Documentation

TauCeti.NumberTheory.ModularForms.Conjugate

The conjugate form f_ρ(τ) = conj (f (-conj τ)) #

The conjugate of a modular form f is f_ρ(τ) = conj (f (-conj τ)). It is holomorphic, and if f = ∑ aₙ qⁿ then f_ρ = ∑ conj (aₙ) qⁿ. Mathlib's slash action of GL(2, ℝ) builds complex conjugation into the action of a matrix of negative determinant, so f_ρ = f ∣[k] J for the reflection J = !![-1, 0; 0, 1] (UpperHalfPlane.J), and Mathlib's ModularForm.translate makes f_ρ a modular form for J⁻¹ 𝒢 J. Conjugation by J changes the signs of the off-diagonal entries, so it preserves Γ₀(N) and Γ₁(N).

On Γ₁(N) the conjugate form carries the nebentypus χ to its complex conjugate χ⁻¹, since the diamond eigenvalues are roots of unity, and it intertwines the Hecke operators Tₙ: their coefficient formula has real coefficients apart from the values of χ. So the conjugate of a Hecke eigenform of nebentypus χ is a Hecke eigenform of nebentypus χ⁻¹ with the complex-conjugate eigenvalues. This is the form that the Fricke involution relates a newform to: for a newform f of level N, f ∣ W_N is a multiple of f_ρ (Miyake, Theorem 4.6.15), the multiple being the Atkin–Li pseudo-eigenvalue.

Main definitions #

Main results #

References #

@[simp]

The matrix J = !![-1, 0; 0, 1] is its own inverse.

Slashing by J conjugates the value at the reflected point J • τ = -conj τ, in every weight.

The q-parameter at the reflected point J • τ = -conj τ is the conjugate of the q-parameter at τ.

Membership in J⁻¹ 𝒢 J, spelled out as a conjugation condition.

A subgroup contained in its conjugate J⁻¹ 𝒢 J by the involution J is equal to it.

A strict period of 𝒢 is a strict period of J⁻¹ 𝒢 J: conjugating !![1, h; 0, 1] by J gives its inverse !![1, -h; 0, 1].

Slashing by J conjugates the q-expansion coefficients. For a modular form f of weight k on 𝒢 and a strict period h of 𝒢, the q-expansion of f ∣[k] J = conj ∘ f ∘ (τ ↦ -conj τ) is that of f with every coefficient conjugated.

The congruence subgroups #

The J-conjugate !![a, -b; -c, d] of !![a, b; c, d] ∈ SL(2, ℤ).

Equations
  • γ.conjJ = ⟨!![↑γ 0 0, -↑γ 0 1; -↑γ 1 0, ↑γ 1 1], ⋯⟩
Instances For
    @[simp]
    theorem Matrix.SpecialLinearGroup.coe_conjJ (γ : SpecialLinearGroup (Fin 2) ℤ) :
    ↑γ.conjJ = !![↑γ 0 0, -↑γ 0 1; -↑γ 1 0, ↑γ 1 1]

    conjJ leaves the lower-right entry, hence the diamond label, alone.

    Nebentypus #

    noncomputable def ModularForm.conj {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : ModularForm 𝒢 k) :
    ModularForm 𝒢' k

    The conjugate form f_ρ(τ) = conj (f (-conj τ)) of a modular form f for 𝒢, as a modular form for any 𝒢' conjugated into 𝒢 by J = !![-1, 0; 0, 1]. It is the slash of f by the determinant -1 matrix J, so its q-expansion has the complex-conjugate coefficients (ModularForm.qExpansion_conj).

    Equations
    Instances For
      theorem ModularForm.coe_conj {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : ModularForm 𝒢 k) :
      @[simp]
      theorem ModularForm.conj_apply {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : ModularForm 𝒢 k) (τ : UpperHalfPlane) :
      (conj hJ f) τ = (starRingEnd ℂ) (f (UpperHalfPlane.J • τ))
      @[simp]
      theorem ModularForm.conj_conj {𝒢 : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢 ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : ModularForm 𝒢 k) :
      conj hJ (conj hJ f) = f

      The conjugate form is an involution.

      @[simp]
      theorem ModularForm.conj_zero {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) :
      conj hJ 0 = 0
      @[simp]
      theorem ModularForm.conj_add {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f g : ModularForm 𝒢 k) :
      conj hJ (f + g) = conj hJ f + conj hJ g
      @[simp]
      theorem ModularForm.conj_smul {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} [𝒢.HasDetOne] [𝒢'.HasDetOne] (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (c : ℂ) (f : ModularForm 𝒢 k) :
      conj hJ (c • f) = (starRingEnd ℂ) c • conj hJ f
      noncomputable def ModularForm.conjₗ {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} [𝒢.HasDetOne] [𝒢'.HasDetOne] (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) :

      The conjugate form, as a ℂ-antilinear map.

      Equations
      Instances For
        @[simp]
        theorem ModularForm.conjₗ_apply {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} [𝒢.HasDetOne] [𝒢'.HasDetOne] (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : ModularForm 𝒢 k) :
        (conjₗ hJ) f = conj hJ f
        theorem ModularForm.qExpansion_conj {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} {h : ℝ} (hh : 0 < h) (hΓ : h ∈ 𝒢.strictPeriods) (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : ModularForm 𝒢 k) :

        The q-expansion of the conjugate form has the conjugate coefficients: a_n(f_ρ) = conj (a_n(f)).

        The conjugate of a form of nebentypus χ has nebentypus χ⁻¹: f_ρ ∈ M_k(N, χ⁻¹).

        noncomputable def CuspForm.conj {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : CuspForm 𝒢 k) :
        CuspForm 𝒢' k

        The conjugate cusp form f_ρ(τ) = conj (f (-conj τ)), the cusp-form counterpart of ModularForm.conj.

        Equations
        Instances For
          theorem CuspForm.coe_conj {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : CuspForm 𝒢 k) :
          @[simp]
          theorem CuspForm.conj_apply {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : CuspForm 𝒢 k) (τ : UpperHalfPlane) :
          (conj hJ f) τ = (starRingEnd ℂ) (f (UpperHalfPlane.J • τ))
          @[simp]

          The conjugate cusp form is the conjugate of the underlying modular form.

          @[simp]
          theorem CuspForm.conj_conj {𝒢 : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢 ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : CuspForm 𝒢 k) :
          conj hJ (conj hJ f) = f

          The conjugate cusp form is an involution.

          @[simp]
          theorem CuspForm.conj_zero {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) :
          conj hJ 0 = 0
          @[simp]
          theorem CuspForm.conj_add {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f g : CuspForm 𝒢 k) :
          conj hJ (f + g) = conj hJ f + conj hJ g
          @[simp]
          theorem CuspForm.conj_smul {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} [𝒢.HasDetOne] [𝒢'.HasDetOne] (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (c : ℂ) (f : CuspForm 𝒢 k) :
          conj hJ (c • f) = (starRingEnd ℂ) c • conj hJ f
          noncomputable def CuspForm.conjₗ {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} [𝒢.HasDetOne] [𝒢'.HasDetOne] (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) :

          The conjugate cusp form, as a ℂ-antilinear map.

          Equations
          Instances For
            @[simp]
            theorem CuspForm.conjₗ_apply {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} [𝒢.HasDetOne] [𝒢'.HasDetOne] (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : CuspForm 𝒢 k) :
            (conjₗ hJ) f = conj hJ f
            theorem CuspForm.qExpansion_conj {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} {h : ℝ} (hh : 0 < h) (hΓ : h ∈ 𝒢.strictPeriods) (hJ : 𝒢' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (f : CuspForm 𝒢 k) :

            The q-expansion of the conjugate cusp form has the conjugate coefficients.

            theorem CuspForm.conj_levelRaise {𝒢 𝒢' : Subgroup (GL (Fin 2) ℝ)} {k : ℤ} {𝒢₁ 𝒢₁' : Subgroup (GL (Fin 2) ℝ)} [𝒢'.HasDetOne] [𝒢₁'.HasDetOne] {d : ℕ} [NeZero d] (h : 𝒢' ≤ ConjAct.toConjAct (TauCeti.scaleGL d)⁻¹ • 𝒢) (hJ : 𝒢₁' ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢') (hJ' : 𝒢₁ ≤ ConjAct.toConjAct UpperHalfPlane.J⁻¹ • 𝒢) (h' : 𝒢₁' ≤ ConjAct.toConjAct (TauCeti.scaleGL d)⁻¹ • 𝒢₁) (f : CuspForm 𝒢 k) :

            Conjugation commutes with the level-raising operators: (V_d f)_ρ = V_d (f_ρ), since the reflection τ ↦ -conj τ commutes with τ ↦ d τ.

            Nebentypus and Hecke operators #

            The conjugate of a cusp form of nebentypus χ has nebentypus χ⁻¹: f_ρ ∈ S_k(N, χ⁻¹).

            noncomputable def CuspForm.conjCharSpace (k : ℤ) {N : ℕ} (χ : (ZMod N)ˣ →* ℂˣ) :

            The conjugate cusp form as an antilinear equivalence S_k(N, χ) ≃ S_k(N, χ⁻¹). Its inverse is again f ↦ f_ρ, now on S_k(N, χ⁻¹).

            Equations
            • One or more equations did not get rendered due to their size.
            Instances For
              @[simp]
              theorem CuspForm.coe_conjCharSpace_apply {k : ℤ} {N : ℕ} {χ : (ZMod N)ˣ →* ℂˣ} (f : ↥(cuspFormCharSpace k χ)) :
              ↑((conjCharSpace k χ) f) = conj ⋯ ↑f
              @[simp]
              theorem CuspForm.coe_conjCharSpace_symm_apply {k : ℤ} {N : ℕ} {χ : (ZMod N)ˣ →* ℂˣ} (f : ↥(cuspFormCharSpace k χ⁻¹)) :
              ↑((conjCharSpace k χ).symm f) = conj ⋯ ↑f
              theorem CuspForm.coe_conjCharSpace_conjCharSpace {k : ℤ} {N : ℕ} {χ : (ZMod N)ˣ →* ℂˣ} (f : ↥(cuspFormCharSpace k χ)) :
              ↑((conjCharSpace k χ⁻¹) ((conjCharSpace k χ) f)) = ↑f

              Conjugating twice, first on S_k(N, χ) and then on S_k(N, χ⁻¹), gives back the original form.

              The conjugate form intertwines the Hecke operators: for f ∈ S_k(N, χ) and n ≠ 0, T_n (f_ρ) = (T_n f)_ρ, where on the left T_n acts on S_k(N, χ⁻¹). Since the conjugation is antilinear, the conjugate of a T_n-eigenform with eigenvalue λ is a T_n-eigenform with eigenvalue conj λ.

              The conjugate of a Hecke eigenform is an eigenform with the conjugate eigenvalue: if Tₙ f = λ f on S_k(N, χ), then Tₙ f_ρ = conj λ • f_ρ on S_k(N, χ⁻¹).