The two finite-place sign conventions in rational adelic Poisson summation and their agreement on even tests
ProvedMasulloRH.radial_adelic_poisson_sign_invarianceadelic-analysispoisson-summationriemann-hypothesis
Let on the rational adeles and , with the standard character trivial on diagonal rationals. The modified uses the same Fourier sign at the infinite and finite places as in the printed character. Its restriction to has kernel exactly . For every adelic Schwartz–Bruhat test and any additive Haar measure , the translated Poisson formula holds with and the reflected dual lattice , including the usual covolume factor. If , its entire Fourier transform is the same for both signs, and unshifted Poisson summation holds for on the ordinary diagonal. This includes tests invariant under the finite unit ideles.
Preamble
import Mathlib import Definitions.Def_NumberField_StandardGlobalAddCharRat import Definitions.Def_NumberField_AdelicBox import Definitions.Def_NumberField_AdelicFourier set_option autoImplicit false set_option backward.isDefEq.respectTransparency.types false
Formal statement
open NumberField NumberField.StandardAddChar NumberField.AdelicFourier MeasureTheory
open scoped NumberField.AdeleRing
namespace MasulloRH
theorem radial_adelic_poisson_sign_invariance [MeasurableSpace 𝔸[ℚ]] [BorelSpace 𝔸[ℚ]]
(μ : Measure 𝔸[ℚ]) [μ.IsAddHaarMeasure] (f : 𝔸[ℚ] → ℂ) (hf : f ∈ schwartzBruhat ℚ) :
let u : 𝔸[ℚ] := (1,-1)
let χ := psiQ.mulShift u
(∀ q : ℚ, χ (algebraMap ℚ 𝔸[ℚ] q) = 1 ↔ ∃ k : ℤ, 2*q = k) ∧
(∀ x : 𝔸[ℚ], (∑' q : ℚ, f (x + algebraMap ℚ 𝔸[ℚ] q)) =
((μ (AdelicBox.adelicBox ℚ)).toReal : ℂ)⁻¹ *
∑' q : ℚ, fourierIntegral χ μ f (u * algebraMap ℚ 𝔸[ℚ] q) *
χ ((u * algebraMap ℚ 𝔸[ℚ] q)*x)) ∧
((∀ x : 𝔸[ℚ], f (u*x) = f x) →
(∀ w : 𝔸[ℚ], fourierIntegral χ μ f w = fourierIntegral psiQ μ f w) ∧
(∑' q : ℚ, f (algebraMap ℚ 𝔸[ℚ] q)) =
((μ (AdelicBox.adelicBox ℚ)).toReal : ℂ)⁻¹ *
∑' q : ℚ, fourierIntegral χ μ f (algebraMap ℚ 𝔸[ℚ] q)) := by sorry
end MasulloRHSource
Masullo, Examples A.3.1.5–A.3.1.8 and Theorem 2.2.3.2. The opposite-sign issue at rational arguments was identified independently in F13 of the mccorvie/riemann-masullo errata, https://github.com/mccorvie/riemann-masullo/blob/main/ERRATA.md. The finite-sign-even sector and reflected-lattice formula are proved using the established adelic Poisson theorem 4340c7f0.