*(Historical record, filed 14 July 2026, before the v2.0 rename: "resolution" herein is the framework term since renamed "cohering", and claim labels are as of v1.6. Preserved unedited.)*

# Constructive Red-Team: The Container Problem × Space Immanence Governance Fit

**Referent:** `The_Container_Problem_v0.1.md` (working note v0.1, 14 July 2026, published in this corpus)
**Against:** `site/claims.json` (v1.6), `site/state.json`, `site/agents.html` (evidence non-transfer rule), `test/state-drift.test.mjs`, `test/claims-json.test.mjs`, `site/objections.html`, `site/research.html` (Study 5), `site/readings.html` (Dōgen reading).
**Date:** 14 July 2026. Prepared as red-team input; the note asked to be attacked, not admired.

**Literature verification status.** Claims relied on below were checked against the web (see Appendix A). Verified: Starmans & Bloom 2012 (eyes-location, including the chest-eyed alien); Bertossa et al. 2008 (Point Zero; blind subjects included, "midway behind the eyes"); Forstmann & Burgmer, "The Cartesian Folk Theater" (JEP: General; confined prefrontal localisation); Gamma & Metzinger MPE-92M (PLOS ONE 2021; the four factor names quoted in the note are real factor names; N=1,403 usable of 3,627); Okumura, *Realizing Genjōkōan* (the illumination/darkness passage is genuine Genjōkōan commentary); Graziano's "ethereal" description (near-verbatim: "an ethereal thing... vaguely localisable to space inside us"). Partially verified: Chalmers's meta-problem catalogue — his problem-intuition taxonomy (explanatory/gap, metaphysical, quality, epistemic/acquaintance) indeed contains no container intuition as such, but see §1.4 on the "unclaimed move" claim; the count "fifteen canvassed solutions" is approximately right (the follow-up paper consolidates to 7+6). Unverified/contestable: "Graziano's schema has no reason to reshape with contemplative training" — I could not confirm AST makes no such prediction; AST holds the awareness model to be a plastic, updatable schema, so this rival-elimination claim needs a citation before the trajectory is billed as uniquely discriminating. The "Cartesian Folk Theater" N≈2,057 figure was not independently confirmed (paper verified, headline N not).

---

## 1. Firewall analysis

### 1.1 The transfer temptation, node by node

The dangerous sentence is in the note's second paragraph:

> "On this account the container error (already central to Space Immanence v1.5 as diagnosis) acquires a generative mechanism: it is what partial illumination produces by default."

"Acquires" flows credit in the wrong direction. Read naively: mechanism confirmed → the diagnosis "has" its mechanism → the diagnosis is supported. That inference is invalid at almost every node, and at two nodes it has the sign backwards. Node by node:

**S1 (Container-logic distorts both problems, confidence 0.8).** The primary temptation, because it is the node the note name-checks. What B+C confirmation would actually show: (i) a separable non-spatial container factor exists in self-report (B), and (ii) a medium-shaped encoding is compression-optimal for an unpresentable remainder (C). What S1 claims: that the *hard problem* and the *spacetime-emergence problem* are both distorted by container-logic. B+C say nothing about spacetime, nothing about whether philosophers' formulations inherit the schema, and nothing about whether the problems are thereby *distorted* rather than merely accompanied by the intuition. The confidencePolicy in claims.json is explicit that 0.8 prices only the weak reading ("the container schema is load-bearing in the relevant cognition") with the strong reading "held pending the Part I production-presupposes-containment argument." A confirmed mechanism does not supply that argument. **Partial transfer is legitimate here, and only here — see §1.3.**

**S3 (Investigate consciousness structurally).** Depends on S1; inflation arrives transitively if S1 inflates. No direct edge; block by blocking S1 inflation.

**S4 (Contemplative phenomenology is data).** Temptation: if the residue instrument (B) extracts stable, discriminating structure from introspective reports, that "vindicates introspection as data." But S4's falsifier is about two-orientation clustering, which B does not touch. B succeeding is generic method validation — grade it, at most, the way the swarm study is graded: discipline-level, `bearsOnClaims: []`.

**M1 (the fold).** The strongest *aesthetic* temptation. "Ground that funds the current figure but cannot itself be shown" reads like the fold's inward orientation; partial illumination reads like the aperture. But darkening is a representational/compression story about self-modeling systems — it is a Metzinger/Graziano-family mechanism, fully at home in orthodox physicalism. The note concedes this in its own kill-check: "'Introspection misrepresents its own basis' is the reigning consensus programme." A confirmed C-result is therefore *tier-neutral at best* for M1, and arguably strengthens the deflationary toolkit that obj12 (illusionism) draws on. Any transfer must be blocked absolutely.

**M2 (Predictive processing implements the fold).** The mechanism is phrased in attention/prediction/salience vocabulary, so a formal result in C will look like "PP machinery doing fold-work." C's formal object is a compression optimum over encodings; none of the four fold-criteria appear in it. Block.

**M5 (the fold becomes lived through resolution) — and the trajectory.** Point → container → collapse maps temptingly onto M5's practice-phenomenology claim, and shares the MPE collapse endpoint. But the note itself already refuses the endpoint transfer: "Container collapse at the far end of practice is a consistency check we must pass, not credit we can claim." The same sentence must be written for the whole trajectory: B confirming the stage sequence supports the darkening mechanism's ordering prediction, not the claim that practice is "transparency to resolving." Different explananda; block.

**P2 (Inside-ness is the act of the fold's resolving).** **The sign error.** The container note is *about* inside-ness, so readers will treat mechanism confirmation as support for P2. But P2 says inside-ness is the real structural character of resolving from within; darkening says the container-feel is a *default compression artifact*. These compete. If darkening explains the sense of an inside as cheaply as it explains the sense of a container, P2 loses its explanandum to a deflationary rival. A confirmed C-result is closer to being adverse evidence for P2 than support. This is the single most important thing the firewall must state explicitly, because it is the transfer nobody will suspect needs blocking in the *negative* direction.

**P1 (spacetime as self-reference resolving), S2/P3 (convergent pressure).** No evidential channel at all; only the vocabulary bridge (§4). The temptation for S2/P3 is rhetorical — "an unclaimed move nobody else derives" spun as the framework being ahead of the convergence. Block trivially by never listing the container programme as a domain under S2.

### 1.2 Does the note already lean on fold-coherence as support?

Strictly: **no evidential lean, two adjacency leans.** The word "fold" never appears in the note, and the note contains one exemplary act of firewall discipline, worth preserving verbatim as the template:

> "Container collapse at the far end of practice is a consistency check we must pass, not credit we can claim."

The two leans:

1. The parenthetical already quoted — "the container error **(already central to Space Immanence v1.5 as diagnosis)** acquires a generative mechanism." This is pedigree, not evidence, but it is the one place the note wires itself to the claim graph, and "acquires" primes exactly the transfer §1.1 blocks. When public, rewrite to fix the direction: *"a candidate generative mechanism for the intuition the v1.5 diagnosis targets; confirming the mechanism would not confirm the diagnosis, and the two are graded separately."*

2. The sources line — "Okumura, *Realizing Genjōkōan* (provenance of the illumination/darkening framing)." An authority/aesthetics lean, not an evidential one; §4 treats it. Note also "pre-registered survey, **swarm-study mould**" (Study B) borrows the swarm instrument's earned credibility — acceptable as method citation, but the swarm's credibility is *procedural* (prereg, ledger, published correction); the borrow is only honest if the container programme adopts the whole procedure, including the public results ledger and drawer state (§2).

### 1.3 The one legitimate edge, argued exactly

The site's evidence semantics: "A future result counts toward a claim only if it tests a prediction derived from that claim specifically." S1's own falsifier already wires it to container-intuition empirics: "The dissociation study finds container intuitions and phenomenal-realism intuitions move together..." — i.e., the site has *already committed* that patterns in container-intuition data can count **against** S1. Evidence channels must be symmetric or they are a one-way ratchet, which is its own dishonesty: if B finding the container factor inseparable from body/frame/vocabulary would damage S1's weak reading (it would — the "relevant cognition" would contain no distinct container schema to be load-bearing), then the mirrored positive result must be allowed to count **for** the weak reading, at matching, explicitly capped grade.

So: **one edge, B → S1, both directions, weak reading only, supporting-not-establishing, pre-registered arm only.** Grade cap: S1 stays at 0.8 on confirmation (the number already prices the weak reading; confirmation de-risks it, it does not raise the strong reading's standing). And the symmetric commitment must be published *in advance*: if B fires drawer-condition 1 or 3, that is logged as negative evidence on S1's weak reading in the results ledger — the firewall blocks credit-laundering, not blame. **C transfers to no node in either direction as support** (mechanism confirmation is rival-compatible, §1.1/M1), with one exception worth pre-registering: a *hostile* C-result (non-medium optimum) is mild adverse evidence for nothing in the graph — the graph nowhere asserts the mechanism — which is precisely why C-confirmation can support nothing in it either. A (MPE reanalysis) has no kill authority by the note's own statement, so A-results must never appear on any claim page as support; pin that with a test (§2.4).

### 1.4 A gap in the "unclaimed move" claim

The note: "the container intuition does not appear in Chalmers's catalogue of problem intuitions at all" — verified as consistent with his taxonomy. But among Chalmers's canvassed *solutions* sits "introjection and the phenomenological fallacy" (Place 1956): the mistake of treating properties of the representation as properties of an inner field. That is container-adjacent machinery in the canonical catalogue. The specific derivation (empty surrounding medium *from* the unpresentable remainder, under compression) does look unclaimed; the broader "container as introspective artifact" is foreshadowed. The public version should cite Place and narrow the priority claim to the derivation, or a referee will do it less kindly.

---

## 2. Quarantine design

### 2.1 Principle

Container nodes must **not** enter `claims[]`. Three mechanical reasons on top of the epistemic one: `test/claim-graph.test.mjs` and `test/claims-json.test.mjs` assert `claims[]` is a verbatim mirror of `ClaimGraph.NODES` (12 nodes, pinned tier counts 4/5/3); the invariant ("no strong or medium claim depends on a speculative claim") is stated over that graph and a "pre-theoretical" tier would muddy it; and S1's confidence is test-pinned at 0.8 (`claims-json.test.mjs`), which is already your best anti-inflation guard — keep it.

Recommendation: a **separate machine object**, `site/container.json`, with its own node namespace, own version counter, own results ledger, and an explicit two-directional transfer contract. A one-line pointer from `claims.json` (`relatedProgrammes`) and from `llms.txt`. Mixing programme objects into claims.json invites exactly the adjacency the firewall exists to prevent.

### 2.2 JSON shape (sketch)

```json
{
  "artifact": "Container Problem programme — machine layer",
  "programme": "container-problem",
  "status": "private",
  "statusVocabulary": ["private", "public-pilot", "active", "drawered", "promoted"],
  "version": "0.1",
  "framework": {
    "name": "Space Immanence",
    "frameworkVersion": "1.6",
    "relation": "Shares the container diagnosis as subject matter, not as support. Publishing or confirming results here changes no tier, no confidence number, and no claim text unless a transferRule fires."
  },
  "vocabulary": {
    "darkening": "Placeholder for a formal object not yet in hand (author's own caveat). Instruments, item wording, and preregistration documents use neutral operational terms only.",
    "provenance": "Okumura, Realizing Genjōkōan — aesthetic lineage, explicitly not evidence; see namingRisk."
  },
  "nodes": [
    { "id": "C1", "label": "Existence: a non-spatial container factor separable from body-boundedness and perceptual spatial frame",
      "tier": "pre-theoretical", "killStudy": "B",
      "falsifier": "No separable third factor under the pre-registered discriminant-validity criterion, in the experienced strata (see drawerConditions[0])." },
    { "id": "C2", "label": "Trajectory: point-observer → container → collapse, ordered by introspective experience",
      "tier": "pre-theoretical", "dependsOn": ["C1"], "killStudy": "B",
      "falsifier": "Failure of the pre-registered ordered contrast naive < trained; collapse limb tested separately (drawerConditions[1])." },
    { "id": "C3", "label": "Mechanism: medium-shaped optimum (homogeneous, surrounding, empty — formally defined) for encoding the unpresentable remainder under compression",
      "tier": "pre-theoretical", "killStudy": "C",
      "falsifier": "Non-medium optimum across the pre-registered constraint family (drawerConditions[3])." },
    { "id": "C4", "label": "Dissociation: container-report stability tracks figure-ground manipulation, not body-model manipulation",
      "tier": "pre-theoretical", "dependsOn": ["C1"], "killStudy": "D",
      "falsifier": "Crossed magnitude predictions fail at pre-registered power." }
  ],
  "noTransfer": "No result in this programme attaches to any Space Immanence claim node (S1–S4, M1–M5, P1–P3), open hinge, or tier unless an entry in transferRules names the edge. In particular: confirmation of C3 is compatible with illusionism and the attention schema theory and supports neither S1's strong reading, nor M1, M2, M5, P1, nor P2 — for P2 it functions as a deflationary rival, not support. Treating a container-programme result as evidence for the fold is the exact conflation the swarm instrument's anti-goals forbid, in a nearer disguise.",
  "transferRules": [
    { "from": "C1", "to": "S1", "direction": "both",
      "reading": "weak only (container schema load-bearing in the relevant cognition)",
      "grade": "supporting, never establishing; S1.confidence stays ≤ 0.8 on confirmation",
      "condition": "Pre-registered confirmatory arm of Study B only. Symmetric commitment: if drawerConditions[0] or [2] fires, that is logged in this file's results[] AND surfaced as negative evidence on S1's weak reading in the framework's results ledger and changelog.",
      "mechanism": "On firing, an entry is added to claims.json evidence[] with attachesTo:'claims', bearsOnClaims:['S1'], carrying grade and nonTransfer text per the existing schema; no silent confidence edit." }
  ],
  "drawerConditions": [ "...tightened wording from §3, each with a machine-checkable 'criterion' string..." ],
  "results": [],
  "resultsLedgerContract": "Same shape and honesty rules as swarm-results-ledger.json: preregistered flag, status vocabulary, exploratory labelling, outcome text, public correction discipline."
}
```

### 2.3 state.json implications

Add one block; keep the framework version untouched (the v1.6 precedent — "a discipline-arm release: no metaphysical change" — is the template: container releases must never bump the framework version or tier counts):

```json
"programmes": {
  "container-problem": {
    "status": "private",
    "version": "0.1",
    "latestResult": null,
    "drawered": false,
    "transfersFired": []
  }
}
```

### 2.4 Drift-test extensions (concrete)

New suite `test/container-programme.test.mjs`:
1. **Namespace quarantine:** no `^C\d+$` id appears in `claims.json` `claims[]`, any `dependsOn`, any `openHinges[].targets`, or `ClaimGraph.NODES`. (The mechanical form of the firewall.)
2. **No-transfer clause present and thick:** `noTransfer.length > 200`; every `transferRules[].to` is a real claim id; every rule carries `direction: "both"` and non-empty symmetric `condition` text — one-way (credit-only) rules fail the test.
3. **Drawer conditions fireable:** ≥ 4 entries, each with a `criterion` string naming a statistic and threshold (`length > 60`), mirroring the "a test that cannot fail is not a test" assertion already in `state-drift.test.mjs` for obj11.
4. **Study A containment:** no page and no ledger entry may cite the MPE-92M reanalysis as support for any C-node or claim node — pin as the AUC-retraction test pins phrasing: assert no surface matches `/MPE[- ]92M[^.]*(supports|confirms|evidence for)/i`.

Extensions to `test/state-drift.test.mjs`:
5. Programme status agrees across `state.json` ↔ `container.json` ↔ the container page's status line (mirror of the hinge-status test).
6. While `status` ∈ {public-pilot, active}: the container page and `llms.txt` must contain the no-transfer sentence (pin a stable sentence verbatim, exactly as the swarm firewall sentence works on agents.html).
7. If `transfersFired` is non-empty: the corresponding `evidence[]` entry exists in claims.json with `attachesTo:"claims"`, `bearsOnClaims:["S1"]`, and the changelog names it (mirror of the latestEvidence test).
8. Keep — do not weaken — the existing `claims-json.test.mjs` pin `S1.confidence === 0.8`. Any container-driven change to that number must arrive as a deliberate two-file edit that turns a test red first. That is the anti-inflation ratchet working as designed.

---

## 3. Falsification-condition audit

The note's drawer clause: "The idea goes in the drawer if any of the following hold." Disjunctive kill is the right shape. Each condition, graded for decisiveness (can it actually fire?), testability as stated, and escape hatches:

**Condition 1 — "no separable container factor in B after controlling for body and frame."**
*Decisiveness:* moderate. *Testability:* under-specified — "separable factor" is analyst-dependent (extraction method, rotation, loading and cross-loading thresholds, discriminant criterion all unfixed). *Escape hatches:* two. (a) The item-blame hatch: the note commands item-level attack before fielding — good — but that legitimizes the post-hoc move "the items were bad, refield." (b) **The population hatch, which is structural:** the primary arm is non-meditators, but the note's own trajectory says the container "should only appear once attention begins interrogating the remainder." A null in the naive arm is therefore *predicted by the theory being tested*, and the author can honestly decline to fire the condition. As stated, condition 1 may be unfireable exactly where the data will first arrive.
*Tightened wording:* "The existence claim dies if no separable third factor emerges **in the introspectively-experienced strata** under the pre-registered discriminant criterion (CFA nested-model comparison with thresholds fixed before fielding), with the item pool frozen after one adversarial revision round; a second null after that single permitted revision fires the condition regardless of item-quality objections. Separately: the factor appearing **equally strongly in naive strata** fires the trajectory kill (condition 2), not a save."

**Condition 2 — "container loading flat or decreasing across introspective-experience strata."**
*Decisiveness:* **broken as stated — it contradicts the theory it polices.** The trajectory predicts rise then collapse; a *decreasing* loading in the highest-practice (MPE-adjacent) stratum is the theory's own prediction, yet the condition counts "decreasing" as a kill. As written the author can be forced to drawer a confirmed theory, which means in practice the condition will be reinterpreted post hoc — the worst kind of leak, an ambiguity that guarantees a judgment call at fire time. Also "flat" is undefined: a non-significant trend in an underpowered study is not "flat."
*Tightened wording:* "The trajectory dies if the pre-registered ordered contrast (naive < intermediate-trained) fails at the pre-set minimum effect size, or if equivalence testing (TOST at a pre-registered SESOI) shows the naive-to-trained slope is flat. The collapse limb (decrease in the MPE-adjacent stratum only) is a separate pre-registered prediction whose failure kills the collapse limb, not the existence claim. Cross-sectional survivorship is acknowledged as unresolvable in B; the condition fires on B's data anyway."

**Condition 3 — "container reports tracking metaphor inventory across languages rather than experience."**
*Decisiveness:* weak as stated. *Testability:* "tracking X rather than Y" is a model comparison with no specified models, and the predictors are confounded (languages with contemplative traditions have both the metaphor inventory and the practitioners). *Escape hatch — the fatal one:* the cross-linguistic arm is "**where feasible**." A kill condition gated on an optional study arm is decorative; if the arm is never fielded the condition can never fire.
*Tightened wording:* "The cross-linguistic arm is mandatory before the existence claim exits pilot grade. Pre-registered model comparison: container-factor loading regressed on individual introspective experience versus a language-level MIND-IS-A-CONTAINER productivity index constructed (Kövecses-style, from corpora) and frozen before any B data are seen. The condition fires if the language-level predictor dominates under the pre-set model-selection criterion and the individual-level predictor adds no explanatory power at the pre-registered threshold."

**Condition 4 — "a non-medium-shaped optimum in C under adversarially fixed constraints."**
*Decisiveness:* lowest of the four; the note itself says "'cheapest' in C is rigging-prone." *Testability:* three undefined load-bearing terms — "medium-shaped" (the three signatures are named but not formalised), "optimum" (global? approximate? robust?), and "stated compression constraints" (stated by whom, revisable when?). *Escape hatches:* (a) a hostile optimum blamed on wrong constraints, then re-run — nothing forbids re-runs; (b) the author defines the encoding space, so medium-shape can be the only smooth option by construction — one red-team round over an author-defined space is thin protection; (c) post-hoc reading of any diffuse encoding as "medium." The note's own discipline ("probing LLMs for container talk is contamination") is good but guards a different door.
*Tightened wording:* "Before any run: the three signature predicates (homogeneous, surrounding, empty) are defined formally; a family of ≥ N cost functions and encoding spaces is fixed, at least half nominated by the red team; the whole constraint file is hash-committed publicly. The condition fires if the medium-shaped encoding fails to be optimal (under the pre-defined optimality criterion) in a pre-registered majority of the family. No constraint may be revised after any member has run; every run publishes, hostile or not."

**Missing from the set:** (i) no condition covers Study D at all, despite D carrying crossed magnitude predictions; (ii) no deadline or staleness clause — a drawer that can be indefinitely deferred is a shelf; (iii) "in the drawer" has no public semantics. When the programme goes public, drawering must be a published state (`status: "drawered"` in container.json/state.json, drift-tested), exactly as the swarm's nulls are published in the results ledger. A private drawer next to a public programme is the drift the state.json machinery exists to prevent. (iv) Per §1.3: firing conditions 1 or 3 must also be logged as negative evidence on S1's weak reading — the symmetric commitment, stated in advance.

---

## 4. Naming risk

"Darkening" is currently nowhere in the served corpus — `grep` finds no "darken"/"Okumura" in the working papers, references.bib, or site/. The import is therefore a *choice* the container programme is making now, and it imports two things at once: Okumura's (Dōgen's) authority and the framework's aesthetics. The site's readings.html already positions Dōgen as "the closest fit on the spacetime side the proposal has" and finds Genjōkōan "unusually close to the fold." Naming the mechanism from the *Genjōkōan* ("when one side is illuminated, the other side is dark") builds, in advance, exactly the coherence channel §1 exists to block: if the mechanism is confirmed, the confirmation arrives pre-dressed in the vocabulary of the framework's favourite classical ally, and readers will experience mechanism-confirmation as fold-vindication without a single evidential step. The transfer will happen in the connotation layer, beneath the firewall.

Places where the vocabulary pre-answers a question the studies are supposed to ask:

1. **"The unpresentable remainder" / "the residue."** Study B's whole purpose is to test whether the residue *exists* ("the strongest surviving objection: the residue may not exist"). Naming B's target "the remainder" before B runs presupposes B's answer. Concretely dangerous at the item level: items drafted from darkening vocabulary will manufacture the factor. The note correctly levels this exact charge at the MPE-92M ("built from contemplative vocabulary on trained populations, so neither outcome settles the existence question") — and does not notice that it applies to its own instrument. Rule: no mechanism vocabulary in item wording; items phrased in the three operational contrasts only.

2. **"Untrained cognition encodes the darkened remainder as a point-observer."** The trajectory paragraph describes the *data* (Starmans & Bloom; Bertossa; Cartesian Folk Theater) in mechanism terms, as if the mechanism were already operating in the findings. Those studies found a located egocenter; "darkened remainder" is interpretation wearing the costume of description.

3. **"Illumination" vs. the MPE-92M "Luminosity" factor.** Study A will harvest candidate items from factors including *Luminosity* — contemplative light-vocabulary — using a theory whose mechanism is named in light-vocabulary. Any apparent convergence between instrument and theory is partly shared lexicon. The note's own standard ("probing LLMs for container talk is contamination, not convergence") applies: Study A meeting its own metaphor in the factor names is the same contamination one step removed.

4. **"Figure-ground dissolution (Ganzfeld)" in Study D.** The manipulation is labelled with the mechanism, so a positive result is described as mechanism-confirmation before the inference is made. Call the arm "unstructured-field viewing" in the prereg; let "figure-ground" be a conclusion.

5. **"The container error."** In a note whose Studies B–D are officially neutral on whether the intuition is an error (they test existence, trajectory, mechanism — not falsity, and not S1's distortion claim), "error" pre-judges both the empirical question and the philosophical one the studies explicitly do not test.

6. **"No rival on the board predicts the full sequence."** The sequence is stated in the theory's own vocabulary, which partly explains why no rival "predicts" it; and the specific claim that Graziano's schema "has no reason to reshape with contemplative training" is unverified and doubtful (AST's schema is plastic by design). Before the trajectory is billed as the discriminating signature, state the rivals' predictions in *their* vocabularies, or the discrimination is partly lexical.

**Recommendation.** Two-register discipline, enforced: "darkening" stays an internal placeholder (the note already says so — "a placeholder for a formal object we do not yet have" — extend that from prose to practice); preregistrations, instruments, and the public machine layer use neutral operational names; the public page states the Okumura provenance as aesthetic lineage with an explicit disclaimer (the `vocabulary` block in §2.2); and the drift test in §2.4(6) pins the disclaimer to the page.

---

## Appendix A — verification ledger

| Claim in note | Status | Source |
|---|---|---|
| Starmans & Bloom 2012, self at the eyes, alien with chest-eyes | Verified | Cognition 123(2):313–318; minddevlab.yale.edu PDF |
| Bertossa et al. 2008 "Point Zero", blind subjects, head/behind-eyes | Verified | Perceptual & Motor Skills 107:323–335; PubMed 19093595 |
| "Cartesian Folk Theater", confined-brain-region localisation | Verified (headline N≈2,057 not independently confirmed) | Forstmann & Burgmer, JEP: General; PubMed 34928681 |
| Gamma & Metzinger MPE-92M, PLOS ONE 2021, factor names as quoted, OSF data | Verified (factors: Witness Consciousness; Emptiness and Non-egoic Self-awareness; Sensory Perception in Body and Space; Luminosity all real) | journals.plos.org 10.1371/journal.pone.0253694 |
| Sytsma & Machery 2010; Sytsma & Ozdemir 2019 "No Problem" | Verified (titles/venues consistent) | PhilPapers/JCS listings |
| Chalmers meta-problem: no container intuition in the catalogue | Consistent with verified taxonomy; but Place's "phenomenological fallacy" appears among canvassed solutions — priority claim should be narrowed (§1.4) | consc.net (403; verified via secondary: Imprint Academic editorial, "How Can We Solve the Meta-Problem", josephlevine.net) |
| Okumura, Realizing Genjōkōan, illumination/darkness passage | Verified | Wisdom Publications; Genjōkōan commentary sources |
| Graziano "ethereal essence vaguely localisable in space" | Verified near-verbatim ("an ethereal thing... vaguely localisable to space inside us") | Prospect Magazine profile; Graziano Lab AST papers |
| "Graziano's schema has no reason to reshape with contemplative training" | **Unverified / contestable** | — |