# Connectivity manifest — what actually connects, end to end **This file is the single source of truth for connectivity claims.** No document in this repository (README, INDEX, module docstrings, commit messages) may assert a stronger end-to-end connection than a row here marked **CONNECTED** with a named backing theorem. `scripts/check-connectivity.sh` enforces the parts of this that can be checked statically. Created 2026-07-07 after a provenance audit found the top-level "one posited object yields both pillars" claim was **not** realized in the code. ## The honest one-paragraph status **(Updated 2026-07-15 — supersedes the 2026-07-07 "not realized" wording, which predated fixes C1–C7 and the Σ-layer and contradicted the CONNECTED rows below.)** The FORWARD chain is now realized end-to-end on genuine objects. A single genuine `Φ ≠ id` Kähler sector supports BOTH pillars together — the Schrödinger form (`H ≠ 0`) AND the Born frequencies from sampling its own Liouville measure — first on `rotationSetup` (C4, L6) and, at GENERAL `N` with a genuine many-to-one `π = Prod.fst` (`Σ = ℂℙ^{N-1} × T²`, fibres `= T²`), on `manyToOneSchrodingerSetup` with arbitrary Hermitian `H` (C7, L8). The SigmaLayer (the projective-sector ontology, Paper C) (2026-07) then puts isolated Hamiltonian dynamics AND de-isolating measurement AND time-indexed records AND the Born content AND the conditional state update on ONE ontic model `(Σ, μL, Φ, π)` — the unified capstone `unified_projectiveSector_capstone` (L9). So "a single posited Kähler sector yields both pillars, and one ontic model carries dynamics + measurement + records + Born + update" is CONNECTED. What the sector is **POSITED** — Σ is the floor, and **deriving it is a non-question** (the retired "SO-1 / L7 Born-from-flow" gap). Born *arises from* the Ω-region volume ratios + typicality; that IS the reconstruction, not a "derive the sector" debt. The goal now is to **complete the reconstruction of QM from Σ and Ω-regions**; the record layer that does this is now **formalized in Lean** (2026-07-25, `record-layer-plan.md §4`; measurement as `context + unknown microstate → record` on the base×fibre Σ, Born = the LLN over the unknown microstate). The residual is context-fixing the *pinned closure's* regions (still the prep-indexed `bornRegion ψ'`); the near frontier is now the **extensions** (continuous spectra, relativistic locality, identical particles). ⚠️ *Frontier vocabulary superseded 2026-09-05:* `specs/POSITS.md` ("What 'frontier' means here") separates **permanent boundary** / **open mathematics** / **open foundations**. The surviving foundational frontier is Posit 1's discharge condition — deriving `IsTorusGenerated` from the de-isolation dynamics. The extensions are breadth, not the frontier. (Paper C **Axiom A5** = projectability, never "the origin of Σ".) The Kähler-geometry fields (L1) now carry their genuine pointwise/linear core (the concrete field `kahler_pointwise : IsFubiniStudyKahler`, proved; F-04 tightening 2026-08-06); only the manifold residual (`dω = 0`, top-power volume identity) was an interpretive posit until 2026-09-10 (W1, `LF4/SectorManifold.lean`: L1 DISCHARGED) (no Mathlib exterior-calculus API). ## The intended chain and per-link status Intended dependency chain: ``` Kähler geometry (ω, complex structure) → compact ontic Σ + Liouville measure → deterministic measure-preserving flow Φ on Σ → projected flow φ_t on ℂℙ^{N-1} (via π, the descent equation `projectable`) → { Born rule: frequencies → FS-volume = ‖⟨e_i,ψ⟩‖² } → { Schrödinger: φ_t = exp(-itH)· on rays } ``` | Link | Claim | Status | Backing theorem / the gap | |---|---|---|---| | **L1** | Kähler geometry `⇒` the sector's fields | **DISCHARGED 2026-09-10 (W1). History: PARTIAL (2026-07-07 fix C5; pointwise Kähler core DISCHARGED 2026-07-19; fields made CONCRETE 2026-08-06, F-04): both geometry fields are concrete structure fields any consumer can unpack; the manifold residual is now proved and wired** | `liouville_isProbability` carries the formalizable core — `μ_FS` is a *normalized* volume (probability measure), now an instance — exposed by `unitaryFlowSetup_liouville_isProbability` (`LF4/NonTrivialSetup.lean`). The Kähler condition is the concrete field `kahler_pointwise : IsFubiniStudyKahler N` — the pointwise Fubini–Study Kähler-compatibility triple (`g = re⟪·,·⟫`, `ω = im⟪·,·⟫`, `J = i•·`, with `J² = -1`, `ω = g∘J`, `g = ω∘J`, `ω` a `(1,1)`-form, `ω u (Ju) = ‖u‖²`), PROVED axiom-free (`CSD.LF4.isFubiniStudyKahler` / `isFubiniStudyKahler`, `Mathlib/Analysis/InnerProductSpace/KahlerForm.lean`). The **flat** closedness `dω = 0` is now PROVED on the tangent model (2026-08-06, `Kahler.extDeriv_fundamentalFormAlt` in `Mathlib/Analysis/InnerProductSpace/KahlerClosed.lean`, via the generic `extDeriv_const`); The **manifold** residual — closedness and the top-power identity `ω^{∧(N-1)}/(N-1)! = μ_FS` ON `ℂℙ^{N-1}` itself — is PROVED (2026-09-07/08: `Projectivization.fsForm_mextDeriv`, `fsForm_isSymplectic`, `fsVolume_eq_smul_fsMeasure`; Kähler 2026-09-10: `fsForm_isKahler`) and WIRED to the sectors (2026-09-10, `LF4/SectorManifold.lean`: `unitaryFlowSetup_liouvilleMeasure_eq_fsVolumeNormalized`, `fsVolumeNormalized_isForcedKahlerVolume`, `unitaryFlowSetup_isKahler_liouville`, `manyToOneSetup_isKahler_liouville`). The structure's fields are unchanged; on the `ℂℙⁿ` instances they are the standard objects. | | **L2** | Σ + `Φ` + `π` `⇒` a well-defined projected flow | **CONNECTED (interface) but see L4** | The descent field `projectable : π(Φ_t x) = φ_t(π x)` is a genuine field and is consumed by `CSD.LF4.sigmaFlow_schrodinger_form` (`LF4/PhaseLift.lean`). Enforced by `scripts/check-sector-linkage.sh`. | | **L3** | projected flow `⇒` Schrödinger form | **CONNECTED (2026-07-07, fix C2)** | `CSD.LF4.rotationSetup_schrodinger_form` (`LF4/RotationSchrodinger.lean`) fires `sigmaFlow_schrodinger_form` on the genuine `Φ ≠ id` `rotationSetup`: the FS-isometry (unitary flow), coboundary (`b = 1`, trivial cocycle since `R(s+t) = R(s)R(t)`), and C¹ (`R'(τ) = R(τ)·J`) data are all discharged on a non-trivial flow, recovering `H = iJ = σ_y ≠ 0` (`rotationSetup_generator_ne_zero`). The first fully-instantiated `H ≠ 0` Schrödinger statement — no longer trivial-witness-only. **General-`N` (2026-07-19):** `CSD.LF4.manyToOneSchrodingerSetup_schrodinger_derived` (`LF4/ManyToOneSchrodingerDerived.lean`) exercises the same C¹-Stone derivation on the REAL family at general `N` with ARBITRARY Hermitian `H`: `schrodingerUnitary_hasDerivAt` discharges the smoothness datum `U' t = U t·(-iH)` (via `hasDerivAt_exp_smul_const` under the `C^*` L2-operator norm), and `Matrix.StoneC1.eq_exp_of_hasDeriv` recovers `U t = exp(t•A)`, `A = -iH` — so the general-`N` `manyToOneSchrodingerSetup_schrodinger_form` (`rfl`) is backed by an actual derivation, not only the `N=2 σ_y` or `A=0` witnesses. AxiomAudit-pinned. | | **L4** | a genuine `Φ ≠ id` `KahlerOnticSetup` inhabitant exists | **CONNECTED (2026-07-07, fix C1)** | `CSD.LF4.rotationSetup` (`LF4/NonTrivialSetup.lean`) is a `KahlerOnticSetup 2` whose projected flow is the `ℂℙ¹` rotation `R(t)`; `rotationSetup_projectedFlow_ne_id` proves `∃ t, projectedFlow t ≠ id` (at `t = π/2`, `[e₀] ↦ [e₁]`). The general constructor `unitaryFlowSetup N U p₀` builds one from any unitary family (measure-preserving via `fsMeasure_smul_invariant`). NB: `kFlow` was the wrong tool — it translates a `T²` fibre and so acts trivially on rays (`projectedFlow = id`). | | **L5** | sector `⇒` Born frequencies (structural) | **CONNECTED (structural, 2026-07-07, fix C4)** | `unitaryFlowSetup_born_frequency` (`LF4/BothPillars.lean`) states `born_frequency_convergence_N` with the sampling law = the SECTOR FIELD `d.liouvilleMeasure` (definitionally `fsMeasure p₀`), so the Born theorem now references the same object the Schrödinger chain consumes. CAVEAT: structural sharing only — the trials `X` still SAMPLE the measure i.i.d.; they are not evolved by `d.flow`, and the weights are not derived from the dynamics (that is L7 = C6). **STRENGTHENED 2026-07-25 (ontic-grounded, general):** `SigmaLayer/OnticBornFrequency.lean` — `born_frequency_from_ontic_sampling` states the Born frequency for **any** `ProjectiveSector` with the pushforward bridge, sampling the **ONTIC** `μ_L = D.muL` (not the projective `μ_FS`); the projective law and Born weight are DERIVED, and `onticBornVolume_eq` names the bridge identity **Born = ontic typicality volume** (`μ_L(π⁻¹Ωᵢ) = ‖⟨eᵢ,ψ⟩‖²`). So the *only* trial hypothesis is now at the ontic floor. It does **not** close L7: that the trials are i.i.d.-`μ_L` (vs. flow-evolved) remains the floor (SO-1 / repeated-preparation typicality — deriving it from a single flow's ergodicity is deliberately not the CSD route). | | **L6** | ONE object underlies both pillars | **CONNECTED (2026-07-07, fix C4)** | `rotationSetup_both_pillars` (`LF4/BothPillars.lean`) proves, for the SINGLE `rotationSetup p₀`, BOTH (A) Schrödinger (`π(Φ_t x) = exp(-itH)·π(x)`, `H=σ_y`) AND (B) Born (frequencies from sampling `(rotationSetup p₀).liouvilleMeasure` → `‖⟨eᵢ,ψ⟩‖²`). One `KahlerOnticSetup` instance, both pillars. | | **L7** | Born weights derived FROM the deterministic flow | **N/A — deriving Σ is a non-question; but *constraining* Σ is real work** | "Derive the sector/Born from the flow" is retired: **Σ is the floor; deriving it is a non-question.** Born *arises from* the Ω-region volume ratios + typicality — that IS the reconstruction. The theorems once filed here are **constraint work on Σ** (Σ is hidden, so pin it down as tightly as possible, from above): `SigmaLayer/LocalisedTypicality.lean` shows the full `U(N)` symmetry *forces* `μ_FS` (Paper B's uniqueness, in Lean) — Σ's measure is not arbitrary; `SigmaLayer/SectorPostulateNoGo.lean` (`flow_admits_invariant_ne_fubiniStudy`) records only that a single *epistemic* flow does not time-average to `μ_FS` — expected, irrelevant to CSD's mechanism (typicality is repeated-preparation ignorance over `Ω₀`). Legitimate constraint work, not a *research frontier*. | | **L8** | ONE object with a GENUINE many-to-one `π` (Paper C A3) underlies both pillars | **CONNECTED (2026-07-08, fix C7)** | `manyToOneRotationSetup_both_pillars` (`LF4/ManyToOnePillars.lean`) proves, for the SINGLE `manyToOneRotationSetup p₀` — whose `Σ = ℂℙ¹ × T²`, `π = Prod.fst` is genuinely many-to-one (fibres `= T²`, `manyToOneSetup_pi_not_injective`) AND whose projected flow is a non-trivial ray rotation (`manyToOneRotationSetup_projectedFlow_ne_id`) — BOTH (A) Schrödinger (`π(Φ_t x) = exp(-itH)·π(x)`, `H=σ_y`, inherited from `rotationSetup_schrodinger_form`) AND (B) Born (frequencies from sampling `kMuL p₀` and scoring the FIBRED region `π⁻¹'(bornRegion ψ i)` → `‖⟨eᵢ,ψ⟩‖²`, via `manyToOneSetup_born_frequency`). At GENERAL `N` with ARBITRARY Hermitian `H`: `manyToOneSchrodingerSetup_both_pillars` (`U t = exp(-itH)`). Removes the `π = id` degeneracy of L6/`rotationSetup_both_pillars` — the Paper-C A3 caveat below. AxiomAudit-pinned (foundational triple, Gleason-free). SAME standing gap as L5/L7: the Born trials still SAMPLE `kMuL` i.i.d. rather than being evolved by `d.flow`, and the fibre flow is trivial (the flow moves only the base ray). | | **L9** | ONE ontic model carries isolated dynamics AND de-isolating measurement AND records AND Born AND state update | **CONNECTED (2026-07-17: assembled into one tiered closure `FiniteQMClosure`)** | `unified_projectiveSector_capstone` (`SigmaLayer/UnifiedMeasurement.lean`) BUNDLES the dynamics + measurement CORE on the SAME `(Σ = ℂℙ^{M}×T², μL = μFS ⊗ vol, π = Prod.fst)`: the ISOLATED dynamics is the genuine `exp(-itH)` Hamiltonian flow (`productDynamics`, measure-preserving + Schrödinger-projectable + `π_*μL = μFS`) AND a `DeisolationModel` over that same dynamics realises the de-isolation MEASUREMENT (`(measurementFlow p, θ)`, measure-preserving, a.e.-defined pointer readout `vnDeisolationModel_ae_total`, record establishment). All the remaining pieces are now ASSEMBLED with it INTO one tiered record **`FiniteQMClosure` / `unifiedFiniteQMClosure`** (`SigmaLayer/FiniteQMClosure.lean`): time-physical records — INSTANTIATED on the unified model (`unifiedFlowedSemantics` = `flowedSemantics` over the isolated `productDynamics` flow; `unified_records_persistence`: Born weight conserved + flow-covariant; `unifiedFlowedSemantics_zero`: the static `vnRecordSemanticsProd` is the `t=0` slice), #5; post-outcome preparation with proven nonzero measure (`PostMeasurement.appendFact`); Born-frequency convergence — stated ON the model (`unified_born_frequency`: i.i.d. trials of `productDynamics.muL`, frequency in `π⁻¹(bornRegion i)` → `‖⟨eᵢ,ψ⟩‖²`, for EVERY unit `ψ` — `hpos` retired via the `_uncond` engine), #2; and the state update = record conditioning = Lüders update — GENUINELY proved (`ConditioningLuders.conditioning_luders_effect_equivalence`: the ontic record-conditioning and the Lüders update give the same conditional prediction for every pointer-basis effect, via the proved weight agreement `onticRegion_measure_eq_born`; the earlier `ConditioningLink.luders_record_conditioning_correspondence` only conjoined the two Bayesian halves and ASSERTED their agreement); and mixed-state Born (`MixedOntic.mixed_ontic_born_weight` = the `mixed_born` field, #8 C: the classical mixture over `ρ`'s spectral ensemble of ontic Born-region measures reproduces `Tr(ρ Eᵢ)`, so the model carries mixed-state Born content, not only the pure `ψ`). `FiniteQMClosure` carries these eleven as fields (the tenth/eleventh being mixed-state Born as a weight `mixed_born` and as an a.s. frequency `mixed_born_frequency` / `unified_mixed_born_frequency`, `SigmaLayer/MixedFrequency.lean`: a two-stage mixture LLN converging to `Tr(ρ Eᵢ)`), each discharged by its source lemma; the Choice-A posit, QM adapters, and open residue are documented in the module header (no field is `sorry`) — so the tiers stay honest. AxiomAudit-pinned. Consumes the posited sector — Σ is the floor (deriving it is a non-question); the record layer is now built; the frontier is now the extensions, not this. **Addendum 2026-08-02 — L9 SUPERSEDED BY A STRONGER WITNESS:** `unifiedArenaClosure` (`RecordLayer/UnifiedArena.lean`) carries the whole reconstruction on **one arena with one Liouville measure family** — the lifted `exp(-itH)` and the record-creating `swapEvolve` preserve the *same* `arenaLiouville`, records are context-fixed and dynamically created (superseding the preparation-indexed `vnPointerOutcome` readout), Lüders is the *dynamical* pushforward (superseding conditioning-as-prediction), and the isolate→measure→isolate round trip is a theorem (`arena_round_trip`). The `FiniteQMClosure` witness above stands unchanged as the historical operational capstone. | | **L10** | records of ONE flow on the arena of UA-18 `⇒` two observers' radar coordinates related by a boost, within an error that vanishes along a family | **CONNECTED for the coordinates and the interval, on an open set of pointer angles containing the box of radius `1/(12000 k²)` about the start (2026-11-17)** | `csd_operational_lorentz_from_records` (`RecordLayer/UnifiedRadarGeometry.lean`), `wallRunAt_orbit` (`RecordLayer/UnifiedRecordSpacetimeBridge.lean`). The kinematics is that of the arena's Dirac carriers (`specs/POSITS.md`, Posit 11): a reconstruction, not a derivation. Audit: `specs/unified-m6-audit.md`. UO-16 was marked complete by the author on 2026-11-17, with the audit's findings as its scope. | **Net (updated 2026-07-08 after C1+C2+C4+C7):** **L2–L6 + L8 CONNECTED.** A single genuine `Φ ≠ id` object (`rotationSetup`) supports BOTH pillars: the Schrödinger form (`H = σ_y ≠ 0`) AND the Born frequencies (sampling its own `liouvilleMeasure`), proved together in `rotationSetup_both_pillars`. And (C7, L8) a single object with a GENUINE many-to-one `π` (`Σ = ℂℙ¹ × T²`, `π = Prod.fst`, fibres `= T²`) with a non-trivial projected ray flow supports both pillars too (`manyToOneRotationSetup_both_pillars`), removing the `π = id` degeneracy and matching Paper C's A3 fibred-projection shape. Remaining: - **L1** — the Kähler-geometry fields now carry genuine formalizable content (`kahler_pointwise : IsFubiniStudyKahler`, the proved pointwise FS Kähler compatibility; `liouville_isProbability` = normalized volume; concrete fields since the 2026-08-06 F-04 tightening). The **manifold** residual (closedness `dω=0`, top-power volume identity) is proved and wired since 2026-09-10 (`LF4/SectorManifold.lean`, W1) — L1 DISCHARGED. - **L7 — N/A (a non-question).** "Born weights derived FROM the flow" is retired: Σ is the floor, and the Born trials sampling the sector's measure is not a gap — Born *arises from* the Ω-region volume ratios + typicality, which IS the reconstruction. (Paper C Axiom A5 = projectability, not "the origin of Σ".) The record layer is now built; the frontier is now the extensions, not this. ⚠️ *Frontier vocabulary superseded 2026-09-05:* `specs/POSITS.md` ("What 'frontier' means here") separates **permanent boundary** / **open mathematics** / **open foundations**. The surviving foundational frontier is Posit 1's discharge condition — deriving `IsTorusGenerated` from the de-isolation dynamics. The extensions are breadth, not the frontier. So the honest headline is now: **"a single posited Kähler sector object yields both the Born rule and Schrödinger dynamics"** is CONNECTED (structurally); the remaining frontier is deriving that sector — and the Born trials — from the deterministic flow itself. ⚠️ *Corrected 2026-09-05:* as written this is the **retired "derive Σ"** framing — a non-question (`specs/CSD-CHARTER.md`), and `SectorPostulateNoGo` proves a deterministic flow does not pin the sector. What is legitimately open is Posit 1's discharge condition (derive the *rate field* from the de-isolation dynamics), not deriving the sector. **Paper-C cross-check caveat (2026-07-08).** Paper C's architecture (Axiom A3) requires the projection `π : Σ → ℂℙ^{N-1}` to be **smooth many-to-one** — `Σ` is strictly LARGER than ray space, and each epistemic ray `[ψ]` is the image of a whole fibre `π⁻¹([ψ])` of ontic microstates (the fibre coordinates carry the Liouville / absolute-continuity structure). The both-pillars object `rotationSetup` uses **`π = id`** — the DEGENERATE one-to-one case, `Σ = ` ray space, fibres = points. So the C1/C2/C4 theorems, while valid, do NOT exercise Paper C's genuine many-to-one ontic projection; they run on `Σ = ℂℙ^{N-1}` directly. A genuine many-to-one `π` DOES exist in the corpus (`KSigma = ℂℙ^{N-1} × T²`, `π = Prod.fst`, fibres `= T²`, pushforward bridge `Prod.fst_* kMuL = μ_FS` — `KahlerFlow.lean`), but (i) it is on the older `SectorData` track, not `KahlerOnticSetup`, and (ii) its flow `kFlow` acts trivially on rays (`projectedFlow = id`), so Schrödinger on it is trivial. **No single object yet has BOTH a many-to-one `π` AND a non-trivial projected flow** — that is the fix **C7** (below), the honest next step to match Paper C's A3. Also note: our Schrödinger route is Wigner-rigidity + phase-lift + Stone, NOT Paper C's Ashtekar–Schilling (holomorphic-vector-field) derivation — same endpoint, different path (⚠️ both paths exist since 2026-09-12, see the note below). Paper C itself (§A/§3.6) states `Σ`, `π`, and the quantum-effective sector (selected by A5) are **assumed, not derived** — **Σ is the floor** (deriving it is a non-question). So our posited-sector scope matches Paper C's own. **Update (2026-07-08, fix C7 — A3 caveat RESOLVED, link L8).** The "no single object has BOTH a many-to-one `π` AND a non-trivial projected flow" gap is now closed. `LF4/ManyToOnePillars.lean` builds `manyToOneSetup U p₀`: a `KahlerOnticSetup` on the fibred `Σ = ℂℙ^{N-1} × T²` with `π = Prod.fst` (genuinely many-to-one — `manyToOneSetup_pi_not_injective`, fibres `= T²`) and a flow `Φ_t (p, θ) = (U t • p, θ)` that ROTATES THE BASE RAY (so `projectedFlow` is the genuine ray action, non-trivial for the rotation, unlike `kFlow`). Both pillars fire on the concrete `N = 2` witness (`manyToOneRotationSetup_both_pillars`, link L8): Schrödinger inherited verbatim from `rotationSetup_schrodinger_form` (identical base ray action), Born via `manyToOneSetup_born_frequency` scoring the FIBRED region `π⁻¹'(bornRegion ψ i)` — whose `kMuL`-volume equals the base Born weight because the fibre volume is normalized (`Prod.fst_* kMuL = μ_FS`, `Measure.fst_prod` / `k_measure_bridge`), reducing to `born_frequency_convergence_N` on the projected trials. This is now on the `KahlerOnticSetup` track (not the older `SectorData` track), so it composes with the Schrödinger chain. **Not addressed by C7 (nor a gap):** "weights-from-flow" (L7) is a non-question — Σ is the floor; the Born trials sampling `kMuL` is not a debt (Born arises from the Ω-region volume ratios). C7 fixes the *projection shape*; the frontier is now the extensions (the record layer is built). ⚠️ *Frontier vocabulary superseded 2026-09-05:* `specs/POSITS.md` ("What 'frontier' means here") separates **permanent boundary** / **open mathematics** / **open foundations**. The surviving foundational frontier is Posit 1's discharge condition — deriving `IsTorusGenerated` from the de-isolation dynamics. The extensions are breadth, not the frontier. Our route is still Wigner, not Ashtekar–Schilling. ⚠️ *Superseded 2026-09-12 (Q29(e)), reconciled 2026-09-14 (KG-3):* the geometric route exists too — the Hamiltonian flow of the projected energy `−2⟨H⟩` for the Fubini–Study symplectic form IS `exp(−itH) •` (`hamiltonianFlow_schrodingerHamiltonian`), and on the sector `manyToOneSchrodingerSetup_projectedFlow_eq_hamiltonianFlow`; both routes produce the same `schrodingerUnitary`. ## What may be claimed (until a link flips to CONNECTED) **(The "both pillars on one posited sector" claims are CONNECTED forward results, so they are permitted; the forbidden line is: do not claim the sector/Σ is *derived*, and do not sell the epistemic witness as completing the reconstruction. Σ is the floor.)** - ✅ "CSD machine-verifies the Born rule as an FS-typicality volume (on `ℂℙ^{N-1}` with `μ_FS`), Gleason-free, general `N`, incl. POVMs." — a real theorem about the FS measure. - ✅ "GIVEN a posited Kähler ontic sector, CSD derives finite-dim Schrödinger dynamics via Wigner, at general `N` with arbitrary Hermitian `H`." — a real conditional theorem (`sigmaFlow_schrodinger_form`), now instantiated non-trivially (`manyToOneSchrodingerSetup`, `H` arbitrary). - ✅ "A single posited Kähler sector object yields BOTH the Born rule AND Schrödinger dynamics" — CONNECTED (L6/L8), FORWARD, at general `N` (`manyToOneSchrodingerSetup_both_pillars`). Must carry the posited-sector caveat (Σ is the floor; the sector is posited, not derived). - ✅ "One ontic model `(Σ, μL, Φ, π)` carries isolated dynamics, de-isolating measurement, time-indexed records, Born frequencies and the conditional/Lüders state update" — CONNECTED (L9: `FiniteQMClosure` / `unifiedFiniteQMClosure` assembles the dynamics + measurement core, records #5, Born-freq #2 and conditioning=Lüders #3/#4 into one tiered closure), FORWARD. *(Since 2026-08-02 the stronger one-arena form is also available: `unifiedArenaClosure`, see the L9 row addendum.)* - ✅ "The thermodynamics track (TH1–TH4) is proved." — real, independent. - ❌ "CSD DERIVES the sector `(π, G)` / the Born weights FROM the deterministic flow" / "the sector is no longer posited" — **FORBIDDEN**. Σ is the floor; the sector is posited, not derived (deriving it is a non-question). Every CONNECTED claim above is FORWARD (consumes the posited sector). - ❌ Any phrasing implying the Kähler differential geometry (closed 2-form, complex structure) is constructed / load-bearing — **FORBIDDEN** (L1: no Mathlib Kähler API; the VOLUME is forced, the differential-geometric fields are demoted interpretive posits). ## The fix course (sequenced) **Phase 1 — make the Schrödinger pillar non-trivial (bounded).** - **C1 (fixes L4) — DONE 2026-07-07** (`LF4/NonTrivialSetup.lean`). Built `unitaryFlowSetup N U p₀` (a `KahlerOnticSetup` from any unitary family, measure-preserving via `fsMeasure_smul_invariant`) and the concrete `rotationSetup` at `N = 2` (the `ℂℙ¹` rotation flow), with `rotationSetup_projectedFlow_ne_id : ∃ t, projectedFlow t ≠ id`. NB: `kFlow` was NOT reusable — it acts trivially on rays (`projectedFlow = id`); the fix uses a projective *unitary* flow instead. - **C2 (fixes L3 off-trivial) — DONE 2026-07-07** (`LF4/RotationSchrodinger.lean`). `rotationSetup_schrodinger_form` fires `sigmaFlow_schrodinger_form` on `rotationSetup`: FS-isometry (unitary flow), coboundary (`b=1`, trivial cocycle via `R(s+t)=R(s)R(t)`), C¹ (`R'=R·J`) all discharged; recovers `H = iJ = σ_y ≠ 0`. The Schrödinger form now holds on a genuine `Φ≠id` flow with `H≠0`. - **C3 (done 2026-07-07).** Retract the overclaiming docs; add `scripts/check-connectivity.sh` enforcing non-trivial-instance + no-overclaim. **Phase 2 — one object, and Kähler content (medium).** - **C4 (fixes L5/L6 structurally) — DONE 2026-07-07** (`LF4/BothPillars.lean`). `unitaryFlowSetup_born_frequency` states the Born law with sampling measure = the sector field `d.liouvilleMeasure` (defeq `fsMeasure`); `rotationSetup_both_pillars` proves Schrödinger ∧ Born on the SINGLE `rotationSetup` object. Structural sharing only — weights-from-flow is C6. - **C5 (fixes L1) — DONE 2026-07-07** (`LF4/NonTrivialSetup.lean`). Made `IsLiouvilleKahlerVolume` carry genuine content (`IsProbabilityMeasure μ_FS`, the normalized-volume core) and consumed it (`unitaryFlowSetup_liouville_isProbability`). `IsKahlerSector` honestly demoted to an unformalizable interpretive posit (no Mathlib Kähler API at the time; the corpus built the API 2026-09-07/11 and L1 is now DISCHARGED, see the L1 row) — documented in `PLACEHOLDERS.md` and the structure docstring; no doc implies the Kähler differential geometry is load-bearing. **Phase 2b — genuine many-to-one projection (Paper C A3; medium).** - **C7 (fixes the A3 caveat, link L8) — DONE 2026-07-08** (`LF4/ManyToOnePillars.lean`). `manyToOneSetup U p₀` is a `KahlerOnticSetup` on `Σ = ℂℙ^{N-1} × T²` with the genuine many-to-one `π = Prod.fst` (`manyToOneSetup_pi_not_injective`) AND a non-trivial projected ray flow (`Φ_t` rotates the base by `U t`); both pillars fire on the `N = 2` rotation witness (`manyToOneRotationSetup_both_pillars`) — Schrödinger inherited from `rotationSetup_schrodinger_form`, Born via `manyToOneSetup_born_frequency` scoring the fibred region `π⁻¹'(bornRegion)`. AxiomAudit-pinned (foundational triple). Removes the `π = id` degeneracy; does NOT touch L7 (Born still samples `kMuL`; fibre flow trivial). **Phase 3 — the record layer (kinematic interface BUILT 2026-07-25; MD-1 CLOSED 2026-08-31, residuals `R-015` boundary / `R-016` open) + the extensions.** ⚠️ *Corrected 2026-07-28 (external review): general-`N` A7, the de-isolation `H_int`, and the A1-sector status of the fibred Σ are all open — see `BACKLOG.md` and `record-layer-plan.md`.* - **C6 (retired as stated).** "Derive the Born weights FROM the deterministic flow" is a **non-question** — Σ is the floor; Born *arises from* the Ω-region volume ratios + typicality (that is the reconstruction). The **record layer** is now **formalized in Lean** (`record-layer-plan.md §4`); its residual is context-fixing the *pinned closure's* regions (still the preparation-indexed `bornRegion ψ'`), and realizing **the record** as the ontic selection in Σ (which region the trajectory realizes) — the Ω-regions stay epistemic — `specs/CSD-CHARTER.md`, `future-work.md` (record layer / MD-1). Paper C Axiom A5 = projectability, not "the origin of Σ". **Gate:** a link's row flips to CONNECTED only when a named `sorry`-free, AxiomAudit-pinned theorem backs it. The README claim upgrades only then.