almanac ← the almanac

Human-facing editorial referee report — goalv0a.tex and proofv0a.tex

Lens: SANDBOX_STANDARD §1 criterion 6 (human-facing editorial referee), at full weight. Route A makes these manuscripts the outward almanac artifact, so the reader I refereed for is a competent mathematician outside this program, reading the PDF with no access to the sandbox record.

Date: 2026-08-12. Subjects (as read):

file sha256 lines
informal/goalv0a.tex 7c46f318f07d2703c94595a8a867bd297b65d48f15eee4e6a66d06e815b1d44d 111
informal/proofv0a.tex c42cbf8fc204bc04773dc308e9dd4c29efb65d64a3ebc1a2eeaacc6feb875cd1 414

Context read: versions/goalv0a_SIGNED_2026-07-19.tex (byte-frozen anchor, sha 91a50173…5a7e), versions/goalv0a_SIGNED_2026-07-19_current-notation.tex, versions/proofv0a_pre-notation_2026-08-10.tex, REFEREE_REPORT_PROOFv0a.md (hostile mathematical referee, CONFIRM-WITH-FINDINGS), CITATION_VERIFICATION_2026-08-09.md, ../SCOPE_CONTRACT.md, ../goalv0a.lock.json, ../APPROVAL_STATE.json, proofv0a.ledger.json + proofv0a.current.report.json, and the cited source at corpus/literature/arxiv_src/1601.06110/{main.tex,main.bbl} plus its Library transcript.

This report is advisory and read-only. No manuscript byte was touched.

VERDICT (also restated at the end)

PASS-WITH-EDITS. 2 BLOCKING · 15 EDITORIAL · 12 NIT. The mathematics is sound and the statement has not drifted from the signed anchor (mechanical proof below). Both BLOCKING findings are one-sentence fixes.

Semantic-drift check vs the signed anchor

VERDICT: NO DRIFT. The live goalv0a.tex is a pure notation transform of the byte-frozen signed anchor, verified mechanically, not by eye.

Method: I applied the lock's declared dictionary (goalv0a.lock.json → notation_migration_2026_08_10.dictionary) to the signed anchor's own bytes — \Az→\At, \Kz→\Ft, A_z→A_t, K_z→F_t, then Q→q, q→v, z→t on standalone math tokens, then the product index t→s inside Definition 5 — and diffed the result against the live edition with the three declared non-mathematical items normalised away:

  1. the added migration note paragraph (live lines 25–30),
  2. \newtheorem{goal} → \newtheorem*{goal} (post-signing style rule: the Goal is unnumbered),
  3. \boxed{…} removed from clause (ii) (post-signing style rule: the Goal is displayed plainly),
  4. \date{July 19, 2026 --- draft, unsigned} → \date{}.

Result: byte-identical. Every ring, every relation, every quantifier, all three clauses. In particular: A_t = Z[v^{±1}, t^{±1}] is the signed A_z = Z[q^{±1}, z^{±1}] under the dictionary — the family parameter was renamed (z ↦ t), not squared, and the base ring is the same ring. K_{n,+}K_{n,-}=t^{\,n} is the signed =z^{\,n}. ev_χ(Ŷ)=q^{χ} is the signed =Q^{χ}. Clause (iii)'s sum is the signed sum. No weakening, no strengthening, no hypothesis added, no convention changed.

I ran the same transform against versions/proofv0a_pre-notation_2026-08-10.tex. The migration of the proof is likewise mathematically inert; the only non-mechanical differences are (a) the added edition note, (b) the index rename t→s, (c) three places where an over-eager earlier pass had renamed Harman–Hopkins' own parameter q to v and the live edition correctly restores their symbol ($q$-Pascal, $q$-binomials, $q$-deformed, \mathcal R_q^{+}), (d) a repaired $Z[…]$ → $\Z[…]$ typo at line 99, (e) a \qedhere display upgrade at 241–244. All improvements.

One caveat, and it is the report's first BLOCKING finding: the sentence that announces the migration to the reader (goalv0a.tex:26–28, proofv0a.tex:34–36) describes the dictionary in a way that invites the reader to conclude the opposite of what I just verified. See B1.

BLOCKING (a reader is misled)

B1 — the migration sentence asserts t = z², with z undefined, and thereby invites the reader to infer a ring change that did not happen

goalv0a.tex:26–28:

Signed July 19, 2026; this is the notation-migrated edition of August 10, 2026 (same statement, current symbols: the former $Q,q,z$ are written $q,v,t$, with $q=v^{2}$ and $t=z^{2}$).

proofv0a.tex:34–36 carries the identical clause.

Three defects compound here:

  1. z is the only symbol in either manuscript that is used and never defined. It occurs exactly twice in the two documents outside this sentence — nowhere in goalv0a.tex, and once in proofv0a.tex:384 ("our square root of $t$"), which presupposes the very definition that is missing. First-use completeness fails on the manuscript's first line of prose.
  2. The parallel construction is false. "$q=v^{2}$" restates the signed frame's own Q=q^{2} under the rename and is correct. "$t=z^{2}$" has no counterpart in the signed frame: the signed z had no square root. Set side by side as a matched pair, the two look like the two halves of one dictionary, which they are not.
  3. The natural misreading is exactly the drift the reader is checking for. A reader who takes "the former z is written t, with t=z²" at face value concludes that the new family parameter is the square of the signed one, hence A_t = Z[v^{±1}, z^{±2}] ⊊ Z[q^{±1}, z^{±1}] = A_z — a strictly smaller base ring, i.e. a different (weaker) theorem. That conclusion is wrong, but the sentence is what produces it. The one paragraph whose job is to certify "same statement" is the one paragraph that undermines it.

Proposed rewording (both files, identical clause):

Signed July 19, 2026; this is the notation-migrated edition of August 10–11, 2026. The statement is unchanged — only symbols were renamed, by the Sandbox-wide dictionary $Q\mapsto q$, $q\mapsto v$, $z\mapsto t$: the relation formerly written $Q=q^{2}$ now reads $q=v^{2}$, and the family parameter formerly written $z$ is now written $t$.

If the program wants the four-symbol current frame on the record, append one sentence rather than folding it into the dictionary:

(In the current Sandbox-wide frame, $v$ and $z$ denote the square roots $v^{2}=q$ and $z^{2}=t$; only $v$ and $t$ occur below.)

Owner must confirm which reading is intended before editing — if t = z² was meant as a relation between the signed z and the current t, then my drift verdict above and the lock's own dictionary are in conflict with the manuscript, and that is an AMEND-class question, not an editorial one. Everything I measured says it was not meant that way.

B2 — the Attribution remark attributes Harman–Hopkins' Lusztig identification to the wrong object

proofv0a.tex:383–386:

The node-preserving affine substitution $\Yc=1+(q-1)x$ (their \S4 $z$-variable --- their symbol, unrelated to our square root of $t$ --- which is remarked there to give the Cartan part of Lusztig's integral form) carries the grid node $x=[n]_q$ to $\Yc=q^{n}$ …

The which attaches to "their §4 $z$-variable", so the manuscript says HH remark that the $z$-variable gives the Cartan part of Lusztig's integral form. They do not. The source (corpus/literature/arxiv_src/1601.06110/main.tex:417, verbatim) says:

Going one step further, if we formally adjoin square roots $K^2 = z$ and $v^2 = q$, then the ring $\mathcal{R}q[K,v]$ is equivalent to the Cartan part of Luztig's [sic] integral form of the quantum group $U_2)$.}(\mathfrak{sl

The identification is of R_q[K,v] — the $z$-variable plus the adjoined square roots — not of the $z$-variable. On an outward artifact whose citations are meant to be checkable, a reader who opens the source finds the manuscript's claim inexact. The correction is also more favourable to this program, and CITATION_VERIFICATION_2026-08-09.md step 3 already records why: HH's adjoined K (with K²=z) and v (with v²=q) are literally this manuscript's T^{+} (with (T^{+})²=Ŷ) and v — the program frame is the square-root cover their remark constructs. The manuscript is giving away its own best sentence and misciting at the same time.

The em-dash aside ("their symbol, unrelated to our square root of $t$") should go with it: it leans on the undefined z of B1, and post-migration the collision it defends against no longer exists (their z vs our t are different letters now). The positive statement — their z is our Ŷ — is the one the reader needs.

Proposed rewording:

The node-preserving affine substitution $\Yc=1+(q-1)x$ is exactly their \S4 $z$-variable; it carries the grid node $x=[n]_q$ to $\Yc=q^{n}$ and carries each basis term $\qbinom{x}{k}$ onto $\nu_k(\Yc)$ \emph{term by term} --- the factors $(q-1)$ cancel in every Newton ratio. (They remark further that on adjoining square roots $K^{2}=z$, $v^{2}=q$ --- which is precisely the passage to our $T^{+}$ and our $v$ --- the ring $\mathcal R_q[K,v]$ is equivalent to the Cartan part of Lusztig's integral form of $U_v(\mathfrak{sl}_2)$. Nothing below rests on that remark.)

EDITORIAL (should fix)

E1 — proofv0a.tex:150–151 — clause (ii)'s key equality rests on an injectivity that is never stated

the middle equality because $\Uev=\Nev\subseteq U$ maps into $U\otimes_{\Ft}\operatorname{Frac}(\At)$.

"Maps into" is not enough to license Uev ∩ (U ⊗ Frac(A_t)) = Uev, which requires Uev to sit inside the base-changed algebra as a subset — i.e. requires U → U ⊗_{F_t} Frac(A_t) to be injective. It is (the hostile referee supplied the reason at its item 8), but the manuscript never says so. This is the one place in the document where a careful reader is asked to accept a step the text does not justify, and it is in the shortest, most-read proof.

Proposed: "…the middle equality because $U\hookrightarrow U\otimes_{\Ft}\operatorname{Frac}(\At)$ is injective — $\Ft\hookrightarrow\operatorname{Frac}(\At)=\mathbb Q(v,t)$ and $U$ is free over $\Ft$ — so $\Uev=\Nev\subseteq U$ is literally a subset of the base-changed algebra."

E2 — proofv0a.tex:80 — Φ⁺ is used in a document that never introduces a root system

Because $\Phi^{+}=\emptyset$ at $GL_1$, the \emph{even hybrid family integral form} is …

proofv0a.tex advertises itself as stand-alone (title, and line 33). It defines v, t, q, A_t, F_t, U, T^±, Ŷ, ev_χ, N^ev, W — but never Φ or Φ⁺. A reader without goalv0a.tex meets an undefined symbol carrying the load of the definition.

Proposed: "Because the root system of $GL_1$ is empty ($\Phi=\Phi^{+}=\emptyset$), the …"

E3 — proofv0a.tex:310–340 — Lemma 3's proof states the same argument three times, and says so twice in its own voice

The core argument occupies lines 300–308 and is complete there. It is then repeated as The $t$-spectator argument (310–325), which closes "This is the same conclusion reached above", and again as Base change (327–340), which closes "Equivalently, and most directly:" followed by a verbatim restatement of lines 306–308. The hostile referee already flagged the second repetition as "correct but redundant (not load-bearing)". For the human reader this is the manuscript's largest readability defect: 30 of a 40-line proof are restatement, twice self-declared as such, and it reads as though the author did not trust the argument.

Proposed: keep 280–308 unchanged; replace 310–340 with two sentences, e.g.

\emph{Where $t$ goes.}\; The map $\ev_m$ substitutes only $\Yc\mapsto q^{m}$; it touches neither $v$ nor $t$, so $V$ and $V^{-1}$ have entries in $\Z[q^{\pm1}]$ and the system decouples over the $t$-powers. No flatness or base-change argument is needed: $V^{-1}$ has entries in $\Z[q^{\pm1}]\subseteq\At$ and $\mathbf v\in\At^{d+1}$, so $\mathbf c\in\At^{d+1}$, and in particular honest odd powers of $v$ in the $c_r$ cause no difficulty.

This is the single largest quality gain available and the single largest diff; see the churn analysis.

E4 — proofv0a.tex:375, 387 — the same object typeset two different ways, one of them a tall inline box

The document defines \qbinom (line 23) and uses it correctly six times (196, 199, 216, 236, 239, 242). The Attribution then writes the same $q$-binomial coefficient polynomial as $\begin{bmatrix}x\\k\end{bmatrix}_{q}$ twice. Besides the inconsistency, an inline bmatrix is a two-row box in a text line: it inflates leading in the surrounding paragraph in the PDF. (This is the referee's finding F1 half-discharged — the un-deformed \binom was correctly removed, but the replacement did not use the document's own macro.)

Proposed: $\qbinom{x}{k}$ at both sites.

E5 — proofv0a.tex:236–238 — [k]_q! used, only [k]_q defined

$=\frac{[n+r-1]_q!\,/\,[n-1]_q!}{[r]_q!}=\qbinom{n+r-1}{r}$, where $[k]_q=(q^{k}-1)/(q-1)$ and the common powers of $(q-1)$ cancel.

The factorial is introduced in the display and glossed only for the non-factorial. It reappears at line 377.

Proposed: "…where $[k]_q=(q^{k}-1)/(q-1)$, $[k]_q!=[1]_q[2]_q\cdots[k]_q$, and the common powers of $(q-1)$ cancel."

E6 — proofv0a.tex:314–316 — wrong base ring named for the basis actually used

expand each coordinate along the free $\Z[v^{\pm1}]$-basis of $t$-powers: $c_r=\sum_{k}c_{r,k}\,t^{k}$ with $c_{r,k}\in\mathbb Q(v)$.

A Z[v^{±1}]-basis with coefficients declared in Q(v) is a mismatch: {t^k} is a basis of F_t over Q(v) and of A_t over Z[v^{±1}], and the argument uses both. As written the reader has to repair the sentence before using it.

Proposed: "expand each coordinate along the $t$-powers — a basis of $\Ft$ over $\mathbb Q(v)$, and of $\At$ over $\Z[v^{\pm1}]$ — writing $c_r=\sum_k c_{r,k}t^{k}$ with $c_{r,k}\in\mathbb Q(v)$." (Survives the E3 compression; fold in whichever version is kept.)

E7 — proofv0a.tex:372, 378 — internal Library labels leak into the human-facing document

Their \S1 Proposition (Prop.~\S1.2, \texttt{prop:qbinom}), the … … their \S4 Proposition (Prop.~\S4.3, \texttt{prop:localization}), …

\texttt{prop:qbinom} and \texttt{prop:localization} are LaTeX labels from this program's transcript of the source. They are meaningless to an outside reader and read as machine noise in a typeset paper. Separately, "Prop.~\S1.2" reads as a section number; the source prints "Proposition 1.2".

Proposed: "their Proposition 1.2, the $q$-analogue of …" and "their Proposition 4.3, $\mathcal R_q=\mathcal R_q^{+}\otimes_{\Z[q]}\Z[q^{\pm1}]$, is the bilateral localization." (Both numbers verified against the source.)

E8 — proofv0a.tex:406–410 — \cite{SandboxGL2} is unresolvable to any reader outside the program

\bibitem{SandboxGL2} SandboxA, \emph{The centre of the even hybrid family quantum $GL_2$} (closure record of the signed goal, revision A1+R5), Sandbox record, 2026.

No identifier, no location, no availability statement — yet this is the sole support for the clause (ii) parenthetical (line 136) and for the "rank-one specialization of that transport" claim (line 396), i.e. for the manuscript's entire in-program ancestry. On the outward artifact the reader is told a result exists and given no way to reach it.

Proposed: state what it is and that it is internal, e.g. "…, signed goal center-GL2, closure record (revision A1+R5), Project Sandbox internal record, 2026 (not publicly available)."

Gate caution: the obvious richer fix — naming the certified Lean declarations (nuNewtonIntegral_iff_mem_span, gl2_center_closed_newton_laurent), which CITATION_VERIFICATION_2026-08-09.md verified at their source lines — risks tripping manuscript_gate P2 (no … formalisation language) and P3 (no computational evidence). Test the gate before committing to that variant.

E9 — proofv0a.tex:134, 396 — "rank-one" here, "rank-0" everywhere else in the sandbox

The rank-one, $\Phi^{+}=\emptyset$ degeneration of the even hybrid Harish--Chandra isomorphism … The $GL_1$ statement above is the rank-one specialization of that transport.

SCOPE_CONTRACT.md ("the rank-0 base case of the root GOAL"), APPROVAL_STATE.json ("GL_1 (the rank-0 base case…)") and REPORT.md ("degenerate at rank 0") all say rank 0. Both are defensible (torus rank 1, semisimple rank 0), and neither document defines which it means. A route-A reader who reads the manuscript and then the sandbox page meets both numbers for the same object.

Proposed: on first use, "the rank-one (equivalently, semisimple-rank-zero) degeneration …", or drop the number: "the $GL_1$ degeneration of …". Harmonise with whatever the outward pages say.

E10 — goalv0a.tex carries no % criterion-13: adopted, and the gate says that blocks publication of the edition [publication-gating]

manuscript_gate.py --kind goal on the live file returns PASS, but with:

NOTE: criterion 13 (ancestry) is reported but NOT enforced for this file — it carries no % criterion-13: adopted line. … An edition may not be published while any of its manuscripts still lacks the declaration.

followed by five G9 NOTICEs (lines 42, 55, 68, 78, 89 — every definition, no citation and no \nopredecessor). proofv0a.tex adopted the rule at its line 2 and passes P9 cleanly. Since route A is a publication path, the goal manuscript is currently the thing standing in front of it.

This is not an editorial edit and I do not recommend making it as one. Adding the declaration converts the five NOTICEs into FAILURES, so it must be accompanied by a citation or \nopredecessor on five definitions of a hash-locked, Overseer-signed statement. Route to the Overseer under the SANDBOX_STANDARD §5 amendment protocol, with the observation that four of the five definitions are specializations of the root informal/goal.tex and the fifth (Newton divided classes) has the Harman–Hopkins ancestry the proof already cites.

E11 — goalv0a.tex:58–60 — ev_χ is under-determined as defined (inherited; recommend NO edit)

For $\chi\in X_\Lambda$ let $\mathrm{ev}\chi$ be the $\At$-linear map on even toral elements determined by $\mathrm{ev}\chi(\Yc)=q^{\,\chi}$.

An A_t-linear map on F_t[Ŷ^{±1}] is not determined by its value at Ŷ; multiplicativity is what pins it. The proof silently upgrades this to "the $\Ft$-algebra homomorphism substituting the unit $q^{\chi}$ for $\Yc$" (proofv0a.tex:67–69), which is correct and defines the same N^ev. The looseness is verbatim in the signed anchor — it is not drift, and the hostile referee recorded the same point (its finding 3).

Recommendation: record, do not edit. Repairing it touches the locked statement.

E12 — goalv0a.tex:32–34, 58, 64 — X_Λ is never said to be anything (inherited; recommend NO edit)

the cocharacter lattice is $\Lambda=\mathbb Z$, and $X_\Lambda=\mathbb Z$.

The reader is told the value and never the meaning, then χ is quantified over X_Λ at line 58 and over ℤ at line 64, six lines apart, with no word that they are the same index set. Inherited verbatim from the signed anchor. Record, do not edit (a gloss would change a definition of a locked statement). Worth one line in whatever front-matter the almanac wraps around the PDF.

E13 — proofv0a.tex:36–38 — the opening sentence mislabels what is proved

The result identifies the even hybrid integral form of the family quantum $GL_1$ with its centre and with a Newton (divided-class) lattice;

U^{hyb,ev} is N^ev by definition (line 81, and goalv0a.tex Def 3). So "identifies … with a Newton lattice" describes a definition, not a result. The actual content of clause (iii) is the identification of the evaluation-integrality lattice with the span of the divided classes — which the next clause of the same sentence then correctly names as HH's theorem. The first thing the reader is told about the paper is the one thing the paper does not do.

Proposed: "The result shows that the even hybrid integral form of the family quantum $GL_1$ — cut out by integrality of all evaluations on the grid $q^{\mathbb Z}$ — is an $\At$-subalgebra, is its own centre, and is spanned over $\At$ by the shifted Newton divided classes; clause~(iii) is the bilateral quantum Pólya theorem of Harman and Hopkins, transported to the multiplicative grid of the toral algebra (attribution remark at the end)."

E14 — proofv0a.tex:74 — the referee's finding 3 re-examined: the manuscript is RIGHT, no edit

It is the unique $\At$-linear ring homomorphism with $\ev_\chi(\Yc)=q^{\chi}$, and it fixes $\Ft$ (hence $\At$) pointwise.

REFEREE_REPORT_PROOFv0a.md finding 3 calls this uniqueness claim not "literally forced". I checked it and it is forced: an A_t-linear ring homomorphism fixes Z[v^{±1}] pointwise, and a ring homomorphism into a domain that fixes Z[v] pointwise fixes every a/b ∈ Q(v) (from φ(a/b)·b = φ(a) = a), hence fixes F_t = Q(v)[t^{±1}]; together with ev_χ(Ŷ)=q^χ that determines the map on F_t[Ŷ^{±1}]. No edit. Recorded so the point is not re-raised a third time. (The sentence order would read better as "It fixes F_t (hence A_t) pointwise, and is the unique ring homomorphism doing so with ev_χ(Ŷ)=q^χ" — cosmetic only.)

E15 — proofv0a.tex:162–164 — a bare paragraph doing a remark environment's job

\noindent There is no content here beyond commutativity: at $GL_1$ the rational centre is everything, $\HC$ is the identity, and Weyl invariance is vacuous.

The document declares a remark environment and uses it once (line 368). This paragraph is a remark set as loose body text. (Its content is exactly the honesty item SCOPE_CONTRACT.md requires "wherever v0a is cited", and it discharges it — the objection is only to the packaging.) Wrapping it renumbers the Attribution remark 1→2; nothing in the ledger cites a Remark number.

NIT

Verified clean (checked, no finding — recorded so these are not re-litigated)

  1. "Pólya's 1919 theorem" (proofv0a.tex:371–372) is FAITHFUL to the source, despite the misleading Library bib key polya1915ganzwertige. The source writes \begin{proposition}[P\'olya 1919~\cite{polya1915ganzwertige}] (main.tex:99) and the entry is J. Reine Angew. Math. 149 (1919), 97–116 (main.bbl:266–284). I flagged this as a suspected date error and it survived.
  2. Every Harman–Hopkins quotation is verbatim: the R_q^+ definition, "freely generated as a Z[q]-module by the q-binomial coefficient polynomials", the denominator q^{C(k,2)}[k]_q!, the additive grid x=[n]_q, Proposition 4.3 R_q = R_q^+ ⊗_{Z[q]} Z[q^{±1}], and the node map x=[n]_q ↦ Ŷ=q^n. Numbering 1.2 and 4.3 both check out against the source's per-section counter.
  3. "their parameter $q$ coincides with our $q$ ($=v^{2}$), so their displayed formulas read verbatim on our side" (proofv0a.tex:380–381) is correct and is the migration's genuine payoff — the pre-migration edition had to carry a dictionary here and no longer does.
  4. The term-by-term transport claim (proofv0a.tex:386–388) is exact. I recomputed: ∏_{s<k}(q^k−q^s) = q^{C(k,2)} ∏_{i=1}^{k}(q^i−1), which is precisely HH's denominator, so qbinom(x,k)|_{x=(Ŷ−1)/(q−1)} = ν_k(Ŷ) on the nose and the (q−1) factors do cancel.
  5. The SCOPE_CONTRACT honesty stance is honoured, not oversold. Clause (iii) is named as Harman–Hopkins' theorem in the opening paragraph (line 38), again in the theorem's own parenthetical (178–181), and the in-program GL_2 ancestry is stated at 392–396 with the GL_1 statement called "the rank-one specialization of that transport". The mandatory Weyl caveat is discharged at 162–164. I found no sentence claiming clause (iii) as new.
  6. No legacy symbol survives the migration anywhere in either manuscript. Q, A_z, K_z, K_t and the product index t appear nowhere except inside the migration sentence itself (B1) and in the deliberate reference to HH's own z (B2).
  7. Both files build warning-free (goalv0a.log, proofv0a.log: zero warnings, zero over/underfull boxes, no undefined references). manuscript_gate: goal PASS (8 rules + 1 NOTICE, see E10), proof PASS 9/9 with criterion-13 enforced.
  8. goalv0a.pdf sha cfaa2f86… matches goalv0a.lock.json → current_edition_pdf_sha256.

Which edits are worth the byte-churn

Any byte change to proofv0a.tex forces a ledger document re-pin (document_sha256_at_transcription), a battery re-run, a PDF rebuild, and a gate re-run. You pay that once whether you fix one finding or fifteen — so batch them into a single pass, or do none.

Worth it — one batched pass on proofv0a.tex: B2, E1, E2, E3, E4, E5, E6, E7, E8, E9, E13, plus N5, N6, N8, N9, N10 folded in for free. E3 is the largest diff and the largest reader gain; including it is the whole reason to open the file.

Worth it — one pass on goalv0a.tex, confined to the migration note (lines 25–30): B1, N6. This paragraph is editorial matter added at migration; it is not part of the signed statement, so it can be repaired without the amendment protocol. Update current_edition_tex_sha256 and current_edition_pdf_sha256 in goalv0a.lock.json afterwards.

Not worth it / must not be done as an editorial edit:

Collateral (outside my two subjects; read-only, reported to the owner)

Not findings against the manuscripts, but a route-A reader who follows the manuscript into the record will hit these:

  1. APPROVAL_STATE.json → notation_migration_2026_08_10.current_proof_tex_sha256 = "7110a472…" is STALE. The live proofv0a.tex is c42cbf8f…. The ledger and its report are current (both pin c42cbf8f…, re-pinned by document_revision_note_2026_08_11 after the F_t rename); it is the approval record that was not carried forward. Anyone verifying the manuscript against the approval record today gets a hash mismatch.
  2. CITATION_VERIFICATION_2026-08-09.md addendum (lines 82–92) quotes the same stale 7110a472… and states the dictionary as "A_z→A_t, K_z→K_t" — superseded by the F_t naming of 2026-08-11, which the live manuscripts already use.
  3. REPORT.md still reads "Approval-2 … NOT SIGNED"; APPROVAL_STATE.json records it SIGNED at 2026-07-20T12:59:23Z. The report is dated 2026-07-20 (wave 3) and predates the signature by hours.
  4. Not a defect: the ledger's claim strings are in legacy symbols ([m r]_Q, Z[Q^{pm}], z-spectator). proofv0a.ledger.json → frame explicitly declares this — the expressions are code against the shared engine preload, whose API keeps the legacy names, and the note supplies the translation. Correctly handled.

VERDICT

PASS-WITH-EDITS.

class count
BLOCKING 2
EDITORIAL 15
NIT 12

Semantic-drift check vs the signed anchor: NO DRIFT — the live goalv0a.tex is byte-identical to the mechanical dictionary transform of versions/goalv0a_SIGNED_2026-07-19.tex (sha 91a50173…5a7e), modulo the added migration note and the three declared post-signing style conformances. The proof's migration is likewise mathematically inert.

Both BLOCKING findings are one-sentence repairs (B1 in both files, B2 in the proof). Neither the mathematics nor the statement is in question; what is in question is that the paragraph certifying "same statement" invites the reader to infer a smaller base ring (B1), and that the one place where the manuscript reports what its cited source says, it reports something the source does not say (B2). On an outward artifact both must go before publication. Everything else is quality of exposition, and the compression of Lemma 3 (E3) is the item that would most change how the manuscript reads.