Résumé de lecture, chapitre par chapitre — rapporte ce que le livre dit.
THIS DOCUMENT IS NOT A CITATION SOURCE — a reading summary, ungraded.
C'est un résumé de lecture, non coté, qui rapporte ce que le livre dit. Il ne porte aucun poids probant et ne fonde aucune affirmation. Toute citation reste ancrée sur le .txt du livre, jamais sur ce résumé.
Being You - A New Science of Consciousness - Seth, Anil (2021) (EN)| Title | Being You: A New Science of Consciousness |
| Author | Anil Seth |
| Year | 2021 |
| Publisher | Faber & Faber (ISBN 978–0–571–33773–6, attested in the text) |
| Type | popular-science essay |
| Source | library/39_Predictive_processing/Being You - A New Science of Consciousness - Seth, Anil (2021) (EN).txt |
| Text sha256 (16) | a697c3b8e8def1ce |
| Volume | ≈ 100,526 words (dossier) / 103,726 words (wc) |
| Structure | Prologue + 4 parts (I Level, II Content, III Self, IV Other) = 13 numbered chapters + Epilogue |
| Table of contents | ✅ READ (reconstructed from the book's contents page and re-verified against the section headings in the body) |
Note on structure. The associated DOSSIER declared a fallback ("a single block, insufficient segmentation"): its automatic detection failed. The structure below was reconstructed by hand — the book's "Contents" page is intact in the first 220 lines, and each part heading (LEVEL, CONTENT, SELF, OTHER) and chapter heading appears as an isolated line in the body, which made clean chapter boundaries recoverable. No part-level fallback was needed: the summary really is chapter by chapter.
Prologue
Part I — LEVEL (level of consciousness)
Part II — CONTENT (contents of consciousness)
Part III — SELF (the self)
Part IV — OTHER (beyond the human)
Epilogue
Seth proposes a science of consciousness that sidesteps Chalmers's "hard problem" rather than claiming to solve it. His reformulation — the "real problem" — accepts that experience exists and sets itself the task of explaining, predicting, and controlling the phenomenological properties of experience from cerebral and bodily mechanisms, without seeking to explain why the universe contains consciousness at all. It is a methodological shift, not a metaphysical solution.
The book is built in four movements. Level deals with the level of consciousness (from anaesthesia through coma to wakefulness) and with the possibility of measuring it, drawing on the analogy with temperature and on Tononi's integrated information theory (IIT) — which Seth admires but judges, in its "hard problem" version, untestable. Content develops the central thesis: perception is a "controlled hallucination", the brain being a Bayesian prediction machine that projects its best guesses onto the world rather than passively recording its signals.
Self extends this logic to the self: the self itself is a perception, a controlled hallucination of a particular kind. Seth anchors the self in the living body via interoceptive inference and the "beast machine" theory — emotions and mood are perceptions oriented toward physiological control and survival, not toward discovering the world. Here he connects Friston's free-energy principle and a deflationary reading of free will. Other finally opens onto animal consciousness and machine consciousness, where Seth remains cautious and anti-anthropocentric.
The tone is that of a researcher who bounds his own claims. Seth repeats that "the way things seem is often a poor guide to how they actually are", but applies the same caution to his own theses: controlled hallucination is an explanatory framework, the beast machine is not a demonstration that life is necessary for consciousness, and functionalism (hence machine consciousness) remains for him an open question, with no decisive argument either way.
In the 39_Predictive_processing folder, this book counts as an accessible popular exposition of predictive processing applied to consciousness and the self — useful for the overview and the vocabulary ("controlled hallucination", "beast machine", active inference, interoception), but it is not a primary evidence source: it is the narrative of a research programme, mixing lab results, syntheses of others' work, and avowed speculation.
"anaesthesia is the art of turning people into objects" (Prologue)
"the way things seem is often a poor guide to how they actually are" (ch. 1)
"addressing the real problem requires explaining why a particular pattern of brain activity – or other physical process – maps to a particular kind of conscious experience, not merely establishing that it does" (ch. 1)
"The real problem is distinct from the hard problem, because it is not – at least not in the first instance – about explaining why and how consciousness is part of the universe in the first place" (ch. 1)
"thermodynamics did more than merely establish that mean kinetic energy correlated with temperature – it proposed that this is what temperature actually is" (ch. 2)
"Conscious experiences are not only highly informative, they are also integrated" (ch. 2)
"According to IIT consciousness simply is integrated information" (ch. 3)
"it is precisely the ambition of IIT – to solve the hard problem – that renders its most distinctive claims untestable in practice" (ch. 3)
"what we see, hear, and feel is nothing more than the brain’s ‘best guess’ of the causes of its sensory inputs" (ch. 4)
"the contents of consciousness are a kind of waking dream – a controlled hallucination – that is both more than and less than whatever the real world really is" (ch. 4)
"‘prediction’ in prediction error minimisation is not necessarily about the future" (ch. 5)
"Priors, likelihoods, and posteriors are often called Bayesian ‘beliefs’ because they represent states of knowledge rather than states of the world" (ch. 5)
"When we experience the world as being ‘really out there’, this is not a passive revealing of an objective reality, but a vivid and present projection – a reaching out to the world from the brain" (ch. 6)
"One cannot perceive that which one cannot classify" (ch. 6)
"This is not my mother. This is a woman who looks like my mother, but who is not my mother" (ch. 7)
"The self is another perception, another controlled hallucination, though of a very special kind" (ch. 8)
"The self, too, is not what it might seem to be" (ch. 8)
"affective experiences are not merely shaped by interoceptive predictions but constituted by them" (ch. 9)
"Perception in this situation is a controlling hallucination just as much as it is a controlled hallucination" (ch. 9)
"It’s as close to a ‘theory of everything’ in biology as has yet been proposed" (ch. 10)
"the FEP, despite being touted as a ‘theory of everything’, is not a theory of consciousness" (ch. 10)
"experiences of volition do not reveal the existence of an immaterial self with causal power over physical events" (ch. 11)
"it is a self-fulfilling controlling hallucination related to specific kinds of actions – those actions that seem to come from within" (ch. 11)
"Absence of language is not evidence for absence of consciousness" (ch. 12)
"As we move further out, the evidence becomes sparser and sketchier, and inferences about consciousness more tentative" (ch. 12)
"This is the myth of the golem made silicon" (ch. 13)
Facing a living human brain in the operating theatre, Seth returns to Chalmers's hard problem and reaffirms the real problem approach: cooking "with whatever is in the fridge" (philosophy, neuroscience, psychiatry, AI…) rather than choosing an "ism". A meditative close on mortality: oblivion as a return to the eternity from which each of us emerged.
"the controlled hallucination of being you finally breaks down into nothingness" (Epilogue)
"perhaps it’s not so bad if a little mystery remains" (Epilogue)
It is not primary evidence. Being You is popular science written by a researcher: it weaves together results from his own laboratory (Sussex), syntheses of others' work (Tononi, Friston, Clark, Hohwy), and avowed speculation (the endorsements themselves note "some proposals being as yet unproven"). A reader in the 39_Predictive_processing folder must not draw an evidence grade from it: it is a programme narrative, not a systematic review or a meta-analysis.
Seth does NOT claim to solve the hard problem. The "real problem" methodologically sidesteps it — explaining/predicting/controlling phenomenology — without asserting that it has said why there is experience rather than nothing. He writes it explicitly: the real problem "is not – at least not in the first instance –" about why consciousness is part of the universe. Attributing to the book a solution to the hard problem would be a misreading.
"Controlled hallucination" is a framework, not a mechanistic demonstration. The phrase is a striking reformulation of predictive perception; it renders the intuition, but does not by itself prove the underlying neural mechanism. Likewise, the "beast machine" is a proposed theory: Seth states that he has not shown that life (flesh, blood, biological neurons) is necessary for consciousness.
The book does not settle machine consciousness or functionalism. Seth calls himself agnostic: "there is no knockdown argument either way" on functionalism. He therefore concludes neither that AI will be conscious nor that it will not — and he rejects the equation intelligence = consciousness.
Do not conflate his lab results with the narration. Some demonstrations (binocular rivalry, illusions, "hallucination machine", cardiac interoception) are cited empirical works; other passages are conceptual frameworks or analogies (temperature, free energy, cybernetics). The book does not provide the evidential apparatus for grading each: that belongs to the primary sources (Friston, Hohwy, Clark, Gershman), not to this summary.
sed -n by section boundaries. Sufficient for a faithful 3–6 sentence summary per unit; it is not a line-by-line reading of the entire body.grep. No part-level fallback: the summary really is chapter by chapter, with the real TOC used..txt (whitespace normalised, curly quotes and dashes normalised), via a Python script. No MISS; no invented citation; no evidence level (E) assigned.wc). This document reports what the book says and carries no evidential weight.