Peterfalvi §11: the Wielandt setup, opening layer
DefinitionOddOrder_Peterfalvi_S11_MaximalII_III_IV_WielandtSetupBasicfeit-thompsonfinite-groupsgroup-theoryishida-odd-order
The opening layer of the Wielandt setup in Peterfalvi, Character Theory for the Odd Order Theorem, §11, split from the parent file. It introduces the common setup of Peterfalvi (9.2) for a maximal subgroup of type II, III or IV (TypesIIIIIIVSetup) and the notation , , , and .
Main declarations: TypesIIIIIIVSetup; H (Peterfalvi's H in (9.2)); U (Peterfalvi's U in (9.2)); W1 (Peterfalvi's W_1 in (9.2)); and 4 more.
The bundle also proves 24 supporting lemmas used by later statements.
Definition code
/-
Copyright (c) 2026 Yawara Ishida. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Yawara Ishida
-/
-- From github.com/yawara/odd-order @ 20a6fccbb63c (Apache-2.0; LICENSE at that commit):
-- OddOrder/Peterfalvi/S11_MaximalII_III_IV/WielandtSetupBasic.lean
-- Generated for Prove2Me by deleting declarations; the remaining commands are verbatim except:
-- `private` dropped where another file uses the declaration, and a solution's main theorem
-- renamed `_root_.solution` (platform convention).
import Definitions.Def_OddOrder_Peterfalvi_S10_MinimalSimpleStructure
import Definitions.Def_OddOrder_Peterfalvi_S10_CoherenceWiring
import Definitions.Def_OddOrder_GroupTheory_CoprimeAction
import Definitions.Def_OddOrder_GroupTheory_WielandtFixedPoint
import Definitions.Def_OddOrder_GroupTheory_CoprimeFixedPoints
import Definitions.Def_OddOrder_GroupTheory_MaximalSubgroupTypeConj
import Definitions.Def_OddOrder_GroupTheory_AInvariantPiSubgroups
import Definitions.Def_OddOrder_BG_Ch1_Preliminary_OperatorMaschke
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S15_MF_SetupLemma151
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S15_MF_Theorem152Helpers
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S15_MF_Corollary155
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S15_MF_TIFailure
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S15_MF_WitnessPGroup
import Definitions.Def_OddOrder_GroupTheory_RepresentationTheory_SingerField
import Definitions.Def_OddOrder_GroupTheory_RepresentationTheory_SingerLineBound
import Definitions.Def_OddOrder_GroupTheory_RepresentationTheory_WielandtElabBridge
import Definitions.Def_OddOrder_GroupTheory_RepresentationTheory_CliffordSingleOrbit
import Definitions.Def_OddOrder_Peterfalvi_S08_CoherenceCorePart1
import Definitions.Def_OddOrder_GroupTheory_RepresentationTheory_InducedTransport
import Definitions.Def_OddOrder_GroupTheory_RepresentationTheory_OrbitOnIrr
import Definitions.Def_OddOrder_Mathlib_SchurZassenhausConj
namespace OddOrder.Peterfalvi.S11
open OddOrder.GroupTheory
open OddOrder.RepresentationTheory
open OddOrder.Isaacs
open scoped Pointwise
variable {G : Type*} [Group G]
/-- The common setup of Peterfalvi (9.2): a maximal subgroup of type II, III, or
IV, together with its type-`P` data from (8.4).
The `nontrivial` field records the `TypePNontrivialCore` of (8.6) — `U ≠ 1`, `|W₁|` prime, and the
TI condition — that all of types II, III, IV carry (it is the `common` field of `TypeIIData` /
`TypeIIIData` / `TypeIVData`); it makes the setup a faithful model of "`M` is of type II/III/IV with
*these* data" (so e.g. `U ≠ 1` is available for the present `typeP`, not only for some witness). -/
structure TypesIIIIIIVSetup (M : Subgroup G) where
maximal : M ∈ maximalSubgroups G
typeP : TypePData M
nontrivial : TypePNontrivialCore M typeP
type_alt : IsTypeII M ∨ IsTypeIII M ∨ IsTypeIV M
namespace TypesIIIIIIVSetup
/-- Peterfalvi's `H` in (9.2). Reducible so that the chief-factor carrier `↥data.H ⧸ N` is
defeq-transparently `↥data.typeP.H ⧸ N`, letting instance search find the kernel's normality and the
descended action across the `TypesIIIIIIVSetup`/`TypePData` boundary. -/
@[reducible] def H {M : Subgroup G} (data : TypesIIIIIIVSetup M) : Subgroup G :=
data.typeP.H
/-- Peterfalvi's `U` in (9.2). -/
def U {M : Subgroup G} (data : TypesIIIIIIVSetup M) : Subgroup G :=
data.typeP.U
/-- Peterfalvi's `W_1` in (9.2). -/
def W1 {M : Subgroup G} (data : TypesIIIIIIVSetup M) : Subgroup G :=
data.typeP.W1
/-- Peterfalvi's `W_2` in (9.2). -/
def W2 {M : Subgroup G} (data : TypesIIIIIIVSetup M) : Subgroup G :=
data.typeP.W2
/-- Peterfalvi's `q = |W_1|` in (9.2). -/
noncomputable def q {M : Subgroup G} (data : TypesIIIIIIVSetup M) : ℕ :=
Nat.card ↥data.W1
/-- Two `TypesIIIIIIVSetup`s over the same `M` with the same type-`P` structure are equal: every
other field (`maximal`, `nontrivial`, `type_alt`) is propositional, so the setup is determined by
`typeP` up to proof irrelevance. This identifies a §13 `Hypothesis.s11Setup` (pinned only by
`setup_typeP_eq`) with the producer `toTypesIIIIIIVSetup`, transporting §9 facts stated over the
producer setup into the §13 world. -/
theorem eq_of_typeP_eq {M : Subgroup G} {data data' : TypesIIIIIIVSetup M}
(h : data.typeP = data'.typeP) : data = data' := by
cases data
cases data'
cases h
rfl
end TypesIIIIIIVSetup
section Wielandt93
open OddOrder.Isaacs.Ch06
variable {M : Subgroup G}
theorem derivedInG_le_self (H : Subgroup G) : derivedInG H ≤ H := by
unfold derivedInG; exact Subgroup.map_subtype_le _
/-- `A ⊓ B = ⊥` from a relative complement: if `A, B ≤ K` and `A.subgroupOf K`, `B.subgroupOf K`
are complements in `↥K`, then `A` and `B` meet trivially in `G`. -/
private theorem inf_eq_bot_of_isComplement_subgroupOf {A B K : Subgroup G}
(hA : A ≤ K) (hc : Subgroup.IsComplement' (A.subgroupOf K) (B.subgroupOf K)) :
A ⊓ B = ⊥ := by
have hcomap : (A.subgroupOf K) ⊓ (B.subgroupOf K) = (A ⊓ B).subgroupOf K :=
(Subgroup.comap_inf A B K.subtype).symm
have hd : (A ⊓ B).subgroupOf K = ⊥ := by rw [← hcomap]; exact hc.disjoint.eq_bot
rw [Subgroup.subgroupOf_eq_bot] at hd
exact hd.eq_bot_of_le (inf_le_left.trans hA)
/-- `H ⊓ U = ⊥`: `H` and `U` are complementary in `M' = derivedInG M` (`derived_complement`). -/
theorem typeP_H_inf_U (data : TypePData M) : data.H ⊓ data.U = ⊥ :=
inf_eq_bot_of_isComplement_subgroupOf data.H_le data.derived_complement
/-- `U ⊓ W₁ = ⊥`: `W₁` complements `M' = derivedInG M` in `M` (`M_complement`) and `U ≤ M'`. -/
private theorem typeP_U_inf_W1 (data : TypePData M) : data.U ⊓ data.W1 = ⊥ := by
have hM' : derivedInG M ⊓ data.W1 = ⊥ :=
inf_eq_bot_of_isComplement_subgroupOf (derivedInG_le_self M) data.M_complement
refine le_bot_iff.mp (le_trans ?_ hM'.le)
exact inf_le_inf_right _ data.U_le
/-- The fixed-point-free condition of Definition (8.4): for `w ∈ W₁#`, no nontrivial `u ∈ U` is
fixed by conjugation by `w`. `C_U(w) ⊆ M' ⊓ C_G(w) = W₂` (`centralizer_W1`) and
`U ⊓ W₂ ⊆ U ⊓ H = ⊥`
(`W₂ ≤ H`, `H ⊓ U = ⊥`). -/
private theorem typeP_W1_fpf_U (data : TypePData M) {w : G} (hw : w ∈ data.W1) (hwne : w ≠ 1)
{u : G} (hu : u ∈ data.U) (heq : w * u * w⁻¹ = u) : u = 1 := by
have hcomm : w * u = u * w := mul_inv_eq_iff_eq_mul.mp heq
have hmemC : u ∈ Subgroup.centralizer ({w} : Set G) :=
Subgroup.mem_centralizer_singleton_iff.mpr hcomm.symm
have hin : u ∈ derivedInG M ⊓ Subgroup.centralizer ({w} : Set G) := ⟨data.U_le hu, hmemC⟩
rw [data.centralizer_W1 w hw hwne] at hin
have huH : u ∈ data.H := (data.W2_le hin).1
have : u ∈ data.H ⊓ data.U := ⟨huH, hu⟩
rw [typeP_H_inf_U data] at this
exact this
/-- **Definition (8.4)**: `U W₁` is a Frobenius group with kernel `U` and complement `W₁`
(for types II–IV, where `U ≠ 1`; at type V the kernel `U = 1` degenerates). -/
theorem typeP_uW1_frobenius (data : TypePData M) (hU : data.U ≠ ⊥) :
IsFrobeniusGroup ↥(data.U ⊔ data.W1)
(data.U.subgroupOf (data.U ⊔ data.W1)) (data.W1.subgroupOf (data.U ⊔ data.W1)) := by
set L := data.U ⊔ data.W1 with hL
have hUnorm : (data.U.subgroupOf L).Normal :=
(Subgroup.normal_subgroupOf_iff_le_normalizer le_sup_left).mpr
(sup_le Subgroup.le_normalizer data.W1_normalizes_U)
have hdisj : Disjoint (data.U.subgroupOf L) (data.W1.subgroupOf L) := by
have hc : (data.U.subgroupOf L) ⊓ (data.W1.subgroupOf L) = (data.U ⊓ data.W1).subgroupOf L :=
(Subgroup.comap_inf data.U data.W1 L.subtype).symm
rw [disjoint_iff, hc, typeP_U_inf_W1 data, Subgroup.bot_subgroupOf]
have hsup : (data.U.subgroupOf L) ⊔ (data.W1.subgroupOf L) = ⊤ := by
rw [← Subgroup.subgroupOf_sup le_sup_left le_sup_right, Subgroup.subgroupOf_self]
refine ⟨hUnorm, ?_, ?_, ?_, ?_⟩
· refine Subgroup.isComplement'_of_disjoint_and_mul_eq_univ hdisj ?_
have := hUnorm
rw [← Subgroup.normal_mul, hsup, Subgroup.coe_top]
· rw [Ne, Subgroup.subgroupOf_eq_bot]
exact fun hd => hU (hd.eq_bot_of_le le_sup_left)
· rw [Ne, Subgroup.subgroupOf_eq_bot]
exact fun hd => data.W1_nontrivial (hd.eq_bot_of_le le_sup_right)
· intro a haA hane n hnN hnne hconj
have hwW1 : (a : G) ∈ data.W1 := Subgroup.mem_subgroupOf.mp haA
have huU : (n : G) ∈ data.U := Subgroup.mem_subgroupOf.mp hnN
have haG : (a : G) ≠ 1 := fun h => hane (Subtype.ext h)
have hnG : (n : G) ≠ 1 := fun h => hnne (Subtype.ext h)
have hconjG : (a : G) * (n : G) * (a : G)⁻¹ = (n : G) := Subtype.ext_iff.mp hconj
exact hnG (typeP_W1_fpf_U data hwW1 haG huU hconjG)
/-- `U W₁ ≤ M ≤ N_G(H)`: the conjugation action of `U W₁` on `H = M_F`. -/
theorem typeP_uW1_le_normalizer_H (data : TypePData M) :
data.U ⊔ data.W1 ≤ Subgroup.normalizer (data.H : Set G) := by
have hMnorm : M ≤ Subgroup.normalizer (data.H : Set G) := by
rw [data.H_eq]; exact OddOrder.BG.Ch4.S15.maxNilpotentNormalHall_le_normalizer M
refine sup_le ?_ (data.W1_le.trans hMnorm)
exact (data.U_le.trans (derivedInG_le_self M)).trans hMnorm
/-- The conjugation action of `U W₁ ≤ N_G(H)` on `H = M_F`. -/
noncomputable def typeP_conjAction (data : TypePData M) :
↥(data.U ⊔ data.W1) →* MulAut ↥data.H :=
letI : MulDistribMulAction ↥(data.U ⊔ data.W1) ↥data.H :=
MulDistribMulAction.compHom (M := ↥(Subgroup.normalizer (data.H : Set G))) ↥data.H
(Subgroup.inclusion (typeP_uW1_le_normalizer_H data))
MulDistribMulAction.toMulAut ↥(data.U ⊔ data.W1) ↥data.H
theorem typeP_conjAction_apply (data : TypePData M) (a : ↥(data.U ⊔ data.W1)) (x : ↥data.H) :
((typeP_conjAction data a x : ↥data.H) : G) = (a : G) * (x : G) * (a : G)⁻¹ := rfl
/-- The action of `U W₁` on `H = M_F` is coprime: `|H| ⟂ |U W₁| = |U| · |W₁|`. `H` is a Hall
subgroup of `M` (`maxNilpotentNormalHall_isHall`), and the index `[M : H] = [M : M'] · [M' : H]
= |W₁| · |U| = |U W₁|` (the relative-index tower with the `M_complement`/`derived_complement`
splittings and the Frobenius product `|U W₁| = |U| · |W₁|`). -/
theorem typeP_coprime_H_uW1 [Finite G] (data : TypePData M) (hU : data.U ≠ ⊥) :
Nat.Coprime (Nat.card ↥data.H) (Nat.card ↥(data.U ⊔ data.W1)) := by
have hHleM : data.H ≤ M := data.H_le.trans (derivedInG_le_self M)
-- `[M' : H] = |U|` and `[M : M'] = |W₁|` from the two complement splittings.
have hUidx : (data.H.subgroupOf (derivedInG M)).index = Nat.card ↥data.U := by
rw [data.derived_complement.symm.index_eq_card,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe data.U_le).toEquiv]
have hW1idx : ((derivedInG M).subgroupOf M).index = Nat.card ↥data.W1 := by
rw [data.M_complement.symm.index_eq_card,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe data.W1_le).toEquiv]
-- `[M' : H] · [M : M'] = [M : H]` (relative-index tower), so `[M : H] = |U| · |W₁|`.
have htower : (data.H.subgroupOf (derivedInG M)).index * ((derivedInG M).subgroupOf M).index
= (data.H.subgroupOf M).index :=
Subgroup.relIndex_mul_relIndex data.H (derivedInG M) M data.H_le (derivedInG_le_self M)
rw [hUidx, hW1idx] at htower
-- `|U W₁| = |U| · |W₁|` from the Frobenius complement.
have hcard : Nat.card ↥(data.U ⊔ data.W1) = Nat.card ↥data.U * Nat.card ↥data.W1 := by
rw [← (typeP_uW1_frobenius data hU).isComplement.card_mul_card,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe le_sup_left).toEquiv,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe le_sup_right).toEquiv]
-- `H` is Hall in `M`: `|H| ⟂ [M : H] = |U| · |W₁| = |U W₁|`.
have hHall := OddOrder.BG.Ch4.S15.maxNilpotentNormalHall_isHall M
rw [← data.H_eq] at hHall
have hcop_idx := hHall.coprime_index
rw [Nat.card_congr (Subgroup.subgroupOfEquivOfLe hHleM).toEquiv] at hcop_idx
rw [hcard, htower]
exact hcop_idx
/-- The coprime Frobenius action of `U W₁` (kernel `U`, complement `W₁`) on `H = M_F`, by
conjugation — the carrier for Wielandt's formula (9.1). -/
noncomputable def typeP_coprimeAction [Finite G] (data : TypePData M) (hU : data.U ≠ ⊥) :
CoprimeFrobeniusAction ↥(data.U ⊔ data.W1) ↥data.H where
U := data.U.subgroupOf (data.U ⊔ data.W1)
E := data.W1.subgroupOf (data.U ⊔ data.W1)
frobenius := typeP_uW1_frobenius data hU
H_solvable := by
rw [data.H_eq]
have := OddOrder.BG.Ch4.S15.maxNilpotentNormalHall_isNilpotent M
exact IsNilpotent.to_isSolvable
φ := typeP_conjAction data
coprime_order := typeP_coprime_H_uW1 data hU
/-- The image in `G` of the fixed subgroup of `K ≤ U W₁` under the conjugation action is the
concrete centralizer `C_H(K) = H ⊓ C_G(K)`. -/
theorem typeP_fixedSubgroup_map (data : TypePData M) {K : Subgroup G}
(hK : K ≤ data.U ⊔ data.W1) :
(fixedSubgroup (typeP_conjAction data) (K.subgroupOf (data.U ⊔ data.W1))).map data.H.subtype
= data.H ⊓ Subgroup.centralizer (K : Set G) := by
ext x
constructor
· rintro ⟨h, hh, rfl⟩
refine ⟨h.2, Subgroup.mem_centralizer_iff.mpr (fun g hg => ?_)⟩
have hfix := hh ⟨g, hK hg⟩ (Subgroup.mem_subgroupOf.mpr hg)
have hconj : (g : G) * (h : G) * (g : G)⁻¹ = (h : G) := by
have := Subtype.ext_iff.mp hfix; rwa [typeP_conjAction_apply] at this
exact mul_inv_eq_iff_eq_mul.mp hconj
· rintro ⟨hxH, hxC⟩
refine ⟨⟨x, hxH⟩, fun l hl => ?_, rfl⟩
have hlx : (l : G) * x = x * (l : G) :=
Subgroup.mem_centralizer_iff.mp hxC (l : G) (Subgroup.mem_subgroupOf.mp hl)
have hval : ((typeP_conjAction data l) ⟨x, hxH⟩ : G) = x := by
rw [typeP_conjAction_apply]; change (l : G) * x * (l : G)⁻¹ = x
rw [hlx, mul_inv_cancel_right]
exact Subtype.ext hval
/-- The order of the fixed subgroup of `K ≤ U W₁` equals `|C_H(K)| = |H ⊓ C_G(K)|`. -/
theorem typeP_card_fixedSubgroup (data : TypePData M) {K : Subgroup G}
(hK : K ≤ data.U ⊔ data.W1) :
Nat.card ↥(fixedSubgroup (typeP_conjAction data) (K.subgroupOf (data.U ⊔ data.W1)))
= Nat.card ↥(data.H ⊓ Subgroup.centralizer (K : Set G)) := by
rw [← typeP_fixedSubgroup_map data hK]
exact Nat.card_congr
(Subgroup.equivMapOfInjective _ data.H.subtype data.H.subtype_injective).toEquiv
/-- **`C_H(W₁) = W₂`** (Peterfalvi (9.2)): `W₂ = C_{M'}(w)` for any `w ∈ W₁#` (`centralizer_W1`),
and since `W = W₁ W₂` is cyclic, `W₂` centralizes all of `W₁`. -/
theorem typeP_H_inf_centralizer_W1 (data : TypePData M) :
data.H ⊓ Subgroup.centralizer (data.W1 : Set G) = data.W2 := by
obtain ⟨w, hwW1, hwne⟩ := (data.W1.bot_or_exists_ne_one).resolve_left data.W1_nontrivial
apply le_antisymm
· intro h hh
obtain ⟨hhH, hhC⟩ := hh
have : h ∈ derivedInG M ⊓ Subgroup.centralizer ({w} : Set G) := by
refine ⟨data.H_le hhH, Subgroup.mem_centralizer_singleton_iff.mpr ?_⟩
exact (Subgroup.mem_centralizer_iff.mp hhC w hwW1).symm
rwa [data.centralizer_W1 w hwW1 hwne] at this
· refine le_inf ?_ ?_
· intro x hx; exact (data.W2_le hx).1
· intro x hxW2
rw [Subgroup.mem_centralizer_iff]
intro g hgW1
have hxW : x ∈ data.W := by rw [data.W_eq]; exact Subgroup.mem_sup_right hxW2
have hgW : g ∈ data.W := by rw [data.W_eq]; exact Subgroup.mem_sup_left hgW1
exact (S06.commute_of_mem_of_isCyclic data.W_cyclic hgW hxW).eq
/-- **Peterfalvi (9.3), the quantitative core** via Wielandt's fixed-point formula (9.1):
`|C_H(U W₁)|^q · |H| = |W₂|^q · |C_H(U)|` with `q = |W₁|` and `C_H(W₁) = W₂`. The Frobenius
group `U W₁` (kernel `U`) acts coprimely on `H = M_F`, and the three fixed-point subgroups of
Wielandt's formula are the concrete centralizers. -/
theorem typeP_wielandt_order_relation [Finite G] (data : TypePData M) (hU : data.U ≠ ⊥) :
Nat.card ↥(data.H ⊓ Subgroup.centralizer ((data.U ⊔ data.W1 : Subgroup G) : Set G))
^ Nat.card ↥data.W1 * Nat.card ↥data.H
= Nat.card ↥data.W2 ^ Nat.card ↥data.W1
* Nat.card ↥(data.H ⊓ Subgroup.centralizer (data.U : Set G)) := by
have key := wielandt_fixedPoint_frobenius (typeP_coprimeAction data hU)
have hEcard : Nat.card ↥(typeP_coprimeAction data hU).E = Nat.card ↥data.W1 := by
change Nat.card ↥(data.W1.subgroupOf (data.U ⊔ data.W1)) = Nat.card ↥data.W1
exact Nat.card_congr (Subgroup.subgroupOfEquivOfLe le_sup_right).toEquiv
have hUfix : Nat.card ↥(typeP_coprimeAction data hU).fixedByU
= Nat.card ↥(data.H ⊓ Subgroup.centralizer (data.U : Set G)) := by
change Nat.card ↥(fixedSubgroup (typeP_conjAction data) (data.U.subgroupOf (data.U ⊔ data.W1)))
= _
exact typeP_card_fixedSubgroup data le_sup_left
have hEfix : Nat.card ↥(typeP_coprimeAction data hU).fixedByE = Nat.card ↥data.W2 := by
change Nat.card ↥(fixedSubgroup (typeP_conjAction data) (data.W1.subgroupOf (data.U ⊔ data.W1)))
= _
rw [typeP_card_fixedSubgroup data le_sup_right, typeP_H_inf_centralizer_W1 data]
have hUEfix : Nat.card ↥(typeP_coprimeAction data hU).fixedByUE
= Nat.card ↥(data.H ⊓ Subgroup.centralizer ((data.U ⊔ data.W1 : Subgroup G) : Set G)) := by
change Nat.card ↥(fixedSubgroup (typeP_conjAction data) ⊤) = _
rw [← Subgroup.subgroupOf_self (data.U ⊔ data.W1)]
exact typeP_card_fixedSubgroup data le_rfl
rw [hUEfix, hEcard, hEfix, hUfix] at key
exact key
/-- `C_H(U W₁) ≤ W₂`: the fixed points of the whole Frobenius group lie in `C_H(W₁) = W₂`. -/
theorem typeP_centralizer_uW1_le_W2 (data : TypePData M) :
data.H ⊓ Subgroup.centralizer ((data.U ⊔ data.W1 : Subgroup G) : Set G) ≤ data.W2 := by
rw [← typeP_H_inf_centralizer_W1 data]
exact inf_le_inf_left data.H (Subgroup.centralizer_le (SetLike.coe_subset_coe.mpr le_sup_right))
/-- `|H| ⟂ |W₁|`: `H` is Hall in `M` and `|W₁| = [M : M']` divides `[M : H]`. -/
theorem typeP_coprime_H_W1 [Finite G] (data : TypePData M) :
Nat.Coprime (Nat.card ↥data.H) (Nat.card ↥data.W1) := by
have hHleM : data.H ≤ M := data.H_le.trans (derivedInG_le_self M)
have hHall := OddOrder.BG.Ch4.S15.maxNilpotentNormalHall_isHall M
rw [← data.H_eq] at hHall
have hcop_idx := hHall.coprime_index
rw [Nat.card_congr (Subgroup.subgroupOfEquivOfLe hHleM).toEquiv] at hcop_idx
have hW1idx : ((derivedInG M).subgroupOf M).index = Nat.card ↥data.W1 := by
rw [data.M_complement.symm.index_eq_card,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe data.W1_le).toEquiv]
have hdvd : Nat.card ↥data.W1 ∣ (data.H.subgroupOf M).index := by
rw [← hW1idx]; exact Subgroup.index_dvd_of_le (Subgroup.comap_mono data.H_le)
exact hcop_idx.coprime_dvd_right hdvd
/-- `|U| ⟂ |W₁|`: from the Frobenius group `U W₁`. -/
theorem typeP_coprime_U_W1 [Finite G] (data : TypePData M) (hU : data.U ≠ ⊥) :
Nat.Coprime (Nat.card ↥data.U) (Nat.card ↥data.W1) := by
have h := (typeP_uW1_frobenius data hU).coprime_card_kernel_complement
rwa [Nat.card_congr (Subgroup.subgroupOfEquivOfLe le_sup_left).toEquiv,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe le_sup_right).toEquiv] at h
/-- **Peterfalvi (8.5.b)**: if `U ≠ 1`, then `U` does not centralize `H`.
If `U ≤ C(H)`, then `F(M) = H ⊔ (U ⊓ C(H)) = H ⊔ U = M'` (`fitting_eq`,
`derivedInG_eq_fitting_sup_U`), so `M' = derivedInG M` is nilpotent (the image of the Fitting
subgroup). But `M'` is also a normal Hall subgroup of `M` (`|M'| = |H|·|U|` is coprime to
`[M : M'] = |W₁|`, since `H` is Hall and `U W₁` is Frobenius), so `M' ≤ M_F = H`
(`le_maxNilpotentNormalHall`); then `U ≤ M' ≤ H` with `H ⊓ U = ⊥` forces `U = ⊥`. -/
theorem typeP_U_not_centralizes_H [Finite G] (data : TypePData M) (hU : data.U ≠ ⊥) :
¬ data.U ≤ Subgroup.centralizer (data.H : Set G) := by
intro hUC
apply hU
have hM'M : derivedInG M ≤ M := derivedInG_le_self M
-- `M' = H ⊔ U` and `F(M) = M'`.
have hM'eq : derivedInG M = data.H ⊔ data.U := by
rw [data.derivedInG_eq_fitting_sup_U, data.H_eq]
have hFit : (OddOrder.Isaacs.Ch01.fitting ↥M).map M.subtype = derivedInG M := by
rw [data.fitting_eq, inf_eq_left.mpr hUC, hM'eq]
-- `M'` is nilpotent (image of `F(↥M)`).
have : Group.IsNilpotent ↥(OddOrder.Isaacs.Ch01.fitting (↥M : Type _)) :=
OddOrder.Isaacs.Ch01.fitting.isNilpotent
have hM'nil : Group.IsNilpotent ↥(derivedInG M) := by
rw [← hFit]
exact Group.nilpotent_of_mulEquiv (Subgroup.equivMapOfInjective (OddOrder.Isaacs.Ch01.fitting
↥M)
M.subtype M.subtype_injective)
-- `M'.subgroupOf M` is the commutator subgroup of `↥M`: normal and nilpotent.
have hM'sub : (derivedInG M).subgroupOf M = commutator ↥M :=
Subgroup.comap_map_eq_self_of_injective M.subtype_injective (commutator ↥M)
have hM'norm : ((derivedInG M).subgroupOf M).Normal := by rw [hM'sub]; infer_instance
have : Group.IsNilpotent ↥((derivedInG M).subgroupOf M) :=
Group.nilpotent_of_mulEquiv (Subgroup.subgroupOfEquivOfLe hM'M).symm
-- `M'` is a normal Hall subgroup of `M`, so `M' ≤ M_F = H`.
have hidxM' : ((derivedInG M).subgroupOf M).index = Nat.card ↥data.W1 := by
rw [data.M_complement.symm.index_eq_card,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe data.W1_le).toEquiv]
have hcop : Nat.Coprime (Nat.card ↥(derivedInG M)) (Nat.card ↥data.W1) := by
rw [show Nat.card ↥(derivedInG M) = Nat.card ↥data.H * Nat.card ↥data.U by
rw [← data.derived_complement.card_mul_card,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe data.H_le).toEquiv,
Nat.card_congr (Subgroup.subgroupOfEquivOfLe data.U_le).toEquiv]]
exact (Nat.Coprime.mul_right (typeP_coprime_H_W1 data).symm
(typeP_coprime_U_W1 data hU).symm).symm
have hcardSub : Nat.card ↥((derivedInG M).subgroupOf M) = Nat.card ↥(derivedInG M) :=
Nat.card_congr (Subgroup.subgroupOfEquivOfLe hM'M).toEquiv
have hM'hall : OddOrder.Isaacs.Ch03.IsHallSubgroup
(Nat.card ↥(derivedInG M)).primeFactors ((derivedInG M).subgroupOf M) := by
refine ⟨fun p hp => hcardSub ▸ hp, fun p hp hpM' => ?_⟩
rw [hidxM'] at hp
have hp1 : p ∣ 1 :=
hcop ▸ Nat.dvd_gcd (Nat.dvd_of_mem_primeFactors hpM') (Nat.dvd_of_mem_primeFactors hp)
exact (Nat.prime_of_mem_primeFactors hp).ne_one (Nat.dvd_one.mp hp1)
have hM'leH : derivedInG M ≤ data.H := by
rw [data.H_eq]
exact OddOrder.BG.Ch4.S15.le_maxNilpotentNormalHall hM'M hM'norm inferInstance hM'hall
have hUleH : data.U ≤ data.H := (le_sup_right.trans hM'eq.ge).trans hM'leH
rw [← inf_of_le_left hUleH, inf_comm, typeP_H_inf_U data]
/-- **Peterfalvi (8.5.b), first part**: `[U, U]` centralizes `H`, so `Ū = U/C_U(H)` is abelian — the
structural input to Peterfalvi (9.7) case (b).
`[U,U] ⊆ M'' = ⁅M', M'⁆` (`commutator_mono`, `U ≤ M' = derivedInG M`), and `M'' ⊆ F(M) =
H ⊔ (U ⊓ C_M(H))` (`secondDerived_le_fitting`). Since `[U,U] ⊆ U`, `U ⊓ H = ⊥`, and `K = U ⊓
C_M(H)`
centralizes (hence normalizes) `H` — so `↑(H ⊔ K) = ↑H · ↑K` — any `x ∈ [U,U]` writes `x = h·c` with
`h ∈ H`, `c ∈ K ⊆ U`; then `h = x·c⁻¹ ∈ U ⊓ H = ⊥`, so `x = c ∈ K ⊆ C_M(H)`. -/
theorem typeP_commutator_U_centralizes_H (data : TypePData M) :
⁅data.U, data.U⁆ ≤ Subgroup.centralizer (data.H : Set G) := by
set K := data.U ⊓ Subgroup.centralizer (data.H : Set G) with hK
-- `derivedInG J = ⁅J, J⁆`, so `M'' = secondDerivedInAmbient M = ⁅M', M'⁆`.
have hderiv : ∀ J : Subgroup G, derivedInG J = ⁅J, J⁆ := fun J => by
change (commutator ↥J).map J.subtype = ⁅J, J⁆
rw [commutator_def, Subgroup.map_commutator]
simp only [← MonoidHom.range_eq_map, Subgroup.range_subtype]
have hUU_M'' : ⁅data.U, data.U⁆ ≤ secondDerivedInAmbient M := by
change ⁅data.U, data.U⁆ ≤ derivedInG (derivedInG M)
rw [hderiv (derivedInG M)]
exact Subgroup.commutator_mono data.U_le data.U_le
have hUU_HK : ⁅data.U, data.U⁆ ≤ data.H ⊔ K := hUU_M''.trans data.secondDerived_le_fitting
have hUU_U : ⁅data.U, data.U⁆ ≤ data.U :=
Subgroup.commutator_le.mpr fun a ha b hb => by
rw [commutatorElement_def]
exact Subgroup.mul_mem _ (Subgroup.mul_mem _ (Subgroup.mul_mem _ ha hb)
(Subgroup.inv_mem _ ha)) (Subgroup.inv_mem _ hb)
-- `K ≤ C_M(H) ≤ N(H)`, so `↑(H ⊔ K) = ↑H · ↑K`.
have hKnorm : K ≤ Subgroup.normalizer data.H :=
inf_le_right.trans (Subgroup.centralizer_le_normalizer (data.H : Set G))
intro x hx
have hxU : x ∈ data.U := hUU_U hx
have hxHK : x ∈ (↑(data.H ⊔ K) : Set G) := hUU_HK hx
rw [Subgroup.coe_mul_of_right_le_normalizer_left data.H K hKnorm] at hxHK
obtain ⟨h, hh, c, hc, heq⟩ := Set.mem_mul.mp hxHK
have hcU : c ∈ data.U := (inf_le_left : K ≤ data.U) hc
have hhU : h ∈ data.U := by
have hheq : h = x * c⁻¹ := by rw [← heq]; group
rw [hheq]; exact Subgroup.mul_mem _ hxU (Subgroup.inv_mem _ hcU)
have hh1 : h = 1 := by
have hmem : h ∈ data.U ⊓ data.H := ⟨hhU, hh⟩
rw [inf_comm, typeP_H_inf_U data] at hmem
exact Subgroup.mem_bot.mp hmem
have hxc : x = c := by rw [← heq, hh1, one_mul]
rw [hxc]
exact (inf_le_right : K ≤ Subgroup.centralizer (data.H : Set G)) hc
end Wielandt93
/-- Conjugation preserves coatoms (local copy; avoids a cross-area dependency on BG Ch3). -/
private theorem isCoatom_conj_smul {g : G} {M : Subgroup G} (h : IsCoatom M) :
IsCoatom (MulAut.conj g • M) := by
have hMeq : M = MulAut.conj g⁻¹ • (MulAut.conj g • M) := by
rw [← mul_smul, ← map_mul, inv_mul_cancel, map_one, one_smul]
constructor
· intro htop
apply h.1
rw [eq_top_iff]
intro x _
have hx : g * x * g⁻¹ ∈ MulAut.conj g • M := by
rw [htop]; exact Subgroup.mem_top _
rw [Subgroup.mem_pointwise_smul_iff_inv_smul_mem] at hx
simpa [MulAut.smul_def, mul_assoc] using hx
· intro b hb
have hle : M ≤ MulAut.conj g⁻¹ • b := by
rw [hMeq]; exact Subgroup.pointwise_smul_le_pointwise_smul_iff.mpr hb.le
have h1 : M < MulAut.conj g⁻¹ • b := by
refine lt_of_le_of_ne hle (fun heq => hb.ne ?_)
rw [heq, ← mul_smul, ← map_mul, mul_inv_cancel, map_one, one_smul]
have h2 := h.2 _ h1
have h3 := congrArg (fun K => MulAut.conj g • K) h2
simpa [← mul_smul, ← map_mul, mul_inv_cancel] using h3
/-- A maximal subgroup of a minimal simple group is self-normalizing (`N_G(M) = M`): `N_G(M) ⊇ M`
is a coatom, so it is `M` or `⊤`; the latter makes `M` normal, hence `⊥` or `⊤` by simplicity —
impossible for a nontrivial proper maximal subgroup. -/
private theorem maximal_normalizer_eq_self [Finite G] (hG : OddOrder.BG.IsMinimalSimpleOdd G)
{M : Subgroup G} (hM : M ∈ maximalSubgroups G) :
Subgroup.normalizer (M : Set G) = M := by
have hco : IsCoatom M := hM
refine le_antisymm ?_ Subgroup.le_normalizer
by_contra hle
have hlt : M < Subgroup.normalizer (M : Set G) :=
lt_of_le_of_ne Subgroup.le_normalizer (fun heq => hle heq.ge)
have hnormal : M.Normal := Subgroup.normalizer_eq_top_iff.mp (hco.2 _ hlt)
rcases hG.simple.eq_bot_or_eq_top_of_normal M hnormal with hb | ht
· exact hG.ne_bot_of_isCoatom hco hb
· exact hco.1 ht
/-- **§8 input to (9.3)** — Peterfalvi (8.6.b II) + (8.12.b): for a maximal subgroup `M` of type II,
the Frobenius kernel `U` acts fixed-point-freely on `H = M_F`, i.e. `C_H(U) = 1`.
*Proof.* If `C_H(U) ≠ 1` then, taking `X = U#` (`C_G(X) = C_G(U)`), Peterfalvi (8.12.b)
(`S10.typeI_or_typeII_centralizer_unique_hall`) makes `M` the unique maximal subgroup containing
`C_G(U)`. Since `g ∈ N_G(U)` fixes `C_G(U)` (it normalizes `U`), conjugation by `g` carries that
unique maximal to itself, so `g` normalizes `M`; as `M` is self-normalizing,
`N_G(U) ⊆ M`. This contradicts (8.6.b II) (`S10.typeII_normalizer_not_le_of_typePData`), so
`C_H(U) = 1`. -/
theorem typeII_centralizer_U_eq_bot [Finite G] (hG : OddOrder.BG.IsMinimalSimpleOdd G)
{M : Subgroup G} (data : TypePData M) (hM : M ∈ maximalSubgroups G) (hII : IsTypeII M) :
data.H ⊓ Subgroup.centralizer (data.U : Set G) = ⊥ := by
by_contra hne
have hP2 : OddOrder.BG.Ch4.S14.IsTypeP2 M :=
((OddOrder.BG.Ch4.S16.proposition_type_classification hG hM).2.1.mp hII)
obtain ⟨dataII⟩ := hII
-- `data.U ≠ ⊥`: `|U| = [M' : M_F]` is independent of the type-`P` witness, and the `IsTypeII`
-- witness has `U ≠ 1`.
have hU_ne : data.U ≠ ⊥ := by
have hcard : Nat.card ↥data.U = Nat.card ↥dataII.typeP.U := by
rw [data.card_U_eq_index, dataII.typeP.card_U_eq_index]
refine fun hbot => dataII.common.1 (Subgroup.card_eq_one.mp ?_)
rw [← hcard, hbot, Subgroup.card_bot]
-- `C_G(U#) = C_G(U)` (the identity is centralized by everyone).
have hCeq : Subgroup.centralizer (sharpSubgroup data.U)
= Subgroup.centralizer (data.U : Set G) := by
refine le_antisymm ?_ (Subgroup.centralizer_le Set.sdiff_subset)
intro g hg
rw [Subgroup.mem_centralizer_iff] at hg ⊢
intro h hh
rcases eq_or_ne h 1 with rfl | h1
· simp
· exact hg h ⟨hh, h1⟩
-- `X = U#` is nonempty and `C_H(X) ≠ 1`, so (8.12.b) makes `M` the unique maximal `∋ C_G(X)`.
obtain ⟨x, hxU, hx1⟩ := (Subgroup.bot_or_exists_ne_one data.U).resolve_left hU_ne
have hUleM : data.U ≤ M := data.U_le.trans (Subgroup.map_subtype_le _)
have hCHne : maxNilpotentNormalHall M ⊓ Subgroup.centralizer (sharpSubgroup data.U) ≠ ⊥ := by
rw [← data.H_eq, hCeq]; exact hne
have hUhall : Ch03.IsHallSubgroup
((OddOrder.BG.Ch4.S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ) (data.U.subgroupOf M) :=
OddOrder.Peterfalvi.S10Interface.isHall_kappaSigmaCompl_of_isTypeP2_complement hG hM hP2
hUleM data.derivedInG_eq_fitting_sup_U data.fitting_inf_U_eq_bot
obtain ⟨hCleM, hUniq⟩ :=
OddOrder.Peterfalvi.S10.typeI_or_typeII_centralizer_unique_hall hG hM (Or.inr ⟨dataII⟩)
hUleM hUhall (sharpSubgroup data.U) ⟨x, hxU, hx1⟩ le_rfl hCHne
-- `N_G(U) ⊆ M`.
have hNleM : Subgroup.normalizer (data.U : Set G) ≤ M := by
intro g hg
have hgU : MulAut.conj g • data.U = data.U := conj_smul_eq_self_of_mem_normalizer hg
have hgC : MulAut.conj g • Subgroup.centralizer (sharpSubgroup data.U)
= Subgroup.centralizer (sharpSubgroup data.U) := by
rw [centralizer_pointwise_smul, image_sharpSubgroup, hgU]
have hgM : MulAut.conj g • M = M := by
have hCle' : Subgroup.centralizer (sharpSubgroup data.U) ≤ MulAut.conj g • M := by
rw [← hgC]; exact Subgroup.pointwise_smul_le_pointwise_smul_iff.mpr hCleM
exact hUniq.eq_of_isCoatom_of_le (isCoatom_conj_smul hM) hCle' hM hCleM
have hmem : g ∈ Subgroup.normalizer (M : Set G) := by
rw [Subgroup.mem_normalizer_iff]
intro h
have key : MulAut.conj g • h ∈ MulAut.conj g • M ↔ h ∈ M :=
Subgroup.smul_mem_pointwise_smul_iff
rw [hgM] at key
simpa [MulAut.smul_def, MulAut.conj_apply] using key.symm
rwa [maximal_normalizer_eq_self hG hM] at hmem
exact OddOrder.Peterfalvi.S10.typeII_normalizer_not_le_of_typePData hG data hM ⟨dataII⟩ hNleM
/-- **§8 input to (9.3)** — Peterfalvi Theorem (8.8): for a maximal subgroup `M` of type III or IV,
the order `|W₂|` is prime.
*Discharge route:* (8.8) supplies a type-II maximal `S` with `|S : S'| = |W₂|`
(`OddOrder.Peterfalvi.S12.Hypothesis.exists_typeII_maximal_with_w2`, resting on
`theorem88_caseB_holds`), and a type-II cyclic factor has prime order
(`theorem88_caseB_prime_orders`). The §8/§12 obligation. -/
theorem typeIIIorIV_W2_prime [Finite G] (hG : OddOrder.BG.IsMinimalSimpleOdd G)
{M : Subgroup G} (data : TypePData M) (hM : M ∈ maximalSubgroups G)
(hIIIIV : IsTypeIII M ∨ IsTypeIV M) :
(Nat.card ↥data.W2).Prime := by
-- (8.8)/(8.13) for `M` (`S10.exists_typeII_maximal_with_w2_of_typeP`): there is a Type-II maximal
-- subgroup `S` with `|S : [S,S]| = |W₂|`. A Type-II maximal's cyclic factor `W₁(S)` has prime
-- order (8.6.a, carried by `TypePNontrivialCore`) and equals that index
-- (`card_W1_eq_derived_index`), so `|W₂|` is prime. (Mirrors `S12.Hypothesis.w2_prime`.)
obtain ⟨S, -, hSII, hindex⟩ :=
OddOrder.Peterfalvi.S10.exists_typeII_maximal_with_w2_of_typeP hG data hM
(hIIIIV.elim Or.inl (fun h => Or.inr (Or.inl h)))
obtain ⟨dataII⟩ := hSII
have hcard : Nat.card ↥dataII.typeP.W1 = Nat.card ↥data.W2 := by
rw [dataII.typeP.card_W1_eq_derived_index]; exact hindex
rw [← hcard]
exact dataII.common.2.1
/-- **Peterfalvi (9.3)**: the order and centralizer alternatives for type II
versus types III/IV.
*Proof.* The order relations are the Wielandt fixed-point identity (9.1)
`typeP_wielandt_order_relation` (`|C_H(U W₁)|^q · |H| = |W₂|^q · |C_H(U)|`) specialised by the
fixed-point-free §8 inputs. For type II, `C_H(U) = 1` (Peterfalvi (8.6.b II)+(8.12.b): a nontrivial
`C_H(U)` would put `N_G(U)` both inside and outside `M`); then `C_H(U W₁) ⊆ C_H(U) = 1` and the
identity gives `|H| = |W₂|^q`. For types III/IV, `|W₂| = p` is prime (Theorem (8.8): there is a
type-II maximal `S` with `|S : S'| = |W₂|`), and then `C_H(U W₁) = 1` is *derived* from Peterfalvi
(8.5.b) (`typeP_U_not_centralizes_H`): `C_H(U W₁) ⊆ C_H(W₁) = W₂` has order dividing the prime `p`,
so
if nontrivial it equals `W₂` and the identity forces `|H| = |C_H(U)|`, i.e. `U` centralizes `H`,
contradicting (8.5.b); the identity then gives `|H| = p^q · |C_H(U)|`. The remaining §8 obligations
are exactly: type II `C_H(U) = 1`, and types III/IV `|W₂|` prime and `U ≠ 1` — the order arithmetic
and the type III/IV `C_H(U W₁) = 1` reduction are the content discharged here. -/
theorem typeII_III_IV_order_relations [Finite G] (hG : OddOrder.BG.IsMinimalSimpleOdd G)
{M : Subgroup G} (data : TypesIIIIIIVSetup M) :
(IsTypeII M →
data.H ⊓ Subgroup.centralizer (data.U : Set G) = ⊥ ∧
Nat.card ↥data.H = Nat.card ↥data.W2 ^ data.q) ∧
((IsTypeIII M ∨ IsTypeIV M) →
∃ p : ℕ, p.Prime ∧ Nat.card ↥data.W2 = p ∧
data.H ⊓ Subgroup.centralizer ((data.U ⊔ data.W1 : Subgroup G) : Set G) = ⊥ ∧
Nat.card ↥data.H =
p ^ data.q * Nat.card ↥(data.H ⊓ Subgroup.centralizer (data.U : Set G))) := by
have hH_ne : data.typeP.H ≠ ⊥ := fun heq => data.typeP.H_noncyclic (heq ▸ inferInstance)
refine ⟨fun _hII => ?_, fun _hIII_IV => ?_⟩
· -- **Type II.** §8 input: `C_H(U) = 1` (Peterfalvi (8.6.b II) + (8.12.b)), cited as the
-- `typeII_centralizer_U_eq_bot` obligation.
have hU : data.typeP.U ≠ ⊥ := data.nontrivial.1
have hCU : data.typeP.H ⊓ Subgroup.centralizer (data.typeP.U : Set G) = ⊥ :=
typeII_centralizer_U_eq_bot hG data.typeP data.maximal _hII
-- `C_H(U W₁) ⊆ C_H(U) = 1`.
have hmono : Subgroup.centralizer ((data.typeP.U ⊔ data.typeP.W1 : Subgroup G) : Set G)
≤ Subgroup.centralizer (data.typeP.U : Set G) :=
Subgroup.centralizer_le (SetLike.coe_subset_coe.mpr le_sup_left)
have hCUW : data.typeP.H
⊓ Subgroup.centralizer ((data.typeP.U ⊔ data.typeP.W1 : Subgroup G) : Set G) = ⊥ :=
le_bot_iff.mp (hCU ▸ inf_le_inf_left data.typeP.H hmono)
refine ⟨hCU, ?_⟩
have key := typeP_wielandt_order_relation data.typeP hU
rwa [hCUW, hCU, Subgroup.card_bot, one_pow, one_mul, mul_one] at key
· -- **Types III/IV.** Sole §8 input: `|W₂| = p` prime (Theorem (8.8)). `U ≠ 1` is the setup's
-- `nontrivial` core, and `C_H(U W₁) = 1` is then *derived* from (8.5.b).
obtain ⟨p, hp_prime, hpW2⟩ : ∃ p : ℕ, p.Prime ∧ Nat.card ↥data.typeP.W2 = p :=
⟨_, typeIIIorIV_W2_prime hG data.typeP data.maximal _hIII_IV, rfl⟩
have hU : data.typeP.U ≠ ⊥ := data.nontrivial.1
-- `C_H(U W₁) = 1`: it lies in `C_H(W₁) = W₂` of prime order `p`; were it nontrivial it would
-- equal `W₂`, and Wielandt's identity would force `|H| = |C_H(U)|`, i.e. `U` centralizes `H`,
-- against (8.5.b) (`typeP_U_not_centralizes_H`).
have hCUW : data.typeP.H
⊓ Subgroup.centralizer ((data.typeP.U ⊔ data.typeP.W1 : Subgroup G) : Set G) = ⊥ := by
by_contra hne
have hdvd : Nat.card ↥(data.typeP.H
⊓ Subgroup.centralizer ((data.typeP.U ⊔ data.typeP.W1 : Subgroup G) : Set G))
∣ Nat.card ↥data.typeP.W2 :=
Subgroup.card_dvd_of_le (typeP_centralizer_uW1_le_W2 data.typeP)
have hcardCUW : Nat.card ↥(data.typeP.H
⊓ Subgroup.centralizer ((data.typeP.U ⊔ data.typeP.W1 : Subgroup G) : Set G)) = p := by
rw [hpW2] at hdvd
refine (hp_prime.eq_one_or_self_of_dvd _ hdvd).resolve_left (fun h1 => hne ?_)
exact Subgroup.card_eq_one.mp h1
have key := typeP_wielandt_order_relation data.typeP hU
rw [hcardCUW, hpW2] at key
have hHC : Nat.card ↥data.typeP.H
= Nat.card ↥(data.typeP.H ⊓ Subgroup.centralizer (data.typeP.U : Set G)) :=
Nat.eq_of_mul_eq_mul_left (pow_pos hp_prime.pos _) key
have hHleCU : data.typeP.H ≤ Subgroup.centralizer (data.typeP.U : Set G) := by
have : data.typeP.H ⊓ Subgroup.centralizer (data.typeP.U : Set G) = data.typeP.H :=
Subgroup.eq_of_le_of_card_ge inf_le_left hHC.le
rw [← this]; exact inf_le_right
exact typeP_U_not_centralizes_H data.typeP hU (Subgroup.le_centralizer_iff.mp hHleCU)
refine ⟨p, hp_prime, hpW2, hCUW, ?_⟩
have key := typeP_wielandt_order_relation data.typeP hU
rwa [hCUW, hpW2, Subgroup.card_bot, one_pow, one_mul] at key
end OddOrder.Peterfalvi.S11
Source
Yawara Ishida, Lean 4 formalization of the Feit–Thompson (odd order) theorem, github.com/yawara/odd-order @ 20a6fccbb63c (Apache-2.0), https://github.com/yawara/odd-order/blob/20a6fccbb63c0b138acda9f7fafd798c42f9441b/OddOrder/Peterfalvi/S11_MaximalII_III_IV/WielandtSetupBasic.lean