From 905deb720bc79457f418909dddd11cca9baafa15 Mon Sep 17 00:00:00 2001 From: Dylan Roy Date: Mon, 28 Sep 2026 02:16:30 -0500 Subject: [PATCH 1/2] Preserve recovered C022 and Q0 correction history Add 13 byte-exact historical files in a separate review-only intake: five C022 originals, Q0 ledger and scope context, plus KIMI-AUD-006 and 006b. No current scientific record or proof body is changed. --- .../c022/C022 Observed Update.json | 53 + .../c022/Constants Table Canonical C022.md | 51 + .../c022/Registry Reconciliation C022.md | 50 + .../c022/c022 barrier.py.txt | 153 + .../c022/c022 fd.json | 315 + .../context/C102_DETERMINISTIC_REDUCTION.md | 199 + .../C102_FINITE_JET_NONDEGENERACY_copy2.md | 275 + .../context/Lemma I Package.md | 36 + .../context/Lemma MODZONE CountLaw Package.md | 134 + .../context/Q0_LEDGER.md | 5245 +++++++++++++++++ .../context/R0UNIF and LBRestate.md | 55 + .../KIMI-AUD-006_task7_v34_facewise_audit.txt | 132 + .../KIMI-AUD-006b_amendment_closure.txt | 80 + 13 files changed, 6778 insertions(+) create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/c022/C022 Observed Update.json create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/c022/Constants Table Canonical C022.md create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/c022/Registry Reconciliation C022.md create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/c022/c022 barrier.py.txt create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/c022/c022 fd.json create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/context/C102_DETERMINISTIC_REDUCTION.md create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/context/C102_FINITE_JET_NONDEGENERACY_copy2.md create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/context/Lemma I Package.md create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/context/Lemma MODZONE CountLaw Package.md create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/context/Q0_LEDGER.md create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/context/R0UNIF and LBRestate.md create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/side24-reviews/KIMI-AUD-006_task7_v34_facewise_audit.txt create mode 100644 incoming/chatgpt-c022-q0-recovery-20260928/side24-reviews/KIMI-AUD-006b_amendment_closure.txt diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/c022/C022 Observed Update.json b/incoming/chatgpt-c022-q0-recovery-20260928/c022/C022 Observed Update.json new file mode 100644 index 000000000..3da577407 --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/c022/C022 Observed Update.json @@ -0,0 +1,53 @@ +{ + "cycle": "C022 (proof-obligation closure)", + "date": "2026-07-09", + "snapshot_sha256": "c8fe1ed1f93c6c1a81874002b84f06881a6c2fb84ffde85bd2a9cb718755cf4a", + "headline": "OBL-LB-ARCH -- the single load-bearing proof item -- is CLOSED at derived-and-verified grade (two cited-standard steps flagged inline): Lambda(y*) > 0 by five mp certificates with 38-61-order margins at BOTH rungs, and AO(y*) > 0 by an explicit RKHS barrier g0 whose nine pins are exact, whose Morse structure was censused completely (six criticals, two independent boundedness certificates), whose separatrix connection was flow-verified to machine terminals (branch- -> M exactly, branch+ -> E* at f = 1.676 exactly, monotone value profile), and whose FROZEN-PIN RIGIDITY (every conditional-ensemble field keeps criticals at M, S, y* with exact values b, b-ell, b-ell/2) makes the persistence radius MACROSCOPIC: eps0 ~ 0.008, not ell-scale. The support step needs no density lemma: z0 = g0 - m9 is a Cameron-Martin point by construction (Bogachev 3.6.1). Combined with Lemma FD's certified far ceiling (P(beta-far) <= 12.41 r^3 at the d0 = 0.3 split, rung-independent), the r^3 EXPONENT is now pinned two-sided at proof-adjacent grade: positivity below, explicit O(r^3) envelope above, sharp constant C* = 0.96 +- 0.05 measured. OBL-FAR-DECAY far-field core, OBL-BETA-FAR fixed-radius core, OBL-R0-UNIF, and OBL-LB-RESTATE are also discharged; the moderate-zone r-uniform sharpening of the beta ceiling is re-registered with precise scope.", + + "certificates": { + "protocol": "mpmath dps 60, eigsy, PASS iff lambda_min > 10^3 x residual; typed-set interior point (H_M, H_S, H_Y) = ((-1.5,0,-1.6),(1.2,0,-1.4),(1.0,0,-1.2)), ball 0.2, strict margins (2.40, 3.10, 1.68, 1.20)", + "r0.05 @ y* = (-0.063, 0.012)": {"C1_grad_marginal": [1.156e-06, "10^60.8"], "C2_needle": ["s_t = 0.0676 ell", "Pwin = 1.000000"], "C3_hess6": [2.603e-10, "10^52.0"], "C4_hess9": [1.948e-13, "10^51.6"], "C5_gram9": [1.321e-12, "10^48.5"], "C6_gram18": [9.727e-21, "10^40.1"]}, + "r0.025 @ y* = (-0.030, 0.007)": {"C1": [8.939e-08, "10^60.8"], "C2": ["s_t = 0.0864 ell", "Pwin = 0.999999"], "C3": [4.068e-12, "10^51.4"], "C4": [3.095e-14, "10^51.6"], "C5": [3.333e-14, "10^46.9"], "C6": [3.728e-23, "10^37.7"]}, + "verdict": "ALL PASS at both rungs -- Lambda(y*) > 0 certified" + }, + + "barrier": { + "construction": "g0 = nine-pin minimal-norm RKHS interpolant (bump footprints subtracted exactly in the mp solve) + elder bump c_E = 0.9 at x_E = (-0.492, 0.147) [r=0.05] / (-0.438, 0.170) [r=0.025]; pins exact by construction; ||g0||_H = 3.4294 / 3.070; max|w| = 4.96e5 / 5.36e6 -> float128 eval error ~5e-14 / 5e-13", + "construction_kill": "the 15/18-pin variant (Hessian pins included) REJECTED-AS-BUILT: max|w| = 1.27e14 destroyed evaluability at window precision and its natural y*-Hessian was max-typed; recorded as a construction dead end (not a registry kill); the 9-pin route is the construction of record", + "morse_verification_r0.05": "complete census on [-2.4,2.4]x[-1.2,1.6]: exactly 6 criticals: E* (-1.16755, 0.859248) f=1.676208 max (min|eig| 1.974); +x artifact max (1.3437, 0.6480) f=1.938; M f=1.200000 max (det +0.0196, min|eig| 0.0299); y* f=1.199990 saddle (det -0.0345, eig (+0.0467, -0.7379), unstable dir (-0.919, 0.393)); S f=1.199979 saddle (det -0.0449); far min f=-0.564. ALL THREE PINNED POINTS NATURALLY TYPED -- no shaping bumps needed", + "boundedness": "two independent certificates: (i) decay envelope |g0| <= Sigma|w| (d+2)^2 e^{-d^2/2} < 1.19 for hull-dist >= 6.0 (Sigma|w| = 1.023e6); (ii) float128 annulus grid [-7,7]^2 minus window: max g0 = 0.8650, Lipschitz-corrected <= 0.9086 < 1.19. {g0 >= 1.19} bbox = [-1.88, 2.19] x [-0.17, 1.56], strictly interior", + "connection": "branch- -> M exactly (f = 1.200000, |grad| = 9e-16); branch+ -> E* exactly (f = 1.676208, |grad| = 1e-14); value profile monotone, crosses the 1.31 capture level at (-0.536, 0.356); {g0 >= 1.30} has exactly 2 components (E*'s, +x's), disjoint by Morse theory (no saddle above 1.19998), crossing clearance to the +x component = 0.986", + "margins": "min|grad g0| on {g0 >= 1.19} minus six 0.05-balls = 0.00802; ||g0||_C2(K) = 2.82; frozen-pin rigidity: (h, grad h) = 0 at M, S, y* for every conditional-ensemble field -> exact values forever, ball self-protection (no new criticals inside pinned balls for eps < lambda_min: 0.0299/0.0666/0.0467), typing preserved for eps < 0.026 -> eps0 ~ 0.008 MACROSCOPIC", + "replication_r0.025": "same topology: 6 criticals (E* at (-1.007, 1.034) f = 1.7693; +x max f = 1.614), naturally typed (det margins 0.0036-0.0082), both terminals verified (branch+ -> E*, branch- -> M); softer margins eps0 ~ 0.003; full margin scan not repeated (optional per freeze)", + "incident": "the first r=0.025 replication run FAILED-AS-RUN: the module-global ell was not patched, so the S pin targeted b - ell(0.05) (f-err -1.82e-5 = exactly the ell discrepancy, giant artifact structure +-15). Root-caused, fixed (B.ell = ell2), re-run clean. Recorded as an instrument-patch incident, closed same-session", + "support_step": "z0 = g0 - m9 is in the conditional Cameron-Martin space BY CONSTRUCTION (both terms in H; z0 has zero 9-pins); P(||Z - z0||_C2(K) < eps0) > 0 by Gaussian small-ball positivity at CM points (Bogachev Thm 3.6.1) -- no density lemma needed for the event; C2(K)-density of the BF RKHS (polynomials x Gaussian + pin correction) recorded for the full-support remark" + }, + + "lemma_FD": { + "exact_tables": "19-row certified pointwise band-critical intensity bound c(d), d in [0.05, 6.0], both rungs, mp dps 50, max over 3 directions; rung-agreement < 0.5% at every d >= 0.3 (pair is pointlike -> physical r-uniformity of the far field)", + "dual_weights": "||G6^{-1} v||_1 = 6.556e4 (r=0.05) / 5.183e5 (r=0.025); envelope useful only from d ~ 6 -- the exact table carries [0.05, 6], the envelope the tail", + "band_intensity_constant": "E[|det H| | f=b, grad=0] = 2.41290 exact (GH 80^3); C_b = 0.07457 per unit area per unit level; crude ceiling saturates at 0.1236 (x1.66 Cauchy-Schwarz inflation, recorded)", + "far_ceiling": "P(beta-far) <= C(d0) r^3 with C(0.3) = 12.41 (rung-independent, rigorous for all r <= 0.05); C(2r) = 84.2 (r=0.05) / 334.9 (r=0.025) per-rung; the moderate zone (2r, d0) is NOT r-uniform as-written (crude Hessian bound ~ d^-4; true near mass is Lambda-scale ~6 ell, two orders below) -- the r-uniform sharpening is the re-scoped residual", + "three_piece_cover": "d <= 2r: C020 rigidity (8e-42) + corridor KR (2.6e-8); (2r, d0): per-rung table rows incl. seam rows c(0.05), c(0.07) at r=0.025; d >= d0: the rung-independent ceiling. No gap at r=0.05 (2r = d0 = 0.10)", + "gamma_LOC_corollary": "far-field stabilization certified: all one-point conditional quantities within 2.2% of unconditional at d >= 3, < 1e-4 by d = 5, enveloped beyond 6; tube-local part remains architecture (unchanged scope)", + "sandwich": "the rigorous two-sided exponent statement: 0 < c(r,b) <= (1-q)/r^3 <= 0.96 + 12.41 + eps ~ 13.4 at the anchor rungs -- Theta(r^3) with exponent proof-adjacent, constant measured" + }, + + "obligations": { + "closed": [ + "OBL-LB-ARCH -> derived-and-verified-CLOSED (two cited-standard steps flagged: Bogachev 3.6.1 small-ball; Morse-Smale basin continuity). THE load-bearing item. Lemma_LB_ARCH_Package.md", + "OBL-FAR-DECAY (far-field core) -> discharged derived-and-verified; tube-local part remains architecture (scope unchanged). Lemma_FD_Package.md", + "OBL-BETA-FAR (fixed-radius far core) -> discharged derived-and-verified; moderate-zone r-uniform sharpening re-registered with precise scope (see new). Lemma_FD_Package.md", + "OBL-R0-UNIF -> discharged (bookkeeping): continuity + compactness + universal-interpolation off the analytic null set; mp anchors 9.73e-21 / 3.73e-23. R0UNIF_and_LBRestate.md Part A", + "OBL-LB-RESTATE -> discharged: Lemma LB scoped supersede-in-scope to moderate-r phenomenology; asymptotic chain of custody = LB0 -> R0 -> LB-ARCH -> FD. Part B" + ], + "new": [ + "OBL-BETA-MODZONE (successor to OBL-BETA-FAR residual, precisely scoped): replace the Cauchy-Schwarz Hessian ceiling on d in (2r, d0) by the near-fold |det H| suppression to obtain an r-uniform moderate-zone bound; per-rung numbers already rigorous", + "OBL-LB-RATE (optional): rigorous sharp lower rate via the C012/C013 exact limit objects; NOT required for Theta(r^3)" + ], + "carried": ["OBL-NEEDLE-TREND (non-load-bearing)", "OBL-LOOP-LOCATE (bookkeeping)", "OBL-LBB-1", "OBL-C019-2", "gamma-LOC tube-local architecture (inside Lemma UB0 scope)"], + "program_state": "Theorem A: measurement-complete (C021) + lower-side positivity CLOSED at derived grade + rigorous O(r^3) upper envelope (C022). Remaining to full proof grade: OBL-BETA-MODZONE, the basin-continuity formalization if referee-demanded, gamma-LOC tube-local, and manuscript unification. Theorem B: unaffected, its ledger stands as of C015." + }, + + "artifacts": ["C022_ProofObligations_snapshot.json + C022_freeze_hash.txt", "c022_certs_r05.json", "c022_certs_r025.json", "c022_cert18_r0.05.json", "c022_cert18_r0.025.json", "c022_barrier.py", "c022_barrier_census.json", "c022_barrier_extended.json", "c022_barrier_flow.json", "c022_barrier_margins.json", "c022_barrier_global.json", "c022_barrier_r025.json", "c022_fd.json", "Lemma_LB_ARCH_Package.md", "Lemma_FD_Package.md", "R0UNIF_and_LBRestate.md"] +} diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/c022/Constants Table Canonical C022.md b/incoming/chatgpt-c022-q0-recovery-20260928/c022/Constants Table Canonical C022.md new file mode 100644 index 000000000..1b220f107 --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/c022/Constants Table Canonical C022.md @@ -0,0 +1,51 @@ +# The Constants of the r³ Law — canonical table as of C022 (2026-07-09) + +Single source of truth; supersedes Constants_Table_Canonical_C021.md (priors retained; scope chains +explicit). b = 1.2, κ = 1, ℓ = r³/6, BF testbed. [PINNED]/[PALM] conventions as before. + +**Headline (C022):** the r³ exponent is now **pinned two-sided at proof-adjacent grade**: +lower side by Lemma LB-ARCH (rigorous positivity, derived-and-verified, two cited-standard steps), +upper side by the measured integral + Lemma FD's certified envelope. Assembled: +**0 < c(r, b) ≤ (1 − q)/r³ ≤ 0.96 + 12.41 + ε ≈ 13.4** at the anchor rungs, with the sharp +constant **C\* = 0.96 ± 0.05** [measured, two rungs, clean gates]. The single load-bearing proof +item (OBL-LB-ARCH) is closed. + +## Exact +(unchanged from the C021 edition: pair-Hessian Gram (1/540)r¹⁸; Lemma P Gram 13824; 9×9 free-jet +3,981,312; pinned laws; E|det H| = 2.414589; ρ_saddle/ρ_max = 0.030513/0.044111; octave 1/8; +ρ₆(0) = 0.196817; the λ disintegration identity; the Λ window-integral closed form; the near-M +null functional with λ_min = 5.4×10⁻²⁰ at r = 0.025.) +New exact-grade entry: **E[|det H| | f = b, ∇f = 0] = 2.41290** (unconditional level-b +Hessian-determinant mean, GH 80³; distinct object from the 2.414589 Palm constant — both stand). + +## Certified (C022, mp dps 50–60, residual-verified) +| object | r = 0.05 | r = 0.025 | margin | +|---|---|---|---| +| C1 ∇-marginal λ_min | 1.156×10⁻⁶ | 8.939×10⁻⁸ | 10^60.8 | +| C2 needle s_t/ℓ; P_win | 0.0676; 1.000000 | 0.0864; 0.999999 | exact | +| C3 6-pin Hessian-cond λ_min | 2.603×10⁻¹⁰ | 4.068×10⁻¹² | 10^51–52 | +| C4 9-pin Hessian-cond λ_min | 1.948×10⁻¹³ | 3.095×10⁻¹⁴ | 10^51.6 | +| C5 9-pin Gram λ_min | 1.321×10⁻¹² | 3.333×10⁻¹⁴ | 10^47–49 | +| C6 18-pin 2-jet Gram λ_min | 9.727×10⁻²¹ | 3.728×10⁻²³ | 10^38–40 | +| barrier ‖g₀‖_H; max\|w\| | 3.4294; 4.96×10⁵ | 3.070; 5.36×10⁶ | — | +| barrier ε₀ (persistence) | ≈ 0.008 (macroscopic) | ≈ 0.003 | — | +| ‖G₆⁻¹v‖₁ (FD dual weights) | 6.556×10⁴ | 5.183×10⁵ | — | +| FD far ceiling C(d₀ = 0.3) | 12.41·r³ | 12.41·r³ | rung-independent | +| FD far ceiling C(d₀ = 2r) | 84.2·r³ | 334.9·r³ | per-rung | + +## Measured — asymptotic-regime, two rungs [PALM] (unchanged from C021) +C\*(0.05) = 0.9702 → 0.9728 (ballfix) ± 0.055; C\*(0.025) = 0.9461 ± 0.066; combined +**C\* = 0.96 ± 0.05**; ∫Λ dA ratio 8.23 ≈ 8; qual/raw ≥ 0.9997; rim ~r^2.5; β dead both rungs +(support edge 300ℓ → 894ℓ); gates 6/6 at C021. + +## Measured — moderate-r brackets (unchanged; phenomenology; Lemma LB scoped here per C022 Part B) + +## The assembled statement (supersedes the C021 paragraph) +**1 − q(r, b) = Θ(r³): the exponent is two-sided at proof-adjacent grade — lower side +Lemma LB-ARCH [derived-and-verified; cited-standard: Bogachev 3.6.1, MS basin continuity], upper +side E[N] (measured 0.96·r³) + β-near (≤ 2.6×10⁻⁸ + 8×10⁻⁴²) + β-far (≤ 12.41·r³ certified, +d₀ = 0.3, rung-independent) + ledger (< 1.5×10⁻³ rel.). The sharp constant C\* = 0.96 ± 0.05 is +measured (two pre-registered rungs).** Chain of custody: LB0 [Proved] → R0 [+ R0-UNIF discharged] +→ Lemma LB-ARCH [C022] → Lemma FD [C022]. Remaining to full proof grade: OBL-BETA-MODZONE +(r-uniform moderate zone), basin-continuity formalization (if demanded), γ-LOC tube-local, +manuscript unification. Optional sharpenings: OBL-LB-RATE (exact limit objects). diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/c022/Registry Reconciliation C022.md b/incoming/chatgpt-c022-q0-recovery-20260928/c022/Registry Reconciliation C022.md new file mode 100644 index 000000000..b3418d71c --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/c022/Registry Reconciliation C022.md @@ -0,0 +1,50 @@ +# Registry_Reconciliation — obligations ledger and record hygiene (as of C022, 2026-07-09) + +Supersedes Registry_Reconciliation_C021.md; companion to Constants_Table_Canonical_C022.md, +Lemma_LB_ARCH_Package.md, Lemma_FD_Package.md, R0UNIF_and_LBRestate.md, +C022_Observed_Update.json. Supersede-never-overwrite throughout. + +## 1. Obligations ledger (complete, current) + +| ID | content | status | discharge / carrier | +|---|---|---|---| +| OBL-LB-ARCH | rigorous positivity of ∫Λ·AO dA | **CLOSED (C022)** at derived-and-verified; two cited-standard steps flagged inline (Bogachev 3.6.1 small-ball at CM points; Morse–Smale basin continuity). Five certificates (margins 10^38–10^61, both rungs) + explicit RKHS barrier with frozen-pin rigidity (ε₀ ≈ 0.008 macroscopic) + CM support step. | Lemma_LB_ARCH_Package.md | +| OBL-FAR-DECAY | quantitative off-pair decorrelation (γ-LOC) | **far-field core DISCHARGED (C022)** derived-and-verified (19-row exact tables to d = 6, both rungs, envelope beyond); tube-local part remains architecture, scope unchanged | Lemma_FD_Package.md §2 | +| OBL-BETA-FAR | rigorous far-zone β ceiling | **fixed-radius core DISCHARGED (C022)**: P(β-far) ≤ 12.41·r³ at d₀ = 0.3, rung-independent, all r ≤ 0.05; per-rung three-piece cover complete (no gap at r = 0.05). Residual re-scoped → OBL-BETA-MODZONE | Lemma_FD_Package.md §3 | +| **OBL-BETA-MODZONE** | r-uniform moderate-zone (2r, d₀) sharpening: replace Cauchy–Schwarz Hessian ceiling by the near-fold \|det H\| suppression | **NEW (C022)** — the successor residual, precisely scoped; per-rung numbers already rigorous (84.2 / 334.9 ·r³) | Lemma_FD §3 scoping | +| OBL-R0-UNIF | 18×18 nondegeneracy uniformity | **DISCHARGED (C022)**: continuity + compactness + universal interpolation off the analytic null set; mp anchors 9.73e−21 / 3.73e−23 | R0UNIF_and_LBRestate.md Part A | +| OBL-LB-RESTATE | scope Lemma LB to moderate-r | **DISCHARGED (C022)**: supersede-in-scope note; asymptotic chain of custody = LB0 → R0 → LB-ARCH → FD | Part B | +| **OBL-LB-RATE** | rigorous sharp lower rate via C012/C013 limit objects | **NEW (C022, optional)** — not required for Θ(r³) | LB-ARCH §6 | +| OBL-NEEDLE-TREND | needle-ratio drift 0.068 → 0.086 | carried (non-load-bearing; one third-rung datum) | C021 | +| OBL-LOOP-LOCATE | exhibit/bound-away loop population | carried (bookkeeping) | UB0 Addendum A | +| OBL-LBB-1 | two-point at d ∈ {0.02, 0.05, 0.1} | carried | — | +| OBL-C019-2 | MID mechanism / polish hygiene | carried | — | + +Previously discharged (unchanged): OBL-UB-PALM, OBL-C019-1, OBL-CLIP-ARCH, OBL-C017-1/2, +OBL-LB-ADJ (superseded), File-5 ghost debt; OBL-RIM-REFINE optional; κ_β moot. + +## 2. C022 record notes + +- **Construction dead end (not a registry kill):** the 15/18-pin barrier variant rejected-as-built + (max|w| = 1.27×10¹⁴, evaluability destroyed; natural y*-Hessian max-typed). The 9-pin + + shaping-bump route is the construction of record. +- **Instrument incident (closed same-session):** first r = 0.025 barrier replication ran with the + unpatched module-global ℓ (S-pin f-err −1.82×10⁻⁵ = exactly the ℓ discrepancy). Root-caused, + fixed, re-run clean; both runs retained in the record. +- **Honest-scoping note:** the FD moderate-zone bound is per-rung rigorous but not r-uniform + as-written (crude ceiling ~d⁻⁴ vs true Λ-scale mass); the gap is now a *named, scoped* + obligation rather than an implicit one. + +## 3. Program state (of record) + +**Theorem A:** measurement-complete (C021) **+** lower-side positivity closed at derived grade +**+** rigorous O(r³) upper envelope (C022). The Θ(r³) statement holds with exponent at +proof-adjacent grade and constant measured. Remaining to full proof grade: OBL-BETA-MODZONE, +basin-continuity formalization (if referee-demanded), γ-LOC tube-local, manuscript unification. +**Theorem B:** unaffected; its ledger stands as of C015. +**Submission paths:** experimental-mathematics package — CLEAR (assembly only); methodology +paper — CLEAR; exact-lemma cluster — CLEAR; proof-grade Theorem A — the four items above. + +## 4. Reconciliation notes carried unchanged +C014 dual-account resolution; C014p2 snapshot-absence flag; date-field corrections; C020 ballfix +supersede chain (0.9702 → 0.9728) — as in the C020/C021 registries. diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/c022/c022 barrier.py.txt b/incoming/chatgpt-c022-q0-recovery-20260928/c022/c022 barrier.py.txt new file mode 100644 index 000000000..5495ee18b --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/c022/c022 barrier.py.txt @@ -0,0 +1,153 @@ +# c022_barrier.py -- C022 S3: explicit RKHS barrier g0 for the AO-positivity step. +# g0 = 15-pin minimal-norm interpolant (2-jets at M and S, 1-jet01 at y*) + elder bump, +# coefficients solved in mp; evaluation in float128 via Hermite formulas (abs err ~1e-7, +# adequate against ell = 2.08e-5 margins); census float128 + mp polish of candidates. +import numpy as np, json, sys, time +from mpmath import mp, mpf, matrix as mpM +import mpmath +import c019_flow as C +mp.dps = 60 +t0 = time.time() +LD = np.longdouble + +b = 1.2; r = 0.05; ell = r**3/6 +M = (-0.025, 0.0); S = (0.025, 0.0); Y = (-0.063, 0.012) +XE = (-0.492, 0.147) +J2 = [(0,0),(1,0),(0,1),(2,0),(1,1),(0,2)] +FUN = [(a, M) for a in J2[:3]] + [(a, S) for a in J2[:3]] + [(a, Y) for a in J2[:3]] +# shaping bumps: (amplitude, center) -- elder + typing shapers; footprints subtracted in build() +BUMPS = [(0.9, (-0.492, 0.147))] + +# ---- Hermite tables for D^alpha e^{-t^2/2} = (-1)^|a| He_a(t) e^{-t^2/2} ---- +def He(n, t): + if n == 0: return np.ones_like(t) + if n == 1: return t + if n == 2: return t*t-1 + if n == 3: return t**3-3*t + if n == 4: return t**4-6*t*t+3 + if n == 5: return t**5-10*t**3+15*t + raise ValueError + +def dK(bx, by, ax, ay, x, y, zx, zy): + """D^{(bx,by)}_x D^{(ax,ay)}_z K(x,z), K = exp(-|x-z|^2/2), vectorized in x.""" + ux = x - zx; uy = y - zy + n1 = ax + bx; n2 = ay + by + sgn = (-1)**(ax+ay) * (-1)**(n1+n2) + return sgn * He(n1, ux)*He(n2, uy)*np.exp(-(ux*ux+uy*uy)/2) + +def dK_mp(ba, aa, xp, zp): + (bx, by), (ax, ay) = ba, aa + ux = mpf(xp[0])-mpf(zp[0]); uy = mpf(xp[1])-mpf(zp[1]) + def He1(n, t): + return [mpf(1), t, t*t-1, t**3-3*t, t**4-6*t*t+3, t**5-10*t**3+15*t][n] + n1, n2 = ax+bx, ay+by + return (-1)**(ax+ay) * (-1)**(n1+n2) * He1(n1, ux)*He1(n2, uy)*mp.e**(-(ux*ux+uy*uy)/2) + +def build(vstar, bumps): + """Solve the 9 pin conditions (targets minus all bump footprints) in mp.""" + tg = [b, 0, 0, b-ell, 0, 0, vstar, 0, 0] + n = len(FUN) + G = mpM(n, n) + for i, (ai, xi) in enumerate(FUN): + for j, (aj, xj) in enumerate(FUN): + G[i, j] = C.cov_mp(ai, xi, aj, xj) + tvec = mpM(n, 1) + for i, (ai, xi) in enumerate(FUN): + acc = mpf(tg[i]) + for (cb, zb) in bumps: + acc -= mpf(cb)*dK_mp(ai, (0,0), xi, zb) + tvec[i, 0] = acc + w = mp.lu_solve(G, tvec) + return [w[i, 0] for i in range(n)], G + +class G0: + def __init__(self, w, bumps): + self.w = np.array([LD(mp.nstr(x, 25)) for x in w]) + self.bumps = [(LD(cb), (LD(zb[0]), LD(zb[1]))) for (cb, zb) in bumps] + def jets(self, x, y, order=2): + x = np.asarray(x, dtype=LD); y = np.asarray(y, dtype=LD) + outs = {} + for (bx, by) in ([(0,0),(1,0),(0,1)] + ([(2,0),(1,1),(0,2)] if order >= 2 else [])): + acc = np.zeros_like(x) + for wi, (ai, zi) in zip(self.w, FUN): + acc = acc + wi*dK(bx, by, ai[0], ai[1], x, y, LD(zi[0]), LD(zi[1])) + for (cb, zb) in self.bumps: + acc = acc + cb*dK(bx, by, 0, 0, x, y, zb[0], zb[1]) + outs[(bx, by)] = acc + return outs + +def census(g, x0=-1.05, x1=0.55, y0=-0.45, y1=0.55, step=0.02): + xs = np.arange(x0, x1, step); ys = np.arange(y0, y1, step) + Xg, Yg = np.meshgrid(xs, ys) + P = np.stack([Xg.ravel(), Yg.ravel()], 1).astype(LD) + for _ in range(30): + j = g.jets(P[:, 0], P[:, 1]) + gx, gy = j[(1,0)], j[(0,1)] + hxx, hxy, hyy = j[(2,0)], j[(1,1)], j[(0,2)] + det = hxx*hyy - hxy*hxy + det = np.where(np.abs(det) < 1e-14, 1e-14, det) + dx = ( hyy*gx - hxy*gy)/det + dy = (-hxy*gx + hxx*gy)/det + n = np.hypot(dx.astype(float), dy.astype(float)) + sc = np.where(n > 0.05, 0.05/np.maximum(n, 1e-30), 1.0) + P[:, 0] -= dx*sc; P[:, 1] -= dy*sc + j = g.jets(P[:, 0], P[:, 1]) + gn = np.hypot(j[(1,0)].astype(float), j[(0,1)].astype(float)) + ok = (gn < 1e-9) & (P[:, 0] > x0) & (P[:, 0] < x1) & (P[:, 1] > y0) & (P[:, 1] < y1) + cand = P[ok] + uniq = [] + for p in cand: + if not any(np.hypot(float(p[0]-q[0]), float(p[1]-q[1])) < 1e-4 for q in uniq): + uniq.append(p) + rows = [] + for p in uniq: + j = g.jets(np.array([p[0]]), np.array([p[1]])) + f = float(j[(0,0)][0]); hxx = float(j[(2,0)][0]); hxy = float(j[(1,1)][0]); hyy = float(j[(0,2)][0]) + det = hxx*hyy - hxy*hxy; tr = hxx + hyy + typ = 'max' if (det > 0 and tr < 0) else ('min' if det > 0 else 'saddle') + rows.append(dict(x=float(p[0]), y=float(p[1]), f=f, det=det, tr=tr, type=typ)) + return rows + +def flow(g, p0, d0, h=0.004, cap=20000): + p = np.array(p0, dtype=LD) + LD(0.002)*np.array(d0, dtype=LD) + for it in range(cap): + j = g.jets(np.array([p[0]]), np.array([p[1]])) + gx = float(j[(1,0)][0]); gy = float(j[(0,1)][0]) + gn = np.hypot(gx, gy) + if gn < 1e-7: + jj = g.jets(np.array([p[0]]), np.array([p[1]])) + det = float(jj[(2,0)][0]*jj[(0,2)][0]-jj[(1,1)][0]**2); tr = float(jj[(2,0)][0]+jj[(0,2)][0]) + return dict(term=[float(p[0]), float(p[1])], f=float(jj[(0,0)][0]), + typ='max' if (det > 0 and tr < 0) else 'other', steps=it) + d1 = np.array([gx, gy])/gn + pm = p + LD(h/2)*d1.astype(LD) + jm = g.jets(np.array([pm[0]]), np.array([pm[1]])) + gmx = float(jm[(1,0)][0]); gmy = float(jm[(0,1)][0]); gmn = max(np.hypot(gmx, gmy), 1e-30) + p = p + LD(h)*np.array([gmx/gmn, gmy/gmn], dtype=LD) + if abs(float(p[0])) > 2.5 or abs(float(p[1])) > 2.5: + return dict(term=[float(p[0]), float(p[1])], f=None, typ='esc', steps=it) + return dict(term=[float(p[0]), float(p[1])], f=None, typ='cap', steps=cap) + +if __name__ == '__main__': + vstar = b - ell/2 + import ast + bumps = ast.literal_eval(sys.argv[1]) if len(sys.argv) > 1 else BUMPS + w, G = build(vstar, bumps) + wn = max(abs(float(x)) for x in w) + print('bumps: %s' % str(bumps), flush=True) + print('coefficients solved: max|w| = %.3e (float128 eval abs-err scale ~ %.1e)' % (wn, wn*1e-19), flush=True) + g = G0(w, bumps) + # pin verification in float128 + for nm, p, tv in [('M', M, b), ('S', S, b-ell), ('Y*', Y, vstar)]: + j = g.jets(np.array([p[0]]), np.array([p[1]])) + print('%s: f-err %.2e |grad| %.2e H=(%.3f, %.3f, %.3f)' % ( + nm, float(j[(0,0)][0])-tv, np.hypot(float(j[(1,0)][0]), float(j[(0,1)][0])), + float(j[(2,0)][0]), float(j[(1,1)][0]), float(j[(0,2)][0])), flush=True) + print('census... (%.0fs)' % (time.time()-t0), flush=True) + rows = census(g) + rows.sort(key=lambda r: -r['f']) + for rr in rows: + print(' crit (%.4f, %.4f) f=%.6f %s det=%.3f tr=%.3f' % ( + rr['x'], rr['y'], rr['f'], rr['type'], rr['det'], rr['tr']), flush=True) + json.dump(dict(w_max=wn, census=rows), open('c022_barrier_census.json', 'w'), indent=1) + print('(%.0fs)' % (time.time()-t0)) diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/c022/c022 fd.json b/incoming/chatgpt-c022-q0-recovery-20260928/c022/c022 fd.json new file mode 100644 index 000000000..afa494b35 --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/c022/c022 fd.json @@ -0,0 +1,315 @@ +{ + "band_intensity": { + "E_absdet": 2.4129, + "C_b": 0.07457 + }, + "fd_r0.05": { + "w6_l1": 65560.01988862328, + "table": [ + { + "d": 0.3, + "c": 144.82158636577662, + "mu_f": 1.2011647351870498, + "s_f": 0.0024900040517053306, + "det_gg": 0.000400593602989338, + "Edet_bound": 0.1268141573010898 + }, + { + "d": 0.5, + "c": 17.898754697187922, + "mu_f": 1.205515895130968, + "s_f": 0.011889672461930938, + "det_gg": 0.006845081989680566, + "Edet_bound": 0.30880711504008007 + }, + { + "d": 0.75, + "c": 4.013575837832931, + "mu_f": 1.2177634065736984, + "s_f": 0.04232628866536052, + "det_gg": 0.050440465823565636, + "Edet_bound": 0.6562163511273142 + }, + { + "d": 1.0, + "c": 1.7219742239975702, + "mu_f": 1.2368210557305395, + "s_f": 0.10738306394604896, + "det_gg": 0.1567256663408461, + "Edet_bound": 1.2227389214480997 + }, + { + "d": 1.25, + "c": 1.0293703654186088, + "mu_f": 1.2480159897667646, + "s_f": 0.22495468207971925, + "det_gg": 0.2908178078897732, + "Edet_bound": 2.012056542618649 + }, + { + "d": 1.5, + "c": 0.7076728532971432, + "mu_f": 1.2008348538972744, + "s_f": 0.4055082097442327, + "det_gg": 0.3962631541527714, + "Edet_bound": 2.8450817924144065 + }, + { + "d": 2.0, + "c": 0.6080124963025609, + "mu_f": 1.4205144340617868, + "s_f": 0.5447476309991697, + "det_gg": 0.5146097534309881, + "Edet_bound": 4.06161975830341 + }, + { + "d": 2.5, + "c": 0.2971052986948908, + "mu_f": 0.6476033788760626, + "s_f": 0.9021729950300372, + "det_gg": 0.6831825518625313, + "Edet_bound": 4.208642795998544 + }, + { + "d": 3.0, + "c": 0.16491496204480732, + "mu_f": 0.19231414584172563, + "s_f": 0.9881819473184879, + "det_gg": 0.9026028193014997, + "Edet_bound": 4.101184958304723 + }, + { + "d": 3.5, + "c": 0.13269022842588626, + "mu_f": 0.05072498184794291, + "s_f": 0.9990403216928669, + "det_gg": 0.985935252300623, + "Edet_bound": 4.0176499145468165 + }, + { + "d": 4.0, + "c": 0.12533950013096662, + "mu_f": 0.010801105566206037, + "s_f": 0.9999532797249878, + "det_gg": 0.998981173140425, + "Edet_bound": 4.001495983960994 + }, + { + "d": 4.5, + "c": 0.12389032082082672, + "mu_f": 0.0017541460103518297, + "s_f": 0.9999987040359813, + "det_gg": 0.9999611868853993, + "Edet_bound": 4.0000615286251495 + }, + { + "d": 5.0, + "c": 0.12365790408747042, + "mu_f": 0.0002159942705439981, + "s_f": 0.9999999794982796, + "det_gg": 0.9999991968433524, + "Edet_bound": 4.000001267420025 + }, + { + "d": 5.5, + "c": 0.12362877547659193, + "mu_f": 2.023474175544616e-05, + "s_f": 0.9999999998133401, + "det_gg": 0.9999999907661107, + "Edet_bound": 4.00000001342751 + }, + { + "d": 6.0, + "c": 0.12362598739571923, + "mu_f": 1.4472052743955408e-06, + "s_f": 0.9999999999990141, + "det_gg": 0.9999999999400464, + "Edet_bound": 4.000000000073266 + } + ], + "annulus_integral": 69.64289064674838, + "eta_env_d6": 0.009822383290447156, + "tail": 4.8193571278059935, + "ceiling_over_ell": 505.13341792863224, + "ceiling_over_r3": 84.18890298810537, + "bridge": [ + { + "d": 0.1, + "c": 14401.98635384807 + }, + { + "d": 0.15, + "c": 2628.675565442333 + }, + { + "d": 0.2, + "c": 786.7026118296612 + }, + { + "d": 0.25, + "c": 309.61799485717114 + } + ], + "annulus_integral_from_0.10": 500.31406080082627 + }, + "fd_r0.025": { + "w6_l1": 518320.00998764904, + "table": [ + { + "d": 0.3, + "c": 143.342730850061, + "mu_f": 1.2012336169702231, + "s_f": 0.0025002895937896287, + "det_gg": 0.00040063436660786394, + "Edet_bound": 0.12760590342680678 + }, + { + "d": 0.5, + "c": 17.834767424341234, + "mu_f": 1.2056237994321084, + "s_f": 0.011909040776826337, + "det_gg": 0.006845722880262285, + "Edet_bound": 0.30941948795905017 + }, + { + "d": 0.75, + "c": 4.007544704555363, + "mu_f": 1.2179196375254357, + "s_f": 0.04236190409408311, + "det_gg": 0.050443991884948004, + "Edet_bound": 0.6567267513247166 + }, + { + "d": 1.0, + "c": 1.720769031765548, + "mu_f": 1.2370220161965901, + "s_f": 0.10744286705908841, + "det_gg": 0.15673096806706194, + "Edet_bound": 1.2232903202091179 + }, + { + "d": 1.25, + "c": 1.0290635887157338, + "mu_f": 1.248242051099781, + "s_f": 0.22504668003573988, + "det_gg": 0.29081758234793226, + "Edet_bound": 2.0126735912001545 + }, + { + "d": 1.5, + "c": 0.7076043089823709, + "mu_f": 1.2010296862590155, + "s_f": 0.40562934983751947, + "det_gg": 0.3962606553502001, + "Edet_bound": 2.8456502342687013 + }, + { + "d": 2.0, + "c": 0.607837438756751, + "mu_f": 1.4205707191504906, + "s_f": 0.5449488070521139, + "det_gg": 0.5146385101798758, + "Edet_bound": 4.06198748065292 + }, + { + "d": 2.5, + "c": 0.2970409093278798, + "mu_f": 0.6474103636630819, + "s_f": 0.9022508244341291, + "det_gg": 0.6832796723502417, + "Edet_bound": 4.20886006965871 + }, + { + "d": 3.0, + "c": 0.16489439848587148, + "mu_f": 0.1922358410661015, + "s_f": 0.9881937595523874, + "det_gg": 0.9026657169026432, + "Edet_bound": 4.10122302760376 + }, + { + "d": 3.5, + "c": 0.13268246688094804, + "mu_f": 0.05069783202599579, + "s_f": 0.9990415396628527, + "det_gg": 0.9859493593002564, + "Edet_bound": 4.017651994099563 + }, + { + "d": 4.0, + "c": 0.1253355184148558, + "mu_f": 0.0107932673123052, + "s_f": 0.9999533555811538, + "det_gg": 0.9989825878641119, + "Edet_bound": 4.001495626784607 + }, + { + "d": 4.5, + "c": 0.12388736186830045, + "mu_f": 0.00175247973532731, + "s_f": 0.9999987066694256, + "det_gg": 0.9999612571324513, + "Edet_bound": 4.000061486712816 + }, + { + "d": 5.0, + "c": 0.12365516106351655, + "mu_f": 0.00021573548686516038, + "s_f": 0.9999999795492296, + "det_gg": 0.9999991986672239, + "Edet_bound": 4.000001265895345 + }, + { + "d": 5.5, + "c": 0.12362606678573625, + "mu_f": 2.0205009984876394e-05, + "s_f": 0.9999999998138969, + "det_gg": 0.9999999907917243, + "Edet_bound": 4.0000000134027545 + }, + { + "d": 6.0, + "c": 0.12362328275782555, + "mu_f": 1.4446528228704224e-06, + "s_f": 0.9999999999990176, + "det_gg": 0.9999999999402455, + "Edet_bound": 4.000000000073069 + } + ], + "annulus_integral": 69.3069679989126, + "eta_env_d6": 0.07599093433871817, + "tail": 5.135145214312845, + "ceiling_over_ell": 2009.597930405415, + "ceiling_over_r3": 334.9329884009025, + "bridge": [ + { + "d": 0.1, + "c": 13442.455325375806 + }, + { + "d": 0.15, + "c": 2521.156761498776 + }, + { + "d": 0.2, + "c": 768.2047282365804 + }, + { + "d": 0.25, + "c": 305.0061137793145 + } + ], + "annulus_integral_from_0.10": 478.24499371315966, + "seam_rows": [ + { + "d": 0.05, + "c": 238874.6333449837 + }, + { + "d": 0.07, + "c": 59030.91776454546 + } + ], + "annulus_integral_from_2r": 2004.462785191102 + } +} \ No newline at end of file diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/context/C102_DETERMINISTIC_REDUCTION.md b/incoming/chatgpt-c022-q0-recovery-20260928/context/C102_DETERMINISTIC_REDUCTION.md new file mode 100644 index 000000000..25917befb --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/context/C102_DETERMINISTIC_REDUCTION.md @@ -0,0 +1,199 @@ +# C102 Deterministic Elder-Rule Defect Dichotomy + +**Grade:** `PROVEN-HERE` + +## Hypotheses + +Let \(f\) be a Morse function on a compact surface with distinct critical +values. Assume its gradient flow is Morse–Smale, so every ascending branch of +an index-one saddle terminates at a local maximum. + +Let \(M\) be a local maximum and \(S\) an index-one saddle such that one +ascending branch of \(S\) terminates at \(M\). Put + +\[ +f(M)=b, +\qquad +f(S)=b-\ell, +\qquad +\ell>0. +\] + +Let \(D(M)\) be the death saddle paired with \(M\) by superlevel +\(H_0\) persistence. + +Define: + +\[ +\Pi += +\left\{ +\begin{array}{c} +\text{there exists an index-one saddle }S' +\text{ with}\\ +f(S) window interceptor: + CLOSED-HERE UNDER R0 + +unregistered global path class: + NONE UNDER THE HYPOTHESES + +external SARD-G specialist acceptance: + PENDING +``` diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/context/C102_FINITE_JET_NONDEGENERACY_copy2.md b/incoming/chatgpt-c022-q0-recovery-20260928/context/C102_FINITE_JET_NONDEGENERACY_copy2.md new file mode 100644 index 000000000..1bea9e62e --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/context/C102_FINITE_JET_NONDEGENERACY_copy2.md @@ -0,0 +1,275 @@ +# C102 Finite-Jet Fourier Nondegeneracy + +**Grade:** `DERIVED-EXACT` +**Model:** exact normalized periodized Bargmann–Fock field on +\(\mathbb T_L^2\) +**Numerical eigenvalue floor:** not claimed + +## 1. Exact Fourier representation + +The covariance is + +\[ +K_L(x) += +\frac{ +\sum_{k\in\mathbb Z^2} +e^{-2\pi^2|k|^2/L^2} +e^{2\pi i k\cdot x/L} +}{ +\sum_{k\in\mathbb Z^2} +e^{-2\pi^2|k|^2/L^2} +}. +\] + +Every spectral weight + +\[ +w_k=e^{-2\pi^2|k|^2/L^2} +\] + +is strictly positive. + +Let + +\[ +\Lambda f += +\sum_{j=1}^{m} +\sum_{\alpha\in A_j} +c_{j,\alpha} +\partial^\alpha f(x_j), +\] + +where the points \(x_j\) are distinct modulo \(L\mathbb Z^2\). Its variance is + +\[ +\operatorname{Var}(\Lambda f) += +\frac1{Z_L} +\sum_{k\in\mathbb Z^2} +w_k +\left| +\sum_{j,\alpha} +c_{j,\alpha} +\left(\frac{2\pi i k}{L}\right)^\alpha +e^{2\pi i k\cdot x_j/L} +\right|^2. +\] + +Because every \(w_k>0\), + +\[ +\operatorname{Var}(\Lambda f)=0 +\] + +if and only if + +\[ +E(k_1,k_2) += +\sum_j +p_j(k_1,k_2) +z_j^{k_1}y_j^{k_2} += +0 +\] + +for every \((k_1,k_2)\in\mathbb Z^2\), where + +\[ +z_j=e^{2\pi i x_{j,1}/L}, +\qquad +y_j=e^{2\pi i x_{j,2}/L}, +\] + +and each \(p_j\) is a polynomial. + +## 2. One-dimensional exponential-polynomial lemma + +### Lemma + +Let \(z_1,\ldots,z_s\) be distinct nonzero complex numbers and let +\(p_1,\ldots,p_s\) be complex polynomials. If + +\[ +\sum_{j=1}^s p_j(n)z_j^n=0 +\] + +for every \(n\ge0\), then every \(p_j\) is the zero polynomial. + +### Proof + +For \(|t|\) sufficiently small, + +\[ +F(t) += +\sum_{n\ge0} +\left( +\sum_j p_j(n)z_j^n +\right)t^n += +\sum_j +\sum_{n\ge0} +p_j(n)(z_jt)^n. +\] + +If \(\deg p_j=d_j\), the inner generating function is rational and has its +only possible pole at + +\[ +t=z_j^{-1}, +\] + +with order at most \(d_j+1\). Distinct \(z_j\) give distinct pole locations. + +The assumed zero sequence gives \(F(t)\equiv0\) in a neighborhood of zero, +and hence as a rational function. Every principal part at every distinct pole +must vanish. The principal part at \(z_j^{-1}\) is zero only when +\(p_j=0\). Therefore all \(p_j\) vanish. \(\square\) + +The confluent-Vandermonde example checked by the companion executable is + +\[ +\det V += +z_1z_2(z_1-z_2)^4, +\] + +the two-node, first-derivative instance of the same independence mechanism. + +## 3. Two-dimensional independence + +Group the terms of \(E(k_1,k_2)\) by equal first-coordinate phase \(z_j\). + +For each fixed \(k_2\), the one-dimensional lemma in \(k_1\) shows that every +coefficient polynomial attached to each distinct \(z\)-group vanishes. + +Within one \(z\)-group, the torus points are distinct, so their second phases +\(y_j\) are distinct. Apply the one-dimensional lemma again in \(k_2\). +Every polynomial \(p_j\) vanishes. + +Finally, + +\[ +p_j(k_1,k_2) += +\sum_\alpha +c_{j,\alpha} +\left(\frac{2\pi i}{L}\right)^{|\alpha|} +k_1^{\alpha_1}k_2^{\alpha_2}. +\] + +A zero polynomial has all coefficients zero, so every +\(c_{j,\alpha}=0\). + +Hence no nonzero linear combination of distinct finite derivative evaluations +has zero variance. + +## 4. Theorem + +### Finite-jet nondegeneracy theorem + +For the exact periodized Bargmann–Fock field, every finite family + +\[ +\left\{ +\partial^{\alpha_\nu}f(x_\nu) +\right\}_{\nu=1}^N +\] + +with distinct point/multi-index pairs and distinct underlying points where +appropriate has a positive-definite Gaussian Gram matrix. + +Equivalently, the derivative-evaluation functionals are linearly independent +in the Gaussian Hilbert space. + +This proves the exact-model version of \(H2'\) for arbitrary finite derivative +order, not merely the order-four families required by the Q0 theorem. + +## 5. Divided-difference corollary + +Let \(J\) be a base finite-jet vector with positive-definite covariance +\(\Sigma_J\), and let + +\[ +Y=TJ +\] + +for a deterministic matrix \(T\). Then + +\[ +\Sigma_Y=T\Sigma_JT^*. +\] + +If \(T\) has full row rank, then for every nonzero \(v\), + +\[ +v^*\Sigma_Yv += +(T^*v)^*\Sigma_J(T^*v)>0. +\] + +Thus every divided-difference frame is nondegenerate once its coefficient +transform has full row rank. + +The companion exact checker records: + +- corrected six-pin pair-frame determinant + \[ + \det T_r=-r^{-5}\ne0; + \] +- coalesced six-jet covariance determinant + \[ + 12; + \] +- generic near-frame covariance determinant + \[ + \frac{s^8(c^2+s^2)(3c^2+s^2)}{24}; + \] +- axis-frame determinant + \[ + \frac{(1-\eta^2/4)^{10}}{5760}; + \] +- collar-gradient determinant + \[ + Y^4(4X^2+Y^2). + \] + +The zero sets of the last two generic formulas are precisely the boundary +faces assigned to separate axis, transverse, or collision charts. + +## 6. Conditional Gaussian corollary + +Let \(X\) and \(Y\) be finite jet families. If the joint Gram matrix of +\((X,Y)\) is positive definite, then the Schur complement + +\[ +\operatorname{Cov}(Y\mid X) += +\Sigma_Y-\Sigma_{YX}\Sigma_X^{-1}\Sigma_{XY} +\] + +is positive definite. + +Therefore every fixed-\(r\), distinct-point conditional covariance used by the +Kac–Rice formulas is nonsingular. In collapsing limits, positivity is checked +after the named divided-difference transformation, where the limiting +determinants above apply. + +## 7. Status + +```text +H2' for exact periodized BF: + CLOSED + +fixed-r distinct-point jet Gram matrices: + CLOSED + +named C099-C101 scaled limit frames: + CLOSED + +outward numerical eigenvalue floors: + NOT-CLAIMED +``` diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/context/Lemma I Package.md b/incoming/chatgpt-c022-q0-recovery-20260928/context/Lemma I Package.md new file mode 100644 index 000000000..5ec0e33ef --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/context/Lemma I Package.md @@ -0,0 +1,36 @@ +# Lemma I, Resolved: the Inner-Zone Count and the Completed r³ Architecture for Theorem A +**Artifact class:** lemma-resolution package + zone-architecture consolidation (C009). +**Chain:** C009 snapshot 2e363140…9dc3; observed ledger (companion, hashed). Inputs: Lemma T certified table (C007-B), FlatSaddle Tail Lemma + verification (C007-C, exponent 4.959), bi-graded general-τ laws + O2 closure (C008), window_lemma_v2_patch (Sections R, P, A′, B, I). + +## 1. The lemma as posed, and its resolution + +**Lemma I (window_lemma_v2_patch, Section I).** Under (H1+)–(H2′), E[#{critical points in B_{Cr²}(S) ∪ B_{Cr²}(M), ∉ {M,S}} | pair] ≤ Cr^β, some β > 0. Candidate given there: β = 4. Stated reduction: boundedness of the conditional 3-point/2-point Kac–Rice ratio, requiring nondegeneracy of the leading coefficient of the jet-covariance determinant under the two-scale degeneration |x_S − x_M| ~ r, |y − x_S| ~ r². + +**Resolution.** β = **5**, strictly better than the candidate and than the β ≥ 3 needed for the r³ rate: +(i) *Nondegeneracy of the blow-up.* The two-scale jet covariance is the C007-B certified exact table (gates G1–G4); its general-τ extension is the C008 bi-graded table. The load-bearing Gram — the 6×6 conditional covariance of the pair-Hessian block given E_lin — has determinant **(1/540)·r¹⁸ + O(r¹⁹)** (exact, this cycle): strictly positive leading coefficient at finite order, which is precisely the required blow-up nondegeneracy and discharges the density-positivity part of (U3). The remaining (U3) piece (bounded density of the third-derivative functional ξ) follows from Lemma H directly. +(ii) *The count.* The pair-Palm expectation (which is what "given the pair" means: both member determinants in the weight) of the r²-ball third-point count is O(r⁵) by the Flat-Saddle Tail Lemma with every ledger input now a certified coefficient, and is **numerically verified at 4.959** (unwindowed; 4.967 windowed) over r ∈ [0.014, 0.2]. The M-side ball is symmetric. The r⁵ arises as candidate-β=4 (area × bounded ratio) *plus one power* from the M-side determinant drag |det H_M| ≍ κr² on the contributing near-monkey-saddle set — the ratio is not merely bounded but vanishing after correct weight bookkeeping. +(iii) *Assumption grade.* (U1) conditional Kac–Rice validity ← Lemma H nondegeneracy of the conditional Gaussian + standard a.s.-Morse arguments [obligation, mechanical]. (U2) L^{2+ε} uniformity of the O_P(r) corrections ← finite Gaussian moments under λ₈ < ∞ of (H1+) [obligation, mechanical]. (U3) ← discharged as above, exact constant on record. (U4) ← Palm-base compactness [obligation, bookkeeping]. + +## 2. Two corrections to window_lemma_v2_patch (registered, superseding) + +**C-1 (Localization statement).** The claim that third critical points in B_{3r} concentrate within Cr² of {M, S} "up to exp(−c/r²) tails" is a conditional-at-fixed-w statement (the R3-class error already catalogued in the FlatSaddle kill log). The Palm-averaged law in r² ≪ τ ≲ r is polynomial: **E[N(shell τ)] ≍ r·τ² = r^{2γ+1}**, certified inputs, verified at γ = 2 and γ = 3/2 (C008). Assembly-harmless: the polynomial mass totals ≍ r³, boundary-dominated. + +**C-2 (Lemma A′ scope).** A′'s t-floor argument (|t² − a| ≥ cτ² for τ ≥ 3r) fails on the vertical strips |t ∓ r/2| ≪ r, where the f_t conditional mean vanishes at every distance. Measured on the strip {|t − r/2| ≤ r/2, s ∈ [r/2, 4r]}: windowed pair-Palm count with **tail exponent 2.948** (unwindowed 3.014) — polynomial r³-class where the skeleton claimed exp-smallness, at a stable ≈ 0.24 of the C008 annulus total. A′ is scope-corrected to the off-strip annulus, where the floor argument is degeneracy-proof (the fold modulus κ is *fixed* by the Palm base, so the suppression constant cannot degenerate under the |det|-biased average — the flat direction is μ, which the strips expose and the off-strip t-floor does not involve). + +## 3. The completed zone architecture (all measured or graded) + +| zone | law | grade | +|---|---|---| +| inner balls B_{Cr²}(S) ∪ B_{Cr²}(M) | Θ(r⁵)-bounded, window const ≈ 0.51 | derived (certified inputs) + verified (4.959/4.967) | +| shells r² ≪ τ ≤ r/2 | r^{2γ+1} per shell; total r³, top-octave 0.86–1.00 | derived + verified at two γ; successive slopes → 2.96 | +| peak zone τ ≍ r (strip base [r/2, r]) | r³-class, ≈ 0.24 × annulus; window acceptance ≈ 0.02 | measured (tail 2.95) | +| strip decay s ∈ [r, O(1)] | geometric per-octave decay ×0.08–0.14 | measured (three rungs) | +| far floor, absolute O(1) distances | C(L)·ℓ (Lemma G) | graded per v2 patch; magnitude cross-checked: measured [8r,16r] octave 3.7e-7 at r=0.1 vs ledger estimate ~5e-7 ✓ | +| off-strip annulus [3r, δ] | exp(−c/τ²), κ-pinned floor | derived (v2 A′, scope-corrected) | +| band/loop | ≤ exp(−c/r²) + inner | v2 B1–B2 | + +**Assembly.** 1 − q(r, b) ≤ C(L)ℓ + C·r³ (peak + shells) + C·r⁵ (inner) + exp-small = **O(r³)**, uniformly for b in compacts. Every polynomial component is now either verified numerically against certified-coefficient theory or measured directly; the exp-small components carry degeneracy-proof constants. Remaining formal obligations: (U1), (U2), (U4)-type discharges; W-far/Lemma-G uniformity bookkeeping; Sublemma R0 full write-out. The matching lower bound (the exact law) remains open and now owns three measured constants: the 0.51 inner window coin, the 0.83–0.85 boundary acceptance, and the 0.24 strip/annulus ratio. + +## 4. Process record + +The P-L2 commitment "strip exponent ≥ 4" FAILED-AS-WRITTEN (measured 2.95): root cause, regime conflation in the pre-commitment ledger — the deep-strip (s ≫ r) flatness suppression was applied to a domain dominated by its s ≍ r edge, where the peak-zone physics rules. The pre-committed adversarial branch ("≤ 3 ⟹ re-architect") executed: the peak zone is now an explicit, measured component and the rate conclusion is unchanged and strengthened. Clause (i) (ratio monotonicity) PASSED. The turnover beyond the peak and the far-floor crossover were then measured rather than asserted. diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/context/Lemma MODZONE CountLaw Package.md b/incoming/chatgpt-c022-q0-recovery-20260928/context/Lemma MODZONE CountLaw Package.md new file mode 100644 index 000000000..55c8712a1 --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/context/Lemma MODZONE CountLaw Package.md @@ -0,0 +1,134 @@ +# Lemma MODZONE-CL — the exact moderate-zone band-critical count law, the Mahalanobis ridge, and the re-scoping of the β moderate zone + +**Cycle:** C023 / C023b · **Date:** 2026-07-09 +**Freezes:** C023_MODZONE_snapshot.json sha256 `94e4372b2287766a9ab53124a00ef014fc338073578ede91be51f1ea73e7d089`; +C023b_Exact_snapshot.json sha256 `f0d08becb81081f367f2e5ccf56daaa610a6a2a5add5db88c956e85f72af065e` +**Obligation addressed:** OBL-BETA-MODZONE → adjudicated FAILED-AS-WRITTEN on both registered +routes (P-W1, P-W2) and **superseded** by OBL-BETA-RELEVANCE (§6). The cycle's positive residue +is an exact, four-rung-certified count law (§2–3), a structural discovery about the raw Palm law +(§4), and two archive-correction flags (§5). +**Data:** c023_modzone_r{0.05,0.025,0.0125,0.00625}.json, c023b_exact_r{…}.json, +c023b_adjudication_raw.json, c023c_mahalanobis.json. + +--- + +## §1. The two adjudications (both branch iii, recorded without rescue) + +**P-W1 (sup-decoupled sharpening).** Frozen statistic: c_sharp = φ_∇(0; no mean term) · +sup_v(f-density) · sup_v E[|det H| | f = v, ∇ = 0, pins]. Four-rung integrals +I_sup = 272.5 / 902.6 / 4000 / 15758 at r = 0.05 / 0.025 / 0.0125 / 0.00625; successive +exponents 1.73, 2.15, 1.98. **Branch (iii): γ ≈ 1.95 ≥ 0.3 — FAILED-AS-WRITTEN.** Post-mortem +(recorded, not rescued): (a) the sup-decoupling; (b) the gradient-density factor omitted the +conditional-mean term e^{−½μ_g'Σ⁻¹μ_g} — a **valid but loose** ceiling (factor ≤ 1). The same +omission is present in the C022 Lemma-FD table; those entries remain valid ceilings (inequality +direction preserved) and the fixed-d₀ constants stand; recorded as a looseness note. + +**P-W2 (exact Kac–Rice count).** Successor statistic: the exact band-critical intensity +c̄(d) = ⟨θ⟩[ φ_∇(0; **with** mean term) · (1/ℓ)∫_band φ_f(v)·E(|det H| ∣ f = v, ∇ = 0, pins) dv ], +so that ∫ c̄·2πd dd · ℓ **equals** E[# band criticals in the annulus | 6-pin Palm] (Kac–Rice +identity — no inequality except maxima ≤ all types). Refined log-spaced grids, log-trapezoid: + +| r | I_exact = E[#]/ℓ over (2r, 0.3] | +|---|---| +| 0.05 | **12.43** | +| 0.025 | **43.83** | +| 0.0125 | **161.7** | +| 0.00625 | **621.9** | + +Power fit **γ = 1.882, residual rms 0.022** (successive 1.818, 1.884, 1.943). +**Branch (iii) — FAILED-AS-WRITTEN for the total-count route.** This is now known to be a +property of the truth, not of a bound: the exact expected moderate-zone band-critical count in +ℓ-units diverges like r^{−1.9}. + +## §2. The exact count law (positive residue; derived-and-verified) + +The intensity separates into an inner self-similar profile and an outer fixed profile: + +- **Inner collapse.** c̄(s·r)·r⁴ → ĉ(s) with a four-rung collapse to 0.5% at the inner edge: + ĉ(2) = 6.74/6.63/6.61/**6.60** ×10⁻³ across the rungs; cross-rung agreement at matching s to + 3 digits (e.g. s = 2.2: 2.44×10⁻³ at both r = 0.0125 and 0.00625). ĉ(s) falls ≈ s^{−10}: + samples (2, 6.6e−3), (2.2, 2.44e−3), (2.5, 7.44e−4). +- **Outer profile.** At fixed d, c̄ converges (r ≤ d/4) to c∞(d): 109.5 / 33.1 / 14.4 / 4.70 at + d = 0.10 / 0.15 / 0.20 / 0.30 — local slope ≈ d^{−2.9}, c∞ ≈ 0.14·d^{−2.9}. +- **Assembly.** I_exact(r) ≈ 2π∫ĉ(s)s ds·r^{−2} + outer ≈ **0.019·r^{−2}** + O((2r)^{−0.9}); + verification at r = 0.00625: 0.019/3.9×10⁻⁵ ≈ 487 plus outer ≈ 130 gives 617 ≈ 621.9 ✓. + +**Per-rung salvage (rigorous numbers at the anchor rungs).** Because I_exact is exact, the +moderate-zone β ceiling per rung is: r = 0.05: ≤ 12.43ℓ = **2.07·r³**; r = 0.025: ≤ 43.8ℓ = +**7.31·r³**. Hence the complete per-rung upper envelope at the two flagship rungs: +(1 − q)/r³ ≤ 0.96 + 12.41 (far, d₀ = 0.3) + 2.07 / 7.31 (moderate, exact) + β_near +(≤ 2.6×10⁻⁸ + 8×10⁻⁴²) + ledger ≈ **15.5 / 20.7** — rigorous at each anchor. Only the r → 0 +uniformity of the moderate term is open, and §1 proves it cannot come from counting. + +## §3. Why the count diverges: the Mahalanobis ridge (structural discovery) + +Adversarial verification of the mechanism (c023c_mahalanobis.json), exact mp, three rungs: + +- **q_min(d) := min_θ μ_g'Σ_∇∇⁻¹μ_g ≈ 0.720–0.727, constant** — rung-collapse 0.2–0.9% at + fixed s ∈ {2, 3, 4}, flat in s, minimizer θ* ≈ 1.31–1.49 rad (a crest-line tilted slightly + M-ward of perpendicular). The suppression factor along the crest is e^{−q/2} ≈ **0.70 per + station** — polynomial physics; there is **no exp(−c/τ²) mechanism in the raw 6-pin Palm law**. +- **Transverse gradient sd is linear:** sd(f_⊥′ | pair) ≈ κ_g·d with κ_g ≈ 2.7/r·…, i.e. + sd_gy(s·r) ≈ 2.72·r·g(s), doubling across rungs at fixed s and growing ~d within a rung. + The archived τ³ transverse-gradient scaling holds only for the 2-jet-removed residual / + on-axis; the raw law's transverse gradient freedom is first-order (the free Hessians at the + pins rotate the gradient at rate ~d — as an analytic-kernel Taylor argument predicts). +- On-axis the archived scalings are confirmed: sd(f | pair)_axis ~ τ⁴, exit mean ~ κτ³/3, + longitudinal sd_gx ~ τ^{2.6–3}: the on-axis physics is as File-1 §5.1–5.2 describes. The + failure of suppression is strictly a transverse/off-axis phenomenon. + +**Retro-consistency.** 45,000+ measured flows across C020/C021 produced **zero** β events +(terminal support edge 894ℓ) despite this abundant band-critical population. Conclusion: β +suppression is a **relevance/selectivity** phenomenon (the one S⁺ branch almost never selects a +band-max as its terminal), not a scarcity phenomenon. The deterministic key: the S⁺ ascending +branch is unique, hence Σ_maxima 1{branch terminates at m} = 1 always — so the correct estimand +is the **terminal-height law** P(f(T) ∈ band ∧ ρ_T ∈ zone), not a count. + +## §4. What the divergent population is + +The band-critical mass sits on the crest-line where the conditional mean-gradient's Mahalanobis +norm is O(1): the mean surface's level-b contour through the pair, along which the mean value +stays within O(ℓ)-to-O(d³) of b while both the value sd (perpendicular ~d²) and the transverse +gradient sd (~d) are polynomially large. A β-mod event would force the **entire branch** into +the closed ℓ-band (ascent + terminal-in-band ⟹ f ∈ (b − ℓ, b] along the whole branch), i.e. an +ascending path of length ≥ d − r/2 inside an ℓ-slab; near-critical band points are abundant on +the crest, so path-rigidity/entropy arguments fail in the inner moderate zone. For **fixed** d₀, +the path-rigidity route re-derives an O(ℓ) bound (noted; consistent with Lemma FD's fixed-d₀ +ceiling); it is the (2r, d₀) inner-moderate zone that requires selectivity. + +## §5. Archive-correction flags (supersede-never-overwrite; scope-precise) + +1. **File-1 §5.2 (raw-law reading) and File-3 Proposition 5.2** claim the annulus band count is + O(ℓ) via the Pinning-scaling suppression exp(−c/τ²) with sd(f_s | pair) ≤ Cτ³. The exact + four-rung computation above **contra-indicates the raw-count reading**: the raw transverse + gradient sd is ~d, the crest Mahalanobis is O(1), and E[N_ann | pair] = Θ(r^{−1.9})·ℓ. + The July-5 status "PROVEN-MODULO: B1 chaining" for the annulus/loop suppression is + **downgraded to OPEN-CONTRA-INDICATED for the raw-count reading**. Scope caveat, recorded + honestly: the File-1/File-3 route conditions differently (midpoint 2-jet removed and chained + separately), so this is an archive **correction/scope flag**, not necessarily a disproof of + the chained version; but no chaining can rescue a statement about the raw expectation + E[N_ann | pair] itself, which is now exactly computed and divergent in ℓ-units. +2. **Lemma B1's circle-band mechanism** (loop absorption via exp(−c/r²) on the 3r-circle): the + perpendicular stations of the circle are only polynomially suppressed under the raw law + (mean deficit ~ μd²/2 against sd ~ d²: ratio O(1)). B1's conclusion may survive via the + multi-station product (many near-independent O(ℓ/r²)-cheap stations), but its stated + mechanism is flagged; carried as a scoped audit item inside OBL-BETA-RELEVANCE. + +## §6. The superseding obligation + +**OBL-BETA-RELEVANCE** (successor of OBL-BETA-MODZONE, registered): prove +P(f(T) ∈ (b − ℓ, b] ∧ ρ_T ∈ (2r, d₀)) ≤ C·r³ — equivalently an O(ℓ) bound on the *S-adjacent* +band-maximum intensity — by a selectivity mechanism. Candidate mechanisms recorded at freeze +of the successor (not adjudicated here): (i) slab-traversal geometry: the branch must ascend a +length-(d − r/2) path inside the ℓ-slab whose transverse ribbon width is ~ℓ/|∇m_⊥| ~ ℓ/(κ_g d); +(ii) crest-decorrelation: the crest-line's field decorrelates along its length at the kernel +scale, making the joint in-slab event multiplicatively costly; (iii) the exact terminal-height +law via the AO-machinery of Lemma LB-ARCH run in reverse (upper barrier). The measured truth +this must explain: terminal support edge = 894ℓ (C021, 20k+ flows/rung). + +## §7. Labels + +§1: adjudications, FAILED-AS-WRITTEN ×2, recorded per protocol. §2: derived-and-verified +(exact KR identity + deterministic mp arithmetic + four-rung collapse). §3: derived-and-verified +(three-rung collapse 0.2–0.9%). §4: derived (mechanism narrative on top of §2–3 facts). +§5: adjudicated contra-indication with explicit scope caveats. §6: registered obligation. diff --git a/incoming/chatgpt-c022-q0-recovery-20260928/context/Q0_LEDGER.md b/incoming/chatgpt-c022-q0-recovery-20260928/context/Q0_LEDGER.md new file mode 100644 index 000000000..898c52ad6 --- /dev/null +++ b/incoming/chatgpt-c022-q0-recovery-20260928/context/Q0_LEDGER.md @@ -0,0 +1,5245 @@ +# Q0 LEDGER — CONSOLIDATED CORRECTION, KILL, SUPERSESSION, AND PRE-REGISTRATION RECORD + +**Discipline:** append-only. Nothing here is ever edited; corrections to corrections are appended. +**Scope:** every correction ledger, adjudication, kill, supersession, reconciliation, discharge, and cycle pre-registration freeze from the 488-file archive, verbatim, in chronological order — followed by the consolidation-event entries (E-CONS-1..6) and the duplicate/twin tables. +**Companions:** Q0_MASTER.md (frozen canon), q0_machine.json (registries, contracts, data, 488-file manifest), q0_verify.py (executable verification layer). + +## PART A — CORRECTION AND ADJUDICATION SPINE (verbatim, chronological) + + +--- + + + +# C024 — Corrections of the reviewer's own record + freeze: independent estimator verification and factor localization of the r³ law +**Date:** 2026-07-10 · **Author:** the in-loop reviewer (Claude), correcting its own prior-turn statements per supersede-never-overwrite. + +## Part 0 — Corrections (filed before any computation) + +**E1 — SUPERSEDED: my C023-turn summary of the upper side.** I stated Lemma MODZONE gave "a rigorous O(r³) ceiling on the moderate zone … modulo (ND′)" and concluded "upper rate: proven modulo (ND′)." Wrong on both counts against the C023 record: (i) C023 proved the exact moderate-zone count DIVERGES (γ = 1.882, four rungs); the per-rung masses 2.07/7.31·r³ grow in r³ units; the r-uniform upper side is OPEN (OBL-BETA-RELEVANCE). (ii) The (ND′) dependency was my own import, not C023's. Corrected status: positivity proven (C022); per-rung envelopes rigorous at the two anchors; BOTH r-uniform directions of Θ(r³) open — upper via terminal selectivity, lower via the rate of ∫Λ·AO. + +**E2 — VOID: my "Fix it"-turn evidence.** The scaling probe (ratio 6.05 vs 4.0) used the slice-at-b intensity — the estimand retired by C020 G-U2 (the needle). The "0.3σ/0.0σ centered" numbers used raw-f moments, not the gradient-conditioned (μ_t, s_t). Both void. The centering conclusion stands only via C022's C2 certificate (P_win = 1.000000, correct object). My "the anisotropic blow-up is not clean" is UNSUPPORTED-BY-EVIDENCE (undetermined), and is hereby withdrawn as a finding. + +## Part 1 — Freeze: independent Λ-exact reimplementation (verification instrument) + +New implementation from scratch: mp-Hermite analytic-BF covariances (dps 40), mp Schur conditioning, float64+eigh-clip MC for Hessian factors. No code shared with c019/c020/c021 instruments. Estimand definitions copied from Lemma UB0 §5 verbatim: Λ(y) = φ₂(∇f(y)=0 | 𝒥₆, with mean) · [Φ((b−μ_t)/s_t) − Φ((b−ℓ−μ_t)/s_t)] · E[W₃1_typed | 9 pins @ v*=clip(μ_t)] / E[W₂1_typed | 𝒥₆]; typing (detM>0, trM<0, detS<0, detY<0); window (b−ℓ, b); pins ((b,0,0),(b−ℓ,0,0),(v*,0,0)). + +**Anchors and pass bands (recorded values of the program; deterministic unless marked MC):** +| ID | object | station (r; y) | recorded | band | +|---|---|---|---|---| +| A1 | s_t/ℓ | 0.05; (−0.063, 0.012) | 0.0676 | ±0.0015 | +| A2 | P_win | same | 1.000000 | ±1e−4 | +| A3 | s_t/ℓ | 0.025; (−0.030, 0.007) | 0.0864 | ±0.0015 | +| A4 | s_t/ℓ | 0.025; (−0.0315, 0.006) | 0.0670 | ±0.0015 | +| A5 | C1 λ_min(Σ_∇) | 0.05 y*; 0.025 y* | 1.156e−6; 8.939e−8 | rel ±2% | +| A6 | C3 λ_min(6-pin Hess 6×6) | both y* | 2.603e−10; 4.068e−12 | rel ±5% | +| A7 | C4 λ_min(9-pin Hess 9×9 @v*) | both y* | 1.948e−13; 3.095e−14 | rel ±5% | +| A8 | C5 λ_min(9-pin Gram) | both y* | 1.321e−12; 3.333e−14 | rel ±5% | +| A9 (MC) | E[W₂1_typed \| 6] | r = 0.025 | 0.00202046 ± 2e−6 | ±1% | +| A10 (MC) | Λ | 0.025; (−0.0315, 0.006) | 0.090989 | rel ±6% | + +Adjudication rule: any deterministic-anchor failure ⇒ STOP, convention audit, report discrepancy as the finding; Phase 2 conclusions blocked. A9/A10 in (6%, 12%] ⇒ FAILED-AS-WRITTEN with diagnosis; > 12% ⇒ instrument-discrepancy alarm. + +## Part 2 — Freeze: factor localization at the fixed rescaled station +Station ỹ* = (−0.76, 0.24) in (y − M)/r coordinates — identically the C022 r=0.05 y* and the C021 r=0.025 sanity point. Rungs r ∈ {0.05, 0.025, 0.0125} (the 0.0125 rung is NEW, labeled as such). Report per rung: φ₂, det Σ_∇∇, Mahalanobis q, (μ_t−b)/ℓ, s_t/ℓ, P_win, den = E[W₂1|6], num = E[W₃1_sad|9], num_max = E[W₃1_max|9] (the β-side object), R_H = num/den, Λ; the EXACT (Schur, no MC) 9-pin conditional mean Hessians and their dets. +**Planted gate G-D1:** den(r=0.05) = 0.00808 ± 8% — the r² law predicted from the recorded r=0.025 anchor via the mechanism |det H_M| ≈ κr·|f_ss| forced by the ∇-pin divided differences (f_tt(M) = −(r/2)f_ttt + O(r²) with f_ttt cubic-pinned). Failure kills the mechanistic account, not the data. +**Stability gate G-D2′:** per-factor log₂-slopes consistent across the two octaves to ±0.35. +All component exponents are RECORDED DISCOVERY (no committed split); indicative hypothesis H1 (recorded, not gated): R_H ~ r¹-class via the shared depletion of all three dets. + +## Part 3 — Freeze: needle-sweep profiles (L2 inputs) +μ_t(y) along both axis directions through ỹ*, ±5 steps of 0.1r, three rungs; report (μ_t−b)/ℓ and the max sweep slope in ℓ-per-r units; derived ribbon width ℓ/|∇μ_t|. Recorded discovery. + +## Part 4 — Deliverable +Lemma LB-RATE architecture package: the reduction, the L1/L2/L3 certificate families with computed per-rung inputs, the honest open core, and the terminal-height-law correspondence between the two frontiers (classified as a structural correspondence of estimand families, NOT an isomorphism; prior-art search pending — no network in this environment). No claim of closure. + + + + + +--- + + + +# Corridor_Mechanism_KILL — kill-registry entry: the O(1)-halo deployment of the LB0 mechanism + +**Registry class: FAILED-AS-WRITTEN / SUPERSEDED-IN-SCOPE** (no rescue). Date 2026-07-09. +Scope chain: C017 (assembly) → C018 (kill evidence) → C019 (replacement architecture measured). +What is killed is the **deployment region**, not the mechanism: Lemma LB0 remains **Proved** and the +exact conditional Kac–Rice computation behind c₁ remains **Computed**; both are retained as +correct-but-misdeployed and re-enter corridor-locally. + +## 1. The claim as written (C017) + +C017 assembled: 1 − q ≥ c₁ · c₂ · r³ with + c₁ = 0.113 = (1/6) ∫_{Ω_halo} λ^{pinned}(y) dA over the opposite-halfplane halo at O(1) distances, + c₂ = inf_{Ω_halo} (P_adj · P_older), field-estimated c₂ ≈ 0.93, +yielding 1 − q ≥ 0.105·r³ "at field c₂", with the rigorous c₂ > 0 left as the sole obligation +(OBL-LB-ADJ). The deployment presumed the trigger mass and the separatrix connection both live in the +O(1) halo. + +## 2. The kill evidence + +- **C018 (A-side):** the pair-Palm ascent-adjacency A(y) ≡ 0 at every measured O(1) station (24-point + polar grid, ρ ∈ [0.8, 4.0], all θ; trigger-weighted ⟨A⟩ = 0): the ascent connection to M does not + exist at O(1) — c₂(Ω_halo) = 0 as a measured fact. P-L2 adjudicated branch (iii): corridor-scale + collapse. The C017 assembled number 0.105·r³ is therefore **vacuous as deployed** (0.113 × 0 = 0). +- **C018 corollary:** the Palm-corrected trigger coefficient over the halo is c₁^{Palm} = 0.0643 ± + 0.0008 (P-L1), vs the pinned 0.113 — the pinned/Palm weighting-scope error, first caught here, + foreshadowing OBL-UB-PALM. +- **C019 (λ-side):** the window-density factor independently annihilates the intermediate zone + (Mahalanobis q = 416 at (0.2, 0.025); log φ ~ −10⁶..−10¹⁰ on the axis; λ ~ e⁻²⁰⁰ at inner rungs) — + even if ascent were present, no trigger mass reaches from the corridor to O(1): a **death valley** + between the arch (τ ≲ 2r) and the O(1) zone, with λ recovering to ~0.005 only at ρ ≳ 0.8 where A = 0. +- **Corridor scale correction:** C018's committed r^{3/2}-scale probes (P-L4/P-L5-band/P-L6-band) all + FAILED-AS-WRITTEN; C019 measured the true support: an arch/filament at τ ≈ 0.75r–1.5r behind/over M — + the scale is **O(r)**, not O(1) and not r^{3/2}. + +## 3. What survives, and where it goes + +| object | status | redeployment | +|---|---|---| +| Lemma LB0 (window-saddle + separatrices ⟹ D(M) ≠ S) | **Proved**, untouched | applied on the arch, where both separatrix conditions hold with measured probability 0.992–1.000 | +| exact conditional Kac–Rice machinery behind c₁ | **Computed**, correct | becomes the factor-1/factor-2 algebra of the lam3 disintegration (C019) | +| c₁ = 0.113 (pinned, halo) | superseded-in-scope | replaced by the Palm arch integral: (1/6)∫λ·A·O dA = **1.10 ± 0.09** (E1_extended, C019) | +| c₂ = inf(P_adj·P_older) over Ω_halo | killed (measured 0) | replaced by the measured on-arch product A·O = 0.992–1.000; the rigorous positivity obligation migrates | +| OBL-LB-ADJ (prove c₂ > 0 on Ω_halo) | superseded | → **OBL-LB-ARCH**: prove ∫_arch λ·(A·O) dA > 0 (positivity, not sharpness) — strictly easier: A = 1 on the arch is a separatrix-identity statement (the filament IS the D(M)-witness locus), reducible to LB0 + R0 + a nondegeneracy at one arch point | +| Bonferroni step (OBL-C017-2) | discharged | LB_Bonferroni.md (this delivery) | + +## 4. Reconciliation with the standing record + +- **C010** (field, moderate r): the G-trigger lower bound 0.18·r³ = (1/6)·p_∞·0.93 is a *gain-channel + subset* of the arch count; 0.18 < 1.10 ✓ (a weaker bound, not a contradiction). C010's preemption- + given-trigger 0.93 is the moderate-r shadow of the on-arch A·O ≈ 1. +- **C013** (Lemma ADJ): pair-Palm-weighted adjacency 1.0000 at all rungs — consistent with (and + explained by) the corridor-local A ≡ 1; C013's adjacency was measured *at pair scale*, i.e., inside + the corridor, never at O(1). +- **C008** (τ-shell law): outermost-shell dominance of the pinned annulus count — the pinned shadow of + the same outward mass migration whose Palm completion is the arch; the C008 annulus (τ ≤ r/2) captures + the arch's inner flank only, which is why C_annW = 0.188 under-counts the Palm mass ×~6 and C_strip + ×~25 (→ OBL-UB-PALM). +- **C016**: the O(1)-zone R(ρ) profile (suppressed ρ < 1, peak ρ ≈ 1.5) probed 10× outside the arch; + no overlap, no conflict; its "asymptotic far intensity = unconditional" conclusion stands for the far + channel. +- **q(r) field law**: c_eff(0.7) ≈ 1.31, c_eff(0.4) ≈ 1.7 bracket the asymptotic 1.10 from above as + moderate-r laws should. The arch coefficient is the first *asymptotic-slice* measurement of the sharp + constant. + +## 5. Registry lines (verbatim for the ledger) + +- KILL: C017 regional deployment {c₁(halo,pinned) × c₂(halo)} — FAILED-AS-WRITTEN (c₂ measured 0; + trigger mass absent at O(1) under Palm). Evidence: C018_Observed_Update (P-L2 branch iii, ⟨A⟩ = 0), + C019_Observed_Update (death valley; arch). +- KILL: corridor scale r^{3/2} (C018 committed P-L4/5-band/6-band) — FAILED-AS-WRITTEN; measured scale + O(r) (arch at 0.75r–1.5r). +- RETAIN: LB0 [Proved]; conditional-KR machinery [Computed]; C013 adjacency [Measured, in-corridor]. +- SUPERSEDE: c₁ 0.113(pinned,halo) → 0.0643 (Palm, halo; C018 P-L1, no load) → arch coefficient 1.10 + (Palm, corridor-local; C019 E1_extended) — scope chain explicit, nothing overwritten. +- NEW OBLIGATIONS: OBL-LB-ARCH (positivity proof), OBL-UB-PALM (upper-bound constants under Palm), + OBL-C019-1/2, OBL-LBB-1, OBL-R0-UNIF (see Registry_Reconciliation.md). + + + + + +--- + + + +# C014-KILL: The Third-Point Max Failure Channel Vanishes as r → 0 (moderate-r contamination corrected) +**Artifact class:** kill-registry entry + correction (corrections supersede, do not overwrite). **Chain:** freeze C014p2 724b2636…8a7b; validated instrument joint_sim3.py; confirming runs at r ∈ {0.2, 0.1, 0.05}, tilts m ∈ {±1/2, ±3/2}. + +## 1. What triggered it +The C014 part-2 Palm integrator returned num3 = 0 (empty numerator) at both tilts, while the validated failure fraction J was clearly nonzero. Instead of reporting Φ = 0, the empty numerator was treated as a diagnostic. Radial analysis of the partners resolved it. + +## 2. Root cause (measured, not argued) +Points x_M = (−r/2, 0), x_S = (+r/2, 0) sit at separation r, which in r²-units is **1/r**. At r = 0.2 that is exactly **5 r²**. The C014 partner search validity extended to 10 r² = 2r, so **M lay inside the third-point search ball**. Measured max-partner distances: median **5.0 r² at both tilts, i.e. exactly M**. The "max third-point" was the pair's own maximum; "S-separatrix branch captured by the max-partner" was substantially "branch reaches M" = **adjacency, misclassified as failure**. The P-K2 gate passed only because both the part-1 and the corrected part-3 instruments carry the identical contamination — agreement between two instruments does not certify freedom from a shared physical mis-scoping. + +## 3. Decisive small-r test (M excluded) +With the search restricted below M (validity < 8 r², partner ≥ r/4 from M): +| r | M at | max-partner fraction | location | +|---|---|---|---| +| 0.2 | 5 r² | 0.09 / 0.21 | 5.0 r² = **M** (contamination) | +| 0.1 | 10 r² | 0.013 / 0.016 | 7–8 r² (M-basin edge) | +| 0.05 | 20 r² | **0.000 / 0.000** | none | +Exhaustive whole-gap sweep at r = 0.05 (tilts ±1/2, ±3/2; radii to 16 r²): **max third-points at frac ≤ 0.0003**, the lone residual at 13 r² (M-basin outskirt). As the scales separate, the nearest non-S critical is always a **saddle or a min**. Mechanism: the natural nearby criticals of a tilted saddle are the transverse cubic's second root (a saddle) and mins; a genuine extra max requires higher-order structure and is O(r^{>0})-suppressed. + +## 4. What is KILLED / SUPERSEDED +- **C014 J = 0.089 / 0.199 as a "deep-tail failure constant" → SUPERSEDED.** It is an r = 0.2 M-contamination artifact; the genuine small-r third-point failure fraction through this channel is **→ 0** (only max-partners produce branch-capture failure, and max third-points vanish). +- **"J = p_max exactly" → re-scoped.** True as an r = 0.2 identity, but p_max there was M's contamination fraction; p_max → 0 at small r, so it is not a deep-tail law. +- **C014 part-1 "deep-tail adjacency ≈ 0.23–0.28" revision → RETRACTED.** adjM was deflated by the spurious partner-vs-M competition for the same M-region events. **C013's clean pair-adjacency ≈ 0.92 on the |μ_S| ≤ r set (gated D-A1′) is REINSTATED** as the operative B2 tail-adjacency number. + +## 5. What SURVIVES +- **The joint instrument (partner sweep + two-phase germ flow) is validated** — the P-K2 regression gate passes (batched J = 0.089/0.199 vs part-1 0.091/0.198, |diff| ≤ 0.002). It is a correct tool that was run at a contaminated regime; the analytic-seed partner search (part-2) was separately a bug and is retired in favor of the angular sweep. +- **The two-sided LOWER BOUND is unchanged.** Its positive constant was never from a third-point channel — it comes from the C010 field-level preemption measurement P(D≠S | G∈(0,ℓ)) = 0.93–1.00 × the inner window count, giving 1 − q ≥ ~0.18 r³. That stands untouched. +- **Positive structural content:** at small r the third point, when present, is a saddle or a min — neither disrupts the elder-rule S–M pairing (a saddle does not trap an ascending separatrix; a min is H0-superlevel-inert). So **third points do not generically break the pairing**; adjacency is robust. This *reinforces* that the genuine failures are the r-scale channels C010 measured (window-saddle ≈ 64%, W₂ ≈ 26%), not an r²-third-point disruption. + +## 6. Corrected inner-law status +E_Palm[N_fail-inner] is set by the r-scale mechanism (C010), not a third-point factor. The inner window count C_innerW = 11.9 (C012) is multiplied by the field-measured preemption, not by a third-point J. The two-sided statement is exactly as before the C014 attempt: **0.18 r³ ≲ 1 − q ≤ [2(C_annW + C_strip) + C_far(L,b)] r³ (1+o(1))**; C014's contribution is a *null* refinement (no extra channel), logged. + +## 7. Epistemic labels +Finding (max third-point vanishes as r → 0): **[Verified-numerically]**, robust across tilts/radii, with a clear suppression mechanism; a proof would bound the conditional intensity of a max in the r²-annulus as O(r^{k>0}). The kill of the C014 J/adjacency numbers: **[Retracted]**, root-caused. C013 reinstatement: **[Verified]** (unchanged, D-A1′-gated). + + + + + +--- + + + +# C032 — Lemma TC: the Terminal-Counting Supersession; the Mean-Rim Discovery; the AB Negative Result +**Freeze:** 1487c20c…2cae. **Provenance:** C031's G2 selection-obstacle analysis exposed that selection is unnecessary — the C025 step "terminal ≥ v* = b − ℓ/2 ≥ b − δ₀" discarded the ℓ-scale start-height information. + +## Lemma TC (established at derived-structure + station-certified grade, modulo R0) +Modulo R0 (terminal of an ascent is a local max — the architecture item every AO statement already inherits): {outward terminal ≤ b} ⊆ {separatrix loop: outward branch terminates at M} ∪ {∃ local max with value ∈ [v*, b] on T²₂₄ ∖ collar(M)}. Hence, with the band width 0.49993·ℓ (m = −0.4999290, C026): + +P(terminal ≤ b | 9 pins) ≤ P(sep-loop) + **2.1·(ℓ/2)** [B₅; C030 KR-MB] + **ρ_mx(1.2)·0.49993·ℓ·(L² − π·25)·E_sup** [exterior] = P(sep-loop) + **2.03·r³** + +with ρ_mx(1.2) = 0.043685 DERIVED (C030 G-F1 two-route) and **E_sup = 1.0235 station-certified this cycle (G-H1 PASS):** conditional/unconditional band-max intensity ratio ∈ [0.98, 1.014] at d ∈ {5, 6, 8, 12} × 2 angles × 2 rungs (deviations = MC noise; v = 1 − O(1e-8) by d = 5, pull monotone → exact-zero underflow by d = 8); torus periodization ≤ e^{−72}, negligible. The sep-loop event is FOLDED into the flow-topology named family. The magic is the start height: the outward branch launches at mid-window, so only an ℓ/2-thin terminal band can fail — the far term is O(r³) **by counting alone**, no crossing selection, no Palm disintegration, no barrier in the load path. + +## Supersession ledger (corrections supersede, never overwrite) +- **OBL-FAR-COMPOSE → CLOSED-BY-SUPERSESSION** (registered at C031 §4; both routes moot for the assembly; the Rice constant 2.43 remains as context). +- **C028 far-ascent theorem (p̄(1/5) ≤ 1 − c) and C031 R3′ (conditional-barrier transfer, 36× margin) → STANDALONE results:** correct, gate-passed, no longer load-bearing for AO. +- **Lemma FD → repositioned:** exterior-transfer evidence (now superseded in that role by the direct G-H1 intensity ratios) + the quantitative channel. +- **OBL-P0-FLOOR → moot for the assembly** (standalone-optional). +- **Reconciliation:** the measured p̄(0.05) = 0.0334 and the Rice-composed 0.081 were honest bounds for the δ₀-framed event; the framing was lossy because δ₀ was fixed while the true start height is b − ℓ/2. No prior number is wrong; the architecture is superseded by a sharper inclusion. + +## The mean-rim discovery (vacuity check FAILED-AS-WRITTEN → finding) +The freeze's sup-of-mean check (m₉ ≤ b, max at M) FAILED: **m₉ exceeds b at every station outside the pin cluster** — 1.2007 at d = 0.2, rising to a rim of **1.40–1.57 at d = 1.3** (anisotropic: higher toward the arch side, angle 2.35), decaying to 0.0003 by d = 5 (C031 table). The check's premise (my pre-freeze narrative) was wrong; the instrument was right. The 9-pin cluster sits on the shoulder of a mean bump — the MEAN side of the same jet-cluster physics whose VARIANCE side is the C030 horizon. Consequences: (i) the above-b-saddle O(1) count is explained (the mean itself is above b in the zone); (ii) AO ≈ 1 acquires its mechanism (the outward ascent climbs mean-uphill; qualification is the default; capture 1.0000 explained at mean-field level); (iii) the sep-loop event requires descending the rim back to exactly b at M — strongly disfavored (evidence for the named item, not a closure); (iv) band-maxima in the rim zone require downward fluctuations against tiny v — the C030 transition-zone numbers are conservative. + +## The AB negative result (G-H2; pre-registered P1 CONFIRMED, band hit) +Zone-integrated crude count E[N_sad(f > b, 0.15 < |x| < 1.5) | 9] = **0.0325 (r = 0.05) / 0.0260 (r = 0.025), rung ratio 1.25 ∈ [0.5, 2]** — O(1) in r, not ≈ 8. Station profile: exact-zero underflow for d ≤ 0.35 (rigidity: mean 1.2007-pinned above the threshold's reach... the field there is deterministically ABOVE b — no saddle-below), 1e-12 at d = 0.7, growing to 1e-2 at d = 1.3 where the rim's fluctuations live. **The counting route for the above-b-saddle diversion channel is DEAD** (my pre-freeze mechanism story — gap ∝ r·d² — is superseded by the rim mechanism); the channel remains in the flow-topology family with this quantitative context: any closure must use the pass-connectivity/flow structure, not saddle counting. + +## The new AO assembly and LB-RATE +**AO ≥ 1 − WP(0.213·r³) − TC(2.03·r³) − [flow-topology named family] − [KR validity + collar exclusion] , modulo R0** += **1 − 2.24·r³ − named**. At r = 0.025: AO ≥ 0.999965 − named; at r = 0.05: ≥ 0.999720 − named (G-H3 PASS: ≪ the superseded 0.081; consistent with the direct C021 AO ≈ 1). +**LB-RATE:** 1 − q ≥ C*(r)·r³·(1 − 2.24·r³)·(1 − O(r³)) ≈ **0.946·r³ at r = 0.025 (measured grade)** — the assembly now MATCHES the direct two-sided measurement C* ≈ 0.96 ± 0.06 instead of undershooting at 0.87; limit form 0.9091·r³·(1 − o(1)) with the named set. + +## The revised named set (11 items; net −2, one family consolidated) +1. R0 / MS–Sard covering (architecture; carries TC's "terminal is a max"). 2. γ-LOC tube-local (architecture). 3. (ND′) chain note (upper side). 4. KR validity + collar exclusion (WP, KR-MB, TC-exterior). 5. **Flow-topology in the rigid zone** (consolidated family: ridge split + separatrix loop + above-b-saddle diversion; quantitative context: crude count ~0.03 r-free; mean-rim disfavors all three). 6. OBL-BETA-RELEVANCE (upper side, out of scope). 7. 9-pin certificate tightening. 8. Collar residual |ỹ| < 0.15. 9. Station-density/on-grid class (C027 grid; R3′ slab; G-H1/G-H2 stations). 10. Λ-side grid→continuum formality. 11. Prior-art searches (network-gated; now ALSO covering the mean-rim structure — analytic-kernel interpolant theory is a known field; NO novelty language used). +Removed: OBL-FAR-COMPOSE (superseded), OBL-P0-FLOOR (moot-for-assembly, standalone-optional). + +## Gate ledger +G-H1 PASS (10 stations, both rungs, ratio band + monotonicity). G-H2: P1 CONFIRMED at band center (ratio 1.25); **sup-of-mean vacuity check FAILED-AS-WRITTEN → the mean-rim finding registered** (premise wrong, instrument right); tail quadrature 7 nodes × 6k draws per station, 20 stations. G-H3 PASS (both consistency clauses). Scope: BF, b = 6/5, L = 24, arch station; two angles per distance (angular sweep optional). + + + + + +--- + + + +# Correction Addendum to window_lemma_v2_patch (C009) — supersedes in part, never overwrites +1. **Section I (Lemma I): RESOLVED**, β = 5 at derived+verified grade; see Lemma_I_Package.md. The candidate β = 4 was an underestimate (M-side determinant drag adds one power). The sympy verification plan proposed in the patch was executed and extended far beyond (certified exact table, gates, two-weighting measurement). +2. **Section I (Localization): SUPERSEDED in its exp(−c/r²) form.** Conditional-at-fixed-w only; the Palm-averaged shell law is polynomial r^{2γ+1} (C008). All downstream assembly conclusions survive; the exp claim is retired to the kill-log class of R3. +3. **Section A′: SCOPE-CORRECTED.** Valid off the vertical strips through M and S; on the strips the count is polynomial (measured tail 2.95 on [r/2, 4r], geometric decay beyond, far-floor crossover at absolute O(1)). The κ-pinned t-floor makes the off-strip suppression degeneracy-proof; the strips expose the μ-flat direction and are now an explicit measured component of the peak zone. +4. **Assembly line replaced:** "C r^β (inner)" → the seven-zone table of Lemma_I_Package.md §3; conclusion 1 − q = O(r³) unchanged, now with every polynomial component verified or measured. + + + + + +--- + + + +# Correction Addendum to window_lemma_v2_patch (C009) — supersedes in part, never overwrites +1. **Section I (Lemma I): RESOLVED**, β = 5 at derived+verified grade; see Lemma_I_Package.md. The candidate β = 4 was an underestimate (M-side determinant drag adds one power). The sympy verification plan proposed in the patch was executed and extended far beyond (certified exact table, gates, two-weighting measurement). +2. **Section I (Localization): SUPERSEDED in its exp(−c/r²) form.** Conditional-at-fixed-w only; the Palm-averaged shell law is polynomial r^{2γ+1} (C008). All downstream assembly conclusions survive; the exp claim is retired to the kill-log class of R3. +3. **Section A′: SCOPE-CORRECTED.** Valid off the vertical strips through M and S; on the strips the count is polynomial (measured tail 2.95 on [r/2, 4r], geometric decay beyond, far-floor crossover at absolute O(1)). The κ-pinned t-floor makes the off-strip suppression degeneracy-proof; the strips expose the μ-flat direction and are now an explicit measured component of the peak zone. +4. **Assembly line replaced:** "C r^β (inner)" → the seven-zone table of Lemma_I_Package.md §3; conclusion 1 − q = O(r³) unchanged, now with every polynomial component verified or measured. + +## Entry 5 (C011) +**Lemma B1 KILLED** (band event is typical: IVT on ∂B_{3r} with f(M) = b inside and transverse dips below b−ℓ; the exp(−c/r²) claim is off by everything — kill registry K-B1). **Corollary B2 replaced by loop absorption:** A_loop = {M₂ ∈ C at (b−ℓ)+} ⊆ {a window saddle exists}, since the max-min pass between M and M₂ lies in (b−ℓ, b); the loop term vanishes from the assembly, which simplifies to 1 − q ≤ E[N_w] + P(A₀) = O(r³). Field-level confirmation of the loop channel: C010 records, died-below-S ∧ f(M₂)-∉-window, rate 0.0014 at r = 0.7. See Obligations_Registry_Discharge.md §8. + + + + + +--- + + + +# The Obligation Registry and Formal Discharges (C011) +**Artifact class:** proof-grade consolidation. **Chain:** C011 snapshot a7b84d5a…4f91; companion ledger hashed. This document enumerates every open formal obligation across the Theorem A chain (FlatSaddle U1–U5, v2-patch R0/H/G/P/B1–B2, C008 general-γ transfers), discharges what is dischargeable at proof grade with the certified constants as named inputs, registers one kill with its replacement, and updates the master dependency graph. + +## 0. Registry summary + +| obligation | source | status after C011 | +|---|---|---| +| (U1) conditional Kac–Rice | FlatSaddle §5 | **DISCHARGED** (§2) | +| (U2) L^{2+ε} remainders | FlatSaddle §5 | **DISCHARGED** (§3) | +| (U3) tail-density Gram | FlatSaddle §5 | DISCHARGED C009: det = (1/540)r¹⁸; ξ-part via Lemma H | +| (U4) Palm-base uniformity | FlatSaddle §5 | **DISCHARGED** (§4) | +| (U5) adjacency ≤ 1 | FlatSaddle §5 | closed by definition (upper bound) | +| R0/R2 separatrix + merge | v2 patch | **DISCHARGED** (§1) | +| Lemma G far-zone uniformity | v2 patch | **DISCHARGED** (§5) | +| Lemma H | v2 patch | **PROVED for BF/positive spectral density** (§6); general (H2′) cited-with-conditions | +| Lemma P finite write-out | v2 patch QED-plan | OPEN, finite; precise sympy spec registered (§7) | +| Lemma B1 (band) | v2 patch | **KILLED** (§8); replaced by B1′ | +| Corollary B2 (loop) | v2 patch | **REPLACED**: A_loop absorbed into the window count (§8) | +| general-γ (U1)–(U4) | C008 | transfer by §§2–4 verbatim with τ = r^γ (all inputs certified bi-graded) | + +## 1. R0/R2: separatrices and merge identification — discharged + +(a) *A.s. Morse.* Under Lemma H the joint law of (∇f, ∇²f)(y) is nondegenerate for every y; by the Bulinskaya-type argument (Azaïs–Wschebor), a.s. no point has ∇f = 0 with det ∇²f = 0, and f is a.s. C² Morse on the torus. Critical values are a.s. pairwise distinct: for each pair type, Kac–Rice over pairs (y₁, y₂) with the codimension-one constraint f(y₁) = f(y₂) has vanishing expected count under Lemma-H nondegeneracy of (∇f(y₁), ∇f(y₂), f(y₁) − f(y₂)). +(b) *Separatrices.* At a nondegenerate saddle, the ∇f-flow's unstable manifold is one-dimensional with two branches (stable-manifold theorem); on a compact surface every forward orbit's ω-limit is a critical point (Łojasiewicz not needed — Morse suffices), of strictly higher value. A.s. no saddle-to-saddle connection: fix the countable family of saddle pairs; a connection is destroyed by an arbitrarily small Cameron–Martin shift in a direction altering the value gap along the connecting orbit; conditional on the orthogonal complement, the CM coordinate's law is absolutely continuous and the connection holds on a Lebesgue-null set of it, so P = 0. Hence both branches a.s. terminate at maxima: M₂ is well-defined. +(c) *R2.* Passing the single nondegenerate critical value f(S) = b−ℓ attaches a 1-cell at S to the sublevel-complement; the attaching endpoints are the two ascending-separatrix germs, so the superlevel components joined (or self-joined) at S are exactly those of the two terminal maxima. Merge identification proved. + +## 2. (U1): conditional Kac–Rice — discharged + +The pair conditioning is on six linear functionals L(f) = a; the conditional field is m_a + h with h Gaussian, covariance K minus a rank-6 correction. For counts over any region R excluding shrinking balls B_{ρ}(M) ∪ B_{ρ}(S): (i) paths a.s. C² (analytic kernel); (ii) the conditional law of ∇f(y) is nondegenerate for y ∉ {M, S} — degeneracy would make ∇f(y) an a.s. linear function of six functionals, excluded by Lemma H — with the degeneration rate as y → x_S given *exactly* by the certified gradient-block minors: Var_t Var_s − Cts² = τ⁴r²z_s⁴(1 + O(r + τ/r)) > 0 off the axis, with axis-collision integrability established by the C007-B cutoff-insensitivity record; (iii) E|det ∇²f| moments finite (Gaussian). Azaïs–Wschebor's conditional Kac–Rice then applies on every such R, and the ρ ↓ Cr² limit is dominated by the (U2) moment bounds. The three-point/two-scale application is precisely the C007/C008 assembly; its validity input is the C009 Gram (1/540)r¹⁸. + +## 3. (U2): L^{2+ε} uniformity — discharged + +Every O_P(·) remainder in (F1)–(F4) and in the certified mean/covariance expansions is an explicit polynomial (degree ≤ 4) in a fixed Gaussian vector of derivatives of order ≤ 4 at ≤ 3 points, with covariance entries given by the certified series, uniformly bounded on the Palm compact; λ₈ < ∞ (H1+) covers the fourth-order entries. By Gaussian hypercontractivity, ‖P(X)‖_{2+ε} ≤ C(deg, ε)‖P(X)‖₂, and the ‖·‖₂ scalings are the certified coefficients termwise. Uniformity over the tail set follows from the bounded conditional densities of (U3). ∎ + +## 4. (U4): Palm-base uniformity — discharged + +Conditional means are affine in a = (b, 0, b−ℓ, 0, 0, 0); conditional covariances are a-independent. Every ledger constant is an integral of continuous-in-(b, κ) integrands against Gaussian envelopes independent of (b, κ) on compacts; dominated convergence gives continuity and uniform integrability. ∎ + +## 5. Lemma G: far-zone window count, uniform — discharged + +E[N_w(far)] = ∫_{|y−x₀|≥δ} E[|det∇²f(y)| 1_win | ∇f(y)=0, pair] p_{∇f(y)|pair}(0) dy. Two inputs: (a) v(δ) := inf_{far} Var(f(y) | ∇f(y), pair functionals) > 0 — the variance is continuous in y, strictly positive pointwise (an eight-functional degeneracy is excluded by Lemma H), and the domain is compact (torus), so the inf is attained and positive; hence the conditional window probability is ≤ ℓ/√(2πv(δ)) uniformly. (b) The remaining integrand integrates to at most the conditional total-critical-point intensity, finite and continuous by the same nondegeneracy. Hence E[N_w(far)] ≤ C(δ, L)·ℓ, uniform on b-compacts. ∎ + +## 6. Lemma H for the testbed class — proved + +For BF the spectral density is everywhere positive. A linear relation Σ c_{j,α} ∂^α f(x_j) = 0 a.s. forces Σ c_{j,α}(ik)^α e^{ik·x_j} = 0 for a.e. k; the left side is an entire function of k (finite sum of polynomials times distinct exponentials), which vanishes identically only if all c = 0 (linear independence of {k^α e^{ik·x_j}} for distinct x_j). Hence joint nondegeneracy of any finite derivative family at distinct points. For general (H2′), the same holds whenever the spectral measure is not supported on a real-analytic variety annihilating such a family; adopted as the stated form of (H2′). ∎ + +## 7. Lemma P: the one remaining finite computation + +Open, with the exact spec now registered: compute (sympy, exact) the 6×6 covariance of the on-axis jet coefficients (c_{t³}, c_{t⁴}, c_{t²s}, c_{t³s}, c_{ts²}·grade, …) conditional on the four on-axis pin constraints {h, h_t at ±r/2} = 0, verify the r → 0 limit is positive-definite (the two-point analogue of the C009 (1/540) computation), yielding the pinned variance bounds sd(h) ≤ Cρ³, sd(h_t) ≤ Cρ²·ρ on the strip used by the off-strip A′ argument. Registered as the C012 candidate; the assembly's polynomial terms do not depend on it (only the off-strip exponential grade does). + +## 8. The B1 kill and what replaces it + +**Kill (K-B1, IVT argument).** B1 claimed P(∃ y ∈ ∂B_{3r}: f(y) ∈ (b−ℓ, b]) ≤ C exp(−c/r²). False: f(M) = b with M's window component of O(1) diameter forces f > b−ℓ somewhere on ∂B_{3r} with probability ~ 1, the transverse arcs dip below b−ℓ, and continuity on the circle forces window values at the crossovers. The band event is *typical*. Root cause class: a containment bound whose complementary event was analyzed at fixed generic geometry (the μ ≍ 1 transverse suppression) — the same μ-flat-direction blindness as the R3/Localization kills. + +**Replacement (B1′, and B2 absorbed).** By R2, A_loop = {the S-passage is a self-merge} = {M₂ ∈ C at (b−ℓ)+}. If so, M and M₂ are connected in {f > b−ℓ}, and the max-min pass between them is attained (a.s.) at a saddle with value in (b−ℓ, b) — strictly below b because at level b−ε the components of M and M₂ are disjoint germs of their maxima. Hence **A_loop ⊆ {a window saddle exists} ⊆ {N_w ≥ 1}**: the loop event is *absorbed into the window count already bounded at O(r³)*; no separate band/loop budget exists in the assembly. Field-level confirmation: the C010 died-below-S ∧ f(M₂)-∉-window subpopulation (the predicted loop signature, partner elder in all members) is observed at rate 0.0014 at r = 0.7 — one event, exactly the Θ(ℓ)-class rarity expected. + +**Corrected assembly.** 1 − q ≤ P(N_w ≥ 1) + P(A₀) ≤ E[N_w] + P(A₀) with E[N_w] = O(r³) by the seven-zone table (C009) and A₀ (a second critical point of value ≥ b inside C) a window-free O(ℓ)-type event bounded by the same counting at the level b. The exponential-zone lemmas (off-strip A′, Lemma P) now affect only interior refinements, not the rate. + +## 9. Master dependency graph after C011 + +PROVED: R0/R2; (U1)–(U5) [with (U3)'s exact constant]; Lemma G; Lemma H (testbed class); B1′/loop absorption. DERIVED + VERIFIED: inner r⁵ law; γ-family shell law; annulus r³ boundary-dominance; the r³ assembly; Lemma LB preemption. MEASURED: window constants (0.51 core, 0.83–0.85 boundary, 0.24 strip ratio, acceptance profiles); failure taxonomy; p∞(b) ≈ 1.19 first estimate; loop-channel existence. OPEN: Lemma P finite computation (spec §7); the c·ℓ deep-scaling verification of Lemma LB (importance-sampled instrument); the exact-constant/two-sided inner law (B2 adjacency, with its three target constants); saddle-channel constant c_saddle. + + + + + +--- + + + +# Lemma P, Discharged Exactly: Cubic Pinning, the Slab Scope, and the Octave Law (C012) +**Artifact class:** final formal-obligation discharge. **Chain:** C012 snapshot 3473ca55…426c; companion ledger hashed. With this document the obligation registry of C011 §0 has no remaining OPEN formal entries; what remains open in the program is research (constants, two-sided law), not proof debt. + +## 1. The limiting σ-algebra (the mechanism, made exact) + +As r → 0 the six pair functionals {f, f_t, f_s at (±r/2, 0)} converge, through their symmetric/antisymmetric divided differences, to the pinning of **{f, f_t, f_tt, f_ttt, f_s, f_ts}(x₀)** — value, full t-jet to *third* order, and the transverse 1-jet. This is the cubic pinning: the on-axis field is observed to cubic order, so the on-axis residual opens at the quartic; the transverse direction is observed only to first order, so the transverse residual opens at f_ss — which is why the slab is polynomial and only the off-slab is exponential. + +## 2. Certified variance laws (two-route: hand algebra committed first, engine authoritative) + +All conditional variances proven **even in r** (P6: the conditioning set {±r/2} is r ↔ −r invariant; every odd-kr coefficient vanishes identically in the bi-graded output). Leading laws, exact: + +| quantity | law | hand route | engine | +|---|---|---|---| +| Var(f_t(τ,0) \| A), on-axis | (2/3)τ⁶ | CondVar(f_tttt \| f, f_tt) = 105 − 81 = 24; ÷36 | 2/3 ✓ | +| Var(f(τ,0) \| A), on-axis | τ⁸/24 | same 24; ÷(4!)² | 1/24 ✓ | +| Var(f_t(0,σ) \| A), transverse | σ⁴/2 | CondVar(f_tss \| f_t, f_ttt) = 3 − 1 = 2; ÷4 | 1/2 ✓ | +| Var(f_s(0,σ) \| A), transverse | 2σ² | CondVar(f_ss \| f, f_tt) = 2 | 2 ✓ | +| general direction, Var(f_t) | (0,4)-coeff = z_s²(z_s² + 4z_t²)/2 | — | exact; vanishes on-axis, handing to τ⁶ | + +Consequences used downstream, now at proof grade: on-axis sd(f_t) ~ τ³ against the mean floor κτ² gives the exp(−c/τ²) axis suppression; on-axis sd(f) ~ τ⁴ against the fold mean κτ³ gives the value-side suppression; the anisotropic 1/(|sinθ|+τ) degeneracy of 2-jet-only conditioning is removed exactly where claimed. + +## 3. P5: the limiting nondegeneracy — exact + +The conditional covariance of the six lowest unpinned jet coordinates (f_ss, f_tss, f_sss, f_ttss, f_ttts, f_tttt) given the six pinned functionals has diagonal **(2, 2, 6, 6, 6, 24)** and determinant **13824 > 0**, with all leading principal minors positive. This is the "limiting 6×6 covariance nondegenerate" of the v2 QED-plan, discharged with an exact integer. The diagonal's entries are precisely the constants that appeared independently in the certified S-centered table (Var(f_ss|A) = 2), in P3's input, and in P1/P2's input — three-route consistency. + +## 4. Scope sharpening of A′ (registered refinement of the C009 correction) + +Because the transverse residual opens at the *free* f_ss (variance 2), the exponential zone is **{|t| ≥ Cr} ∩ {τ ≤ δ}** — off the vertical *slab*, not merely off the two strips: the mid-line t ≈ 0 belongs to the polynomial zone as well. The C009/C010 strip measurements covered t ∈ [0, r] (mirror-symmetric to [−r, 0]), so the measured polynomial coverage is exactly the slab; no gap remains between the measured polynomial zone and the now-proven exponential zone. + +## 5. The octave law (derivation catching up to a measurement) + +In the deep slab (s = τ ≫ r): p_{f_t}(0) ~ 1/τ² (transverse σ⁴/2-law), p_{f_s}(0) ~ e^{−μ²/4}/(√2τ) (free-f_ss channel), value-window factor ℓ/(σ²-scale) ~ ℓ/τ², slab width ~ τ; assembling per octave: **count(τ-octave) ≍ ℓ/τ³**, geometric ratio (1/2)³ = **0.125 per octave**. The C010 measured ratios 0.08–0.14 bracket it. A law derived after its own measurement, matching it — registered as the bonus result of the cycle. + +## 6. Status + +Lemma P: **DISCHARGED** (variance laws exact and two-route; nondegeneracy exact; r-parity exact). The obligation registry is now clear of proof debt: every lemma in the Theorem A chain is proved, derived-with-certified-inputs-and-verified, or measured, with the open items being research objects (constants, the deep-r LB scaling, the two-sided inner law). + + + + + +--- + + + +# Registry_Reconciliation — obligations ledger and record hygiene (as of C020, 2026-07-09) + +Supersedes Registry_Reconciliation.md (C019 edition); companion to Constants_Table_Canonical_C020.md, +Lemma_UB0.md, and C020_Observed_Update.json. Supersede-never-overwrite: nothing below deletes an archive +entry; every change is a scoped supersession. + +## 1. Obligations ledger (complete, current) + +| ID | content | status | discharge / carrier | +|---|---|---|---| +| MS-Sard / R0 | a.s. Morse–Smale for the conditioned flow | derived-architecture | MS_Shift_R0_Closure.md; residue → OBL-R0-UNIF | +| OBL-R0-UNIF | uniformity of the 18×18 nondegeneracy + constant chase | OPEN (bookkeeping grade) | carried | +| OBL-UB-PALM | re-derive upper-bound channel constants under Palm weighting | **DISCHARGED (C020)** — by *replacement*, not re-derivation: the pinned zone-bound route (annulus/strip/far constants) is retired; the two-sided constant is one Palm integral, E[N_qual] = ∫Λ·AO dA, with the β channel measured dead and the ledger within budget. The pinned constants become scaffolding-only (table note). | Lemma_UB0.md; C020_Observed_Update | +| OBL-C019-1 | second independent λ method at arch scale; E1 band [0.9, 1.3] adjudication | **DISCHARGED (C020)** — G-U1 two-covariance-model concordance (ratios 0.989–1.003, 6 stations); the unified constant 0.970 lies inside the committed band; the slice value 1.097 is superseded in scope (estimand change, G-U2-driven). | C020 gates | +| OBL-C019-2 | MID other-branch mechanism; marginal-typed polish artifact | OPEN (instrument hygiene) | carried | +| OBL-LBB-1 | two-point measurement at d ∈ {0.02, 0.05, 0.1} on the arch | OPEN | carried (not executed in C020; the O(r⁶) pair bound rests on the C018 d ≥ 0.2 values + the LB mechanism) | +| OBL-LB-ARCH | rigorous positivity of ∫Λ·AO dA | OPEN — the sole remaining lower-bound obligation; basis strengthened by C020 (Λ-exact estimator, two-model concordance, rim map) | Corridor_Mechanism_KILL §3; Lemma_UB0 §6 | +| **OBL-BETA-FAR** | rigorous ceiling for the far-zone window-edge-terminal β channel (∫_far Λ_max·B_far) | **NEW (C020)** | route: adjacency-type decay + the measured support edge (Gmin = 0.0062 = 300ℓ) | +| **OBL-LB-RESTATE** | restate Lemma LB's p_∞ mechanism as moderate-r; asymptotic lower bound carried by α alone | **NEW (C020)** | fork branch (iii) consequence | +| **OBL-RIM-REFINE** | rim quadrature densification + Pre-map refinement (~2% interpolation systematic) | **NEW (C020)** | optional escape-radius study attached | +| **OBL-CLIP-ARCH** | quantify the ≤ 2% arch systematic from the retired clip margins (re-run two arch rows margin-free) | **NEW (C020)** | error-budget hygiene | +| OBL-FAR-DECAY | quantitative off-pair decorrelation (γ-LOC formalization) | OPEN | carried; now load-bearing for Lemma_UB0's architecture step | +| κ_β (gain-channel multiplier) | — | **MOOT (C020)** under fork branch (iii): c_β ≤ 2.6×10⁻⁸; no multiplier enters the constant | C020_Observed_Update | +| File-5 B2-positivity ghost debt | — | DISCHARGED (C015) | LemmaP_Discharge_2.md | +| OBL-C017-1 / OBL-C017-2 / OBL-LB-ADJ | — | DISCHARGED / DISCHARGED / SUPERSEDED (C018–C019, unchanged) | prior registry | + +## 2. FAILED-AS-WRITTEN ledger additions (C020; recorded without rescue) + +- **G-C1 (continuity, 3.52σ at ε = 0.03):** the committed monotone upper-tail convention + 1σ z-binning + bias the strata-assembled bulk CDF against the direct run. Both records stand; the small-ε ladder + (load-bearing) has zero tail contribution and is unaffected. +- **G-U2 (window flatness):** the flatness premise itself was false (needle σ_t = 0.068ℓ at the arch, + 10⁻⁴ℓ near M). Drove the in-cycle estimator supersession (slice → Λ-exact). The gate as committed was + mis-designed; recorded as such. +- **P-U2 (correction ledger, inner budget):** blown by the 372-coefficient alarm; resolved into (a) the + genuine rim channel (+0.143, unified into α) and (b) the null-rigidity incident (below). + +## 3. Instrument incident record (C020) + +**Null-rigidity off-manifold conditioning (near-M zone).** The 9-pin Gram at δ ≈ 0.003–0.005 from M has +λ_min = 2.2×10⁻¹⁷ with null functional (f(M) − f(y))/√2; the retired 0.02ℓ interior clip margin displaced +the factor-2 height 57 null-sd off the compatible manifold, inflating E[W₃1] ×4×10⁴ and manufacturing +22σ far-field artifacts (measured; linear response 0.4σ-of-mountain per null-sd). Exposed by physical +forensics (f = 21–23 at escape positions). Repair: margin-free v* = clip(μ_t, window). All near-M records +superseded by compatible-ensemble reruns (*_v2 files); distorted records retained. Arch rows unaffected +(≤ 2% residual, budgeted, → OBL-CLIP-ARCH). Class: v2.1-lesson recurrence (near-degenerate conditioning ++ off-manifold target = silent ensemble distortion). + +## 4. Reconciliation notes carried unchanged + +C014 part-2 dual-account resolution (operative: KILL account (B)); C014p2 snapshot-absence flag; +C015/C016/C017 date-field corrections — all as in the C019 registry, unchanged. + + + + + +--- + + + +# Registry_Reconciliation — obligations ledger and record hygiene (as of C021, 2026-07-09) + +Supersedes Registry_Reconciliation_C020.md; companion to Constants_Table_Canonical_C021.md, +Lemma_UB0.md + Addendum A, C021_Observed_Update.json. Supersede-never-overwrite throughout. + +## 1. Obligations ledger (complete, current) + +| ID | content | status | discharge / carrier | +|---|---|---|---| +| OBL-LB-ARCH | rigorous positivity of ∫Λ·AO dA | **OPEN — the single load-bearing item to proof-grade Theorem A.** Basis now maximal: Λ-exact estimator, two-model concordance at two rungs, whole-support qualification, clean gates. Reduction on file: LB0 + R0 + finite-dimensional Gaussian-positivity certificates + one open-tube-event step. | Lemma_UB0 §6; Corridor_Mechanism_KILL §3 | +| OBL-FAR-DECAY | quantitative off-pair decorrelation (γ-LOC) | OPEN (proof grade) | Lemma_UB0 §4 | +| OBL-BETA-FAR | rigorous far-zone β ceiling | OPEN (proof grade) | C020 registry | +| OBL-R0-UNIF | 18×18 nondegeneracy uniformity + constant chase | OPEN (bookkeeping) | MS_Shift_R0_Closure §6 | +| OBL-CLIP-ARCH | quantify the retired-clip arch systematic | **DISCHARGED (C021)**: margin-free re-runs, ratios 1.0009/1.0000 → ≤ 0.1%; r = 0.05 bar tightens to ±0.055 | c021_cliparch.json | +| OBL-RIM-REFINE | rim quadrature/Pre densification | **DOWNGRADED to optional (C021)**: rim share 0.026 and vanishing ~r^2.5 — refinement no longer moves C\*; structural interest only | P-V2 branch (ii) | +| OBL-LB-RESTATE | restate Lemma LB's p_∞ as moderate-r | OPEN (writing task) | carried | +| OBL-LBB-1 | two-point at d ∈ {0.02, 0.05, 0.1} | OPEN | carried | +| OBL-C019-2 | MID mechanism / polish artifact hygiene | OPEN | carried | +| **OBL-NEEDLE-TREND** | classify the peak needle-ratio drift 0.068 → 0.086 (constant vs r-trend) | **NEW (C021, non-load-bearing)**: one third-rung datum suffices | G-V3 note | +| **OBL-LOOP-LOCATE** | exhibit or bound-away the loop population (unmeasured at both rungs) | **NEW (C021, bookkeeping)**: the C011 absorption currently bounds an unpopulated event | Lemma_UB0 Addendum A §A.3 | + +Previously discharged (unchanged): OBL-UB-PALM, OBL-C019-1, OBL-C017-1/2, OBL-LB-ADJ (superseded), +File-5 ghost debt; κ_β moot. + +## 2. Fork outcomes of record (C021) + +P-V1 → **branch (i)**: |Δ| = 2.5% ≤ 10% — constant stable across the octave; the frozen +measurement-completeness interpretation is adopted verbatim. P-V2 → branch (ii): rim ~r^2.5. +P-V3 → branch (i): β dead, edge deepens (894ℓ). + +## 3. Incident/caveat closures (C021) + +- **C020 isM-ball caveat → CLOSED by re-measurement** (c020_ballfix.py, 4 points, same seeds): + the r = 0.05 "loop core" was ball-asserted; measured Pre = 0.81–1.00. C\*(0.05) supersede chain: + 0.9702 (as-adjudicated) → 0.9728 (ballfix-corrected); both stand. Lemma UB0 §3 dichotomy retired + as measured geometry (Addendum A); qual/raw ≥ 0.9997 at both rungs. +- **Clean gate ledger**: C021 is the program's first cycle with every committed gate passing + (6/6). Recorded as a protocol-health datum, not a relaxation of the FAILED-AS-WRITTEN norm. + +## 4. Measurement-completeness statement (of record) + +Per P-V1(i): no open item on the Theorem A path is a measurement. Submission path as experimental +mathematics with the pre-registered verification record: **CLEAR** (assembly task: manuscript +unification). Path to proof grade: OBL-LB-ARCH (load-bearing) + OBL-FAR-DECAY + OBL-BETA-FAR + +OBL-R0-UNIF. Theorem B's program is unaffected by this cycle (its own ledger stands as of C015). + +## 5. Reconciliation notes carried unchanged +C014 dual-account resolution; C014p2 snapshot-absence flag; date-field corrections — as in the +C019/C020 registries. + + + + + +--- + + + +# Registry_Reconciliation — obligations ledger and record hygiene (as of C022, 2026-07-09) + +Supersedes Registry_Reconciliation_C021.md; companion to Constants_Table_Canonical_C022.md, +Lemma_LB_ARCH_Package.md, Lemma_FD_Package.md, R0UNIF_and_LBRestate.md, +C022_Observed_Update.json. Supersede-never-overwrite throughout. + +## 1. Obligations ledger (complete, current) + +| ID | content | status | discharge / carrier | +|---|---|---|---| +| OBL-LB-ARCH | rigorous positivity of ∫Λ·AO dA | **CLOSED (C022)** at derived-and-verified; two cited-standard steps flagged inline (Bogachev 3.6.1 small-ball at CM points; Morse–Smale basin continuity). Five certificates (margins 10^38–10^61, both rungs) + explicit RKHS barrier with frozen-pin rigidity (ε₀ ≈ 0.008 macroscopic) + CM support step. | Lemma_LB_ARCH_Package.md | +| OBL-FAR-DECAY | quantitative off-pair decorrelation (γ-LOC) | **far-field core DISCHARGED (C022)** derived-and-verified (19-row exact tables to d = 6, both rungs, envelope beyond); tube-local part remains architecture, scope unchanged | Lemma_FD_Package.md §2 | +| OBL-BETA-FAR | rigorous far-zone β ceiling | **fixed-radius core DISCHARGED (C022)**: P(β-far) ≤ 12.41·r³ at d₀ = 0.3, rung-independent, all r ≤ 0.05; per-rung three-piece cover complete (no gap at r = 0.05). Residual re-scoped → OBL-BETA-MODZONE | Lemma_FD_Package.md §3 | +| **OBL-BETA-MODZONE** | r-uniform moderate-zone (2r, d₀) sharpening: replace Cauchy–Schwarz Hessian ceiling by the near-fold \|det H\| suppression | **NEW (C022)** — the successor residual, precisely scoped; per-rung numbers already rigorous (84.2 / 334.9 ·r³) | Lemma_FD §3 scoping | +| OBL-R0-UNIF | 18×18 nondegeneracy uniformity | **DISCHARGED (C022)**: continuity + compactness + universal interpolation off the analytic null set; mp anchors 9.73e−21 / 3.73e−23 | R0UNIF_and_LBRestate.md Part A | +| OBL-LB-RESTATE | scope Lemma LB to moderate-r | **DISCHARGED (C022)**: supersede-in-scope note; asymptotic chain of custody = LB0 → R0 → LB-ARCH → FD | Part B | +| **OBL-LB-RATE** | rigorous sharp lower rate via C012/C013 limit objects | **NEW (C022, optional)** — not required for Θ(r³) | LB-ARCH §6 | +| OBL-NEEDLE-TREND | needle-ratio drift 0.068 → 0.086 | carried (non-load-bearing; one third-rung datum) | C021 | +| OBL-LOOP-LOCATE | exhibit/bound-away loop population | carried (bookkeeping) | UB0 Addendum A | +| OBL-LBB-1 | two-point at d ∈ {0.02, 0.05, 0.1} | carried | — | +| OBL-C019-2 | MID mechanism / polish hygiene | carried | — | + +Previously discharged (unchanged): OBL-UB-PALM, OBL-C019-1, OBL-CLIP-ARCH, OBL-C017-1/2, +OBL-LB-ADJ (superseded), File-5 ghost debt; OBL-RIM-REFINE optional; κ_β moot. + +## 2. C022 record notes + +- **Construction dead end (not a registry kill):** the 15/18-pin barrier variant rejected-as-built + (max|w| = 1.27×10¹⁴, evaluability destroyed; natural y*-Hessian max-typed). The 9-pin + + shaping-bump route is the construction of record. +- **Instrument incident (closed same-session):** first r = 0.025 barrier replication ran with the + unpatched module-global ℓ (S-pin f-err −1.82×10⁻⁵ = exactly the ℓ discrepancy). Root-caused, + fixed, re-run clean; both runs retained in the record. +- **Honest-scoping note:** the FD moderate-zone bound is per-rung rigorous but not r-uniform + as-written (crude ceiling ~d⁻⁴ vs true Λ-scale mass); the gap is now a *named, scoped* + obligation rather than an implicit one. + +## 3. Program state (of record) + +**Theorem A:** measurement-complete (C021) **+** lower-side positivity closed at derived grade +**+** rigorous O(r³) upper envelope (C022). The Θ(r³) statement holds with exponent at +proof-adjacent grade and constant measured. Remaining to full proof grade: OBL-BETA-MODZONE, +basin-continuity formalization (if referee-demanded), γ-LOC tube-local, manuscript unification. +**Theorem B:** unaffected; its ledger stands as of C015. +**Submission paths:** experimental-mathematics package — CLEAR (assembly only); methodology +paper — CLEAR; exact-lemma cluster — CLEAR; proof-grade Theorem A — the four items above. + +## 4. Reconciliation notes carried unchanged +C014 dual-account resolution; C014p2 snapshot-absence flag; date-field corrections; C020 ballfix +supersede chain (0.9702 → 0.9728) — as in the C020/C021 registries. + + + + + +--- + + + +# Registry_Reconciliation — obligations ledger and record hygiene (as of C023, 2026-07-09) + +Supersedes Registry_Reconciliation_C022.md; companion to Constants_Table_Canonical_C023.md, +Lemma_MODZONE_CountLaw_Package.md, C023_Observed_Update.json. Supersede-never-overwrite. + +## 1. Kill registry — new entries (FAILED-AS-WRITTEN, no rescue) + +| ID | route | verdict | +|---|---|---| +| **P-W1 (C023)** | r-uniform moderate zone via the sup-decoupled certified ceiling c_sharp | branch (iii): γ ≈ 1.95 over four rungs (I = 272.5/902.6/4000/15758). Post-mortem recorded: sup-decoupling + gradient-density mean-term omission (C022/C023 ceilings remain valid, loose). | +| **P-W2 (C023b)** | r-uniform moderate zone via the **exact** Kac–Rice band-critical count | branch (iii): γ = 1.882, rms 0.022 (I_exact = 12.43/43.83/161.7/621.9). The divergence is exact truth: **no count-based route can deliver O(r³) in the moderate zone.** | + +## 2. Obligations ledger (movements) + +| ID | status | notes | +|---|---|---| +| OBL-BETA-MODZONE | **SUPERSEDED (C023)** | both registered routes killed; the estimand was wrong (count vs relevance) | +| **OBL-BETA-RELEVANCE** | **NEW (C023)** — the single mathematical frontier of Theorem A's r-uniform upper side | prove P(f(T) ∈ band ∧ ρ_T ∈ (2r, d₀)) ≤ C·r³ via terminal selectivity. Deterministic anchor: Σ_maxima 1{terminal} = 1 (unique S⁺ branch). Candidate mechanisms recorded: slab-traversal geometry (ribbon width ~ℓ/(κ_g d) over length d − r/2), crest decorrelation, reverse-barrier AO upper bound. Must reproduce the measured support edge 894ℓ. Includes the B1-mechanism audit item. | +| OBL-LB-ARCH, FAR-DECAY core, BETA-FAR fixed-d₀ core, R0-UNIF, LB-RESTATE | closed (C022), unchanged | fixed-d₀ far constant C(0.3) = 12.41 unaffected by the C023 corrections | +| OBL-LB-RATE (opt), OBL-NEEDLE-TREND, OBL-LOOP-LOCATE, OBL-LBB-1, OBL-C019-2, γ-LOC tube-local | carried unchanged | | + +## 3. Archive corrections (scope-precise; supersede-in-scope) + +1. **File-1 §5.2 (raw-law reading) + File-3 Proposition 5.2** — the claimed O(ℓ) annulus band + count via Pinning-scaling suppression exp(−c/τ²): **OPEN-CONTRA-INDICATED** for the raw-count + reading. Empirical basis (exact mp, four rungs): raw transverse gradient sd ~ κ_g·d (not τ³); + crest Mahalanobis q_min ≈ 0.72 constant; E[N_ann | pair] = Θ(r^{−1.9})·ℓ exactly. Scope + caveat: the chained / midpoint-2-jet-removed variant conditions differently and is not per se + disproven — but no chaining rescues the raw expectation. The July-5 assembly's annulus term + must be re-routed through OBL-BETA-RELEVANCE. +2. **Lemma B1 circle-band mechanism** — flagged: perpendicular circle stations are polynomially, + not exponentially, suppressed under the raw law. Conclusion possibly rescuable via the + multi-station product; audit carried inside OBL-BETA-RELEVANCE. +3. Positive replacement: the C020–C023 certified structure (near rigidity 8×10⁻⁴², corridor + 2.6×10⁻⁸, fixed-d₀ far 12.41·r³, exact per-rung moderate masses 2.07/7.31·r³) **supersedes** + the contra-indicated annulus route at the anchor rungs with stronger, exact results. + +## 4. C023 record notes + +- Fork discipline: the sup-form freeze (P-W1) was adjudicated on its literal frozen statistic + before the successor was frozen (C023b) — no silent upgrade; both snapshots hashed + pre-compute; all adjudication data deterministic and archived. +- The gradient-density mean-term omission in C022's Lemma-FD table is a **looseness note, not a + correctness incident** (the omitted factor is ≤ 1; all inequalities preserved). +- New certified constants promoted to the canonical table: ĉ(2) = 6.60×10⁻³ (0.5% four-rung), + c∞ ≈ 0.14·d^{−2.9}, q_min = 0.720–0.727, I_exact table, per-rung all-in envelopes 15.5/20.7. + +## 5. Program state (of record) + +**Theorem A:** exponent two-sided at proof-adjacent grade (C022); **rigorous all-in per-rung +envelopes at both anchors (15.5 / 20.7 in r³ units)** with the sharp constant measured +(C\* = 0.96 ± 0.05). Remaining to r-uniform proof grade: **OBL-BETA-RELEVANCE** (the frontier), +basin-continuity formalization (if demanded), γ-LOC tube-local, manuscript unification. +**Theorem B:** unaffected; ledger stands as of C015. +**Submission paths:** experimental-mathematics package, methodology paper, exact-lemma cluster — +all CLEAR (assembly only); the C023 structural discovery (the Mahalanobis ridge + exact count +law + archive correction) is itself a publishable experimental-mathematics unit. + + + + + +--- + + + +# Registry_Reconciliation — obligations ledger and record hygiene (as of C019, 2026-07-09) + +Companion to Constants_Table_Canonical_07-09-26.md and the four lemma packages delivered this cycle. +Supersede-never-overwrite: nothing below deletes an archive entry; every change is a scoped supersession. + +## 1. Obligations ledger (complete, current) + +| ID | content | status | discharge / carrier | +|---|---|---|---| +| MS-Sard / R0 | a.s. Morse–Smale for the conditioned flow | **derived-architecture** | MS_Shift_R0_Closure.md (C015 18×18 input); residue → OBL-R0-UNIF | +| OBL-R0-UNIF | uniformity of the 18×18 nondegeneracy off the analytic null set + constant chase in the C¹ manifold citation | OPEN (bookkeeping grade) | §6 of MS_Shift_R0_Closure.md | +| OBL-C017-1 | Palm-vs-pinned correction of the halo trigger c₁ | **DISCHARGED** (C018 P-L1: c₁^{Palm} = 0.0643 ± 0.0008) | C018_Observed_Update | +| OBL-C017-2 | Bonferroni second-moment control | **DISCHARGED** | LB_Bonferroni.md (E[N(N−1)] = O(r⁶)); residue → OBL-LBB-1 | +| OBL-LBB-1 | two-point measurement at d ∈ {0.02, 0.05, 0.1} on the arch | OPEN (C020 freeze item) | LB_Bonferroni §5 | +| OBL-LB-ADJ | rigorous c₂ > 0 on the O(1) halo | **SUPERSEDED** (halo deployment killed; c₂(halo) measured 0) | → OBL-LB-ARCH | +| OBL-LB-ARCH | rigorous positivity of ∫_arch λ·A·O dA (not sharpness) | OPEN — the sole remaining lower-bound obligation; reduced to LB0 + R0 + one-point nondegeneracy on the filament | Corridor_Mechanism_KILL §3 | +| OBL-UB-PALM | re-derive upper-bound channel constants (annulus/strip/far) under Palm weighting; cover τ ∈ (r/2, 2r) | OPEN — **the program's next load-bearing item**; exponent unaffected | C019_Observed_Update; Constants table ⚠ rows | +| OBL-C019-1 | C020 freeze: arch-adapted grid; second independent λ method at arch scale; falsifiable E1 band [0.9, 1.3] | OPEN | C019_Observed_Update | +| OBL-C019-2 | MID other-branch mechanism; marginal-typed terminal-at-y polish artifact | OPEN (instrument hygiene) | C019_Observed_Update | +| File-5 B2-positivity ghost debt | — | **DISCHARGED by composition** (C015 cycle, LemmaP-general + B2-comp addendum) | LemmaP_Discharge_2.md | + +## 2. Reconciliation of the two C014-part-2 accounts + +The archive contains two part-2 narratives: (A) C014p2_Observed_Update.json (third-point channel as a +live failure mode, joint constants at r = 1/5) and (B) C014_KILL_ThirdPoint_Channel.md (the channel as +a moderate-r artifact: M itself sat inside the search ball at r = 0.2). **Operative account: (B)**, per +the D-A1′-gated M-excluded small-r re-run (max-third-point fraction ≤ 3×10⁻⁴ at r = 0.05). Account (A) +is retained verbatim as the FAILED-AS-WRITTEN record of the artifact, scope-tagged "moderate-r, ball +uncorrected"; its joint constants are demoted to phenomenology (see P-J6 row in the canonical table). +Chain: C013 (0.92 tail adjacency) → C014p1 (0.23–0.28, superseded) → C014-KILL (C013 reinstated) — +explicit and closed. + +## 3. Record-hygiene items (logged, non-blocking) + +- **C014p2 snapshot absence**: the archive holds the chain hash 724b2636… but the frozen snapshot JSON + itself is ABSENT from the project store. The update+KILL pair fixes the content; the absence is + logged so no future cycle claims a freeze it cannot exhibit. Action: none possible retroactively; + flagged. +- **Date-field errors**: C015/C016/C017 update JSONs carry date 2026-07-06; the C017 verification + session established the correct dates (C015/C016: 2026-07-06 execution confirmed; C017: 2026-07-09). + Corrections live here; originals untouched. +- **C016 "≈14" flag**: one prose line in the C016 narrative quotes the enhancement peak as "≈14"; the + frozen data gives R(1.5) ≈ 5. The data value governs; prose slip logged. +- **LB0 citation debt**: the LB0 proof text cites R0 and R2 informally; formal pointers are now + MS_Shift_R0_Closure.md (R0) and the C015 reduction note (R2/Sard step). Bonferroni citation → + LB_Bonferroni.md. LB0's Proved status is unchanged. +- **numpy 2.x**: np.trapz removed; all instruments now use np.trapezoid (silent-failure risk for any + future re-run of old scripts; noted in instrument headers going forward). + +## 4. The C019 instrument-chain lesson (for the methods paper) + +The frozen G1′ pin gate functioned as the sole arbiter through two successive, differently-caused +instrument failures — (i) float64 amplification of kriging weights at cond > 1/ε (v1, 16/40 points, +errors up to 5×10²¹), then (ii) a 10⁻⁶-scale analytic-vs-discrete model mismatch exposed only after mp +whitening (v2.1) — before any science was read. The v2.2 resolution (exact coefficient-space projection: +condition the *simulator's own* Gaussian coordinates, then run stock dynamics) is the general pattern: +**when the estimand is a conditional law of a simulated Gaussian object, condition in the object's exact +coordinates rather than kriging an idealized continuum model onto it.** The same lesson produced lam3: +the frozen λ estimator was unbiased with empty effective support at shell scale (importance collapse, +CI [0, 0] false confidence); the exact disintegration computes the identical estimand with the +small-variance factor in closed form. Both replacement steps were identity-verified against the frozen +forms where those operate (v1/v2/v2.2 flow concordance; lam3 vs frozen-λ vs C018 at the G2′ point, +1.2σ/2.0σ). + +## 5. Standing deliverable map (this delivery) + +C018 set: snapshot + observed update + c018_flow.py + trigger files + 9 flow JSONs + g4_honest + pl5 + +pl6. C019 set: snapshot (sha256 fccf253f…) + observed update + instrument chain (4 flow versions + +lam3 + drivers) + 39-point committed grid + 33-point committed λ + arch data (ridge scan, quadrature, +micro-scans, 8 arch flows, halvings) + results assembly. Lemma packages: MS_Shift_R0_Closure, +LB_Bonferroni, Lemma_P_old, Corridor_Mechanism_KILL. Tables: Constants_Table_Canonical_07-09-26. +This file. All copied to outputs. + + + + + +--- + + + +# C091 CORRECTION LEDGER + +**Date:** 2026-07-16 +**Discipline:** append-only. A correction changes the live dependency graph; it does not erase the historical artifact. + +## E-C091-1 — Upper interpolation gate over-strengthened at C090 + +**Prior form:** require a bound on \(\sup|U''|\). +**Finding:** the linear-interpolation remainder has \((r-a)(r-b)\le0\), so only negative curvature can lift \(U\) above its endpoint chord. +**Live successor:** `UPPER_ONE_SIDED_INTERPOLATION`, requiring \(U''\ge-M\). +**Effect on theorem:** reduces numerical burden; no theorem is weakened. + +## E-C091-2 — H4 one-step Markov product killed + +**Prior form:** use \(q_{step}^{n}\) for an \(n\)-checkpoint corridor. +**Finding:** exact six-pin multivariate Gaussian diagnostics give corridor probabilities larger than \(q_{step}^{n}\) by factors up to about 3.35 in the tested skeletons. +**Classification:** structural probability error, not a rounding issue. +**Live successor:** `H4_PALM_ORTHANT`, using finite-dimensional Gaussian Chernoff bounds, shifted Palm determinant moments, and explicit path entropy. +**Effect on theorem:** the C089 far coefficient \(0.0199\) returns to OPEN until re-certified. + +## E-C091-3 — Bonferroni near-diagonal mark mismatch + +**Prior form:** cite spatial critical-point repulsion to assert a two-point intensity per unit height squared bounded by \(C_{nd}d\), then factor \(\ell^2\). +**Finding:** shrinking-window second-moment theory shows that nearby critical values can cost only one explicit height-window factor; same-type saddle repulsion is unmarked and does not by itself restore \(\ell^2\). +**Classification:** incomplete source-to-estimand transfer. The old claim is quarantined, not declared impossible. +**Live successor:** `BONFERRONI_MARKED_TWO_SCALE`, combining + +\[ +K_{ss}(d)\le C_{rep}d^3\log(e/d) +\] + +with a bounded density for + +\[ +A=(u_1+u_2)/2, +\qquad +Z=(u_2-u_1)/d^3. +\] + +**Effect on theorem:** lower decimal tiers are conditional on the new uniform constants. The one-point \(\Lambda\) machinery remains live. + +## E-C091-4 — Exact torus transfer discharged in declared matrix scopes + +**Prior status:** exact periodized ensemble defined, downstream matrix transfer open. +**Finding:** explicit Neumann/Schur perturbation bounds preserve the local and far covariance floors with large margins. +**Live status:** local and far-station transfer CLOSED for derivative order \(\le4\), pin dimension \(\le9\), target dimension \(\le6\), local radius \(3\), and far distance interval \([5,12]\). +**Residual:** global integrals must partition their domains and declare any larger matrix/order use. + +## E-C091-5 — Upper uncertainty polarity fixed + +**Prior form:** a symmetric \(\pm0.02\) assembly band was attached to an upper theorem. +**Finding:** upper claims consume only the positive side. +**Live rule:** use \(+0.02\) unless the inputs are proved to be already one-sided upper limits. +**Residual:** confidence level, multiplicity, sample size, and construction of the band remain unlocated. + +## Status delta + +```text +CLOSED: + exact torus local transfer + exact torus far-station transfer + one-sided interpolation theorem + upper uncertainty semantics + Gaussian corridor Chernoff theorem + Palm-weighted Chernoff theorem + planar d^3 critical-value-gap law + +KILLED: + H4 q_step^n product rule + +QUARANTINED: + pointwise C_nd d claim as a uniform per-height^2 Bonferroni discharge + +OPEN: + six-pin marked saddle-repulsion constants + H4 numerical path assembly + upper shape certificate + upper uncertainty calibration + +BLOCKED FROM PROMOTION: + finite lower 0.84 + asymptotic lower 0.8501 + continuous upper 0.99/1.01 +``` + + + + + +--- + + + +# C093 LOOSE-ENDS DISPOSITION REGISTER + +**Rule:** every item has one terminal disposition and one explicit effect on +the C092 core. An unnamed `OPEN` status is prohibited in the active release. + +| ID | Item | Final disposition | Successor/owner | Effect on C092 core | +|---|---|---|---|---| +| LE-01 | Exact torus covariance | CORE-CLOSED | C092 | Canonical model | +| LE-02 | Local/far periodization transfer | CORE-CLOSED in declared scope | C092 | None | +| LE-03 | Fixed-L rate theorem | CORE-CLOSED at program grade | C092 | Canonical theorem | +| LE-04 | Qualitative q0=1 theorem | CORE-CLOSED | C092 | Canonical corollary | +| LE-05 | R0/SARD-G internal dependency | CORE-CLOSED | C092 | None | +| LE-06 | Independent SARD-G review | EXTERNAL-REVIEW-TRACK | Q0-REFEREE | No core dependency | +| LE-07 | Source-level marked ND audit | EXTERNAL-REVIEW-TRACK | Q0-REFEREE | No core dependency | +| LE-08 | Interval replacement of GRID moduli | DISPOSITIONED-SUCCESSOR | Q0-REFEREE | No core dependency | +| LE-09 | Full-radius analytic nine-pin atlas | DISPOSITIONED-SUCCESSOR | Q0-REFEREE | No core dependency | +| LE-10 | C089 finite lower 0.8501 | SUPERSEDED as finite theorem | Q0-SHARP | No core dependency | +| LE-11 | C089 upper 0.97 through r=.05 | KILLED by endpoint | None | No core dependency | +| LE-12 | H4 q_step^n product | KILLED by correlated-Gaussian counterexample | None | No core dependency | +| LE-13 | First-interceptor sharpened upper | DISPOSITIONED-SUCCESSOR | Q0-SHARP | No core dependency | +| LE-14 | Infinite-volume and critical-height scaling | DISPOSITIONED-SUCCESSOR | Q0-IV | No core dependency | +| LE-15 | Persistence-density Theorem B | DISPOSITIONED-SUCCESSOR | Q0-B | No core dependency | +| LE-16 | LLM empirical deployment | DISPOSITIONED-SUCCESSOR | Q0-LLM | No core dependency | +| LE-17 | Production LLM guarantee | NOT-CLAIMED | Q0-LLM | No core dependency | +| LE-18 | Scale-free alias TRUTH_CONST | KILLED/RETIRED | None | Replaced by rung tag | +| LE-19 | Absolute output paths | RELEASE-CLOSED | C093 | Portable scripts | +| LE-20 | Exact software environment | RELEASE-CLOSED | C093 | Pinned environment | +| LE-21 | Standalone builder/verifier | RELEASE-CLOSED | C093 | Included | +| LE-22 | Declared-file versus ZIP-entry count | RELEASE-CLOSED | C093 | Explicitly recorded | +| LE-23 | Exact ZIP attestation | RELEASE-CLOSED | C093 | Detached hash record | +| LE-24 | Fresh-extraction execution | RELEASE-CLOSED | C093 | All checks pass | +| LE-25 | Unowned items | CLOSED: ZERO | C093 | Release gate | + +`DISPOSITIONED-SUCCESSOR` means scientifically unfinished but not unfinished +within C092/C093. `NOT-CLAIMED` is a completed boundary decision, not an +omission. + +A successor can reopen C092 only through a versioned correction node containing +the failed core step, mathematical evidence, source and artifact hashes, the +downstream invalidation set, and a replacement contract. + +**Unowned dispositions: 0.** + + + + + +--- + + + +# C094 E-LEDGER — PASS 1 + +**Policy:** append-only. A failure, near-loss, or blocked import is preserved +with evidence, cause, downstream effect, and terminal disposition. + +--- + +## E-C094-1 — Prescribed C093 verification order mutates a manifested file + +**Evidence:** `C094_FRESH_STATE_VERIFICATION.json`. + +**Observed sequence:** + +```text +verify before audit: PASS +audit: PASS, but rewrites Q0_C093_AUDIT_REPORT.json +verify after audit: FAIL, hash and size mismatch +``` + +**Cause:** `audit_q0_c093_release.py` writes its current shallow audit result +over the deep-execution audit report included in the immutable manifest. + +**Downstream impact:** release-pipeline reproducibility only; no mathematical +claim changes. + +**Disposition:** `ERRATUM`. In every C094+ release, immutable-manifest +verification precedes writers, or the audit writes to a new versioned output +path. + +--- + +## E-C094-2 — UB-G upper coefficient rounded in the wrong direction + +**Evidence:** `C094_UBG_ADJUDICATION.json`. + +**Finding:** + +\[ +(0.66+2.82)/0.80=4.35, +\] + +while the frozen display was \(4.3\). The unrounded near/exterior assembly is +approximately \(4.3444138731\). + +**Cause:** downward decimal rounding of an upper bound, with no displayed +offset certificate. + +**Downstream impact:** `UB_G`, the two-sided rate display, endpoint arithmetic, +claim-language files, deprecation aliases, and both theorem hashes. + +**Disposition:** `CORRECTION`. Replacement upper coefficient: \(4.35\). + +--- + +## E-C094-3 — The \(0.8411\) lower coefficient lost its dimensions + +**Evidence:** `C094_LOWER_COEFFICIENT_ADJUDICATION.json`. + +**Finding:** the source writes + +\[ +1-q\ge C^*(r)r^3AO(r)(1-O(r^3)), +\qquad C^*(r)\ge0.8411. +\] + +It does not license + +\[ +1-q\ge0.8411r^3 +\] + +without propagating the subunit AO and Bonferroni factors. + +**Cause:** a first-moment coefficient was promoted to a final probability +coefficient. + +**Downstream impact:** `LOWER_RATE_PG`, `RATE_PROGRAM_GRADE`, every “two-sided +closed theorem” wording. + +**Disposition:** `CORRECTION`. Preserve \(C^*(r)\ge0.8411\); move finite lower +probability claims to Q0-SHARP. + +--- + +## E-C094-4 — C091 same-name files lacked per-file provenance rows + +**Evidence:** `C094_C091_C093_PROVENANCE.json`. + +**Finding:** ten same-name artifacts had changed or release-specific bytes: +four portability edits, one control-character repair, one diagnostic +re-execution, and four identities; seven complete C091 artifacts were +archive-only. + +**Cause:** C093 documented a portability pass globally but did not provide a +per-file same-name delta ledger. + +**Downstream impact:** provenance clarity only. + +**Disposition:** `ERRATUM-CLOSED` by the C094 table. + +--- + +## E-C094-5 — Randomized orthant diagnostics changed at identical byte count + +**Evidence:** same-size `gaussian_corridor_certificate_report.json` comparison. + +**Finding:** twelve scalar leaves changed, exclusively +`orthant_probability_diagnostic` and `orthant_over_naive_product`. Maximum +relative change was approximately \(4.18\times10^{-7}\). The adjudication, +Chernoff bounds, covariance data, counterexample labels, and kill decision +were byte-equivalent as JSON values. + +**Cause:** numerical multivariate-normal CDF diagnostics were re-executed +without a frozen integration state. + +**Downstream impact:** none on theorem inputs; reproducibility metadata was +incomplete. + +**Disposition:** `DIAGNOSTIC-REEXECUTION`. Future diagnostics record all +algorithmic seeds/state or are explicitly treated as non-hashed displays. + +--- + +## E-C094-6 — “Linear repulsion is known” conflated three different claims + +**Evidence:** `C094_CITATION_PROVENANCE.json`. + +**Finding:** + +- Ladgham–Lachièze-Rey describe typed extrema/saddle hard repulsion with three + additional powers at the factorial-moment level; +- Azaïs–Delmas give aggregate neutrality in dimension two; +- neither fact alone supplies the pair-Palm two-height marked-window law. + +**Cause:** spatial/type correlation, marked height density, and conditioned +six-pin transfer were collapsed into one citation label. + +**Downstream impact:** Q0-REFEREE, BR-REP, BR-MARK, Q0-B. + +**Disposition:** `CORRECTION-REGISTERED`; no citation-only mark transfer is +permitted. + +--- + +## E-C094-7 — Nicolaescu identifier correction was stated too broadly + +**Evidence:** `C094_CITATION_PROVENANCE.json`. + +**Finding:** both IDs are real and refer to different papers: + +```text +1101.5990 Critical sets of random smooth functions on compact manifolds +1209.0639 Random Morse functions and spectral geometry +``` + +**Cause:** an operator-line correction treated the IDs as globally +interchangeable. + +**Downstream impact:** bibliography only. + +**Disposition:** `ADJUDICATED-CONFLICT`. Match identifier to title and claimed +content. + +--- + +## E-C094-8 — Rung minimum was consumed as a domain infimum + +**Evidence:** `C094_IMPORTED_PRIORS_RECHECK.json`. + +**Finding:** the C091 allowed-product values at registered rungs exceed 80, +but the full-domain limit is + +\[ +79.9889203915211757\ldots +\] + +and the function crosses 80 at + +\[ +r=0.0018360362779673286\ldots. +\] + +**Cause:** sampled-rung validation of a uniformly quantified interval claim. + +**Downstream impact:** Q0-SHARP finite lower gate and verifier v5. + +**Disposition:** `CORRECTION`. `DOMAIN-INFIMUM` is a first-class active gate. + +--- + +## E-C094-9 — Exact six-pin factorization was nearly lost as an unowned import + +**Evidence:** `C094_SIX_PIN_FACTORIZATION.json`. + +**Finding:** for the exact periodized field, + +\[ +m_L(x,y)=k_L(y)m_L(x,0) +\] + +follows exactly from product-kernel structure and parity-block decoupling. + +**Cause:** the operator-line result was absent from C093 successor inputs. + +**Downstream impact:** BR-MARK and Theorem B design. + +**Disposition:** `IMPORTED-AFTER-REVERIFICATION`, grade `DERIVED-EXACT`. +It reshapes but does not close BR-MARK. + +--- + +## E-C094-10 — Two operator-line measurements lack immutable source artifacts + +**Items:** + +- height-universality / flatness of \(q\) in \(b\); +- \(b/4\,H_{xx}(M)\) Richardson candidate. + +**Cause:** no raw table, exact claim statement, or source hash exists in the +available C091/C093/Q0-X artifacts. + +**Downstream impact:** Q0-IV, Q0-B, Q0-SHARP. + +**Disposition:** `BLOCKED-EXTERNAL`. The unblock condition is receipt of the +immutable source plus independent rerun. + +--- + +## E-C094-11 — C093 declared a two-sided core whose own source retained losses + +**Evidence:** E-C094-2 and E-C094-3. + +**Finding:** the semantic v4 graph could pass because it encoded theorem +statements, not the arithmetic assembly and propagation formulas behind them. + +**Cause:** missing `ASSEMBLY`, `DOMAIN-INFIMUM`, and source-formula +cross-check gates. + +**Downstream impact:** C094 corrected core; Q0-LLM verifier v5. + +**Disposition:** `CORRECTION`. Live C094 root is the upper-rate theorem; +verifier v5 must include the three named gates. + +--- + +## E-C094-12 — C093 release-close terminology overstated Pass-1 reproducibility + +**Finding:** the frozen ZIP is immutable and verifies before mutation, but the +documented four-command sequence is not itself idempotent. + +**Cause:** a writer was placed before hash verification. + +**Downstream impact:** C093 closeout wording only. + +**Disposition:** `ERRATUM`. C093 bytes remain immutable; C094 carries the +replacement procedure. + +--- + +**Pass-1 E-ledger entries:** 12 +**Homeless entries:** 0 + + + + + +--- + + + +# C094 UB-G DISPLAY ADJUDICATION + +**Item:** Directive v2.1 §2.2 +**Disposition:** **CORRECTION branch** +**Frozen C093 archive:** unchanged +**Affected claim IDs:** `UB_G`, `RATE_PROGRAM_GRADE`, `Q0_LIMIT` + +## Finding + +At 60 decimal digits, + +\[ +\frac{0.030449}6(576-9\pi)(1.014) += +2.8185310984873743106224. +\] + +Using the registered near coefficient \(0.657\), + +\[ +\frac{0.657+2.8185310984873743106}{0.80} += +4.344413873109217888277999. +\] + +Using the displayed rounded inputs, + +\[ +\frac{0.66+2.82}{0.80}=4.35. +\] + +The frozen display \(4.3\) is therefore a downward rounding by \(0.05\) from +the displayed-input assembly and by +\(0.0444138731092178883\) from the unrounded-input +assembly. No C092 correction-ledger entry or displayed certificate licenses +that downward movement. + +## Correction + +Replace every live occurrence of the upper coefficient \(4.3\) by \(4.35\) +in the C094 successor release. + +The corrected program-grade theorem is + +\[ +\boxed{ +0.8411r^3 +\le +1-q(r,6/5) +\le +4.35r^3, +\qquad +0 + + +--- + + + +# C094 LOWER-COEFFICIENT DIMENSIONAL ADJUDICATION + +**Directive row:** §2.3(3) +**Disposition:** `ADJUDICATED-CONFLICT` → **CORRECTION** +**Frozen C092/C093 archives:** untouched + +## Source-level formula + +The lower-side source states + +\[ +C^*(r)\ge0.8411 +\] + +and then assembles + +\[ +1-q(r,6/5) +\ge +C^*(r)r^3\,AO(r)\,(1-O(r^3)). +\] + +It separately records + +\[ +AO(r)\ge1-2.24r^3-2e^{-111}-\text{named residues}. +\] + +Therefore \(0.8411\) is a lower coefficient for the qualifying first moment, +not automatically the final defect probability. + +At \(r=0.025\), even before Bonferroni and the other named residues, + +\[ +0.8411\left(1-2.24(0.025)^3\right) += +0.8410705615 +< +0.8411. +\] + +The displayed finite lower theorem cannot follow from the displayed inputs. + +## Correction + +Withdraw the live core claim + +\[ +1-q(r,6/5)\ge0.8411r^3. +\] + +Preserve the correctly typed statement + +\[ +C^*(r)\ge0.8411, +\qquad0 + + +--- + + + +# RECONCILIATION REPORT C094 + +**Pass:** 1 +**Status:** PASS +**Blocking UB-G item:** resolved by CORRECTION +**Frozen C091/C093 bytes:** unchanged + +## Corrected live boundary + +\[ +\boxed{0\le1-q(r,6/5)\le4.35r^3,\quad0 + + +--- + + + +# C094 Cross-Line Reconciliation + +**Rule:** no import without independent re-derivation or re-measurement; conflicts are adjudicated, never averaged. + +| ID | Item | Disposition | Grade | Owner | Result | Evidence | +|---|---|---|---|---|---|---| +| XL-01 | UB-G upper 4.3 vs 4.35 | ADJUDICATED-CONFLICT | CERTIFIED-ARITHMETIC | C094 CORRECTION | 4.3 rejected; 4.35 is the conservative replacement. New roots filed. | `C094_UBG_ADJUDICATION.json` | +| XL-02 | Six-pin transverse conditional-mean factorization | IMPORTED-AFTER-REVERIFICATION | DERIVED-EXACT | Q0-SHARP / Q0-B | m_L(x,y)=k_L(y)m_L(x,0) for the exact periodized field; mean-Hessian identities follow. BR-MARK reshaped, not closed. | `C094_SIX_PIN_FACTORIZATION.json` | +| XL-03 | Lower coefficient 0.8411 dimensional status | ADJUDICATED-CONFLICT | DERIVED-EXACT CONTRACT AUDIT | C094 CORRECTION / Q0-SHARP | 0.8411 retained as C*(r) first-moment coefficient; finite defect lower display withdrawn pending AO/Bonferroni propagation. | `C094_LOWER_COEFFICIENT_ADJUDICATION.json` | +| XL-04 | Height-universality / q flat in b | REGISTERED-SUCCESSOR-ITEM | MEASURED-CANDIDATE | Q0-IV / Q0-B | Raw operator-line data and immutable source hash are not present in the available C093/C091 artifacts; registered BLOCKED-EXTERNAL until supplied and independently rerun. | `C094_PORTFOLIO_CHARTER.json` | +| XL-05 | PALM-SUSC/STAB-UB Stage-0 priors and de-truncation artifact | IMPORTED-AFTER-REVERIFICATION | MEASURED-PROXY | Q0-IV | At b=.2, wrap 0.2433→0.0433→0.00333→0 for L=24,48,96,192; mean area L192/L96=0.93886. Single-value-pin proxy only. | `C094_IMPORTED_PRIORS_RECHECK.json` | +| XL-06 | Selected-set Campbell coefficient lambda | IMPORTED-AFTER-REVERIFICATION | MEASURED-PROXY | Q0-IV / Q0-LLM | Cross-fitted moving-selector lambda rechecks to 4.5087–4.8462; registered as selection-amplification instrument. | `C094_IMPORTED_PRIORS_RECHECK.json` | +| XL-07 | b/4 H_xx(M) Richardson candidate | REGISTERED-SUCCESSOR-ITEM | HYPOTHESIS | Q0-SHARP | Exact candidate statement and raw Richardson table are absent from available artifacts. BLOCKED-EXTERNAL pending source; must be promoted by independent derivation or killed. | `C094_PORTFOLIO_CHARTER.json` | +| XL-08 | Citation content fixes | IMPORTED-AFTER-REVERIFICATION | ESTABLISHED-SOURCE-CHECK | Q0-REFEREE | Ladgham typed law has three additional powers; Azaïs–Delmas aggregate N=2 is neutral; Nicolaescu IDs disambiguated by title. | `C094_CITATION_PROVENANCE.json` | +| XL-09 | Repulsion vs marked-window error family | ADJUDICATED-CONFLICT | DERIVED-SOURCE-AUDIT | Q0-REFEREE / Q0-SHARP / Q0-B | Spatial/type correlation and pair-Palm two-height marked density are separate claims; no citation-only transfer is licensed. | `C094_CITATION_PROVENANCE.json` | +| XL-10 | Bonferroni allowed-product domain infimum | IMPORTED-AFTER-REVERIFICATION | CERTIFIED-ARITHMETIC | Q0-SHARP / verifier v5 | Full-domain infimum is 79.9889203915 as r→0; allowed=80 crossing at r≈0.00183603628. Rung minimum is invalid. | `C094_IMPORTED_PRIORS_RECHECK.json` | +| XL-11 | C093 verification order | ADJUDICATED-CONFLICT | CERTIFIED-REPRODUCTION | C094 release engineering | Audit script mutates manifested report; verify-before-writer or versioned audit output required. | `C094_FRESH_STATE_VERIFICATION.json` | + +## Governing outcomes + +- corrected upper coefficient: **4.35** +- finite lower `0.8411 r^3`: withdrawn from the live C094 core; +- exact six-pin mean factorization: imported at derived-exact grade; +- Bonferroni domain-infimum target: **79.988920391521175719575325883263398928955399268802**; +- every measured import remains explicitly Stage-0/proxy unless its full pair-Palm design is executed. + + + + +--- + + + +# C094 C091 → C093 Inter-Release Provenance Audit + +| File | Classification | C091 SHA | C093 SHA | Cause / disposition | +|---|---|---|---|---| +| `Q0_C091_PROOF_GATE_REDUCTION.md` | CONTROL-CHARACTER-REPAIR | `128ab070189f` | `10e1f4e1fa73` | hidden form-feed interpretation after LaTeX \frac caused split 'rac' lines; C093 repaired the rendered equations; semantic equations restored; no theorem-status change | +| `C091_CORRECTION_LEDGER.md` | IDENTICAL | `b0f806b3ff5c` | `b0f806b3ff5c` | byte-identical across frozen bundles; no ledger action beyond identity record | +| `q0_c091_canonical_contract.json` | C091-ARCHIVE-ONLY | `—` | `—` | C093 closeout selected a provenance subset; the complete file remains immutable in the verified C091 bundle; not overwritten or lost; C091 bundle remains the authority for this artifact | +| `q0_c091_contract_checker.py` | C091-ARCHIVE-ONLY | `—` | `—` | C093 closeout selected a provenance subset; the complete file remains immutable in the verified C091 bundle; not overwritten or lost; C091 bundle remains the authority for this artifact | +| `q0_c091_contract_check_report.json` | C091-ARCHIVE-ONLY | `—` | `—` | C093 closeout selected a provenance subset; the complete file remains immutable in the verified C091 bundle; not overwritten or lost; C091 bundle remains the authority for this artifact | +| `q0_c091_gate_reduction.py` | C091-ARCHIVE-ONLY | `—` | `—` | C093 closeout selected a provenance subset; the complete file remains immutable in the verified C091 bundle; not overwritten or lost; C091 bundle remains the authority for this artifact | +| `q0_c091_gate_reduction_report.json` | C091-ARCHIVE-ONLY | `—` | `—` | C093 closeout selected a provenance subset; the complete file remains immutable in the verified C091 bundle; not overwritten or lost; C091 bundle remains the authority for this artifact | +| `periodized_bf_matrix_transfer.py` | PORTABILITY-EDIT | `91cb008e45df` | `3b1c44449f52` | hard-coded /mnt/data output/module paths replaced by paths relative to __file__; formatting-only changes where present; algorithm and declared mathematical result unchanged | +| `periodized_bf_matrix_transfer_report.json` | IDENTICAL | `75f423ed03bd` | `75f423ed03bd` | byte-identical across frozen bundles; no ledger action beyond identity record | +| `gaussian_corridor_certificate.py` | PORTABILITY-EDIT | `d8bd3f12b341` | `46f373df22bb` | hard-coded /mnt/data output/module paths replaced by paths relative to __file__; formatting-only changes where present; algorithm and declared mathematical result unchanged | +| `gaussian_corridor_certificate_report.json` | DIAGNOSTIC-REEXECUTION | `b6a70cfa846b` | `8b874396913a` | SciPy multivariate-normal orthant diagnostics are numerical algorithm outputs without a frozen random integration state; re-execution changed only diagnostic CDF leaves; all Chernoff candidates, covariance inputs, counterexample labels, invalid-product verdicts, and H4 replacement gates are unchanged | +| `palm_weighted_gaussian_chernoff.py` | PORTABILITY-EDIT | `e2a308393bcd` | `6a32d466d028` | hard-coded /mnt/data output/module paths replaced by paths relative to __file__; formatting-only changes where present; algorithm and declared mathematical result unchanged | +| `palm_weighted_gaussian_chernoff_report.json` | IDENTICAL | `1a46553e7b57` | `1a46553e7b57` | byte-identical across frozen bundles; no ledger action beyond identity record | +| `bf_two_critical_value_gap.py` | PORTABILITY-EDIT | `d0324cdc9b9a` | `0a0b5e24bebe` | hard-coded /mnt/data output/module paths replaced by paths relative to __file__; formatting-only changes where present; algorithm and declared mathematical result unchanged | +| `bf_two_critical_value_gap_report.json` | IDENTICAL | `acf6691a68d9` | `acf6691a68d9` | byte-identical across frozen bundles; no ledger action beyond identity record | +| `q0_c091_corridor_product_audit.png` | C091-ARCHIVE-ONLY | `—` | `—` | C093 closeout selected a provenance subset; the complete file remains immutable in the verified C091 bundle; not overwritten or lost; C091 bundle remains the authority for this artifact | +| `q0_c091_bonferroni_budget.png` | C091-ARCHIVE-ONLY | `—` | `—` | C093 closeout selected a provenance subset; the complete file remains immutable in the verified C091 bundle; not overwritten or lost; C091 bundle remains the authority for this artifact | + +## Same-size corridor report + +- changed scalar leaves: **12** +- only orthant diagnostic fields changed: **True** +- adjudication object unchanged: **True** +- maximum relative diagnostic shift: **4.174e-07** + +The exact Gaussian/Chernoff inputs and every terminal adjudication remain unchanged. The varying fields are explicitly labeled diagnostics in the report; they are not theorem inputs. + +## Governance conclusion + +All changed same-name files are now ledgered. The C091 bundle remains the immutable authority for its own bytes; the C093 variants are new-release portability or repair variants, not silent overwrites. + + + + +--- + + + +# C094 Citation Provenance and Content Adjudication + +## Verified primary sources + +### Ladgham–Lachièze-Rey + +`arXiv:2209.04150`, *Local repulsion of planar Gaussian critical points*. + +The paper distinguishes the aggregate critical-point process from typed +subprocesses. Its discussion states that extrema and saddle subprocesses have +**three additional orders of magnitude** of local repulsion. The Q0 +prior-art surface must not describe this typed law as merely linear. + +### Azaïs–Delmas + +`arXiv:1911.02300`, *Mean number and correlation function of critical points +of isotropic Gaussian fields and some results on GOE random matrices*. + +The abstract states: + +```text +N > 2: attraction +N = 2: neutrality +N = 1: repulsion +``` + +Strong maxima–minima repulsion is separate. This paper does not license a +claim of aggregate planar repulsion. + +### Nicolaescu disambiguation + +```text +arXiv:1101.5990 +Critical sets of random smooth functions on compact manifolds + +arXiv:1209.0639 +Random Morse functions and spectral geometry +``` + +Both identifiers are valid, but for different papers. The v3.2 sentence that +pairs `1209.0639` with *Random Morse functions and spectral geometry* is +consistent. A citation that attaches `1209.0639` to *Critical sets...* must be +corrected to `1101.5990`. + +## Q0-REFEREE rule + +Spatial/type repulsion and the pair-Palm two-height marked-window density are +distinct claims. No external citation discharges `BR-MARK` without an +explicit mark-transfer theorem. + +## Frozen-release scan + +The C093 theorem contract and core verifier do not embed these citations. The +retained C091 audit cites Ladgham–Lachièze-Rey bibliographically. Therefore +the citation corrections alter the successor referee package and the +v3.2-line prior-art wording, not the frozen C093 executable theorem graph. + + + + + +--- + + + +# C094 Q0-X Intake and Verifier-Lineage Continuity + +| Feature | v2 | v3 | v4 | Continuity route | +|---|---:|---:|---:|---| +| proof DAG validation, cycle/supersession checks | ✓ | ✓ | ✓ | imports Grade/ClaimNode/ProofGraph from v3 | +| incremental descendant invalidation and Merkle certificates | ✓ | ✓ | ✓ | v3 dependency + SemanticProofGraph Merkle roots | +| H0 widest-path grounding and exact forest compression | ✓ | ✓ | ✓ | imports SupportEdge/widest_path_grounding from v3 | +| anchored universal-kernel RKHS mismatch | ✓ | ✓ | ✓ | imports anchored_rkhs_mismatch from v3 | +| anisotropic Q-SARD tube bound and effective rank | ✓ | ✓ | ✓ | imports anisotropic_sard_tube_bound/effective_transverse_rank | +| CoverageCertificate | — | ✓ | ✓ | direct import from v3 | +| SelectionAwareRiskLedger and coverage-limited rank | — | ✓ | ✓ | direct import from v3 | +| semantic theorem-contract gates | — | — | ✓ | SemanticProofGraph.validate | +| MARK and COMPOSITION gates | — | — | ✓ | SemanticProofGraph.validate | + +## Result + +- v4 subsumes every validated v2/v3 component in the directive. +- `q0_llm_verifier_v3.py` remains a direct import dependency of v4. +- the clean-extraction v4 validation passed. +- Q0-X v2 is registered as the experiment and theorem-design source for Q0-IV and Q0-LLM. +- v5 additions are new C094 scope, not retroactive claims about v4. + + + + +--- + + + +# C095 E-LEDGER + +**Cycle:** C095 +**Policy:** append-only; supersede-never-overwrite. + +--- + +## E-C095-1 — Gate Framework v1.0 mixed status, assumption role, and warrant + +**Finding:** `Primitive`, `Proven`, `Open`, and `Killed` were presented as one +linear epistemic ladder. + +**Cause:** unlike concepts were compressed into a convenient ordering. + +**Impact:** dependency-grade rules became contradictory and difficult to +implement mechanically. + +**Disposition:** `CORRECTION`. Master v1.1 separates: + +```text +claim status +assumption role +warrant class +reasoning mode +``` + +--- + +## E-C095-2 — Composition witness list omitted UNION_BOUND + +**Finding:** v1.0 required a witness for unions while its closed witness list +did not contain `UNION_BOUND`; the worked example also suggested disjointness. + +**Cause:** products, unions, and monotone assemblies were treated as one +operation. + +**Impact:** an implementation could incorrectly require independence for a +union bound or accept a product under a generic “assembly” label. + +**Disposition:** `CORRECTION`. v1.1 uses operation-specific witnesses and adds +`UNION_BOUND`, `MONOTONE-ASSEMBLY`, and `INTERVAL-ARITHMETIC`. + +--- + +## E-C095-3 — The standalone master omitted C094-mandated numerical gates + +**Finding:** Master v1.0 contained 13 gates but omitted: + +```text +ASSEMBLY +DOMAIN-INFIMUM +BAND-PROVENANCE +``` + +**Evidence:** E-C094-2, E-C094-3, E-C094-5, E-C094-8, and E-C094-11. + +**Disposition:** `CORRECTION`. Master v1.1 contains 16 gates plus the SCHEMA +preflight. + +--- + +## E-C095-4 — COMMON-MODE and HEURISTIC-BRIDGE retrodiction evidence was not self-contained + +**Finding:** v1.0 Appendix A described two examples for each new gate but did +not identify frozen failure IDs, artifact hashes, or the earlier catch +mechanism. + +**Impact:** the gates' logic is implementable, but the meta-claim that they +passed the framework's own retrodiction protocol could not be independently +verified from the accessible release. + +**Disposition:** `PROVISIONAL`. Both gates are implemented. Their +framework-level retrodiction grade remains provisional until the evidence +package is supplied. + +--- + +## E-C095-5 — Initial v5 ASSEMBLY implementation did not consume named residual terms + +**Finding:** the first v5 draft stored `residual_terms` but did not require +those terms to appear in the input map and arithmetic expression. + +**Catch mechanism:** self-application of Gate 14 to the current UB-G object. + +**Impact:** the draft could have repeated the exact failure it was designed to +catch. + +**Disposition:** `CORRECTION`. v5 now fails when any required or residual term +is absent or unused. + +--- + +## E-C095-6 — C094 4.35 UB-G correction still lacked a full residual coefficient + +**Finding:** the C094 base coefficient + +\[ +(0.657+2.8185310984\ldots)/0.80 += +4.3444138731\ldots +\] + +does not itself include the positive Gamma and collar residuals named by the +source. + +The rounded display inputs + +\[ +(0.66+2.82)/0.80=4.35 +\] + +leave zero display-input margin. + +**Impact:** the uniform decimal \(4.35r^3\) is not promotable until + +\[ +\sup_{0