Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

BG §16: Theorems A and B, the component layer

Definition
OddOrder_BG_Ch4_FamilyOfMaximal_S16_MainResults_TheoremAB

by burkh4rt · Nov 12, 2026 · Mathlib 0df444a (Lean v4.33.1)

feit-thompsonfinite-groupsgroup-theoryishida-odd-order

The component layer for Theorems A and B of Bender–Glauberman, Local Analysis for the Odd Order Theorem, §16: the ungated conjuncts of Theorem A, the parts of Theorem B (a Sylow abelian UUU of rank at most 2, the TI-subset statement, tameness) and the bridging lemmas between the κ\kappaκ- and σ\sigmaσ-complements. It consists of lemmas only.

The bundle consists of lemmas only.

The bundle also proves 10 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/BG/Ch4_FamilyOfMaximal/S16_MainResults/TheoremAB.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_BG_Ch4_FamilyOfMaximal_S16_MainResults_Notation

namespace OddOrder.BG.Ch4.S16

open OddOrder.GroupTheory

open OddOrder.Isaacs

open OddOrder.BG.Ch3.S12

open OddOrder.BG.Ch4.S14

open OddOrder.BG.Ch4.S15

open scoped Pointwise

variable {G : Type*} [Group G]

/-- **BG Theorem A(3) decomposition** (mmd L4276): `M = K U M_σ`.  For a maximal `M` with Hall
`κ`-subgroup `K ≤ M` and Hall `(κ ∪ σ)'`-subgroup `U ≤ M`, the three factors join to all of `M`.

Type-F (`K = ⊥`): `M = U M_σ` from the `K = ⊥` `SubgroupESetup` (`E_compl_sup`).  Type-P (`K ≠ ⊥`):
`M' = U M_σ` and `M'` complements `K` in `M` (`typeP_auxiliary_structure`, Lemma 15.1(b)/Theorem
14.7(h)), so `M = M' K = (U M_σ) K`; the complement's `H ⊔ K = ⊤` is pushed from `M` to `G` via
`Subgroup.map M.subtype`.  This is the standalone proof of conjunct 3 used by
`theoremA_maximal_structure_faithful`. -/
theorem typeP_maximal_eq_kappaHall_sup_U_sup_Msigma [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M K U : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hKM : K ≤ M) (hUM : U ≤ M) (hK : Ch03.IsHallSubgroup (S14.kappa M) (K.subgroupOf M))
    (hU : Ch03.IsHallSubgroup ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ) (U.subgroupOf M)) :
    M = K ⊔ U ⊔ OddOrder.BG.Ch3.S10.Msigma M := by
  by_cases hKbot : K = ⊥
  · -- Type-F: `M = U M_σ`.
    obtain ⟨_E₁, _E₂, _E₃, hsetup⟩ :=
      subgroupESetup_of_isHall_kappa_eq_bot hG hM hKM hUM hK hKbot hU
    rw [hKbot, bot_sup_eq, sup_comm]
    exact hsetup.E_compl_sup.symm
  · -- Type-P: `M = M' K`, `M' = U M_σ`.
    have haux := typeP_auxiliary_structure hG hM hKM hUM hK rfl hU
    obtain ⟨hM'eq, _, hcompl, _⟩ := haux.2.2.2.2.1 hKbot
    have hMle : derivedInG M ≤ M := Subgroup.map_subtype_le _
    have hMM' : derivedInG M ⊔ K = M := by
      have htop : (derivedInG M ⊔ K).subgroupOf M = ⊤ := by
        rw [Subgroup.subgroupOf_sup hMle hKM]; exact hcompl.sup_eq_top
      exact le_antisymm (sup_le hMle hKM) (Subgroup.subgroupOf_eq_top.mp htop)
    -- `M = M' K = (U M_σ) K = K U M_σ` (a pure `⊔`-AC rewrite, no lattice `whnf`).
    conv_lhs => rw [← hMM']
    rw [hM'eq, sup_comm (U ⊔ OddOrder.BG.Ch3.S10.Msigma M) K, ← sup_assoc]

/-- **BG Theorem A — the ungated conjuncts** (mmd L4274), as a standalone `sorry`-free lemma.

Bundles the conjuncts used by `theoremA_maximal_structure_faithful` whose upstreams are all proved
transitively
(mirroring `theoremB_U_sylow_abelian_rank_le_two` / `sigma_reps_pairwise_disjoint`):

* A(1) `M_σ` is a `σ(M)`-Hall subgroup (`Msigma_isHall`);
* A(2) `K` cyclic (`typeP_auxiliary_structure`, Lemma 15.1(a)/Theorem 14.7(h));
* A(3) `M ≤ N_G(U M_σ)` (`typeP_auxiliary_structure`: `U M_σ = M` for type-F, `= M' ⊴ M`
  for type-P);
* A(4) `C_U(k) = 1` for `k ∈ K#` (`typeP_hall_inf_centralizer_kappaElement_eq_bot`);
* A(5) `Kstar ≠ ⊥` (`S14.typeP_structure`, with `K = ⊥` collapsing `Kstar = M_σ ≠ ⊥`) and the
  element form `C_M(k) = K ⊔ K*` for `K ≠ ⊥`, `k ∈ K#` (`typeP_centralizer_kappaElement_eq`);
* A(6) `M_F ≤ M_σ ≤ M'` (`maxNilpotentNormalHall_le_Msigma`, `Msigma_le_derived`).

The remaining faithful-monolith conjuncts are supplied separately: `M = K U M_σ` by
`typeP_maximal_eq_kappaHall_sup_U_sup_Msigma`, `M'' ≤ F(M)` by
`derivedDerived_le_fittingInAmbient`, and the `M_F ≠ M_σ` extreme A(8) by
`theoremA8_structure`.

As with the Theorem B(1) precedent the explicit `hKM : K ≤ M`, `hUM : U ≤ M` are added (part of the
BG setup `M = K U M_σ` but not forced by the Hall conditions on
`K.subgroupOf M`/`U.subgroupOf M`). -/
theorem theoremA_ungated_conjuncts [Finite G] (hG : OddOrder.BG.IsMinimalSimpleOdd G)
    {M K Kstar U : Subgroup G} (hM : M ∈ maximalSubgroups G) (hKM : K ≤ M) (hUM : U ≤ M)
    (hK : Ch03.IsHallSubgroup (S14.kappa M) (K.subgroupOf M))
    (hKstar : Kstar = OddOrder.BG.Ch3.S10.Msigma M ⊓ Subgroup.centralizer (K : Set G))
    (hU : Ch03.IsHallSubgroup ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ)
      (U.subgroupOf M)) :
    Ch03.IsHallSubgroup (OddOrder.BG.Ch3.S10.sigma M) (OddOrder.BG.Ch3.S10.Msigma M) ∧
      IsCyclic ↥K ∧
      M ≤ Subgroup.normalizer ((U ⊔ OddOrder.BG.Ch3.S10.Msigma M : Subgroup G) : Set G) ∧
      (∀ k ∈ K, k ≠ 1 → U ⊓ Subgroup.centralizer ({k} : Set G) = ⊥) ∧
      Kstar ≠ ⊥ ∧
      (K ≠ ⊥ → ∀ k ∈ K, k ≠ 1 → M ⊓ Subgroup.centralizer ({k} : Set G) = K ⊔ Kstar) ∧
      S15.MF M ≤ OddOrder.BG.Ch3.S10.Msigma M ∧
      OddOrder.BG.Ch3.S10.Msigma M ≤ derivedInG M := by
  -- `IsTypeP M` from a nonempty `κ`-Hall, packaged for the type-`P` conjuncts.
  have hPofne : K ≠ ⊥ → S14.IsTypeP M := fun hKne =>
    isTypeP_of_isHall_kappa_subgroupOf_ne_bot hK
      (fun h => hKne (by rw [← Subgroup.map_subgroupOf_eq_of_le hKM, h, Subgroup.map_bot]))
  have haux := typeP_auxiliary_structure hG hM hKM hUM hK hKstar hU
  have hKcyc : IsCyclic ↥K := haux.2.1
  refine ⟨OddOrder.BG.Ch3.S10.Msigma_isHall hG hM, hKcyc, haux.1, ?_, ?_, ?_,
    maxNilpotentNormalHall_le_Msigma hG hM, OddOrder.BG.Ch3.S10.Msigma_le_derived hG hM⟩
  · -- A(4): `C_U(k) = 1`.  Vacuous for `K = ⊥`; else type-`P` element centralizer.
    intro k hk hk1
    by_cases hKbot : K = ⊥
    · rw [hKbot, Subgroup.mem_bot] at hk; exact absurd hk hk1
    · exact typeP_hall_inf_centralizer_kappaElement_eq_bot hG hM (hPofne hKbot) hKM hUM hK hKstar hU
        k hk hk1
  · -- A(5): `Kstar ≠ ⊥`.
    by_cases hKbot : K = ⊥
    · -- type-F: `Kstar = M_σ ⊓ C(1) = M_σ ≠ ⊥`.
      subst hKbot
      have hc : Subgroup.centralizer ((⊥ : Subgroup G) : Set G) = ⊤ := by
        rw [Subgroup.coe_bot, Subgroup.centralizer_eq_top_iff_subset]
        exact Set.singleton_subset_iff.mpr (Subgroup.one_mem _)
      rw [hc, inf_top_eq] at hKstar
      rw [hKstar]
      exact OddOrder.BG.Ch3.S10.Msigma_ne_bot hG hM
    · exact (S14.typeP_structure hG hM (hPofne hKbot) hKM hK hKstar hU).2.1
  · -- A(5) element form: `C_M(k) = K ⊔ K*` for `K ≠ ⊥`.
    intro hKne k hk hk1
    exact typeP_centralizer_kappaElement_eq hG hM (hPofne hKne) hKM hK hKstar hU k hk hk1

/-- **BG Theorem B(1)** (mmd L4373): every Sylow subgroup of `U` is abelian of rank at most two.

Standalone, `sorry`-free form of the first conjunct of `theoremB_U_and_A_tame`, mirroring
`sigma_reps_pairwise_disjoint` (the Theorem E precedent).  Two faithfulness adjustments over the
scaffold conjunct:

* the explicit hypothesis `hUM : U ≤ M` — the §12 supporting-subgroup reduction
  (`exists_subgroupESetup_with_le`) genuinely needs `U ≤ M`, and `hU` (a Hall condition on
  `U.subgroupOf M`) does not force it; in BG `U ⊆ M` is part of the setup (`M = K U M_σ`);
* the restriction to a *prime* `p` — "Sylow subgroup" means a `p`-group for prime `p`; for a
  composite `p` the bare `IsPGroup p` hypothesis is met by non-abelian `{q, r}`-groups (e.g. a
  Frobenius `7 ⋊ 3`, possible in odd order), so the unrestricted `∀ p : ℕ` form is false.

Proof (cite-only over §12): a `p`-subgroup `P ≤ U` is a `σ(M)'`-subgroup of `M`, hence lies in a
supporting subgroup `E` with `SubgroupESetup M E …` (`exists_subgroupESetup_with_le`); then
`rank P ≤ rank E ≤ 2` (`rank_le_two`) and `P`, a prime `p`-group, is nilpotent, so abelian by
`nilpotent_sigmaComplement_abelian` (a). -/
theorem theoremB_U_sylow_abelian_rank_le_two [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M U : Subgroup G}
    (hM : M ∈ maximalSubgroups G) (hUM : U ≤ M)
    (hU : Ch03.IsHallSubgroup ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ)
      (U.subgroupOf M)) :
    ∀ p : ℕ, p.Prime → ∀ P : Subgroup G, P ≤ U → IsPGroup p ↥P →
      rank ↥P ≤ 2 ∧ IsMulCommutative ↥P := by
  intro p hp P hPU hPp
  have : Fact p.Prime := ⟨hp⟩
  have hPM : P ≤ M := hPU.trans hUM
  -- `P ≤ U` is a `σ(M)'`-subgroup of `M`.
  have hP_pi : Subgroup.IsPiSubgroup ((OddOrder.BG.Ch3.S10.sigma M)ᶜ) P := by
    intro q hq
    have hqU : q ∈ (Nat.card ↥U).primeFactors :=
      Nat.primeFactors_mono (Subgroup.card_dvd_of_le hPU) Nat.card_pos.ne' hq
    intro hqσ
    exact hU.1 q (by rwa [Nat.card_congr (Subgroup.subgroupOfEquivOfLe hUM).toEquiv])
      (Set.mem_union_right _ hqσ)
  -- §12 supporting subgroup `E ⊇ P`.
  obtain ⟨E, _E₁, _E₂, _E₃, hsetup, hPE, _hE_pi⟩ :=
    exists_subgroupESetup_with_le hG hM hPM hP_pi
  refine ⟨?_, ?_⟩
  · rw [rank_le_iff]
    intro q hq
    exact le_trans (pRank_le_of_injective (Subgroup.inclusion_injective hPE))
      ((rank_le_iff.mp (hsetup.rank_le_two hG)) q hq)
  · exact (nilpotent_sigmaComplement_abelian hG hsetup).1 P hPM hP_pi
      (IsPGroup.isNilpotent hPp)

/-- Centralizer of a cyclic subgroup = centralizer of a generator. -/
private theorem centralizer_zpowers_eq_singleton_B5 (z : G) :
    Subgroup.centralizer ((Subgroup.zpowers z : Subgroup G) : Set G)
      = Subgroup.centralizer ({z} : Set G) := by
  ext c
  simp only [Subgroup.mem_centralizer_iff]
  constructor
  · intro hc y hy
    rw [Set.mem_singleton_iff] at hy; rw [hy]
    exact hc z (SetLike.mem_coe.mpr (Subgroup.mem_zpowers z))
  · intro hc y hy
    obtain ⟨n, rfl⟩ := Subgroup.mem_zpowers_iff.mp (SetLike.mem_coe.mp hy)
    have hzc : Commute z c := hc z (Set.mem_singleton z)
    exact (hzc.zpow_left n).eq

/-- A maximal subgroup of a minimal simple group is self-normalizing (`N_G(M) = M`). -/
private theorem normalizer_eq_self_of_maximal_B5 [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

/-- Conjugation of `M` by one of its own elements fixes `M`. -/
private theorem conj_smul_self_of_mem {M : Subgroup G} {g : G} (hg : g ∈ M) :
    MulAut.conj g • M = M := by
  ext z
  rw [Subgroup.mem_pointwise_smul_iff_inv_smul_mem]
  have hinv : (MulAut.conj g)⁻¹ • z = g⁻¹ * z * g := by
    rw [← map_inv, MulAut.smul_def, MulAut.conj_apply, inv_inv]
  rw [hinv]
  constructor
  · intro hz
    have := mul_mem (mul_mem hg hz) (inv_mem hg)
    simpa [mul_assoc] using this
  · intro hz
    exact mul_mem (mul_mem (inv_mem hg) hz) hg

/-- **BG Lemma 15.1(c) / Theorem B(4), general element form**: for a maximal subgroup `M`, a Hall
`(κ ∪ σ)ᶜ`-subgroup `U ≤ M`, and a nonidentity `(κ(M) ∪ σ(M))ᶜ`-element `y ∈ M` whose
`M_σ`-centralizer is nontrivial, `C_G(y)` lies in the unique maximal subgroup `M`:
`ℳ(C_G(y)) = {M}`.

Type-independent core behind both `typeI_centralizer_le_and_unique` (Peterfalvi 8.12.b) and Theorem
B(5).  `⟨y⟩` is a `(κ ∪ σ)ᶜ`-subgroup of solvable `M`, so Hall D/C conjugate it (by `τ ∈ M`) into
`U`; Lemma 15.1(c) (`typeP_hall_small_subgroup_cyclic_tau2`) pins `ℳ(C_G(y^τ)) = {M}`; conjugating
back by `τ ∈ M = N_G(M)` recovers `ℳ(C_G(y)) = {M}`. -/
theorem uniqueMaximal_of_kappaSigmaCompl_element [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M U : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hUM : U ≤ M)
    (hU : Ch03.IsHallSubgroup ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ) (U.subgroupOf M))
    {y : G} (hyM : y ∈ M) (hy1 : y ≠ 1)
    (hyπ : IsPiElement (S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ y)
    (hyC : OddOrder.BG.Ch3.S10.Msigma M ⊓ Subgroup.centralizer ({y} : Set G) ≠ ⊥) :
    maximalSubgroupsContaining (Subgroup.centralizer ({y} : Set G)) = {M} := by
  classical
  have : Group.IsSolvable ↥M := hG.isSolvable_of_mem_maximalSubgroups hM
  set yM : ↥M := ⟨y, hyM⟩ with hyMdef
  -- (i) `⟨y⟩` is a `(κ ∪ σ)ᶜ`-subgroup of `M`.
  have hordeq : orderOf y = orderOf yM :=
    orderOf_injective M.subtype M.subtype_injective yM
  have hZπ : ∀ q ∈ (Nat.card ↥(Subgroup.zpowers yM)).primeFactors,
      q ∈ (S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ := by
    intro q hq
    rw [Nat.card_zpowers, ← hordeq] at hq
    obtain ⟨hqp, hqdvd, -⟩ := Nat.mem_primeFactors.mp hq
    exact hyπ q (Nat.mem_primeFactors.mpr ⟨hqp, hqdvd, (orderOf_pos y).ne'⟩)
  -- (ii) Hall D + C inside `M`: conjugate `y` into `U`.
  obtain ⟨W, hWhall, hZW⟩ := OddOrder.Isaacs.Ch03.hall_D (G := ↥M) hZπ
  obtain ⟨t, htmap⟩ := OddOrder.Isaacs.Ch03.hall_C (G := ↥M) hWhall hU
  set τ : G := (t : G) with hτdef
  have hτM : τ ∈ M := t.2
  set y' : G := τ * y * τ⁻¹ with hy'def
  have hy'U : y' ∈ U := by
    have hyW : yM ∈ W := hZW (Subgroup.mem_zpowers yM)
    have hmem : (MulAut.conj t) yM ∈ W.map (MulAut.conj t).toMonoidHom := ⟨yM, hyW, rfl⟩
    rw [htmap] at hmem
    have := Subgroup.mem_subgroupOf.mp hmem
    simpa [hy'def, hτdef, MulAut.conj_apply] using this
  have hy'1 : y' ≠ 1 := by
    simp only [hy'def, ne_eq, conj_eq_one_iff]; exact hy1
  -- (iii) transported witness `M_σ ⊓ C_G(⟨y'⟩) ≠ ⊥`.
  obtain ⟨z, hzmem, hz1⟩ := (Subgroup.bot_or_exists_ne_one _).resolve_left hyC
  obtain ⟨hzMσ, hzC⟩ := Subgroup.mem_inf.mp hzmem
  have hz'1 : τ * z * τ⁻¹ ≠ 1 := by simpa [conj_eq_one_iff] using hz1
  have hz'Mσ : τ * z * τ⁻¹ ∈ OddOrder.BG.Ch3.S10.Msigma M := by
    have hM_le_NMσ : M ≤ Subgroup.normalizer (OddOrder.BG.Ch3.S10.Msigma M : Set G) := by
      rw [OddOrder.BG.Ch3.S10.Msigma]; exact le_normalizer_opiCoreInG _ _
    have hMσ_norm : ((OddOrder.BG.Ch3.S10.Msigma M).subgroupOf M).Normal :=
      (Subgroup.normal_subgroupOf_iff_le_normalizer
        (OddOrder.BG.Ch3.S10.Msigma_le M)).mpr hM_le_NMσ
    have hzMmem : (⟨z, OddOrder.BG.Ch3.S10.Msigma_le M hzMσ⟩ : ↥M) ∈
        (OddOrder.BG.Ch3.S10.Msigma M).subgroupOf M := Subgroup.mem_subgroupOf.mpr hzMσ
    have := Subgroup.mem_subgroupOf.mp (hMσ_norm.conj_mem _ hzMmem t)
    simpa [hτdef] using this
  have hz'C : τ * z * τ⁻¹ ∈ Subgroup.centralizer ({y'} : Set G) := by
    have hcomm : Commute y z := Subgroup.mem_centralizer_iff.mp hzC y (Set.mem_singleton y)
    have hcomm' : Commute y' (τ * z * τ⁻¹) := by
      have := hcomm.map (MulAut.conj τ : G →* G)
      simpa [hy'def, MulAut.conj_apply] using this
    exact Subgroup.mem_centralizer_iff.mpr fun w hw => by
      rw [Set.mem_singleton_iff] at hw; subst hw; exact hcomm'.eq
  have hCne : OddOrder.BG.Ch3.S10.Msigma M ⊓
      Subgroup.centralizer ((Subgroup.zpowers y' : Subgroup G) : Set G) ≠ ⊥ := by
    rw [centralizer_zpowers_eq_singleton_B5]
    intro hbot
    have hmem : τ * z * τ⁻¹ ∈ OddOrder.BG.Ch3.S10.Msigma M ⊓
        Subgroup.centralizer ({y'} : Set G) := Subgroup.mem_inf.mpr ⟨hz'Mσ, hz'C⟩
    rw [hbot, Subgroup.mem_bot] at hmem
    exact hz'1 hmem
  -- (iv) Lemma 15.1(c) pins `ℳ(C_G(y')) = {M}`.
  have hB : maximalSubgroupsContaining (Subgroup.centralizer ({y'} : Set G)) = {M} := by
    have h := (typeP_hall_small_subgroup_cyclic_tau2 hG hM hUM hU
      (Subgroup.zpowers_le.mpr hy'U)
      (by simpa [Subgroup.zpowers_eq_bot] using hy'1) hCne).1
    rwa [centralizer_zpowers_eq_singleton_B5] at h
  -- (v) conjugate back by `τ ∈ M = N_G(M)`.
  have hCconj : Subgroup.centralizer ({y'} : Set G)
      = MulAut.conj τ • Subgroup.centralizer ({y} : Set G) := by
    rw [centralizer_pointwise_smul]
    congr 1
    simp [hy'def, MulAut.conj_apply, Set.image_singleton]
  rw [Set.eq_singleton_iff_unique_mem]
  refine ⟨⟨hM, ?_⟩, ?_⟩
  · -- `C_G(y) ≤ M`.
    have hCy'M : Subgroup.centralizer ({y'} : Set G) ≤ M := by
      have := hB ▸ (Set.mem_singleton M)
      exact (mem_maximalSubgroupsContaining.mp this).2
    intro c hc
    have hc' : τ * c * τ⁻¹ ∈ Subgroup.centralizer ({y'} : Set G) := by
      rw [hCconj]; exact ⟨c, hc, by simp [MulAut.conj_apply]⟩
    have := hCy'M hc'
    have hback := mul_mem (mul_mem (inv_mem hτM) this) hτM
    simpa [mul_assoc] using hback
  · -- uniqueness.
    intro L hL
    obtain ⟨hLco, hCL⟩ := hL
    have hLconj : MulAut.conj τ • L ∈
        maximalSubgroupsContaining (Subgroup.centralizer ({y'} : Set G)) := by
      refine ⟨isCoatom_pointwise_smul (MulAut.conj τ) hLco, ?_⟩
      rw [hCconj]
      exact Subgroup.pointwise_smul_le_pointwise_smul_iff.mpr hCL
    rw [hB, Set.mem_singleton_iff] at hLconj
    calc L = MulAut.conj τ⁻¹ • (MulAut.conj τ • L) := by
            rw [← mul_smul, ← map_mul, inv_mul_cancel, map_one, one_smul]
      _ = MulAut.conj τ⁻¹ • M := by rw [hLconj]
      _ = M := conj_smul_self_of_mem (inv_mem hτM)

/-- **BG Theorem B(5)** (mmd L4373; schematic proof Lemma 15.1(c)): the set `A(M) − M_σ` is a
TI-subset of `G` (with normalizer `M`).

`A(M) = \widehat{M_σ} ∩ U M_σ`; an element `a ∈ A(M) − M_σ` has a nontrivial `(κ ∪ σ)ᶜ`-part
`w = a_π` (nontrivial because `a ∉ M_σ`, using that `a ∈ U M_σ` is a `κ(M)′`-element), and
`ℳ(C_G(w)) = {M}` by `uniqueMaximal_of_kappaSigmaCompl_element` (Lemma 15.1(c)).  If `g` conjugates
`a ∈ A(M) − M_σ` to another such element, then `w` and `g w g⁻¹` are both `π`-parts of elements of
`A(M) − M_σ` (`piPart` is conjugation-equivariant), so `ℳ(C_G(g w g⁻¹)) = {M}` too; but
`C_G(g w g⁻¹) = g C_G(w) g⁻¹`, so `g⁻¹ M g` is a maximal subgroup over `C_G(w)`, forcing
`g⁻¹ M g = M`, i.e. `g ∈ N_G(M) = M`. -/
theorem theoremB_A_minus_Msigma_isTISubset [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M K U : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hKM : K ≤ M) (hUM : U ≤ M)
    (hK : Ch03.IsHallSubgroup (S14.kappa M) (K.subgroupOf M))
    (hU : Ch03.IsHallSubgroup ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ) (U.subgroupOf M)) :
    IsTISubset (ASet M U \ (OddOrder.BG.Ch3.S10.Msigma M : Set G)) M := by
  classical
  set π : Set ℕ := (S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ with hπdef
  have hMσM : OddOrder.BG.Ch3.S10.Msigma M ≤ M := OddOrder.BG.Ch3.S10.Msigma_le M
  -- normality of `U ⊔ M_σ` in `M` (Theorem A(3)), used only for the `κ′`-characterization.
  have hnorm : ((U ⊔ OddOrder.BG.Ch3.S10.Msigma M).subgroupOf M).Normal :=
    (Subgroup.normal_subgroupOf_iff_le_normalizer (sup_le hUM hMσM)).mpr
      (theoremA_ungated_conjuncts hG hM hKM hUM hK rfl hU).2.2.1
  -- **`uniqB`**: for `a ∈ A(M) − M_σ`, the `π`-part `w = a_π` has `ℳ(C_G(w)) = {M}`.
  have huniqB : ∀ a ∈ ASet M U \ (OddOrder.BG.Ch3.S10.Msigma M : Set G),
      maximalSubgroupsContaining (Subgroup.centralizer ({piPart π a} : Set G)) = {M} := by
    intro a ha
    obtain ⟨haA, haMσ⟩ := ha
    simp only [ASet, hatMsigma, Set.mem_inter_iff, Set.mem_ofPred_eq, SetLike.mem_coe] at haA
    obtain ⟨⟨haM, haC⟩, haUMσ⟩ := haA
    -- `w = a_π`, a power of `a`, is a `π`-element of `M`.
    obtain ⟨-, -, -, -, -, hwz, -⟩ := piPart_spec π a
    have hwM : piPart π a ∈ M := (Subgroup.zpowers_le.mpr haM) hwz
    have hwπ : IsPiElement π (piPart π a) := isPiElement_piPart π a
    -- `w ≠ 1`: else `a` is a `(κ ∪ σ)`-element and (being a `κ′`-element of `U M_σ`) a `σ`-element,
    -- hence `a ∈ M_σ`, contradicting `a ∉ M_σ`.
    have hw1 : piPart π a ≠ 1 := by
      intro hw0
      have haκσ : IsPiElement (S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M) a := by
        have := isPiElement_compl_of_piPart_eq_one hw0
        rwa [hπdef, compl_compl] at this
      have haκ' : IsPiElement (S14.kappa M)ᶜ a :=
        (S14.mem_U_sup_Msigma_iff_isPiElement_kappa_compl hG hM hUM hU hnorm haM).mp haUMσ
      have haσ : IsPiElement (OddOrder.BG.Ch3.S10.sigma M) a := by
        intro p hp
        exact (haκσ p hp).resolve_left (haκ' p hp)
      exact haMσ ((S14.mem_Msigma_iff_isPiElement_sigma hG hM haM).mpr haσ)
    -- `M_σ ⊓ C_G(w) ≠ ⊥`: the witness centralizing `a` also centralizes its power `w`.
    have hwC : OddOrder.BG.Ch3.S10.Msigma M ⊓ Subgroup.centralizer ({piPart π a} : Set G) ≠ ⊥ := by
      obtain ⟨z, hzmem, hz1⟩ := (Subgroup.bot_or_exists_ne_one _).resolve_left haC
      obtain ⟨hzMσ, hzC⟩ := Subgroup.mem_inf.mp hzmem
      obtain ⟨n, hn⟩ := Subgroup.mem_zpowers_iff.mp hwz
      have hzCw : z ∈ Subgroup.centralizer ({piPart π a} : Set G) := by
        refine Subgroup.mem_centralizer_iff.mpr fun w hw => ?_
        rw [Set.mem_singleton_iff] at hw; subst hw
        have hcomm : Commute a z := Subgroup.mem_centralizer_iff.mp hzC a (Set.mem_singleton a)
        rw [← hn]; exact (hcomm.zpow_left n).eq
      exact fun hbot => hz1 (Subgroup.mem_bot.mp
        (hbot ▸ Subgroup.mem_inf.mpr ⟨hzMσ, hzCw⟩))
    exact uniqueMaximal_of_kappaSigmaCompl_element hG hM hUM hU hwM hw1 hwπ hwC
  -- TI: `g` overlaps `A(M) − M_σ` with its conjugate `⇒ g ∈ M`.
  intro g hg
  obtain ⟨a, ha, hga⟩ := hg
  have hMa := huniqB a ha
  have hMga := huniqB (g * a * g⁻¹) hga
  rw [piPart_conj] at hMga
  set w : G := piPart π a with hwdef
  -- `C_G(g w g⁻¹) = conj g • C_G(w)`.
  have hCconj : Subgroup.centralizer ({g * w * g⁻¹} : Set G)
      = MulAut.conj g • Subgroup.centralizer ({w} : Set G) := by
    rw [centralizer_pointwise_smul]; congr 1
    simp [MulAut.conj_apply, Set.image_singleton]
  rw [hCconj] at hMga
  -- `conj g • C_G(w) ≤ M`, so `C_G(w) ≤ conj g⁻¹ • M`.
  have hCleM : MulAut.conj g • Subgroup.centralizer ({w} : Set G) ≤ M :=
    (mem_maximalSubgroupsContaining.mp (hMga ▸ Set.mem_singleton M)).2
  have hle : Subgroup.centralizer ({w} : Set G) ≤ MulAut.conj g⁻¹ • M := by
    have h := Subgroup.pointwise_smul_le_pointwise_smul_iff (a := MulAut.conj g⁻¹) |>.mpr hCleM
    rwa [← mul_smul, ← map_mul, inv_mul_cancel, map_one, one_smul] at h
  -- `conj g⁻¹ • M` is a maximal subgroup over `C_G(w)`; by `ℳ(C_G(w)) = {M}` it equals `M`.
  have hmemM : (MulAut.conj g⁻¹ • M) ∈
      maximalSubgroupsContaining (Subgroup.centralizer ({w} : Set G)) :=
    ⟨isCoatom_pointwise_smul _ hM, hle⟩
  rw [hMa, Set.mem_singleton_iff] at hmemM
  -- `g⁻¹ ∈ N_G(M) = M`, hence `g ∈ M`.
  have hg1norm : g⁻¹ ∈ Subgroup.normalizer (M : Set G) :=
    mem_normalizer_of_conj_smul_eq_self hmemM
  rw [normalizer_eq_self_of_maximal_B5 hG hM] at hg1norm
  simpa using inv_mem hg1norm

/-- **`κ(M) = π(M) ∖ σ(M)` from a trivial Hall `(κ ∪ σ)ᶜ`-complement** (the prime-set half of the
type-`P₁` ⟺ `U = ⊥` characterization).  If the Hall `(κ(M) ∪ σ(M))ᶜ`-subgroup of `M` is trivial,
then `(κ ∪ σ)ᶜ` contains no prime dividing `|M|` (the Hall index condition `hU.2` with index `|M|`),
so `π(M) ⊆ κ(M) ∪ σ(M)`; combined with `κ(M) ⊆ π(M) ∖ σ(M)` (every `κ`-prime is a non-`σ` divisor of
`|M|`) this gives `κ(M) = π(M) ∖ σ(M) = sigmaComplementPrimes M`.

Hoisted above `theoremC_paired_structure` (2026-06-21, lane F) so its conjunct 12 (`U = ⊥ → |K*|`
prime) can derive `IsTypeP1 M` from `U = ⊥`.  Together with
`isTypeP1_kappaSigma_compl_hall_subgroupOf_eq_bot` (below) it is the prime-set core of BG's
"`M ∈ M_{P₁} ⟺ U = ⊥`": modulo `IsTypeP` (`κ ≠ ∅`), `U = ⊥ ⟺ M` is type `P₁`. -/
theorem kappa_eq_sigmaComplementPrimes_of_hall_subgroupOf_eq_bot [Finite G] {M U : Subgroup G}
    (hU : Ch03.IsHallSubgroup ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ) (U.subgroupOf M))
    (hUbot : U.subgroupOf M = ⊥) :
    S14.kappa M = S14.sigmaComplementPrimes M := by
  refine Set.Subset.antisymm (fun p hpκ => ?_) (fun p hpσ' => ?_)
  · -- `κ ⊆ π ∖ σ`: a `κ`-prime is non-`σ` (`τ₁ ∪ τ₃`) and divides `|M|` (rank-one `P ≤ M`).
    obtain ⟨hpp, hτ, P, hPelem, hPM, _⟩ := hpκ
    have hpσ : p ∉ OddOrder.BG.Ch3.S10.sigma M := by
      rcases hτ with h | h
      · exact ((mem_tau1_iff M p).mp h).1
      · exact ((mem_tau3_iff M p).mp h).1
    have hpπ : p ∈ S14.piSet M := by
      refine Nat.mem_primeFactors.mpr ⟨hpp, ?_, Nat.card_pos.ne'⟩
      have hPcard : Nat.card ↥P = p := by rw [hPelem.2, pow_one]
      exact hPcard ▸ Subgroup.card_dvd_of_le hPM
    exact ⟨hpπ, hpσ⟩
  · -- `π ∖ σ ⊆ κ`: `U = ⊥` forces every `|M|`-prime into `κ ∪ σ`; non-`σ` ones land in `κ`.
    obtain ⟨hpπ, hpσ⟩ := hpσ'
    by_contra hpκ
    refine hU.2 p (by rw [hUbot, Subgroup.index_bot]; exact hpπ) ?_
    rintro (h | h)
    · exact hpκ h
    · exact hpσ h

/-- **`M` is type `P₂` from a nontrivial Hall `(κ ∪ σ)ᶜ`-complement** (the type-`P₂` side of the
`M ∈ M_{P₁} ⟺ U = ⊥` characterization): a type-`P` maximal subgroup whose Hall `(κ(M) ∪ σ(M))ᶜ`-
subgroup `U` is nontrivial has `κ(M) ≠ σ'(M)`, i.e. is type `P₂`.  Dual to
`isTypeP1_kappaSigma_compl_hall_subgroupOf_eq_bot`: a prime `p ∣ |U|` lies in `(κ ∪ σ)ᶜ ∩ π(M)`,
so if `κ = σ' = π ∖ σ` then `p ∈ κ` (it is a non-`σ` divisor of `|M|`), contradicting `p ∉ κ`. -/
theorem isTypeP2_of_hall_subgroupOf_ne_bot [Finite G] {M U : Subgroup G}
    (hP : S14.IsTypeP M)
    (hU : Ch03.IsHallSubgroup ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ) (U.subgroupOf M))
    (hUne : U.subgroupOf M ≠ ⊥) :
    S14.IsTypeP2 M := by
  refine ⟨hP, fun hκσ => ?_⟩
  obtain ⟨p, hpp, hpdvd⟩ := Nat.exists_prime_and_dvd
    (by rwa [ne_eq, ← Subgroup.card_eq_one] at hUne)
  have hpcompl : p ∈ (S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ :=
    hU.1 p (Nat.mem_primeFactors.mpr ⟨hpp, hpdvd, Nat.card_pos.ne'⟩)
  have hpπ : p ∈ S14.piSet M :=
    Nat.mem_primeFactors.mpr ⟨hpp,
      hpdvd.trans (Subgroup.card_subgroup_dvd_card (U.subgroupOf M)), Nat.card_pos.ne'⟩
  have hpσ : p ∉ OddOrder.BG.Ch3.S10.sigma M := fun h => hpcompl (Set.mem_union_right _ h)
  have hpκ : p ∈ S14.kappa M := by rw [hκσ]; exact ⟨hpπ, hpσ⟩
  exact hpcompl (Set.mem_union_left _ hpκ)

end OddOrder.BG.Ch4.S16
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/BG/Ch4_FamilyOfMaximal/S16_MainResults/TheoremAB.lean

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me