SuperGrokTOE · premise register · S371a
Premises
What this surface is. A register of the premises this programme has DECLARED — postulates, adopted premises, measured inputs, conventions, loaded correspondences, fences and open requirements. Every statement is quoted from its source — verbatim from the sealed text, or transcribed by the house from the TeX where the extraction was damaged, as each page's statement-quality row says — and carries that source's locator. No premise on this page has been adjudicated by the house, and none is defended here. A premise is what an argument RESTS ON, not what it establishes.
33 premises are registered. 13 of them are reached by nothing in the corpus — not because nothing depends on them, but because only 35 of the 202 CLAIM records (17.3%) declare any dependency at all.
★ CORRECTION, S315. This page printed “16 of them are reached by nothing”, and printed “none” in the reached transitively column for PRM-023 and PRM-024, until 2026-09-08. The honest figures are 14, and 1 each.
premise_closure was computed once at S312 20:26Z; the LIB2-019 → PRM-023 and LIB2-020 → PRM-024 edges were declared at 21:57Z, ninety-one minutes later, and the field was never rebuilt. ★★ Worse than the numbers: the box below this one explained those two “none”s by saying “because the record declaring it has no dependency chain of its own”. LIB2-019 and LIB2-020 each carry uses_input → LIB2-018. They have a dependency chain; the sentence was a reason invented for an artefact. It was written at S313 to explain the residue of a defect the Grok lane had just made the house fix — the house fixed the column and then explained what was left over wrongly. The general sentence stands; the two named records are gone from it. Found by the promise-sweep gate (scripts/promise_sweep.py), S315. The root cause was that the closure had no re-runnable derivation at all: premise_edges_apply.py is a one-shot S312 migration that hardcodes its session, repopulates a deprecated field and rewrites all 357 records, so it could not be run again. scripts/premise_closure_recompute.py now exists and wrote exactly the two records that moved. Nothing above is edited away; this note is the record that it changed.★ CORRECTION, S314. The paragraph above printed “only 23 of the 202 CLAIM records (11.4%)” until 2026-09-07. The query behind it counted
depends_on and uses_input and omitted rests_on_premise — the one edge type this surface exists to display. Six CLAIM records carried an explicit rests_on_premise edge to a premise in the table below and were counted as declaring nothing: LIB2-018, LIB2-108, LIB2-122, LIB2-173, LIB2-182, LIB2-356. The honest figure is the one now printed. The same query produced a second wrong number on the graph page; the ChatGPT lane found one instance by arithmetic on a dispatch (D1262) and the Grok lane found the other on the live page (D1264). Both builders are fixed. The sentence above is the corrected one; this note is the record that it changed.These are incidence figures, not completeness. The fraction of records carrying a dependency edge says nothing about the fraction of their actual dependencies that were recorded. Even complete incidence could omit a premise shared by every claim, and its apparent exposure would still read as zero. Nothing on this surface bounds what was never written down — which is why no risk, blast-radius or isolation number is published here. 0 records have had their dependency enumeration reviewed.
An empty premise closure means UNKNOWN, not INDEPENDENT. A claim that reaches no premise has not been shown to rest on none. It has declared nothing.
★ The two count columns answer different questions and must not be read as one. Declared directly is how many records carry an explicit
★ The two count columns answer different questions and must not be read as one. Declared directly is how many records carry an explicit
rests_on_premise edge to this premise. Reached transitively is how many records reach it by walking dependencies as well. A premise can be declared by a record and reached by nobody — that happens when the record declaring it has no dependency chain of its own, so nothing walks into it. Until S313 this table printed only the second number under a heading that promised the first, and so printed “none” for two premises a record explicitly rests on. Found by the Grok lane, cold, at D1263; confirmed by the house.Read no position on this surface as a warrant.The register
| id | premise | kind | source | declared directly | reached transitively | details |
|---|---|---|---|---|---|---|
| PRM-001 | Neutral signature | ADOPTED_PREMISE | suite p.16 | none | none | note |
| PRM-002 | Split Jordan carrier J3(O_s) (indefinite trace form) | ADOPTED_PREMISE | suite p.247 | 1 | 1 | note · corrected (R81 D1) |
| PRM-003 | AX-COT | POSTULATE | suite p.318 | none | none | note · corrected (R81 D2) |
| PRM-004 | AX-MG | POSTULATE | suite p.300 | none | none | note · corrected (R81 D3) |
| PRM-005 | AX-MOS | POSTULATE | suite p.316 | 1 | 3 | note · corrected (R81 D4) |
| PRM-006 | AX6_pol — polarisation postulate (weak-angle orbit selector) | POSTULATE | suite p.77 | 1 | 1 | note · corrected (R81 D5) |
| PRM-007 | Action shape (the adopted principle Φ_diag) | ADOPTED_PREMISE | suite p.314 | none | none | note · corrected (R81 D6) |
| PRM-008 | Alpha as declared input | MEASURED_INPUT | suite p.264 | 1 | 2 | note · corrected (R81 D7) |
| PRM-009 | Basepoint / tensor equivalence | FENCE | suite p.332 | none | none | note · corrected (R81 D8) |
| PRM-010 | Boundary datum | LOADED_CORRESPONDENCE | suite p.175 | 1 | 1 | note · corrected (R81 D9) |
| PRM-011 | Colour real form as declared input | ADOPTED_PREMISE | suite p.7 | 1 | 2 | note · corrected (R81 D10) |
| PRM-012 | Correspondence map algebra to SM | LOADED_CORRESPONDENCE | suite p.2 | 3 | 10 | note |
| PRM-013 | DET-7 structural postulate | POSTULATE | suite p.171 | 5 | 8 | note · register note |
| PRM-014 | Finite regulator | CONVENTION | suite p.311 | 2 | 2 | note · corrected (R81 D11) |
| PRM-015 | Module-naturality premise | OPEN_REQUIREMENT | suite p.365 | none | none | note · corrected (R81 D12) |
| PRM-016 | Parity pattern is not a spin structure | FENCE | suite p.284 | none | none | note · corrected (R81 D13) |
| PRM-017 | CKM first-row Route B (selected against kaon data) | LOADED_CORRESPONDENCE | suite p.49 | 3 | 3 | note · corrected (R81 D14) |
| PRM-018 | X_vac = diag(φ², 1, 0) — legacy vacuum normalization (house file) | POSTULATE | house file | none | none | note · corrected (R81 D15) |
| PRM-019 | m_tau as registered input | MEASURED_INPUT | suite p.339 | 3 | 4 | note · corrected (R81 D16) |
| PRM-020 | Cascade prescription | POSTULATE | house file | none | none | note · corrected (R81 D17) |
| PRM-021 | Power assignment n = 4, 5 | POSTULATE | house file | none | none | note · corrected (R81 D18) |
| PRM-022 | DSE / Faddeev closure | POSTULATE | house file | none | none | note · corrected (R81 D19) |
| PRM-023 | Route A inversion symmetry (palindromic spectrum) | ADOPTED_PREMISE | suite p.168 | 1 | 1 | note |
| PRM-024 | Route B golden-unit restriction (positive algebraic integers in Q(√5), abc = 1) | ADOPTED_PREMISE | suite p.171 | 2 | 2 | note · corrected (R81 D20) |
| PRM-025 | AX1 — golden unit | POSTULATE | suite p.10 | none | none | note · registered (R81 D15) |
| PRM-026 | Det² selector object (historical; retired as a selector of AX6_pol) | POSTULATE | suite p.77 | none | none | note · registered (R81 D5) |
| PRM-027 | W01 — ultraviolet boundary attachment (y_t(M_Pl) = 1) | LOADED_CORRESPONDENCE | suite p.145 | 1 | 1 | — |
| PRM-028 | v_EW = 246.22 GeV — electroweak calibration anchor | MEASURED_INPUT | suite p.355 | 1 | 1 | — |
| PRM-029 | Λ_G₂ ≈ 260 MeV — external dimensional calibration | MEASURED_INPUT | suite p.36 | 2 | 3 | — |
| PRM-030 | Koide relation K = 2/3 — external empirical selection rule | ADOPTED_PREMISE | suite p.328 | 1 | 1 | — |
| PRM-031 | Normal ordering (the neutrino mass-ordering selector) | ADOPTED_PREMISE | suite p.82 | 1 | 4 | — |
| PRM-032 | Dirac-neutrino extension (the Dirac-ν benchmark assumption) | ADOPTED_PREMISE | suite p.88 | 1 | 2 | — |
| PRM-033 | Propagation interface (the R_d ladder read as propagation eigenstates) | LOADED_CORRESPONDENCE | suite p.83 | 1 | 4 | — |
Premises once listed here as used but not registered (7, registered at R84)
From S312 to S328 this section listed premises the corpus uses that were not in this register. The PI ruled them registered (R84, S328). Each row names the register entry, the records the house declared to rest on it, and the pairings it refused with the reason (an edge exists only where a record’s stated result assumes the premise).
| premise | held as | declared to rest on it | refused |
|---|---|---|---|
| PRM-027 W01 — ultraviolet boundary attachment (y_t(M_Pl) = 1) | W01 ultraviolet boundary attachment | LIB2-121, LIB2-139 | LIB2-120: The record STATES the attachment ("Reading the carrier eigenvalue as yt(MPl) = 1 is the declared W01 attachment, requiring …"; domain "The W01 ultraviolet boundary attachment and its three unproved identifications"). A record that states a premise does not rest on it (R78-a). It is named in PRM-027's curation note instead. Its dependents LIB2-121 and LIB2-139 carry their own edges. |
| PRM-028 v_EW = 246.22 GeV — electroweak calibration anchor | vEW = 246.22 GeV as a measured input | LIB2-121 | LIB2-124: A usage inventory: the record reports that the chains use three anchors (domain "The formula-usage inventory, not a derivation of the dimensional inputs"). Its result does not assume the value of v_EW; it lists it. Same shape as R78-b LIB2-182 and R82-b LIB2-174. Records that consume the anchor carry their own edges. |
| PRM-029 Λ_G₂ ≈ 260 MeV — external dimensional calibration | Lambda_G2 ~ 260 MeV, the fitted hadronic anchor | LIB2-128, LIB2-129 | LIB2-124: As LIB2-124 → PRM-028: an inventory of anchors, not a result that assumes Λ_G2. Refused under R78-a. LIB2-128 and LIB2-129 carry the edges. |
| PRM-030 Koide relation K = 2/3 — external empirical selection rule | The external empirical Koide relation K = 2/3 | LIB2-127 | none |
| PRM-031 Normal ordering (the neutrino mass-ordering selector) | Normal ordering (the ordering selector) | LIB2-060 | none |
| PRM-032 Dirac-neutrino extension (the Dirac-ν benchmark assumption) | The Dirac-neutrino extension (Dirac-nu selector) | LIB2-108 | none |
| PRM-033 Propagation interface (the R_d ladder read as propagation eigenstates) | The propagation interface | LIB2-060 | none |
How the edges were made
The 33 record→premise edges are editorial house judgments, each read from the record's own statement, hypotheses and domain, and each carrying its reasoning. They are typed rests_on_premise, never depends_on: a record→record dependency and a record→premise grounding are different relations, and merging them would let the coverage figure be inflated by the very edges just added. They are not extracted from the suite and they are not a lane's claim.
Generated from library_v2.jsonl + PRM_REGISTER_S312.jsonl. Declared volatile fields (R36): BUILD_STAMP.
BUILD_STAMP S371a · 2026-09-26 22:53Z · cut 9dcc6ce9a8ee3e02