Gap-08 — Inflation spectrum: the gate anchor ledger

The honest one-line: Gap-08 has two genuine, target-blind narrowing wins — one candidate inflaton is killed by computation and the surviving plateau's tilt direction is read off the frozen split with no dial — and on the ratified board the gate closes as DISSOLVED-GIVEN-root · RESOLVED +0: decisively, it lives inside a sector the programme has declared out of scope, so even a perfect spectrum match would stay a diagnostic consistency, never a derivation; the residual family stays shown.

This page is the gate-specific instantiation of the anchoring method: it takes the master anchor and the four bridges and applies them, object by object, to one gate. Every exact thing Gap-08 touches gets its own row — its status, what it is allowed to claim, and what it is forbidden to claim. It follows the same eleven-part shape and the same universal eight-column table as the canonical SG-4 ledger; it adapts honestly to this gate's nature — a conditional / diagnostic gate inside the excluded cosmology sector, whose central structural datum is a theorem-candidate, not a banked output.

1. Gate status header

• Gate-level Gap-08 roll-up: DISSOLVED-GIVEN-root · RESOLVED +0 (board-canonical, ratified 2026-07-08; closure-of-record: the live gate dossier) · CANDIDATE (testable). Direction: held. Binding label: Conditional / Diagnostic; EXCLUDED cosmology sector (Axiom I / SG-10 §9.3.2); the earlier roll-up under the superseded least-closed-residual rule read OPEN (cascade).
• Taxonomy reconciliation (2026-07-05): On the ratified 2026-07-08 board, the gate-level status is DISSOLVED-GIVEN-root · RESOLVED +0 (excluded sector) — in-scope $\lambda^2=1/6$ is CLOSED-NEGATIVE (the $K=24$ denominator is geometry-forced by the breathing-mode normalization $d(d+2)/2$ at $d=6$; the numerator is a selected-term model choice, so $\lambda^2=1/6$ is a convention that is retired, not a forced output), and the inflation sector is a declared-excluded input — read as terminal reached + residuals shown. The residual family listed on this page (A1-GEOM, III-b, B3, $c_{\rm loop}$, $N_*$, $A_s$, the LiteBIRD bet, and the scope wall) remains listed and carried unchanged; the earlier “OPEN (cascade) · CANDIDATE” roll-up above reflects the superseded least-closed-residual rule, not different facts.
• What "local/partial closed leg" means here. Gap-08 has no anomaly-style closed arithmetic identity. Its analog of a closed local leg is the forward chain run from the frozen split to a falsifiable spectrum — the geometry, given the frozen split $n=(6,2,1)$, hands a single surviving inflaton and a convention-independent falsifier: $$ \text{(frozen split)} \;\longrightarrow\; \{\rho \text{ DEAD},\ \sigma \text{ survives}\}\;\longrightarrow\; \lambda^2 = \frac{4}{K_{\sigma\sigma}} = \frac16 \;\Longrightarrow\; \text{red tilt, pre-registered }r\text{-band}. $$
• Status, split so it cannot be misread:
  • Win #1 — $\rho$-inflaton DEAD (KK-stiff): DERIVED-GIVEN-E / qualitative grade — a computed verdict; a certificate-grade kill is an optional sharpening, not required and not claimed.
  • Win #2 — plateau slope / tilt direction $\lambda^2 = 1/6$: CERTIFICATE-CONDITIONAL / THEOREM-CANDIDATE — dial-free given the frozen split at the arithmetic layer, but the forcing of $K_{\sigma\sigma}=24$ fails closed on an unread frozen source (BLOCKER A1-GEOM). It is a candidate computed verdict, not a banked derivation.
  • The $\kappa_{13}$ gravitational anchor: DERIVED-CLOSED (closed-by-computation).
  • The breathing axis: REDUCED-TO-AXIOM / closed-by-source.
  • The K13-2 normalization (III-a parity bit; III-b $(9/11)^2$ factor), $c_{\rm loop}$, $N_*$, $A_s$: AXIOM-OPEN / OPEN / BLOCKED, as marked below.
  • The excluded-sector scope wall: AXIOM-OPEN (atomic by kind) — the gate's dominant terminal; never closeable in-scope.

Given the frozen split, the geometry does real narrowing: it computes one candidate field dead and fixes the tilt direction with no dial. What stays open is everything past those two moves — the forcing of the slope, the magnitude (which relocates onto $\ge 2$ un-traced conventions rather than reducing), the absolute amplitude, the e-fold window, the ~2030 measurement, and the scope wall the gate can never cross.

selection ≠ derivation · given-E ≠ derivation-of-E · frozen/reproducible ≠ proven-unique.

2. Frozen inputs (what Gap-08 stands on, not what it produces)

• Frozen branch hashes dcc66f1b2685 / manifest a5b1e6f9d951. The 13D K₆ branch is READ-ONLY; nothing in this gate mutates or closes it. These hashes are audit anchors — they certify which object was tested and that it cannot be quietly retuned. They do not validate the physics.
• Dimension split $n=(6,2,1)$ — a facet of the frozen branch, the claimed source of $K_{\sigma\sigma}$ and hence the slope. GIVEN-E / upstream-inherited, not Gap-08-derived.
• K13-2 datum $81/22$ — FROZEN, source-confirmed at corpus line L10268; arithmetic re-verified exact: $81/22=(11/2)\cdot(9/11)^2=9^2/(2\cdot 11)$. AUDIT-anchored at the byte level; its physics meaning is the gate's central open residual (see §8).
• Measured comparators (Planck $n_s\approx 0.965$; bound $r<0.036$; $A_s\approx 2.1\times10^{-9}$; $N_{\rm eff}=3.044$): held strictly as comparison targets, never inputs. The $\ge 1$ measured-invariant floor anchor is paid by transfer-onto, not anchor-on.

3. Object anchors (given-E / upstream)

The structures the gate acts on, all inherited from the frozen geometry:

• Two candidate inflaton directions $\{\rho,\sigma\}$ — the geometry's own moduli, read off the split (not posited).
• The breathing modulus $\sigma$ — controls the overall scale of the compact internal space; the surviving candidate.
• The breathing-mode potential (corpus §II): $$ V(\sigma) = c_{\rm KK}\,e^{-4\sigma} + c_{\rm bdry}\,e^{-2\sigma} + c_{\rm Wilson}\cos\theta_W\,e^{-4\sigma} + c_{\rm loop}\,e^{-6\sigma}. $$ In the corrected (v10) regime this supports an exponential plateau — the regime giving small $r$ and a near-universal tilt.

Status: GIVEN-E / upstream-inherited — the fields, the split, and the potential's form enter Gap-08 as input. Gap-08 does not derive the spectrum $E$. Every "passes/projects" below is a statement about this geometry, not a derivation of it.

4. Root and master-anchor traceability

Deep roots that are load-bearing for Gap-08:

Physical equivalence / invariance is not a primary load-bearing anchor for the Gap-08 narrowing.

Master anchors in play: finite invariant ledgers · no unpaid labels · the frozen branch · given-$E$ · the declared K13-2 conventions · open-residual discipline.

5. The Gap-08 anchor ledger (the universal table)

6. The construction — the two wins, at full depth

The geometry touches cosmology through one forward chain — no new geometry is searched, and no in-scope claim may rest on it.

Win #1 — the $\rho$-inflaton is computed DEAD (KK-stiff). Of the two candidate directions the split offers, $\rho$ has an effective mass at the Kaluza–Klein scale, far above the slow-roll scale, so it physically cannot inflate. KK-stiffness is a target-blind discriminator: it converts "which field inflates?" — normally a free choice — into a computed narrowing from two candidates to one. Honest grade: qualitative; the countersign slot is pending; a certificate-grade kill is the optional sharpening, not required because the $\sigma$-survivor carries the gate. The NT-1 $\chi$-leg run gate-halted on a scale artifact (VOID both directions = inconclusive, not evidence either way); the $\rho$ verdict stands independently.

Win #2 — the dial-free plateau slope. For the $\sigma$ exponential plateau, $$ \lambda^2 = \frac{4}{K_{\sigma\sigma}},\qquad \text{(claimed)}\ K_{\sigma\sigma}=24 \ \Longrightarrow\ \lambda^2 = \frac{4}{24} = \frac16, $$ with no dial available once the split is fixed. The tilt direction — a red tilt ($n_s<1$) — is geometry-fixed given the split. The corpus checks this at three layers and fails closed at the third:

1. Arithmetic layer (checkable, trivially true). $\lambda^2=1/6 \Leftrightarrow K_{\sigma\sigma}=24$ is pure bookkeeping. (Verified exact: $4/24=1/6$; note $4/9\ne 1/6$.) This is algebra, not physics.
2. Plateau-phenomenology layer (order-of-magnitude sanity check). A plateau gives $n_s\approx 1-2/N_*$; at $N_*\approx 55$, $n_s\approx 1-2/55\approx 0.9636$ — the same Planck-dead-center region as the banked band — and $r$ at the $10^{-2}$–$10^{-3}$ order. A sanity check, NOT a derivation.
3. Geometry layer — FAIL-CLOSE (BLOCKER A1-GEOM). That $K_{\sigma\sigma}=24$ is forced by the split and not chosen requires reading the frozen $\sigma$-kinetic reduction source (the D.1 region, $\approx$ corpus line L10268), which is not in scope and must not be fabricated. Until read in scope and verified, $\lambda^2=1/6$ stays a THEOREM-CANDIDATE.

Diagnostic — the result is specific, not trivial. The slope's robustness profile discriminates: for an exponential plateau the tilt is dominated by the $-2/N_*$ term and is only weakly sensitive to $\lambda^2$, which is exactly why $n_s$ is recorded as "near-universal across branches" — while $r$ spans $[3.5,36]\times10^{-3}$, a factor of $\sim 10$, across the four conventions. That $n_s$ stays pinned while $r$ fans out by an order of magnitude is what makes the structure a real geometric datum rather than a convention-driven artifact: the direction is forced; the magnitude relocates. A second specificity check: the distinct-quantity caution $24/9=8/3$ is non-trivial and non-zero — $K_{\sigma\sigma}=24$ (slope) and $K=99/11=9$ (normalization) are not the same object, and the $8/3$ bridge must be accounted, not assumed away.

The forward-chain "obstruction" split. The local narrowing is $$ O_{\rm Gap08,local}(\text{split}) = \big(\rho\text{ DEAD},\ \lambda^2 = 1/6\big), $$ closed only as narrowing given the split (qualitative + theorem-candidate). The full gate additionally carries the magnitude conventions, $N_*$, $A_s$, the measurement, and the scope wall — none closed here.

Traceability: the construction, the three-layer slope check, and the re-verified arithmetic are reproduced in the Gap-08 dossier §3.3–3.6.

7. The $\kappa_{13}$ anchor and the BLOCKED amplitude — the closed sub-win, in full

The absolute amplitude $A_s$ rides two inputs; the gravitational coupling that anchors it is closed-by-computation: $$ \kappa_{13}^2 = \frac{1}{\pi^4\, M_U^9\, M_{\rm Pl}^2}\qquad\text{(closed form, corpus-verbatim, frozen radii).} $$ This is a genuine sub-win: pure geometry, no dial; the fold contributes a $\pm$factor-2 band. But the absolute $A_s$ remains BLOCKED on two named inputs: (i) K13-2 — which operative branch, fixed by the $\sigma$-map yaml read (B3); and (ii) $c_{\rm loop}$ — the $\sigma^{-6}$ loop coefficient (inherited from Gap-04). The corpus standing non-claim is verbatim: "No absolute amplitude is claimed in this document." A tilt + ratio with no amplitude is not a spectrum. When B3 + $c_{\rm loop}$ discharge, the frozen $A_s$ relation is evaluated and consistency-checked against Planck $A_s\approx 2.1\times10^{-9}$ — held as a comparison target only, with a $\pm$factor-2 fold tolerance (so no per-entry fitting leaks in; a mismatch beyond that band is a recorded tension, not a silent re-tune).

A scoping fact carried verbatim: with $c_{\rm bdry}$ negative across its band, the well is unconditional in $c_{\rm loop}$ — $c_{\rm loop}$ cannot threaten the existence of the well and does not enter $n_s$, $r$, or the union falsifier; its Gap-08 role is confined to $A_s$.

Traceability: the $\kappa_{13}$ closed form, the BLOCKED $A_s$, and the $c_{\rm loop}$ scoping are reproduced in the Gap-08 dossier §3.7, §3.2.

8. The K13-2 normalization, split into honest objects — "one bit became two un-traced freedoms"

The single phrase "the K13-2 convention" hid more than one claim. The deep audit found the $\sigma$ magnitude does not reduce, it RELOCATES — the most important anti-promotion fact in this gate:

1. III-a — the K13-2 parity bit. The $\sigma$-kinetic denominator: D.1 uses $(D-2)=11$; the textbook 4D reduction would use $(D_4-2)=2$. Status: AXIOM-OPEN / OWNER-DECISION (value-free) — retained as D.1, intentional, recorded; external-to-AI (no source records authorial intent to read), with a demotion clause live. Support cited (NOT proof): the integer normalization $K=99/11=9$ (the textbook denominator would give the non-integer $99/2=49.5$). Support, not proof.
2. III-b — the $(9/11)^2=81/121$ normalization factor. A separate, newly-exposed freedom. The arithmetic spine, re-verified exact: $81/22=(11/2)\cdot(9/11)^2 = 9^2/(2\cdot 11)$. The $(11/2)$ is the raw denominator ratio $(D-2)/(D_4-2)=11/2$; the extra $(9/11)^2$ is mentioned by neither convention axis, and a corpus-wide search for "$81/121$"/"$(9/11)^2$" in this context returns nothing. Status: AXIOM-OPEN / REDUCE_FURTHER (un-source-traced) — carried as the named posit AXIOM-K13-2-RESIDUAL-NORMALIZATION-FACTOR, OPEN byte-trace; it demotes to a derived consequence of the SHAPE-SPLIT only if an external read shows it follows from $n=(6,2,1)$ target-blind.
3. The breathing axis (III-a's partner). REDUCED-TO-AXIOM / closed-by-source — the D.1.0 uniform-breathing declaration resolves it.

Net. The gate carries $\ge 2$ local conventions, not one — it RELOCATES, it does not reduce. The four convention branches span $r$ across $[3.5,36]\times10^{-3}$ (factor $\sim 10$), so the magnitude is convention-anchored, not formula-forced — while the direction and the union falsifier are convention-robust across the four NAMED branches.

Traceability: the III-a / III-b split, the relocation finding, and the re-verified $81/22=(11/2)\cdot(9/11)^2$ arithmetic are reproduced in the Gap-08 dossier §3.8.

9. Open residuals — grouped by family

None is closed by the two narrowing wins. Each is a separate row above; grouped here:

Family A — the forcing of the central datum (theory, AI-pursuable): - A1-GEOM — is $K_{\sigma\sigma}=24$ ($\lambda^2=1/6$) forced by the split? CERTIFICATE-CONDITIONAL / fail-closed on the unread frozen source at L10268. - III-b — source-trace $(9/11)^2=81/121$. AXIOM-OPEN / REDUCE_FURTHER. Co-closes with A1-GEOM and removes the PROVISIONAL flag on the band.

Family B — the magnitude / amplitude conventions: - B3 — read the frozen $\sigma$-map yaml to fix the operative branch (T6 vs T_u). PENDING (yaml not located/read in scope). - $c_{\rm loop}$ — the $\sigma^{-6}$ FRG-2 coefficient. OPEN; gates only $A_s$; the shared Gap-04 ↔ Gap-08 ↔ Gap-09 residue (export, do not re-derive). - $A_s$ — the absolute amplitude. BLOCKED on $c_{\rm loop}$ + B3.

Family C — reproducibility & cascade: - $N_*$ — the e-fold window the bands assume. OPEN / NOT STATED; set by Gap-09 reheating ($T_{\rm RH}\in[2.4\times10^{12},4\times10^{15}]$ GeV).

Family D — the external adjudicator: - LiteBIRD $r$ (~2030) — OPEN (observation bet). $r\notin[3.5,36]\times10^{-3}\Rightarrow$ candidate dies (a refute is a valid close); $r$ inside $\Rightarrow$ the measured value reads off which named convention nature uses ("the convention becomes an observable").

Family E — the binding frame (not a closeable hole): - Excluded-sector scope wall (SG-10 §9.3.2). AXIOM-OPEN, atomic by kind — the dominant terminal; disclose, never "close."

Dissolved unicorns (NOT open weaknesses — shared ceilings): the $\sigma\to-\infty$ UV corner (perturbatively undecidable, tied to the 122-orders Λ hardness that is OPEN for everyone; NON-BLOCKING); a source proof that $(D-2)=11$ was intended (unprovable in principle; OWNER-RULED); and "$\lambda^2=1/6$ is forced under ANY convention / no other geometry could reproduce this" (an absolute universal-negative; the bounded banked claim is forced given the frozen split and the named conventions, with a convention-independent falsifier).

10. Anti-claims (what this page refuses to say)

• Gap-08 does not derive the spectrum $E$, and the gate cannot close in-scope — cosmology is an excluded sector (Axiom I / SG-10 §9.3.2). Formally, even a perfect $n_s/r$ match is a diagnostic consistency, never a derivation, and may never support an in-scope claim.
• $\lambda^2=1/6$ is not a derived/banked output — it is a THEOREM-CANDIDATE, fail-closed at A1-GEOM. The arithmetic identity $4/24=1/6$ is not physics; the physics is whether $K_{\sigma\sigma}=24$ is forced.
• $n_s$ and $r$ are not predictions in the reserved sense — they are PROJECTED at CANDIDATE grade; the Planck $n_s$ agreement is a consistency-check (comparator held target-only).
• The $\rho$-death is not certificate-grade — it is the qualitative KK-stiff verdict; the NT-1 $\chi$-leg is VOID-both-directions (no value asserted).
• $A_s$ is not computed — it is BLOCKED on $c_{\rm loop}$ + the yaml read.
• The σ magnitude does not reduce to one convention — it relocates onto $\ge 2$ un-traced freedoms (III-a, III-b); the union band is PROVISIONAL-on-III-b, not "convention-independent regardless of ANY bookkeeping choice."
• The small-integer coincidences ($K=99/11=9$; the $8/3$ factor; expressibility hit-counts) are not evidence the gate banks — they are open consistency questions / unverified illustrative figures, carried as support, not proof. (The illustrative hit-counts carry no logical-spine weight.)
• The frozen-branch hashes are audit anchors; they do not validate the physics.
• A captured log is not an independent countersign; local narrowing is not whole-gate closure.

11. Specialist closure plan

Each open residual is a concrete, finite, target-blind work-package. A refuting result is a valid close. Every numeric target must pass the no-reverse-engineering from the measured value firewall (writable WITHOUT the comparators).

1. A1-GEOM ($\lambda^2=1/6$ forcing) — read the frozen $\sigma$-kinetic reduction at L10268 in scope, target-blind: confirm or refute that $K_{\sigma\sigma}=24$ follows from $n=(6,2,1)$ under the frozen kinetic metric, and state the $K_{\sigma\sigma}$-vs-$K$ relation (the $8/3$ bridge). Success: $\lambda^2=1/6$ becomes a banked theorem (or Reduce to Axiom AXIOM-BREATHING-KINETIC-NORMALIZATION, value-free, if the source merely records the convention). Refuting close (valid): the source does not single out $K_{\sigma\sigma}=24$ ⇒ the slope is a declared convention range. Do not fabricate the reduction.
2. III-b ($81/121$) — external mechanical read of L10268 to source-trace the $(9/11)^2$ factor. Success: it demotes to a SHAPE-SPLIT consequence ⇒ the band becomes non-provisional and co-closes A1-GEOM. Refuting close: it is a free convention ⇒ a named value-free posit, band stays a disclosed range. (The single highest-value AI-pursuable next move.)
3. B3 ($\sigma$-map yaml) — identify the exact frozen yaml + key under a freeze hash, confirm K₆-only vs uniform breathing, read the single bit. Success: operative branch fixed (decision-grade). Fail-close (valid): the frozen file is not produced ⇒ B3 stays PENDING. Freeze-before-compare; do not claim a read that did not happen (the yaml was not located/read in scope).
4. $c_{\rm loop}$ — run the pre-registered FRG-2 computation at the frozen radii inside the window $0Export to the moduli/loop specialist (shared Gap-04 residue); do not re-derive inside Gap-08. Refuting close: outside the window, or an $A_s$ mismatch beyond the band ⇒ a recorded tension.
5. $N_*$ — publish the e-fold window the bands assume; show consistency with the Gap-09 reheating band; publish $n_s(N_*,\lambda^2)$ and the four-branch $r$-integration. Refuting close: no $N_*$ window consistent with Gap-09 ⇒ a recorded cross-gate tension.
6. LiteBIRD $r$ (~2030) — keep the four bands frozen (freeze-before-compare); await the measurement. Refuting close: $r$ outside $[3.5,36]\times10^{-3}$ ⇒ the candidate dies — an honest close of the bet. Even a confirming reading discharges the falsifier, NOT the scope wall.
7. Scope wall — disclose it as the dominant terminal; there is no "close," only honest disclosure. The boundary is script-enforced at deploy.

Closing Families A–C upgrades the gate from "two narrowing wins" toward a reproducible candidate spectrum — and even then, only given the geometry, and only ever as a diagnostic consistency inside the excluded sector.

Completion tests for this page

Tests passed (required presence, all met): gate roll-up DISSOLVED-GIVEN-root · RESOLVED +0 · CANDIDATE · the closed local narrowing $O_{\rm Gap08,local}(\text{split})=(\rho\text{ DEAD},\lambda^2=1/6)$ · $\rho$-death DERIVED-GIVEN-E/qualitative · $\lambda^2=1/6$ THEOREM-CANDIDATE/A1-GEOM fail-close · "spectrum $E$ not derived" · frozen hashes AUDIT ONLY · the three-layer slope check · diagnostic ($n_s$ near-universal while $r$ fans factor $\sim10$; $24/9=8/3$) · $\kappa_{13}^2$ DERIVED-CLOSED · breathing axis REDUCED-TO-AXIOM · III-a OWNER-DECISION · III-b REDUCE_FURTHER · $c_{\rm loop}$ OPEN (gates only $A_s$) · $N_*$ OPEN · $A_s$ BLOCKED · union falsifier (POSIT atomic, band PROVISIONAL) · LiteBIRD observation bet · scope wall AXIOM-OPEN/atomic · every exact object as its own row · the gate's anti-claims.

Tests failed (required absence, all held): no claim that Gap-08 is closed / in-scope · $E$ derived · $\lambda^2=1/6$ banked · $n_s/r$ asserted as predictions · $\rho$-death certificate-grade · $A_s$ computed · band convention-independent regardless of ANY bookkeeping · coincidences sold as derivations · hashes validate physics · local narrowing = whole-gate closure · any reader-visible build-process vocabulary.

Open items: A1-GEOM · III-b · B3 · $c_{\rm loop}$ · $N_*$ · $A_s$ · LiteBIRD (~2030) · the excluded-sector scope wall (disclosed, never closeable in-scope).

Assumptions made: none beyond the dossier — every number, band, hash, sign, and equation is traceable to the cited dossier sections; nothing was invented. No status on this page was upgraded.

This gate follows the canonical eleven-part shape and the universal eight-column table of the SG-4 ledger.

See also: the anchoring method · the master anchor · Layer 2 — no unpaid exact labels (why a relocated convention is not a reduced one) · Layer 4 — carrier-forcing & the given-E wall · the SG-10 ledger (the excluded-sector scope wall) · the full Gap-08 dossier.