All three taken, plus the ontology ruling. Every defect you found was in B6's own emission layer, which is the right place for a new layer's defects to be.
Gate: grid v1.46.0 (93 entries / 383 citations) · negative battery 320/320 · film 45/45 · lowering 23/23 · emission 9/9 over 9 fixtures · bridge 48/48 · derive 45/45 · realm 24/24 · harness 14/14 · runner 3/3. cert_id a08ee15d… byte-identical — forty-first consecutive round.
Confirmed, including your ids:
add(const 2, const 3) ctmpl-d0105d4f…
add(input x, input y) {x:2,y:3} ctmpl-d0105d4f… SAME
x + (x + y) {x:2,y:3} ctmpl-efba3154… DIFFERENT2 + 3 == 3 + 2, so it is exactly the symmetric fixture INSTANTIATION_FALSIFIERS I-4c rejects in its own words — a test whose output cannot reveal the defect it is named for. I reported the collapse as the interesting finding and the collapse was an artifact of the mislabelling.
Your fix taken exactly: kept under the honest name two-port collapse, because what it actually proves is worth having, and the real I-4c added beside it — already a distinct closed template, so nothing was invented to force distinctness. Nine fixtures, eight distinct domain values, and E-9 now reports …, 3, 5, 7 — the 7 is the real I-4c computing correctly.
The surviving theorem is the one you named: emission cannot depend on provenance instantiation erased.
Confirmed, and it is the worse of the two reporting defects because it lands in RESULTS.txt where a reviewer who runs nothing else sees only that. Every quantity in the final line is now a field written by the case that measured it:
QUOTIENT-VISIBLE STRUCTURAL DRIFT … REFUSED on 9/9
ALPHA-RENAMING AND LABEL PERMUTATION LEAVE IT UNCHANGED on 9/9, as L-BYTES-1 requires
The structural alternate differs in id on 6/6 applicable and agrees in decoded outcome on 6/6No new harness species, as you said. The stale header prose is corrected too.
TARGET_ENCODING.dup_label_policy justified the allocation as semantic "because the label reaches the canonical signature". It does not, and B6's own battery measured that. An encoding identity cannot be justified by a difference its own codomain erases.
The encoding now commits to label equality and freshness structure — collapsing two distinct dups onto one label changes whether DUP-SUP= or DUP-SUP! fires, and that is semantic — and says nothing about the integers representing it.
What moved, and what did not:
MOVED TARGET_EXECUTABLE_ENCODING_SEM_ID xenc-7ce8f33f… → xenc-25464e50…
MOVED EMISSION_SEM_ID esem-aab30fb5… → esem-a88100f1…
SAME LOWERING_SEM_ID · INSTANTIATION_SEM_ID · TARGET_TEMPLATE_ENCODING_SEM_ID · DECODE_SEM_IDThe executable encoding's semantics changed and nothing else did — which is what makes this a correction rather than a re-cutting of the chain. B6's pair is kept in SUPERSEDED_LABEL_SEMANTICS_SEM_IDS.
I took your first design, not the second. No target_term_bytes_id in the receipt. Emission proves the semantic target term; exact bytes belong beside executable_artifact_id as provenance. The law now says so explicitly, including what the receipt does not prove.
And your ladder is in the law, replacing my pair:
exact emitted BYTES ──quotient alpha and label spelling──▶ target_term_sem_id
──execute, normalise, decode─────────▶ outcome_sem_idderivation.emission-conformance@2 supersedes @1, kept as history.
E-F1 uses an unconstrained quotient-visible drift — bump a church numeral — and therefore covers 9/9 with no applicability caveat. E-8 keeps the commutativity-dependent equivalence and covers the 6 add-bearing fixtures. No replacement equivalence invented before `sub` exists, as you ruled.
Two of my own new assertions matched substrings rather than the thing:
the M-10 wiring check counted mentions of phase_pin_lint.py, so true "$BASE/phase_pin_lint.py"
disarmed the guard and left it green;
the adversary-split check tested /driftEmit/, which notTheDriftEmitter satisfies.
Both were caught by the negative case written beside the assertion, on its first run. That is the argument for writing the forgery next to the claim rather than after it — but two instances in one session of the guard matches a substring of its own disarming is a shape, not an accident.
I am not proposing an M-11. You said stop at ten and I agree; I am recording it in the ledger so that if it recurs a third time there is a countable record rather than a memory. If you think that is already enough to act on, say so and I will bring it back.
C-side replay — verifier diversity / a small native proof consumer.
Canonical-locus alias precedence.
film-projection-not-unique has no direct negative fixture.
film-too-many-frames reached by no term tried; guard witnessed by a -DMAXFRAMES=4 build.
Source-refusal ↔ instantiation-refusal preservation.
E-8's equivalence witness when `sub` arrives — deferred by ruling, not forgotten.
b61-review.zip — verify.sh runs 26 checks and generates RESULTS.txt from that run.
unzip b61-review.zip && cd b61-review && ./verify.shRead: governance/round-11-ledger.md §339–346 · governance/emission_conformance.mjs (the ladder is in the header, the adversaries are driftEmit / equivEmit) · governance/lowering.mjs TARGET_ENCODING.label_semantics and SUPERSEDED_LABEL_SEMANTICS_SEM_IDS.