Spelled out at the level of eigenvalues: the empirical spectral distribution of
ProvedRademacherWigner.esd_secondMomentaether-catalogprobability
Spelled out at the level of eigenvalues: the empirical spectral distribution of
W/√N has second moment exactly 1 - 1/N.
theorem RademacherWigner.esd_secondMoment(g : Config N) (hN : 0 < N) :
(1 / (N : ℝ)) *
∑ i, ((W_isHermitian g).eigenvalues i / Real.sqrt (N : ℝ)) ^ 2 = 1 - 1 / (N : ℝ) := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Probability/WignerSemicircleLawLowOrder.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Probability/WignerSemicircleLawLowOrder.lean
import Mathlib
import Definitions.Def_Probability_WignerRademacherEnsemble
import Theorems.Thm_RademacherWigner_W_isHermitian
/-
Copyright (c) 2026 Harmonic. All rights reserved.
Released under Apache 2.0 license.
# The semicircle law at orders two and four for the Rademacher Wigner ensemble
Combining
* `Probability.WignerSemicircleMoments` (moments of the semicircle law are Catalan
numbers),
* `Probability.WignerTraceBridge` (empirical spectral moments = normalised traces),
* `Probability.WignerRademacherEnsemble` (exact trace moments of the ensemble),
this file proves the moment-method form of the Wigner semicircle law at the first
two nontrivial orders:
* `esd_secondMoment` : the second moment of the empirical spectral distribution of
`W/√N` equals `1 - 1/N` for **every** realisation (perfect self-averaging), and
converges to `∫ x² dsc(x) = C₁ = 1`;
* `expect_normalizedMoment_four` : the expected fourth moment equals
`(N-1)(2N-3)/N²`, converging to `∫ x⁴ dsc(x) = C₂ = 2`;
* `eventually_secondMoment_close` : a (deterministic, hence in-probability)
concentration statement for the second moment.
-/
open Matrix BigOperators Filter Topology
open RademacherWigner
variable {N : ℕ}Formal statement
theorem RademacherWigner.esd_secondMoment(g : Config N) (hN : 0 < N) :
(1 / (N : ℝ)) *
∑ i, ((W_isHermitian g).eigenvalues i / Real.sqrt (N : ℝ)) ^ 2 = 1 - 1 / (N : ℝ) := by sorrySource