Bender–Glauberman Appendix C: hypotheses (A) and (B) of Theorem C
DefinitionOddOrder_BG_AppC_HypothesisBfeit-thompsonfinite-groupsgroup-theoryishida-odd-order
Formalizes the hypotheses of Theorem C in Bender–Glauberman Appendix C, §1 (p. 145). Let be primes satisfying condition (A), ; let be the additive group of and the norm-one subgroup of , acting on by multiplication, giving the semidirect product with prime-field line .
Main definitions: conditionA (the cyclotomic condition), primeLine (), primeLineElement and primeLineGenerator (the elements of for a scalar and its distinguished generator), HypothesisBAbstract (Hypothesis (B), in the book's abstract form) and FieldNormalizerData (Hypothesis (B) together with condition (A)).
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/AppC_HypothesisB.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_NormSet
namespace OddOrder.BG.AppC
open scoped Pointwise
/-- **BG Appendix C, Condition (A)**: the cyclotomic factor attached to
`F_{p^q}` is coprime to `p - 1`.
By Remark (I) this is equivalent to `q ∤ p - 1` (`NormSet.conditionA_iff_not_dvd`). -/
def conditionA (p q : ℕ) : Prop :=
Nat.Coprime ((p ^ q - 1) / (p - 1)) (p - 1)
/-- **BG Appendix C**, the prime-field line `P₀ ≤ P`: the image in `P = 𝔽_{p^q}` of the additive
group of `𝔽_p`, i.e. the `𝔽_p`-span of `1`, viewed inside `H = P ⋊ U`. -/
noncomputable def primeLine (p q : ℕ) [Fact p.Prime] :
Subgroup (NormSet.normOneFrobeniusGroup p q) :=
NormSet.normOneFrobeniusSubspaceKernel p q
(Submodule.span (ZMod p) ({(1 : GaloisField p q)} : Set (GaloisField p q)))
/-- The `𝔽_p`-span of `1` in `𝔽_{p^q}` is the prime field, of cardinality `p`. -/
theorem card_span_singleton_one (p q : ℕ) [Fact p.Prime] :
Nat.card (Submodule.span (ZMod p) ({(1 : GaloisField p q)} : Set (GaloisField p q))) = p := by
have hinj : Function.Injective
(LinearMap.toSpanSingleton (ZMod p) (GaloisField p q) 1) := by
intro a b hab
have : (algebraMap (ZMod p) (GaloisField p q)) a =
(algebraMap (ZMod p) (GaloisField p q)) b := by
simpa [LinearMap.toSpanSingleton_apply, Algebra.smul_def] using hab
exact FaithfulSMul.algebraMap_injective (ZMod p) (GaloisField p q) this
rw [LinearMap.span_singleton_eq_range]
have : Nat.card (LinearMap.range (LinearMap.toSpanSingleton (ZMod p) (GaloisField p q) 1)) =
Nat.card (ZMod p) := (Nat.card_congr (Equiv.ofInjective _ hinj)).symm
rw [this, Nat.card_eq_fintype_card, ZMod.card]
/-- **`|P₀| = p`**: the prime-field line has order `p`. This is what the Section 16 development
used to take from its own hypothesis (`p = |W₂|`); it is in fact a consequence of the setup, since
`P₀` is the image of the prime field `𝔽_p` inside `P = 𝔽_{p^q}`. -/
theorem card_primeLine (p q : ℕ) [Fact p.Prime] : Nat.card (primeLine p q) = p := by
rw [primeLine, NormSet.normOneFrobeniusSubspaceKernel,
Subgroup.card_map_of_injective SemidirectProduct.inl_injective]
exact card_span_singleton_one p q
/-- The element of the prime-field line `P₀ ≤ P ≤ H` corresponding to a scalar `c : ZMod p`,
i.e. `algebraMap c ∈ 𝔽_p ⊆ 𝔽_{p^q}` read inside the additive kernel. -/
noncomputable def primeLineElement (p q : ℕ) [Fact p.Prime] (c : ZMod p) :
NormSet.normOneFrobeniusGroup p q :=
SemidirectProduct.inl (Multiplicative.ofAdd (algebraMap (ZMod p) (GaloisField p q) c))
/-- The distinguished nonidentity element of the prime-field line `P₀`, corresponding to
`1 : 𝔽_{p^q}` in BG Appendix C. -/
noncomputable def primeLineGenerator (p q : ℕ) [Fact p.Prime] :
NormSet.normOneFrobeniusGroup p q :=
SemidirectProduct.inl (Multiplicative.ofAdd (1 : GaloisField p q))
/-- **BG Appendix C, Hypothesis (B)** (p. 145), in the book's `p, q, G`-abstract form: no
Peterfalvi Section 16 configuration anywhere in the statement.
Verbatim: *"There is a monomorphism `σ : H → G`, a finite abelian `p'`-subgroup `Q` of `G`, and an
element `y ∈ Q` such that `σ(P₀)` normalizes `Q` and `σ(P₀)^y` normalizes `U`."* Here
`H = P ⋊ U` is `NormSet.normOneFrobeniusGroup p q`, `P₀` is `primeLine p q`, and `U` is
identified with its image `σ(U)` inside `G` (Remark (VI)).
The conjugation is written `MulAut.conj y • σ(P₀) = y σ(P₀) y⁻¹`; since `Q` is a subgroup,
feeding it `y⁻¹` recovers the book's `σ(P₀)^y = y⁻¹ σ(P₀) y` verbatim, so the two readings define
the same hypothesis.
The concrete Peterfalvi Section 16 instance of (B) along the Feit--Thompson spine is
`OddOrder.Peterfalvi.S16.FieldNormalizerData`. What is still missing to run Theorem C off the
abstract form is the implication *(B) ⟹ `hrel`* (BG Lemma C.3), currently available only through
Section 16; see issue 0151.
`hypothesisBAbstract_sl2` (in `OddOrder.BG.AppC_SL2Example`) is the book's Remark (II) witness,
showing this hypothesis is satisfiable (`p = 2`, `G = SL(2, 2^q)`). -/
structure HypothesisBAbstract (p q : ℕ) [Fact p.Prime] (G : Type*) [Group G] where
/-- The monomorphism `σ : H = P ⋊ U → G`. -/
sigma : NormSet.normOneFrobeniusGroup p q →* G
/-- `σ` is a monomorphism. -/
sigma_injective : Function.Injective sigma
/-- The finite abelian `p'`-subgroup `Q ≤ G`. -/
Q : Subgroup G
/-- `Q` is finite. -/
Q_finite : Finite Q
/-- `Q` is abelian. -/
Q_commutative : IsMulCommutative Q
/-- `Q` is a `p'`-group. -/
Q_pPrime : ¬ p ∣ Nat.card Q
/-- The distinguished element `y ∈ Q`. -/
y : G
/-- `y` lies in `Q`. -/
y_mem_Q : y ∈ Q
/-- `σ(P₀)` normalizes `Q`. -/
primeLine_normalizes_Q :
(primeLine p q).map sigma ≤ Subgroup.normalizer (Q : Set G)
/-- `σ(P₀)^y` normalizes `σ(U)`. -/
primeLine_conj_normalizes_U :
MulAut.conj y • ((primeLine p q).map sigma) ≤
Subgroup.normalizer
(((NormSet.normOneFrobeniusComplement p q).map sigma : Subgroup G) : Set G)
section Data
variable {p q : ℕ} [Fact p.Prime] {G : Type*} [Group G]
/-- **Field-normalizer data**: hypothesis (B) together with condition (A), with the three
subgroups `σ(P)`, `σ(U)`, `σ(P₀)` given names.
This is the shape in which the Lemma C.1--C.3 development consumes Theorem C's hypotheses: the
proofs constantly refer to the ambient subgroups `P = σ(P)`, `U = σ(U)` and `W₂ = σ(P₀)` of `G`,
so it is convenient to carry them as fields with their defining equations rather than to unfold
`Subgroup.map` everywhere.
The packaging adds **nothing**: `HypothesisBAbstract.toFieldNormalizerData` builds it from (B)
and (A) alone, taking the three subgroups to be literally the images (all three equations are
`rfl`). So a `FieldNormalizerData` is exactly "(A) and (B)". -/
structure FieldNormalizerData (p q : ℕ) [Fact p.Prime] (G : Type*) [Group G]
extends HypothesisBAbstract p q G where
/-- The image `P = σ(P)` of the additive kernel. -/
P : Subgroup G
/-- The image `U = σ(U)` of the norm-one complement. -/
U : Subgroup G
/-- The image `W₂ = σ(P₀)` of the prime-field line. -/
W2 : Subgroup G
/-- `P` is the image of the additive kernel. -/
sigma_P_eq_P : (NormSet.normOneFrobeniusKernel p q).map sigma = P
/-- `U` is the image of the norm-one complement. -/
sigma_U_eq_U : (NormSet.normOneFrobeniusComplement p q).map sigma = U
/-- `W₂` is the image of the prime-field line. -/
sigma_P0_eq_W2 : (primeLine p q).map sigma = W2
/-- `q` is prime (the standing hypothesis of Theorem C). -/
q_prime : q.Prime
/-- `p` is odd. By **Remark (V)** ("by (A), we can assume `p` and `q` are odd") this costs
nothing: `le_of_conditionA_of_not_odd` disposes of the even case outright. -/
p_odd : Odd p
/-- Condition (A). -/
cyclotomic_coprime : conditionA p q
namespace FieldNormalizerData
/-- `W₂ = σ(P₀)` normalizes `Q`: clause (B) read through `sigma_P0_eq_W2`. -/
theorem W2_normalizes_Q (data : FieldNormalizerData p q G) :
data.W2 ≤ Subgroup.normalizer (data.Q : Set G) := by
rw [← data.sigma_P0_eq_W2]
exact data.primeLine_normalizes_Q
/-- **`|W₂| = p`**: the transported prime line has order `p`, because `W₂ = σ(P₀)`, `σ` is
injective and `|P₀| = p` (`card_primeLine`). The Section 16 development previously took this
from its own hypothesis field `p = |W₂|`; it is a consequence of (B). -/
theorem card_W2 (data : FieldNormalizerData p q G) : Nat.card data.W2 = p := by
rw [← data.sigma_P0_eq_W2, Subgroup.card_map_of_injective data.sigma_injective]
exact card_primeLine p q
/-- `W₂^y` normalizes `U`: clause (B) read through `sigma_P0_eq_W2` and `sigma_U_eq_U`. -/
theorem W2_conj_y_normalizes_U (data : FieldNormalizerData p q G) :
MulAut.conj data.y • data.W2 ≤ Subgroup.normalizer (data.U : Set G) := by
rw [← data.sigma_P0_eq_W2, ← data.sigma_U_eq_U]
exact data.primeLine_conj_normalizes_U
end FieldNormalizerData
end Data
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_HypothesisB.lean