Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

BG §15: OpO_pOp​-core centralizer, opening layer

Definition
OddOrder_BG_Ch4_FamilyOfMaximal_S15_MF_OpicoreCentralizerBasic

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

feit-thompsonfinite-groupsgroup-theoryishida-odd-order

The opening layer of the OpO_pOp​-core centralizer part of §15 of Bender–Glauberman, Local Analysis for the Odd Order Theorem. It consists of lemmas only, mostly about the failure of F(M)F(M)F(M) to be TI: then MF=MσM_F=M_\sigmaMF​=Mσ​, M′≤F(M)M'\le F(M)M′≤F(M) and E3=1E_3=1E3​=1; conversely F(M)F(M)F(M) is TI when MF≠MσM_F\ne M_\sigmaMF​=Mσ​ or for type P2\mathcal{P}_2P2​.

The bundle consists of lemmas only.

The bundle also proves 6 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/S15_MF/OpicoreCentralizerBasic.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_S15_MF_PisetBetaDisjoint

namespace OddOrder.BG.Ch4.S15

open OddOrder.GroupTheory

open OddOrder.Isaacs

open OddOrder.BG.Ch3.S12

open OddOrder.BG.Ch4.S14

open scoped Pointwise

open scoped IsMulCommutative

open scoped commutatorElement

variable {G : Type*} [Group G]

/-- **BG Theorem 15.7(a), rank-theoretic core** (mmd L4192-4198): if `F(M)` is not a TI-subgroup
of `G`, then no prime divides `M_F` and lies in `β(M)`.

The `≥ 3` side is `three_le_pRank_mf_of_mem_beta` (any `r ∈ π(M_F) ∩ β(M)` has `r_r(M_F) ≥ 3`); the
proof below establishes the complementary `< 3` bound `pRank (M_F) r < 3`, the genuinely deep §15
content.  **Both sides are now fully proved — this theorem is sorry-free and axiom-clean.**

**Building blocks (this file):** the setup
`exists_notMem_inf_conj_fitting_ne_bot_of_not_fittingIsTI` (step 1: `g ∉ M`,
`X = F(M) ⊓ F(M)^g ≠ ⊥`)
and `rank_lt_three_of_le_two_maximals` (step 7 core: a subgroup in two distinct maximals has rank
`< 3`).  The assembly, with the upstream lemmas it uses:

* **(step 3, `p ∈ σ(M)`)** pick `p ∈ π(X)`, `X₁ ≤ X` of order `p`
  (`le_opiCoreInG_singleton_of_isPGroup_of_le_nilpotent`: `X₁ ≤ O_p(F(M))`).  If `p ∉ σ(M)` then
  `O_p(F(M)) ≤ O_{σ'}(F(M))` is cyclic (`fitting_decomposition`), so `X₁` is the unique order-`p`
  subgroup, hence characteristic and normal in both `M` and `M^g`;
  `normalizer_eq_of_normal_of_mem_maximal`
  (S08, currently `private`) forces `M^g = M`, contradicting `g ∉ M`. ⟹ `p ∈ σ(M)`. *(fiddly
  sub-step:
  cyclic group ⟹ unique/characteristic order-`p` subgroup.)*
* **(step 5, `p ∉ β(M)`)** `X₁ ≤ O_p(M) ≤ M_σ` (`opiCoreInG_singleton_le_Msigma_of_mem_sigma`) and
  `X₁ ≤ F(M)^g ≤ M^g`, so `X₁ ≤ M_σ ⊓ M^g`; Lemma 12.17 (`Msigma_inf_conj_isBetaCompl`) ⟹
  `p ∉ β(M)`,
  hence `r ≠ p` for the `β`-prime `r` (so the `r = p` case is vacuous).
* **(step 6, `C_G(X₁) ⊄ M`)** `fusion_control_of_mem_sigma` part (e) with `p ∈ σ(M)`, `X₁ ≤ M`,
  `conj g⁻¹ • X₁ ≤ M`.
* **(step 7, `rank C_{M_F}(X₁) < 3`)** `C_G(X₁) < ⊤` (else `X₁ ≤ Z(G) = ⊥`, simple), so a coatom
  `N ⊇ C_G(X₁)` exists with `N ≠ M`; `C_{M_F}(X₁) ≤ M ⊓ N` ⟹ `rank_lt_three_of_le_two_maximals`.
* **(step 8, bridge)** `O_r(M_F) ≤ C_{M_F}(X₁)`
  (`commute_of_coprime_orderOf_of_isNilpotent`, `r ≠ p`,
  both in nilpotent `F(M)`), so `r_r(M_F) = r_r(O_r(M_F)) ≤ rank C_{M_F}(X₁) < 3`.

Combined with `mf_eq_msigma_of_piSet_inf_beta_disjoint` this yields the `M_F = M_σ` conclusion of
Theorem 15.7(a), i.e. the `FittingIsTI` clause of Theorem A(8). -/
theorem piSet_mf_inf_beta_disjoint_of_not_fittingIsTI [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hnotTI : ¬ FittingIsTI M) :
    ∀ q : ℕ, q ∈ S14.piSet (MF M) → q ∉ OddOrder.BG.Ch3.S10.beta M := by
  intro r hrπ hrβ
  -- The `≥ 3` side (proved): `r ∈ π(M_F) ∩ β(M) ⟹ r_r(M_F) ≥ 3`.
  have h3 : 3 ≤ pRank ↥(MF M) r := three_le_pRank_mf_of_mem_beta hG hM hrπ hrβ
  have hrp : r.Prime := Nat.prime_of_mem_primeFactors hrπ
  have : Fact r.Prime := ⟨hrp⟩
  refine absurd h3 (not_le.mpr ?_)
  -- Setup: `g ∉ M`, `X = F(M) ⊓ F(M)^g ≠ ⊥`.
  obtain ⟨g, hgM, hXne⟩ := exists_notMem_inf_conj_fitting_ne_bot_of_not_fittingIsTI hnotTI
  -- A prime `p ∈ π(X)`, and `p ∈ σ(M)` (step 3).
  have hXcard : Nat.card ↥(fittingInAmbient M ⊓ MulAut.conj g • fittingInAmbient M) ≠ 1 :=
    fun h => hXne (Subgroup.card_eq_one.mp h)
  obtain ⟨p, hp, hpdvd⟩ := Nat.exists_prime_and_dvd hXcard
  have : Fact p.Prime := ⟨hp⟩
  have hpσ : p ∈ OddOrder.BG.Ch3.S10.sigma M :=
    mem_sigma_of_prime_dvd_card_inf_conj_fitting hG hM hgM hp hpdvd
  -- An order-`p` subgroup `X₁ ≤ X`.
  obtain ⟨x, hxord⟩ := exists_prime_orderOf_dvd_card'
    (G := ↥(fittingInAmbient M ⊓ MulAut.conj g • fittingInAmbient M)) p hpdvd
  set X₁ : Subgroup G :=
    (Subgroup.zpowers x).map (fittingInAmbient M ⊓ MulAut.conj g • fittingInAmbient M).subtype
    with hX₁def
  have hX₁leX : X₁ ≤ fittingInAmbient M ⊓ MulAut.conj g • fittingInAmbient M :=
    hX₁def ▸ Subgroup.map_subtype_le _
  have hX₁card : Nat.card ↥X₁ = p := by
    rw [hX₁def, Subgroup.card_map_of_injective
      (fittingInAmbient M ⊓ MulAut.conj g • fittingInAmbient M).subtype_injective,
      Nat.card_zpowers, hxord]
  have hX₁ne : X₁ ≠ ⊥ := fun h => hp.one_lt.ne' (by rw [← hX₁card, h, Subgroup.card_bot])
  have hX₁pg : IsPGroup p ↥X₁ := IsPGroup.of_card (n := 1) (by rw [hX₁card, pow_one])
  have hX₁F : X₁ ≤ fittingInAmbient M := hX₁leX.trans inf_le_left
  have hX₁cF : X₁ ≤ MulAut.conj g • fittingInAmbient M := hX₁leX.trans inf_le_right
  have hX₁M : X₁ ≤ M := hX₁F.trans (OddOrder.BG.Ch2.S08.fittingInG_le M)
  -- Step 5: `p ∉ β(M)` (via `X₁ ≤ M_σ ⊓ M^g` and Lemma 12.17), hence `r ≠ p`.
  have hpσ_sub : ({p} : Set ℕ) ⊆ OddOrder.BG.Ch3.S10.sigma M := by
    intro q hq; rw [Set.mem_singleton_iff] at hq; rw [hq]; exact hpσ
  have hX₁Mσ : X₁ ≤ OddOrder.BG.Ch3.S10.Msigma M := by
    have h1 : X₁ ≤ opiCoreInG ({p} : Set ℕ) (fittingInAmbient M) :=
      OddOrder.BG.Ch2.S08.le_opiCoreInG_singleton_of_isPGroup_of_le_nilpotent
        (OddOrder.BG.Ch2.S08.fittingInG_isNilpotent M) hX₁F hX₁pg
    have h2 : opiCoreInG ({p} : Set ℕ) (fittingInAmbient M)
        ≤ opiCoreInG (OddOrder.BG.Ch3.S10.sigma M) (fittingInAmbient M) :=
      Subgroup.map_mono (OddOrder.Isaacs.Ch03.oPiCore_mono hpσ_sub _)
    have h3' : opiCoreInG (OddOrder.BG.Ch3.S10.sigma M) (fittingInAmbient M)
        = fittingInAmbient (OddOrder.BG.Ch3.S10.Msigma M) :=
      opiCoreInG_sigma_fittingInAmbient_eq_fittingInAmbient_Msigma
    exact (h1.trans h2).trans (h3' ▸ OddOrder.BG.Ch2.S08.fittingInG_le _)
  have hX₁cM : X₁ ≤ MulAut.conj g • M :=
    hX₁cF.trans (Subgroup.pointwise_smul_le_pointwise_smul_iff.mpr
      (OddOrder.BG.Ch2.S08.fittingInG_le M))
  have hpβ : p ∉ OddOrder.BG.Ch3.S10.beta M :=
    OddOrder.BG.Ch3.S12.Msigma_inf_conj_isBetaCompl hG hM hgM p
      (Nat.mem_primeFactors.mpr ⟨hp,
        hX₁card ▸ Subgroup.card_dvd_of_le (le_inf hX₁Mσ hX₁cM),
        Nat.card_pos.ne'⟩)
  have hrnep : r ≠ p := fun h => hpβ (h ▸ hrβ)
  -- Step 6: `C_G(X₁) ⊄ M` (Theorem 10.1(e)).
  have hconj_g_inv : MulAut.conj g⁻¹ • X₁ ≤ M := by
    have hle : MulAut.conj g⁻¹ • X₁ ≤ MulAut.conj g⁻¹ • (MulAut.conj g • fittingInAmbient M) :=
      Subgroup.pointwise_smul_le_pointwise_smul_iff.mpr hX₁cF
    rw [← mul_smul, ← map_mul, inv_mul_cancel, map_one, one_smul] at hle
    exact hle.trans (OddOrder.BG.Ch2.S08.fittingInG_le M)
  have hCGnotM : ¬ Subgroup.centralizer (X₁ : Set G) ≤ M := by
    intro hCG
    have he := (OddOrder.BG.Ch3.S10.fusion_control_of_mem_sigma hG hM hpσ hX₁ne hX₁pg).2.2.2.2
    exact hgM (by simpa using inv_mem (he hX₁M hCG g⁻¹ hconj_g_inv))
  -- Step 7: `rank (C_{M_F}(X₁)) < 3` (`C_{M_F}(X₁)` lies in `M` and in a coatom `N ≠ M`).
  obtain ⟨x₀, hx₀ne⟩ := Subgroup.ne_bot_iff_exists_ne_one.mp hX₁ne
  have hCGlt : Subgroup.centralizer (X₁ : Set G) < ⊤ :=
    lt_of_le_of_lt
      (Subgroup.centralizer_le (Set.singleton_subset_iff.mpr x₀.2))
      (OddOrder.BG.Ch2.S09.centralizer_singleton_lt_top hG
        (fun h => hx₀ne (Subtype.ext h)))
  obtain ⟨N, hNco, hCGN⟩ :=
    (eq_top_or_exists_le_coatom (Subgroup.centralizer (X₁ : Set G))).resolve_left hCGlt.ne
  have hNmax : N ∈ maximalSubgroups G := mem_maximalSubgroups.mpr hNco
  have hNneM : M ≠ N := fun h => hCGnotM (h ▸ hCGN)
  have hrank3 : rank ↥(MF M ⊓ Subgroup.centralizer (X₁ : Set G)) < 3 :=
    rank_lt_three_of_le_two_maximals hG hM hNmax hNneM
      (inf_le_left.trans (maxNilpotentNormalHall_le M)) (inf_le_right.trans hCGN)
  -- Step 8: `r`-elements of `M_F` centralize `X₁` (coprime, `F(M)` nilpotent), so the `r`-Sylow of
  -- `M_F` lies in `C_{M_F}(X₁)`; hence `r ∤ [M_F : C_{M_F}(X₁)]` and
  -- `r_r(M_F) = r_r(C_{M_F}(X₁)) < 3`.
  have : Group.IsNilpotent ↥(fittingInAmbient M) := OddOrder.BG.Ch2.S08.fittingInG_isNilpotent M
  have hcentr : ∀ A : Subgroup G, A ≤ MF M → IsPGroup r ↥A →
      A ≤ Subgroup.centralizer (X₁ : Set G) := by
    intro A hAMF hAr a ha
    rw [Subgroup.mem_centralizer_iff]
    intro y hy
    have haF : a ∈ fittingInAmbient M := (hAMF.trans (maxNilpotentNormalHall_le_fittingInG M)) ha
    have hyF : y ∈ fittingInAmbient M := hX₁F hy
    obtain ⟨i, hi⟩ := (IsPGroup.iff_orderOf.mp hAr) ⟨a, ha⟩
    obtain ⟨j, hj⟩ := (IsPGroup.iff_orderOf.mp hX₁pg) ⟨y, hy⟩
    have e1 : orderOf (⟨a, haF⟩ : ↥(fittingInAmbient M)) = r ^ i :=
      (orderOf_injective (fittingInAmbient M).subtype
        (fittingInAmbient M).subtype_injective ⟨a, haF⟩).symm.trans
        ((orderOf_injective A.subtype A.subtype_injective ⟨a, ha⟩).trans hi)
    have e2 : orderOf (⟨y, hyF⟩ : ↥(fittingInAmbient M)) = p ^ j :=
      (orderOf_injective (fittingInAmbient M).subtype
        (fittingInAmbient M).subtype_injective ⟨y, hyF⟩).symm.trans
        ((orderOf_injective X₁.subtype X₁.subtype_injective ⟨y, hy⟩).trans hj)
    have hcop : Nat.Coprime (orderOf (⟨a, haF⟩ : ↥(fittingInAmbient M)))
        (orderOf (⟨y, hyF⟩ : ↥(fittingInAmbient M))) := by
      rw [e1, e2]; exact Nat.Coprime.pow _ _ ((Nat.coprime_primes hrp hp).mpr hrnep)
    have hcomm := OddOrder.GroupTheory.commute_of_coprime_orderOf_of_isNilpotent
      (x := (⟨a, haF⟩ : ↥(fittingInAmbient M))) (y := ⟨y, hyF⟩) hcop
    have := congrArg (Subtype.val) hcomm.eq
    simpa using this.symm
  -- The `r`-Sylow `P'` of `M_F` lies in `C_{M_F}(X₁) = M_F ⊓ C_G(X₁)`.
  obtain ⟨P⟩ : Nonempty (Sylow r ↥(MF M)) := inferInstance
  set P' : Subgroup G := (P : Subgroup ↥(MF M)).map (MF M).subtype with hP'def
  have hP'MF : P' ≤ MF M := hP'def ▸ Subgroup.map_subtype_le _
  have hP'r : IsPGroup r ↥P' := by
    obtain ⟨n, hn⟩ := IsPGroup.iff_card.mp P.2
    exact IsPGroup.of_card (by
      rw [hP'def, Subgroup.card_map_of_injective (MF M).subtype_injective, hn])
  have hP'C : P' ≤ MF M ⊓ Subgroup.centralizer (X₁ : Set G) :=
    le_inf hP'MF (hcentr P' hP'MF hP'r)
  have hidx : ¬ r ∣ ((MF M ⊓ Subgroup.centralizer (X₁ : Set G)).subgroupOf (MF M)).index := by
    intro hdvd
    have hP'sub : P'.subgroupOf (MF M) ≤ (MF M ⊓ Subgroup.centralizer (X₁ : Set G)).subgroupOf (MF
        M) :=
      Subgroup.subgroupOf_mono (MF M) hP'C
    have hdvd2 : r ∣ (P'.subgroupOf (MF M)).index :=
      hdvd.trans (Subgroup.index_dvd_of_le hP'sub)
    have hPeq : P'.subgroupOf (MF M) = (P : Subgroup ↥(MF M)) :=
      hP'def ▸ Subgroup.comap_map_eq_self_of_injective (MF M).subtype_injective _
    rw [hPeq] at hdvd2
    exact P.not_dvd_index hdvd2
  calc pRank ↥(MF M) r
      = pRank ↥(MF M ⊓ Subgroup.centralizer (X₁ : Set G)) r :=
        (OddOrder.GroupTheory.pRank_eq_of_le_of_not_dvd_index inf_le_left hidx).symm
    _ ≤ rank ↥(MF M ⊓ Subgroup.centralizer (X₁ : Set G)) := pRank_le_rank r
    _ < 3 := hrank3

/-- **`M_F = M_σ` from `¬FittingIsTI`** (the `M_F = M_σ` conclusion of BG Theorem 15.7(a)):
combine the rank-theoretic core `piSet_mf_inf_beta_disjoint_of_not_fittingIsTI` with the
Theorem 15.2(b) endgame `mf_eq_msigma_of_piSet_inf_beta_disjoint`. -/
theorem mf_eq_msigma_of_not_fittingIsTI [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hnotTI : ¬ FittingIsTI M) : MF M = OddOrder.BG.Ch3.S10.Msigma M :=
  mf_eq_msigma_of_piSet_inf_beta_disjoint hG hM
    (piSet_mf_inf_beta_disjoint_of_not_fittingIsTI hG hM hnotTI)

/-- **BG Theorem 15.7(c), `M' ≤ F(M)` for `¬FittingIsTI M`** (Coq `nonTI_Fitting_structure` `sM'F`):
the derived subgroup of a maximal subgroup whose Fitting subgroup is *not* `TI` is nilpotent, hence
`≤ F(M)`.

`M_F = M_σ` (`mf_eq_msigma_of_not_fittingIsTI`), and `M_β ≤ M_σ = M_F` is a `β(M)`-group
(`Mbeta_isPiGroup`; `β ⊆ α ⊆ σ` via `alpha_subset_sigma`), so `π(M_β) ⊆ π(M_F) ∩ β(M) = ∅`
(`piSet_mf_inf_beta_disjoint_of_not_fittingIsTI`), forcing `M_β = 1`. Then `M'/M_β ≅ M'` is
nilpotent
(`derivedQuotientMbeta_isNilpotent`), and the nilpotent normal `M'` lies in the Fitting subgroup
(`le_fittingInAmbient_of_subgroupOf_normal_of_isNilpotent`).  Sole upstream gate of the `E₃ = 1`
chain feeding Corollary 15.9's cyclic Frobenius complement (with
`tau3_eq_empty_of_derivedInG_le_fittingInAmbient` + `E3_eq_bot_of_tau3_eq_empty`). -/
theorem derivedInG_le_fittingInAmbient_of_not_fittingIsTI [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hnotTI : ¬ FittingIsTI M) : derivedInG M ≤ fittingInAmbient M := by
  classical
  have : Group.IsSolvable ↥M := hG.isSolvable_of_mem_maximalSubgroups hM
  have hMFMσ : MF M = OddOrder.BG.Ch3.S10.Msigma M := mf_eq_msigma_of_not_fittingIsTI hG hM hnotTI
  -- `M_β ≤ M_σ`: `M_β` is a normal `σ(M)`-subgroup (`β ⊆ α ⊆ σ`).
  have hMβσ : Subgroup.IsPiSubgroup (OddOrder.BG.Ch3.S10.sigma M)
      (OddOrder.BG.Ch3.S10.Mbeta M) := fun p hp =>
    OddOrder.BG.Ch3.S10.alpha_subset_sigma hG hM (OddOrder.BG.Ch3.S10.Mbeta_isPiGroup M p hp).1
  have hMβnorm : ((OddOrder.BG.Ch3.S10.Mbeta M).subgroupOf M).Normal := by
    rw [OddOrder.BG.Ch3.S10.Mbeta, OddOrder.BG.Ch3.S10.opiCoreInG_subgroupOf]; infer_instance
  have hMβMσ : OddOrder.BG.Ch3.S10.Mbeta M ≤ OddOrder.BG.Ch3.S10.Msigma M :=
    le_opiCoreInG_of_normal_of_isPiSubgroup (OddOrder.BG.Ch3.S10.Mbeta_le M) hMβnorm hMβσ
  -- `M_β = ⊥`: `π(M_β) ⊆ π(M_F) ∩ β(M) = ∅`.
  have hMβbot : OddOrder.BG.Ch3.S10.Mbeta M = ⊥ := by
    rw [← Subgroup.card_eq_one]
    by_contra hne
    obtain ⟨p, hp, hpdvd⟩ := Nat.exists_prime_and_dvd hne
    have hpMβ : p ∈ (Nat.card ↥(OddOrder.BG.Ch3.S10.Mbeta M)).primeFactors :=
      Nat.mem_primeFactors.mpr ⟨hp, hpdvd, Nat.card_pos.ne'⟩
    have hpβ : p ∈ OddOrder.BG.Ch3.S10.beta M := OddOrder.BG.Ch3.S10.Mbeta_isPiGroup M p hpMβ
    have hpMF : p ∈ S14.piSet (MF M) := by
      change p ∈ (Nat.card ↥(MF M)).primeFactors
      rw [hMFMσ]
      exact Nat.primeFactors_mono (Subgroup.card_dvd_of_le hMβMσ) Nat.card_pos.ne' hpMβ
    exact piSet_mf_inf_beta_disjoint_of_not_fittingIsTI hG hM hnotTI p hpMF hpβ
  -- `M'` nilpotent: `M'/M_β ≅ M'` (`M_β = ⊥`), and `M'/M_β` is nilpotent (Corollary 10.7).
  have heq : (OddOrder.BG.Ch3.S10.Mbeta M).subgroupOf (derivedInG M) = ⊥ := by
    rw [hMβbot, Subgroup.bot_subgroupOf]
  have hnorm2 : ((OddOrder.BG.Ch3.S10.Mbeta M).subgroupOf (derivedInG M)).Normal := by
    rw [heq]; infer_instance
  have := OddOrder.BG.Ch3.S10.derivedQuotientMbeta_isNilpotent hG hM
  have hM'nil : Group.IsNilpotent ↥(derivedInG M) :=
    Group.nilpotent_of_mulEquiv ((QuotientGroup.quotientMulEquivOfEq heq).trans
        QuotientGroup.quotientBot)
  -- `M'` normal in `M` + nilpotent ⟹ `M' ≤ F(M)`.
  have hid : (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 [hid]; infer_instance
  exact le_fittingInAmbient_of_subgroupOf_normal_of_isNilpotent (Subgroup.map_subtype_le _) hM'norm

/-- **BG Theorem 15.7(d), `E₃ = 1`** (Coq `nonTI_Fitting_structure` `E3_1`):
the `τ₃(M)`-Hall factor of
any `σ(M)'`-complement `E`-setup is trivial when `F(M)` is not `TI`.  Composes the three parts:
`M' ≤ F(M)` (`derivedInG_le_fittingInAmbient_of_not_fittingIsTI`, part (c)) → `τ₃(M) = ∅`
(`tau3_eq_empty_of_derivedInG_le_fittingInAmbient`) → `E₃ = ⊥` (`E3_eq_bot_of_tau3_eq_empty`).  With
`τ₂(M) = ∅` (Theorem 15.8) forcing `E₂ = 1` too, `E = E₁` is cyclic (`E1_isCyclic`) — the cyclic
Frobenius complement of Corollary 15.9. -/
theorem E3_eq_bot_of_not_fittingIsTI [Finite G] (hG : OddOrder.BG.IsMinimalSimpleOdd G)
    {M E E₁ E₂ E₃ : Subgroup G} (hM : M ∈ maximalSubgroups G) (hnotTI : ¬ FittingIsTI M)
    (hsetup : OddOrder.BG.Ch3.S12.SubgroupESetup M E E₁ E₂ E₃) : E₃ = ⊥ :=
  E3_eq_bot_of_tau3_eq_empty hsetup
    (tau3_eq_empty_of_derivedInG_le_fittingInAmbient hG hM
      (derivedInG_le_fittingInAmbient_of_not_fittingIsTI hG hM hnotTI))

/-- **BG Theorem A(8), the `FittingIsTI` clause** (mmd L4274, schematic proof: Theorem 15.7(a)(b)):
if `M_F ≠ M_σ`, then `F(M)` is a TI-subgroup of `G`.  This is the contrapositive of the
`M_F = M_σ` conclusion of Theorem 15.7(a) (`mf_eq_msigma_of_not_fittingIsTI`): if `F(M)` failed to
be TI, then `M_F` would equal `M_σ`.  Discharges the last (and deepest) conjunct of Theorem A(8),
modulo the single rank-theoretic residual `piSet_mf_inf_beta_disjoint_of_not_fittingIsTI`. -/
theorem fitting_isTI_of_mf_ne_msigma [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hne : MF M ≠ OddOrder.BG.Ch3.S10.Msigma M) : FittingIsTI M := by
  by_contra h
  exact hne (mf_eq_msigma_of_not_fittingIsTI hG hM h)

/-- **`IsTypeP2 M → FittingIsTI M`** — the type-classification conjunct (a) of BG Theorem 15.7
(mmd L4244): a type-`P₂` maximal subgroup has a TI Fitting subgroup.  Equivalently, every maximal
`M` with `¬FittingIsTI M` lies in `M_F ∪ M_{P₁}` (is type `F` or `P₁`, never `P₂`) — conjunct (a)
of `fitting_not_ti_cases`, separated out here because it is all that BG's §16 (Theorem C(10) /
Proposition 16.1) needs, and it is provable from the landed §15 pieces alone.

Proof (BG L4244): suppose `¬FittingIsTI M`.  Then `M_F = M_σ`
(`mf_eq_msigma_of_not_fittingIsTI`) and `π(M_F) ∩ β(M) = ∅`
(`piSet_mf_inf_beta_disjoint_of_not_fittingIsTI`).  But a type-`P₂` maximal subgroup has
`σ(M) = β(M)` (Proposition 14.2(g) = the type-`P₂` clause of `typeP_structure`), and `M_σ ≠ 1`
provides a prime `q ∈ π(M_σ) = π(M_F)` with `q ∈ σ(M) = β(M)`, contradicting the disjointness. -/
theorem fittingIsTI_of_isTypeP2 [Finite G] (hG : OddOrder.BG.IsMinimalSimpleOdd G)
    {M : Subgroup G} (hM : M ∈ maximalSubgroups G) (hP2 : S14.IsTypeP2 M) :
    FittingIsTI M := by
  classical
  by_contra hnotTI
  have : Group.IsSolvable ↥M := hG.isSolvable_of_mem_maximalSubgroups hM
  -- Hall `κ(M)`- and `(κ ∪ σ)ᶜ`-subgroups `K`, `U` of `M`, as Proposition 14.2 needs.
  obtain ⟨K', hK'⟩ := Ch03.hall_E_exists (G := ↥M) (S14.kappa M)
  set K : Subgroup G := K'.map M.subtype with hKdef
  have hKM : K ≤ M := hKdef ▸ Subgroup.map_subtype_le K'
  have hKeq : K.subgroupOf M = K' :=
    hKdef ▸ Subgroup.comap_map_eq_self_of_injective M.subtype_injective K'
  have hK : Ch03.IsHallSubgroup (S14.kappa M) (K.subgroupOf M) := hKeq ▸ hK'
  obtain ⟨U', hU'⟩ := Ch03.hall_E_exists (G := ↥M) ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ)
  have hUeq : (U'.map M.subtype).subgroupOf M = U' :=
    Subgroup.comap_map_eq_self_of_injective M.subtype_injective U'
  have hU : Ch03.IsHallSubgroup ((S14.kappa M ∪ OddOrder.BG.Ch3.S10.sigma M)ᶜ)
      ((U'.map M.subtype).subgroupOf M) := by rw [hUeq]; exact hU'
  -- Proposition 14.2(g): a type-`P₂` maximal subgroup has `σ(M) = β(M)`.
  have hσβ : OddOrder.BG.Ch3.S10.sigma M = OddOrder.BG.Ch3.S10.beta M :=
    ((S14.typeP_structure hG hM hP2.1 hKM hK rfl hU).2.2.2.2.1 hP2).1
  -- `¬FittingIsTI`: `M_F = M_σ` and `π(M_F) ∩ β(M) = ∅`.
  have hMFeq : MF M = OddOrder.BG.Ch3.S10.Msigma M := mf_eq_msigma_of_not_fittingIsTI hG hM hnotTI
  have hdisj := piSet_mf_inf_beta_disjoint_of_not_fittingIsTI hG hM hnotTI
  -- `M_σ ≠ 1` gives a prime `q ∈ π(M_σ) = π(M_F)`, in `σ(M) = β(M)`: the contradiction.
  have hMσne1 : Nat.card ↥(OddOrder.BG.Ch3.S10.Msigma M) ≠ 1 := by
    rw [ne_eq, Subgroup.card_eq_one]; exact OddOrder.BG.Ch3.S10.Msigma_ne_bot hG hM
  obtain ⟨q, hqp, hqdvd⟩ := Nat.exists_prime_and_dvd hMσne1
  have hqπMσ : q ∈ S14.piSet (OddOrder.BG.Ch3.S10.Msigma M) :=
    Nat.mem_primeFactors.mpr ⟨hqp, hqdvd, Nat.card_pos.ne'⟩
  have hqσ : q ∈ OddOrder.BG.Ch3.S10.sigma M :=
    (OddOrder.BG.Ch3.S10.Msigma_isHall hG hM).1 q hqπMσ
  have hqMF : q ∈ S14.piSet (MF M) := by rw [hMFeq]; exact hqπMσ
  exact hdisj q hqMF (hσβ ▸ hqσ)

end OddOrder.BG.Ch4.S15
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/S15_MF/OpicoreCentralizerBasic.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