Documentation

Mathlib.NumberTheory.NumberField.FinitePlaces

Finite places of number fields #

This file defines finite places of a number field K as absolute values coming from an embedding into a completion of K associated to a non-zero prime ideal of š“ž K.

Main Definitions and Results #

Tags #

number field, places, finite places

The norm of a maximal ideal as an element of ā„ā‰„0 is > 1

The norm of a maximal ideal as an element of ā„ā‰„0 is ≠ 0

The v-adic absolute value on K defined as the norm of v raised to negative v-adic valuation

Equations
Instances For
    Equations
    • One or more equations did not get rendered due to their size.
    @[deprecated Valuation.instRankOneCompletion (since := "2026-01-05")]
    Equations

    A finite place of a number field K is a place associated to an embedding into a completion with respect to a maximal ideal.

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

      Return the finite place defined by a maximal ideal v.

      Equations
      Instances For

        A predicate singling out finite places among the absolute values on a number field K.

        Equations
        Instances For
          theorem NumberField.isFinitePlace_iff {K : Type u_1} [Field K] [NumberField K] (v : AbsoluteValue K ā„) :
          IsFinitePlace v ↔ ∃ (w : FinitePlace K), ↑w = v

          The norm of the image after the embedding associated to v is equal to the v-adic absolute value.

          The norm of the image after the embedding associated to v is equal to the norm of v raised to the power of the v-adic valuation.

          The norm of the image after the embedding associated to v is equal to the norm of v raised to the power of the v-adic valuation for integers.

          The v-adic absolute value satisfies the ultrametric inequality.

          The v-adic absolute value of a natural number is ≤ 1.

          The v-adic absolute value of an integer is ≤ 1.

          The v-adic norm of an integer is 1 if and only if it is not in the ideal.

          The v-adic norm of an integer is less than 1 if and only if it is in the ideal.

          Equations
          theorem NumberField.FinitePlace.coe_apply {K : Type u_1} [Field K] [NumberField K] (v : FinitePlace K) (x : K) :
          v x = ↑v x

          For a finite place w, return a maximal ideal v such that w = finite_place v .

          Equations
          Instances For
            theorem NumberField.FinitePlace.pos_iff {K : Type u_1} [Field K] [NumberField K] {w : FinitePlace K} {x : K} :
            0 < w x ↔ x ≠ 0
            @[simp]
            theorem NumberField.FinitePlace.mk_eq_iff {K : Type u_1} [Field K] [NumberField K] {v₁ vā‚‚ : IsDedekindDomain.HeightOneSpectrum (RingOfIntegers K)} :
            mk v₁ = mk vā‚‚ ↔ v₁ = vā‚‚

            The equivalence between finite places and maximal ideals.

            Equations
            Instances For
              theorem NumberField.FinitePlace.maximalIdeal_inj {K : Type u_1} [Field K] [NumberField K] (w₁ wā‚‚ : FinitePlace K) :
              w₁.maximalIdeal = wā‚‚.maximalIdeal ↔ w₁ = wā‚‚
              theorem NumberField.FinitePlace.mulSupport_finite_int {K : Type u_1} [Field K] [NumberField K] {x : RingOfIntegers K} (h_x_nezero : x ≠ 0) :
              (Function.mulSupport fun (w : FinitePlace K) => w ↑x).Finite
              theorem NumberField.FinitePlace.mulSupport_finite {K : Type u_1} [Field K] [NumberField K] {x : K} (h_x_nezero : x ≠ 0) :
              theorem NumberField.FinitePlace.add_le {K : Type u_1} [Field K] [NumberField K] (v : FinitePlace K) (x y : K) :
              v (x + y) ≤ max (v x) (v y)