session 2026-08-08: REVIEWED-97 (vi disposition) placed + REVIEWED-98 draft; PENDING-114 authorized, PENDING-115 filed

claude/governance/REVIEWED-97-draft-2026-08-08.md

@@ -0,0 +1,27 @@
+## REVIEWED-97 — PENDING-113 — Disposition (vi): `voice:` for anonymous, traditional and scripturally-claimed matter
+
+**Date:** 2026-08-08
+**Decision:** DISPOSED. Unblocks remediation-order step 1; step 2 (re-tag the Mauss blocks) may proceed per item, under the structure below. Step 3 (`citable: true`) stays gated behind step 2, as ruled.
+**Full reasoning:** `studium-engine/docs/voice-non-individual-origin-2026-08-08.md`. The compressed rule below is **not** a substitute for it — whoever writes the next per-source note reasons from that document, not from two tokens.
+**Provenance:** jurist design-gate 2026-08-08 (category vocabulary + two-job structure); executor correction of the field placement, from substrate the jurist cannot read (PENDING-86); steward's bibliographic-vs-theological distinction is the ground of the split.
+
+**The disposition — four slots, each doing one job.** The category vocabulary is the jurist's, unchanged; what changed is where it lives.
+
+- **identity — `voice:`.** Who is convoked. `voice:` is the retrieval key (`retrieve.py` filters `d.voice = ?`), so one value IS one speaker: measured, `glidden` already spans 5 sources and `weil` spans 2, correctly, because one person is behind each. Individual-author voices aggregate at the person; traditional and scriptural matter has no such person, so **identity lives at the work level** — `havamal`, `quran-taghabun` — or at the **passage** level where there is no work (the Trobriand formulae). Read as unique per *quoted work*, not per host; `quoted_by` distinguishes hosts. Executor names the values.
+- **relation — `quoted_by:`.** Unchanged from REVIEWED-96 Q3: a quoted span grounds the host's REPRODUCTION, never the quoted author's AUTHORSHIP. This, not the voice name, is what limits a thin voice built from an excerpt.
+- **category — closed pair, and NOT in the convocation key.** `traditional` (no author claimed, by anyone, within the source's own tradition) · `non-individual-origin` (the tradition itself holds this is not a human composition, by whatever mechanism it names).
+- **account — per-source prose note.** The specific claim each tradition makes about its own text's origin does not compress into a token; it goes in prose, one per item.
+
+**Why the category may not sit in `voice:`.** Putting it there would make `traditional` a single convocable speaker spanning Old Norse gnomic verse and Sanskrit epic, with no mechanism to separate them again — the same flattening the ruling exists to prevent, committed in the mechanism instead of the taxonomy.
+
+**Binding caution.** Do NOT assign the remaining items — Havámál, Trobriand formulae, Mahābhārata, Brahmanic passages — to `traditional` by elimination. At least one (the Trobriand formulae, tied in their own cosmology to ancestral or mythic origin) may need `non-individual-origin`, or a note complicating either bucket. Each item gets its own considered note before tagging. **Writing the note is the work; the two-way choice is not a substitute for it.**
+
+**Correction carried into this entry.** The text is **Surah LXIV (`at-Taghābun`), vv. 15–18** — not CXIV (`an-Nās`), which is a different six-verse surah. Mauss's own line reads *"la fameuse Sourate LXIV, « déception mutuelle »"*. The mislabelling originated in the executor's sidecar title and propagated into REVIEWED-96 (Q4), PENDING-113 and session memory; those three name the wrong surah and should be read with this correction. The earlier worked example was built on the "Say" (قُل) formula, which opens an-Nās and does **not** appear in the passage Mauss quotes — that evidence is withdrawn, not transplanted. The general ground (the Qur'án presents itself as divine speech) is untouched. Mauss's *"donnée à La Mecque"* is contestable — at-Taghābun's Meccan/Medinan classification is disputed — and the note **records the dispute rather than resolving it**.
+
+**Not disposed here.** All twelve (vi) blocks are translated matter and `role` is single-valued, so each is *quoted-in* and *translated-from* at once. REVIEWED-96 filed that composition as open against Ungaretti-in-Harrison, as an edge case; it covers the whole of (vi). The disposition is decided; it cannot be **applied** until that composition is ruled.
+
+**Scope.** Disposes (vi) only. REVIEWED-96 Q1–Q6 and PENDING-113's Q2/Q4/Q5 conditions are untouched. The Harrison/Mark finding raised alongside this is filed separately as **PENDING-114**, not as evidence for this entry.
+
+**Numbering note (PENDING-110).** This is `REVIEWED-97`, the next in its own sequence — not `REVIEWED-113`. The registers are independent; PENDING-110 REJECTED renumbering-to-align, and numbering this 113 would skip 97–112 and entrench the false expectation that the sequences match. The heading names its PENDING per that item's convention (b).
+
+---

claude/governance/REVIEWED-98-draft-2026-08-08.md

@@ -0,0 +1,23 @@
+## REVIEWED-98 — PENDING-114 — Scripture quoted inside a host text, unmarked: authorize the validated census
+
+**Date:** 2026-08-08
+**Decision:** **(b) AUTHORIZED**, with **(c) attached exactly as proposed** for whatever (b)'s validated recall cannot reach. (a) — fix the Harrison span and stop — REJECTED.
+**Provenance:** steward, 2026-08-08, on the executor's PENDING-114. Reasoning recorded here rather than compressed, per PENDING-108/113: a ruling that exists only in conversation evaporates.
+
+**Ground 1 — (a) would repeat, in miniature, the error already named in this same thread.** Fixing Harrison alone treats a class-shaped problem with a span-shaped fix. That is not an analogy to `118f411`; it is the same taxonomy error, identified hours earlier the same day and now written into the register against it.
+
+**Ground 2 — the scale context, which PENDING-114 understates.** Harrison's quotation is invisible to any sidecar-based detector because it belongs to the intra-line class measured 2026-08-07 at **~6,455 marked quotation runs across 8 sources, ~94% of them intra-line** and therefore not expressible at the current section granularity. Harrison is **not an isolated miss — it is the first confirmed hit in a population that is this corpus's blind spot for precisely the property the corpus exists to guarantee.** First-order gap, not a curiosity.
+
+**Ground 3 — why (b) rather than merely thorough.** (b) is built so as not to repeat the marker census's own failure. It does not propose trusting a signal; it proposes **demonstrating recall against a hand-scored known-positive / known-negative pair** — Harrison's actual quotation against Weil's mentions — before any corpus-wide claim is made. That is the discriminating-instance discipline (REVIEWED-83 A1) applied to the instrument *before* it is applied to the corpus.
+
+**Ground 4 — the payoff, and the proposed doctrine earning its keep forward.** Whatever (b) finds is **marked, not fenced.** That is a materially different act from `118f411`: attribution removes nothing from the test population, it improves it. A correctly-marked Harrison/Mark instance becomes a **better** positive control than an unmarked one, not a destroyed one. REVIEWED-96's proposed doctrine — *a fix that enforces a property can destroy the population that tests it* — is here running **forward in time**, as a design constraint on remediation rather than as a post-hoc finding. This is the first occasion on which it has done real work.
+
+**On (c) as the attached fallback.** If the validated detector's recall does not reach some part of the class, **disclosing that limit is a legitimate stopping point, not a failure to execute (b) properly** — the same standing (c) has in REVIEWED-48's gate-to-abstain, which was ruled doctrine-consistent rather than merely tolerated. (c) is not a consolation branch; it is the honest-degradation route (Constraint #4) and is authorized on its own terms.
+
+**Binding on execution.** No corpus-wide claim — finding *or* all-clear — may be reported from the detector until its recall is demonstrated on the hand-scored pair. The marker census already run is a **failed instrument**, not a count, and is not to be cited as either evidence or reassurance.
+
+**Scope.** Disposes PENDING-114 only. REVIEWED-96, REVIEWED-97 and PENDING-115 are untouched; PENDING-115 remains a separate blocker on remediation step 3.
+
+**Pacing.** Execution timing left to the executor as a proportionate-to-energy call, explicitly not decided by the substantive case above.
+
+---