Peterfalvi (5.6) / Isaacs 7.14: the general mixed-degree coherence adjoin
DefinitionOddOrder_Peterfalvi_S08_GeneralAdjoinfeit-thompsonfinite-groupsgroup-theoryishida-odd-order
Peterfalvi, Character Theory for the Odd Order Theorem, §5, (5.6), whose underlying theorem is Isaacs, Character Theory of Finite Groups, Theorem 7.14 (mathcomp extend_coherent). The Feit–Thompson main text consumes the mixed-degree adjoin in a general form.
Main definitions: memberExtensionDecomposition_general (general member decomposition) and decompositionDaFromDiff_general (general decomposition for ), generalizing the earlier versions.
The bundle also proves 2 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/Peterfalvi/S08_GeneralAdjoin.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_Peterfalvi_S07_BridgeCoherent
import Definitions.Def_OddOrder_Peterfalvi_S08_CoherenceCorePart1_CoherentAdjoin
namespace OddOrder.Peterfalvi.S07
open OddOrder.RepresentationTheory
variable {L G : Type*} [Group L] [Group G]
/-- **`ℤ[T] ⊆ ℤ[S, A]` when `T ⊆ ℤ[S, A]`.** The `A`-supported lattice `zSupportedSpan S A`
(`= ℤ[S] ∩ CF(L, A)`) is closed under `ℤ`-combinations, so the span of a subset it contains stays
inside it. This is what lets the general adjoin's `hisom` (needed on the difference lattices
`ℤ[{χ − χ̄, χ − a·χ₁}]` etc.) be supplied at the Feit–Sibley call site from
`Hypothesis.tau_isometry_diff` (which preserves `⟨·,·⟩` on all of `ℤ[𝒮, A]`): every difference the
engine feeds has its two endpoints in `𝒮`, hence in `ℤ[𝒮, A]`. -/
theorem zSpan_subset_zSupportedSpan {S : Set (ClassFunction L ℂ)} {A : Set L}
{T : Set (ClassFunction L ℂ)} (hT : ∀ s ∈ T, s ∈ zSupportedSpan (L := L) S A) :
(Submodule.span ℤ T : Set (ClassFunction L ℂ)) ⊆ zSupportedSpan (L := L) S A := by
intro φ hφ
induction hφ using Submodule.span_induction with
| mem s hs => exact hT s hs
| zero => exact zero_mem_zSupportedSpan
| add x y _ _ ihx ihy => exact add_mem_zSupportedSpan ihx ihy
| smul c x _ ih =>
refine ⟨Submodule.smul_mem _ c ih.1, ?_⟩
intro t ht
apply ih.2
rw [ClassFunction.mem_support] at ht ⊢
intro hx0
apply ht
rw [← Int.cast_smul_eq_zsmul ℂ c x, ClassFunction.smul_apply, hx0, mul_zero]
/-- **General `inner_Y_extension_member_eq`, rational-degree (scaled-difference) form**
(generalizes `inner_Y_extension_member_eq`, `S08_CoherenceCorePart1/CoherentAdjoin.lean`, to
**arbitrary member degrees** — Isaacs 7.14 / Peterfalvi Lemma 1(a) needs only the anchor
divisibility `χ₀(1) ∣ ψ(1)`, not `χ₀(1) ∣ χ(1)` for every member, so the ratio `χⱼ(1)/χ₁(1)` is
generally
rational).
The integer version used the degree-matched difference `χⱼ − aⱼ·χ₁` (needs `aⱼ = χⱼ(1)/χ₁(1) ∈ ℕ`);
here the **scaled difference `d₁·χⱼ − dⱼ·χ₁`** (`d₁ = χ₁(1)`, `dⱼ = χⱼ(1)`, integer coefficients,
degree `d₁·dⱼ − dⱼ·d₁ = 0`, hence in `ℤ[S₁, A]`) is used instead, so the `ν`-isometry on `ℤ[S₁]`
still applies (`hfound`). The conclusion is stated in the **division-free scaled form**
`d₁·⟨Y, νχⱼ⟩ = a·d₁·⟨χ₁,χⱼ⟩ − dⱼ·(a·m₁ + μ)` (`μ = ⟨τ(χ−a·χ₁), νχ₁⟩`); the caller
(`crux1_of_memberFamily_general`) divides by `d₁ ≠ 0` to obtain `⟨Y,νχⱼ⟩ = a·[j=1] − λ·(dⱼ/d₁)`. -/
theorem inner_Y_extension_member_eq_general
{τ : IntegralCharacterMap L G} {A : Set L}
[Fintype L] [Fintype G] [Invertible (Nat.card L : ℂ)] [Invertible (Nat.card G : ℂ)]
{S₁ : Set (ClassFunction L ℂ)}
(hS₁ : IsCoherent τ S₁ A)
(χ : ClassFunction L ℂ) {chi1 cj : ClassFunction L ℂ} {a : ℕ} (d1 dj : ℕ)
{Xχ Y : ClassFunction G ℂ}
(hYeq : Y = Xχ - τ (χ - a • chi1))
(hXortho : ClassFunction.inner Xχ (hS₁.extension cj) = 0)
(hfound : ClassFunction.inner (τ (χ - a • chi1)) (hS₁.extension (d1 • cj - dj • chi1)) =
ClassFunction.inner (χ - a • chi1) (d1 • cj - dj • chi1))
(hχcj : ClassFunction.inner χ cj = 0)
(hχchi1 : ClassFunction.inner χ chi1 = 0)
{m₁ : ℂ} (hchi1chi1 : ClassFunction.inner chi1 chi1 = m₁) :
(d1 : ℂ) * ClassFunction.inner Y (hS₁.extension cj) =
(a : ℂ) * (d1 : ℂ) * ClassFunction.inner chi1 cj -
(dj : ℂ) * ((a : ℂ) * m₁ + ClassFunction.inner (τ (χ - a • chi1))
(hS₁.extension chi1)) := by
-- `⟨Y, νχⱼ⟩ = −⟨τ(χ−a·χ₁), νχⱼ⟩` (`Xχ ⊥ νχⱼ`).
have hY0 : ClassFunction.inner Y (hS₁.extension cj) =
-ClassFunction.inner (τ (χ - a • chi1)) (hS₁.extension cj) := by
rw [hYeq, ClassFunction.inner_sub_left, hXortho, zero_sub]
-- LHS of `hfound`: `⟨τ(χ−a·χ₁), ν(d₁·χⱼ − dⱼ·χ₁)⟩ = d₁·⟨·,νχⱼ⟩ − dⱼ·⟨·,νχ₁⟩`.
have hνexp : hS₁.extension (d1 • cj - dj • chi1) =
(d1 : ℂ) • hS₁.extension cj - (dj : ℂ) • hS₁.extension chi1 := by
rw [map_sub, map_nsmul, map_nsmul, ← Nat.cast_smul_eq_nsmul ℂ d1 (hS₁.extension cj),
← Nat.cast_smul_eq_nsmul ℂ dj (hS₁.extension chi1)]
have hlhs : ClassFunction.inner (τ (χ - a • chi1)) (hS₁.extension (d1 • cj - dj • chi1)) =
(d1 : ℂ) * ClassFunction.inner (τ (χ - a • chi1)) (hS₁.extension cj) -
(dj : ℂ) * ClassFunction.inner (τ (χ - a • chi1)) (hS₁.extension chi1) := by
rw [hνexp, ClassFunction.inner_sub_right,
OddOrder.RepresentationTheory.inner_smul_right,
OddOrder.RepresentationTheory.inner_smul_right, star_natCast, star_natCast]
-- RHS of `hfound`: the source expansion.
have hrhs : ClassFunction.inner (χ - a • chi1) (d1 • cj - dj • chi1) =
-((a : ℂ) * (d1 : ℂ)) * ClassFunction.inner chi1 cj + (a : ℂ) * (dj : ℂ) * m₁ := by
rw [← Nat.cast_smul_eq_nsmul ℂ a chi1, ← Nat.cast_smul_eq_nsmul ℂ d1 cj,
← Nat.cast_smul_eq_nsmul ℂ dj chi1]
simp only [ClassFunction.inner_sub_left, ClassFunction.inner_sub_right,
ClassFunction.inner_smul_left, OddOrder.RepresentationTheory.inner_smul_right,
hχcj, hχchi1, hchi1chi1, star_natCast]
ring
rw [hlhs, hrhs] at hfound
-- `d₁·⟨τ(χ−a·χ₁),νχⱼ⟩ = dⱼ·μ − a·d₁·⟨χ₁,χⱼ⟩ + a·dⱼ·m₁`; multiply `hY0` by `d₁`.
rw [hY0]
ring_nf
ring_nf at hfound
linear_combination -hfound
/-- **General member `ψ = 0` decomposition** (generalizes `memberExtensionDecomposition`,
`S08_YsetInner`). For a member `χ ∈ S₁` (with `χ̄ ∈ S₁`) and its `R(χ)` family, the (5.5)
decomposition of `χ` uses the **running extension `ν = hS₁.extension` as the auxiliary isometry**
(not `τ`): `ν χ ∈ ℤ[Irr G]` (`hνZ`, since `χ ∈ S₁`) whereas `τ χ` need not be integral, and `ν`
agrees with `τ` on the supported difference `χ − χ̄`. The `R(χ)` family is w.r.t. `τ`
(supplied from `Hypothesis.difference_image`). -/
noncomputable def memberExtensionDecomposition_general
{τ : IntegralCharacterMap L G} {A : Set L}
[Fintype L] [Fintype G] [Invertible (Nat.card L : ℂ)] [Invertible (Nat.card G : ℂ)]
{S₁ : Set (ClassFunction L ℂ)}
(hS₁ : IsCoherent τ S₁ A) {χ : ClassFunction L ℂ}
(imageFamily : OrthonormalCharacterImageFamily (L := L) (G := G) τ χ)
(hdiffsupp : (χ.conj - χ).support ⊆ A)
(hχ_S1 : χ ∈ S₁) (hχbar_S1 : χ.conj ∈ S₁)
(hνZ : hS₁.extension χ ∈ ZIrr G)
(hχχbar : ClassFunction.inner χ χ.conj = 0) :
CharacterPsiDecomposition (L := L) (G := G) τ χ 0 := by
classical
have hχmem : χ ∈ Submodule.span ℤ S₁ := Submodule.subset_span hχ_S1
have hχbarmem : χ.conj ∈ Submodule.span ℤ S₁ := Submodule.subset_span hχbar_S1
have hle : Submodule.span ℤ ({χ - χ.conj, χ - 0} : Set (ClassFunction L ℂ)) ≤
Submodule.span ℤ S₁ := by
rw [Submodule.span_le]
intro s hs
simp only [Set.mem_insert_iff, Set.mem_singleton_iff] at hs
rcases hs with rfl | rfl
· exact Submodule.sub_mem _ hχmem hχbarmem
· rw [sub_zero]; exact hχmem
have hdiffsupported : χ - χ.conj ∈ zSupportedSpan (L := L) S₁ A :=
mem_zSupportedSpan_iff.mpr ⟨Submodule.sub_mem _ hχmem hχbarmem, by
rw [show χ - χ.conj = -(χ.conj - χ) from by abel, ClassFunction.support_neg]
exact hdiffsupp⟩
exact CharacterPsiDecomposition.ofProjection imageFamily hS₁.extension
(fun φ ζ hφ hζ => hS₁.extension_inner_eq φ ζ (hle hφ) (hle hζ))
(hS₁.extends_on_supported _ hdiffsupported)
(by rw [sub_zero]; exact hνZ)
(by simp) (by simp) hχχbar
/-- **General χ decomposition for `χ − a·χ₁`** (generalizes `decompositionDaFromDadeOfDiff`,
`S07_Coherence/FamilyBundleDade`). Uses `ofProjection` with the auxiliary isometry `τ` itself,
whose inner-preservation is needed only on the supported **difference** lattice `{χ − χ̄, χ − a·χ₁}`
(`hisom`), and the `ZIrr`-membership only on `χ − a·χ₁` (`hτaχ1Z`). Works for an unsupported `χ`
(e.g. `χ = Ind θ`). -/
noncomputable def decompositionDaFromDiff_general
{τ : IntegralCharacterMap L G} {A : Set L} {Samb : Set (ClassFunction L ℂ)}
[Fintype L] [Fintype G] [Invertible (Nat.card L : ℂ)] [Invertible (Nat.card G : ℂ)]
{χ chi1 : ClassFunction L ℂ} {a : ℕ}
(imageFamily : OrthonormalCharacterImageFamily (L := L) (G := G) τ χ)
(hisom : ∀ (T : Set (ClassFunction L ℂ)), (∀ s ∈ T, s ∈ zSupportedSpan (L := L) Samb A) →
∀ φ ζ : ClassFunction L ℂ, φ ∈ Submodule.span ℤ T → ζ ∈ Submodule.span ℤ T →
ClassFunction.inner (τ φ) (τ ζ) = ClassFunction.inner φ ζ)
(hdiffmem : χ - χ.conj ∈ zSupportedSpan (L := L) Samb A)
(hadiffmem : χ - a • chi1 ∈ zSupportedSpan (L := L) Samb A)
(hτaχ1Z : τ (χ - a • chi1) ∈ ZIrr G)
(hχaχ1 : ClassFunction.inner χ (a • chi1 : ClassFunction L ℂ) = 0)
(hχbaraχ1 : ClassFunction.inner χ.conj (a • chi1 : ClassFunction L ℂ) = 0)
(hχχbar : ClassFunction.inner χ χ.conj = 0) :
CharacterPsiDecomposition (L := L) (G := G) τ χ (a • chi1) := by
classical
have hSdiff : ∀ s ∈ ({χ - χ.conj, χ - a • chi1} : Set (ClassFunction L ℂ)),
s ∈ zSupportedSpan (L := L) Samb A := by
intro s hs
simp only [Set.mem_insert_iff, Set.mem_singleton_iff] at hs
rcases hs with rfl | rfl
· exact hdiffmem
· exact hadiffmem
exact CharacterPsiDecomposition.ofProjection imageFamily τ
(fun φ ζ hφ hζ => hisom _ hSdiff φ ζ hφ hζ)
rfl hτaχ1Z hχaχ1 hχbaraχ1 hχχbar
end OddOrder.Peterfalvi.S07
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/Peterfalvi/S08_GeneralAdjoin.lean