Documentation

TauCeti.NumberTheory.NumberField.Cyclotomic.Finrank

The degree of a cyclotomic extension of a number field #

Mathlib computes the degree of an m-th cyclotomic extension either over (IsCyclotomicExtension.Rat.finrank) or over a base for which the m-th cyclotomic polynomial is already known to be irreducible (IsCyclotomicExtension.finrank). Neither is directly usable over a general number field K, where irreducibility of Φ_m over K is exactly what has to be established.

This file supplies the arithmetic criterion: if m is coprime to the discriminant of K, then

[M : K] = φ m for M / K an m-th cyclotomic extension with K a number field.

The mechanism is linear disjointness rather than a direct irreducibility argument. Inside M, the two subfields ℚ(ζ) and (the image of) K have coprime discriminants, so Mathlib's NumberField.linearDisjoint_of_isGalois_isCoprime_discr makes them linearly disjoint; their compositum is M, so the degree of M over K equals the degree of ℚ(ζ) over , which is φ m. Coprimality of the discriminants is where the hypothesis is spent, via the divisibility input below.

Main results #

Implementation notes #

The hypothesis is ((NumberField.discr K).natAbs).Coprime m, a statement about the base field. It is the condition an arithmetic caller can actually arrange — e.g. by choosing m to be a prime unramified in K — whereas the resulting intersection or irreducibility conditions would have to be re-derived from it at each use.

Only the base K carries a NumberField hypothesis. M is finite over K by IsCyclotomicExtension.finiteDimensional, hence a number field on its own, so demanding [NumberField M] of the caller would be an avoidable hypothesis. The degree itself comes from linear disjointness of ℚ(ζ) and the image of K inside M, whose compositum is M.

Adapted from the Birkbeck–Brasca Chebotarev density project.

A prime dividing the discriminant of an m-th cyclotomic extension of divides m.

The cyclotomic degree over a number field base. If M / K is an m-th cyclotomic extension with K a number field and m coprime to discr K, then [M : K] = φ m.

Coprimality to discr K stands in for irreducibility of Φ_m over K, and is the hypothesis an arithmetic caller can arrange directly. Only the base K need be a number field.