Documentation

TauCeti.AlgebraicGeometry.ModularCurve.Rigidifier.FreeAction

Free actions on simultaneous moduli schemes #

Let 𝒫 and 𝒬 be moduli problems over Ell/B, and let a group G act on 𝒬. If an object V = (E_V/𝕄) of Ell/B represents the simultaneous problem (𝒫, 𝒬), then G acts on V through its action on the second factor (ModuliProblem.simultaneousAction): by the Yoneda lemma (CategoryTheory.Functor.RepresentableBy.autMulEquiv), g acts as the automorphism Οƒ_g of V such that pulling back the universal pair (Ξ±, Ξ²) along Οƒ_g gives (Ξ±, g β€’ Ξ²).

In Katz and Mazur's proof that a relatively representable, affine, rigid moduli problem 𝒫 is representable (KM 4.7.0), 𝒬 is a rigidifying problem, for instance the naive full level-three problem [Ξ“(3)] with its action of GLβ‚‚(𝔽₃), and 𝒫 is recovered as the quotient of the moduli scheme 𝕄(𝒫, 𝒬) = V.base by G. The key input is that this action is free: if 𝒫 is rigid and G acts freely on 𝒬 over nonempty bases, then for g β‰  1 the automorphism of 𝕄(𝒫, 𝒬) induced by Οƒ_g fixes no point with values in a nonempty scheme (ModuliProblem.IsRigid.eq_one_of_comp_simultaneousAction_base_eq). Indeed, a point x : T ⟢ 𝕄(𝒫, 𝒬) fixed by g gives an automorphism Ο„ of the pulled-back curve E_T over T which carries the pulled-back structure (Ξ±_T, Ξ²_T) to (Ξ±_T, g β€’ Ξ²_T). Rigidity of 𝒫 forces Ο„ = 1, so g β€’ Ξ²_T = Ξ²_T, and freeness of the action on 𝒬 gives g = 1.

Rigidity of 𝒫 cannot be dropped. For the trivial problem 𝒫 = [Ξ“(1)] and 𝒬 = [Ξ“(3)], the automorphism [-1] of every elliptic curve carries a level-three basis (P, Q) to (-P, -Q), so the matrix -1 of GLβ‚‚(𝔽₃) acts trivially on 𝕄([Ξ“(3)]), although it acts freely on level-three bases.

When 𝕄(𝒫, 𝒬) is affine, freeness says equivalently that for g β‰  1 the elements Οƒ_g^*(a) - a generate the unit ideal of its ring of functions (ModuliProblem.IsRigid.span_range_appTop_simultaneousAction_sub_eq_top, through AlgebraicGeometry.Scheme.Hom.span_range_appTop_sub_eq_top_iff). Pullback reverses composition, so a group G acts on this ring by g β€’ a = Οƒ_{g⁻¹}^*(a) (ModuliProblem.simultaneousSectionsAction), and for g β‰  1 the elements g β€’ a - a generate the unit ideal (ModuliProblem.IsRigid.span_range_smul_sub_eq_top). For a finite group this is the hypothesis under which the projection to the invariant quotient is a finite Γ©tale torsor (TauCeti.AffineInvariantQuotient.etale_projection, TauCeti.AffineInvariantQuotient.isPullback_of_free).

For the naive full level-N problem [Ξ“(N)], on which GLβ‚‚(β„€/Nβ„€) acts freely over nonempty bases (ModuliProblem.eq_one_of_gammaNaiveAction_hom_app_eq_self), this gives the freeness of the action of GLβ‚‚(β„€/Nβ„€) on 𝕄(𝒫, [Ξ“(N)]) for every rigid 𝒫 (ModuliProblem.IsRigid.eq_one_of_comp_simultaneousAction_gammaNaiveAction_base_eq), and its form for functions on an affine 𝕄(𝒫, [Ξ“(N)]) (ModuliProblem.IsRigid.span_range_gammaNaiveAction_smul_sub_eq_top).

Main definitions #

Main results #

References #

The action on the object representing a simultaneous problem. If G acts on 𝒬 and V represents the simultaneous problem (𝒫, 𝒬), then G acts on V through its action on the second factor: g acts as the automorphism Οƒ_g of V along which the universal pair (Ξ±, Ξ²) pulls back to (Ξ±, g β€’ Ξ²) (homEquiv_comp_simultaneousAction_hom).

Instances For

    The automorphism Οƒ_g of the object V representing (𝒫, 𝒬) acts on the pair of level structures classified by an arrow f : X ⟢ V through the action of g on the second factor.

    The inverse of the automorphism Οƒ_g of the object V representing (𝒫, 𝒬) acts on the pair of level structures classified by an arrow f : X ⟢ V through the inverse of the action of g on the second factor.

    A rigid problem makes a free action on a rigidifier free on the simultaneous moduli scheme (the freeness step in Katz and Mazur's proof of Theorem 4.7.0). Let 𝒫 be rigid, let G act on 𝒬 freely over nonempty bases, and let V represent (𝒫, 𝒬). If a point x : T ⟢ V.base with values in a nonempty scheme T is fixed by the automorphism of 𝕄(𝒫, 𝒬) = V.base induced by g, then g = 1.

    Freeness on an affine simultaneous moduli scheme, in terms of functions. Let 𝒫 be rigid, let G act on 𝒬 freely over nonempty bases, and let V represent (𝒫, 𝒬) with 𝕄(𝒫, 𝒬) = V.base affine. For g β‰  1, the elements Οƒ_g^*(a) - a of Ξ“(𝕄(𝒫, 𝒬), ⊀) generate the unit ideal. For a group, IsRigid.span_range_smul_sub_eq_top restates this for the action g β€’ a = Οƒ_{g⁻¹}^*(a), in the form taken by the invariant quotient by a finite group (TauCeti.AffineInvariantQuotient.etale_projection).

    GLβ‚‚(β„€/Nβ„€) acts freely on 𝕄(𝒫, [Ξ“(N)]) for every rigid 𝒫. Let V represent the simultaneous problem (𝒫, [Ξ“(N)]) with the naive full level-N problem, on which GLβ‚‚(β„€/Nβ„€) acts by change of basis. If a point of V.base with values in a nonempty scheme is fixed by a matrix g, then g = 1. For N = 3 this is the freeness of the action of GLβ‚‚(𝔽₃) on the moduli scheme of 𝒫 rigidified by naive level-three structures.

    @[instance_reducible]

    The action on the functions of the simultaneous moduli scheme. If a group G acts on 𝒬 and V represents (𝒫, 𝒬), then G acts on the ring of functions Ξ“(𝕄(𝒫, 𝒬), ⊀) by pullback along inverses, g β€’ a = Οƒ_{g⁻¹}^*(a) (simultaneousSectionsAction_smul_def). For finite G and affine 𝕄(𝒫, 𝒬), TauCeti.AffineInvariantQuotient forms the quotient by this action.

    Instances For
      @[simp]

      Under simultaneousSectionsAction R ρ, g acts on a function on 𝕄(𝒫, 𝒬) by pullback along the automorphism of 𝕄(𝒫, 𝒬) induced by Οƒ_{g⁻¹}.

      Freeness on an affine simultaneous moduli scheme, for the action on functions. Let 𝒫 be rigid, let a group G act on 𝒬 freely over nonempty bases, and let V represent (𝒫, 𝒬) with 𝕄(𝒫, 𝒬) = V.base affine. For g β‰  1, the elements g β€’ a - a of Ξ“(𝕄(𝒫, 𝒬), ⊀) generate the unit ideal, for the action simultaneousSectionsAction. For finite G this is the freeness hypothesis of TauCeti.AffineInvariantQuotient.etale_projection and TauCeti.AffineInvariantQuotient.isPullback_of_free.

      GLβ‚‚(β„€/Nβ„€) acts freely on the functions of an affine 𝕄(𝒫, [Ξ“(N)]) for every rigid 𝒫. Let V represent the simultaneous problem (𝒫, [Ξ“(N)]) with V.base affine. For a matrix g β‰  1, the elements g β€’ a - a of Ξ“(V.base, ⊀) generate the unit ideal, for the action simultaneousSectionsAction. For N = 3 this is the freeness hypothesis of the invariant quotient of the moduli scheme of 𝒫 rigidified by naive level-three structures by GLβ‚‚(𝔽₃).