# n207 — The audit ledger II: passes 4–7 (§§23–56), each claim accepted or challenged

**Date:** 2026-07-11 · **Method as n206:** every claim identified; executables spot-checked before adjudication (n123's U2 gate verified unable to fail; n175/n159 executables verified absent; n113 verified hanging). A = accept · A* = accept with nuance · CH = challenged.

## Pass 4 — anharmonic layer, Mercury, pump (§§23–29)

| § | Claim | Verdict |
|---|---|---|
| 23 | Derivative IBP identity exact; but η_N = 0 is a lattice identity, not the relativistic Nordtvedt parameter — LLR/MICROSCOPE are not "structural passes" of it | **A** |
| 24 | The published Mercury script is broken (calls the divergent Newton solver its own docs say was replaced; prints a sign after non-convergence); Δβ = 2.163ξ recoverable analytically | **A — verified** (script hangs; gate defect recorded; repair queued: fail-closed + the declared first-order implementation) |
| 25 | **ξ = 0.2312 is outside controlled perturbation theory** (2nd order = 3.04× 1st; no stable nonlinear branch); Δβ is an imported dictionary, no orbit computed | **A** (parameters row voided) |
| 26 | ΔP is a regulated auxiliary damping-power correction, not Λ; finite positive regulator limit **supported by the auditor's own sweep** | **A** (the verification is banked) |
| 27 | **The Mercury–Λ chain is vacuous even in sign**: ΔP ∝ ξ² is positive for either sign of ξ — Mercury contributes nothing | **A** (the index chain language dismantled — the sharpest single catch of pass 4) |
| 28–29 | Verdicts + 5 gates | **A** (adopted) |

## Pass 5 — tensor sector (§§30–38)

| § | Claim | Verdict |
|---|---|---|
| 30 | TT null space correct; the weights are a load-bearing *amendment* (introduced after the uniform failure) | **A** (n197 already records the correction; both readings true — the isotropy equation has no freedom, and it is an amendment) |
| 31 | **A TT projector is not GR's constraint algebra**: no π_ij, no Poisson brackets, no first-class closure; the covariant 10−4−4 does not follow from the spatial 6−4; constraint/gauge roles risk double-counting | **A — n197's "constraint theorem" walked back** to what it is: a valid projector construction. FP leg 2 re-marked partial. The canonical construction is audit gate 38.1 |
| 32 | Helicity ±2 tautological once TT is imposed; the little group's null rotations unproven | **A** (C2 demoted to representation bookkeeping) |
| 33 | n175 has no executable in the repo; masslessness imposed by tuning u″, thermostat not modeled | **A — verified absent**; reproducibility defect recorded; the sandpile dynamics simulation stays queued |
| 34 | The h·T identity is a spatial scalar-field identity, not universal coupling | **A** (n177 scope) |
| 35 | c_g fails decisively; **the bare-gradient bargain is calibration on the predicted observable — frozen condition E — and is rejected**; deferral = open-ended escape | **A — the n203 proposal is withdrawn.** Registered counter-position (not an evasion): data-pinned bare parameters are how G enters GR; but the *frozen gate's* law governs this sector, and the frozen wording forbids it. The one legitimate route: a preregistered dressed-vacuum computation, now the sector's decisive calculation |
| 36 | Weinberg–Deser cannot supply their own premises | **A** (long since adopted) |
| 37–38 | Verdict + 5 gates | **A** |

## Pass 6 — topology wing (§§39–47)

| § | Claim | Verdict |
|---|---|---|
| 39 | Handles-as-particles needs dynamics; **indestructibility is conditional on frozen adjacency, which is never derived** — proton stability and baryon conservation do not follow from static topology | **A** (entry 37 re-scoped: the neutron bet is conditional on — and wagers — the frozen-adjacency conjecture) |
| 40 | **n123's U2 gate cannot fail** (verified: `'PASS' if soften else 'reported'`); the tube holds only N = 1; the tested cycle is length 7 (odd), conflicting with the claimed even six-cycle; no executable produces the quoted 3.634 | **A — verified**; gate defect recorded; keV band demoted on predictions page |
| 41 | Antiperiodic boundary conditions ≠ fermions; the FR homotopy chain (4 steps) absent | **A** (already admitted; "structurally zero" softened by inheritance) |
| 42 | Alice quantization incomplete (no gauge group, declared phase law, odd-cycle vs bipartite tension); confinement not derived | **A** (n158 demoted to a declared-normalization argument) |
| 43 | The keV band is dimensional anchoring, not a mass prediction (no dispersion, no inertial mass for the winding) | **A** (applied with §40) |
| 44 | σ/m ≤ 3×10⁻¹² is conditional dimensional analysis, not normalized scattering theory | **A** (28d's kill conditions stand as exposures; the number is an estimate) |
| 45 | The original KZ python fails (log(0)/NaN — gate defect); the Rust analog is valid; abundance unpredicted | **A** (n136 script defect recorded; n200 stands) |
| 46–47 | Verdict + 5 gates | **A** |

## Pass 7 — electromagnetism (§§48–56)

| § | Claim | Verdict |
|---|---|---|
| 48 | Coulomb sector reproducible (auditor's rerun: 0.36%); the mobile-charge/one-site-translation problem | **A** (the transport debt, already filed — now with the auditor's confirmation of the statics) |
| 49 | **Clock invisibility is not U(1) gauge invariance** (no redundancy, no Ward identities; Wilson loops would be physical) | **A — the flagship phrase "gauge invariance arriving as a theorem" is retracted** (index rewritten: clock-invisibility theorem, gauge reading withdrawn) |
| 50 | **Charge superselection conflicts with the dark-state trunk** — and n171's charge-blind verdict had already killed n69's mechanism internally | **A — retracted on the index.** An internal contradiction we should have caught: recorded as such |
| 51 | W's static zero is not a photon pole; the m_γ bound cannot apply; n159 has no executable | **A — verified absent**; results entry rescoped |
| 52 | Minimal coupling asserted through the missing covariance | **A** (parameters row: vertex = declared definition) |
| 53 | α fails ×45; an S_T repair needs an unexplained 1/2500 stiffness hierarchy conflicting with the invoked Lorentz structure | **A** (the S_T hope is no longer a plan, it is a named problem) |
| 54 | |q| ≤ m is the triangle inequality + electron calibration — bookkeeping, not a bound | **CH — the ledger's one substantive challenge:** once the unit is calibrated on a single species, the bound on *every other* species is a real constraint (a charge-e particle lighter than the electron is forbidden — falsifiable). Conceded: this presupposes the consumption bookkeeping maps linearly to (q, m), which is the conjecture. Registered as a disagreement for the reviewer's reply |
| 55–56 | Verdict + 5 gates | **A** |

## Combined scoreboard (n206 + n207: all 56 sections)

**64 distinct claims: 61 accepted (12 verified independently before acceptance), 3 challenged** (#8 the repo does contain dynamical single-pole realizations; #6 the sign structure remains model content; #54 calibrate-once-predict-many). Casualties of passes 4–7, all applied: the Mercury calibration voided; the Mercury–Λ chain dismantled (vacuous in sign); the constraint theorem walked back to a projector construction; the bargain withdrawn; the keV band demoted to unreproduced dimensional anchoring; "gauge invariance as a theorem" and "charge superselection derived" retracted; the m_γ theorem rescoped as static; three gate-implementation defects recorded (n113, n123-U2, n136). The audit's twenty new gates join the ten from passes 1–3 as the program's queue.

**What the whole audit leaves standing, in one paragraph:** exact lattice identities (IBP, TT projector, fidelity metric, constrained minimum), reproducible statics (Coulomb-class, 0.36%), a clean analog KZ exponent, the odd/even mediator symmetry, the dark-subspace theorem in its narrow form, and a research program with fifty-six numbered ways to die. That is what the site now says it is.