Tate's local Euler characteristic formula, prime-to-ℓ case #
Let F/ℚ_p be finite, ℓ a prime, and A a finite smooth discrete ZMod ℓ-representation of
G_F on which an open normal subgroup of index prime to ℓ acts trivially. This file proves
Tate's formula χ_F(A) = φ_F(A) for such A
(localEulerCharacteristic_eq_localCardNorm_of_index_coprime):
#H⁰(F, A) · #H²(F, A) / #H¹(F, A) = ‖#A‖_F.
This is the case of the formula to which the modular Artin theorem and Shapiro's lemma reduce the
general one (localEulerCharacteristic_eq_localCardNorm_of_indFDRep).
Shrink the subgroup to an open normal W that also fixes μ_ℓ, with index still prime to ℓ
(exists_openNormalSubgroup_le_muNRep_ρ_eq_self_of_coprime), and write A, and μ_ℓ as M, as
inflations of representations of G = G_F ⧸ W. Then, by the coprime index,
dim H⁰(F, A) = dim Aᴳ(finrank_continuousCohomology_zero_fdGalRepOfQuotient);dim H¹(F, A) = dim (H¹(W, 𝔽_ℓ) ⊗ A)ᴳ(finrank_continuousCohomology_one_fdGalRepOfQuotient);dim H²(F, A) = dim (M^∨ ⊗ A)ᴳ(finrank_continuousCohomology_two_fdGalRepOfQuotient).
Equivariant Kummer theory computes the middle term at the fixed field L of W:
dim (H¹(W, 𝔽_ℓ) ⊗ A)ᴳ = dim (M^∨ ⊗ A)ᴳ + dim Aᴳ + [ℓ = p] [F : ℚ_p] dim A
(finrank_invariants_kummerH1FiniteRepresentation_tprod_of_isUnit and
finrank_invariants_kummerH1FiniteRepresentation_tprod_of_finitePadicExtension, stated for
Gal(L/F) and read here through G ≃* Gal(L/F)). Comparing with ‖ℓ‖_F, which is 1 for
ℓ ≠ p and p^{-[F : ℚ_p]} for ℓ = p, gives χ_F(A) = φ_F(A).
Main results #
TauCeti.ClassFieldTheory.finrank_continuousCohomology_zero_fdGalRepOfQuotient:dim H⁰(F, A) = dim Aᴳfor an inflated representation.TauCeti.ClassFieldTheory.finrank_continuousCohomology_one_fdGalRepOfQuotient:dim H¹(F, A) = dim (H¹(W, 𝔽_ℓ) ⊗ A)ᴳfor an inflated representation,Wof index prime toℓ.TauCeti.ClassFieldTheory.localEulerCharacteristic_eq_localCardNorm_of_index_coprime:χ_F(A) = φ_F(A)when an open normal subgroup of index prime toℓacts trivially onA.
References #
- J. Neukirch, A. Schmidt, K. Wingberg, Cohomology of Number Fields, 2nd ed., proof of (7.3.1).
- J. S. Milne, Arithmetic Duality Theorems, 2nd ed., I, proof of Theorem 2.8.
- The statement and the outline of its proof are adapted from split 6 of the pull-request split
plan in
LOOKAHEAD.mdon the Tau Ceti lookahead branchlookahead/ClassFieldTheory/euler-characteristic-mixedofroed-math/TauCeti, commit217e838c7c3ccfe5f1c3874537459b0527c37625.
dim H⁰(F, A) = dim Aᴳ for an inflated representation: H⁰ is the module of invariants
(ContinuousCohomology.zeroIso), and an element is fixed by G_F exactly when it is fixed by
its quotient G = G_F ⧸ W.
dim H¹(F, A) = dim (H¹(W, 𝔽_ℓ) ⊗ A)ᴳ for an inflated representation, when the index of
W is prime to ℓ: restriction to W identifies H¹(F, A) with the G-invariants of
H¹(W, A) = H¹(W, 𝔽_ℓ) ⊗ A (finrank_H1_eq_finrank_representationInvariants), G = G_F ⧸ W
acting on H¹(W, 𝔽_ℓ) by conjugation.
Tate's local Euler characteristic formula, prime-to-ℓ case. Let F/ℚ_p be finite and
ℓ a prime. If an open normal subgroup of G_F of index prime to ℓ acts trivially on a finite
smooth discrete ZMod ℓ-representation A, then χ_F(A) = φ_F(A), that is
#H⁰(F, A) · #H²(F, A) / #H¹(F, A) = ‖#A‖_F.