introvoyz045's picture
Migrated from GitHub
4a56cd6 verified
Raw History Blame Contribute Delete
1.95 kB
import OAI.AlgebraicGeometry.GroupSchemes.PrimeDichotomy
import OAI.AlgebraicGeometry.GroupSchemes.DualPoints
/-! # First Level -/
namespace OAI.AlgebraicGeometry.GroupSchemes
noncomputable section
/-! The height-one first level is split étale or has split étale Cartier dual. -/
open _root_.CategoryTheory _root_.AlgebraicGeometry Opposite WithConv _root_.CategoryTheory.MonObj
universe u
variable {k : Type u} [Field k] [IsAlgClosed k]
variable {p : ℕ} [Fact p.Prime] [CharP k p]
/-- In the connected case, the first-level Cartier dual is split étale. -/
theorem FiniteEtale.PDivisibleGroup.first_dual_split_of_local (G : FiniteEtale.PDivisibleGroup k p 1)
(hfirst : IsLocalRing (G.level 1).algebra) :
Algebra.IsFiniteSplit k (G.level 1).cartierDual.algebra := by
have hdim : Module.finrank k (G.level 1).algebra = p := by
simpa only [Nat.mul_one, pow_one] using G.rank 1
have hdimD : Module.finrank k (G.level 1).cartierDual.algebra = p :=
(G.level 1).cartierDual_rank.trans hdim
have hpow : ∀ x : FiniteEtale.affine k (CommAlgCat.of k k) ⟶ (G.level 1).cartierDual.scheme,
x ^ p = 1 := by
intro x
exact @DualPoints.FiniteEtale.dualOfHopf_points_pow k (G.level 1).algebra
_ _ (G.level 1).coordinateHopf (G.level 1).coordinateHopf_cocomm (G.level 1).finite
p _ _ hfirst hdim x
exact ((G.level 1).cartierDual.local_or_split p hdimD hpow).resolve_left
(G.dual_not_local_of_first_local hfirst)
/-- The actual first level in height one is étale or of multiplicative type. -/
theorem FiniteEtale.PDivisibleGroup.first_split_or_dual_split (G : FiniteEtale.PDivisibleGroup k p 1) :
Algebra.IsFiniteSplit k (G.level 1).algebra ∨
Algebra.IsFiniteSplit k (G.level 1).cartierDual.algebra := by
rcases G.first_local_or_split with hlocal | hsplit
· exact Or.inr (G.first_dual_split_of_local hlocal)
· exact Or.inl hsplit
end
end OAI.AlgebraicGeometry.GroupSchemes