*(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.)*

# Response to the Author of "The Container Problem"
**From the red team · Space Immanence · 14 July 2026**
**Referent: your working note v0.1. Full findings: `The_Container_Problem_Redteam_v0.1.md` and the three companion documents. This is the cover letter: what happened when we did what you asked.**

---

You circulated the note to be attacked, not admired. We attacked it with six lenses, merged twenty-two attacks into eighteen, and put every serious one in front of referees instructed to refute it, with the rule that literature claims had to be verified, not remembered. Here is the outcome, stated the way you'd state it.

## The note did not die

No attack killed the programme. Four survived every priority assault we could mount, and you should know exactly how narrow the survival is, because it is your remaining asset: **nobody derives a homogeneous-surrounding-empty medium from the remainder of selection under resource constraints.** We searched hard for a prior claim. Metzinger derives empty-space phenomenology from the system modelling its own *capacity* (the MPE programme you cite has a positive generative model — the epistemic-openness work — that your note never engages). The Projective Consciousness Model derives the *perspectival frame* — content-structured, neither homogeneous nor empty. Jaynes derives the container from *language*. Dennett diagnoses the theater without a sub-personal generative mechanism. Place named the fallacy in 1956. Five near-claims, one clearing, one field wide. It is real, and it is yours — but only under the narrow description. Every broader sentence in your priority case ("unaddressed problem intuition," "nowhere as derived mechanism," "no rival predicts the full sequence") is false as written, and the report documents each falsifier.

Your concessions were load-bearing. Every attack drafted against ground you had already yielded scored zero and was retired. The kill-check section did exactly what it was built to do. Keep writing that way.

## Your "Attack here first" predictions, scored

You predicted three attack surfaces. Here is how your self-model fared:

1. **"The residue may not exist."** Correctly predicted as the strongest objection, and it is where the sharpest design work now sits. The items document drafted your instrument and then killed a fifth of it — including NSC6, a gap-observation item that is a disguised pointing-out instruction and would have produced your predicted practice-correlated loading *artefactually*. The construct resists itemisation: two clean container items survived out of eight. That thinness is a finding. The decision table separating "no factor exists" from "our items failed" is built; without it, your drawer-condition 1 was unfireable.
2. **"The trajectory could be a vocabulary artefact."** You under-called this one. It is worse than vocabulary: it is *curriculum*. The classical formless-attainment sequence — point, boundless space, beyond space — is a two-millennia-old syllabus, taught procedurally, cross-linguistically (the MPE sample anchoring your collapse endpoint is ~80% Vipassana and Zen practitioners across five languages). Your cross-linguistic vocabulary control is structurally blind to a procedural confound delivered identically in every language. And Jaynes upgrades vocabulary to *construction*: if metaphor builds the phenomenology, "metaphor inventory" and "experience strata" are the same variable and B's dissociation collapses. The one test that discriminates you from both rivals — practitioners of ground-interrogating lineages that never use space instruction (pure noting, kōan work) should still develop the container stage on your account, and should not on theirs — appears nowhere in your note. Commit to it in v0.2 or the trajectory is a fact about syllabi.
3. **"'Cheapest' in C is rigging-prone."** Confirmed, and answered. The constraint specification you assigned to this red team is written and ready to freeze: ten encoding families, four cost functions with two-sided bias analyses, twelve experimenter degrees of freedom with binding mitigations, executable signature tests, ties resolving against you. One result from that work you must absorb now, before any formalization: **compression alone can never select the medium.** A medium is a scalar plus an index set plus enclosure topology; every cost function strictly prefers the scalar. Only task demands can produce your optimum — which makes task-ensemble selection the master degree of freedom, and makes your sentence "under compression constraints the default encoding... is a homogeneous, surrounding, empty medium" false as literally stated. The claim survives only in conditional-existence form. The spec has the honest reformulation; take it or strike C's confirm role.

## What you did not predict

The largest finding was not on your list, because it is inside the note rather than in the literature: **your own flip contradicts your own mechanism paragraph, and your falsification conditions are wired to the version you abandoned.** "By default" and "only once attention begins interrogating the remainder" cannot both stand. Study C as specified tests the default claim; drawer-condition 4 fires on exactly the outcome your trajectory version predicts for the unconditioned case. As written, either you drawer a claim your own theory says should fail that test, or condition 4 never binds. A test that cannot fail is not a test — your project wrote that sentence before you did.

And here is the part you should find encouraging, because we did: three lines of attack that never touched each other converged on the same repair. The internal-logic referee, the scalar-dominance analysis in the C spec, and the clinical record on how brains actually handle their representational limits (the blind spot is completed, not darkened; the neglected hemifield has no edge; Anton's patients confabulate rather than perceive a void — the empirical default for the unrepresented is *nothing*) all force one rewrite: **point-index by default; medium conditional on the remainder becoming the target of modelling.** The absence-perception literature then hands you a gift your note didn't know it had: absence is positively represented only under an actively deployed template — which is your flip, stated as someone else's finding. The conditional mechanism predicts a *transition* (the C spec's Appendix A gives it a parameter), and a predicted transition is a sharper, more falsifiable object than a default. v0.1 is wrong in a way that makes v0.2 stronger. That is the best outcome an attack can produce.

## Also broken, fix cheaply

- Stage 1 rests on folk-*theory* data (drawing consciousness on brain images; third-party closeness judgments) that your own Sytsma citation disqualifies — and the Cartesian Folk Theater is a located *theater*, so the citation cuts both ways. Type your evidence; restate the point as a prediction B's free-report arm must confirm.
- B's trajectory kill-condition is the survivorship confound you yourself used to demote Study A, one paragraph earlier — and drawer-condition 2 counts "decreasing in the top stratum" as a kill when your own theory predicts it. Restrict B's kill authority to existence, or commit to the longitudinal arm.
- "Cannot itself be shown" is false against James's fringe and Gurwitsch's field; weaken to "cannot be shown as figure" and specify where the structured near-remainder hands off to your allegedly medium-encoded far one.
- All four drawer conditions leak; the governance document has tightened wordings, plus the conditions you forgot (Study D, staleness, public drawer semantics, and a falsification condition for the priority claim itself — your method for negative existentials has failed you twice by your own narrative; it does not get a free pass the third time).
- Before anything goes public: the quarantine. One sentence in your note wires the programme to the claim graph ("the container error... acquires a generative mechanism") and it flows credit the wrong way. And the transfer nobody would think to block: a confirmed darkening mechanism is a deflationary *rival* to P2, not support. The firewall must block the negative direction too.

## What we ask of v0.2

The report's §8 has the full eleven points in priority order. The non-negotiables: the conditional mechanism rewrite with the two-regime C; the tradition-contrast commitment; the neighbour map with per-neighbour discriminating sentences; a priority falsification condition with a stated search protocol; and the C constraint file frozen — hash-committed — before a single line of formalization exists.

You wrote a note that made a decisive review possible: commitments stated plainly, kill authority assigned, concessions made in advance. That discipline is why we can tell you precisely what is dead, what is wounded, and what is yours. The clearing is one field wide. Fence it honestly and it might hold.

— The red team

*(Adjudication trails, verification statuses, and the disclosure of which sections carry reduced independent review are in the main report.)*