Bender–Glauberman Appendix C, Lemma C.3: the transported configuration
DefinitionOddOrder_BG_AppC_LemmaC3_Setupfeit-thompsonfinite-groupsgroup-theoryishida-odd-order
Sets up the configuration in Bender–Glauberman Appendix C, Lemma C.3 (§3, pp. 148–152). Under hypothesis (B) the monomorphism transports the concrete Frobenius group into . The file develops the resulting configuration: the element , the isomorphism , the decomposition with , that is characteristic in , and the irreducibility bridge.
Main definitions: normSetGeneratorRelation (the norm relation from the generator relation), normSetTwistedNormOneStep (Step 4 narrowed to the norm-one units), s (the BG element transported to ), and finiteness of the transported prime line of order and of (finite_W2, finite_Q, finite_W2_sup_Q).
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/AppC_LemmaC3_Setup.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_AppC_FrobeniusBasics
namespace OddOrder.BG.AppC
open scoped Pointwise
/-- **BG Appendix C, Lemma C.3**: the norm relation produced by the generator-relation argument.
For every `a` in the norm set `E`, `N(2a − 1) = 1`; BG then converts this finite-field statement
into `a⁻¹ ∈ E`. -/
def normSetGeneratorRelation (p q : ℕ) [Fact p.Prime] : Prop :=
∀ a : GaloisField p q, a ∈ NormSet.normSetE p q →
NormSet.normN p q ((2 : GaloisField p q) * a - 1) = 1
/-- **BG Appendix C, Lemma C.3, Step 4**, narrowed to the norm-one unit group `U`. This matches
the actual BG Step 4 action by conjugation with `t`; no extension to all of `𝔽_{p^q}ˣ` is
required. -/
def normSetTwistedNormOneStep (p q : ℕ) [Fact p.Prime] : Prop :=
∃ φ : MulAut (NormSet.normOneUnits p q),
φ ^ p = 1 ∧ NormSet.normSetETwistedNormOneStep p q φ
variable {p q : ℕ} [Fact p.Prime] {G : Type*} [Group G]
namespace FieldNormalizerData
/-- The BG element `s ∈ P₀#`, transported to `G` through the concrete
field-normalizer monomorphism. -/
noncomputable def s (data : FieldNormalizerData p q G) : G :=
data.sigma (primeLineGenerator p q)
/-- The transported element `s` lies in Peterfalvi's subgroup `W₂ = σ(P₀)`. -/
theorem s_mem_W2 (data : FieldNormalizerData p q G) :
data.s ∈ data.W2 := by
rw [← data.sigma_P0_eq_W2]
exact ⟨primeLineGenerator p q,
primeLineGenerator_mem p q, rfl⟩
/-- The transported element `s` is nontrivial. -/
theorem s_ne_one (data : FieldNormalizerData p q G) :
data.s ≠ 1 := by
intro hs
exact primeLineGenerator_ne_one p q
(data.sigma_injective (by simpa [s] using hs))
/-- The transported prime-line generator has `p`-th power equal to `1`. -/
theorem s_pow_p_eq_one (data : FieldNormalizerData p q G) :
data.s ^ p = 1 := by
simpa [s] using congrArg data.sigma (primeLineGenerator_pow_p p q)
/-- The transported generator `s` has exact order `p`. -/
theorem s_orderOf_eq_p (data : FieldNormalizerData p q G) :
orderOf data.s = p := by
have hdiv : orderOf data.s ∣ p :=
orderOf_dvd_of_pow_eq_one data.s_pow_p_eq_one
rcases (Fact.out : Nat.Prime p).eq_one_or_self_of_dvd (orderOf data.s) hdiv with h | h
· exact False.elim (data.s_ne_one (orderOf_eq_one_iff.mp h))
· exact h
/-- The transported prime line `W₂ = σ(P₀)` is finite, of order `p`. Together with the
hypothesis (B) field `Q_finite` this is all the finiteness the Lemma C.3 development needs --
in particular the ambient group `G` is never assumed finite. -/
instance finite_W2 (data : FieldNormalizerData p q G) : Finite data.W2 :=
Nat.finite_of_card_ne_zero (by rw [data.card_W2]; exact (Fact.out : Nat.Prime p).ne_zero)
/-- Hypothesis (B)'s `Q` is finite, as an instance. -/
instance finite_Q (data : FieldNormalizerData p q G) : Finite data.Q := data.Q_finite
/-- The underlying set of `W₂` is finite. -/
theorem setFinite_W2 (data : FieldNormalizerData p q G) : (↑data.W2 : Set G).Finite :=
Set.finite_coe_iff.mp (finite_W2 data)
/-- The underlying set of `Q` is finite (hypothesis (B)). -/
theorem setFinite_Q (data : FieldNormalizerData p q G) : (↑data.Q : Set G).Finite :=
Set.finite_coe_iff.mp (finite_Q data)
/-- `W₂Q` is finite: `W₂` normalizes `Q` by (B), so the join is the product set `W₂ · Q` of two
finite sets. This is what lets the Step 3 Sylow argument run without assuming `G` finite. -/
theorem setFinite_W2_sup_Q (data : FieldNormalizerData p q G) :
((data.W2 ⊔ data.Q : Subgroup G) : Set G).Finite := by
rw [Subgroup.coe_mul_of_left_le_normalizer_right data.W2 data.Q data.W2_normalizes_Q]
exact data.setFinite_W2.mul data.setFinite_Q
instance finite_W2_sup_Q (data : FieldNormalizerData p q G) :
Finite ↥(data.W2 ⊔ data.Q) := data.setFinite_W2_sup_Q.to_subtype
/-- The transported prime line `W₂ = σ(P₀)` is generated by the distinguished
nonidentity element `s`. This is the `P₀ = ⟨s⟩` input in BG Appendix C. -/
theorem W2_eq_zpowers_s (data : FieldNormalizerData p q G) :
data.W2 = Subgroup.zpowers data.s := by
have hsle : Subgroup.zpowers data.s ≤ data.W2 :=
Subgroup.zpowers_le.mpr data.s_mem_W2
have hcard : Nat.card ↥data.W2 ≤ Nat.card ↥(Subgroup.zpowers data.s) := by
rw [Nat.card_zpowers, data.s_orderOf_eq_p, data.card_W2]
exact (Subgroup.eq_of_le_of_card_ge hsle hcard).symm
/-- The concrete norm-one complement transported through `σ` onto Peterfalvi's
subgroup `U`. -/
noncomputable def normOneUnitsToU (data : FieldNormalizerData p q G) :
NormSet.normOneUnits p q →* data.U :=
{ toFun := fun u =>
⟨data.sigma (SemidirectProduct.inr u), by
rw [← data.sigma_U_eq_U]
exact ⟨SemidirectProduct.inr u, ⟨u, rfl⟩, rfl⟩⟩
map_one' := by
ext
simp
map_mul' := by
intro u v
ext
simp }
/-- The transported norm-one complement map is injective. -/
theorem normOneUnitsToU_injective (data : FieldNormalizerData p q G) :
Function.Injective data.normOneUnitsToU := by
intro u v h
have hsig : data.sigma (SemidirectProduct.inr u) = data.sigma (SemidirectProduct.inr v) :=
congrArg Subtype.val h
exact SemidirectProduct.inr_injective (data.sigma_injective hsig)
/-- The transported norm-one complement map is onto Peterfalvi's `U`. -/
theorem normOneUnitsToU_surjective (data : FieldNormalizerData p q G) :
Function.Surjective data.normOneUnitsToU := by
intro u
have hu : (u : G) ∈ (NormSet.normOneFrobeniusComplement p q).map data.sigma := by
rw [data.sigma_U_eq_U]
exact u.property
rcases hu with ⟨x, hxU, hx⟩
rcases hxU with ⟨u0, rfl⟩
refine ⟨u0, ?_⟩
ext
exact hx
/-- The concrete norm-one unit group is isomorphic to Peterfalvi's subgroup `U`
through the field-normalizer embedding. -/
noncomputable def normOneUnitsEquivU (data : FieldNormalizerData p q G) :
NormSet.normOneUnits p q ≃* data.U :=
MulEquiv.ofBijective data.normOneUnitsToU
⟨data.normOneUnitsToU_injective, data.normOneUnitsToU_surjective⟩
end FieldNormalizerData
end OddOrder.BG.AppC
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/AppC_LemmaC3_Setup.lean