Skip to content

Prove adjacent gamma interlacing transfer - #325

Open
PerAlexandersson wants to merge 21 commits into
mainfrom
codex-gamma315
Open

Prove adjacent gamma interlacing transfer#325
PerAlexandersson wants to merge 21 commits into
mainfrom
codex-gamma315

Conversation

@PerAlexandersson

@PerAlexandersson PerAlexandersson commented Aug 3, 2026

Copy link
Copy Markdown
Owner

Summary

Completes issue #315 along the exact proof of Hoster--Stump Proposition 2.5.

  • corrects the challenge endpoint to include the source's nonnegative gamma
    coefficient and nonzero exact-degree hypotheses;
  • formalizes equations (2.1)--(2.2), including reciprocal root pairs and the
    exceptional -1 multiplicity;
  • transports ordered preferred-root lists through
    gammaRootMap ρ = ρ / (1 + ρ)^2;
  • proves the reusable backend prec_gammaTransform_succ_iff;
  • replaces def GammaAdjacentInterlacingTransferStatement : Prop with a
    checked theorem wrapper;
  • removes the proved transfer theorem from StrategyInputs.

The endpoint derives the backend's nonzero constant coefficients from the
source's nonzero exact-degree hypotheses via
coeff_ambient_gammaTransform; it introduces no target-shaped assumption.

Source

E. Hoster and C. Stump, Chow polynomials of simplicial posets with positive
h-vector are real-rooted
, arXiv:2508.15538, Proposition 2.5, equations
(2.1)--(2.2) (paper-cache record 90).

Validation

  • focused RealRooted.Challenges.HosterStump build: 8620/8620 jobs;
  • full aggregate build: 8951/8951 jobs.

All reported warnings are pre-existing and confined to untouched modules.

Closes #315.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Source-shape follow-up: adding HasNonnegCoeffs gamma and HasNonnegCoeffs delta fixes the main Proposition 2.5 mismatch. For an exact source transcription, also exclude the Lean-only d = 0, f = 0 edge: the paper assumes deg f = d and deg g = d+1, so both polynomials are nonzero, whereas Lean has natDegree 0 = 0. Add f ≠ 0 (and preferably g ≠ 0 symmetrically/documentarily) or state explicitly that the theorem deliberately proves a harmless zero extension rather than claiming exact-degree hypotheses alone encode the source assumptions.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Commit a56cfe67 adds the exact -1 root-multiplicity identity from Hoster--Stump Proposition 2.5, equation (2.2), with a source link in the Lean declaration. It derives the identity directly from the source factorization and the proved nonvanishing of the minimal core. No broader adjacent-interlacing endpoint is asserted in this slice, and no local build was run; comparator CI remains the validation oracle.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Commit b3b2b2c6 adds the Proposition 2.5 ordered-root transport bridge. The recursive proof was checked in isolation with AXLE against Lean 4.31 and uses only strict monotonicity of gammaRootMap on (-1,0). No local Lake build or post-edit reread was run; comparator CI remains authoritative. Per the project cadence, the next continuation is cleanup/dedup/golf only.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Cleanup-only commit b545c654: deduplicated the base-case handling in interleaves_map_gammaRootMap_iff using the project shim's existing simp lemmas. Net change: 3 insertions, 6 deletions, with no API or mathematical change. Comparator CI remains authoritative.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Commit 332fd47d fixes the CI-confirmed forward reference by relocating interleaves_map_gammaRootMap_iff after its dependencies. It also adds the exact reciprocal quadratic factor and its composition with gammaTransform_X_sub_C_mul_two, following Hoster--Stump Proposition 2.5, equation (2.1). AXLE accepted the isolated new lemmas under Lean 4.31; comparator CI remains authoritative.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Commit 3a5451d8 adds the arbitrary-multiplicity iteration of the equation (2.1) quadratic factor and its reciprocal-pair specialization. This is the algebraic engine for the pending rootMultiplicity/multiset equality. AXLE accepted the isolated proofs under Lean 4.31. Per cadence, the next continuation is cleanup-only.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Cleanup-only commit 4a8403a0: retained the one-root API while deduplicating its proof through the m = 1 case of the arbitrary-multiplicity theorem. No mathematical or API change. Comparator CI remains authoritative.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Commit 93739cfb adds the exact equation (2.1) root-multiplicity theorem and repairs dependency order for the power-iteration block. The complete proof was accepted in isolation by AXLE under Lean 4.31. The remaining source packaging is the filtered root-multiset identity, then reciprocal and -1 reconstruction.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

New proof slice: exists_mem_Ioo_gammaRootMap_eq formalizes the inverse/surjectivity part of Hoster--Stump (2.1) on (-1,0) via IVT on x-y(1+x)^2. AXLE Lean 4.31 accepted the isolated proof (request b860c65c-bdc7-4708-9165-08ee260ec3eb). No local Lake build or post-edit validation was run; comparator CI is authoritative. Next continuation is cleanup-only by cadence; afterward the route is the filtered root-multiset identity and equation (2.2).

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Cleanup-only slice pushed: shortened exists_mem_Ioo_gammaRootMap_eq with no API, hypothesis, or mathematical change. AXLE accepted the golfed theorem (request 78958025-b0f1-49b6-8162-17634412a0b8); its automated simplifier found no additional safe cleanup (request 905bc05a-30f2-4801-92b7-46ef1d0aa7c8). No local Lake build or post-edit validation ran; comparator CI remains authoritative.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Compiled proof slice pushed. Added roots_eq_map_filter_roots_gammaTransform, the exact multiplicity-preserving root-multiset form of Hoster--Stump (2.1), and repaired two earlier proof defects exposed by local compilation. Verification: AXLE accepted the generic multiset argument (request c0c50105-6125-4e66-a2ba-38d7836fc729) and reciprocal repair (request cbec6cc6-2b03-490c-85fd-753100f7eb1f); the focused external-cache build lake ... build RealRooted.GammaRealRoots then completed successfully with no warnings in the touched module. Erik (sqrt-of-2) currently has no open PRs. Remaining endpoint: equation (2.2) root-list reconstruction and adjacent interlacing transfer.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Pushed a9d1073f, following Hoster--Stump Proposition 2.5, equations (2.1) and (2.2). This slice generalizes reciprocal-pair root multiplicity to every negative root other than -1 and proves roots_gammaTransform_eq_reciprocal_add_neg_one_add: the full transform root multiset is the reciprocal branch below -1, exactly d - 2 * γ.natDegree copies of -1, and the preferred branch in (-1,0). The focused external-cache build of RealRooted.GammaRealRoots passed. PR #322 still owns the challenge endpoint file, so this PR does not touch it.

@PerAlexandersson

PerAlexandersson commented Aug 4, 2026

Copy link
Copy Markdown
Owner Author

Mandatory cleanup continuation pushed. The proof of roots_gammaTransform_eq_reciprocal_add_neg_one_add now unfolds its local preferred-root multiset once and removes four duplicated full RHS expansions and terminal rfl steps, with no theorem-surface or mathematical changes. The focused RealRooted.GammaRealRoots build passed.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Pushed the next Proposition 2.5 proof slice. GammaRealRoots.lean now contains the ordered reciprocal/center/preferred root completion and two proved fixed-case equivalences matching Hoster--Stump after equations (2.1)--(2.2): equal gamma degrees add one central -1 on the right and reduce to ListAlternates; successive gamma degrees remove one central -1 on the right and reduce to ListInterlaces. The proof uses the existing List.Interleaves merged-chain API. The focused module build passed.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Pushed the reusable endpoint prec_gammaTransform_succ_iff, a complete formalization of Hoster--Stump Proposition 2.5 for adjacent gamma transforms. The backward direction derives the equal/successive gamma-degree split from the two-sided -1 root-multiplicity bound rootMultiplicity_bounds_of_prec; both directions canonicalize preferred roots, apply equation (2.1), reconstruct full roots via equation (2.2), and use the two proved center cases. The focused dependency build completed successfully: 8619/8619 jobs.

@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Mandatory cleanup continuation pushed. Removed the duplicated private ListInterlaces root-count proof block from GarloffWagner.lean; rootMultiplicity_sub_one_le_of_prec_right remains available unchanged as a one-line projection from the new generic rootMultiplicity_bounds_of_prec. Focused RealRooted.GarloffWagner build passed all 8618 jobs.

@PerAlexandersson
PerAlexandersson marked this pull request as ready for review August 4, 2026 05:21
@PerAlexandersson

Copy link
Copy Markdown
Owner Author

Completed endpoint pushed as e6c8320a after rebasing onto merged PR #322.

GammaAdjacentInterlacingTransferStatement is now a checked theorem, not a
def ... : Prop placeholder. Its constant-coefficient premises are derived
from nonzero exact degree and the supplied gamma expansions before applying
prec_gammaTransform_succ_iff. The redundant StrategyInputs assumption
was removed.

Validation passed:

  • focused challenge build: 8620/8620 jobs;
  • full aggregate build: 8951/8951 jobs.

@PerAlexandersson PerAlexandersson changed the title Correct adjacent gamma interlacing transfer hypotheses Prove adjacent gamma interlacing transfer Aug 4, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Prove adjacent-degree gamma interlacing transfer

1 participant