Place REVIEWED-138 and REVIEWED-139 — the count in both units, and the errata

REVIEWED-138 (PENDING-168) settles the doctrine's count by declaring its UNIT
before its number, and declaring both: four named instances, seven occurrences —
or eleven under a per-control reading of instance four. The steward counts the
occurrence the executor declined; the byline records which hand did which, since
the party that declined it is the party that drafted the entry counting it.
PENDING-180 is discharged. The structural remedy stays open.

REVIEWED-139 records two line citations in placed rulings that no longer hold —
wrap.py:236 -> :256, moved by 7948c09 today; mumble.py:123 -> :139, moved by
3d45e3a — the second of which was stale on the day it was written, invalidated by
the very commit its ruling was ruling on. Recorded by joining, never by editing.

Placed by the steward; committed by the executor, which is this file's standing
practice (six prior placements carry the same trailer). Content untouched.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01T1i5VRfHjD79hfaXjWsBXA

REVIEWED.md

@@ -3358,4 +3358,95 @@ caught by a second instrument, not by review.**
 5. ⚠ **The assertion-versus-file gap** (§6). `source_lacks`/`source_has` assemble the **assertion's** needle from parts; nothing stops a comment elsewhere in the file planting a contiguous copy. Demonstrated in this sitting, caught only by mutation. **The repaired mechanism is not yet whole**, and §6 declares it verified.
 6. **The render-volume hypothesis**, first to test, plausibility flagged as the hazard.
 
-**For a reader in a month, two things.** **The pile-up was never evidence about the cap** — it was evidence about a condition nobody has named. **And the field that would have survived deletion is one opaque integer**, content-free by construction, which did not exist because nobody had asked the question this sitting's instrument just asked.
+**For a reader in a month, two things.** **The pile-up was never evidence about the cap** — it was evidence about a condition nobody has named. **And the field that would have survived deletion is one opaque integer**, content-free by construction, which did not exist because nobody had asked the question this sitting's instrument just asked.
+
+## REVIEWED-138 — PENDING-168 — The count in both units: four instances, seven occurrences (or eleven)
+**Date:** 2026-09-10
+**Decision:** AUTHORIZED as to the count. **The unit is declared before the number, and both units are declared.** PENDING-168's doctrine stands at the
+strength stated.
+
+**Ruled by:** steward — the unit is the steward's per REVIEWED-137 §7.
+**Drafted by:** executor, at the steward's instruction. The count of instance two's fourth occurrence is the steward's act; the executor declined it and
+says so below.
+**Authorized by:** steward, 2026-09-10.
+
+**THE UNIT, FIRST.** The count is stated in two units because the doctrine makes two claims with different denominators, and one number cannot carry both.
+Neither figure is to be quoted without its unit.
+
+- **Named instances: FOUR.** Unit = a distinct governed condition that relied on remembering and was broken. This is the denominator for the claim *a
+condition that can only be honoured by remembering will be broken*.
+- **Occurrences: SEVEN — or ELEVEN under a per-control reading of instance four.** Unit = a single breach event. This is the denominator for the claim that
+the breach recurs, including under active watch.
+
+**The seven, composed, so the arithmetic is auditable rather than asserted:**
+
+| instance | occurrences |
+|---|---|
+| the staleness case written inside the section naming the staleness pattern | 1 |
+| the self-planted needle | **4** |
+| REVIEWED-128 condition 3 read for content, 2 h 33 min 24 s after receipt | 1 |
+| five hook-allowlist controls registered after the tally that would have reported their failure | 1 — **or 5** |
+| **total** | **7 — or 11** |
+
+The four occurrences of instance two: written twice while watching for it (2026-08-27); `tarbuckle-seam.py`'s positive-form control, filed as PENDING-180;
+and the comment in PENDING-180 ADDENDUM 1's own repair, which re-planted the needle inside the sentence explaining it. **The executor declined to count the
+fourth, correctly — the drafting hand choosing the unit that suits its number is the move the deferral exists to block. The steward counts it.** It is the
+third written by a party actively watching for the shape, and on the doctrine's own terms it is the strongest datum the item holds.
+
+⚠ **ONE UNIT IS INHERITED, NOT DERIVED, AND IS MARKED AS SUCH.** Instance four is counted as **one** occurrence though it names five controls. That reading
+comes from PENDING-180's stated base of five occurrences and is adopted here without re-derivation. **This is why the figure is written as "seven, or eleven
+under a per-control reading" wherever it appears** — the range is inseparable from the number, because a bare "seven" will be quoted and a caveat will not
+travel with it. **The named-instance figure is four under either reading.**
+
+⚠ **THE DETECTOR'S UNIT IS NOT THE CLAIM'S UNIT, and this entry is the occasion to say so.** The deferred decision `pending-168-count-unit-declared` fires
+on the literal `— PENDING-168 —` appearing in REVIEWED.md. That detects **that a ruling was placed on PENDING-168** — not **that the count unit was
+declared**. It is correct today only because this entry does both. Since PENDING-168 stays open on the remedy, a later ruling on that remedy would fire the
+same needle and report a declaration that had not occurred. **The trigger is a header string standing in for a declaration; it is the same substitution the
+item counts.**
+
+**Two corrections land with this ruling and are not severable from it:**
+1. PENDING-168's summary states *"care failed in seven hours."* The interval, measured from receipt, is **2 h 33 min 24 s** (REVIEWED-136). The doctrine's
+*"hours, not weeks"* **tightens**; nothing in it weakens.
+2. PENDING-180 and its ADDENDUM 1 are part of the evidence base being ruled on, not later commentary on it.
+**PENDING-180 is DISCHARGED.** Its (c) census → (b) `source_has()` → (a) repair were executed in that order and the repair is mutation-verified. Its
+deferred decision `pending-168-count-unit-declared` is satisfied by this entry.
+
+**WHAT THIS RULING DOES NOT SETTLE, stated so it is not read as broader than it is.** PENDING-168's structural remedy — **(a) write-only until a date** or
+**(b) encrypted at rest to a steward-held key** — is **NOT ruled here and remains open**, as does the question of whether it should be built by the party it
+constrains. This entry settles the count and its unit. **The item stays open on the remedy.**
+
+**If AUTHORIZED:** Proceed. Tag any commit touching the count with REVIEWED-138. A future citation of this doctrine's evidence base must carry its unit
+**and its range**; a count stated without them is not to be read as established.
+
+The errata entry. Filed standalone rather than as an amendment to either ruling: it corrects citations in two entries, so keying it to one misfiles the
+other. Re-key it if you'd rather it hang off 136.
+
+## REVIEWED-139 — ERRATA — Two line citations in placed rulings do not hold, and one never did
+**Date:** 2026-09-10
+**Decision:** RECORDED. The citations below are corrected **by this entry joining the record, not by editing the entries that carry them.** No in-place
+revision has been made to REVIEWED-136 or REVIEWED-137.
+
+**Ruled by:** jurist (Claude.app), on a trace run by the executor with explicit argv and reported with its failure mode.
+**Authorized by:** steward, 2026-09-10.
+
+- **REVIEWED.md:2993** (REVIEWED-136) cites `tarbuckle-wrap.py:236`
+    — the control asserting 12 words are rejected.
+    It is at **`:256`**. Moved by **`7948c09`**, 2026-09-10, by the executor.
+- **REVIEWED.md:3104** (REVIEWED-137) cites `tarbuckle-mumble.py:123`
+    — `log_rejection()` opening the log in append mode.
+    It is at **`:139`**. Moved by **`3d45e3a`**, 2026-09-09.
+
+⚠ **THE SECOND CITATION WAS STALE ON THE DAY IT WAS WRITTEN.** The append-mode open sat at `:123` from `b95ee73` (2026-08-25) through `97a0cb3`, and moved
+to `:139` at **`3d45e3a`** — the condition-G commit REVIEWED-136 itself ordered. REVIEWED-137 was placed at `3d340f6`, **after** it. **A ruling cited a line
+invalidated by the very commit it was ruling on, the same evening.** This is not decay over time; it is a citation that never held.
+
+**Why this is recorded rather than repaired.** A placed ruling is amended by a recorded act. Correcting a number inside a placed entry would leave no trace
+that the entry had ever been wrong, and the register would then hold a citation that looks as though it had always been right — which is the failure the
+register-integrity check was built for (2026-08-07, REVIEWED-87 overwritten by its own amendment).
+
+**The general finding, carried to PENDING-183 item 11, which changes from observational to load-bearing:** the register cites code by line, lines move, and
+**nothing re-checks any citation.** Two placed rulings are wrong today; `tarbuckle-seam.py` moved twice in two days. ⚠ **Recommended and NOT authorized
+here: anchor register citations to symbols or content rather than line numbers** — the reasoning that makes `source_has()` assemble its needle from parts. A
+line number in a placed ruling is a claim with a half-life measured in days.
+
+**If AUTHORIZED:** Proceed. Future citations of code in this register carry a symbol or a quoted fragment; a bare `file:line` is to be read as provisional.