Documentation

TauCeti.NumberTheory.Multiquadratic.Quadratic.RamifiedPrime.Product

The relation between the ramified primes of a quadratic field #

Let K = ℚ(√d) be a quadratic number field, presented by θ : 𝓞 K with minpoly ℤ θ = X ^ 2 - d and Algebra.adjoin ℚ {θ} = ⊤, with d squarefree. Every rational prime p dividing d ramifies, so p 𝓞 K = 𝔭 ^ 2 for the unique prime 𝔭 of 𝓞 K above it (TauCeti.Multiquadratic.map_span_eq_sq_of_dvd_fundamentalDiscriminant). Squaring the principal ideal (θ) gives (d), which is the product of those p 𝓞 K; cancelling the square in the factorisation monoid of ideals leaves

(θ) = ∏_{p ∣ d} 𝔭_p.

This is an explicit relation of genus theory. The classes [𝔭_p] of the ramified primes are 2-torsion (NumberField.classGroupMk0_sq_eq_one_of_mem_ramifiedPrimes), so they span an elementary-2 subgroup of Cl(𝓞 K), a priori of order at most 2 ^ t with t the number of ramified primes (ncard_ramifiedPrimes_eq_card). The identity above says that the product of those spanning classes over p ∣ d is trivial, so they are generated by all but one prime factor of d and span at most 2 ^ (t - 1) classes. The matching lower bound is not proved here. In the 2-rank formula rank = t - 1 of the Multiquadratic roadmap this is where the - 1 comes from; for real quadratic fields that exact formula concerns the narrow class group, while this file only gives an upper bound in the ordinary class group.

Because the ramified primes are exactly those dividing fundamentalDiscriminant d, which is d or 4 * d, the product here runs over p ∣ d. It covers all ramified primes unless d ≡ 3 (mod 4), in which case exactly the ramified prime 2 is omitted. That is why the counting statement takes the ambient set of ramified primes as a parameter s and only asks that it contain the prime factors of d: the relation lives on the prime factors of the radicand, while the classes being counted are those of all the ramified primes.

The primes above the rational primes involved may be supplied by the caller as a family (P for the ideal identity, Q for the class-group statements). The theorem exists_span_singleton_eq_prod_primeFactors also chooses a family with the nonzero-divisor packaging needed by ClassGroup.mk0.

The classical source for the relation and its use in genus theory is D. A. Cox, Primes of the Form x² + ny², Chapter 3, and F. Lemmermeyer, Reciprocity Laws, Chapter 6.

Main results #

A prime factor of the radicand is a ramified prime of ℚ(√d).

theorem TauCeti.Multiquadratic.span_singleton_eq_prod_primeFactors {K : Type u_1} [Field K] [NumberField K] {θ : NumberField.RingOfIntegers K} {d : ℤ} (hmin : minpoly ℤ θ = Polynomial.X ^ 2 - Polynomial.C d) (hgen : ℚ[↑θ] = ⊤) (hsf : Squarefree d) (P : ℕ → Ideal (NumberField.RingOfIntegers K)) (hprime : ∀ p ∈ d.natAbs.primeFactors, (P p).IsPrime) (hover : ∀ p ∈ d.natAbs.primeFactors, (P p).LiesOver (Ideal.span {↑p})) :
Ideal.span {θ} = ∏ p ∈ d.natAbs.primeFactors, P p

The square root of the radicand generates the product of the ramified primes dividing it. Let K = ℚ(√d) with d squarefree, and for every prime factor p of d let P p be the prime of 𝓞 K above p. Then the principal ideal generated by θ is the product of those primes:

(θ) = ∏_{p ∣ d} 𝔭_p.

Squaring both sides gives (d) = ∏_p p 𝓞 K, which holds because p 𝓞 K = 𝔭_p ^ 2 at every ramified prime and d is squarefree; the ideals of a Dedekind domain form a factorisation monoid with no 2-torsion, so the squares may be cancelled.

theorem TauCeti.Multiquadratic.exists_span_singleton_eq_prod_primeFactors {K : Type u_1} [Field K] [NumberField K] {θ : NumberField.RingOfIntegers K} {d : ℤ} (hmin : minpoly ℤ θ = Polynomial.X ^ 2 - Polynomial.C d) (hgen : ℚ[↑θ] = ⊤) (hsf : Squarefree d) :
∃ (Q : ℕ → ↥(nonZeroDivisors (Ideal (NumberField.RingOfIntegers K)))), (∀ p ∈ d.natAbs.primeFactors, (↑(Q p)).IsPrime) ∧ (∀ p ∈ d.natAbs.primeFactors, (↑(Q p)).LiesOver (Ideal.span {↑p})) ∧ Ideal.span {θ} = ∏ p ∈ d.natAbs.primeFactors, ↑(Q p)

There is a family of prime ideals above the prime factors of d for which the square root relation holds. The ideals are packaged as non-zero-divisors, so the same family can be used in the class-group results below.

theorem TauCeti.Multiquadratic.prod_classGroupMk0_eq_one {K : Type u_1} [Field K] [NumberField K] {θ : NumberField.RingOfIntegers K} {d : ℤ} (Q : ℕ → ↥(nonZeroDivisors (Ideal (NumberField.RingOfIntegers K)))) (hmin : minpoly ℤ θ = Polynomial.X ^ 2 - Polynomial.C d) (hgen : ℚ[↑θ] = ⊤) (hsf : Squarefree d) (hprime : ∀ p ∈ d.natAbs.primeFactors, (↑(Q p)).IsPrime) (hover : ∀ p ∈ d.natAbs.primeFactors, (↑(Q p)).LiesOver (Ideal.span {↑p})) :
∏ p ∈ d.natAbs.primeFactors, ClassGroup.mk0 (Q p) = 1

An explicit relation of genus theory. In Cl(𝓞 K) for K = ℚ(√d) with d squarefree, the product of the classes of the primes above the prime factors of d is trivial: their product is the principal ideal (θ) by span_singleton_eq_prod_primeFactors.

Together with the fact that each such class is 2-torsion (NumberField.classGroupMk0_sq_eq_one_of_mem_ramifiedPrimes) this bounds the elementary-2 subgroup they span by 2 ^ (t - 1) elements.

theorem TauCeti.Multiquadratic.prod_classGroupMk0_eq_prod_sdiff {K : Type u_1} [Field K] [NumberField K] {θ : NumberField.RingOfIntegers K} {d : ℤ} (Q : ℕ → ↥(nonZeroDivisors (Ideal (NumberField.RingOfIntegers K)))) (hmin : minpoly ℤ θ = Polynomial.X ^ 2 - Polynomial.C d) (hgen : ℚ[↑θ] = ⊤) (hsf : Squarefree d) (hprime : ∀ p ∈ d.natAbs.primeFactors, (↑(Q p)).IsPrime) (hover : ∀ p ∈ d.natAbs.primeFactors, (↑(Q p)).LiesOver (Ideal.span {↑p})) {S : Finset ℕ} (hS : S ⊆ d.natAbs.primeFactors) :
∏ p ∈ S, ClassGroup.mk0 (Q p) = ∏ p ∈ d.natAbs.primeFactors \ S, ClassGroup.mk0 (Q p)

Complementary sub-products of the primes above the prime factors of d have the same class. For a subset S of the prime factors of d, the class of ∏_{p ∈ S} 𝔭_p equals the class of the complementary product ∏_{p ∉ S} 𝔭_p. This is prod_classGroupMk0_eq_one combined with the 2-torsion of the individual classes, and it is the concrete form of an explicit relation among the generators of the elementary-2 subgroup spanned by the ramified-prime classes.

theorem TauCeti.Multiquadratic.natCard_closure_image_classGroupMk0_le {K : Type u_1} [Field K] [NumberField K] {θ : NumberField.RingOfIntegers K} {d : ℤ} (Q : ℕ → ↥(nonZeroDivisors (Ideal (NumberField.RingOfIntegers K)))) (hmin : minpoly ℤ θ = Polynomial.X ^ 2 - Polynomial.C d) (hgen : ℚ[↑θ] = ⊤) (hsf : Squarefree d) {s : Finset ℕ} (hram : ∀ p ∈ s, p ∈ NumberField.ramifiedPrimes K) (hprime : ∀ p ∈ s, (↑(Q p)).IsPrime) (hover : ∀ p ∈ s, (↑(Q p)).LiesOver (Ideal.span {↑p})) (hs : d.natAbs.primeFactors ⊆ s) (hd : 1 < d.natAbs) :
Nat.card ↥(Subgroup.closure ((fun (p : ℕ) => ClassGroup.mk0 (Q p)) '' ↑s)) ≤ 2 ^ (s.card - 1)

The classes of the ramified primes generate a subgroup of order at most 2 ^ (t - 1). Let s be a finite set of ramified primes of K = ℚ(√d) containing every prime factor of d, with Q p the prime of 𝓞 K above p ∈ s. Then the subgroup of Cl(𝓞 K) generated by their classes has at most 2 ^ (#s - 1) elements.

The classes are involutions, so the subgroup they generate is the set of sub-products, of which there are at most 2 ^ #s; the relation prod_classGroupMk0_eq_one over the prime factors of d removes one generator (TauCeti.natCard_closure_image_le_two_pow_card_sub_one). The hypothesis 1 < |d| excludes d = -1, the one radicand with no prime factor and hence no relation.

This is only an upper bound in the ordinary class group; no matching lower bound is asserted.