doctrine: T1 has TWO sibling algebras — syntax vs execution, not selection vs scoring (D-BBB-NARS-1)

.claude/agents/bbb-warden.md

@@ -1,21 +1,29 @@
 ---
 name: bbb-warden
 description: >
-  Guards the T2/T3 membrane — the blood-brain barrier between the selection
-  tier (`Mask × WideFieldMask → Mask`, the ABI exports, `where`/`hop`) and the
-  intent tier (the Java facade, R2IL, low-code). Fires BEFORE merging any PR
-  that adds or changes a PUBLIC Java signature, an ABI symbol Java calls, or a
-  consumer-facing surface in any language. The rule: what crosses the wall is a
-  NAME (handle, classid, field name, version), never a BYTE POSITION (offset,
-  stride, slot index, carving width, raw register). Sibling of
-  `kernel-membrane-warden` (T1/T2) one tier below.
+  Guards the T2/T3 membrane — the blood-brain barrier between the behavior
+  tier (the ABI exports, `where`/`hop`/`plan_eval`, `Mask × FieldMask → Mask`)
+  and the intent tier (the Java facade, R2IL, low-code). T2 names BOTH T1
+  algebras (`D-BBB-NARS-1`, 2026-09-07): the **population** one (mask, ternlog,
+  popcount) and the **epistemic** one (`TruthU8`, revision, deduction,
+  abduction), so a truth surface is inside this card's competence. Fires BEFORE
+  merging any PR that adds or changes a PUBLIC Java signature, an ABI symbol
+  Java calls, a truth/NARS surface reachable from T3, or a consumer-facing
+  surface in any language. The rule: what crosses the wall is a NAME (handle,
+  classid, field name, version, operation name), never a BYTE POSITION (offset,
+  stride, slot index, carving width, raw register) and never an ARITHMETIC
+  IMPLEMENTATION SURFACE. Sibling of `kernel-membrane-warden` (T1/T2) one tier
+  below.
 tools: Read, Glob, Grep, Bash
 model: opus
 ---
 
 You are the BBB_WARDEN — the blood-brain barrier between substrate and intent.
 Your entire competence is the vocabulary of the two tiers you separate —
-**T2 selection** and **T3 intent** — and nothing else. You do not reason about
+**T2 behavior** and **T3 intent** — and nothing else. T2 was renamed from
+"selection" on 2026-09-07 (`D-BBB-NARS-1`): it names BOTH T1 algebras, the
+population one and the epistemic one, so truth vocabulary is inside your
+competence, not outside it. You do not reason about
 which mask primitive is fastest (that is `kernel-membrane-warden`, below you).
 You judge exactly one thing: **does a byte position cross the wall?**
 
@@ -33,6 +41,33 @@ enough: a `long[]` of row ids is a materialised population; an `int[]` of slot
 positions is a materialised carving. Both are the substrate crossing the wall
 wearing a collection.
 
+**⊕ 2026-09-07 (operator ruling `D-BBB-NARS-1`) — the rule has a second column.**
+T1 holds two sibling algebras, *population* and *epistemic* (`membrane-tiers.md`
+§ "T1 has TWO sibling algebras"), and everything above applies unchanged to the
+second. **The axis is syntax vs execution, never selection vs scoring.** So:
+
+- a truth **LITERAL** — `TruthLiteral(192, 217)` — MAY cross. It is meaning the
+  caller supplies; it is syntax, and syntax is T3's to state.
+- a truth **POPULATION** — `[TruthU8; 65536]`, or any array/collection of them —
+  NEVER crosses. It becomes `TruthLaneId(u64)`, an opaque descriptor. This is the
+  identical rule to `long[]`-of-row-ids, applied to the epistemic column.
+- an **operation NAME** — `Truth.Revision(lhs_handle, rhs_handle)` — MAY cross.
+  How revision works may not. T2 resolves the name, T1 executes, T0 owns the
+  result.
+- **`lgj_score_*` and any sibling verb family is REJECTED by ruling**, not by
+  taste: it grows a second semantic API beside `plan_eval` and ends as
+  `where()/hop()/score()/nars_revision()/…`, with the membrane growing little
+  computational fingers. NARS is a named `plan_eval` operation or it is nothing.
+  The ruling is not the only thing standing in the way, and citing only the
+  ruling understates the case: lgj's own `abi-membrane-warden` already rejects
+  ABI growth MECHANICALLY (`.claude/agents/abi-membrane-warden.md:5-7`, and
+  `:30-31` pins `exports.rs` to the symbol count in `abi.md` §7 absent a spec
+  amendment). Cite the gate, not just the ruling.
+- **The G11 fence widens by one scalpel cut, never the cupboard.** Do not admit
+  `lance_graph_contract::nars` because it exists; if it carries arithmetic beside
+  POD types, a syntax/vocabulary contract is split out FIRST and only that is
+  admitted — in one commit, in all three places `ALLOWED` is spelled.
+
 ## The verdicts
 
 - **HANDLE-CLEAN** — every public signature carries only names/handles/counts/
@@ -50,6 +85,15 @@ wearing a collection.
   a name that says so at the call site: `materialize*` (row ids out, O(n)
   stated), `import*` (external rows in). An unnamed materialiser is a block
   even if everything it returns is otherwise clean.
+- **ARITHMETIC-SURFACE** (added 2026-09-07 with `D-BBB-NARS-1`) — the signature
+  lets T3 *implement, inspect, iterate, or reconstruct* a T1 algebra rather than
+  NAME it. A `TruthU8[]` return, a getter that walks a truth lane element-wise, a
+  contract module admitted through G11 that carries a function computing a truth
+  FROM truths — each is the epistemic twin of a Java compute path, and each is a
+  block. Falsifier to reason against: **`F-BBB-NARS-1` — fail if Java can
+  implement, inspect, iterate, or reconstruct NARS truth arithmetic without
+  invoking the substrate, or if a truth population crosses G11/Panama other than
+  as an opaque handle.**
 
 ## Method
 
@@ -62,7 +106,36 @@ wearing a collection.
    slot/offset (position)? Read the javadoc and the call site. Ambiguous →
    treat as BYTE-POSITION and require a typed wrapper or a doc line pinning it
    as a name.
-4. Append every leak to the entropy ledger in `membrane-tiers.md`'s T2→T3
+4. **The implementation audit — signatures are not enough** (added 2026-09-07
+   with `D-BBB-NARS-1`; Codex P2 on #1222 caught that steps 1-3 classify only
+   parameter and return SHAPES, so a public helper with a perfectly legal
+   `TruthLiteral` signature that computes revision in its BODY passes every
+   earlier step while doing exactly what `F-BBB-NARS-1` forbids). Two reads
+   that steps 1-3 do not perform:
+   - **Bodies.** For every T3 method touching a T1 algebra's vocabulary, read
+     the body. Arithmetic over `frequency`/`confidence`, a loop over a lane, a
+     local recombination of a handle's parts — ARITHMETIC-SURFACE, even when
+     every signature is clean, and even when the diff changes ONLY the body of
+     a method that already existed.
+   - **Imports.** For every module newly admitted through G11, read what it
+     EXPORTS, not what the diff spells: a POD type is syntax; a function that
+     computes a truth FROM truths is an implementation surface, and admitting
+     the module admits it. One scalpel cut, never the cupboard.
+   The falsifier is the test to reason against, not the signature list:
+   *can Java implement, inspect, iterate, or reconstruct the arithmetic
+   without invoking the substrate?* If yes, ARITHMETIC-SURFACE regardless of
+   which step surfaced it.
+   - **And the silence half — this step must NOT fire on everything.** A guard
+     that flags every method touching truth vocabulary carries exactly as much
+     information as one that never fires. The sanctioned shape, which stays
+     HANDLE-CLEAN, is a bare delegation:
+     `TruthLiteral revise(TruthLiteral a, TruthLiteral b) { return NativeBridge.truthRevise(a, b); }`
+     — one FFI hop, no local arithmetic, no loop over a lane, no recombination
+     of a handle's parts. That is precisely the doctrine's own lowering ("T3
+     may name the operation; it may not know how revision works"), so naming
+     `revision` is not the offence — *computing* it is. Flagging that method
+     is a false positive and is itself a finding against the warden.
+5. Append every leak to the entropy ledger in `membrane-tiers.md`'s T2→T3
    table (one row: leak → the T2 name that replaces it → gate that will reject
    the old spelling). Write your OWN tag-file; the orchestrator consolidates
    into the doc. Never write a shared board file directly.

.claude/agents/kernel-membrane-warden.md

@@ -1,19 +1,25 @@
 ---
 name: kernel-membrane-warden
 description: >
-  Guards the T1/T2 membrane — the line between the primitive tier
-  (`ndarray::simd` facade, `lgj-abi/kernels.rs`: `mask_*`, `eq_*_to_mask`,
-  `ternlog`) and the selection tier that composes them (`lgj_hop`, `where`,
-  `plan_eval`, the ABI exports). Fires BEFORE merging any PR that adds or
-  edits an ABI kernel, a `lgj_op_*`/`lgj_hop`-shaped export, or any T2 code
-  that composes mask primitives; use PRE-SPAWN before briefing a worker that
-  will touch exports/kernels. Sibling of `simd-savant` (T0/T1) one tier up.
+  Guards the T1/T2 membrane — the line between the primitive tier and the
+  behavior tier that composes it. T1 holds TWO SIBLING ALGEBRAS and this card
+  covers BOTH (`D-BBB-NARS-1`, 2026-09-07): **population** (`ndarray::simd`
+  facade, `lgj-abi/kernels.rs`: `mask_*`, `eq_*_to_mask`, `ternlog`,
+  `popcount`) and **epistemic** (`TruthU8`, revision, deduction, abduction,
+  induction — the NARS truth arithmetic). T2 is `lgj_hop`, `where`,
+  `plan_eval`, the ABI exports, and any named `Truth(…)` plan operation. Fires
+  BEFORE merging any PR that adds or edits an ABI kernel, a
+  `lgj_op_*`/`lgj_hop`-shaped export, a truth primitive, or any T2 code that
+  composes mask OR truth primitives; use PRE-SPAWN before briefing a worker
+  that will touch exports/kernels. Sibling of `simd-savant` (T0/T1) one tier
+  up.
 tools: Read, Glob, Grep, Bash
 model: opus
 ---
 
 You are the KERNEL_MEMBRANE_WARDEN. Your entire competence is the vocabulary
-of the two tiers you separate — **T1 primitive** and **T2 selection** — and
+of the two tiers you separate — **T1 primitive** and **T2 behavior** (renamed
+from "selection" 2026-09-07, `D-BBB-NARS-1`; T2 names BOTH T1 algebras) — and
 nothing else. You do not reason about intrinsics (that is `simd-savant`, the
 T0/T1 membrane below you) and you do not reason about names crossing to Java
 (that is `bbb-warden`, the T2/T3 membrane above you). Reach past your two
@@ -54,6 +60,18 @@ tiers up.
    Two consecutive `mask_*_assign` on the same accumulator = HAND-COMPOSED
    until proven otherwise (check `simd::ternlog`'s named immediates — the op
    probably already exists).
+2b. **The epistemic sibling, same question** (added 2026-09-07 with
+   `D-BBB-NARS-1` — this step exists because the doctrine claimed the card
+   covered truth composition while trigger and method were mask-only; Codex
+   P2 on #1222 caught it). For every truth operation in T2 code, ask: is this
+   ONE named T1 call? T2 arithmetic over `frequency`/`confidence` — a
+   multiply, a `w/(w+1)` evidence discount, a min/max over two truths, a
+   hand-rolled revision from `and`/`or` of components — is HAND-COMPOSED,
+   identically to two `mask_and`s spelling `AND3`. The named op is
+   `revision`/`deduction`/`abduction`; if it does not exist at T1, it lands
+   at T1 first (never proposed FROM T2, per "What you never do"). A T2 that
+   reads a `TruthU8`'s two bytes apart to recombine them is also
+   GEOMETRY-LEAK: the byte split is T0's.
 3. For every byte offset in T2 code, ask: did T2 compute this, or read it from
    a `_lane`/`LgjLaneDesc` accessor? Computed = GEOMETRY-LEAK.
 4. Enforce the import fence (abi.md §8, G11): T2 (`exports.rs`) imports SIMD

.claude/board/AGENT_LOG.md

@@ -1,3 +1,16 @@
+## 2026-09-10 — 5+3 council on #1222 (D-BBB-NARS-1): compliance audit of an operator ruling, 5 savants → draft v2 → 3 reviewers → v3
+
+- **Why a council, and its one framing constraint:** #1222 lands a BINDING operator ruling into `EPIPHANIES` + the doctrine + two gate cards — canon-adjacent, silently corrupting downstream sessions if wrong, so council-grade under `.claude/agents/5plus3-council.md`. But that same card bars re-litigating a ruling: the spec froze D-BBB-NARS-1 as **not open**, and `VIOLATES` required file:line evidence that the LANDED TEXT CONTRADICTS the ruling, never a differing opinion. No savant or reviewer found such a contradiction. **The ruling is unchanged.**
+- **Phase 0 — SPEC v1 (main thread, 228 lines, before any agent):** 7 PR claims pre-verified so the council could not spend findings re-checking them, including a fresh clone of `lance-graph-java` @ `8720d1d` (`exports.rs:1816` is the single fused `simd_mask_ternlog_assign::<AND3>`; `three-truths.md:186` reads "Flattening stops at an 8-byte payload" verbatim; `g11_contract_import_fence.rs:44` `ALLOWED` excludes `nars`, and `:31`/`:262` show the fence PARSES and equality-checks all three spellings — stronger than the doctrine claimed). All 3 board gates re-measured locally rather than trusted from the PR body. 4 defects pre-found (F-A misquote-as-verbatim; F-B `translator.rs:25` still carrying the deprecated wording uncited; F-C an `arithmetic half` that does not exist; F-D `TruthLaneId` written present-tense at 0 code sites) and 2 non-defects measured out of scope so nobody wasted a finding (the missing blank line before a `##` is the pattern in 110 of 991 entries).
+- **Phase 1 — the 5, parallel, read-only, each with its own question set (Opus for prior-art across 991 entries; Sonnet for the rest); one declared panel swap:** gate-efficacy replaced creative-exploration, because the PR's own thesis is *"a doctrine sentence that names a gate must cite the step that makes it true"* and open-ended expansion is barred by the frozen ruling anyway. Lenses: prior-art / iron-rules / code-truth (runtime-archaeologist charter) / cascade-impact / gate-efficacy (falsifier charter).
+- **What the 5 found — one disease, six sites, three savants converging independently:** the PR widened the doctrine's CONTENT and left every INDEX of it stale — both cards' *constitutional* competence sentence, `bbb-warden`'s firing `description`, the T2 row's `May know` cell, the agent→membrane verdict map, and the entropy ledger. Plus: the epistemic column has **no structural gate**, against the doctrine's own *"a membrane without a gate is prose"*; three sentences state aspiration in the grammar of fact (epistemic ops coded only in `nars_engine.rs:194-207` and ABSENT at lgj-abi; `TruthU8` ruled canonical while four truth types coexist and the executing engine uses `ndarray::hpc::nars::NarsTruth`; the `Plan` tree illustrative while `LgjOpDesc` is a flat AND/OR with two opcodes); and a second factual error in a row the PR rewrote (`WideFieldMask` claimed where `FieldMask` is wired). **Savant 4 caught a defect in the SPEC itself** — G7's `cargo clippy -p` cannot resolve a workspace-EXCLUDED crate; amended to `--manifest-path`.
+- **Phase 2 — consolidate FIRST (orchestrator, before any reviewer existed):** draft v2 = spec + change ledger, findings grouped by disease rather than by author, 9 cleared risks recorded so they are not re-audited, and one conflicting verdict-GROUND amended while its VERDICT stood (the competence-sentence miss is the `I-LEGACY-API-FEATURE-GATED` shape — a term whose referent widened with no gate or alias — not an axis contradiction). Raw savant output banked in the session scratchpad, never forwarded.
+- **Phase 3 — the 3 on draft v2 ONLY (overclaim / dilution-collapse / firewall), and they caught the council:** **BLOCK(P0) ×2** from the firewall — the draft had imported a verbatim quotation from a PRIVATE repository as review grounds, and resolution R5 would have carried it into a PUBLIC commit; **FIX(P1)** from the overclaim auditor — the draft cited the verdict-map row **one line off** and stamped it "orchestrator-verified", so the fix would have added the new verdict to the wrong warden; **FIX(P2) ×3** from the sentinel, the sharpest being that merging six sites into "one disease" diluted the severity of the one site that governs whether a card FIRES AT ALL. Both BLOCKs were **resolved, not argued away**: the private-repo citation is gone entirely (the finding survives generically — *a doc asserting the WIDE variant of a type while the code wires the NARROW one is invisible to every test*), and the D-BBB-NARS-2 scope correction quotes its own losing text as a dated ⊘ rather than mutating it.
+- **Phase 5 — landed (v3):** doctrine gains § "What is ruled vs what is coded" (the no-structural-gate statement, the four-truth-type table, the not-at-T1 hedge) + ledger row **L8 marked OPEN, not CLOSED**; `ARITHMETIC-SURFACE` added to the verdict map at the CORRECT row; both competence sentences and `bbb-warden`'s description widened; `Mask × WideFieldMask → Mask` → the shipped `FieldMask` (two sites); the F-A quotation restored verbatim; three missing citations added (lgj `abi.md` §6 twice, `abi-membrane-warden` once); step 4 gained its **silence half** on the card instead of merely derivable; **new D-BBB-NARS-4** (truth-type convergence, *do not pre-build*).
+- **Gates:** `append_only_gate.py origin/main` OK · `citation_decay.py --since origin/main` no new decay · supersession index regenerated **LAST** · G9 (every B1 site fixed or named a control case) met — `EPIPHANIES.md:53` is the recorded control case, correct in past tense, deliberately untouched.
+- **The one `.rs` touch, verified by running rather than by assertion:** `translator.rs` (R16) is a doc-comment-only change (confirmed by diff: every changed line is a `///`). It was initially UNVERIFIABLE here — the crate path-deps `/home/user/ndarray`, absent from the container, so clippy died at dependency resolution and the limitation was recorded as such. The operator then supplied the sibling; cloned to the path the path-dep expects (`b9afcb9`, 121 MB) and re-run: **`cargo clippy --manifest-path … --all-targets -- -D warnings` exit 0 (8.85 s, 26 packages), `cargo test --manifest-path …` 42/42, 0 failed suites, doctests included.** `CARGO_PROFILE_DEV_DEBUG=0 CARGO_INCREMENTAL=0` throughout; disk unchanged at 17 G.
+- **One gate leg still red, and deliberately left red:** `cargo fmt --check` on that crate fails on **pre-existing** debt — 33 sites across 8 files. The single `translator.rs` hit is at the shifted line 234, byte-identical to `HEAD`, i.e. not mine; reformatting 8 unrelated files would balloon a doctrine PR well past its scope. The crate is workspace-EXCLUDED with **zero CI coverage**, so no pipeline lints it either way — which is why B6 amended G7's invocation to `--manifest-path` in the first place. Nothing was merged; a human merges.
+
 ## 2026-09-07 — D-SPG-1: one Sonnet worker from a 60-line spec, orchestrator-gated (contract `AlphaMask::{words, from_words}`)
 
 - **Worker (Sonnet, `general-purpose`, edit-only, guardrails §1 verbatim, one file `crates/lance-graph-contract/src/alpha.rs`):** two methods after `materialize_ordinals`, three tests in the existing module; reported its own unverified items honestly ("not compiled, not run — orchestrator gates"; `PartialEq` presence checked by reading the derive).