Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
159 changes: 159 additions & 0 deletions reviews/d5_consumption_bridge_20260928/BRIDGE.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,159 @@
# Consumption bridge for recorded D5 / D1 pieces

Scientific effect: **NONE**. This file records exact consumption edges among
already-reviewed statements and reusable finite identities. It does not accept
Math-#116 continuum claims, Math-#118, Math-#120, catalog C6, or C8.

Reviewer: xAI / Grok. Same GitHub account as other lanes; no organizational
independence is claimed.

Base: Math- `main` at `4f6abcb05bd59a7770f12cb7f79640d8b7a9a734`.

## 1. Recorded ACCEPT pieces used as inputs

| ID | Statement | Record |
|---|---|---|
| P2 | punctured-pin first moment, all-height nested disk | `reviews/d5_local_collar_20260928/PUNCTURED_PIN_PROOF.md` |
| collar C1 | `E N_j(r E) \le C r^3 \|E\|` on compact collar `C(\u03b7,R)` | Math-#119 |
| scaled-ball C2 | same bound on scaled ball of fixed `R` minus pins | Math-#119, corollary of P2 + collar C1 at `η=1/4` |
| I3, I4 | height-window shells and intermediate annulus | Math-#117 |
| D4 A | remote first moment `O(r^3 \|E\|)` on `D_\u03c1` | `frontiers/remote_window_20260924/PROOF.md` |
| I5 | `E N_{I_r}(T^2 \setminus \{M,S\}) \le C r^3` | Math-#117: C2 + I4 + D4 A |
| F− | `P(A_r) \ge c r^3` | Math-#110 remote A+E |
| F+ | `P(A_r) \le C r^3` | Math-#117 Markov on I5 |
| Pη | `E N_\u03b7(N_\u03b7-1) \le C_\u03b7 r^5` at fixed `η>0` | Math-#119 / #110 |
| H1–H3 | generic TV identities | Math-#115 |
| H4 | remote singleton law at fixed `D_\u03c1` | Math-#115 |
| Thm A (1.1) | `0 \le 1-p_r \le C r^3` compact B,K, frames | parent (1.1), review `reviews/d1_theorem_a_nonauthor_20260928/` |

Event `A_r = \{ N_{I_r}(T^2 \setminus \{M,S\}) \ge 1 \}`.

## 2. Closed downstream edges

### E1. Window first moment (already recorded)

I5 ← C2 + I4 + D4 A.

Height window on I4 is load-bearing. All-height intermediate `O(r^3)` is false
for the shell method and is not an input.

### E2. Planar event probability (already recorded)

F+ ← Markov(I5), F ← F− ∧ F+.

In `d=2`,

c r^3 ≤ P(A_r) ≤ C r^3.

This is an event probability on the torus minus the two pins. It is not
`E N(N-1)` and not an elder-pairing theorem.

### E3. Remote singleton at fixed `D_\u03c1` (already recorded)

H4 ← H1–H3 + D4 A + #110 Corollary D.

Scope is a deterministic positive-volume `E \subset D_\u03c1`. Not torus-wide
uniqueness, not a height law, not Poisson.

### E4. Compatibility of H4 and planar F

H4 lives on `D_\u03c1`, which excludes a neighborhood of the pins. Planar F counts
any extra window critical point on `T^2 \setminus \{M,S\}`, including the pin
neighborhood that H4 does not see. Both statements can hold at once: remote
configurations are close to Bernoulli at rate `O(r^5)\|E\|`, while the event
that *some* extra window point exists anywhere off the pins is `Θ(r^3)`.

### E5. Thin-tube first moment is not an open obligation

The thin-tube region sits inside any fixed scaled ball. C2 already bounds that
first moment. The tube note remains a more precise intensity; it is not an
unfinished first-moment gate for I5 or F.

### E6. Compact-mark selection-difference *upper* bound

Parent `imports/lifetime_parent_20260925/UNIFORM_MATRIX_CAP_AND_LIFETIME.md`
§11:

r dr/dℓ = (1/3) k^{-2/3} ℓ^{-1/3}, (11.1)
ν_cand(ℓ) = ℓ^{-1/3} ∫ A_{(ℓ/k)^{1/3}} / (3 k^{2/3}). (11.2)

On a compact mark window `B × K` with `k \ge k_- > 0`, `A_r` is bounded by the
parent A3 majorant and `∫_K k^{-5/3} dk < ∞`. Theorem A (1.1) gives
`0 \le 1-p_r \le C r^3`. Substituting `r = (ℓ/k)^{1/3}` into the difference
form of (11.2) yields

0 ≤ ν_cand(ℓ) − ν_eld(ℓ) ≤ C_{B,K} ℓ^{2/3}.

This uses only the *upper* half of (1.1). It does not use Math-#116 C1, does
not claim sharpness `Θ(ℓ^{2/3})`, and does not extend to unrestricted marks
(catalog C7).

### E7. Off-pin second moment does not upgrade to the torus

Split, for any `η>0`,

N = N_η + N_{<η},
E[N(N-1)] = E[N_η(N_η-1)] + E[N_{<η}(N_{<η}-1)] + 2 E[N_η N_{<η}].

Pη bounds the first summand. The other two summands require a pair floor with
at least one witness in `{\|X\|<η}`. Cauchy–Schwarz on I5 does not produce
those terms. Catalog C6 remains open. The elementary comparison
`N(N-1) \ge 2 1{N\ge2}` is unconditional; applying it to a proposed
`P(N\ge2) \ge c r^3` is conditional on catalog C2 and is not used here.

## 3. Reusable finite identities

These are polynomial / calculus identities. They may be consumed by a later
continuum review. They are not Gaussian theorems.

### L1. Collar B-frame Gram

det(BB^T) = u^2 v^4 + v^6/4 = v^4(4u^2+v^2)/4.

Nonnegative, and positive exactly when `v != 0`. It vanishes on the entire
axis `v = 0`, including `(u,v) = (1,0)`. The collar source treats that axis
separately; its reduced B-frame estimate is used only away from it.

### L2. Overlap integral

For `k > 0` and `t != 0`, the finite identity is

∫_0^∞ s (k − s^3 |t|)_+ ds = (3/10) k^{5/3} |t|^{-2/3}.

Direct antiderivative on `s \le (k/|t|)^{1/3}`:
`k s_*^2/2 − |t| s_*^5/5` with `s_*^3 = k/|t|`.

At `t = 0`, `k > 0`, the integral diverges; no finite value at that point is
asserted by the displayed formula.

### L3. Pinned cubic and path

G_k = k(2X^3 − 3X/2 − 1/2 − 3Z^2/4 − X Z^2).

Pins: `G(±1/2, 0)` and `∇G(±1/2, 0) = 0`, with values `0` and `−k`.
Path `(−1/2,0) → (−3/4,0) → (−3/4,9/4) → (−1,9/4)` takes values
`0`, `−7k/32`, `−7k/32`, `+17k/64`. Extra critical points at
`X=−3/4`, `Z^2=15/8`: height `−7k/32`, `det H = −15 k^2/2`.

### L4. Gap Beta moments

Density `(10/9) g^{-1/3}(1-g)` on `(0,1)` is Beta`(2/3,2)`.
Mean `1/4`. Variance `9/176`.
Joint density `(5/9)|a-b|^{-1/3}` on `(0,1)^2` is a different object from this
scalar gap law.

### L5. Factorial split

`E[N(N-1)]` expands as in E7. Pointwise `N(N-1) \ge 2 1{N\ge2}`.

## 4. Explicit non-edges

- Math-#116 C1–C5 continuum steps, including `1-p_r \ge c r^3` and the pair kernel `K_{ij}`.
- Catalog C6 torus-wide second factorial moment.
- Catalog C7 unrestricted difference rate.
- Catalog C8 numerical constants.
- Math-#118 P15 (separate program).
- Math-#120 remote height decoupling (C5 refinement; not a corollary of H4).
- Feeding planar F or I5 into H4 as a torus-wide uniqueness theorem.

No STATUS / GRAPH / lemma_closed / prize change.
38 changes: 38 additions & 0 deletions reviews/d5_consumption_bridge_20260928/DOMAIN_CORRECTION.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,38 @@
# Domain correction to the finite identities in BRIDGE.md

Author: OpenAI / ChatGPT, foreground review, 2026-09-28.
Scientific effect NONE. No governing status or parent mathematical proof changes.

Original review: xAI / Grok, Math- commit
9cf35211ff53c770e30c2b69fbe742accd2acc64. Original BRIDGE.md:
6024 bytes; Git blob21d38b3b1ae94fab3270a0c7b85182b528964d06;
SHA2563248439014601c41d46c20ab55150d17b06bd3313d2f208afc26e30faedcf23f.
The original is retained in Git history. Review5344767944 and pickup5879018889
scope this correction; it does not relabel Grok's other analytic dispositions.

L1: v^4(4u^2+v^2)/4 is positive exactly when v is nonzero, not whenever (u,v)
is nonzero. At (1,0) it is zero. The actual collar source at
4f6abcb05bd59a7770f12cb7f79640d8b7a9a734,
reviews/d5_local_collar_20260928/COLLAR_PROOF.md Sections6-7, handles the axis
separately and uses the reduced B-frame where |v|>=r^(1/3)>0. No change to that
source is required or made.

L2: the displayed overlap integral is finite for k>0 and t!=0. At t=0,k>0,
it is the integral of ks over the half-line and diverges. This excluded-point
statement does not invalidate later integrability in t; it prevents a false
pointwise finite value from being consumed.

Corrected BRIDGE.md:6361 bytes;
Git blob7e1ef1c1d9f6df7f8781c3332fcb069d01c41dd9;
SHA256f4bc7a843a4763241e36fb14bf5920fbddd767c5d3f62a864549e0d6b9cdaaf9.
Only these two finite-identity domains changed. All other bridge content,
README.md and RESULTS.json are unchanged.

Actual local verification:5 controls per normal/optimized Python mode. The
exact original source produced2 intended assertion failures in each mode;
the corrected source passes5/5 in each. Controls check the domain text,
Gram factorization including axial examples, exact overlap integration at
rational cube cutoffs, and equality to the scoped textual correction.
This is not a full-repository or continuum theorem replay. The test program,
original/corrected bytes and logs are retained in the conversation evidence.
Required hosted repository checks must be read separately before integration.
6 changes: 6 additions & 0 deletions reviews/d5_consumption_bridge_20260928/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,6 @@
# D5/D1 consumption bridge — review record

Scientific effect: NONE.
Does not change STATUS, GRAPH, lemma_closed, prizes, PROOF_INDEX, or author sources.

Read BRIDGE.md.
15 changes: 15 additions & 0 deletions reviews/d5_consumption_bridge_20260928/RESULTS.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
{
"scientific_effect": "NONE",
"review_status": "RECORD_ONLY",
"I5": "ACCEPT_COROLLARY",
"planar_F": "ACCEPT_THETA_R3_EVENT",
"H4": "ACCEPT_AT_D_RHO",
"P_eta": "ACCEPT",
"C1_collar": "ACCEPT_EXISTENTIAL",
"C2_scaled_ball": "ACCEPT_COROLLARY",
"compact_mark_difference_upper": "ACCEPT_GIVEN_THEOREM_A_1_1",
"thin_tube_first_moment": "SUPERSEDED_BY_C2",
"C6_torus_wide": "NOT_ACCEPTED",
"C1_elder_lower": "CANDIDATE",
"continuum_verified": false
}
Loading