Deligne's integral representation over the canonical coefficient completion
OpenMTT.Eigenform.exists_adic_matrix_representationLet , let , let be prime, and let be a normalized algebraic cuspidal Hecke eigenform of level , weight and nebentype . Fix an embedding of the algebraic numbers into . Let be the field generated by all Fourier coefficients and nebentype values, let be the prime selected by the embedding, and put
There exists a representation on an integral lattice with a chosen basis,
such that every reduction modulo factors through a finite extension of , inertia acts trivially outside , and at each prime every arithmetic Frobenius lift satisfies
The values on the right are mapped from into . The formal statement uses a monoid homomorphism into matrices; its values are automatically invertible because its source is a group. Congruence continuity is expressed at every ideal power by GaloisFactorsThroughFiniteLevel. The inertia and Frobenius conditions use valuation subrings of the actual algebraic closure, with residue Frobenius .
This is the characteristic-zero Deligne theorem on a stable integral lattice. Deligne–Serre Theorem 6.1 supplies the representation over for any eigenform of weight at least two; it does not require newness or a trivial nebentype. Its compact image preserves an integral lattice (§6.12). The ring of integers of this local field identifies with the displayed adic completion; choosing a basis gives the stated form. The finite-level congruence assertions follow from continuity.
This theorem is the remaining deep input for reducing to the mission's residual representation. It asserts a full integral representation in characteristic zero, rather than assuming the existence of the requested residual kernel field.
import Definitions.Def_KN_EigenformResidualGaloisRepresentationV2 import Definitions.Def_MTT_EigenformCoefficientCompletion import Definitions.Def_GaloisRep_Residual set_option autoImplicit false noncomputable section open NumberField
/-- Deligne's representation on a stable integral lattice at the chosen coefficient prime. -/
theorem MTT.Eigenform.exists_adic_matrix_representation
{N k p : ℕ} [Fact p.Prime] (hN : 0 < N) (hk : 2 ≤ k)
(ι : MTT.Qbar →+* ℂ) (f : MTT.Eigenform N k ι)
(ιp : MTT.Qbar →+* ℂ_[p]) :
∃ ρ : (MTT.Qbar ≃ₐ[ℚ] MTT.Qbar) →*
Matrix (Fin 2) (Fin 2) (f.coefficientCompletion ιp),
(∀ n : ℕ, GaloisFactorsThroughFiniteLevel
(((AdicCompletion.evalₐ (f.coefficientPrime ιp) n).toRingHom.mapMatrix
(m := Fin 2)).toMonoidHom.comp ρ)) ∧
(∀ (l : ℕ), l.Prime → Nat.Coprime l (N * p) →
∀ A : ValuationSubring MTT.Qbar, A.LiesOverPrime l →
∀ σ ∈ A.inertiaSubgroupIn ℚ, ρ σ = 1) ∧
(∀ (l : ℕ), l.Prime → Nat.Coprime l (N * p) →
∀ A : ValuationSubring MTT.Qbar, A.LiesOverPrime l →
∀ σ : MTT.Qbar ≃ₐ[ℚ] MTT.Qbar, A.IsFrobeniusAt σ l →
Matrix.trace (ρ σ) =
algebraMap (𝓞 f.coefficientField) (f.coefficientCompletion ιp)
(f.integralCoeff hN hk l) ∧
Matrix.det (ρ σ) =
algebraMap (𝓞 f.coefficientField) (f.coefficientCompletion ιp)
(f.integralNebentype (l : ZMod N) *
(l : 𝓞 f.coefficientField) ^ (k - 1))) := by
sorry