Place REVIEWED-135 AMENDMENT 1 — §8's stated dependency was false

Steward placement, committed. Joined beneath REVIEWED-135, which stands unedited: the
false sentence remains where it stands, superseded in place and visible, per
REVIEWED-132 §6 and the amendment's own PLACEMENT clause.

Verified after placement: the header registers as id+marker with ident REVIEWED-135,
and the register-integrity check raised NO new finding — the parent uses the em-dash
form and parses correctly, so the hazard the placement instructions routed around was
real in general and absent in this case. The `·` form was avoided correctly, though for
a partly stale reason: it does not merely go unseen, it parses to `compound` with the
ident mangled to a bare `REVIEWED`, and the `###` form the instruction also named is no
longer invisible at all.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01X3L79vgAnt1x2kxvf23Qt7

REVIEWED.md

@@ -2697,4 +2697,200 @@ Distance to threshold, 19.
 
 **If AUTHORIZED:** apply the suspension; nothing else in the counting path is
 touched. Report N-now with its real/mumble split at the next wake. Set
-2026-10-15 as a return-for-ruling, not an expiry.
+2026-10-15 as a return-for-ruling, not an expiry.
+
+## REVIEWED-135 — PENDING-133 — The fr dispositioning pass authorized, and three stale records corrected by dated act
+**Date:** 2026-09-05
+**Decision:** AUTHORIZED — PENDING-133 AMENDMENT 1's corrected scope, with five conditions.
+**Rules on:** PENDING-133 AMENDMENT 1. PENDING-133's parent block is not ruled — it was
+withdrawn by its proposer on 2026-08-10 and is recorded, not decided, at condition 6.
+
+**Notes:** What is authorized here is the running of a pass, not a doctrine. The rule the pass
+applies is PENDING-134's whose-proposition test, already ruled; the pass's scope correction was
+accepted at REVIEWED-116 point 6. Neither is reopened. What had never been authorized is the
+pass itself, and it has been the sole obstacle to `ratio_A_to_B` since 2026-08-10.
+
+⚠ **This ruling's header names an id, and PENDING-145's mechanism will therefore suppress every
+block under PENDING-133 — the withdrawn parent and the amendment both.** That is the correct
+outcome here only because both blocks close together. It is not a precedent: the same header
+over a live sibling block would hide it, which is what PENDING-145 and PENDING-146 exist to fix.
+
+1. **The pass is authorized as scoped by AMENDMENT 1**: every fr grounded span read for
+   reported speech, then dispositioned where present under PENDING-134's test.
+
+2. **CONDITION — the denominator is established, not inherited.** AMENDMENT 1's own finding is
+   that the earlier bound trusted P7's marker set and that nothing had checked it. The pass
+   therefore enumerates the fr grounded spans from the corpus itself and **reports the number it
+   found**, which is currently on the record only as 'nine or ten'. A pass that inherits a prior
+   pass's set reproduces the defect that occasioned it.
+   ⚠ The 21 unmarked inline guillemet spans are AMENDMENT 1's *evidence* that marking passes
+   miss inline cases. They are not the scope. Do not silently widen to them; if the enumeration
+   reaches them by its own criterion, say so.
+
+3. **CONDITION — the narrow ratification holds.** PENDING-134's H2 ratified the nested-voice
+   case and expressly refused the general principle that a marker names a span property while
+   the claim fixes the disposition. The pass touches F4 dispositions only. It does not
+   re-disposition F3, F5, F7, F8 or F10, and it proposes nothing about them.
+
+4. **CONDITION — the defeater is recorded, not resolved.** PENDING-134 (b) names its falsifier:
+   a span on which the test yields a disposition two independent readers both reject on reading.
+   If the pass meets one, it **records it and completes**; it does not adjudicate, and it does
+   not adjust the test to accommodate it. A doctrine that absorbs its own counterexample during
+   the pass that was to test it has no defeater at all.
+
+5. **CONDITION — `ratio_A_to_B` is a separate act and is not authorized here.** REVIEWED-116
+   point 5 permits one re-derivation. The pass makes it available; spending it is a further
+   steward act, taken against a completed and reported disposition set. One-shot instruments are
+   not spent in the same breath as the work that unblocks them.
+
+6. **THE DATED CORRECTION OF THREE PLACED RECORDS.** REVIEWED-132 §6 reserves normalization of
+   a placed record to a dated steward act stating what changed and why. This is that act.
+   - **PENDING-133 (parent).** `Awaiting: Steward authorization` → `WITHDRAWN by the proposer
+     2026-08-10; no ask outstanding. See PENDING-134.` *Why:* the withdrawal is recorded in the
+     body's Status line and the Awaiting line was never updated, so every reader has shown a
+     withdrawn proposal as awaiting the steward for four weeks — and it was nearly authorized
+     as such on 2026-09-05.
+   - **PENDING-133 AMENDMENT 1.** Awaiting → discharged by this ruling.
+   - **PENDING-134.** `Awaiting: Steward authorization, after the MCP restart…` → the ruling
+     that disposes of it, plus `disclosure placed 2026-09-05 at 9221dd8`. *Why:* H1's condition
+     was satisfied by REVIEWED-117 and the record never said so. **Deferred to condition 7 —
+     do not write this line until the ruling id is known.**
+
+7. **PRECONDITION, and the pass does not run before it is answered.** Name the `REVIEWED-N` that
+   disposes of PENDING-134. It is in `PENDING-archive.md` on a 56/56 raw-header verification,
+   but if the matching header is REVIEWED-116's slash-joined form, the widened reader
+   deliberately declines to read it and PENDING-134 is archived on a match no instrument here
+   will make. **One line, and it either closes this or reopens PENDING-134.**
+
+8. **Scope, and what is testimony.** PENDING-133 and its AMENDMENT 1, PENDING-134, REVIEWED-133
+   and PENDING-137 read verbatim this session; the amendment reached only by `governance_read`
+   at a line number supplied by `governance_search`, since `governance_item` returns the parent
+   block alone — PENDING-175, live in the moment it mattered. REVIEWED-116 and REVIEWED-121 are
+   read **only as quoted** in the items above; no independent read. The corpus figures, the
+   commit ids, and the claim that act three is placed are executor testimony. This ruling does
+   not depend on the corpus figures — it depends on the pass never having run, which AMENDMENT 1
+   and today's measurement both state.
+
+**If AUTHORIZED:** answer condition 7 first. Then run the pass under conditions 2–4 and report
+the enumeration with its number. Then apply the three record corrections at condition 6 as one
+dated edit citing this ruling. `ratio_A_to_B` returns for its own act.
+
+## REVIEWED-135 — AMENDMENT 1 — §8's stated dependency was false; conditions 1–4 were satisfied before the ruling was written
+**Date:** 2026-09-05
+**Amends:** REVIEWED-135, which stands unedited. This entry is joined to it, not placed over it.
+
+**THE FALSE CLAIM, VERBATIM, AND WHAT IT WAS.** REVIEWED-135 §8 closed: *'This ruling does not
+depend on the corpus figures — it depends on the pass never having run, which AMENDMENT 1 and
+today's measurement both state.'*
+
+PENDING-133 AMENDMENT 1 states no such thing. It corrects a scope and records an outstanding
+authorization; it says nothing about whether any pass had run. The claim had **one** source —
+the executor's report of the same day — and was written as two, one of them a document.
+
+⚠ **The failure mode, named because the writing conceals it.** A relay was strengthened into
+corroboration. Two independent statements is a stronger warrant than one, and the sentence
+therefore reads as *more* diligent than an honest single citation would have — so nothing in the
+prose signals the defect. This occurred in the sentence whose sole function is to declare what
+the ruling rests on. It is the jurist's instance of the day's third absence-read: the executor
+inferred 'never run' from a missing marker, and this ruling converted that inference into a
+standing premise.
+
+**WHAT ACTUALLY SATISFIED CONDITIONS 1–4.** The pass authorized by REVIEWED-116 point 5 — *'an
+identification pass that writes nothing, marking after the vocabulary is ruled'* — ran on
+2026-08-13 (`docs/fr-reported-speech-identification-pass-2026-08-13.md`, `de1c34b`). A pass
+authorized to write nothing leaves no trace; the absence was what completion predicted.
+
+It was replicated **blind** on 2026-09-05: all eight grounded spans read independently before the
+2026-08-13 document was found. That makes it an unplanned replication rather than a re-reading of
+a known answer, and it is the stronger evidence of the two.
+
+1. **Condition 2 — satisfied, and it did the work it was written for.** The denominator is **8**,
+   enumerated from the corpus, not the 'nine or ten' carried on the record. The earlier pass read
+   nine because L1551 was still grounded when it ran. Result: zero reported speech in the eight;
+   every guillemet is coinage or technical vocabulary, self-marked as such at `« totaux »` and
+   `« potlatch »`; the `dit` at L1549 is adjectival.
+
+2. **Condition 2 — the finding it produced, which the surrounding items had leaned against.**
+   Within the grounded set, P7's F4 tagging was **correct and complete**: F4 applied to exactly
+   the two spans carrying reported speech and to no others. AMENDMENT 1 was right to refuse to
+   inherit P7's marker set and right about the principle; the check vindicated the pass it
+   doubted. Both facts belong on the record.
+
+3. **The positive control, and its limit.** L926 and L1551 both flag; the grounded maximum quoted
+   run is 22 characters against 1144 and 234. Detection is demonstrated.
+   ⚠ **Exhaustiveness is not.** L926 flags on run length alone — the attributing-verb channel
+   returns nothing on it — so the two channels are not equally sensitive, and a span carrying
+   reported speech with neither a listed verb nor a long run would be missed. Recorded, not
+   repaired.
+   ⚠ **Why the zero survives that limit:** it does not rest on the reader alone. All eight spans
+   were read directly, and a direct read is not channel-limited. The instrument's gap bounds what
+   the *scan* proves, not what the *pass* found.
+   ⚠ The length separation is corroboration only. The 2026-08-13 pass refused to promote it —
+   *'a correlation observed on eleven spans in one author, NOT a rule'* — and Arendt and Weil
+   remain unexamined. This amendment does not promote it either.
+
+4. **Conditions 3 and 4 — satisfied vacuously, and the vacuity is the finding.** No F4 remained
+   to disposition, so nothing touched F3, F5, F7, F8 or F10. No defeater candidate arose, and
+   none could have: `reader_disposition_pre_doctrine` is not populable retroactively on inherited
+   spans.
+
+5. **Condition 5 — untouched and unamended.** It was written to survive exactly this: the reason
+   to hold a one-shot instrument is independent of why the work in front of it was delayed.
+   `ratio_A_to_B` remains VOID and unspent, and its spending is a separate steward act.
+
+6. **Condition 6 — proceeds.** The three record corrections are independently true and nothing in
+   them turned on when the pass ran. Condition 7's precondition is discharged. PENDING-134's
+   corrected line records `REVIEWED-116 point 5` and the disclosure placed 2026-09-05 at
+   `9221dd8`.
+
+7. **NEW CONDITION — the standing disclosure overstates this amendment's effect and is corrected
+   with it.** PENDING-134's whose-proposition doctrine **governs zero spans and has never fired**:
+   L926 was retracted at REVIEWED-118 and L1551 reclassified at REVIEWED-119, both by independent
+   routes, and the remaining eight carry nothing for it to reach. The set it would have governed
+   was already empty when it landed.
+   The post-hoc-amendment concern is that membership shifted after the spans were read, in a
+   direction that made a bar passable. For **this** amendment it shifted nothing. The standing
+   note on future recall figures implies otherwise and must say so: the F4 amendment had no
+   effect on membership; the **cell-constant** amendment is the one carrying a direction, and
+   REVIEWED-133 condition 2 already records that direction as inference rather than measurement.
+   Distinguish the two rather than reporting 'amended post-hoc on 2026-08-10' undifferentiated.
+
+8. **NEW CONDITION — L934 is recorded as a disagreement, not resolved into concurrence.** Two
+   readers differed on whether `« foyer d'origine »` — *'ce que Hertz appelait…'* — is an
+   attributing cue; the second deferred to the first on re-reading, that attributing a coinage is
+   not attributing an utterance. This is **not** PENDING-134 (b)'s defeater, which requires both
+   readers rejecting. It is the **only exercise the doctrine has ever had**, and it is worth more
+   on the record as a near-instance than as agreement. File it against (b) as the sole datum.
+
+9. **Scope, and what is testimony.** PENDING-133 and its AMENDMENT 1, PENDING-134, REVIEWED-133,
+   REVIEWED-135 and PENDING-137 read verbatim. REVIEWED-116, REVIEWED-118, REVIEWED-119 and
+   REVIEWED-121 are read **only as quoted** in the items above — including REVIEWED-116 point 5's
+   authorizing language and the two retraction routes. No independent read. The corpus figures,
+   the character counts, the commit ids, the 2026-08-13 document's existence and contents, and
+   the regex findings in the placement note are executor testimony. ⚠ The correction in this
+   amendment was supplied by the party the original ruling ruled on, which is the configuration
+   working rather than a defect — but it is not independent verification, and the record should
+   not read as though it were. ⚠ The result is also a pleasant one: the F4 amendment turning out
+   to have moved nothing makes the fr disclosure more defensible than it was. Pleasantness is not
+   evidence, and it is the condition under which a record is waved through.
+
+**PLACEMENT.** Joined **beneath** REVIEWED-135, which is not edited. §8's false sentence stays
+where it stands, superseded in place and visible.
+
+Header form is `## REVIEWED-135 — AMENDMENT 1`: it classifies as id+marker with ident
+`REVIEWED-135`, verified 2026-09-05. ⚠ PENDING-139's leg (A) is cited nowhere here —
+`RE_HEAD_LINE` now matches `^#{2,3}` and the `###` form is no longer invisible, so that
+measurement is stale and the item needs re-measuring before it is ruled. The `·` form remains
+broken and worse than recorded: it parses to compound with the identifier mangled to a bare
+`REVIEWED`. ⚠ The drift check's report that REVIEWED-131's amendment replaced its record is a
+**false positive** — `RE_ID`'s greedy group backtracks onto the hyphen inside `REVIEWED-131`, so
+the un-amended original is never found. The record is intact and the reader cannot see it. This
+is the parenthetical header form failing in a second instrument, the same shape that broke
+`ruled_pendings` the same morning, both toward hiding a record that exists. REVIEWED-135's parent
+uses the em-dash form and parses to `REVIEWED-135`, so placing beneath it produces no second
+instance.
+
+**AFTER PLACEMENT:** condition 6's three corrections as one dated edit citing REVIEWED-135 as
+amended. Then the standing-disclosure correction at point 7. Then the L934 near-instance filed
+against PENDING-134 (b) as its sole datum. `ratio_A_to_B` returns for its own act under
+condition 5.