Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

BG §15: notation and the Aut-abelian core

Definition
OddOrder_BG_Ch4_FamilyOfMaximal_S15_MF_AutAbelianCore

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

feit-thompsonfinite-groupsgroup-theoryishida-odd-order

The §15 notation layer of Bender–Glauberman, Local Analysis for the Odd Order Theorem and the reusable Aut-abelian core independent of §14. It introduces MFM_FMF​ (the maximal nilpotent normal Hall subgroup of MMM), the Fitting subgroup of MMM in the ambient group, its nonidentity part, and the hypothesis that F(M)F(M)F(M) is a TI-subgroup of GGG, together with the core fact that H′≤CH(N)H'\le C_H(N)H′≤CH​(N) when N⊴HN\trianglelefteq HN⊴H is cyclic.

Main declarations: MF (BG M_F: the maximal nilpotent normal Hall subgroup of M); fittingInAmbient (The Fitting subgroup of M, viewed in the ambient group as in BG §8/§15); fittingSharp (The nonidentity part of the ambient Fitting subgroup of M); FittingIsTI (The BG §15 hypothesis that F(M) is a TI-subgroup of G); and 3 more.

The bundle also proves 16 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/AutAbelianCore.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_S14_TypePCounting_Basics
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S14_TypePCounting_LocalStructure
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S14_TypePCounting_SigmaLengthOne
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S14_TypePCounting_TypePDuality
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S14_TypePCounting_GlobalCounting
import Definitions.Def_OddOrder_BG_Ch1_Preliminary_S03h_Thm38
import Definitions.Def_OddOrder_BG_Ch1_Preliminary_S03g_Thm310ElemAbelian
import Definitions.Def_OddOrder_BG_Ch1_Preliminary_S05_NarrowCharacterization
import Definitions.Def_OddOrder_Mathlib_Subgroup

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 `M_F`**: the maximal nilpotent normal Hall subgroup of `M`. -/
noncomputable abbrev MF (M : Subgroup G) : Subgroup G :=
  maxNilpotentNormalHall M

/-- `M_F ≤ M`: basic containment, directly from the `sSup` construction.  (The §15
well-definedness that the `sSup` is again Hall is `maxNilpotentNormalHall_isHall` below; this
containment is the more elementary half.) -/
theorem maxNilpotentNormalHall_le (M : Subgroup G) : maxNilpotentNormalHall M ≤ M :=
  sSup_le fun _ hN => hN.1

set_option backward.isDefEq.respectTransparency false in
/-- `M` normalizes `M_F` (so `(M_F).subgroupOf M ⊴ M`): the `sSup` of `M`-normal
candidates is again `M`-normal.  Like `maxNilpotentNormalHall_le`, this is part of the
`§14`-independent §15 well-definedness (the Hall property is `maxNilpotentNormalHall_isHall`);
each candidate `N` is fixed by conjugation by `m ∈ M` because `(N.subgroupOf M).Normal`. -/
theorem maxNilpotentNormalHall_le_normalizer (M : Subgroup G) :
    M ≤ Subgroup.normalizer (maxNilpotentNormalHall M) := by
  intro m hm
  refine mem_normalizer_of_conj_smul_eq_self ?_
  unfold maxNilpotentNormalHall
  rw [Subgroup.pointwise_smul_def, (Subgroup.gc_map_comap _).l_sSup, sSup_eq_iSup]
  refine iSup_congr fun N => iSup_congr fun hN => ?_
  rw [← Subgroup.pointwise_smul_def]
  obtain ⟨hNM, hNnorm, -, -⟩ := hN
  exact conj_smul_eq_self_of_mem_normalizer
    (((Subgroup.normal_subgroupOf_iff_le_normalizer hNM).mp hNnorm) hm)

/-- `M_F ⊴ M` in the relative sense `(M_F).subgroupOf M`: the directly usable form of
`maxNilpotentNormalHall_le_normalizer`, matching the normality clause in the defining
predicate of `M_F`. -/
theorem maxNilpotentNormalHall_subgroupOf_normal (M : Subgroup G) :
    ((maxNilpotentNormalHall M).subgroupOf M).Normal :=
  (Subgroup.normal_subgroupOf_iff_le_normalizer (maxNilpotentNormalHall_le M)).mpr
    (maxNilpotentNormalHall_le_normalizer M)

/-- `M_F ≤ F(M)`: the maximal nilpotent normal Hall subgroup lies inside the Fitting subgroup
`F(M)` (`OddOrder.BG.Ch2.S08.fittingInG`, defeq `(Ch01.fitting ↥M).map M.subtype`), because each
candidate `N` is nilpotent and normal in `M` (so `N.subgroupOf M ≤ fitting ↥M`).  `§14`-independent;
the structural bridge between `M_F` and `F(M)` used throughout §15/§16. -/
theorem maxNilpotentNormalHall_le_fittingInG [Finite G] (M : Subgroup G) :
    maxNilpotentNormalHall M ≤ OddOrder.BG.Ch2.S08.fittingInG M := by
  refine sSup_le fun N hN => ?_
  obtain ⟨hNM, hNnorm, hNnil, -⟩ := hN
  have := hNnorm
  have := hNnil
  calc N = (N.subgroupOf M).map M.subtype := (Subgroup.map_subgroupOf_eq_of_le hNM).symm
    _ ≤ OddOrder.BG.Ch2.S08.fittingInG M :=
        Subgroup.map_mono OddOrder.Isaacs.Ch01.nilpotent_normal_le_fitting

/-- **`M_F` is nilpotent** — the §15 well-definedness piece that the defining `sSup` is again
nilpotent.  `M_F ≤ F(M)` (`maxNilpotentNormalHall_le_fittingInG`) and `F(M)` is nilpotent
(image of the nilpotent `fitting ↥M` under the injective `M.subtype`).  `§14`-independent. -/
theorem maxNilpotentNormalHall_isNilpotent [Finite G] (M : Subgroup G) :
    Group.IsNilpotent ↥(maxNilpotentNormalHall M) := by
  have : Group.IsNilpotent ↥(OddOrder.Isaacs.Ch01.fitting (↥M)) :=
    OddOrder.Isaacs.Ch01.fitting.isNilpotent
  have : Group.IsNilpotent ↥(OddOrder.BG.Ch2.S08.fittingInG M) :=
    Group.nilpotent_of_mulEquiv
      (Subgroup.equivMapOfInjective (OddOrder.Isaacs.Ch01.fitting (↥M)) M.subtype
        M.subtype_injective)
  exact Group.nilpotent_of_mulEquiv
    (Subgroup.subgroupOfEquivOfLe (maxNilpotentNormalHall_le_fittingInG M))

/-- **`M_F` absorbs every nilpotent normal Hall subgroup** (`§14`-independent): a subgroup `N ≤ M`
that is `M`-normal (`(N.subgroupOf M).Normal`), nilpotent, and a Hall `π(N)`-subgroup of `M`
(`IsHallSubgroup (Nat.card ↥N).primeFactors (N.subgroupOf M)`) is one of the candidates in the
`sSup` defining `M_F = maxNilpotentNormalHall M`, hence `≤ M_F`.  This is the `le_sSup` half of the
`M_F` characterization (the converse `M_F ≤ M_σ` is `maxNilpotentNormalHall_le_Msigma`); it supplies
`Q ≤ M_F` in Theorem 15.2 once `Q = O_q(M)` is shown to be such a subgroup. -/
theorem le_maxNilpotentNormalHall {M N : Subgroup G} (hNM : N ≤ M)
    (hNnorm : (N.subgroupOf M).Normal) (hNnil : Group.IsNilpotent ↥(N.subgroupOf M))
    (hNhall : OddOrder.Isaacs.Ch03.IsHallSubgroup (Nat.card ↥N).primeFactors (N.subgroupOf M)) :
    N ≤ maxNilpotentNormalHall M :=
  le_sSup ⟨hNM, hNnorm, hNnil, hNhall⟩

/-- If `M_σ` is nilpotent, then `M_σ ≤ M_F`: `M_σ` is then a nilpotent normal Hall subgroup of
`M` (normal `σ`-core, `σ`-Hall by `Msigma_isHall`, nilpotent by hypothesis), hence one of the
candidates in the `sSup` defining `M_F`.  `§14`-independent.  Combined with the (gated)
`M_F ≤ M_σ` of Theorem A this gives `M_F = M_σ ⟺ M_σ` nilpotent (recall `M_F` is always
nilpotent, `maxNilpotentNormalHall_isNilpotent`). -/
theorem Msigma_le_maxNilpotentNormalHall_of_nilpotent [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hnil : Group.IsNilpotent ↥(OddOrder.BG.Ch3.S10.Msigma M)) :
    OddOrder.BG.Ch3.S10.Msigma M ≤ maxNilpotentNormalHall M := by
  have := hnil
  have hle := OddOrder.BG.Ch3.S10.Msigma_le M
  have hcard : Nat.card ↥((OddOrder.BG.Ch3.S10.Msigma M).subgroupOf M) =
      Nat.card ↥(OddOrder.BG.Ch3.S10.Msigma M) :=
    Nat.card_congr (Subgroup.subgroupOfEquivOfLe hle).toEquiv
  apply le_sSup
  refine ⟨hle, ?_, ?_, ?_⟩
  · rw [OddOrder.BG.Ch3.S10.Msigma_subgroupOf]; infer_instance
  · exact Group.nilpotent_of_mulEquiv (Subgroup.subgroupOfEquivOfLe hle).symm
  · obtain ⟨hHcard, hHidx⟩ := OddOrder.BG.Ch3.S10.Msigma_isHall hG hM
    refine ⟨fun q hq => by rwa [hcard] at hq, fun q hq hqπ => ?_⟩
    obtain ⟨hqp, hqd, -⟩ := Nat.mem_primeFactors.mp hq
    have hdvd : ((OddOrder.BG.Ch3.S10.Msigma M).subgroupOf M).index ∣
        (OddOrder.BG.Ch3.S10.Msigma M).index :=
      Subgroup.relIndex_dvd_index_of_le hle
    exact hHidx q
      (Nat.mem_primeFactors.mpr ⟨hqp, hqd.trans hdvd, Subgroup.index_ne_zero_of_finite⟩)
      (hHcard q hqπ)

/-- **`M_F ≤ M_σ`** (BG §15, mmd L4116 "it is easy to see ... `M_F` ... lies in `M_σ`").
Every nilpotent normal Hall subgroup `N` of `M` is a `σ(M)`-group: for a prime `p ∣ |N|`, the
Sylow `p`-subgroup of `N` is characteristic in `N` (nilpotent) hence normal in `M`, and a full
Sylow `p`-subgroup of `M` (`N` Hall), so it maps to `O_p(M)`; minimality of `M` and simplicity
of `G` give `N_G(O_p(M)) ≤ M`, i.e. `p ∈ σ(M)`.  Then `N ≤ O_{σ(M)}(M) = M_σ` by maximality of
the `σ(M)`-core.  `§14`-independent. -/
theorem maxNilpotentNormalHall_le_Msigma [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G) :
    maxNilpotentNormalHall M ≤ OddOrder.BG.Ch3.S10.Msigma M := by
  refine sSup_le fun N hN => ?_
  obtain ⟨hNM, hNnorm, hNnil, hNHall⟩ := hN
  have := hNnorm
  have := hNnil
  set N' : Subgroup ↥M := N.subgroupOf M with hN'def
  have hcardN' : Nat.card ↥N' = Nat.card ↥N :=
    Nat.card_congr (Subgroup.subgroupOfEquivOfLe hNM).toEquiv
  -- `N'` is a `σ(M)`-group: every prime divisor of `|N|` lies in `σ(M)`.
  have hpi : OddOrder.Isaacs.Ch03.Subgroup.IsPiGroup (OddOrder.BG.Ch3.S10.sigma M) N' := by
    intro p hp
    have hpN : p ∈ (Nat.card ↥N).primeFactors := by rwa [hcardN'] at hp
    have : Fact p.Prime := ⟨Nat.prime_of_mem_primeFactors hp⟩
    have hpM : p ∈ (Nat.card ↥M).primeFactors :=
      Nat.mem_primeFactors.mpr ⟨‹Fact p.Prime›.1,
        (Nat.dvd_of_mem_primeFactors hpN).trans
          (hcardN' ▸ Subgroup.card_subgroup_dvd_card N'), Nat.card_pos.ne'⟩
    -- `p ∤ [M : N']` (Hall), so the `p`-part of `|M|` is concentrated in `N'`.
    have hpidx : ¬ p ∣ N'.index := fun hdvd =>
      hNHall.2 p (Nat.mem_primeFactors.mpr
        ⟨‹Fact p.Prime›.1, hdvd, Subgroup.index_ne_zero_of_finite⟩) hpN
    have hfact : (Nat.card ↥M).factorization p = (Nat.card ↥N').factorization p := by
      conv_lhs => rw [← Subgroup.card_mul_index N']
      rw [Nat.factorization_mul Nat.card_pos.ne' Subgroup.index_ne_zero_of_finite,
        Finsupp.add_apply, Nat.factorization_eq_zero_of_not_dvd hpidx, add_zero]
    -- The Sylow `p`-subgroup of the nilpotent `N'` is characteristic, hence normal in `M`,
    -- and (Hall) a full Sylow `p`-subgroup of `M`.
    obtain ⟨SN⟩ := (inferInstance : Nonempty (Sylow p ↥N'))
    have hSNnorm : (SN : Subgroup ↥N').Normal :=
      OddOrder.Isaacs.Ch01.Sylow.normal_of_isNilpotent SN
    have hSNchar : (SN : Subgroup ↥N').Characteristic :=
      Sylow.characteristic_of_normal SN hSNnorm
    have hScard : Nat.card ↥((SN : Subgroup ↥N').map N'.subtype) =
        p ^ (Nat.card ↥M).factorization p := by
      rw [Subgroup.card_map_of_injective N'.subtype_injective, SN.card_eq_multiplicity, hfact]
    have hmapnorm : ((SN : Subgroup ↥N').map N'.subtype).Normal := inferInstance
    let P : Sylow p ↥M := Sylow.ofCard ((SN : Subgroup ↥N').map N'.subtype) hScard
    have hPnorm : (P : Subgroup ↥M).Normal := hmapnorm
    have hPmap : (P : Subgroup ↥M).map M.subtype = OddOrder.GroupTheory.opiCoreInG {p} M :=
      OddOrder.BG.Ch3.S10.sylowMap_eq_opiCoreInG_singleton_of_normal P hPnorm
    have hPne : OddOrder.GroupTheory.opiCoreInG {p} M ≠ ⊥ :=
      OddOrder.BG.Ch3.S10.opiCoreInG_singleton_ne_bot_of_sylowMap_eq hpM P hPmap
    rw [OddOrder.BG.Ch3.S10.mem_sigma_iff]
    refine ⟨hpM, P, ?_⟩
    rw [hPmap]
    exact OddOrder.BG.Ch2.S09.normalizer_opiCoreInG_singleton_le_maximal_of_ne_bot hG hM hPne
  calc N = N'.map M.subtype := (Subgroup.map_subgroupOf_eq_of_le hNM).symm
    _ ≤ (OddOrder.Isaacs.Ch03.oPiCore (OddOrder.BG.Ch3.S10.sigma M) ↥M).map M.subtype :=
        Subgroup.map_mono hpi.le_oPiCore
    _ = OddOrder.BG.Ch3.S10.Msigma M := rfl

/-- **`M_F = M_σ ⟺ M_σ` is nilpotent** (`§14`-independent).  Forward: `M_F` is always nilpotent
(`maxNilpotentNormalHall_isNilpotent`).  Backward: `maxNilpotentNormalHall_le_Msigma` (always)
and `Msigma_le_maxNilpotentNormalHall_of_nilpotent` give equality by antisymmetry. -/
theorem maxNilpotentNormalHall_eq_Msigma_iff_isNilpotent [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G) :
    maxNilpotentNormalHall M = OddOrder.BG.Ch3.S10.Msigma M ↔
      Group.IsNilpotent ↥(OddOrder.BG.Ch3.S10.Msigma M) := by
  refine ⟨fun h => h ▸ maxNilpotentNormalHall_isNilpotent M, fun hnil => ?_⟩
  exact le_antisymm (maxNilpotentNormalHall_le_Msigma hG hM)
    (Msigma_le_maxNilpotentNormalHall_of_nilpotent hG hM hnil)

/-- **`M_F ≤ M'`** (the `H ⊆ M'` part of BG Corollary 15.5(c), `H = M_F`): the containment chain
`M_F ≤ M_σ ≤ M' ≤ M` via `maxNilpotentNormalHall_le_Msigma` and `Msigma_le_derived`.
`§14`-independent.  ⚠ The other half of 15.5(c), "`M'/M_F` is nilpotent", is **proved** as
`S16.derivedInG_quotient_maxNilpotentNormalHall_isNilpotent` (downstream, so it cannot appear in
a §15 statement without an import cycle).  The old "remains deferred — quotient API" note was
stale; corrected in the issue 0177 §15 audit (2026-08-08). -/
theorem maxNilpotentNormalHall_le_derived [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G) :
    maxNilpotentNormalHall M ≤ derivedInG M :=
  (maxNilpotentNormalHall_le_Msigma hG hM).trans (OddOrder.BG.Ch3.S10.Msigma_le_derived hG hM)

/-- **General helper (§14-independent, reusable).**  In a finite group, the `sSup` of a set `s`
of normal `π`-subgroups is again a `π`-subgroup.  (Mirrors the `Finset.sup_induction` argument of
`Ch03.oPiCore.isPiGroup`, but for an arbitrary set rather than the defining family of `oPiCore`.)
Used for the §15 well-definedness "`M_F` is Hall": the supremum defining `M_F` collapses to a join
of normal subgroups, whose order has no new prime factors. -/
theorem isPiGroup_sSup_of_forall_normal {H : Type*} [Group H] [Finite H] (π : Set ℕ)
    (s : Set (Subgroup H)) (hnorm : ∀ N ∈ s, N.Normal)
    (hpi : ∀ N ∈ s, Ch03.Subgroup.IsPiGroup π N) :
    Ch03.Subgroup.IsPiGroup π (sSup s) := by
  classical
  have hSubF : Fintype {N : Subgroup H // N ∈ s} := Fintype.ofFinite _
  -- Carry both normality and the `π`-group property through the induction.
  set q : Subgroup H → Prop := fun K => K.Normal ∧ Ch03.Subgroup.IsPiGroup π K with hq_def
  suffices hgoal : q (sSup s) by exact hgoal.2
  -- Rewrite the `sSup` as a `Finset.sup` over the (finite) subtype `{N // N ∈ s}`.
  have hsup : sSup s =
      (Finset.univ : Finset {N : Subgroup H // N ∈ s}).sup (fun N => (N.val : Subgroup H)) := by
    rw [Finset.sup_eq_iSup]
    simp only [iSup_pos, Finset.mem_univ]
    rw [← sSup_eq_iSup']
  rw [hsup]
  refine Finset.sup_induction (p := q) ?_ ?_ ?_
  · -- `⊥` is a normal `π`-group (no prime factors).
    refine ⟨inferInstance, ?_⟩
    intro r hr
    simp only [Subgroup.card_bot, Nat.primeFactors_one, Finset.notMem_empty] at hr
  · -- closure under join of two normal `π`-subgroups.
    rintro a₁ ⟨ha₁N, ha₁Pi⟩ a₂ ⟨ha₂N, ha₂Pi⟩
    have := ha₁N
    have := ha₂N
    exact ⟨inferInstance, Ch03.Subgroup.IsPiGroup.sup_of_normal ha₁Pi ha₂Pi⟩
  · -- each generator is a normal `π`-group by hypothesis.
    intro b _
    exact ⟨hnorm b.val b.2, hpi b.val b.2⟩

/-- **BG §15 well-definedness: `M_F` is a Hall subgroup of `M`** (general finite-group fact,
§14-independent; no minimal-simple hypothesis needed).  `M_F = maxNilpotentNormalHall M` is, as a
relative subgroup `(M_F).subgroupOf M ⊴ M`, a `π(M_F)`-Hall subgroup of `M`, where
`π(M_F) = (Nat.card ↥(M_F)).primeFactors`.

Proof: `(M_F).subgroupOf M` is the join (in `↥M`) of the normal subgroups `N.subgroupOf M`
ranging over the candidates `N` of the `sSup` defining `M_F`; this join has no prime factor beyond
those of the candidates (`isPiGroup_sSup_of_forall_normal`).  Hence every prime `p ∣ |M_F|` divides
some candidate `|N₀|`, and that candidate is `π(N₀)`-Hall in `↥M` with `p ∈ π(N₀)`, so
`p ∤ [M : N₀.subgroupOf M]`; since `N₀.subgroupOf M ≤ (M_F).subgroupOf M`, also `p ∤ [M : M_F]`.
The index of `(M_F).subgroupOf M` therefore shares no prime with `|M_F|`, i.e. `M_F` is Hall. -/
theorem maxNilpotentNormalHall_isHall [Finite G] (M : Subgroup G) :
    Ch03.IsHallSubgroup (Nat.card ↥(maxNilpotentNormalHall M)).primeFactors
      ((maxNilpotentNormalHall M).subgroupOf M) := by
  classical
  set H : Subgroup G := maxNilpotentNormalHall M with hHdef
  set Hbar : Subgroup ↥M := H.subgroupOf M with hHbar
  -- The defining set of `M_F`, and its image in `↥M`.
  set S : Set (Subgroup G) := {N : Subgroup G | N ≤ M ∧ (N.subgroupOf M).Normal ∧
      Group.IsNilpotent ↥(N.subgroupOf M) ∧
      Ch03.IsHallSubgroup (Nat.card ↥N).primeFactors (N.subgroupOf M)} with hSdef
  have hHsSup : H = sSup S := rfl
  set T : Set (Subgroup ↥M) := (fun N => N.subgroupOf M) '' S with hTdef
  -- `card Hbar = card H` (since `H ≤ M`).
  have hHle : H ≤ M := maxNilpotentNormalHall_le M
  have hcardHbar : Nat.card ↥Hbar = Nat.card ↥H :=
    Nat.card_congr (Subgroup.subgroupOfEquivOfLe hHle).toEquiv
  -- `⨆ x ∈ T, map x = ⨆ N ∈ S, N`, since `map (N.subgroupOf M) = N` for `N ≤ M ∈ S`.
  have hiSup : (⨆ x ∈ T, x.map M.subtype) = ⨆ N ∈ S, N := by
    rw [hTdef, iSup_image]
    refine iSup_congr fun N => iSup_congr fun hN => ?_
    exact Subgroup.map_subgroupOf_eq_of_le hN.1
  -- Key identification: `Hbar = sSup T`.
  have hHbar_eq : Hbar = sSup T := by
    apply Subgroup.map_injective M.subtype_injective
    rw [hHbar, Subgroup.map_subgroupOf_eq_of_le hHle, hHsSup,
      (Subgroup.gc_map_comap M.subtype).l_sSup, hiSup, ← sSup_eq_iSup]
  -- Every element of `T` is normal in `↥M` and a `Pri`-group, where `Pri` is the set of primes
  -- occurring in some candidate's order.
  set Pri : Set ℕ := {p | ∃ N ∈ S, p ∈ (Nat.card ↥N).primeFactors} with hPridef
  have hTnorm : ∀ x ∈ T, x.Normal := by
    rintro x ⟨N, hN, rfl⟩; exact hN.2.1
  have hTpi : ∀ x ∈ T, Ch03.Subgroup.IsPiGroup Pri x := by
    rintro x ⟨N, hN, rfl⟩
    intro p hp
    -- `card (N.subgroupOf M) = card N`, so a prime factor of it is one of `N`.
    have hcardN : Nat.card ↥(N.subgroupOf M) = Nat.card ↥N :=
      Nat.card_congr (Subgroup.subgroupOfEquivOfLe hN.1).toEquiv
    exact ⟨N, hN, by rwa [hcardN] at hp⟩
  -- Hence `Hbar` is a `Pri`-group.
  have hHbarpi : Ch03.Subgroup.IsPiGroup Pri Hbar := by
    rw [hHbar_eq]; exact isPiGroup_sSup_of_forall_normal Pri T hTnorm hTpi
  -- Assemble the Hall property.
  refine ⟨fun p hp => ?_, fun p hp hpπ => ?_⟩
  · -- Prime factors of `|Hbar| = |H|` are exactly `π(M_F)`.
    rwa [hcardHbar] at hp
  · -- `p ∈ π(M_F)` and `p ∣ [M : Hbar]`: derive a contradiction.
    have : Fact p.Prime := ⟨Nat.prime_of_mem_primeFactors hp⟩
    -- `p ∈ π(M_F)`, so `p ∣ |Hbar|`, so `p ∈ Π`: pick a candidate `N₀` with `p ∣ |N₀|`.
    have hpHbar : p ∈ (Nat.card ↥Hbar).primeFactors := by
      rw [hcardHbar]; exact hpπ
    obtain ⟨N₀, hN₀, hpN₀⟩ := hHbarpi p hpHbar
    -- `N₀.subgroupOf M` is `π(N₀)`-Hall with `p ∈ π(N₀)`, so `p ∤ [M : N₀.subgroupOf M]`.
    have hpN₀Hall : ¬ p ∣ (N₀.subgroupOf M).index := by
      intro hdvd
      exact hN₀.2.2.2.2 p
        (Nat.mem_primeFactors.mpr ⟨Fact.out, hdvd, Subgroup.index_ne_zero_of_finite⟩) hpN₀
    -- `N₀.subgroupOf M ≤ Hbar`, so `[M : Hbar] ∣ [M : N₀.subgroupOf M]`, so `p ∤ [M : Hbar]`.
    have hN₀_le_H : N₀ ≤ H := hHsSup ▸ le_sSup hN₀
    have hle : N₀.subgroupOf M ≤ Hbar := by
      rw [hHbar]; exact Subgroup.subgroupOf_mono M hN₀_le_H
    exact hpN₀Hall ((Nat.mem_primeFactors.mp hp).2.1.trans (Subgroup.index_dvd_of_le hle))

/-- The Fitting subgroup of `M`, viewed in the ambient group as in BG §8/§15. -/
noncomputable abbrev fittingInAmbient (M : Subgroup G) : Subgroup G :=
  OddOrder.BG.Ch2.S08.fittingInG M

/-- **`F(H) = H` for a nilpotent subgroup `H`** (`§14`-independent, reusable): the Fitting
subgroup of a nilpotent group is the whole group (`fitting ↥H = ⊤`), so its ambient realization
`fittingInAmbient H` is just `H`.  Used by Corollary 15.5 (the `M_F = M_σ` case, where `M_σ` is
nilpotent so `F(M_σ) = M_σ = M_F`) and the `M_F` cyclic ⟹ `F(M)` cyclic step of Corollary 15.6. -/
theorem fittingInAmbient_eq_self_of_isNilpotent [Finite G] {H : Subgroup G}
    [Group.IsNilpotent ↥H] : fittingInAmbient H = H := by
  have : Group.IsNilpotent ↥(⊤ : Subgroup ↥H) :=
    Group.nilpotent_of_mulEquiv Subgroup.topEquiv.symm
  have htop : OddOrder.Isaacs.Ch01.fitting ↥H = ⊤ :=
    top_le_iff.mp OddOrder.Isaacs.Ch01.nilpotent_normal_le_fitting
  change (OddOrder.Isaacs.Ch01.fitting ↥H).map H.subtype = H
  rw [htop, ← MonoidHom.range_eq_map, Subgroup.range_subtype]

/-- **Converse of `fittingInAmbient_eq_self_of_isNilpotent`** (`§14`-independent, reusable): if the
ambient realization of the Fitting subgroup is all of `H` (`fittingInAmbient H = H`), then `H` is
nilpotent.  Together with the forward direction this is `IsNilpotent ↥H ↔ fittingInAmbient H = H`.

Used in Theorem 15.2's step 2 (the contrapositive of Theorem 3.8): from `⁅M_σ, K⁆ = M_σ ⊆ F(M_σ)`
one gets `F(M_σ) = M_σ`, hence `M_σ` nilpotent — contradicting `M_F ≠ M_σ`. -/
theorem isNilpotent_of_fittingInAmbient_eq_self [Finite G] {H : Subgroup G}
    (h : fittingInAmbient H = H) : Group.IsNilpotent ↥H := by
  have htop : OddOrder.Isaacs.Ch01.fitting ↥H = ⊤ := by
    apply Subgroup.map_injective H.subtype_injective
    rw [← MonoidHom.range_eq_map, Subgroup.range_subtype]
    exact h
  have hnil : Group.IsNilpotent ↥(OddOrder.Isaacs.Ch01.fitting ↥H) :=
    OddOrder.Isaacs.Ch01.fitting.isNilpotent
  rw [htop] at hnil
  have := hnil
  exact Group.nilpotent_of_mulEquiv Subgroup.topEquiv

/-- The conjugation action commutator of `H` on itself is the ordinary derived subgroup:
`actionCommutator (MulAut.conj) = commutator H`.  (Both are the closure of the commutator
elements `g * (a g⁻¹ a⁻¹) = ⁅g, a⁆`.) -/
private theorem actionCommutator_conj_eq_commutator {H : Type*} [Group H] :
    OddOrder.Isaacs.Ch04.actionCommutator (MulAut.conj : H →* MulAut H) = commutator H := by
  rw [commutator_eq_closure]
  unfold OddOrder.Isaacs.Ch04.actionCommutator
  congr 1
  ext x
  simp only [Set.mem_ofPred_eq, commutatorSet_def]
  constructor
  · rintro ⟨g, a, rfl⟩
    exact ⟨g, a, by rw [commutatorElement_def, MulAut.conj_apply]; group⟩
  · rintro ⟨g₁, g₂, rfl⟩
    exact ⟨g₁, g₂, by rw [commutatorElement_def, MulAut.conj_apply]; group⟩

open OddOrder.BG.Ch1.OperatorQuotientAction in
/-- **Aut-abelian core**: if `N ⊴ H` is cyclic, then `H' ≤ C_H(N)`.  Conjugation gives a
homomorphism `H → Aut(↥N)` with kernel `C_H(N)`; `↥N` cyclic makes `Aut(↥N)` abelian, so the
derived subgroup `H'` lands in the kernel.  Reuses the BG Thm 4.12 machinery
(`actionCommutator_le_centralizer_of_isCyclic_isAInvariant`) with `φ = MulAut.conj`. -/
theorem commutator_le_centralizer_of_normal_isCyclic {H : Type*} [Group H]
    {N : Subgroup H} [IsCyclic ↥N] (hN : N.Normal) :
    commutator H ≤ Subgroup.centralizer (N : Set H) := by
  have hinv : OddOrder.Isaacs.Ch03.IsAInvariant (MulAut.conj : H →* MulAut H) N := by
    rw [OddOrder.Isaacs.Ch03.isAInvariant_iff_smul_mem]
    intro a g hg
    rw [MulAut.conj_apply]
    exact hN.conj_mem g hg a
  have h := actionCommutator_le_centralizer_of_isCyclic_isAInvariant
    (φ := (MulAut.conj : H →* MulAut H)) (S := N) hN hinv
  rwa [actionCommutator_conj_eq_commutator] at h

/-- **BG §15, `F(M)` cyclic ⟹ `M'' = 1`** (`§14`-independent).  If the ambient Fitting subgroup
`F(M)` is cyclic then `M' ≤ C_M(F(M)) ≤ F(M)` (Aut-abelian core + Fitting self-centralizing),
so `M'` is abelian and `M'' = 1`.  This is the engine of Corollary 15.6's "`M_F` not cyclic"
clause. -/
theorem fittingInAmbient_cyclic_imp_derivedDerived_eq_bot [Finite G] {M : Subgroup G}
    [Group.IsSolvable ↥M] (hcyc : IsCyclic ↥(fittingInAmbient M)) :
    derivedInG (derivedInG M) = ⊥ := by
  -- `fitting ↥M` is cyclic (transport of `hcyc` along `fittingInAmbient M ≅ fitting ↥M`)
  have hfitcyc : IsCyclic ↥(OddOrder.Isaacs.Ch01.fitting ↥M) := by
    have e : ↥(OddOrder.Isaacs.Ch01.fitting ↥M) ≃* ↥(fittingInAmbient M) :=
      Subgroup.equivMapOfInjective (OddOrder.Isaacs.Ch01.fitting ↥M) M.subtype M.subtype_injective
    exact isCyclic_of_surjective e.symm e.symm.surjective
  -- `commutator ↥M ≤ centralizer (fitting ↥M) ≤ fitting ↥M`
  have hcomm_le : commutator ↥M ≤ OddOrder.Isaacs.Ch01.fitting ↥M :=
    (commutator_le_centralizer_of_normal_isCyclic
      (inferInstance : (OddOrder.Isaacs.Ch01.fitting ↥M).Normal)).trans
      OddOrder.GroupTheory.centralizer_fitting_le_fitting
  -- push to the ambient group: `M' = ⁅M, M⁆ ≤ F(M)`
  have hderiv_le : derivedInG M ≤ fittingInAmbient M := Subgroup.map_mono hcomm_le
  -- `F(M)` cyclic ⟹ abelian ⟹ `F(M) ≤ C_G(F(M))`
  have hself : (fittingInAmbient M : Subgroup G) ≤ Subgroup.centralizer (fittingInAmbient M) := by
    have := hcyc
    intro x hx
    rw [Subgroup.mem_centralizer_iff]
    intro y hy
    have h : (⟨y, hy⟩ : ↥(fittingInAmbient M)) * ⟨x, hx⟩ = ⟨x, hx⟩ * ⟨y, hy⟩ := mul_comm' _ _
    calc y * x = ↑((⟨y, hy⟩ : ↥(fittingInAmbient M)) * ⟨x, hx⟩) := by rw [Subgroup.coe_mul]
      _ = ↑((⟨x, hx⟩ : ↥(fittingInAmbient M)) * ⟨y, hy⟩) := by rw [h]
      _ = x * y := by rw [Subgroup.coe_mul]
  -- `M'' = ⁅M', M'⁆ ≤ ⁅F(M), F(M)⁆ = 1`
  rw [show derivedInG (derivedInG M) = ⁅derivedInG M, derivedInG M⁆ from
        Subgroup.map_subtype_commutator (derivedInG M)]
  refine le_bot_iff.mp (le_trans (Subgroup.commutator_mono hderiv_le hderiv_le) ?_)
  exact le_of_eq (Subgroup.commutator_eq_bot_iff_le_centralizer.mpr hself)

/-- **§14-independent engine for BG Corollary 15.6's `K* ⊆ M''` clause.**  Given the type-`P`
complement structure `M = K M'` with `K ∩ M' = 1` (BG Theorem 14.7(h), here as the relative-
complement hypothesis `hcompl` inside `↥M`) and the coprimality of `|M'|` and `|K|`, Lemma 6.3
(`centralizer_inf_le_derivedInG_of_isComplement'`, proved) yields `C_{M'}(K) ⊆ M''`; since
`K* = C_{M_σ}(K) = M_σ ⊓ C_G(K) ⊆ C_{M'}(K)` (because `M_σ ⊆ M'`), this gives `K* ⊆ M''`.

The complement/coprimality data are the only `§14` inputs; everything else is unconditional, so
Corollary 15.6 reduces its `K* ⊆ M''` clause to a single citation once Theorem 14.7(h) (or the
`IsComplement' (derivedInG M) K` strengthening of Lemma 15.1) lands.  Mirrors the Lemma 6.3
transport in `S12_E.Msigma_E_relations`, with `M'` in place of `M_σ` and `K` in place of `E`. -/
theorem Msigma_inf_centralizer_le_derivedDerived_of_isComplement' [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M K : Subgroup G} (hM : M ∈ maximalSubgroups G)
    (hcompl : Subgroup.IsComplement' ((derivedInG M).subgroupOf M) (K.subgroupOf M))
    (hcop : Nat.Coprime (Nat.card ↥((derivedInG M).subgroupOf M))
      (Nat.card ↥(K.subgroupOf M))) :
    OddOrder.BG.Ch3.S10.Msigma M ⊓ Subgroup.centralizer (K : Set G) ≤
      derivedInG (derivedInG M) := by
  classical
  have : Group.IsSolvable ↥M := hG.isSolvable_of_mem_maximalSubgroups hM
  have hMσM : OddOrder.BG.Ch3.S10.Msigma M ≤ M := OddOrder.BG.Ch3.S10.Msigma_le M
  have hMσ_le_deriv : OddOrder.BG.Ch3.S10.Msigma M ≤ derivedInG M :=
    OddOrder.BG.Ch3.S10.Msigma_le_derived hG hM
  have hderivM : derivedInG M ≤ M := Subgroup.map_subtype_le _
  -- `H := (derivedInG M).subgroupOf M = commutator ↥M`, normal in `↥M`.
  have hid : (derivedInG M).subgroupOf M = commutator ↥M :=
    Subgroup.comap_map_eq_self_of_injective M.subtype_injective (commutator ↥M)
  have hHnorm : ((derivedInG M).subgroupOf M).Normal := by rw [hid]; infer_instance
  have hH_le : (derivedInG M).subgroupOf M ≤ commutator ↥M := le_of_eq hid
  -- Lemma 6.3 inside `↥M`:  `C_{↥M}(K') ⊓ H ≤ derivedInG H`.
  have h632 := OddOrder.BG.Ch1.S06.centralizer_inf_le_derivedInG_of_isComplement'
    (G := ↥M) hcompl hH_le hcop
  -- Transport identity:  `(derivedInG H).map M.subtype = derivedInG (derivedInG M)`.
  have hderiv_transport :
      (derivedInG ((derivedInG M).subgroupOf M)).map M.subtype =
        derivedInG (derivedInG M) := by
    rw [show derivedInG ((derivedInG M).subgroupOf M)
          = ⁅(derivedInG M).subgroupOf M, (derivedInG M).subgroupOf M⁆
          from Subgroup.map_subtype_commutator _,
      Subgroup.map_commutator, Subgroup.map_subgroupOf_eq_of_le hderivM,
      show ⁅(derivedInG M : Subgroup G), derivedInG M⁆ = derivedInG (derivedInG M)
          from (Subgroup.map_subtype_commutator _).symm]
  -- Pointwise: every `x ∈ M_σ ⊓ C_G(K)` lands in `derivedInG (derivedInG M)`.
  intro x hx
  obtain ⟨hxMσ, hxC⟩ := hx
  have hxM : x ∈ M := hMσM hxMσ
  have hxmem : (⟨x, hxM⟩ : ↥M) ∈
      Subgroup.centralizer ((K.subgroupOf M : Subgroup ↥M) : Set ↥M)
        ⊓ (derivedInG M).subgroupOf M := by
    refine ⟨Subgroup.mem_centralizer_iff.mpr ?_, ?_⟩
    · intro k' hk'
      have hkK : ((k' : ↥M) : G) ∈ K := Subgroup.mem_subgroupOf.mp hk'
      exact Subtype.ext (Subgroup.mem_centralizer_iff.mp hxC (k' : G) hkK)
    · exact Subgroup.mem_subgroupOf.mpr (hMσ_le_deriv hxMσ)
  have hmapped := Subgroup.mem_map_of_mem M.subtype (h632 hxmem)
  rwa [hderiv_transport] at hmapped

/-- The nonidentity part of the ambient Fitting subgroup of `M`. -/
def fittingSharp (M : Subgroup G) : Set G :=
  sharpSubgroup (fittingInAmbient M)

/-- The BG §15 hypothesis that `F(M)` is a TI-subgroup of `G`. -/
def FittingIsTI (M : Subgroup G) : Prop :=
  IsTISubset (fittingSharp M) (Subgroup.normalizer ((fittingInAmbient M : Subgroup G) : Set G))

/-- The subgroup generated by the centralizers in `U` of nonidentity elements of
`M_sigma`; this packages BG Lemma 15.1(d). -/
noncomputable def centralizerGeneratedBySigma (M U : Subgroup G) : Subgroup G :=
  sSup {C : Subgroup G | ∃ x ∈ sigmaSharp M,
    C = U ⊓ Subgroup.centralizer ({x} : Set G)}

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/AutAbelianCore.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