Shay 819c8b81ac feat(phase3): compositionality, multi-step-reasoning, introspection, cross-domain-transfer v1

Spreads the four remaining Phase 3 lanes to map the full reasoning-
depth surface alongside inference-closure (already landed at e509e0d).
Each lane is a v1 honest probe per the roadmap; engineering work
follows once the full surface is visible.

Results across all five Phase 3 lanes:

  lane                      split        primary signal  foundation
  inference-closure         public/v1    0.0             1.0 / 1.0
  inference-closure         holdouts/v1  0.0             1.0 / 1.0
  compositionality          public/v1   0.0625 (1/16)   1.0 / 1.0
  compositionality          holdouts/v1  0.0             1.0 / 1.0
  multi-step-reasoning      public/v1    0.0             1.0 / 1.0
  multi-step-reasoning      holdouts/v1  0.0             1.0 / 1.0
  introspection             public/v1    0.0 (no api)    n/a
  introspection             holdouts/v1  0.0             n/a
  cross-domain-transfer     public/v1    0.0             1.0 / 1.0
  cross-domain-transfer     holdouts/v1  0.0             1.0 / 1.0

Foundation guarantees (storage + replay) intact across every lane
that has them. The reasoning-depth signal is uniformly zero. The
five lanes triangulate four architectural gaps:

  Gap 1. generate/graph_planner.py has no transitive composition.
  Gap 2. field/propagate.py has no derivable-but-not-asserted recall.
  Gap 3. core/cognition/explain.py module does not exist.
  Gap 4. no structural-pattern recogniser (cross-subdomain transfer).

Gaps 1, 2, 4 cluster on the same code surface and may close together
as a single bounded PR. Gap 3 is independent module-creation work.

Lane scaffolding mirrors inference-closure (contract.md, runner.py,
dev + public/v1 + holdouts/v1 cases.jsonl, baselines/v1_structural_zero.json,
gaps.md). All runners are parallel-safe and use the standard
run_lane(cases, *, config, workers) interface.

Per-lane gaps.md records the engineering shape for v2 plus future
directions worth not forgetting:
  - compositionality/gaps.md: metaphor is compositionality with
    selective property transfer; building it is correctly downstream
    of closing this lane.
  - cross-domain-transfer/gaps.md: metaphor + narrative as
    cross-domain operators; narrative requires the Agency open-scope
    decision to pin first.
  - introspection/gaps.md: explain API is also the substrate for
    first-person narrative self-account.

Recommended v2 sequence in docs/PROGRESS.md:
  1. Pin Agency + Tool-use open-scope decisions (deadline: before
     Phase 3 engineering).
  2. Engineer Gaps 1 + 2 as one bounded PR.
  3. Engineer Gap 3 independently.
  4. Re-author cross-domain-transfer v2 with matched-control
     contract refinement.

Phase 3 v1 exit: 0/5 lanes passing, which is the expected v1 floor.
CLI suites smoke / cognition / teaching pass; no regression on
Phase 2.

2026-05-16 14:48:36 -07:00

2.2 KiB

Raw Blame History

introspection lane — architectural findings (v1)

v1 result

Split	n	explain_api_present	account_nonempty	surface_match	trace_match
public/v1	12	0.0	0.0	0.0	0.0
holdouts/v1	8	0.0	0.0	0.0	0.0

Structural zero by construction: there is no explain callable to import from core.cognition.

Why this is the right v1

A lane that can't run at all is worse than a lane that runs and reports a typed zero. The introspection lane runs today, attempts the import, catches the failure deterministically, and emits four sub-metrics — all zero, all explained. The day someone lands a core/cognition/explain.py module, this lane immediately starts producing real numbers without any test infrastructure change.

Required engineering for v2

The roadmap (docs/capability_roadmap.md Phase 3 work items) is explicit:

A new cognition/explain.py module may be needed for introspection.

Concretely, an explain(result: CognitiveTurnResult) -> str function that:

Reads structured state from the result — intent tag, proposition graph, articulation target, vault hits, identity score.
Composes a deterministic natural-language account that re-states the trajectory in source language. Probably leans on the same realize_semantic machinery currently used for articulation but inverted: surface → structured trace → surface'.
Round-trip property: feeding the account back through the pipeline produces an articulation whose token coverage of the original surface is high. Strict trace-hash equivalence is the ideal but not the v1 bar — surface token overlap ≥ 0.60 is the v1 contract.

Future direction (recorded here so it's not forgotten)

A working introspection API is also the substrate for narrative self-explanation: the same machinery that produces "I answered X because I retrieved Y under intent Z" is what produces an agent's own first-person account of a turn. Per the open scope decision in docs/PROGRESS.md (Agency: responsive vs. goal-directed), this choice should pin before introspection v2 is engineered.

Status

v1 is structural-zero scaffolding. Permanent regression evidence of the missing module.

2.2 KiB Raw Blame History