nucleation

NOTE-1499: Metzinger — The Elephant and the Blind

13 min read

Properties2
StatusSkimmed
PaperLIT-1943 — The Elephant and the Blind: The Experience of Pure Consciousness: Philosophy, Science, and 500+ Experiential Reports

Graph View

Contribution

The book presents the free-text answers of a 2020 online survey of meditators (pp. xvi, 12–20). Respondents described an episode in which “awareness of awareness itself” was salient. Of 1,403 usable respondents, 841 gave usable reports, and more than 500 are printed. They are grouped into 34 loosely thematic chapters, each followed by commentary. Around this material the book states Metzinger’s later account of consciousness. The simplest conscious state, minimal phenomenal experience (MPE), is said to lack a self, a perspective, time and space. It is modelled as the organism’s model of its own capacity to know. The first-person perspective is said to be content of consciousness, not a condition of it (pp. 457–459). The book calls itself a “prolegomenon”, not a monograph (pp. xxvi, 455).

Key insight

Consciousness, on this account, is “knowing that one is alert” (p. 37). It is a model of the organism’s own epistemic openness. A self, or “epistemic agent model”, is a further layer, and it can drop away while that model runs (pp. 302–306). On the account’s own terms, though, what drops away is the egoic self-model, not self-modelling. The model of epistemic openness is itself “a single, integrated model of its own epistemic openness in the form of a highest-level but nonegoic self-model” (p. 477). The book’s slogan is that consciousness is “not subjective”. Read closely, its claim is narrower: consciousness needs no model of a knowing self. It does not say consciousness needs no model of the system at all.

Assumptions

  • Naturalism and representationalism. Pure awareness “is not literally contentless, but … actually represents something” (p. 366). Its content is the system’s epistemic capacity. Metaphysics is set aside: “I am not interested in metaphysics” (p. 463).
  • Probabilistic, not definitional, concepts. “Pure awareness” has “probabilistic, not definitional structure”, a family resemblance with a “hard prototypical core” (p. xxv). MPE is “a region in or path through the phenomenal state space” (pp. 463–464).
  • Report as access to the target. The book takes the reports, together with semantic analysis of traditional texts and the survey’s factor structure, as triangulating on one phenomenon. It names the method heterophenomenology, “not taking the reports as authoritative beyond the seeming itself” (p. 20).
  • Predictive processing. Wakefulness is “a statistical hypothesis in the brain, a neural representation of the probability that veridical perception will actually occur” (p. 41). The minimal model is sketched in active inference (Figure 34.1, pp. 475–477, with Sandved-Smith).

Key results

  • The survey (pp. xvi, 12–20). Online, early 2020, in five languages. Over 3,500 respondents, 1,403 with usable data, a mean age of 52, and more than 77% meditating regularly. There are 92 items rated on 0–100 sliders, and twelve factors chosen “under joint conceptual and statistical considerations” with “only limited explanatory power” (p. 12). Item 84, “The experience of ‘pure awareness’ is the simplest kind of conscious experience I know”, had a median of 80 (p. xix).
  • The phenomenological core (pp. 468–469). Simplicity, silence, wakefulness, clarity and “the signature of knowing”. Wakefulness appears in two factors, factor 4 (Wakeful Presence) and factor 3 (Self-Knowledge, Autonomous Cognizance and Insight). Factor 8, “Emptiness and Nonegoic Self-Awareness”, has its items for self-knowing “negatively correlated with the phenomenology of selfhood” (p. 466). The factor loadings themselves are not printed in the pages read.
  • States and modes (pp. xxi–xxii, 470–472). MPE occurs as a local state (“my mind was crystal clear”) and as global modes: witnessing, nondual awareness, bodiless body-experience and “virtuality”. A “full-absorption episode” is one after which awareness itself “is the only phenomenal feature that can later be reported” (p. xxiii). It is “concurrently ineffable” (p. xxi).
  • Five senses of “purity” (pp. 460–461). Contentlessness, no thought, clarity, suchness and nonduality.
  • The epistemic agent model (pp. 301–306). It is the layer of the human self-model in which a knowing, attending, thinking self is represented. It “is what creates subjectivity in the sense of a strong first-person perspective” (p. 302), and it “is precisely what is completely missing in all full-absorption episodes” (p. 306). On a conservative estimate, people lack a stable one for about two-thirds of the 24-hour cycle (Figure 25.1, p. 305; from Metzinger 2015, “M-Autonomy”).
  • The contraction principle (pp. 79–82). Phenomenality “is a subpersonal property, a property of certain functionally integrated brain states”. Ordinary brains misrepresent it “by contracting it into a transparent conscious self-model, which then forms the origin of a first-person perspective” (p. 80). Nondual states are offered as “well-documented counterexamples” (p. 80). “There is a methodological primacy to subpersonal state consciousness: It can occur without system consciousness, but not vice versa” (p. 81). “The subpersonal world-model in the person’s head is what is truly conscious” (p. 82).
  • Transparency (pp. 362–369). Being No One’s transparency is restated: contents are “represented as realities” because earlier processing stages are unavailable (pp. 362–363). It is applied to pure awareness: if the model of epistemic openness is transparent, its certainty is “transferred to the content” and a real world appears. If it is opaque, “there would be an explicit experience of emptiness and bare wakefulness” (p. 369). “Transparency creates realness; opacity creates knowingness” (p. 368).
  • The minimal model (pp. 476–477). Level 4 of a nested hierarchy models “the system’s global ‘epistemic openness’”. “It is level 4 that really endows the content generated on all lower-order levels with phenomenality” (pp. 476–477). A full-absorption episode is “inference at the highest level, without any perceptual inference at the levels below it” (p. 477). The sketch is “a hypothetical starting point” (p. 477).
  • Two fallacies (pp. 416–418). The E-fallacy takes a felt sense of knowing for knowledge. The C-fallacy takes a state that feels like the nature of consciousness for a finding about it. The C-fallacy has a mirror image: “Just because something is reported as feeling not like an experience at all, this doesn’t license the conclusion that it isn’t” (p. 417).

Claims

Not filled: the note is Skimmed.

Method

A three-way approach (pp. xx, 14): semantic constraints drawn from traditional texts (Metzinger 2020, LIT-1946), exploratory factor analysis of the 92 items, and a “soft qualitative analysis that doesn’t satisfy academic criteria for systematicity and completeness” (p. xxvi). The grouping of reports reflects “the implicit phenomenological intuitions of someone who has meditated twice a day for more than forty-seven years” (pp. 12–14).

Concepts

  • Minimal phenomenal experience (MPE): the simplest conscious experience. The working hypothesis is that it lacks “time representation; self-location …; ownership, agency, and autobiographical self-awareness; and a phenomenally experienced first-person perspective” (p. xviii).
  • Arousal, tonic alertness, wakefulness: a physical, a functional and a phenomenal property. Wakefulness “can be interpreted as a special way of representing tonic alertness” (pp. 36–37).
  • Epistemic openness: “a global form of phenomenal character … a specific form of openness to the world” (p. 41), the “sheer potential to know” (p. 43).
  • Epistemic agent model (EAM): the self-model layer of a knowing self. “A subject is an epistemic agent model that functions as the unit of identification” (p. 306).
  • Unit of identification: whatever experiential content one experiences oneself as. It can be the EAM, the body, or “the space of wakeful awareness” (pp. 361, 470–471).
  • Zero-person perspective: experiencing from MPE, without egoic self-awareness (p. xix).
  • Contraction: the misrepresentation of a state property as a system property (p. 80).
  • Full-absorption episode: an episode after which only awareness itself can be reported. Its types are I, entered from waking; II, from NREM sleep; III, from a lucid dream (pp. xxiii, 220).

Connections

It develops Metzinger 2020 (LIT-1946) at book length, with the survey behind it. The concept of minimal phenomenal selfhood comes from Blanke and Metzinger 2009 (LIT-1274), and its revision on bodiless dreams from Metzinger 2013 (LIT-1871, NOTE-1434). The self-model and transparency come from Being No One (LIT-467, Unavailable), read in the record through LIT-450 and LIT-568. Woods, Windt and Carter (LIT-1351, THEORY-866) are cited for the semantics of “pure consciousness” (p. 545, n. 9). The mind-wandering estimate comes from Metzinger’s “M-Autonomy” (2015). The machine-ethics passage cites Metzinger 2021 (LIT-1947). The active-inference sketch is shared with Laukkonen, Friston and Chandaria (THEORY-882), who also use MPE.

Bearing on the record’s own argument

  • THEORY-063 (unity of experience without a self-model): it qualifies the second clause rather than confirming it. The précis’s minimal case was “Selfless Snapshot Consciousness”, the presence of “one unified world” with no perspective. The précis itself called it “probably empirically empty”. This book offers full-absorption episodes as what such states are like in fact. Three things change. First, the minimal case is not a world experienced as one. In it “the distinction between world and self is not yet represented”, and there is no world-model content at all (p. 477). The 2020 paper adds that it lacks any “phenomenal quality of singularity” (LIT-1946, p. 13). Second, the model that remains is described as “a highest-level but nonegoic self-model” (p. 477). So what the case would show is unity without an egoic self-model, or an epistemic agent model, which is narrower than “no self-model”. Third, the cases nearer to THEORY-063’s selfless world are the nondual modes (pp. 470–471), in which a world is present and subject/object structure is absent. These, too, are known only by report, and the pages read give no measure of integration for them. THEORY-063’s promote_when asks for such states “established by methods that meet its own performative worry”. The book states the worry and leaves it open (p. 19). The inference is the record’s.
  • THEORY-064 (transparent self-model). The book restates transparency as Being No One gave it (pp. 362–363), and extends it from the self-model to the model of epistemic openness (pp. 368–369). On THEORY-064’s prediction that a fully opaque self-model would end selfhood, the book offers no test. Its opaque cases are opacity of awareness, not of the self-model. It reads meditation as “opacity control” (p. 366, after Sandved-Smith et al.), a mechanism of the kind that prediction needs. It gives no measure of opacity.
  • QUESTION-048 (what report evidence can establish). The book is a worked example of report-only evidence, and it states the limits itself. It names self-selection, theory contamination, memory and “response drift” (pp. 16–20), concurrent ineffability (p. xxi), and the autobiographical-memory problem: a report that one was selfless is “highly dubious from a methodological perspective”, since what is remembered is remembered as a self’s (p. 19). It also names the E-fallacy and the C-fallacy (pp. 416–418). The neural programme it proposes, the Triple Triangulation Project (pp. 220–221, 473), would look for “minimally sufficient neural correlates” common to absorption entered from waking, NREM sleep and lucid dreams. Every one of those episodes would still be identified by report. The summary nonetheless says that “full-absorption episodes, nonegoic self-awareness, and nondual modes of consciousness are empirically established phenomena” (p. 457), and that the strong reading of subjectivity “is false” (p. 459). THEORY-1295 states the gap.
  • QUESTION-046 (whether an organized whole is a subject of experience). The contraction principle answers part of the question with a stipulation, not a criterion. The bearer of phenomenality is “the subpersonal world-model”, and system consciousness is a representation of it (pp. 80–82). Phenomenality “can occur without system consciousness, but not vice versa” (p. 81). If that is right, the criterion the question asks for would have to find the conscious states inside a whole, not the whole. Same structure as the record’s distinction between an organization and its members’ states, not shown same mechanism.
  • QUESTION-254 and CLAIM-412 (experience as perspectival interpretive organization). CLAIM-412’s part 3 requires the organization to be perspectival. On this book’s account the simplest experience is aperspectival, though it is still an organization: a model of the system’s own epistemic capacity. If the account were right, part 3 would need either a sense of “perspectival” that the zero-person perspective satisfies, or a weaker condition. TERM-072’s sense 1, an organizational perspective, may be that sense. The book’s “level 4” (p. 477) is a candidate statement of what makes lower-order content experiential. It is stated without report-independent evidence, which is what QUESTION-254 asks for.
  • TERM-082 (senses of consciousness). The three-level split of arousal, tonic alertness and wakefulness (pp. 36–37) refines TERM-082’s sense 4, wakefulness or arousal. The book cites unresponsive wakefulness, a sleep–wake cycle without awareness, as tonic alertness without phenomenal wakefulness (p. 37). On its usage, “wakefulness” names a phenomenal property, so TERM-082’s sense 4 as worded would be split between the functional and the phenomenal.
  • TERM-083 (moral standing) and THEORY-554, THEORY-677. The epilogue separates consciousness from suffering. Artificial consciousness “may involve a consciously experienced sense of self that may in turn generate artificial suffering”. Yet MPE “involves no psychological suffering whatsoever”, and the book asks whether a machine that “reliably remains in a single full-absorption episode” would be wrong to build (pp. 484–485). The book does not ask whether such a machine would have standing. That is filed as QUESTION-368. The epilogue also flags an information hazard: the simplest conscious state “may be the easiest to implement” (p. 484).
  • LIT-1354 (Chandia-Jorquera et al.). Their “pure awareness” scale is defined by this framework’s MPE. The book’s own measure, the questionnaire’s sliders, has the same structure: practitioners rate episodes against the experimenters’ description of the target. The survey’s instructions defined the target as “awareness of awareness itself” (p. xvi). The worry NOTE-1058 raises about a defined scale applies here too, and more strongly: the reports are of past episodes, some from decades earlier (p. 18).
  • THEORY filed. THEORY-1299 states the account, with the 2020 paper. THEORY-1295 states the evidential gap.
  • No instruction for machine-learning practice.

Limitations

  • Read in part; see the history. The phenomenological chapters between 5 and 33, most of the book, were not read through, and some of the theory sits in their commentaries (chapters 5, 17, 22, 24, 27, 29, 30).
  • Report only. No physiological or behavioural measure is part of the study. The factor analysis is exploratory, its loadings are not printed in the pages read, and the author says future confirmatory work may give “very different results” (p. 12).
  • A count that does not match. The introduction says 841 usable reports and “more than 500” printed (p. xvi). The limitations section speaks of “the 1,183 written phenomenological reports presented in this book” (p. 17). The acknowledgments thank 1,187 meditators (p. xi).
  • The summary claims more than the methods allow (THEORY-1295).
  • A participant-researcher. The grouping is the author’s, a practitioner of 47 years, and he grants that it carries biases “I am completely unaware of” (p. 14).

Open questions

  • Would a report-independent marker separate full absorption from dreamless sleep without report? The Triple Triangulation Project needs one, and the book names none.
  • Does factor 8’s negative relation between nonegoic self-knowing and selfhood survive confirmatory analysis in a sample not recruited through contemplative networks?
  • Is the “nonegoic self-model” of p. 477 a self-model in Being No One’s sense, one the system identifies with, or only a model of the system? The answer decides how THEORY-063’s second clause should read.

Backlinks