design: triangle tenants (frozen/learned/explore) × one-register-two-readings × awareness-surface separation × chess quarantine

.claude/board/EPIPHANIES.md

@@ -1,3 +1,20 @@
+## 2026-07-17 — E-MASLOW-PYRAMID-OF-COGNITION-1 — observation/gestalt FIRST, then SPO 2³ decomposition → CausalEdge64 → NARS candidate design; the ladder type is shipped, the stateful ascent loop is the missing wiring (operator-directed)
+
+**Status:** RULING (design banked, PR #707 §3a) + AUDIT (fleet sweep, same day). Operator: "Maslow pyramid of cognition, observation/gestalt first, then SPO 2³ decomposition, CausalEdge64, design NARS candidates" — each level engages only when the level below fails to settle; the base is wide (the D-TRI-6 EXPECTATION is that most stimuli settle at resonance, O(1) palette lookup), the apex narrow (counterfactual design, fork+simulate). Pyramid width IS the cost amortization. Middle-tier emissions ARE NARS candidates (doctrine `mailbox-kanban-model.md:42`); the upper tier DESIGNS (revises) them, it does not introduce NARS — and the shipped driver runs that revision optionally-per-cycle where the pyramid gates it on non-settlement (a named behavior delta, D-TRI-6-gated).
+
+- **Shipped already:** contract `RungLevel` 0–9 + `RungElevator` (homeostatic: BLOCK streak ↑, FLOW streak ↓ to the dispatched base floor; System-1 felt hints + System-2 gate streaks drive ONE ladder); `pearl_level()` 0–2 observe / 3–5 intervene / 6–9 counterfactual; `causal_mask_bits()` O/PO/SPO (L2/L3 probe-certified, L1 CONVENTION); driver cascade already runs resonance [3] BEFORE the resonance-GATED ≤8 CausalEdge64 emissions [5]; `SpoDistances::all_projections` = 8 projections from 3 table reads (p3b/p6 green); `meaning_axes` volition damping (high rung deliberates, low rung acts).
+- **Missing (= the design):** the stateful ascent loop. thinking-engine `should_elevate` is DEAD code, `Agent.current_rung` never mutated, `LayerId` L1–L10 never dispatched, the contract_bridge selector caps at Analogical. No doc composed the full ordering before §3a (three partials: nail→hammer, gestalt→template@rung, Rubicon schedule + SPO 2³ ladder); persona-vs-rung-ladder O1/O2 are exactly these gaps. Wiring = one `RungElevator` per active thought, kanbanstep-interior; rung's mask selects which projections the 2³ consults; apex (≥6) unlocks the counterfactual machinery. Gated D-TRI-6.
+- **Five "rung" vocabularies confirmed distinct** (semantic-depth 0–9 = the pyramid axis; Pearl 1/2/3 = derived grade; content ladder 0–4 = per-band vocabulary, unwired per O1; temporal epistemic rung = read-side only, NEVER in the pyramid; DTO-ladder rungs = unrelated). Hazards recorded in §3a: L1-mask divergence (contract O vs causal-edge SO), ndarray entropy_ladder bit0=S vs planner bit2=S, scenario.rs "Pearl Rung 3"-for-interventions wobble, thinking-engine's duplicate RungLevel (jc-measure then collapse, D-TSC-1 lesson).
+- Cross-ref: E-OPPONENT-EMULATION-IS-COUNTERFACTUAL-GOALSTATE-1 (the apex distinction, same day).
+
+## 2026-07-17 — E-OPPONENT-EMULATION-IS-COUNTERFACTUAL-GOALSTATE-1 — CORRECTION: the mirror-SoA is counterfactual-goalstate EMULATION (Pearl Imagine), NOT a gestalt read; "feels like a king attack" ≠ "what would the opponent plan" (operator)
+
+**Status:** CORRECTION of the triangle plan §2 first draft (unmerged, fixed in-PR #707 §2a) — the draft filed the opponent mirror-SoA under gestalt awareness. Operator: "feels like a king attack is different from emulation of opponent's thinking based on counterfactual goalstate." The ORIGINAL E-TASKS-SOA-AT-THE-MEMBRANE-1 wording already said "**Counterfactual** and concurrent thinking span compartments"; the draft drifted from it. Three organs on the opponent axis, never blended: episodic replay (what they DID), gestalt resonance (what it FEELS like — resemblance, O(1), no simulation), counterfactual emulation (what they WOULD think given THEIR goalstate — a sibling SoA running the same Rubicon machinery, goalstate swapped).
+
+- **Wires entirely to shipped machinery:** abduce the goalstate (`InferenceType::Abduction`; D-SF-OPPONENT-3's own-L1+residue is the measured shape, and inference is the measured bottleneck per E-SF-TRAP-LURE-GREEN-1); no `GoalState` type exists anywhere — it is phase-implicit in the Rubicon `KanbanColumn`, so the "swap" = seeding the sibling's OWN Planning phase, never a field write; fork = `ScenarioBranch` (fork_seed, interventions, `CounterfactualSynthesis`) + `world/counterfactual.rs::intervene()` (Pearl Rung 3, implemented + tested; `World::fork` hypotheticals not yet per-cycle — the mirror-SoA is that wiring's first customer); emissions ride the −6 mantissa lane under the existing iron invariant "NEVER written as observed SPO truth"; Evaluation compares predicted vs actual move → NARS Revision of the abduced goalstate.
+- **The mirror-SoA is a compartment, not a carrier** — it consumes the §2 awareness surfaces internally under the swapped goalstate; the self-SoA observes it only through cloned witness VALUES (the rs-graph-llm probe-3 shape). On the pyramid it sits at the APEX (RungLevel ≥ 6, pearl_level 3); gestalt sits at the BASE. The pyramid PRESERVES the correction as a base-vs-apex separation, but the distinction itself is CATEGORICAL (O(1) resemblance-lookup vs forked simulation) — position alone does not carry it, since episodic replay also sits low.
+- Probe: D-TRI-5 (emulation vs resonance-only baseline on opponent move prediction). Same-commit fixes: plan §2 table + §2a, INTEGRATION_PLANS entry, FUTURE-DESIGN row (all unmerged in-PR edits; this entry is the merged-history record).
+
 ## 2026-07-17 — E-ACK-ELIMINATED-1 — the ack is removed from everywhere, completely eliminated (operator directive)
 
 **Status:** RULING + SHIPPED same day. Operator: "remove the ack from everywhere, completely eliminate it" — after the concept twice produced tier-crossing designs and, worse, became a lens every session saw everything through. Expelled from code, doctrine, and vocabulary; not relocated, not renamed.

.claude/board/INTEGRATION_PLANS.md

@@ -1,3 +1,25 @@
+## 2026-07-17 — triangle-tenants-gestalt-separation v1 — DESIGN (mint-gated)
+
+**Plan:** `.claude/plans/triangle-tenants-gestalt-separation-v1.md`
+**Scope (operator-directed):** the autopoiesis triangle as THREE value
+tenants (Frozen/Learned/Explore Style — 12 slots x palette256 atom each,
+slot = StyleFamily ordinal OR template step, ClassView-selected: one
+register, two readings); awareness surfaces kept apart (triangle =
+policy; episodic witness + basins = memory; gestalt awareness =
+ResonanceDto/palette resonance; opponent mirror-SoA = counterfactual-
+goalstate EMULATION, a sibling compartment at the pyramid apex — NOT a
+gestalt read, operator correction §2a; chess tenants = quarantined
+domain corpus); the nail->hammer dispatch direction (object resonance
+fetches the tool, never the inverse); the §3a Maslow pyramid of
+cognition (observation/gestalt first -> SPO 2^3 -> CausalEdge64 NARS
+candidates -> NARS candidate DESIGN/revision -> counterfactual apex;
+RungElevator wiring design, audit receipts — the ladder is shipped, the
+stateful ascent loop is the missing wiring); the 12-slot
+redundancy landed old+new and MEASURED with the jc battery before any
+collapse. Probes D-TRI-1..6. NO bytes land — everything rides the next
+batched mint + envelope-auditor gate (with W2a BoardAggregates, the
+Tasks-SoA task classid, and the chess classids as ONE batch).
+
 ## 2026-07-16 — h268-probe-wave v1 — ACTIVE
 
 **Plan:** `.claude/plans/h268-probe-wave-v1.md`

.claude/board/STATUS_BOARD.md

@@ -1,3 +1,16 @@
+## triangle-tenants-gestalt-separation-v1 — triangle tenants, surface separation, chess quarantine
+
+Plan: `.claude/plans/triangle-tenants-gestalt-separation-v1.md`. Design shipped; ALL layout work mint-gated (rides the same batched mint as W2a BoardAggregates + Tasks-SoA task classid + chess classids).
+
+| D-id | Title | Repo | Status | Evidence |
+|---|---|---|---|---|
+| D-TRI-1 | Triangle tenant spec (Frozen/Learned/Explore, 12 slots x palette256) through envelope-auditor T1-T6, behind the batched mint | lance-graph | Queued (LAYOUT-GATED) | plan §1, §6 |
+| D-TRI-2 | 12-family vs 12-step reading agreement: jc battery (ICC, Pearson/Spearman, Cronbach alpha) over real shader cycles; collapse only on measured identity | lance-graph | Queued | plan §4, §6 |
+| D-TRI-3 | Nail->hammer dispatch probe: object resonance -> atom retrieval vs inverted baseline; no inverted read path exists structurally | lance-graph | Queued | plan §3, §6 |
+| D-TRI-4 | Chess<->thinking transfer measurement (validity + reliability) — the quarantine-lift gate | lance-graph + stockfish-rs | Queued | plan §5, §6 |
+| D-TRI-5 | Emulation != resonance: counterfactual-goalstate emulation vs resonance-only baseline on opponent move prediction (operator correction §2a; builds on D-SF-OPPONENT-1/3) | lance-graph + stockfish-rs | Queued | plan §2a, §6 |
+| D-TRI-6 | Pyramid settlement: settle-rung distribution over real shader cycles (base-heavy expected); homeostatic descent verified; elevator threshold jc-calibrated | lance-graph | Queued | plan §3a, §6 |
+
 ## epiphany-integration-2026-07-04-v3 — membranes, parity, unified ruff phases
 
 Plan: `.claude/plans/epiphany-integration-2026-07-04-v3.md`. Full 13-agent review pipeline complete; execution queued behind D1→A1. Cross-repo (ruff/OGAR/lance-graph); consumer-side baton homes per BH-1/2/6.