docs: close G2 float row ledger receipt

[briansrls]

Summary

Records the post-merge Grounding G2 Phase 2 float-row closure authorized after PR #2570 merged at b96a51a2. Updates the R3 schedule, §1.8 ledger rows, and R3 structure notes from HELD/DECLARED wording to the ratified Grounding G2 primitive-row CONSUMER_LANDED + PASSING receipt while keeping broader Real emission and any physical Float = Float64 alias deletion explicitly out of this closure.

Closes #2102

Test plan

• Docs-only change; no code tests run.

[briansrls]

Addressed the cursor narrative-clarity findings in c9ee86058:

• docs/r3-program-plan.md now says PR-F “landed” / “unblocked” in the Grounding-side sequence, matching the closed G2 state.
• docs/r3-design-schedule-2026-05-06.md now marks the G2 dispatch trigger as satisfied and narrows the closure predicate to G2 primitive-row full Rust grounding, with L5 readiness left as downstream on L4 corpus + Shape A multi-target grounding.

Docs-only wording fix; no code tests run.

— sent from proud-raven-495

[briansrls]

Review metadata

• Provider / model: openai-pro / gpt-5-5-thinking
• Commit: b4fa95d4 · Trigger: manual
• Comparison: main @ 756abc39 ... docs/g2-float-ledger-b96a51a2 @ b4fa95d4
• Conversation: View conversation

---

1. Story of the diff

This PR is a documentation/ledger closeout for Grounding G2 after the float-row consumer landed. It changes the R3 schedule from “G2 held on Substrate S7/S8” to “G2 done / consumer landed + passing,” and it records the mechanism: PR #2570’s f32/f64 Rust primitive rows consume the new float/Real substrate shape through ApproximateFieldAlgebra, with pilot mirror and grounding_engine full-list validation (docs/r3-design-schedule-2026-05-06.md:414, docs/r3-design-schedule-2026-05-06.md:430). The program plan and structure docs then propagate that same closure into the relevant R3 predicates, while explicitly preserving downstream scope: broader Real emission demonstrations remain under S9/Shape-A follow-ons, and L5 readiness still waits on L4 corpus plus Shape A multi-target grounding (docs/r3-program-plan.md:244, docs/r3-program-plan.md:850, docs/r3-structure.md:94, docs/r3-design-schedule-2026-05-06.md:433).

2. Invariant categories

1. LAYER MODEL (substrate vs implementation).

Compliant — the diff does not introduce or mutate substrate code; it updates planning/ledger docs to say the already-landed substrate/consumer chain now satisfies Grounding G2. The wording is careful to distinguish primitive-row consumer closure from broader emission work: docs/r3-program-plan.md:244 says the Int<N> / Nat<N> / Real<N> surface has “Grounding G2 primitive-row consumer evidence,” while docs/r3-program-plan.md:850 keeps broader Real demonstrations in S9/Shape-A follow-ons.

1. INVARIANTS.md + modeling-discipline.md.

Compliant — Boundary Discipline requires a declaration, realization, and generated consumer proof before a boundary counts as landed; the diff’s closure claim is tied to Rust f32/f64 rows, grounding_pilot mirror routing, and grounding_engine full-list validation (docs/r3-program-plan.md:244, docs/r3-structure.md:94). That matches the uploaded invariant rule that a boundary counts as landed only when declaration, realization, and generated consumer proof exist. chatgpt-review-db87f1d9-4b40-48…

1. CODING.md.

N/A — diff is documentation-only; it adds no Rust functions, methods, error shapes, helpers, or module organization decisions.

1. TESTING.md.

Compliant — no new test is required in this PR because it is a ledger update over already-landed consumer evidence. The diff names the behavior-level validation it relies on rather than adding a documentation-only test: grounding_engine full-list validation over loaded rust_pilot_primitives (docs/r3-program-plan.md:244) and f32/f64 primitive-row coverage through ApproximateFieldAlgebra (docs/r3-program-plan.md:850). That is consistent with the testing authority’s preference for behavior/interface claims over implementation-detail assertions. chatgpt-review-035c1fc8-3df9-49…

1. LOCKED DESIGN DECISIONS.

Compliant — the diff preserves the locked thesis direction that grounding is structural and target primitives are modeled through declared algebraic inhabitance rather than name-keyed shortcuts. The added G2 text says the rows consume ApproximateFieldAlgebra and explicitly leaves L5 / Shape A follow-ons open (docs/r3-design-schedule-2026-05-06.md:433, docs/r3-program-plan.md:850), which avoids turning primitive-row closure into a broader cross-target closure claim. The thesis authority frames grounding completeness as structural algebra-homomorphism search with fail-closed refusal when a primitive cannot be grounded. chatgpt-review-ebfc3c95-3dfc-40…

1. TRACKED vs UNTRACKED DEBT.

Compliant — no new scaffold, TODO, or temporary implementation path is introduced. The only remaining work is explicitly bounded and routed: downstream L5 readiness depends on L4 corpus + Shape A multi-target grounding (docs/r3-design-schedule-2026-05-06.md:433), and broader Real emission demonstrations remain in S9/Shape-A follow-ons (docs/r3-program-plan.md:850, docs/r3-structure.md:94).

2.5. Top-down PM intent review

Compliant — the PR preserves the PM-level intent rather than diluting it. The highest-level intent is structural grounding and mechanical emission from declared facts, not a manual “float rows exist” checkbox; the diff keeps closure scoped to “G2 primitive-row full Rust grounding” and explicitly does not claim L5 readiness or all Real emission work (docs/r3-design-schedule-2026-05-06.md:433, docs/r3-program-plan.md:850). That aligns with the thesis claim that target primitive grounding is structural and fail-closed, and with the broader Shape A direction that programming-language targets are emitted from language specs rather than per-target compiler paths. chatgpt-review-ebfc3c95-3dfc-40…

3. Verdict

APPROVE — I do not see a diff-cited invariant violation. The PR is a narrow, consistent ledger closeout: it marks G2 closed where the consumer evidence is named, and it preserves the remaining S9/Shape-A/L5 dependencies instead of overclaiming them.