The cohomological dimension of a pro-p group is detected on 𝔽_p #
Let G be a compact pro-p group. Its p-cohomological dimension cd_p G asks for the vanishing
of Hⁱ(G, M) in every degree i > n and for every discrete p-primary torsion G-module M.
For a pro-p group the single module 𝔽_p with trivial action suffices:
cd_p G ≤ nif and only ifHⁿ⁺¹(G, 𝔽_p) = 0.
Two reductions compose. Cohomological dimension is detected in the single degree n + 1 on the
finite discrete p-primary modules
(TauCeti.cohomologicalDimensionAt_le_iff_forall_finite_subsingleton_succ, by dimension
shifting), and for a pro-p group the vanishing of a fixed degree on those modules is detected on
the trivial modules of order p
(TauCeti.IsProP.forall_subsingleton_continuousCohomology_iff_forall_natCard_eq, by dévissage
along the trivial filtration). A trivial discrete G-module of order p is 𝔽_p up to a
G-equivariant isomorphism, so its cohomology is cohomFp p G (n + 1)
(TauCeti.subsingleton_continuousCohomology_iff_subsingleton_cohomFp_of_natCard_eq).
This is a reduction and not a definition: cd_p G is the standard invariant, and the equivalence
is the content of dévissage. The pro-p hypothesis is used. For the symmetric group S₃ at
p = 3, H¹(S₃, 𝔽₃) = Hom(S₃, 𝔽₃) vanishes, while cd_3 S₃ is not 0: by Shapiro's lemma
H¹(S₃, 𝔽₃[S₃/C₃]) = H¹(C₃, 𝔽₃) ≠ 0, and 𝔽₃[S₃/C₃] is a finite 3-primary module.
The equivalence is restated for <, = and = ⊤: n < cd_p G exactly when Hⁿ⁺¹(G, 𝔽_p) ≠ 0,
cd_p G = n exactly when Hⁿ⁺¹(G, 𝔽_p) = 0 and Hⁿ(G, 𝔽_p) ≠ 0, and cd_p G = ⊤ exactly when
Hⁿ(G, 𝔽_p) ≠ 0 for every n.
The file also records Tate's criterion for a profinite pro-p group: if Hⁿ(G, -) is right exact
on the finite discrete G-modules killed by p, then cd_p G ≤ n. The vanishing of Hⁿ⁺¹ on
those modules is
TauCeti.subsingleton_continuousCohomology_succ_of_forall_coeffMap_proj_surjective, and
dévissage extends it to all finite p-primary modules. In degree 2 the right exactness follows
from duality: if Tate's duality map α₂ : H²(G, M) → Hom(H⁰(G, Hom(M, N)), H²(G, N)) is bijective
for every finite discrete M killed by p, then cd_p G ≤ 2
(TauCeti.IsProP.cohomologicalDimensionAt_le_two_of_forall_dualityMap2_bijective). This is how
the cohomological dimension of a Demushkin group is bounded (Serre, Structure de certains
pro-p-groupes, §9.1).
Conversely, cd_p G ≤ 2 makes H²(G, -) right exact on the twisted coefficients I(χ)/pⁱ of a
continuous character χ : G →ₜ* ℤ_pˣ of a compact group G: every reduction
H²(G, I(χ)/pⁿ) → H²(G, I(χ)/pʲ) is surjective, because the next term H³ of the long exact
sequence vanishes
(TauCeti.ZModTwist.surjective_explicitCoeff2_reduce_of_cohomologicalDimensionAt_le_two).
Main results #
TauCeti.IsProP.cohomologicalDimensionAt_le_iff_forall_smul_eq_self,TauCeti.IsProP.cohomologicalDimensionAt_le_iff_forall_natCard_eq_smul_eq_self: the pro-preduction ofcd_p G ≤ nto the finite trivial modules killed byp, resp. of orderp.TauCeti.IsProP.cohomologicalDimensionAt_le_iff_subsingleton_cohomFp: the pro-preduction ofcd_p,cd_p G ≤ n ↔ Hⁿ⁺¹(G, 𝔽_p) = 0.TauCeti.IsProP.lt_cohomologicalDimensionAt_iff_nontrivial_cohomFp,TauCeti.IsProP.cohomologicalDimensionAt_eq_iff,TauCeti.IsProP.cohomologicalDimensionAt_eq_top_iff: the<,=and= ⊤forms.TauCeti.IsProP.cohomologicalDimensionAt_le_of_forall_coeffMap_proj_surjective: Tate's criterion: right exactness ofHⁿ(G, -)on the finite discreteG-modules killed bypgivescd_p G ≤ n.TauCeti.IsProP.cohomologicalDimensionAt_le_two_of_forall_dualityMap2_bijective: Tate's duality criterion: bijectivity ofα₂on the finite discreteG-modules killed bypgivescd_p G ≤ 2.TauCeti.ZModTwist.surjective_explicitCoeff2_reduce_of_cohomologicalDimensionAt_le_two: for a compact group withcd_p G ≤ 2, every reductionH²(G, I(χ)/pⁿ) → H²(G, I(χ)/pʲ)of the twisted coefficients is surjective.
References #
- J.-P. Serre, Galois Cohomology, Ch. I, §4.1, Prop. 21.
- J.-P. Serre, Structure de certains pro-
p-groupes (d'après Demuškin), Séminaire Bourbaki, exp. 252 (1963), §9.1. - J. Neukirch, A. Schmidt, K. Wingberg, Cohomology of Number Fields, 2nd ed., (3.3.2).
- H. Koch, Galois Theory of
p-Extensions, Springer (2002), Definition 5.1.
The pro-p reduction of cd_p, on the trivial modules of order p. For a compact pro-p
group G, cd_p G ≤ n exactly when Hⁿ⁺¹(G, A) vanishes for every discrete G-module A of
order p with trivial action.
The pro-p reduction of cd_p, on the finite trivial modules killed by p. For a compact
pro-p group G, cd_p G ≤ n exactly when Hⁿ⁺¹(G, A) vanishes for every finite discrete
G-module A killed by p with trivial action, that is for every finite elementary abelian
p-group with trivial action.
The pro-p reduction of cd_p to the single module 𝔽_p. For a compact pro-p group
G, cd_p G ≤ n exactly when Hⁿ⁺¹(G, 𝔽_p) vanishes.
For a compact pro-p group G, cd_p G ≤ n exactly when Hᵐ(G, 𝔽_p) vanishes for every
m > n.
For a compact pro-p group G, n < cd_p G exactly when Hⁿ⁺¹(G, 𝔽_p) is nontrivial.
For a compact pro-p group G, n + 1 ≤ cd_p G exactly when Hⁿ⁺¹(G, 𝔽_p) is
nontrivial.
cd_p G = n on 𝔽_p. For a compact pro-p group G, cd_p G = n exactly when
Hⁿ⁺¹(G, 𝔽_p) vanishes and Hⁿ(G, 𝔽_p) does not.
Infinite cd_p on 𝔽_p. For a compact pro-p group G, cd_p G = ⊤ exactly when
Hⁿ(G, 𝔽_p) is nontrivial in every degree.
Tate's criterion for the cohomological dimension of a pro-p group. Let G be a profinite
pro-p group. If for every short exact sequence 0 → A → B → C → 0 of finite discrete G-modules
killed by p the map Hⁿ(G, B) → Hⁿ(G, C) is surjective, then cd_p G ≤ n: right exactness
kills Hⁿ⁺¹ on those modules, and dévissage extends the vanishing to every finite discrete
p-primary module.
Tate's duality criterion for cd_p G ≤ 2. Let G be a profinite pro-p group and N a
discrete G-module. If Tate's duality map in degree two,
α₂ : H²(G, M) → Hom(H⁰(G, InternalHom G M N), H²(G, N)), is bijective for every finite discrete
G-module M killed by p, then cd_p G ≤ 2: duality makes H²(G, -) right exact on those
modules (TauCeti.ContCohomology.DiscreteShortExact.coeffMap_proj_surjective_of_dualityMap2), and
right exactness bounds the dimension. For an infinite Demushkin group and N = 𝔽_p the hypothesis
is Tate's perfect duality, and the conclusion is the upper bound in cd_p G = 2.
cd_p G ≤ 2 makes H²(G, -) right exact on the twisted coefficients: for a compact group
G with cd_p G ≤ 2, every reduction H²(G, I(χ)/pⁿ) → H²(G, I(χ)/pʲ), j ≤ n, is surjective,
because the next term H³(G, I(χ)/pⁿ⁻ʲ) of the long exact sequence vanishes.