Schoenberg's theorem: Fourier transform of a Pólya frequency function has only real zeros
OpenDeBruijnNewman.schoenberg_polya_fourier_real_zerosSchoenberg's theorem (I. J. Schoenberg, 1947, "On totally positive functions, distributions and Fourier transforms"): the Fourier transform of a Pólya frequency function has only real zeros. Precisely: if K : ℝ → ℝ is a Pólya frequency function (the translation kernel (x,y) ↦ K(x−y) is totally positive of every order), K is integrable, and K is not identically zero, then the entire function F(z) = ∫ K(t) e^{izt} dt has no non-real zeros.
This is the analytic heart of de Bruijn's base case (N. G. de Bruijn, "The roots of trigonometric integrals", Duke Math. J. 1950). De Bruijn's proof that H_{1/2} has only real zeros factors through exactly this theorem: the hard analytic step (de Bruijn's kernel lemma, a separate node) is that the kernel defining H_{1/2} — u ↦ e^{u²/2}Φ(u) — is a Pólya frequency function; Schoenberg's theorem then turns that total positivity into real-rootedness of its Fourier transform, i.e. of H_{1/2} (DeBruijnNewman.debruijn_base_case, 9d8af170-793b-4eb3-89af-f0602096501f). Combined with forward heat-flow monotonicity (the de Bruijn–Newman ray structure), this single implication yields H_t real-rooted for every t ≥ 1/2, fixing the right endpoint Λ ≤ 1/2 for the Rodgers–Tao argument.
import Mathlib
namespace DeBruijnNewman
/-- A function `K : ℝ → ℝ` is a Pólya frequency function if the translation
kernel `(x, y) ↦ K (x - y)` is totally positive of every order: every
finite minor formed on strictly increasing nodes is nonnegative. -/
def IsPolyaFrequency (K : ℝ → ℝ) : Prop :=
∀ (n : ℕ) (x y : Fin n → ℝ),
(∀ i j : Fin n, i < j → x i < x j) →
(∀ i j : Fin n, i < j → y i < y j) →
0 ≤ (Matrix.of fun i j => K (x i - y j)).det
end DeBruijnNewman
namespace DeBruijnNewman
theorem schoenberg_polya_fourier_real_zeros
(K : ℝ → ℝ) (hK : IsPolyaFrequency K)
(hKint : MeasureTheory.Integrable K MeasureTheory.volume)
(hne : ∃ t : ℝ, K t ≠ 0) :
∀ z : ℂ,
MeasureTheory.integral MeasureTheory.volume
(fun t => (K t : ℂ) * Complex.exp (Complex.I * z * (t : ℂ))) = 0 →
z.im = 0 := by
sorry
end DeBruijnNewman