Removal of the spatial cutoff for weakly coupled P(phi)_2

Status

What is proved, what is conditional, and what is open. Nothing on this page is a summary of intent: each line names the statement or the gate it comes from.

Part I — conditional, and the condition is named

Theorem 12.2 (Main Theorem) is proved as an implication from the imports A1–A8 enumerated in §2.4. It is not a proof of those imports. Seven of the eight are literal quotations of printed theorems. The eighth is not:

A8(vi), written U_GJS — the cutoff-uniform form of the Glimm–Jaffe–Spencer moment estimate (Ann. of Math. 100 (1974), Thm. 1.1.8) on the slab family used here.

Since the scope amendment of 5 September 2026, U_GJS is documented as imported at the scope at which the source itself states and uses it, with the passages on pp. 594–597 and 629 quoted verbatim in A8(vii′), rather than as a quantifier strengthened beyond the source. It remains a named hypothesis of Theorem 12.2. The manuscript does not re-derive the cluster expansion: Appendix C audits the source mechanism and proves two analytic prerequisites only. See imports.

Of the 61 numbered statements, 24 reach A8 through their own proof chain and therefore inherit this conditionality; the rest do not. Which are which is computed, not asserted — see the imports A8 column of the result inventory. The inverse-energy filter (Lemma 9.1), for instance, is independent of A8 entirely.

Negative scope, stated by the manuscript itself

No uniqueness of ω_∞ among all locally normal ground states. No sharp exponential rate. No claim outside the weak-coupling region.

Gate record (Part I)

gate requirement status
G0 compiles; every cite/ref resolves; every assumption has a “Used in” line passed 10 Aug 2026
G1 / G2 every ⟨i⟩⟨j⟩ step of the Lamport sources maps to a unique environment or display; no unmapped step, no orphan passed 10 Aug 2026
G3a / G3b constants recomputed by hand; forbidden-string check passed 10 Aug 2026
G4 citation ↔ dossier bijection; regression greps passed 10 Aug 2026
G5 3 × pdflatex, zero unresolved; every page rendered and inspected passed 10 Aug 2026
G6 adversarial pass; zero confirmed findings, or all fixed and re-verified passed 10 Aug 2026

The full records, including the findings G6 raised and their dispositions, are in part_i/README.md.

Open audit item C5

No second human reader has yet checked the source page images against the transcriptions in §2. The itemized checklist is part_i/C5_reading_record.md; all nine verdict fields are blank. Items 8–9 were added on 5 September 2026 for the affine-path application of Theorem 1.1.7 on p. 595 and the derivation of Theorem 1.1.7 from Theorem 1.1.8 on pp. 596–597.

Closing C5 does not change the status of U_GJS. It is a check of transcription fidelity, not of the mathematics.

Part II — open at gate (II.11)

Nothing in Part I depends on Part II. The preparatory scoping pass is part_ii/gate_II11_scoping.pdf; phases V0–V4 are recorded in strategy.md.

claim status
finest-scale diagonal norm convergence disproved
exact finite-torus interacting model and ground state constructed
interacting fixed-coarse projective cluster state along a cofinal subnet proved unconditionally
MMST free fixed-coarse / projective convergence proved in the source, strengthened here to an explicit D4 power rate
exact free gap, group-velocity bound, Wick logarithm proved in the Lamport note
Battle–Federbush as a source for OAR-uniform CE1/CE2 not available — the cited note treats different phase-cell variables
full-sequence interacting convergence and identification with the continuum P(φ)₂ state open, exactly at the coarse Weyl characteristic-functional limit
interacting scale-uniform gap, local second moment, Lieb–Robinson open
the cited Nachtergaele–Raz–Schlein–Sims theorem inapplicable to the Wick quartic: it assumes V′ ∈ L¹(ℝ)
the former path (1−s)H_N + sH undefined without a common Hilbert-space/form realization; removed from the corrected program

The record itself

This is an AI-assisted research record and an experiment in open science. Nothing in it has been refereed. The author’s own line-by-line check of the machine-assisted material is still in progress. Every phase, gate and adversarial round is published in strategy.md, including the rounds that found errors in earlier rounds’ repairs.