Cofinal actual-quartet windows with all-accuracy finite selected certificates
OpenConnesWeilArithmetic.cofinal_quartet_finite_certificatesFor every actual critical-strip zeta zero ρ and every real lower support threshold B, choose one positive radius T strictly greater than B. At that SAME T, for every positive accuracy δ there is a finite reflection-closed set F of actual zeros containing the original quartet of ρ, whose complete omitted positive-column squared-norm tail is strictly less than δ and whose original selected coefficient correction is nonnegative. The quartet and selected negative synthesis are unchanged as F varies; the regularized inverse and full-tail allowance use the same δ. T is chosen before δ. No RH, simple-zero, independent-column, positivity, or separation hypothesis is assumed. This is an unproved research goal, not a certified inequality. Its exact equivalence with Mathlib RH has been checked using the existing certified cofinal-window and finite-certificate reductions.
import Definitions.Def_ConnesGreen_finite_selected_correction import Definitions.Def_ConnesGreen_original_quartet set_option autoImplicit false open Complex ConnesRZ ConnesRZFrontier ConnesRZQuartet ConnesGreen open WeilDefect WeilDefect.ConnesNative open scoped BigOperators InnerProductSpace lp ENNReal Classical ComplexOrder Topology noncomputable section
theorem ConnesWeilArithmetic.cofinal_quartet_finite_certificates :
∀ ρ : CriticalZeros, ∀ B : ℝ,
∃ T : ℝ, ∃ hT : 0 < T, B < T ∧
∀ δ : ℝ, 0 < δ → ∃ F : Finset CriticalZeros,
quartet ρ ⊆ F ∧ (∀ τ ∈ F, reflectedZero τ ∈ F) ∧
(∑' τ : {τ : CriticalZeros // τ ∉ F},
‖positiveGreenColumn (fun ζ => sourceEmbed T (actualGreenSource ζ)) τ.1‖ ^ 2) < δ ∧
0 ≤ canonicalFiniteSelectedCorrection T hT (quartet ρ) F δ := by sorry
Read-back
What the Lean code literally says, in plain math · GPT-6
For every complex number with and , and for every real number , there exist a real number and a witness of such that and the following holds for every real : there exists a finite subset of satisfying , , , and . Here is reflection, is a finite set with repeated values removed, and every index denotes a distinct zero value together with its strip-and-zero property, rather than a separate index for each occurrence of its multiplicity. The finite set need only be closed under , not under conjugation; it contains the whole quartet. To define all the columns and the operator in this assertion, let be Lebesgue measure restricted to , let , and give its usual Hilbert direct-sum norm. For any function , write for its equivalence class if it is almost everywhere strongly measurable and has finite seminorm with respect to , and for the zero vector otherwise. Define , where the derivative is the total derivative used in the definitions, and let be the closure in of the complex linear span of all for smooth, compactly supported whose topological support is contained in the open interval . This closed subspace, with the inner product inherited from , is the physical carrier of every column below. Let be the orthogonal projection and let , where . This source is exactly the one obtained in the definitions by setting and using . Write for the natural-number vanishing order, which is the natural-number conversion of the extended analytic order (its total definition returns zero for infinite order or failure of analyticity), and put , with the real square root acting as a complex scalar. The positive and negative pair columns are exactly and in . For any finite set , let with its ordinary coefficient inner product, one coordinate for each zero value in ; define by and by . With Hilbert-space adjoints taken using these coefficient and physical inner products, the asserted correction is the bounded complex-linear endomorphism of
where is the ring inverse, equal to the two-sided inverse when its argument is a unit and zero otherwise; in the finite coefficient space displayed here, makes invertible. The inequality means that is self-adjoint and for every , including zero. The tail sum uses all zero values outside , their physical column norms, and no extra multiplicity factor beyond the square-root weights already in the columns; the definitions establish summability of these squared norms for positive , so it is the usual convergent sum. The single may depend on and but is chosen before and works for every positive ; may depend on all of , with no nesting or monotonicity requirement and no upper bound on its size. No assumption places off the line , assumes a simple zero, or requires a nonzero imaginary part: coincidences within the quartet are permitted, a reflection-fixed zero has , and linearly dependent or zero columns retain their original coefficient coordinates. The quartet always contains , hence and the required are nonempty, although the tail index set can be empty, in which case its sum is zero. There is no assertion at or , and the universal clause over would be vacuous if were empty.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.