Tits's simplicity criterion for the normal closure of a split unipotent subgroup #
Let G carry a Tits system (B, N) with irreducible Weyl group, and let U be a solvable
subgroup normalized by B with B ≤ U (B ∩ N). Neither U ≤ B nor equality is assumed; the
split case B = U (B ∩ N), as for the unipotent radical of the Borel subgroup of a split BN-pair,
is the main example. Put
Z = ⋂_{g : G} g B g⁻¹, H⁺ = normal closure of U in G.
The group G itself need not be perfect or generated by the conjugates of U: for the finite
groups of Lie type, G is the group of fixed points of a Steinberg endomorphism and H⁺ is the
subgroup generated by its root subgroups, typically the derived subgroup. This file proves
Tits's simplicity criterion for H⁺, under the assumption that H⁺ is perfect:
- every normal subgroup of
H⁺lies inZ ∩ H⁺or is all ofH⁺; - if moreover
Z ∩ U = 1, thenZ ∩ H⁺is the centre ofH⁺; - hence
H⁺ / (Z ∩ H⁺)is simple whenever it is nontrivial, and whenH⁺ = [G, G]the derived central quotient[G, G] / Z([G, G])is simple whenever it is nontrivial; - if
Z ∩ U = 1, thenH⁺ / (Z ∩ H⁺)is nontrivial as soon asU ≠ 1. WhenH⁺ = [G, G], the derived central quotient is nontrivial as soon asU ≠ 1, for any perfect[G, G](TauCeti.Group.IsPerfect.nontrivial_quotient_center).
Either H⁺ ≤ Z, and everything is degenerate, or H⁺ B = G by Tits's theorem on normal
subgroups. In the second case G = H⁺ (B ∩ N), so (B, N) pulls back to a Tits system on H⁺
(TauCeti.TitsSystem.comap) with the same Weyl group and simple reflections, in which the
conjugates of U generate. Tits's simplicity theorem for that Tits system gives the first
statement, because the normal core of B ∩ H⁺ in H⁺ is Z ∩ H⁺. If Z ∩ U = 1, then Z
commutes with U, hence with each conjugate of U, so Z ∩ H⁺ is central; conversely the centre
of the perfect group H⁺ cannot be all of H⁺ unless H⁺ is trivial.
Main results #
TauCeti.TitsSystem.le_normalCore_subgroupOf_or_eq_top_of_normal: the normal-subgroup theorem forH⁺.TauCeti.TitsSystem.center_normalClosure_eq_normalCore_subgroupOf: the centre ofH⁺isZ ∩ H⁺whenZ ∩ U = 1.TauCeti.TitsSystem.isSimpleGroup_quotient_normalCore_subgroupOf:H⁺ / (Z ∩ H⁺)is simple whenever it is nontrivial.TauCeti.TitsSystem.isSimpleGroup_derivedCentralQuotient: ifH⁺ = [G, G], the derived central quotient ofGis simple whenever it is nontrivial.TauCeti.TitsSystem.nontrivial_quotient_normalCore_subgroupOf:H⁺ / (Z ∩ H⁺)is nontrivial whenZ ∩ U = 1andU ≠ 1.
References #
- N. Bourbaki, Lie Groups and Lie Algebras, Chapters 4--6, Chapter IV, §2, no. 7.
- B. Conrad, handout on Tits systems for Math 249B (Stanford, Winter 2016), Theorem 2.4.
- R. W. Carter, Simple Groups of Lie Type (1972), Chapter 11.
The normal closure of a split unipotent subgroup #
If Z = ⋂ g B g⁻¹ meets U ≠ 1 trivially, then the quotient H⁺ / (Z ∩ H⁺) of the normal
closure H⁺ of U is nontrivial, since otherwise U ≤ H⁺ ≤ Z.
Tits's theorem on normal subgroups of the normal closure of U. Suppose the Weyl group is
irreducible, U is a solvable subgroup normalized by B with B ≤ U (B ∩ N), and the normal
closure H⁺ of U is perfect. Then every normal subgroup of H⁺ either lies in Z ∩ H⁺, where
Z = ⋂ g B g⁻¹ is the normal core of B, or is all of H⁺.
The centre of the normal closure of U. Under the hypotheses of
TauCeti.TitsSystem.le_normalCore_subgroupOf_or_eq_top_of_normal, if moreover Z ∩ U = 1, then
the centre of the normal closure H⁺ of U is Z ∩ H⁺.
Tits's simplicity criterion for the normal closure of U. Under the hypotheses of
TauCeti.TitsSystem.le_normalCore_subgroupOf_or_eq_top_of_normal, the quotient H⁺ / (Z ∩ H⁺)
of the normal closure H⁺ of U by its intersection with Z = ⋂ g B g⁻¹ is simple whenever it is
nontrivial.
Simplicity of the derived central quotient from a split Tits system. Suppose the Weyl group
is irreducible, U is a solvable subgroup normalized by B with B ≤ U (B ∩ N) and
Z ∩ U = 1 for Z = ⋂ g B g⁻¹, the normal closure of U is the derived subgroup [G, G], and
[G, G] is perfect. Then the derived central quotient [G, G] / Z([G, G]) is simple whenever it
is nontrivial.