BG §16: Proposition 16.1 taxonomy and Theorems I/II (output for Peterfalvi)
DefinitionOddOrder_BG_Ch4_FamilyOfMaximal_S16_MainResults_TaxonomyOutputfeit-thompsonfinite-groupsgroup-theoryishida-odd-order
The Proposition 16.1 taxonomy and Theorems I and II of Bender–Glauberman, Local Analysis for the Odd Order Theorem, §16, in the form consumed by Peterfalvi's text. It consists of lemmas only, including theoremII_conjunct1_of_inputs and the tame embedding of Theorem II.
The bundle consists of lemmas only.
The bundle also proves 5 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/TaxonomyOutput.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_TypeBridges
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S16_MainResults_LocalTaxonomy
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]
/-- **Assembly for BG Theorem II (Ti)** (mmd L4546--L4550), as a `sorry`-free,
axiom-clean *gated-endpoint skeleton*.
The mmd proof decomposes `A_0(M)` into the disjoint pieces `M_σ`, `A(M) - M_σ`,
and `A_0(M) - A(M)`, observes that cross-piece elements have distinct orders (so
are never `G`-conjugate), and concludes:
* within `M_σ`, `G`-conjugacy is `M`-conjugacy by **Theorem D(1)** (`hD1`);
* within either TI piece (**Theorem B(5)**/`hTI_B`, **Theorem C(9)**/`hTI_C`),
`G`-conjugacy forces the conjugator into `M` by the TI condition.
This bundles those three inputs plus the cross-piece exclusion `hPieceInv`
(`G`-conjugate elements of `X` share `M_σ`- and `A(M)`-membership — the formal
content of "distinct orders across pieces") and discharges (Ti) with no `sorry`
of its own. The remaining gated obligation when this is applied is exactly
`hPieceInv` (BG Theorem E prime-structure of the pieces). -/
theorem theoremII_conjunct1_of_inputs {M K U : Subgroup G}
(hD1 : ∀ x ∈ OddOrder.BG.Ch3.S10.Msigma M, ∀ y ∈ OddOrder.BG.Ch3.S10.Msigma M,
(∃ g : G, y = g * x * g⁻¹) → ∃ m ∈ M, y = m * x * m⁻¹)
(hTI_B : IsTISubset (ASet M U \ (OddOrder.BG.Ch3.S10.Msigma M : Set G)) M)
(hTI_C : IsTISubset (A0Set M K \ ASet M U) M)
{X : Set G} (hX : X = ASet M U ∨ X = A0Set M K)
(hPieceInv : ∀ x ∈ X, ∀ y ∈ X, (∃ g : G, y = g * x * g⁻¹) →
(x ∈ OddOrder.BG.Ch3.S10.Msigma M ↔ y ∈ OddOrder.BG.Ch3.S10.Msigma M) ∧
(x ∈ ASet M U ↔ y ∈ ASet M U)) :
∀ x ∈ X, ∀ y ∈ X, (∃ g : G, y = g * x * g⁻¹) → ∃ m ∈ M, y = m * x * m⁻¹ := by
intro x hxX y hyX hconj
obtain ⟨g, hg⟩ := hconj
obtain ⟨hMσiff, hAiff⟩ := hPieceInv x hxX y hyX ⟨g, hg⟩
by_cases hxMσ : x ∈ OddOrder.BG.Ch3.S10.Msigma M
· -- Both in `M_σ`: Theorem D(1).
exact hD1 x hxMσ y (hMσiff.mp hxMσ) ⟨g, hg⟩
· -- `x ∉ M_σ`, hence `y ∉ M_σ`; `x, y` lie in a common TI piece.
have hyMσ : y ∉ OddOrder.BG.Ch3.S10.Msigma M := fun h => hxMσ (hMσiff.mpr h)
have hxMσ' : x ∉ (OddOrder.BG.Ch3.S10.Msigma M : Set G) := by rwa [SetLike.mem_coe]
have hyMσ' : y ∉ (OddOrder.BG.Ch3.S10.Msigma M : Set G) := by rwa [SetLike.mem_coe]
by_cases hxA : x ∈ ASet M U
· -- `x, y ∈ A(M) - M_σ` (TI piece, Theorem B(5)): the conjugator lies in `M`.
have hyA : y ∈ ASet M U := hAiff.mp hxA
exact ⟨g, hTI_B g ⟨x, ⟨hxA, hxMσ'⟩, hg ▸ ⟨hyA, hyMσ'⟩⟩, hg⟩
· -- `x ∉ A(M)`: only possible for `X = A_0(M)`.
rcases hX with hXA | hXA0
· exact absurd (hXA ▸ hxX) hxA
· -- `x, y ∈ A_0(M) - A(M)` (TI piece, Theorem C(9)).
have hyA : y ∉ ASet M U := fun h => hxA (hAiff.mpr h)
exact ⟨g, hTI_C g ⟨x, ⟨hXA0 ▸ hxX, hxA⟩, hg ▸ ⟨hXA0 ▸ hyX, hyA⟩⟩, hg⟩
/-- **Assembly for BG Theorem II** (mmd L4548), as a *gated-endpoint skeleton* (`sorry`-free in its
own body). `A(M)`/`A_0(M)` are tamely embedded — the BG form of the centralizer-control input used
by Peterfalvi (8.12)--(8.13).
The body is the full Theorem II proof, citing the §16 structure theorems A--D and Proposition 16.1
inline (those were `sorry`-stated when this skeleton was written and have since been proved). What
the skeleton isolates are the two obligations *beyond* that standard A--D suite, as named hypotheses
(cf. `theoremII_conjunct1_of_inputs`):
* `hPieceInv` — the conjunct-1 cross-piece exclusion: `G`-conjugate elements of `X` share `M_σ`-
and `A(M)`-membership (the "distinct orders across pieces" content of BG Theorem E);
* `hMaxUnique` — the conjunct-3 uniqueness `|ℳ(C_G(x))| = 1` for an escaping centralizer
(BG §9--§10 Uniqueness), which pins the Type I/II maximal overgroup of `C_G(x)` to Theorem
D(4)'s `N(x)`.
The wrapper `theoremII_tame_embedding` cites this with both obligations discharged. -/
theorem theoremII_tame_embedding_of_inputs [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))
{X : Set G} (hX : X = ASet M U ∨ X = A0Set M K)
(hPieceInv : ∀ x ∈ X, ∀ y ∈ X, (∃ g : G, y = g * x * g⁻¹) →
(x ∈ OddOrder.BG.Ch3.S10.Msigma M ↔ y ∈ OddOrder.BG.Ch3.S10.Msigma M) ∧
(x ∈ ASet M U ↔ y ∈ ASet M U))
(hMaxUnique : ∀ x : G, x ∈ X → x ≠ 1 →
¬ Subgroup.centralizer ({x} : Set G) ≤ M →
∀ N₁ N₂ : Subgroup G,
N₁ ∈ maximalSubgroupsContaining (Subgroup.centralizer ({x} : Set G)) →
N₂ ∈ maximalSubgroupsContaining (Subgroup.centralizer ({x} : Set G)) → N₁ = N₂) :
(∀ x ∈ X, ∀ y ∈ X,
(∃ g : G, y = g * x * g⁻¹) → ∃ m ∈ M, y = m * x * m⁻¹) ∧
let D : Set G := {x | x ∈ X ∧ x ≠ 1 ∧ ¬ Subgroup.centralizer ({x} : Set G) ≤ M}
-- BG Thm II: `D ⊆ A(M)` (not merely `D ⊆ X`); a genuine claim when `X = A_0(M)`.
D ⊆ ASet M U ∧
∀ x : G, x ∈ D →
∃! N : Subgroup G,
N ∈ maximalSubgroupsContaining (Subgroup.centralizer ({x} : Set G)) ∧
(OddOrder.GroupTheory.IsTypeI N ∨ OddOrder.GroupTheory.IsTypeII N) := by
classical
-- Abbreviate the escaping set `D`.
set D : Set G := {x | x ∈ X ∧ x ≠ 1 ∧ ¬ Subgroup.centralizer ({x} : Set G) ≤ M} with hDdef
-- **Core gated reduction (mmd L4546-L4548).** `A_0(M)` is the disjoint union of `M_σ`,
-- `A(M) - M_σ`, and `A_0(M) - A(M)`; the latter two are `TI`-subsets of `G` with normalizer
-- `M` (Theorem B(5) and Theorem C(9)), so every element of them has its `G`-centralizer inside
-- `M`. Hence an `x ∈ X` with `C_G(x) ⊄ M` must lie in `M_σ`, i.e. `D ⊆ M_σ#`.
--
-- This step needs the Hall data behind Theorem B(5)/C(9), which the *statement* of Theorem II
-- does not carry (its `K`, `U` are free, not pinned to the `(κ ∪ σ)'`-Hall / `κ`-Hall factors).
-- It is therefore isolated as a gated input; once it (and the dual-piece `TI` facts) land with
-- their Hall hypotheses, `D ⊆ A(M)` (conjunct 2) becomes pure citation, as below.
have hDsub : D ⊆ S14.sigmaSharp M := by
intro x hxD
obtain ⟨hxX, hx1, hxc⟩ := hxD
-- `x ∈ M_σ#`: it suffices to show `x ∈ M_σ` (we already have `x ≠ 1`).
simp only [S14.sigmaSharp, sharpSubgroup, Set.mem_sdiff, SetLike.mem_coe,
Set.mem_singleton_iff]
refine ⟨?_, hx1⟩
by_contra hxnMσ
-- `x ∉ M_σ`; the coerced form, and the TI piece for `A(M) - M_σ` (Theorem B(5)).
have hxnMσ' : x ∉ (OddOrder.BG.Ch3.S10.Msigma M : Set G) := by rwa [SetLike.mem_coe]
have hTIB : IsTISubset (ASet M U \ (OddOrder.BG.Ch3.S10.Msigma M : Set G)) M :=
theoremB_A_minus_Msigma_isTISubset hG hM hKM hUM hK hU
rcases hX with hXA | hXA0
· -- `X = A(M)`: `x ∈ A(M) - M_σ`, so `C_G(x) ≤ M` (Theorem B(5)) contradicts `C_G(x) ⊄ M`.
exact hxc (hTIB.centralizer_le ⟨hXA ▸ hxX, hxnMσ'⟩)
· -- `X = A_0(M)`.
have hxA0 : x ∈ A0Set M K := hXA0 ▸ hxX
by_cases hxA : x ∈ ASet M U
· -- `x ∈ A(M) - M_σ`: Theorem B(5) again.
exact hxc (hTIB.centralizer_le ⟨hxA, hxnMσ'⟩)
· -- `x ∈ A_0(M) - A(M)`.
by_cases hKbot : K = ⊥
· -- **Type-F** (`K = ⊥`): `A_0(M) = \widehat{M_σ} ⊆ M = U M_σ` (Theorem A(3)),
-- so `x ∈ A(M)`, contradicting `x ∉ A(M)`.
refine hxA ⟨hxA0.1, ?_⟩
have hxM : x ∈ M := hxA0.1.1
have hA3 : M = K ⊔ U ⊔ OddOrder.BG.Ch3.S10.Msigma M :=
(theoremA_maximal_structure_faithful hG hM hKM hUM hK rfl hU).2.2.1
have hx' : x ∈ (K ⊔ U ⊔ OddOrder.BG.Ch3.S10.Msigma M : Subgroup G) := hA3 ▸ hxM
rw [hKbot, bot_sup_eq] at hx'
exact hx'
· -- `K ≠ ⊥`: TI by Theorem C(9), giving `C_G(x) ≤ M`.
obtain ⟨_, _, _, _, _, _, _, _, _, hTIC, _, _⟩ :=
theoremC_paired_structure hG hM hKbot hKM hUM hK rfl hU
exact hxc (hTIC.centralizer_le ⟨hxA0, hxA⟩)
-- **`D ⊆ A(M)` (conjunct 2).** `D ⊆ M_σ#` and `M_σ# ⊆ A(M)`: a nonidentity `x ∈ M_σ` lies in
-- `\widehat{M_σ}` (`sigmaSharp_subset_hatMsigma`) and in `M_σ ≤ U M_σ`, so `x ∈ A(M)`.
have hMσsharp_sub_A : S14.sigmaSharp M ⊆ ASet M U := by
intro x hx
refine ⟨sigmaSharp_subset_hatMsigma M hx, ?_⟩
have hxMσ : x ∈ OddOrder.BG.Ch3.S10.Msigma M := (Set.mem_sdiff _).mp hx |>.1
exact (le_sup_right : OddOrder.BG.Ch3.S10.Msigma M ≤
(U ⊔ OddOrder.BG.Ch3.S10.Msigma M : Subgroup G)) hxMσ
refine ⟨?_, hDsub.trans hMσsharp_sub_A, ?_⟩
· -- **Conjunct 1 (Ti) (mmd L4546-L4550).** Assembled from Theorem D(1) (`M_σ` fusion),
-- Theorem B(5)/C(9) (the two TI pieces), and the cross-piece exclusion `hPieceInv`, via
-- `theoremII_conjunct1_of_inputs`. Only `hPieceInv` remains gated (BG Theorem E).
have hTI_C : IsTISubset (A0Set M K \ ASet M U) M := by
by_cases hKbot : K = ⊥
· -- `K = ⊥` (type F): `A_0(M) = \widehat{M_σ} ⊆ M = U M_σ` (Thm A(3)), so the diff is empty.
intro g hex
obtain ⟨z, ⟨hzA0, hznA⟩, _⟩ := hex
refine absurd ⟨hzA0.1, ?_⟩ hznA
have hA3 : M = K ⊔ U ⊔ OddOrder.BG.Ch3.S10.Msigma M :=
(theoremA_maximal_structure_faithful hG hM hKM hUM hK rfl hU).2.2.1
have hz' : z ∈ (K ⊔ U ⊔ OddOrder.BG.Ch3.S10.Msigma M : Subgroup G) := hA3 ▸ hzA0.1.1
rw [hKbot, bot_sup_eq] at hz'
exact hz'
· obtain ⟨_, _, _, _, _, _, _, _, _, hTIC, _, _⟩ :=
theoremC_paired_structure hG hM hKbot hKM hUM hK rfl hU
exact hTIC
refine theoremII_conjunct1_of_inputs
(theoremD_msigma_conjugacy_and_centralizers hG hM).1
(theoremB_A_minus_Msigma_isTISubset hG hM hKM hUM hK hU) hTI_C hX ?_
-- `hPieceInv`: `G`-conjugate elements of `X` share `M_σ`- and `A(M)`-membership — the
-- "distinct orders across pieces" input of the mmd proof (BG Theorem E), a named obligation.
exact hPieceInv
· -- **Conjunct 3 (mmd L4552).** For `x ∈ D ⊆ M_σ#` with `C_G(x) ⊄ M`, Theorem D(4) gives a
-- unique maximal `N(x) ⊇ C_G(x)` that is of type `F` or `P₂`; Proposition 16.1(a)(b) rewrites
-- this as Type I or Type II. Existence and the type classification are pure citation; the
-- *uniqueness* of the maximal overgroup is the residual gated input (BG §9-§10 Uniqueness).
intro x hxD
obtain ⟨hxX, hx1, hxc⟩ := hxD
have hxMσsharp : x ∈ S14.sigmaSharp M := hDsub ⟨hxX, hx1, hxc⟩
-- Theorem D(4): the `∃! N` with the type-`F`/`P₂` data attached to escaping centralizers.
obtain ⟨_, _, _, hD4⟩ := theoremD_msigma_conjugacy_and_centralizers hG hM
obtain ⟨_R, _hR, N₀, hQN₀, _hQuniq⟩ := hD4 x hxMσsharp hxc
-- Unpack what Theorem II needs from the rich Theorem D(4) predicate `Q N₀`.
obtain ⟨hN₀mem, _, _, _, hN₀type, _⟩ := hQN₀
-- Convert `IsTypeF N₀ ∨ IsTypeP2 N₀` to `IsTypeI N₀ ∨ IsTypeII N₀` (Proposition 16.1(a)(b)).
have hN₀ : N₀ ∈ maximalSubgroups G := (mem_maximalSubgroupsContaining.mp hN₀mem).1
have htype : OddOrder.GroupTheory.IsTypeI N₀ ∨ OddOrder.GroupTheory.IsTypeII N₀ := by
obtain ⟨hIiff, hIIiff, _⟩ := proposition_type_classification hG hN₀
rcases hN₀type with hF | hP2
· exact Or.inl (hIiff.mpr hF)
· exact Or.inr (hIIiff.mpr hP2)
refine ⟨N₀, ⟨hN₀mem, htype⟩, ?_⟩
-- Uniqueness of the maximal overgroup of `C_G(x)`: the named obligation `hMaxUnique`. Theorem
-- D(4) gives uniqueness only for its *full* predicate `Q`; pinning the weaker "maximal
-- overgroup, Type I/II" to the same `N₀` is exactly `|ℳ(C_G(x))| = 1` (BG §9--§10 Uniqueness).
rintro N' ⟨hN'mem, _hN'type⟩
exact hMaxUnique x hxX hx1 hxc N' N₀ hN'mem hN₀mem
/-- **The escaping piece of `A(M)`/`A_0(M)` lands in `M_σ#`** (the `D ⊆ M_σ#` reduction of Theorem
II's conjunct 2, extracted for reuse). An `x ∈ X` (`X = A(M)` or `A_0(M)`) with `x ≠ 1` and
`C_G(x) ⊄ M` must lie in `M_σ`: the dual TI pieces `A(M) - M_σ` (Theorem B(5)) and `A_0(M) - A(M)`
(Theorem C(9), or empty in the type-`F` case via Theorem A(3)) have `G`-centralizer inside `M`. -/
theorem mem_sigmaSharp_of_mem_aSet_of_escape [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))
{X : Set G} (hX : X = ASet M U ∨ X = A0Set M K)
{x : G} (hxX : x ∈ X) (hx1 : x ≠ 1)
(hesc : ¬ Subgroup.centralizer ({x} : Set G) ≤ M) :
x ∈ S14.sigmaSharp M := by
simp only [S14.sigmaSharp, sharpSubgroup, Set.mem_sdiff, SetLike.mem_coe, Set.mem_singleton_iff]
refine ⟨?_, hx1⟩
by_contra hxnMσ
have hxnMσ' : x ∉ (OddOrder.BG.Ch3.S10.Msigma M : Set G) := by rwa [SetLike.mem_coe]
have hTIB : IsTISubset (ASet M U \ (OddOrder.BG.Ch3.S10.Msigma M : Set G)) M :=
theoremB_A_minus_Msigma_isTISubset hG hM hKM hUM hK hU
rcases hX with hXA | hXA0
· exact hesc (hTIB.centralizer_le ⟨hXA ▸ hxX, hxnMσ'⟩)
· have hxA0 : x ∈ A0Set M K := hXA0 ▸ hxX
by_cases hxA : x ∈ ASet M U
· exact hesc (hTIB.centralizer_le ⟨hxA, hxnMσ'⟩)
· by_cases hKbot : K = ⊥
· refine hxA ⟨hxA0.1, ?_⟩
have hxM : x ∈ M := hxA0.1.1
have hA3 : M = K ⊔ U ⊔ OddOrder.BG.Ch3.S10.Msigma M :=
(theoremA_maximal_structure_faithful hG hM hKM hUM hK rfl hU).2.2.1
have hx' : x ∈ (K ⊔ U ⊔ OddOrder.BG.Ch3.S10.Msigma M : Subgroup G) := hA3 ▸ hxM
rw [hKbot, bot_sup_eq] at hx'
exact hx'
· obtain ⟨_, _, _, _, _, _, _, _, _, hTIC, _, _⟩ :=
theoremC_paired_structure hG hM hKbot hKM hUM hK rfl hU
exact hesc (hTIC.centralizer_le ⟨hxA0, hxA⟩)
/-- **`ℳ(C_G(x))` is a singleton for an escaping `σ`-sharp element** (`|ℳ(C_G(x))| = 1`, the BG
§9--§10 uniqueness input of Theorem II's conjunct 3). For `x ∈ M_σ#` with `C_G(x) ⊄ M`, the escape
forces `|𝓜_σ(x)| > 1` (`centralizer_le_of_maximalSigma_le_one`), so the signalizer structure
(`signalizer_structure_of_mem_sigmaSharp`) supplies the type-`F`/`P₂` neighbour `N` over `C_G(x)`
whose `τ₂`-element data feeds the Corollary-14.3 uniqueness
`maximalContaining_centralizer_eq_singleton_of_tau2_element`, pinning `ℳ(C_G(x)) = {N}`. -/
theorem maximalSubgroupsContaining_centralizer_eq_singleton_of_sigmaSharp_escape [Finite G]
(hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G} (hM : M ∈ maximalSubgroups G)
{x : G} (hx : x ∈ S14.sigmaSharp M) (hesc : ¬ Subgroup.centralizer ({x} : Set G) ≤ M) :
∃ N : Subgroup G,
maximalSubgroupsContaining (Subgroup.centralizer ({x} : Set G)) = {N} := by
have hx1 : x ≠ 1 := hx.2
have hxMσ : x ∈ OddOrder.BG.Ch3.S10.Msigma M := hx.1
have hgt : 1 < (S14.maximalSigmaSubgroupsOfElement x).ncard := by
by_contra h
push Not at h
exact hesc (centralizer_le_of_maximalSigma_le_one hG hM hxMσ hx1 h)
obtain ⟨N, hNstruct, -⟩ := signalizer_structure_of_mem_sigmaSharp hG hM hx hgt
obtain ⟨hNmax, hCN, hRne, _hRhall, hxtau2, _hNtype, _hforall⟩ := hNstruct
have hxN : x ∈ N := hCN (Subgroup.mem_centralizer_iff.mpr
(fun y hy => by rw [Set.mem_singleton_iff.mp hy]))
exact ⟨N, maximalContaining_centralizer_eq_singleton_of_tau2_element hG hNmax hxN hx1 hxtau2 hRne⟩
/-- **BG Theorem II** (mmd L4548): `A(M)` and `A_0(M)` are tamely embedded. The BG form of the
centralizer-control input used by Peterfalvi (8.12)--(8.13). Cites the gated-endpoint skeleton
`theoremII_tame_embedding_of_inputs`; **both residual obligations are now discharged** — the BG
Theorem E cross-piece exclusion `hPieceInv` via the order-determined `M_σ`/`A(M)`-membership
(`mem_Msigma_iff_isPiElement_sigma` / `mem_U_sup_Msigma_iff_isPiElement_kappa_compl`), and the BG
§9--§10 maximal-overgroup uniqueness `hMaxUnique` via the signalizer uniqueness
(`maximalSubgroupsContaining_centralizer_eq_singleton_of_sigmaSharp_escape`). The `_of_inputs`
skeleton cites the §16 structure theorems A--D inline; those are proved. -/
theorem theoremII_tame_embedding [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))
{X : Set G} (hX : X = ASet M U ∨ X = A0Set M K) :
(∀ x ∈ X, ∀ y ∈ X,
(∃ g : G, y = g * x * g⁻¹) → ∃ m ∈ M, y = m * x * m⁻¹) ∧
let D : Set G := {x | x ∈ X ∧ x ≠ 1 ∧ ¬ Subgroup.centralizer ({x} : Set G) ≤ M}
-- BG Thm II: `D ⊆ A(M)` (not merely `D ⊆ X`); a genuine claim when `X = A_0(M)`.
D ⊆ ASet M U ∧
∀ x : G, x ∈ D →
∃! N : Subgroup G,
N ∈ maximalSubgroupsContaining (Subgroup.centralizer ({x} : Set G)) ∧
(OddOrder.GroupTheory.IsTypeI N ∨ OddOrder.GroupTheory.IsTypeII N) :=
theoremII_tame_embedding_of_inputs hG hM hKM hUM hK hU hX
-- `hPieceInv`: BG Theorem E cross-piece exclusion. `M_σ` (normal `σ`-Hall) and `U⊔M_σ` (normal
-- `κ′`-Hall, Theorem A(3)) make `M_σ`- and `A(M)`-membership of an element of `M`
-- order-determined
-- (`mem_Msigma_iff_isPiElement_sigma` / `mem_U_sup_Msigma_iff_isPiElement_kappa_compl`), hence
-- conjugation-invariant (`isPiElement_conj`).
(by
have hMσM : OddOrder.BG.Ch3.S10.Msigma M ≤ M := OddOrder.BG.Ch3.S10.Msigma_le M
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
intro x hxX y hyX hconj
obtain ⟨g, hg⟩ := hconj
have hxhat : x ∈ hatMsigma M := by
rcases hX with h | h
· have hm := h ▸ hxX; simp only [ASet, Set.mem_inter_iff] at hm; exact hm.1
· have hm := h ▸ hxX; simp only [A0Set, Set.mem_sdiff] at hm; exact hm.1
have hyhat : y ∈ hatMsigma M := by
rcases hX with h | h
· have hm := h ▸ hyX; simp only [ASet, Set.mem_inter_iff] at hm; exact hm.1
· have hm := h ▸ hyX; simp only [A0Set, Set.mem_sdiff] at hm; exact hm.1
refine ⟨?_, ?_⟩
· rw [S14.mem_Msigma_iff_isPiElement_sigma hG hM hxhat.1,
S14.mem_Msigma_iff_isPiElement_sigma hG hM hyhat.1]
refine ⟨fun h => ?_, fun h => ?_⟩
· rw [hg]; exact S14.isPiElement_conj g h
· rw [show x = g⁻¹ * y * (g⁻¹)⁻¹ from by rw [hg]; group]
exact S14.isPiElement_conj g⁻¹ h
· have hAx : x ∈ ASet M U ↔ x ∈ U ⊔ OddOrder.BG.Ch3.S10.Msigma M := by
simp only [ASet, Set.mem_inter_iff, SetLike.mem_coe]
exact ⟨fun h => h.2, fun h => ⟨hxhat, h⟩⟩
have hAy : y ∈ ASet M U ↔ y ∈ U ⊔ OddOrder.BG.Ch3.S10.Msigma M := by
simp only [ASet, Set.mem_inter_iff, SetLike.mem_coe]
exact ⟨fun h => h.2, fun h => ⟨hyhat, h⟩⟩
rw [hAx, hAy,
S14.mem_U_sup_Msigma_iff_isPiElement_kappa_compl hG hM hUM hU hnorm hxhat.1,
S14.mem_U_sup_Msigma_iff_isPiElement_kappa_compl hG hM hUM hU hnorm hyhat.1]
refine ⟨fun h => ?_, fun h => ?_⟩
· rw [hg]; exact S14.isPiElement_conj g h
· rw [show x = g⁻¹ * y * (g⁻¹)⁻¹ from by rw [hg]; group]
exact S14.isPiElement_conj g⁻¹ h)
-- `hMaxUnique`: BG §9--§10 maximal-overgroup uniqueness `|ℳ(C_G(x))| = 1`, discharged from the
-- signalizer uniqueness
-- (`maximalSubgroupsContaining_centralizer_eq_singleton_of_sigmaSharp_escape`)
-- via the `D ⊆ M_σ#` reduction (`mem_sigmaSharp_of_mem_aSet_of_escape`).
(fun x hxX hx1 hesc N₁ N₂ hN₁ hN₂ => by
obtain ⟨N, hMC⟩ := maximalSubgroupsContaining_centralizer_eq_singleton_of_sigmaSharp_escape
hG hM (mem_sigmaSharp_of_mem_aSet_of_escape hG hM hKM hUM hK hU hX hxX hx1 hesc) hesc
rw [hMC, Set.mem_singleton_iff] at hN₁ hN₂
exact hN₁.trans hN₂.symm)
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/TaxonomyOutput.lean