Documentation

TauCeti.NumberTheory.NumberField.WorkedExamples.GaussianRationals.Ramification.Group

The dyadic ramification groups of ℚ(i) #

For K generated over ℚ by an algebraic integer θ with minpoly ℤ θ = X² + 1, let 𝔭 be the prime of 𝓞 K above 2, and let G_i be its ramification groups in Gal(K/ℚ), the elements acting trivially on 𝓞 K ⧸ 𝔭 ^ (i + 1). This file computes the whole filtration:

G_0 = G_1 = Gal(K/ℚ) ≅ ℤ/2, and G_i = 1 for i ≥ 2.

Every automorphism sends θ to a square root of −1, so to θ or −θ (TauCeti.NumberField.smul_gen_eq_or_eq_neg), and 𝓞 K = ℤ[θ], so by Serre's criterion (TauCeti.Ideal.mem_inertia_iff_of_adjoin_singleton_eq_top) an automorphism σ lies in G_i exactly when σ θ − θ ∈ 𝔭 ^ (i + 1). For the conjugation σ θ − θ = −2θ generates (2) = 𝔭², which lies in 𝔭² but not in 𝔭³.

The two nontrivial groups G_0 and G_1 contribute 1 each to Hilbert's formula v_𝔭(𝔡) = Σ_{i ≥ 0} (#G_i − 1), which gives back the different exponent v_𝔭(𝔡) = 2 (TauCeti.NumberField.GaussianRationals.multiplicity_differentIdeal_eq_two). The jump of the filtration is at 1, not 0: the prime is wildly ramified, so G_1, a 2-group, is not trivial.

Main results #

References #

theorem TauCeti.NumberField.GaussianRationals.mem_ramificationGroup_iff {K : Type u_1} [Field K] [NumberField K] {θ : NumberField.RingOfIntegers K} (hmin : minpoly ℤ θ = Polynomial.X ^ 2 + 1) (hgen : ℚ[↑θ] = ⊤) (𝔭 : Ideal (NumberField.RingOfIntegers K)) [𝔭.IsPrime] [𝔭.LiesOver (Ideal.span {2})] {i : ℕ} {σ : Gal(K/ℚ)} :
σ ∈ Ideal.ramificationGroup Gal(K/ℚ) 𝔭 i ↔ σ = 1 ∨ i ≤ 1

The dyadic ramification filtration of ℚ(i). An automorphism σ lies in the i-th ramification group of the prime above 2 exactly when σ = 1 or i ≤ 1.

Use simp [mem_ramificationGroup_iff hmin hgen 𝔭] to simplify membership. The generator θ does not occur in the membership expression, so simp cannot infer it to apply this theorem as a global rule, regardless of priority.

The ramification groups G_0 and G_1 of the prime above 2 are the whole Galois group of ℚ(i).

The ramification groups G_i of the prime above 2 in ℚ(i) are trivial for i ≥ 2.

G_0 = G_1 ≅ ℤ/2: the ramification groups G_0 and G_1 of the prime above 2 in ℚ(i) have order 2.

Hilbert's different formula, checked in ℚ(i): the sum Σ_{i ≥ 0} (#G_i − 1) over the ramification groups of the prime above 2 is the different exponent v_𝔭(𝔡) = 2, as the general multiplicity_differentIdeal_eq_finsum_card_ramificationGroup_sub_one predicts.