The body as Körper — physical object, perceived from outside — versus the body as Leib — lived body, experienced from within. For most people, the Leib feels more real. For Jane, the Körper has taken over. She sees herself the way others see her — a body moving through space — but she's lost the felt sense of being that body.

What makes DPD philosophically devastating is this: if the self were an "illusion" in any dismissive sense, losing it shouldn't be so bad. But people with DPD describe it as one of the worst things that can happen to you. Jane says she's in "a constant state of grief." This suggests that even if the self is constructed, it's a functional construction that hooks you into reality in a way nothing else does. The philosophers who blithely call the self an illusion should spend an afternoon with someone who's lost it.

DPD also disrupts interoception — patients show reduced ability to detect their own heartbeat and reduced empathy.¹ If Seth is right that the bodily self is built from predictions about internal signals,² this starts to look like a mechanism: lose the interoceptive prediction, lose the sense of embodied selfhood.

Helen: Never Having a Self Until Language Created One

Helen Keller's account of life before her teacher arrived is one of the only first-person reports we have of existence before reflexive self-awareness in a neurotypical brain.³

"I did not know that I am. I lived in a world that was a no-world." She had sensations, associations, habits, even desires — she could feel ice cream on her tongue and make the sign for it. But she insists this wasn't thought: "I had neither will nor intellect. I was carried along to objects and acts by a certain blind natural impetus." She "thought and desired in her fingers" — a phrase that uncannily echoes the Homeric Greek localization of mental functions in body parts, arrived at independently three millennia later by a child in Alabama.

The breakthrough came with language — specifically, with the first-person pronoun. When Keller learned the words "I" and "me," she "found that I was something." Then: "consciousness first existed for me." Not touch, not sensation, not experience — language created the self. "It was the awakening of my soul that first rendered my senses their value."³

This is a radical claim, and it's worth sitting with the possibility that she's wrong about her own history — that she's retroactively imposing a narrative of absence onto experiences she simply can't access from her current vantage point. But her account is remarkably specific and internally consistent, and it resonates with both the phenomenological tradition and the Jaynes/Simler argument that the unified self is a linguistic-cultural construction.

Seth's Taxonomy

Seth proposes at least five separable layers of self-experience:² the bodily self (being a body), the perspectival self (the first-person viewpoint), the volitional self (intention and agency), the narrative self (continuous identity over time, built from autobiographical memory), and the social self (your self-experience refracted through others' minds). In daily life these feel seamlessly integrated. But they can come apart — DPD strips the bodily and perspectival layers while leaving the narrative layer intact. Keller seems to have had proto-versions of the bodily self (she felt things in her body) without any narrative or perspectival self until language provided the scaffolding.

The rubber hand illusion — feeling ownership of a fake hand that's stroked in sync with your real hand — shows the bodily self is constantly being constructed through Bayesian inference. The brain decides what counts as "my body" based on the best available evidence, and that evidence can be hacked. The virtual-hand-pulsing-with-your-heartbeat experiment extends this to interoception: people feel greater ownership of a virtual hand that pulses in sync with their actual heartbeat.²

The Narrative Self: Overrated?

There's a dominant story in psychology and philosophy that the self is fundamentally a narrative — "we are the stories we tell about ourselves," as Dennett, Sacks, and McAdams all claim in various ways. Galen Strawson thinks this is false, and not just for unusual people.⁴

Strawson distinguishes Narrative types (people who naturally experience their lives as stories, living through and into those narratives) from non-Narrative types — people for whom life "never assumes a story-like shape." He's in the second camp. He has no sense of being the author of his life, experiences his thoughts as things that just happen rather than things he generates, and finds Montaigne's comment — "I can find hardly a trace of memory in myself" — more descriptive of his mental life than any narrative theory. He cites Updike's "persistent sensation that I am just beginning" and Fernando Pessoa's "Each moment I feel as if I've just been born into an endlessly new world."⁴

The narrativist consensus is so strong that Schechtman originally argued you "must be in possession of a full and explicit narrative of your life to develop fully as a person." She's since softened this to allow mostly implicit, unconscious narration — which conveniently makes it impossible to disprove. Strawson's objection isn't just that he personally isn't Narrative. It's that the narrativist view pathologizes a large class of perfectly normal people. Mary Midgley writes of people who "have to hold a meeting every time they want to do something slightly difficult, in order to find the self who is capable of undertaking it." Paul Klee described his self as "a dramatic ensemble" — prophetic ancestor, brutal hero, alcoholic bon vivant, lyric muse, pedantic papa — watched with amazement by an I holding a sharpened pen.⁴

This matters because it complicates Seth's taxonomy. If the narrative self is just one layer that can be present or absent without affecting the others — and Strawson seems to function perfectly well without it — then the self is even more modular than Seth's five layers suggest. Some people walk around with a robust bodily and perspectival self but essentially no narrative self, and they're fine. Others, like people with DPD, have the narrative self intact but the bodily self is gone — and they're in agony. The layer that matters most isn't the one the narrativists obsess over.

The Self as Prime Minister

Kevin Simler pushes the modular picture to its logical conclusion. If neurons are selfish agents competing for metabolic resources — an idea he traces from Dennett and Seung — then sub-personal agents (drives, addictions, conscience, curiosity) are coalitions of these selfish neurons, and the self is something like a prime minister: "not powerful in its own right, but because it has managed to become the public face for the most powerful faction."⁵

The self, on this view, is less a designed feature of the brain and more a growth — an emergent agent that develops naturally when a brain teeming with agency is placed in a social environment. Every normal human brain grows one. But "normal" is doing a lot of work there. Tulpa practitioners deliberately cultivate additional person-like agents in their brains. Possession trances install new agents that need to be taught to speak. Schizophrenic voices are agents with their own motives and personalities. Addictions are agents that co-opt your reasoning to serve their purposes — people literally name them ("I call it Stan") and describe negotiating with them.⁵

What Simler's framework does is naturalize the weirdness. If the brain is hospitable to agency the way soil is hospitable to growth, then multiple selves, hallucinatory voices, and possession states aren't mysterious — they're expected outcomes when the growth process goes differently. This connects directly to the bicameral mind theory: the hallucinated gods of Bronze Age humanity weren't malfunctions, they were a different growth pattern for agent hierarchies in brains that hadn't yet developed the cultural technology of unified selfhood.

And Wegner's work on the illusion of conscious will adds another thread: the feeling that "I" am causing my actions may itself be an after-the-fact confabulation. The real causal chain runs from unconscious neural processes to both intention and action simultaneously — we experience the intention, see the action follow, and infer causation. The self isn't the driver; it's the press secretary, constructing a coherent story of agency from events it didn't initiate.⁶

When the Self Breaks Culturally

The Truman Show delusion offers a modern mirror to DPD — not a loss of self but a runaway self that can't stop being the protagonist. Kevin Hall, an Olympic sailor with bipolar disorder, experienced manic episodes in which he believed he was the star of a global TV show, directed by an unseen Director. During a terrifying boating accident that killed a teammate, he couldn't be sure if the death was real or scripted. Google Glass, given to him by his sailing team, became a trigger — it confirmed the surveillance model his mania had built. At Legoland, every coincidence (a Dalí bust, a replica of the Golden Gate Bridge) became a planted clue.⁷

The Gold brothers, psychiatrists who coined the term, noticed something important: delusions track the available cultural technology. Jealousy delusions are more common in Germany than Japan. Delusions of grandeur cluster among wealthy men. Biblical delusions have waned as religious authority recedes. And now, in an era of ubiquitous cameras and livestreaming, patients arrive at Bellevue insisting their lives are being filmed — and they can point to the internet and say "They did write about me." As Joel Gold asked: how do you convince someone they're delusional when the surveillance infrastructure they're describing actually exists?⁷

This is a direct test of the constructionist view of self. If the self is a model the brain builds, then psychosis doesn't generate random noise — it generates culturally shaped models that have gone wrong in culturally specific ways. The Truman Show delusion couldn't have existed before mass media. Bronze Age psychosis looked like hearing the voices of gods. The machinery is the same; the furniture changes.

The Parts and the Demons

IFS therapy takes Simler's prime-minister model and operationalizes it. Internal Family Systems, the fastest-growing psychotherapy framework, treats the mind as containing a calm, wise Self surrounded by "Parts" — sub-minds with their own agendas, emotions, and histories. A Part might be the one that sabotages your relationships, because after your first breakup it decided you should never get close enough to hurt again. Therapy involves discovering these Parts (not inventing them — patients report them as found objects in a trance-like state), learning their motives, and negotiating.⁸

The interesting thing is what happens when the framework breaks. Robert Falconer, a veteran IFS therapist endorsed by IFS's founder, reports that occasionally a Part will refuse the premise. It won't identify as a protective sub-self. It says it's an external entity that entered during a moment of unbearable trauma — a demon, in the traditional vocabulary. Falconer's book documents IFS exorcism procedures, and the IFS founder wrote the foreword.

Scott Alexander's analysis cuts through the metaphysics: people see the inside of their mind the way their culture tells them to. The modern West has a materialistic unitary theory of mind — 1 brain = 1 mind — so distressing internal voices get filed as misfiring circuits. IFS tells you to look for Parts, and people find them, just as the 1980s multiple personality panic told people to look for alters and people developed them. Tell people to look for demons and some will find those too. None of these framings are straightforwardly wrong, because the mind genuinely is multiple in some ways and unitary in others. The question isn't which theory is correct but which produces better outcomes — and the IFS therapists, including the ones doing exorcisms, report outcomes that are hard to dismiss.⁸

The deepest connection: Falconer reports that demons typically enter during moments of overwhelming trauma, when the person will "take help from any corner." This maps onto the prime-minister model — when the ruling coalition collapses under stress, new agents can install themselves. Whether you call them Parts, demons, dissociative alters, or the voices of gods depends on the cultural operating system. The underlying phenomenon — a mind hospitable to multiple agents — is the same one Simler described, Jung mapped, and the bicameral mind theory historicized.

The Shadow Self

Jung's model of the psyche adds historical depth to the modular picture. Where Simler thinks in terms of neural coalitions and Strawson in terms of phenomenological types, Jung proposed structural components — ego, persona, shadow, anima/animus — that interact dynamically. The persona is the public-relations mask; the shadow is everything that doesn't fit behind it. The shadow doesn't go away when repressed — it "begins to take on a life of its own," projected onto others. The qualities we can't stand in other people are, in this framework, the qualities we refuse to see in ourselves.⁹

What makes this more than historical curiosity is how well it maps onto the IFS findings and the prime-minister model. Jung's complexes — "themed organizations in the unconscious mind" that "behave in a rather automatic manner" — are functionally identical to Simler's competing neural coalitions and IFS's Parts. In a healthy person, complexes balance the ego's one-sided perspective. In an unhealthy person, the ego is too weak to integrate them, and they take over — which is exactly what DPD, the Truman Show delusion, and IFS demons look like from different angles.⁹

The individuation process — Jung's term for the lifelong project of integrating conscious and unconscious elements into a functioning whole — is essentially the self-construction process that Seth's taxonomy describes from a neuroscience perspective. Both frameworks agree that the self isn't a given but an achievement, something built through the ongoing integration of competing elements. The difference is that Seth locates the construction in Bayesian inference on interoceptive signals, while Jung locates it in the symbolic negotiation between archetypes. These might be descriptions of the same process at different levels of analysis.

Mirror Neurons and the Social Origin of Self

Ramachandran offers a neurological mechanism for how the self might actually be constructed: mirror neurons turned inward.¹⁰ Mirror neurons fire both when you perform an action and when you watch someone else perform it — they're "mind-reading" neurons that simulate another's behavior internally. "Empathy neurons" in the anterior cingulate fire equally when you're poked with a needle and when you watch someone else being poked. As Ramachandran puts it, "the neuron in question simply doesn't know the difference between it and others."

His provocative claim is that self-awareness evolved after other-awareness, not before. The conventional assumption is that you need a pre-existing self in order to build a theory of other minds. Ramachandran argues the reverse: mirror neurons first evolved to model other primates' intentions (essential for "Machiavellian" social survival), and then the same mechanism was "turned inward" to model one's own mind. Self-awareness is "using mirror neurons for 'looking at myself as if someone else is looking at me.'" It's not coincidental that we say "self-conscious" when we mean conscious of others being conscious of us, or "I am reflecting" when we mean aware of ourselves thinking — the language preserves the social origin of the capacity.

This connects to several threads in the article. If the self is a prime minister maintaining power through social position (Simler's model), and if the neural machinery for selfhood is literally repurposed social-cognition hardware (Ramachandran), then the self is social all the way down — not just in its function but in its evolutionary origin. The prediction this generates is testable: people with deficient mirror neurons should have deficient self-awareness. Autistic children, who do have measurably reduced mirror neuron activity, show correspondingly reduced theory-of-mind capacity — and Ramachandran predicts they should also show reduced introspective ability, a prediction that resonates with what IFS therapy discovers about parts and selves.¹⁰

The Question This Leaves Open

What's the minimum requirement for selfhood? Keller's account suggests language is necessary — but deaf people who never heard speech still report inner "voice" in sign language. Social interaction seems important — Jane seeks physical contact with others "as if I need to be that other person because my own sense of self is not strong enough to sustain me."¹ The body is clearly involved — but the Extended Mind Thesis raises the possibility of selves that extend into tools and environments.

The deepest question: is there a meaningful connection between DPD and contemplative practices that aim to dissolve the self? Buddhist anattā (no-self) sounds superficially similar to depersonalization. But meditators typically describe dissolution as liberating rather than horrifying. If the same phenomenon can be the worst thing (Jane) or the best thing (advanced meditators), then the context and framing of selflessness matters as much as the selflessness itself — which is itself a very Buddhist observation. Dzogchen practitioners would say the difference is between losing the self through malfunction (terrifying) and recognizing the self as empty through insight (freeing) — the same groundlessness, experienced from opposite directions.