Documentation

LeanPool.LehmerE10.Mahler

the Mahler measure of Lehmer's polynomial, in Mathlib's own API. #

Mathlib now has the Mahler measure (Polynomial.mahlerMeasure, Jensen-formula definition, with mahlerMeasure_eq_leadingCoeff_mul_prod_roots), Kronecker's theorem in Mahler form, and Northcott's theorem. This file connects the repository to that API:

The Mahler measure of L is the Salem root μ. #

theorem LehmerE10.classified_prod_max (S : Multiset ) (hS : zS, z = 1 z = mu z = nu) :
(Multiset.map (fun (z : ) => max 1 z) S).prod = mu ^ Multiset.count (↑mu) S

The classified product with max 1 ‖·‖ weights: unit-circle roots contribute 1, ν < 1 contributes 1, and each occurrence of μ contributes μ.

The Mahler measure of Lehmer's polynomial is the Salem root: M(L) = μ. Computed exactly from mahlerMeasure_eq_leadingCoeff_mul_prod_roots and the root classification — this is the number ≈ 1.17628, the smallest known Mahler measure > 1 of an integer polynomial.

M(L) > 1: Lehmer's polynomial is not a product of cyclotomics, quantitatively.

theorem LehmerE10.s5_lt :
s5 < 7 / 10

The Salem square root is below 7/10: s₅² = y₅² − 4 < (21/10)² − 4 < (7/10)².

M(L) < 7/5: an unconditional upper bound from the trace-root interval. (The true value is 1.17628…; sharper bounds need only a tighter sign-change interval for y₅.)

Lehmer's question, stated in Mathlib's vocabulary. #

Both are hypothesis Props: stated, never claimed, never assumed.

Lehmer's conjecture (Lehmer 1933, gap form): Mahler measures of integer polynomials do not accumulate at 1 — there is a uniform gap c > 1 below which the only Mahler measure is 1 itself.

Equations
  • One or more equations did not get rendered due to their size.
Instances For

    The strong form: Lehmer's polynomial attains the minimal Mahler measure > 1.

    Equations
    • One or more equations did not get rendered due to their size.
    Instances For

      The strong form implies the gap form, with the gap witnessed by M(L) itself — legitimate because M(L) > 1 is proven above.

      The Coxeter element has infinite order. #

      The E₁₀ Coxeter element has infinite order. If Cox^n = 1, then since its characteristic polynomial is the irreducible L, the minimal polynomial is L and divides Xⁿ − 1; evaluating at the Salem root μ > 1 gives μⁿ = 1, impossible.

      Contrast: the Coxeter element of a finite Weyl group has order h (the Coxeter number — 30 for E₈); at E₁₀ the spectrum has left the unit circle and no power returns.