Definitions/Def_WeierstrassCurve_ModularityProps.lean
Modularity inputs: unipotency, 3–5 switch, residual modularity lifting
Five predicates on integral Weierstrass models are introduced, all phrased for a chosen model W over \mathbb{Z} rather than for an isomorphism class over \mathbb{Q}.
ModRepHasUnipotent W n asserts the existence of a \mathbb{Q}-algebra automorphism \sigma of \overline{\mathbb{Q}} = AlgebraicClosure ℚ whose induced \mathbb{Z}/n-linear endomorphism u = \rho(\sigma) of the n-torsion submodule of the group of affine points of W base-changed to \overline{\mathbb{Q}} (the endomorphism supplied by galoisRepModuleEnd) satisfies (u-1)(u-1) = 0 together with u - 1 \neq 0: some Galois element acts unipotently but non-trivially.
The remaining four are named statements to be cited as hypotheses: LanglandsTunnellWeightTwo and ModularityLiftingAt p are closed propositions (the latter depending only on a prime), while ModThreeOrFiveIrreducible W and ThreeFiveSwitchCurve W depend on the model W. LanglandsTunnellWeightTwo says: every W with \Delta \neq 0 which is semistable in the project's sense (no prime dividing \Delta divides c_4) and has irreducible mod-3 torsion representation is residually modular at 3, meaning that for some level M > 0 there is a normalised weight-two eigenform on \Gamma_0(M) and a maximal ideal \mathfrak{m} of the integral closure of \mathbb{Z} in \mathbb{C} containing 3 with a_\ell(f) \equiv a_\ell(W) \pmod{\mathfrak{m}} for all primes \ell \nmid \Delta, \ell \nmid M, \ell \neq 3, where a_\ell(W) is computed as \ell + 1 - \#W(\mathbb{F}_\ell) for the reduction of the model. ModThreeOrFiveIrreducible W says that a semistable W with \Delta \neq 0 has irreducible mod-3 or mod-5 torsion representation. ThreeFiveSwitchCurve W says that for such W with irreducible mod-5 representation there is a second model W' with \Delta' \neq 0, semistable, mod-3 irreducible, possessing a non-trivial unipotent in its mod-3 representation, satisfying 3 \mid \Delta' or 3 \nmid a_3(W'), and congruent to W in the sense that 5 \mid a_\ell(W') - a_\ell(W) for every prime \ell \neq 5 good for both. ModularityLiftingAt p says that semistability, \Delta \neq 0, mod-p irreducibility and residual modularity at p imply that W is modular of some level N > 0 divisible by every prime dividing \Delta. A local instance equips \overline{\mathbb{Q}} with classical decidable equality, as required by the torsion-representation definitions.
Relation to Mathlib
Mathlib supplies WeierstrassCurve, CuspForm and the congruence subgroups \Gamma_0(N), but no notion of modularity of an elliptic curve, of residual modularity, or of the mod-n Galois representation on torsion; all predicates here are the project's own, built on its \mathbb{Z}/n-module structure on the n-torsion of the points over \overline{\mathbb{Q}} and on Frobenius traces defined by point counts of reductions.
Where it is used
These predicates are the vocabulary in which modularity of semistable elliptic curves over \mathbb{Q} is assembled: the argument splits according to whether the mod-3 representation is irreducible, using LanglandsTunnellWeightTwo and ModularityLiftingAt 3 in that case, and otherwise ModThreeOrFiveIrreducible, ThreeFiveSwitchCurve and ModularityLiftingAt 5 to transfer the conclusion from an auxiliary curve congruent mod 5. The resulting modularity statement is what is applied to the Frey curve attached to a solution of the Fermat equation.
References
- H. Darmon, F. Diamond and R. Taylor, Fermat's Last Theorem, in: Current Developments in Mathematics 1995, International Press, 1995, 1–154, Theorem 3.48
- J. Tunnell, Artin's conjecture for representations of octahedral type, Bulletin of the American Mathematical Society (N.S.) 5 (1981), 173–175
- K. Rubin and A. Silverberg, Families of elliptic curves with constant mod p representations, in: Elliptic Curves, Modular Forms, and Fermat's Last Theorem (Hong Kong, 1993), International Press, 1995, 148–161
References are suggested automatically and have not been individually verified.
English text generated automatically from the Lean source; the Lean statement is authoritative.
- 55 lines
- 5 declarations
- used in the statements of 2 theorems and imported by 4 proofs
- imports 4 definition modules
Source file: Definitions/Def_WeierstrassCurve_ModularityProps.lean
Imports
Declarations
- def
WeierstrassCurve.ModRepHasUnipotent - def
WeierstrassCurve.LanglandsTunnellWeightTwo - def
WeierstrassCurve.ModThreeOrFiveIrreducible - def
WeierstrassCurve.ThreeFiveSwitchCurve - def
WeierstrassCurve.ModularityLiftingAt
Source
import Mathlib.NumberTheory.FLT.Basic ↗ import Mathlib.RingTheory.IntegralClosure.IsIntegralClosure.Basic ↗ import Mathlib.Algebra.Algebra.Basic ↗ import Definitions.Def_EllipticCurve_FrobeniusTrace import Definitions.Def_FLTPrelim_GaloisRep import Definitions.Def_FLTPrelim_ModularRep import Definitions.Def_FLTPrelim_Modularity set_option autoImplicit false open WeierstrassCurve WeierstrassCurve.Affine.Point open scoped CongruenceSubgroup noncomputable section namespace WeierstrassCurve open CuspForm ModularFormClass noncomputable local instance instDecEqQbarNo2Assembly : DecidableEq (AlgebraicClosure ℚ) := Classical.decEq _ def ModRepHasUnipotent (W : WeierstrassCurve ℤ) (n : ℕ) : Prop := ∃ σ : AlgebraicClosure ℚ ≃ₐ[ℚ] AlgebraicClosure ℚ, (galoisRepModuleEnd (S := ℚ) (K := AlgebraicClosure ℚ) (W.map (Int.castRingHom ℚ)) n σ - 1) * (galoisRepModuleEnd (S := ℚ) (K := AlgebraicClosure ℚ) (W.map (Int.castRingHom ℚ)) n σ - 1) = 0 ∧ galoisRepModuleEnd (S := ℚ) (K := AlgebraicClosure ℚ) (W.map (Int.castRingHom ℚ)) n σ - 1 ≠ 0 def LanglandsTunnellWeightTwo : Prop := ∀ W : WeierstrassCurve ℤ, W.Δ ≠ 0 → W.IsSemistableModel → W.ModRepIsIrreducible 3 → W.IsResiduallyModular 3 def ModThreeOrFiveIrreducible (W : WeierstrassCurve ℤ) : Prop := W.Δ ≠ 0 → W.IsSemistableModel → (W.ModRepIsIrreducible 3 ∨ W.ModRepIsIrreducible 5) def ThreeFiveSwitchCurve (W : WeierstrassCurve ℤ) : Prop := W.Δ ≠ 0 → W.IsSemistableModel → W.ModRepIsIrreducible 5 → ∃ W' : WeierstrassCurve ℤ, W'.Δ ≠ 0 ∧ W'.IsSemistableModel ∧ W'.ModRepIsIrreducible 3 ∧ W'.ModRepHasUnipotent 3 ∧ (¬ W'.IsGoodPrimeFor 3 ∨ ¬ (3 : ℤ) ∣ W'.apOfModel 3) ∧ ∀ ℓ : ℕ, ℓ.Prime → W.IsGoodPrimeFor ℓ → W'.IsGoodPrimeFor ℓ → ℓ ≠ 5 → (5 : ℤ) ∣ (W'.apOfModel ℓ - W.apOfModel ℓ) def ModularityLiftingAt (p : ℕ) : Prop := ∀ W : WeierstrassCurve ℤ, W.Δ ≠ 0 → W.IsSemistableModel → W.ModRepIsIrreducible p → W.IsResiduallyModular p → W.IsModularModelOfConductorLevel end WeierstrassCurve end