If the self were what it presents itself to be — the foundation of conscious experience, the ground that everything else sits on — then its absence should be the absence of everything. Take away the foundation and the building falls. Take away the ground and there is nowhere for anything to stand. This is what foundations do. They are what their dependents depend on.
This is testable. There are three situations in which the self drops while measurements continue to be possible. They were studied by three different research communities, in three different decades, using three different methods, for three different reasons. Anaesthesiology, psychopharmacology, the psychology of skilled performance. None of these research programmes were designed to test whether the self is load-bearing. Each of them, in its own terms, simply followed where the evidence went. The convergence is what this essay is about.
Picture what general anaesthesia actually involves. A person is wheeled into the operating theatre. They are conscious — alert, aware of the room, talking with the anaesthetist, perhaps anxious about what is about to happen to them. The anaesthetist administers an intravenous dose, usually propofol. Within thirty seconds the person is no longer conscious. They cannot be roused by their name being called. They cannot respond to questions. They cannot feel pain. For the next several hours, while major surgery is performed on their body, they are not present.
But the body is. The body continues to do almost everything it was doing.
The heart continues to beat. Heart rate is monitored continuously throughout the procedure, and while anaesthesia depresses heart rate somewhat, the cardiac autonomic system continues to regulate rhythm and output. The lungs continue to breathe — either spontaneously, in lighter forms of anaesthesia, or with mechanical ventilation, and in both cases the chemoreceptors that monitor oxygen and carbon dioxide continue to operate, adjusting respiration as the blood chemistry changes. Body temperature is maintained through the hypothalamus, which continues to monitor core temperature and respond to deviations. Blood pressure is regulated through the autonomic nervous system. The immune system begins repair of the surgical wound before the patient is awake; inflammatory responses to the trauma proceed; coagulation occurs at the incision sites. Hormones continue to be released on their circadian schedules. The body, considered as a biological system, does not know that anyone has gone missing.
This is what general anaesthesia reveals. The body and the self are not the same thing. They were coextensive while the patient was conscious, and the patient experienced them as if they were one thing — I am my body, having my experience. They were not. When the propofol arrives, one of them disappears and the other continues. The disappearance of the self is not the disappearance of the organism. The organism, considered as a biological system, is entirely fine. Whoever was there to inhabit it is gone.
There is a second observation from the operating theatre that strengthens the first. Anaesthesia is sometimes incomplete. About one or two patients in a thousand undergoing general anaesthesia experience episodes during surgery where some component of consciousness persists — what is technically called intraoperative awareness. These cases are rare, but they are well-documented in the anaesthesiology literature, and they are clinically informative. They reveal that what we casually call being under is not a single thing that can be either present or absent. It is several things, each produced by a different neural system, each separately disrupted by the anaesthetic cocktail.
A patient can have residual sensory awareness without being able to move (because the muscle relaxants in the cocktail are still working, regardless of whether the hypnotic component has worked). A patient can have implicit memory of events during surgery without explicit recall — the memory is laid down but cannot be deliberately retrieved. A patient can experience pain during the procedure without forming a lasting memory of having experienced it. A patient can be aware of being in an unfamiliar environment without being able to identify where they are or why.
What this dissociability shows is the deeper observation behind the surgical case. Consciousness is not a single faculty that is either present or absent. Consciousness is a set of integrated components — sensory awareness, the sense of being a subject, the formation of memory, the capacity for action, the felt continuity of self — each one produced by specific neural systems. The self — the felt experience of being someone — is one of these components. It is not the whole. It is not the foundation. It is one specific thing the brain does, and anaesthesia can disable it while leaving the others variously intact, in various combinations, depending on the specific cocktail and the specific patient.
The ordering, in the surgical case, is plain. Under anaesthesia, the self drops and the body persists. This is not a hypothesis under investigation. It is the daily working assumption of every surgical theatre in the world — that the conscious self can be reliably removed for several hours and that the body will continue to function in the interval. Sixty thousand procedures a day in the United States alone are performed on this assumption. The assumption is right, every time.
In 2009, Robin Carhart-Harris and a research team at Imperial College London began a series of studies that placed healthy volunteers in functional MRI scanners and gave them carefully measured intravenous doses of psilocybin — the active compound in psychedelic mushrooms. The team was measuring what the brain actually does under the influence of a classical psychedelic. There was a widespread popular conception, dating back to the 1960s, that psychedelics work by increasing brain activity — opening doors of perception, hyperactivating consciousness, throwing the system into a high-arousal state. What Carhart-Harris and his team found was the opposite.
The first paper, published in 2012 in the Proceedings of the National Academy of Sciences, measured cerebral blood flow in fifteen healthy volunteers before and after intravenous psilocybin. The expectation was that the team would find some regions of the brain becoming more active under the drug. What appeared in the data was decreases. No increases at all. The decreases were localised to specific high-order association cortices — the medial prefrontal cortex behind the forehead, the posterior cingulate cortex deep in the brain's centre, the angular gyrus regions in the parietal lobes. These are the same regions that compose the default mode network — the network identified in TT-03 as the brain's self-network, the system that activates during self-referential thought and consumes a disproportionate share of the brain's energy budget continuously running the experience of being someone. Under psilocybin, this network does not activate. It quiets.
Subsequent studies replicated the finding with BOLD fMRI — a different imaging method measuring different physiological signals. Different research groups, in different countries, using different protocols, found the same effect. Across psilocybin, LSD, the DMT in ayahuasca, and salvinorin from Salvia divinorum, the same network suppresses. The molecular mechanism differs across these compounds — most act on serotonin 5-HT2A receptors, salvinorin acts on kappa-opioid receptors — but the effect on the DMN is consistent. Different molecular pathways. Same network quiets.
This is the same pattern observed under general anaesthesia, by an entirely different molecular route. Propofol acts at GABA-A receptors — the brain's primary inhibitory neurotransmitter system — and broadly increases inhibition across the cortex, the thalamus, and the brainstem. But the inhibition is not uniform. The cortex does not simply go dark. Imaging studies across the last fifteen years have established a specific pattern of cortical disassembly. Higher-order association cortices suppress first. The medial prefrontal cortex, the posterior cingulate cortex, the angular gyrus — the same DMN regions — go quiet before primary sensory cortex shows substantial effects. Primary sensory cortex can continue to register inputs well into anaesthesia; the higher-order processing that integrates inputs into a coherent conscious experience has been suppressed. The signals arrive at the door. There is no one home to read them.
The deeper observation behind both cases is the same. The brain does not lose consciousness because all its neurons shut off. The brain loses consciousness because the integration of activity across regions breaks down. The hierarchy described in TT-02 — the cascading layers of predictions and prediction errors — comes apart from the top down. The deepest layers go first. What remains underneath is sensory machinery still functioning, motor machinery still functioning, regulatory machinery still functioning — but without the integrated whole that knits them into the experience of being someone. The same pattern operates under both anaesthesia and psychedelics, reached by completely different molecular routes, with one specific consequence: the network producing the self quiets first.
What subjects in the psychedelic studies actually report, while this is happening to their brains, is remarkably consistent. Across hundreds of subjects in dozens of studies and across the different classical psychedelics, the reports converge on a small number of features. Subjects describe a softening or loss of the felt boundary between self and world. The sense of being a particular subject located behind the eyes, separate from the room around them, recedes or disappears. The conventional sense of self thins out. In its place, depending on dose and individual, come reports of merging with surroundings, of having no centre, of being everywhere and nowhere at once, of a profound sense that the apparent solidity of being someone has lifted.
The field's name for this phenomenon is ego dissolution. The term has technical precision — it names the specific subjective report of the self-prediction losing its grip, the felt experience of what TT-03 described as a high-precision prior becoming a low-precision one. The use of the word ego here is not the everyday sense (vanity, self-importance) and not the Freudian sense (the negotiating layer between id and superego). It is the narrower technical sense: the sense of being a particular subject at the centre of one's experience. That sense dissolves — softens, thins, in some cases vanishes entirely for a period.
Research groups have measured the correlation between the subjective experience of ego dissolution and the neural changes. Tagliazucchi and colleagues, working with LSD and publishing in Current Biology in 2016, found that the magnitude of ego dissolution subjects reported correlated robustly with the magnitude of disruption to the DMN. The more the network was disrupted, the more pronounced the experience. The two track each other. The network whose activity correlates with self-referential processing under normal conditions is the same network whose suppression correlates with the felt disappearance of the self under psychedelics. The neural signature and the subjective report are not two separate findings. They are the same finding, measured two different ways.
The framework's interpretation of this came later than the empirical findings. In a 2019 paper in Pharmacological Reviews, Carhart-Harris and Friston proposed the REBUS model — Relaxed Beliefs Under pSychedelics. The proposal is straightforward in light of the predictive-processing account of the self developed in TT-03. Psychedelics, on this account, work by reducing the precision-weighting of high-level priors. The priors do not disappear. They become less rigid, more open to update, less able to suppress contradicting evidence arriving from lower levels of the hierarchy. The self, which is the highest-precision prior under normal conditions, loses its precision under the drug. The structure of the hierarchy temporarily flattens. The system continues to function — perception continues, motor control continues, the body persists — but the prediction that organises all the others into the experience of being a particular subject has loosened.
This is the framework's prediction, made independently of the psychedelic data and then matching it. The empirical convergence of two domains — anaesthesia and psychedelics — both producing DMN suppression by completely different molecular routes, both producing the felt drop of the self, both leaving the body intact, is what the predictive-processing framework would have predicted if asked in advance. Whether the framework is the correct interpretation or not, the empirical convergence is real. Two completely different research traditions, working on completely different problems, arrived at the same finding without coordination. The self can be chemically suppressed, by two different chemistries, and the system continues to function. The body does not need the self to operate.
The first two cases were pharmacological. The self drops under chemical intervention; the disruption is imposed from outside the system. The third case is different. Flow states arise endogenously — in the course of skilled action, with no drug administered, with no external chemistry. The same neural signature appears, generated by the brain's own response to a task that fully engages it.
The case in detail. When a skilled musician plays a difficult passage in performance. When a surgeon performs a complex procedure. When an athlete is in the zone. When a writer is fully inside the work. When the chess player has been at the board for two hours and the position has fully entered them. The subjective reports across these very different domains converge on the same set of features: the conventional sense of self thins out, the running commentary in the mind quiets, time alters (sometimes speeding, sometimes slowing), action and awareness merge into a single ongoing process that is no longer experienced as having an agent separate from what is being done. Mihaly Csikszentmihalyi documented this state across decades of research with thousands of subjects in dozens of skilled domains. The reports were unmistakably consistent across activities that had nothing else in common.
When neuroscientists began studying flow with brain imaging, beginning with the work of Ulrich, Keller and colleagues across the 2010s, the neural signature turned out to be the same one that had been documented under anaesthesia and psychedelics. Down-regulation of the core DMN regions. The same medial prefrontal cortex. The same posterior cingulate cortex. The same network that produces the felt experience of being someone. When skilled performers entered flow, the network quieted, and the magnitude of the quieting correlated with the magnitude of the subjective self-loss they reported.
What makes flow distinct from the other two cases is what happens to the performance.
Under anaesthesia, the performance question does not arise — the person is not performing anything, they are unconscious. Under psychedelics, performance on most tasks is impaired or at least altered; the studies do not measure productivity. Under flow, the performance is precisely what is being measured. And the data is unambiguous. The performance does not collapse when the self quiets. The performance improves. The performer is doing better work — more accurate work, more responsive work, faster work, with fewer errors — in the very moments when the self-network is least active.
This is the inversion of what the standard intuition would expect. If the self were the agent of skilled action — the one doing the playing, the surgery, the writing — then losing the self should impair the doing. If the self were the orchestrator of attention, the holder of intention, the centre of control, then dampening the self should produce worse outcomes. The data is the opposite. The doing improves when the self gets out of the way. The mathematics get solved more cleanly. The surgical incision is more precise. The music is more felt. The prose flows. None of this is happening despite the absence of the felt self. All of it is happening because the self has stopped intervening in what the body and the trained capacities can do unaided.
Three domains. Anaesthesiology, psychopharmacology, and the psychology of skilled performance. None of them shares methodology, vocabulary, research community, journal, or funding source. They were not in conversation with each other. They were not testing the same hypothesis. Each one was investigating its own question with its own tools, in its own decade, for its own reasons. Anaesthesiology wanted to make surgery survivable. Psychopharmacology wanted to understand how a class of compounds affects the brain. The psychology of skilled performance wanted to understand what made people excellent at what they did.
What they converged on is the same finding. The self can drop. The body remains. The capacities remain. Under flow, the capacities improve when the self drops. The three independent research traditions produced the same observation, by different methods, on different timelines, with no coordination between them. This is the structural shape this series has been describing since the dandelion. Local rules, applied across independent substrates, converging on a single finding that none of the individual researchers was looking for.
What a system sheds first when resources are constrained tells you what was not load-bearing. The order in which things drop reveals what depended on what. The first things to go are the components that other components do not require. The last things to remain are the components without which the rest cannot operate.
In all three of the domains examined above, the self drops earlier than the body. The DMN suppresses before primary sensory cortex under anaesthesia. The DMN suppresses while motor and autonomic functions remain intact under psychedelics. The DMN suppresses while skilled action proceeds and improves under flow. Across three completely different mechanisms — pharmacological inhibition of cortical integration, pharmacological reduction of high-level prior precision, and endogenous task-driven attentional capture — the self is the first thing to go, and the rest of the system continues to function in its absence. In every available test, the self is the first component to drop and the last component required for the rest to operate.
This is a strong empirical finding. It does not depend on the predictive-processing framework being correct. It does not depend on any particular theory of consciousness. It is what is observed across three independent research traditions, with three different methods, in three different decades, by people who were not in conversation with one another. The ordering is what the ordering is. The framework offered in TT-02 and TT-03 — the self as the highest-precision prior in a predictive hierarchy — is one way of understanding why the ordering takes this specific form. The ordering itself would still be the ordering even if the framework were wrong.
What the ordering tells us is the load-bearing observation of the series so far. The self is not at the foundation. The self is a particular kind of process the brain runs — a continuously updated prediction of being someone, sitting at the top of a predictive hierarchy, present in ordinary experience but not required for the system to function. When the self is removed, by whatever mechanism, the body continues. The capacities continue. The experience that remains, in the cases where reports are available, is not nothing. It is something else.
The next essay turns toward what is reported when the self drops and the system continues. It will not be theoretical — the theory has done its work. The next essay takes seriously what the people who have spent significant time in states of self-dissolution actually report — across psychedelic research, across long-form meditation, across the cardiac-arrest reports, across the contemplative traditions that have spent millennia studying precisely this. The convergence in those reports is the next instance of the count. The investigation, which has been carried so far by science, continues there.