Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Peterfalvi §8 case (B): branch infrastructure and bundles

Definition
OddOrder_Peterfalvi_S08_CaseBAssembly_BranchBundles

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

feit-thompsonfinite-groupsgroup-theoryishida-odd-order

Peterfalvi, Character Theory for the Odd Order Theorem, §8, pp. 30-37, the (6.8.2.3) core of the (6.8) coherence capstone. The file establishes Dade-map canonicality, the column and irreducible branch lemmas, and the bundle constructors consumed by the mixed per-φ\varphiφ dispatch.

Main definitions: CaseBColBundle (the per-θ\thetaθ column-branch bundle) and CaseBIrrBundle (the per-θ\thetaθ irreducible-branch bundle), with supporting lemmas such as dade_H_eq, constituentWeight_eq_apply_one (aθ=θ(1)a_\theta=\theta(1)aθ​=θ(1)) and caseB_hnonlin (positive-weight constituents are non-linear).

The bundle also proves 34 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_CaseBAssembly/BranchBundles.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_S08_CaseBCoherence2

namespace OddOrder.Peterfalvi.S08

open OddOrder.RepresentationTheory

variable {G : Type*} [Group G] [Fintype G] [Invertible (Nat.card G : ℂ)]

variable {L : Subgroup G} [Fintype ↥L] [Invertible (Nat.card ↥L : ℂ)]

variable {H : Subgroup ↥L} [Invertible (Nat.card ↥H : ℂ)]

/-- **A subgroup of order coprime to a normal subgroup's index is contained in it.**  The image of
`K` in `C ⧸ N` has order dividing both `|K|` and `[C : N]` (Lagrange), which are coprime, so the
image
is trivial and `K ≤ ker (mk' N) = N`. -/
theorem le_of_card_coprime_index {C : Type*} [Group C]
    {K N : Subgroup C} [N.Normal] (hcop : Nat.Coprime (Nat.card K) N.index) : K ≤ N := by
  have hker : (QuotientGroup.mk' N).ker = N := QuotientGroup.ker_mk' N
  rw [← hker, ← Subgroup.map_eq_bot_iff, ← Subgroup.card_eq_one]
  have hd1 : Nat.card (K.map (QuotientGroup.mk' N)) ∣ Nat.card K :=
    Subgroup.card_map_dvd K (QuotientGroup.mk' N)
  have hd2 : Nat.card (K.map (QuotientGroup.mk' N)) ∣ N.index :=
    Subgroup.card_subgroup_dvd_card (K.map (QuotientGroup.mk' N))
  exact Nat.dvd_one.mp (hcop ▸ Nat.dvd_gcd hd1 hd2)

omit [Invertible (Nat.card G : ℂ)] in
omit [Fintype ↥L] [Invertible (Nat.card ↥L : ℂ)] in
/-- **The `H`-field of an `S04.Hypothesis` is determined by `(G, A, L)`.**  Two Dade hypotheses on
the
same data agree on every `H(a)`: each is the normal complement of `C_L(a)` in `C_G(a)`
(`centralizer_eq_sup`/`centralizer_disjoint`/`H_normalized`), of order coprime to `|C_L(a)|`
(`centralizer_coprime`).  Working inside `↥C_G(a)`, `[C_G(a) : H(a)] = |C_L(a)|`
(`card_centralizer_eq` + `card_mul_index`), so `le_of_card_coprime_index` gives each `H(a) ≤` the
other, hence equality.  This is the uniqueness behind the (6.8.2.3) map-agreement: since `H` is the
only *data* field of `S04.Hypothesis`, it makes the enlarged datum `h46.dade0` restrict to the base
`hyp.dade` on `H^#`-supported functions (Dade map being `H`-determined). -/
theorem dade_H_eq {A : Set G} (hyp₁ hyp₂ : OddOrder.Peterfalvi.S04.Hypothesis G A L)
    (a : {x : G // x ∈ A}) : hyp₁.H a = hyp₂.H a := by
  suffices h : ∀ p q : OddOrder.Peterfalvi.S04.Hypothesis G A L, p.H a ≤ q.H a from
    le_antisymm (h hyp₁ hyp₂) (h hyp₂ hyp₁)
  intro p q
  have hqC : q.H a ≤ Subgroup.centralizer ({a.1} : Set G) := by
    rw [q.centralizer_eq_sup a]; exact le_sup_left
  have hpC : p.H a ≤ Subgroup.centralizer ({a.1} : Set G) := by
    rw [p.centralizer_eq_sup a]; exact le_sup_left
  -- `(q.H a).subgroupOf C` is normal in `↥C` (C normalizes `q.H a`, by `H_normalized`).
  have hCnorm : Subgroup.centralizer ({a.1} : Set G) ≤ Subgroup.normalizer (q.H a) := by
    intro c hc
    rw [Subgroup.mem_normalizer_iff]
    intro n
    refine ⟨fun hn => q.H_normalized a c hc n hn, fun hn => ?_⟩
    have hmem := q.H_normalized a c⁻¹ (Subgroup.inv_mem _ hc) _ hn
    have heq : c⁻¹ * (c * n * c⁻¹) * (c⁻¹)⁻¹ = n := by group
    rwa [heq] at hmem
  have : ((q.H a).subgroupOf (Subgroup.centralizer ({a.1} : Set G))).Normal :=
    (Subgroup.normal_subgroupOf_iff_le_normalizer hqC).mpr hCnorm
  -- `|C| = |q.H a| · |C_L(a)|`.
  have hcard : Nat.card (Subgroup.centralizer ({a.1} : Set G))
      = Nat.card (q.H a) * Nat.card (OddOrder.Peterfalvi.S04.centralizerIn L a.1) :=
    q.card_centralizer_eq a
  -- `[↥C : (q.H a).subgroupOf C] = |C_L(a)|`.
  have hidx : ((q.H a).subgroupOf (Subgroup.centralizer ({a.1} : Set G))).index
      = Nat.card (OddOrder.Peterfalvi.S04.centralizerIn L a.1) := by
    have hmi := Subgroup.card_mul_index ((q.H a).subgroupOf (Subgroup.centralizer ({a.1} : Set G)))
    rw [Nat.card_congr (Subgroup.subgroupOfEquivOfLe hqC).toEquiv, hcard] at hmi
    exact Nat.eq_of_mul_eq_mul_left Nat.card_pos hmi
  -- `le_of_card_coprime_index` inside `↥C`, then transport back to `G`.
  have hle : (p.H a).subgroupOf (Subgroup.centralizer ({a.1} : Set G))
      ≤ (q.H a).subgroupOf (Subgroup.centralizer ({a.1} : Set G)) := by
    apply le_of_card_coprime_index
    rw [hidx, Nat.card_congr (Subgroup.subgroupOfEquivOfLe hpC).toEquiv]
    exact p.centralizer_coprime a a
  intro x hx
  have h1 : (⟨x, hpC hx⟩ : ↥(Subgroup.centralizer ({a.1} : Set G)))
      ∈ (p.H a).subgroupOf (Subgroup.centralizer ({a.1} : Set G)) :=
    Subgroup.mem_subgroupOf.mpr hx
  exact Subgroup.mem_subgroupOf.mp (hle h1)

omit [Invertible (Nat.card G : ℂ)] in
omit [Fintype ↥L] [Invertible (Nat.card ↥L : ℂ)] in
/-- **Structure extensionality for `S04.Hypothesis` via the data field `H`.**  `H` is the only
`Type`-valued field of `S04.Hypothesis`; the remaining eight fields are `Prop`s, equal by proof
irrelevance.  So two hypotheses on the same `(G, A, L)` whose `H`-fields coincide are equal.  This
is the structural half of the `dade_H_eq` canonicality (the group theory pins `H`, this packages it
back into a structure equality). -/
theorem dadeHypothesis_ext_of_H_eq {A : Set G}
    {p q : OddOrder.Peterfalvi.S04.Hypothesis G A L} (hH : p.H = q.H) : p = q := by
  cases p with
  | mk s1 s2 s3 H1 c1 c2 c3 c4 c5 =>
    cases q with
    | mk t1 t2 t3 H2 d1 d2 d3 d4 d5 =>
      subst hH
      rfl

/-- The integral Dade map `dadeIntegralCharacterMap` ignores its `FullDadeIsometryData` argument:
its definition extends only `hyp.dadeLinearMap`, so the map is a function of the underlying
`S04.Hypothesis` alone.  Hence equal hypotheses give the same map, regardless of the (unused)
isometry data carried alongside. -/
theorem dadeIntegralCharacterMap_congr_hyp {A : Set G}
    {hyp₁ hyp₂ : OddOrder.Peterfalvi.S04.Hypothesis G A L} (h : hyp₁ = hyp₂)
    (d₁ : OddOrder.Peterfalvi.S04.FullDadeIsometryData (G := G) hyp₁)
    (d₂ : OddOrder.Peterfalvi.S04.FullDadeIsometryData (G := G) hyp₂)
    (φ : ClassFunction ↥L ℂ) :
    OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap hyp₁ d₁ φ
      = OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap hyp₂ d₂ φ := by
  subst h; rfl

/-- **(6.8.2.3) map-agreement — the `hmapagree` linchpin.**  On every `H^#`-supported `φ`, the
enlarged `(4.6)` Dade map `dadeIntegralCharacterMap h46.dade0 h46.tau` (on `A₀ = H^# ∪ V^L`)
agrees with the Sibley–Dade `hyp.tau` (the genuine Dade map on `A = H^#`).

Both sides agree on `CF(L, H^#)` with the Dade map of an `S04.Hypothesis` over `(G, H^#, L)`:
restricting `h46.dade0` to `H^#` (`dadeIntegralCharacterMap_restrict_eq_of_support`) yields such a
hypothesis, and by the canonicality `dade_H_eq` it shares its only data field `H` with `hyp.dade`,
hence equals it (`dadeHypothesis_ext_of_H_eq`); since `dadeIntegralCharacterMap` depends only on
the hypothesis (`dadeIntegralCharacterMap_congr_hyp`) and `hyp.tau` unfolds to
`dadeIntegralCharacterMap hyp.dade _`, the two maps agree.  This is the agreement consumed by
`certainTypeSet_isCoherent_tau`
to re-target the certain-type column coherence (built against `h46.dade0`) onto `hyp.tau`. -/
theorem SibleyDadeHypothesis.dade0_map_eq_tau_of_support
    (hyp : SibleyDadeHypothesis G L H)
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L)
    {φ : ClassFunction ↥L ℂ}
    (hφ : φ.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup (sharpImage H) L) :
    OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau φ = hyp.tau φ := by
  have hLemmaD :
      h46.dade0.restrict Set.subset_union_left hyp.dade.L_normalizes_A = hyp.dade :=
    dadeHypothesis_ext_of_H_eq
      (funext fun a => dade_H_eq
        (h46.dade0.restrict Set.subset_union_left hyp.dade.L_normalizes_A) hyp.dade a)
  rw [← dadeIntegralCharacterMap_restrict_eq_of_support h46.dade0 h46.tau
        Set.subset_union_left hyp.dade.L_normalizes_A hφ]
  exact dadeIntegralCharacterMap_congr_hyp hLemmaD _ _ φ

/-- **(6.8.2.3) column-branch map-agreement** (the `hmapagree` conjunct of `CaseBColBundle`),
*unconditional*.  For any nontrivial column `χ₂`, the `hyp.tau`-image of the conjugate column
difference agrees with its enlarged `(4.6)` Dade image
`dadeIntegralCharacterMap h46.dade0 h46.tau`.  This is the column-bundle conjunct that was
*unconstructible* before the canonicality linchpin: the conjugate column is `μ_{χ₂⁻¹}`
(`columnSum_conj_eq`) of equal degree (`columnSum_inv_apply_one`), so the difference is
`H^#`-supported (`columnDiff_support_subset`) and the map-agreement is exactly
`dade0_map_eq_tau_of_support`.  Both bundle-intrinsic column conjuncts — degree equality and this
map-agreement — are thus discharged with no side hypotheses. -/
theorem caseB_column_mapagree
    (hyp : SibleyDadeHypothesis G L H)
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ} (hχ₂ : χ₂ ≠ 1) :
    hyp.tau (OddOrder.Peterfalvi.S06.columnSum h46 χ₂
        - (OddOrder.Peterfalvi.S06.columnSum h46 χ₂).conj)
      = OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau
        (OddOrder.Peterfalvi.S06.columnSum h46 χ₂
          - (OddOrder.Peterfalvi.S06.columnSum h46 χ₂).conj) := by
  rw [OddOrder.Peterfalvi.S06.columnSum_conj_eq]
  exact (hyp.dade0_map_eq_tau_of_support h46
    (OddOrder.Peterfalvi.S06.columnDiff_support_subset h46 hχ₂
      ((inv_ne_one (a := χ₂)).mpr hχ₂)
      (OddOrder.Peterfalvi.S06.columnSum_inv_apply_one h46 χ₂).symm)).symm

/-- **(6.8.2.3) raw `X ⊥ Y`, per certain-type constituent** — Peterfalvi (4.1), source level.
A grid character of a column is orthogonal to every `Y`-member.  Both are irreducible
(`SignedIrreducibleDifferenceFamily.mu` is `IrreducibleCharacter`), and distinct *by degree*: a
`Y`-member has degree `|W₁|` (`Yset_apply_one`) but a grid degree is `≡ ±1 (mod |W₁|)`
(`certainType_degree_modEq`), with `|W₁| ≠ 1`.  So the inner product vanishes. -/
theorem inner_columnFamily_mu_Yset_eq_zero
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hW1 : h46.W1 = hyp.W1)
    [NeZero (Nat.card h46.W1)] [Finite ↥(h46.W1 ⊔ h46.W2)]
    [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {η : ClassFunction ↥L ℂ} (hη : η ∈ hyp.Yset)
    (χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ) (i : Fin (Nat.card h46.W1)) :
    ClassFunction.inner ((h46.columnFamily χ₂).mu i : ClassFunction ↥L ℂ) η = 0 := by
  have : Fintype ↥(h46.W1 ⊔ h46.W2) := Fintype.ofFinite _
  have hηirr := hyp.isIrreducibleCharacter_of_mem_Yset hη
  have hne : ((h46.columnFamily χ₂).mu i : ClassFunction ↥L ℂ) ≠ η := by
    intro heq
    obtain ⟨a, ha⟩ := h46.certainType_degree_modEq χ₂ i
    rw [heq, hyp.Yset_apply_one hη] at ha
    have hcard : (Nat.card h46.W1 : ℂ) = (Nat.card hyp.W1 : ℂ) := by rw [hW1]
    rw [hcard] at ha
    have hw1 : Nat.card hyp.W1 ≠ 1 := fun h => hyp.W1_nontrivial (Subgroup.card_eq_one.mp h)
    have hsign : ((h46.columnFamily χ₂).sign : ℂ)
        = (Nat.card hyp.W1 : ℂ) * (1 - (a : ℂ)) := by linear_combination -ha
    have hsignZ : (h46.columnFamily χ₂).sign = (Nat.card hyp.W1 : ℤ) * (1 - a) := by
      exact_mod_cast hsign
    have hdvd1 : (Nat.card hyp.W1 : ℤ) ∣ 1 := by
      have hdvd : (Nat.card hyp.W1 : ℤ) ∣ (h46.columnFamily χ₂).sign := ⟨1 - a, hsignZ⟩
      rcases (h46.columnFamily χ₂).sign_eq with hs | hs
      · rwa [hs] at hdvd
      · rw [hs] at hdvd; exact (dvd_neg).mp hdvd
    exact hw1 (Nat.dvd_one.mp (by exact_mod_cast hdvd1))
  have hkron := irreducibleCharacter_inner_eq_ite ((h46.columnFamily χ₂).mu i)
    (⟨η, hηirr⟩ : IrreducibleCharacter ↥L)
  rw [if_neg (fun heq => hne (Subtype.ext_iff.mp heq))] at hkron
  simpa using hkron

/-- **(6.8.2.3) raw `X ⊥ Y`, column form** — Peterfalvi (4.1).  A whole certain-type column
(a sum of grid characters) is orthogonal to every `Y`-member, by additivity over the constituents
(`inner_columnFamily_mu_Yset_eq_zero`).  This is the `hpair` orthogonality of the case-(B) `X ∪ Y`
glue, restricted to the reducible column side. -/
theorem inner_columnSum_Yset_eq_zero
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hW1 : h46.W1 = hyp.W1)
    [NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
    [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {η : ClassFunction ↥L ℂ} (hη : η ∈ hyp.Yset)
    (χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ) :
    ClassFunction.inner (OddOrder.Peterfalvi.S06.columnSum h46 χ₂) η = 0 := by
  rw [OddOrder.Peterfalvi.S06.columnSum_def, inner_sum_left]
  exact Finset.sum_eq_zero
    (fun i _ => inner_columnFamily_mu_Yset_eq_zero hyp h46 hW1 hη χ₂ i)

/-- **(6.8.2.3) column-branch orthogonality scalars** (the `hχψ` conjunct of `CaseBColBundle`).
A column is orthogonal to the weighted `Y`-anchor `a • η₁`: pull the natural scalar out
(`inner_smul_right`) and apply the raw column orthogonality `inner_columnSum_Yset_eq_zero`. -/
theorem caseB_column_orthogonal_Yset
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hW1 : h46.W1 = hyp.W1)
    [NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
    [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ) (a : ℕ) :
    ClassFunction.inner (OddOrder.Peterfalvi.S06.columnSum h46 χ₂)
      (a • η₁ : ClassFunction ↥L ℂ) = 0 := by
  rw [← Nat.cast_smul_eq_nsmul ℂ, OddOrder.RepresentationTheory.inner_smul_right,
    inner_columnSum_Yset_eq_zero hyp h46 hW1 hη₁ χ₂, mul_zero]

/-- **(6.8.2.3) column-branch orthogonality scalars, conjugate** (the `hχbarψ` conjunct of
`CaseBColBundle`).  The conjugate column is the inverse column `columnSum h46 χ₂⁻¹`
(`columnSum_conj_eq`), so it too is orthogonal to the weighted `Y`-anchor by
`caseB_column_orthogonal_Yset`. -/
theorem caseB_column_conj_orthogonal_Yset
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hW1 : h46.W1 = hyp.W1)
    [NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
    [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ) (a : ℕ) :
    ClassFunction.inner (OddOrder.Peterfalvi.S06.columnSum h46 χ₂).conj
      (a • η₁ : ClassFunction ↥L ℂ) = 0 := by
  rw [OddOrder.Peterfalvi.S06.columnSum_conj_eq]
  exact caseB_column_orthogonal_Yset hyp h46 hW1 hη₁ χ₂⁻¹ a

/-- **(6.8.2.3) column-anchor difference support.**  Given the degree match at `1`, the column
difference `columnSum h46 χ₂ − a·η₁` is `H^#`-supported.  A column is induced from `H`
(`columnSum_eq_induce_H`, case `h46.K = H`), so this is the `Ind − scalar·η₁` support lemma
`support_indW2_sub_smul_subset_sharpImage` instantiated at the subgroup `H` itself. -/
theorem caseB_column_sub_smul_support
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ) (a : ℕ)
    (h1 : (OddOrder.Peterfalvi.S06.columnSum h46 χ₂) (1 : ↥L) = (a : ℂ) * η₁ (1 : ↥L)) :
    (OddOrder.Peterfalvi.S06.columnSum h46 χ₂ - a • η₁).support ⊆
      OddOrder.Peterfalvi.S04.supportInSubgroup (sharpImage H) L := by
  rw [columnSum_eq_induce_H h46 hHK, ← Nat.cast_smul_eq_nsmul ℂ]
  refine hyp.support_indW2_sub_smul_subset_sharpImage (le_refl H) _ hη₁ (a : ℂ) ?_
  rw [← columnSum_eq_induce_H h46 hHK]; exact h1

/-- **(6.8.2.3) column-branch supports** (the `hSdiff` conjunct of `CaseBColBundle`).  Both running
difference generators are `H^#`-supported: `columnSum − conj` by `columnDiff_support_subset` (the
conjugate is the inverse column, equal degree), and `columnSum − a·η₁` by
`caseB_column_sub_smul_support` (given the degree match `h1`). -/
theorem caseB_column_hSdiff
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    {χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ} (hχ₂ : χ₂ ≠ 1) (a : ℕ)
    (h1 : (OddOrder.Peterfalvi.S06.columnSum h46 χ₂) (1 : ↥L) = (a : ℂ) * η₁ (1 : ↥L)) :
    ∀ s ∈ ({OddOrder.Peterfalvi.S06.columnSum h46 χ₂
        - (OddOrder.Peterfalvi.S06.columnSum h46 χ₂).conj,
        OddOrder.Peterfalvi.S06.columnSum h46 χ₂ - a • η₁} : Set (ClassFunction ↥L ℂ)),
      s.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup (sharpImage H) L := by
  intro s hs
  simp only [Set.mem_insert_iff, Set.mem_singleton_iff] at hs
  rcases hs with rfl | rfl
  · rw [OddOrder.Peterfalvi.S06.columnSum_conj_eq]
    exact OddOrder.Peterfalvi.S06.columnDiff_support_subset h46 hχ₂
      ((inv_ne_one (a := χ₂)).mpr hχ₂)
      (OddOrder.Peterfalvi.S06.columnSum_inv_apply_one h46 χ₂).symm
  · exact caseB_column_sub_smul_support hyp h46 hHK hη₁ χ₂ a h1

/-- **(6.8.2.3) column-branch `ZIrr`-membership** (the `htau1_mema` conjunct of `CaseBColBundle`).
The `hyp.tau`-image of the column difference is a virtual character: the argument
is `H^#`-supported (`caseB_column_sub_smul_support`, given the degree match) and virtual
(`columnSum` a sum of irreducibles, `η₁` irreducible), so the Dade map carries it into `ZIrr G`
(`dadeIntegralCharacterMap_mem_ZIrr_of_supported`, the `(2.6)` integrality of `tau`). -/
theorem caseB_column_htau1_mema
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ) (a : ℕ)
    (h1 : (OddOrder.Peterfalvi.S06.columnSum h46 χ₂) (1 : ↥L) = (a : ℂ) * η₁ (1 : ↥L)) :
    hyp.tau (OddOrder.Peterfalvi.S06.columnSum h46 χ₂ - a • η₁) ∈ ZIrr G := by
  refine OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_mem_ZIrr_of_supported
    hyp.dade (caseB_column_sub_smul_support hyp h46 hHK hη₁ χ₂ a h1) ?_
  refine Submodule.sub_mem _ ?_
    (nsmul_mem (IsIrreducibleCharacter.mem_ZIrr (hyp.isIrreducibleCharacter_of_mem_Yset hη₁)) a)
  rw [OddOrder.Peterfalvi.S06.columnSum_def]
  exact Submodule.sum_mem _ (fun i _ => ((h46.columnFamily χ₂).mu i).mem_ZIrr)

omit [Invertible (Nat.card G : ℂ)] in
omit [Fintype ↥L] [Invertible (Nat.card ↥L : ℂ)] [Invertible (Nat.card ↥H : ℂ)] in
omit [Fintype G] in
/-- **(6.8.2.3) weight reconciliation `aθ = θ(1)`.**  When the source `φ` actually occurs in the
central restriction of `θ` (`0 < constituentWeight`), the multiplicity `aθ = ⟨φ, Res^H_{W₂} θ⟩`
equals the degree `θ(1)`.  Indeed `W₂` is central in `H`, so `Res^H_{W₂} θ = θ(1)·λ` for a unique
linear `λ` (Schur, `exists_central_linear_restriction`); then `aθ = θ(1)·⟨φ, λ⟩` with `⟨φ, λ⟩` a
Kronecker delta (both irreducible), and positivity forces it to be `1`.  This reconciles the
multiplicity weight with the degree ratio `θ(1) = χθ(1)/|W₁|` of (6.8.2.3). -/
theorem constituentWeight_eq_apply_one
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)] [Finite ↥H]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {θ : IrreducibleCharacter ↥H} (hweight : 0 < constituentWeight hφ' θ) :
    (constituentWeight hφ' θ : ℂ) = (θ : ClassFunction ↥H ℂ) 1 := by
  have : Fintype ↥H := Fintype.ofFinite _
  obtain ⟨lam, hlamirr, _hlam1, hres, _⟩ :=
    θ.2.exists_central_linear_restriction (W2.subgroupOf H) hcen
  have hspec := constituentWeight_spec hφ' θ
  rw [hres, OddOrder.RepresentationTheory.inner_smul_right] at hspec
  obtain ⟨d, _hd0, hd⟩ := irreducibleCharacter_apply_one_eq_pos_natCast θ
  rw [hd, star_natCast] at hspec
  have hkron := irreducibleCharacter_inner_eq_ite
    (⟨ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ, hφ'⟩ :
      IrreducibleCharacter ↥(W2.subgroupOf H))
    (⟨lam, hlamirr⟩ : IrreducibleCharacter ↥(W2.subgroupOf H))
  by_cases heq : (⟨ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ, hφ'⟩ :
      IrreducibleCharacter ↥(W2.subgroupOf H)) = ⟨lam, hlamirr⟩
  · rw [if_pos heq] at hkron
    rw [hkron, mul_one] at hspec
    rw [hd]; exact hspec.symm
  · rw [if_neg heq] at hkron
    rw [hkron, mul_zero] at hspec
    exact absurd hspec.symm (by exact_mod_cast hweight.ne')

/-- **(6.8.2.3) column-anchor degree match.**  For a column `columnSum h46 χ₂ = Ind^L_H θ` with the
source occurring in `θ` (`0 < constituentWeight`), the column degree equals the weighted `Y`-anchor
degree: `columnSum(1) = constituentWeight · η₁(1)`.  Indeed `columnSum(1) = (Ind^L_H θ)(1) =
|W₁|·θ(1)` (`induce_apply_one`, `index_H_eq_card_W1`), `η₁(1) = |W₁|` (`Yset_apply_one`), and
`constituentWeight = θ(1)` (`constituentWeight_eq_apply_one`).  This is the `h1` hypothesis that the
support/`ZIrr` column conjuncts need. -/
theorem caseB_column_degree_match
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L)
    [NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
    [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {θ : IrreducibleCharacter ↥H} (hweight : 0 < constituentWeight hφ' θ)
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    {χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ}
    (hcoleq : OddOrder.Peterfalvi.S06.columnSum h46 χ₂
      = ClassFunction.induce H (θ : ClassFunction ↥H ℂ)) :
    (OddOrder.Peterfalvi.S06.columnSum h46 χ₂) (1 : ↥L)
      = (constituentWeight hφ' θ : ℂ) * η₁ (1 : ↥L) := by
  have : Fintype ↥H := Fintype.ofFinite _
  rw [hcoleq, ClassFunction.induce_apply_one, hyp.index_H_eq_card_W1, hyp.Yset_apply_one hη₁,
    constituentWeight_eq_apply_one hW2H hcen hφ' hweight]
  ring

/-- **(6.8.2.3) induced-character degree match** (irreducible-branch form).  For the dispatch weight
`aθ = constituentWeight hφ' θ` (`0 < aθ`), the induced degree matches the weighted `Y`-anchor:
`(Ind^L_H θ)(1) = aθ · η₁(1)`.  Same arithmetic as `caseB_column_degree_match`, for the
induced character directly (no column rewrite). -/
theorem caseB_induce_degree_match
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {θ : IrreducibleCharacter ↥H} (hweight : 0 < constituentWeight hφ' θ)
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset) :
    (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)) (1 : ↥L)
      = (constituentWeight hφ' θ : ℂ) * η₁ (1 : ↥L) := by
  have : Fintype ↥H := Fintype.ofFinite _
  rw [ClassFunction.induce_apply_one, hyp.index_H_eq_card_W1, hyp.Yset_apply_one hη₁,
    constituentWeight_eq_apply_one hW2H hcen hφ' hweight]
  ring

/-- **(6.8.2.3) irreducible-branch anchor difference support** (`CaseBIrrBundle` conjunct).
`Ind^L_H θ − aθ·η₁` is `H^#`-supported, by `support_indW2_sub_smul_subset_sharpImage` at `H`
with the degree match `caseB_induce_degree_match`. -/
theorem caseB_irr_sub_smul_support
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {θ : IrreducibleCharacter ↥H} (hweight : 0 < constituentWeight hφ' θ)
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset) :
    (ClassFunction.induce H (θ : ClassFunction ↥H ℂ) - constituentWeight hφ' θ • η₁).support ⊆
      OddOrder.Peterfalvi.S04.supportInSubgroup (sharpImage H) L := by
  have : Fintype ↥H := Fintype.ofFinite _
  rw [← Nat.cast_smul_eq_nsmul ℂ]
  exact hyp.support_indW2_sub_smul_subset_sharpImage (le_refl H) _ hη₁
    (constituentWeight hφ' θ : ℂ) (caseB_induce_degree_match hyp hW2H hcen hφ' hweight hη₁)

/-- **(6.8.2.3) irreducible-branch `ZIrr`-membership** (`CaseBIrrBundle` conjunct).
`hyp.tau (Ind^L_H θ − aθ·η₁) ∈ ZIrr G`: the argument is `H^#`-supported
(`caseB_irr_sub_smul_support`) and a virtual character (`induce_mem_ZIrr`, `η₁` irreducible), so the
Dade map carries it into `ZIrr G` (`dadeIntegralCharacterMap_mem_ZIrr_of_supported`). -/
theorem caseB_irr_htau1_mema
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {θ : IrreducibleCharacter ↥H} (hweight : 0 < constituentWeight hφ' θ)
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset) :
    hyp.tau (ClassFunction.induce H (θ : ClassFunction ↥H ℂ) - constituentWeight hφ' θ • η₁)
      ∈ ZIrr G := by
  have : Fintype ↥H := Fintype.ofFinite _
  refine OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_mem_ZIrr_of_supported
    hyp.dade (caseB_irr_sub_smul_support hyp hW2H hcen hφ' hweight hη₁) ?_
  exact Submodule.sub_mem _
    (ClassFunction.induce_mem_ZIrr H (IsIrreducibleCharacter.mem_ZIrr θ.2))
    (nsmul_mem (IsIrreducibleCharacter.mem_ZIrr (hyp.isIrreducibleCharacter_of_mem_Yset hη₁)) _)

/-- **(6.8.2.3) irreducible-branch conjugate difference support** (`CaseBIrrBundle` conjunct).
`(Ind^L_H θ).conj − Ind^L_H θ` is `H^#`-supported, with no irreducibility assumption: both terms are
supported on `H` (`support_induce_subset_of_normal`, conj preserves support), and the difference
vanishes at `1` since `(Ind^L_H θ)(1) = |W₁|·θ(1)` is a (real) positive integer fixed by `star`. -/
theorem caseB_irr_conj_diff_support
    (_hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    (θ : IrreducibleCharacter ↥H) :
    ((ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj
        - ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).support ⊆
      OddOrder.Peterfalvi.S04.supportInSubgroup (sharpImage H) L := by
  have : Fintype ↥H := Fintype.ofFinite _
  have hsupp : (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).support ⊆ (H : Set ↥L) :=
    ClassFunction.support_induce_subset_of_normal H _
  intro x hx
  rw [ClassFunction.mem_support, ClassFunction.sub_apply] at hx
  have hxH : x ∈ H := by
    by_contra hxnotH
    have h1 : ClassFunction.induce H (θ : ClassFunction ↥H ℂ) x = 0 := by
      by_contra h; exact hxnotH (hsupp (ClassFunction.mem_support.mpr h))
    exact hx (by rw [ClassFunction.conj_apply, h1, star_zero, sub_zero])
  have hxne : x ≠ 1 := by
    intro hx1
    obtain ⟨d, _, hd⟩ := irreducibleCharacter_apply_one_eq_pos_natCast θ
    refine hx ?_
    rw [hx1, ClassFunction.conj_apply, ClassFunction.induce_apply_one, hd]
    simp [← Nat.cast_mul]
  change (x : G) ∈ sharpImage H
  exact ⟨Subgroup.mem_map.mpr ⟨x, hxH, rfl⟩, fun hx1G => hxne (Subtype.ext hx1G)⟩

/-- **(6.8.2.3) irreducible-branch irreducibility** (`CaseBIrrBundle` conjunct #1, the gate).
When `Ind^L_H θ` is not a (nontrivial) certain-type column (`hnotcol`) and `θ` is nontrivial, the
induced character is irreducible.  Transport `θ` to `↥h46.K` (`subgroupCongr` +
`compHom_of_surjective`), then apply `induce_isIrreducible_of_forall_chiRestrict_ne`: a nontrivial
column source `chiRestrict χ₂` would give a column equal to the induced character
(`induce_restrict_certainType_eq`), contradicting `hnotcol`; and `chiRestrict 1` is trivial
(`certainType_zero_column_anchor`) while the transported `θ` is not. -/
theorem caseB_irr_induce_isIrreducible
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {θ : IrreducibleCharacter ↥H}
    (hθne : (θ : ClassFunction ↥H ℂ) ≠ trivialClassFunction ↥H)
    (hnotcol : ∀ χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ, χ₂ ≠ 1 →
      OddOrder.Peterfalvi.S06.columnSum h46 χ₂
        ≠ ClassFunction.induce H (θ : ClassFunction ↥H ℂ)) :
    IsIrreducibleCharacter (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)) := by
  set e : ↥h46.K ≃* ↥H := MulEquiv.subgroupCongr hHK with he
  set θK : IrreducibleCharacter ↥h46.K :=
    ⟨ClassFunction.compHom e.toMonoidHom (θ : ClassFunction ↥H ℂ),
      IsIrreducibleCharacter.compHom_of_surjective e.surjective θ.2⟩ with hθK
  have hθKval : ∀ (x : ↥L) (hx₁ : x ∈ h46.K) (hx₂ : x ∈ H),
      (θK : ClassFunction ↥h46.K ℂ) ⟨x, hx₁⟩
        = (θ : ClassFunction ↥H ℂ) ⟨x, hx₂⟩ := by
    intro x hx₁ hx₂
    change ClassFunction.compHom e.toMonoidHom (θ : ClassFunction ↥H ℂ) ⟨x, hx₁⟩ = _
    rw [ClassFunction.compHom_apply]
    rfl
  have hindeq : ClassFunction.induce h46.K (θK : ClassFunction ↥h46.K ℂ)
      = ClassFunction.induce H (θ : ClassFunction ↥H ℂ) :=
    OddOrder.Peterfalvi.S04.Hypothesis.induce_congr_of_subgroup_eq hHK hθKval
  rw [← hindeq]
  refine h46.induce_isIrreducible_of_forall_chiRestrict_ne (fun χ₂ hcontra => ?_)
  by_cases hχ₂1 : χ₂ = 1
  · -- `χ₂ = 1`: `chiRestrict 1` is trivial, so `θ ∘ e = 1`, hence `θ = 1` (`e` surjective)
    subst hχ₂1
    have hcompeq : ClassFunction.compHom e.toMonoidHom (θ : ClassFunction ↥H ℂ)
        = ClassFunction.restrict h46.K (trivialClassFunction ↥L) := by
      have h0 : (θK : ClassFunction ↥h46.K ℂ)
          = ClassFunction.restrict h46.K (trivialClassFunction ↥L) := by
        rw [← hcontra, h46.coe_chiRestrict, (h46.certainType_zero_column_anchor).2]
      rw [hθK] at h0; exact h0
    refine hθne (ClassFunction.ext fun h => ?_)
    obtain ⟨k, rfl⟩ := e.surjective h
    have hk := congrArg (fun f : ClassFunction ↥h46.K ℂ => f k) hcompeq
    simp only [ClassFunction.compHom_apply, ClassFunction.restrict_apply,
      trivialClassFunction_apply, MulEquiv.coe_toMonoidHom] at hk
    rw [hk, trivialClassFunction_apply]
  · -- `χ₂ ≠ 1`: `chiRestrict χ₂ = θ_K` forces `columnSum χ₂ = Ind^L_H θ`
    refine hnotcol χ₂ hχ₂1 ?_
    rw [← hindeq, OddOrder.Peterfalvi.S06.columnSum_def,
      ← h46.induce_restrict_certainType_eq χ₂, ← h46.coe_chiRestrict χ₂, hcontra]

/-- `Ind^L_H θ ≠ 1_L`: its degree is a multiple of `|W₁| > 1` (`W₁ ≠ ⊥`), unlike the
degree-`1` trivial character. -/
theorem caseB_induce_ne_trivial
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    (θ : IrreducibleCharacter ↥H) :
    ClassFunction.induce H (θ : ClassFunction ↥H ℂ) ≠ trivialClassFunction ↥L := by
  intro h
  obtain ⟨d, hd0, hd⟩ := irreducibleCharacter_apply_one_eq_pos_natCast θ
  have h1 : ClassFunction.induce H (θ : ClassFunction ↥H ℂ) (1 : ↥L)
      = trivialClassFunction ↥L (1 : ↥L) := by rw [h]
  rw [ClassFunction.induce_apply_one, hyp.index_H_eq_card_W1, hd, trivialClassFunction_apply,
    ← Nat.cast_mul] at h1
  have h2 : Nat.card hyp.W1 * d = 1 := by exact_mod_cast h1
  exact hyp.W1_nontrivial (Subgroup.card_eq_one.mp (Nat.dvd_one.mp ⟨d, h2.symm⟩))

/-- **(6.8.2.3) irreducible-branch non-realness** (`CaseBIrrBundle` conjunct #2).  An irreducible
`Ind^L_H θ` is non-real: it is nontrivial (`caseB_induce_ne_trivial`) and `L` has odd order
(`card_L_odd`), so Peterfalvi (1.1) (`not_isReal_of_ne_trivial_of_odd_card'`) applies. -/
theorem caseB_irr_nonreal
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    {θ : IrreducibleCharacter ↥H}
    (hirr1 : IsIrreducibleCharacter (ClassFunction.induce H (θ : ClassFunction ↥H ℂ))) :
    ¬ ClassFunction.IsReal (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)) := by
  refine OddOrder.RepresentationTheory.not_isReal_of_ne_trivial_of_odd_card'
    hyp.card_L_odd (χ := ⟨_, hirr1⟩) (fun htriv => caseB_induce_ne_trivial hyp θ ?_)
  rw [← IrreducibleCharacter.coe_trivialIrreducibleCharacter (G := ↥L), ← htriv]

/-- **(6.8.2.3) irreducible-branch self-conjugate orthogonality** (`CaseBIrrBundle` conjunct #8).
`⟨Ind^L_H θ, (Ind^L_H θ).conj⟩ = 0`: both `Ind^L_H θ` (`hirr1`) and its conjugate
(`IsIrreducibleCharacter.conj`) are irreducible and distinct (non-real, `caseB_irr_nonreal`), so the
irreducible Kronecker inner product vanishes. -/
theorem caseB_irr_conj_inner
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    {θ : IrreducibleCharacter ↥H}
    (hirr1 : IsIrreducibleCharacter (ClassFunction.induce H (θ : ClassFunction ↥H ℂ))) :
    ClassFunction.inner (ClassFunction.induce H (θ : ClassFunction ↥H ℂ))
        (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj = 0 := by
  set Φ := ClassFunction.induce H (θ : ClassFunction ↥H ℂ) with hΦ
  have hnonreal := caseB_irr_nonreal hyp hirr1
  have hne : (⟨Φ, hirr1⟩ : IrreducibleCharacter ↥L) ≠ ⟨Φ.conj, hirr1.conj⟩ := by
    intro heq
    apply hnonreal
    have h2 := congrArg (fun c : IrreducibleCharacter ↥L => (c : ClassFunction ↥L ℂ)) heq
    change Φ.conj = Φ
    simpa using h2.symm
  have hkron := irreducibleCharacter_inner_eq_ite (⟨Φ, hirr1⟩ : IrreducibleCharacter ↥L)
    (⟨Φ.conj, hirr1.conj⟩ : IrreducibleCharacter ↥L)
  rw [if_neg hne] at hkron
  simpa using hkron

/-- An irreducible character distinct from a `Y`-member is orthogonal to it (both irreducible, so
the inner product is the Kronecker delta).  The raw `X ⊥ Y` for the irreducible branch. -/
theorem inner_irr_Yset_eq_zero
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    {φ η₁ : ClassFunction ↥L ℂ} (hφirr : IsIrreducibleCharacter φ)
    (hη₁ : η₁ ∈ hyp.Yset) (hne : φ ≠ η₁) : ClassFunction.inner φ η₁ = 0 := by
  have hηirr := hyp.isIrreducibleCharacter_of_mem_Yset hη₁
  have hkron := irreducibleCharacter_inner_eq_ite (⟨φ, hφirr⟩ : IrreducibleCharacter ↥L)
    (⟨η₁, hηirr⟩ : IrreducibleCharacter ↥L)
  rw [if_neg (fun heq => hne (by
    simpa using congrArg (fun c : IrreducibleCharacter ↥L => (c : ClassFunction ↥L ℂ)) heq))]
    at hkron
  simpa using hkron

/-- **(6.8.2.3) irreducible-branch anchor orthogonality** (`CaseBIrrBundle` conjunct #6).
The induced character is orthogonal to the weighted `Y`-anchor, given the `X ⊥ Y` distinctness
`Ind^L_H θ ≠ η₁` (a structural input), via `inner_irr_Yset_eq_zero`. -/
theorem caseB_irr_orthogonal_Yset
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    {θ : IrreducibleCharacter ↥H}
    (hirr1 : IsIrreducibleCharacter (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)))
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (hne : ClassFunction.induce H (θ : ClassFunction ↥H ℂ) ≠ η₁) (a : ℕ) :
    ClassFunction.inner (ClassFunction.induce H (θ : ClassFunction ↥H ℂ))
      (a • η₁ : ClassFunction ↥L ℂ) = 0 := by
  rw [← Nat.cast_smul_eq_nsmul ℂ, OddOrder.RepresentationTheory.inner_smul_right,
    inner_irr_Yset_eq_zero hyp hirr1 hη₁ hne, mul_zero]

/-- **(6.8.2.3) irreducible-branch conjugate anchor orthogonality** (`CaseBIrrBundle` conjunct #7).
The conjugate induced character is orthogonal to the weighted `Y`-anchor, given
`(Ind^L_H θ).conj ≠ η₁`: the conjugate is irreducible, so `inner_irr_Yset_eq_zero` applies. -/
theorem caseB_irr_conj_orthogonal_Yset
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    {θ : IrreducibleCharacter ↥H}
    (hirr1 : IsIrreducibleCharacter (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)))
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (hne : (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj ≠ η₁) (a : ℕ) :
    ClassFunction.inner (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj
      (a • η₁ : ClassFunction ↥L ℂ) = 0 := by
  rw [← Nat.cast_smul_eq_nsmul ℂ, OddOrder.RepresentationTheory.inner_smul_right,
    inner_irr_Yset_eq_zero hyp hirr1.conj hη₁ hne, mul_zero]

/-- A `Y`-member is orthogonal to a distinct irreducible (`Y`-member in the *first* slot, the form
the per-`θ` anchor orthogonality `hirrAnc` needs).  Irreducible Kronecker delta. -/
theorem inner_Yset_irr_eq_zero
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    {η ψ : ClassFunction ↥L ℂ} (hη : η ∈ hyp.Yset) (hψirr : IsIrreducibleCharacter ψ)
    (hne : η ≠ ψ) : ClassFunction.inner η ψ = 0 := by
  have hηirr := hyp.isIrreducibleCharacter_of_mem_Yset hη
  have hkron := irreducibleCharacter_inner_eq_ite (⟨η, hηirr⟩ : IrreducibleCharacter ↥L)
    (⟨ψ, hψirr⟩ : IrreducibleCharacter ↥L)
  rw [if_neg (fun heq => hne (by
    simpa using congrArg (fun c : IrreducibleCharacter ↥L => (c : ClassFunction ↥L ℂ)) heq))]
    at hkron
  simpa using hkron

/-- **`Ind^L_H θ ≠ η`** for a `Y`-member `η`, when `θ` is non-linear (`θ(1) ≠ 1`): the induced
degree differs from the `Y`-degree `|W₁|`.  The `X ⊥ Y` distinctness for the irreducible branch. -/
theorem caseB_induce_ne_Yset
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    {θ : IrreducibleCharacter ↥H} (hθ1 : (θ : ClassFunction ↥H ℂ) 1 ≠ 1)
    {η : ClassFunction ↥L ℂ} (hη : η ∈ hyp.Yset) :
    ClassFunction.induce H (θ : ClassFunction ↥H ℂ) ≠ η := by
  intro h
  have h1 : ClassFunction.induce H (θ : ClassFunction ↥H ℂ) (1 : ↥L) = η (1 : ↥L) :=
    by rw [h]
  rw [ClassFunction.induce_apply_one, hyp.index_H_eq_card_W1, hyp.Yset_apply_one hη] at h1
  have hw : (Nat.card hyp.W1 : ℂ) ≠ 0 := by
    have : 0 < Nat.card hyp.W1 := Nat.card_pos
    exact_mod_cast this.ne'
  exact hθ1 (mul_left_cancel₀ hw (by rw [mul_one]; exact h1))

/-- **`(Ind^L_H θ).conj ≠ η`** for `θ` non-linear: the conjugate has the same (real) degree
`|W₁|·θ(1)`, still `≠ |W₁|`. -/
theorem caseB_induce_conj_ne_Yset
    (hyp : SibleyDadeHypothesis G L H) [H.Normal]
    {θ : IrreducibleCharacter ↥H} (hθ1 : (θ : ClassFunction ↥H ℂ) 1 ≠ 1)
    {η : ClassFunction ↥L ℂ} (hη : η ∈ hyp.Yset) :
    (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj ≠ η := by
  intro h
  have h1 : (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj (1 : ↥L) = η (1 : ↥L) :=
    by rw [h]
  obtain ⟨d, _, hd⟩ := irreducibleCharacter_apply_one_eq_pos_natCast θ
  rw [ClassFunction.conj_apply, ClassFunction.induce_apply_one, hyp.index_H_eq_card_W1, hd,
    hyp.Yset_apply_one hη, ← Nat.cast_mul, star_natCast] at h1
  have hw : (Nat.card hyp.W1 : ℂ) ≠ 0 := by
    have : 0 < Nat.card hyp.W1 := Nat.card_pos
    exact_mod_cast this.ne'
  refine hθ1 ?_
  rw [hd]
  have hd1 : (Nat.card hyp.W1 : ℂ) * (d : ℂ) = (Nat.card hyp.W1 : ℂ) * 1 := by
    rw [mul_one, ← Nat.cast_mul]; exact h1
  exact mul_left_cancel₀ hw hd1

omit [Fintype G] [Invertible (Nat.card G : ℂ)] [Fintype ↥L] [Invertible (Nat.card ↥L : ℂ)]
  [Invertible (Nat.card ↥H : ℂ)] in
/-- **(6.8.2.3) positive-weight constituents are non-linear** (the single `hnonlin` dispatch input).
When the central source character `φ` of `W₂ ⊆ ⁅H,H⁆` is **non-trivial**, every constituent `θ`
of `Res^H_{W₂}` (i.e. every `θ` with positive Clifford weight `⟨φ, Res^H_{W₂} θ⟩ > 0`) has
degree `> 1`.

Indeed a degree-one `θ` is trivial on the commutator subgroup `⁅H,H⁆ ⊇ W₂.subgroupOf H`
(`IsIrreducibleCharacter.apply_eq_one_of_mem_commutator_of_apply_one_eq_one`), so
`Res^H_{W₂} θ = 1_{W₂}` and the weight `⟨φ, Res θ⟩ = ⟨φ, 1⟩ = 0` (orthogonality of the
non-trivial irreducible `φ` to the trivial one), contradicting `0 < weight`.

This discharges the structural hypothesis `hnonlin` of `caseB_hirrAnc` (and, through `θ ≠ 1` and
the `Ind θ ≠ η` distinctnesses, of `caseB_irr_bundle`).  The two inputs `W₂ ⊆ ⁅H,H⁆` and `φ ≠ 1`
are the case-(B) data: `φ` is a non-trivial central character (the `X`-side selector) and
`W₂ ≤ ⁅H,H⁆` is the CertainType hypothesis `cert.W2 ≤ ⁅H,H⁆`. -/
theorem caseB_hnonlin [Finite ↥H]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hderiv : W2.subgroupOf H ≤ commutator ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    (hφne : ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ
      ≠ trivialClassFunction ↥(W2.subgroupOf H)) :
    ∀ i : {θ : IrreducibleCharacter ↥H // 0 < constituentWeight hφ' θ},
      (i.val : ClassFunction ↥H ℂ) 1 ≠ 1 := by
  rintro ⟨θ, hweight⟩ h1
  -- Degree-one `θ` is trivial on `⁅H,H⁆ ⊇ W₂.subgroupOf H`, so `Res^H_{W₂} θ = 1`.
  have hrestrict : ClassFunction.restrict (W2.subgroupOf H) (θ : ClassFunction ↥H ℂ)
      = trivialClassFunction ↥(W2.subgroupOf H) := by
    refine ClassFunction.ext fun n => ?_
    rw [ClassFunction.restrict_apply, trivialClassFunction_apply]
    exact θ.2.apply_eq_one_of_mem_commutator_of_apply_one_eq_one h1 (hderiv n.2)
  -- Hence the multiplicity `⟨φ, Res θ⟩ = ⟨φ, 1⟩ = 0` (φ a non-trivial irreducible).
  have hzero : ClassFunction.inner
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ)
      (ClassFunction.restrict (W2.subgroupOf H) (θ : ClassFunction ↥H ℂ)) = 0 := by
    have hne : (⟨ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ, hφ'⟩
        : IrreducibleCharacter ↥(W2.subgroupOf H)) ≠ trivialIrreducibleCharacter _ := by
      intro heq
      apply hφne
      have h := congrArg (fun c : IrreducibleCharacter ↥(W2.subgroupOf H) =>
        (c : ClassFunction ↥(W2.subgroupOf H) ℂ)) heq
      simpa only [IrreducibleCharacter.coe_mk,
        IrreducibleCharacter.coe_trivialIrreducibleCharacter] using h
    rw [hrestrict, ← IrreducibleCharacter.coe_trivialIrreducibleCharacter,
      ← IrreducibleCharacter.coe_mk _ hφ', irreducibleCharacter_inner_eq_ite, if_neg hne]
  exact (constituentWeight_pos_iff hφ' θ).mp hweight hzero

/-- **(6.8.2.3) per-`θ` anchor-vs-constituent orthogonality** (`hirrAnc` of the dispatch).  For a
positive-weight `θ` whose `Ind^L_H θ` is not a column (irreducible branch), the four inner products
of
`η₁, η̄₁` against the induced character and its conjugate vanish: each is a `Y`-member against a
distinct irreducible (`inner_Yset_irr_eq_zero`), distinctness from non-linearity `hnonlin`. -/
theorem caseB_hirrAnc
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (hnonlin : ∀ i : {θ : IrreducibleCharacter ↥H // 0 < constituentWeight hφ' θ},
      (i.val : ClassFunction ↥H ℂ) 1 ≠ 1) :
    ∀ i : {θ : IrreducibleCharacter ↥H // 0 < constituentWeight hφ' θ},
      (∀ χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ, χ₂ ≠ 1 →
          OddOrder.Peterfalvi.S06.columnSum h46 χ₂
            ≠ ClassFunction.induce H (i.val : ClassFunction ↥H ℂ)) →
        ClassFunction.inner η₁ (ClassFunction.induce H (i.val : ClassFunction ↥H ℂ)) = 0
        ∧ ClassFunction.inner η₁ (ClassFunction.induce H (i.val : ClassFunction ↥H ℂ)).conj = 0
        ∧ ClassFunction.inner η₁.conj (ClassFunction.induce H (i.val : ClassFunction ↥H ℂ)) = 0
        ∧ ClassFunction.inner η₁.conj
            (ClassFunction.induce H (i.val : ClassFunction ↥H ℂ)).conj = 0 := by
  have : Fintype ↥H := Fintype.ofFinite _
  intro i hnotcol
  have hθne : (i.val : ClassFunction ↥H ℂ) ≠ trivialClassFunction ↥H := fun heq =>
    hnonlin i (by rw [heq, trivialClassFunction_apply])
  have h1 := caseB_irr_induce_isIrreducible h46 hHK hθne hnotcol
  have hconj := hyp.Yset_closedUnderConjugate hη₁
  refine ⟨inner_Yset_irr_eq_zero hyp hη₁ h1 (caseB_induce_ne_Yset hyp (hnonlin i) hη₁).symm,
    inner_Yset_irr_eq_zero hyp hη₁ h1.conj (caseB_induce_conj_ne_Yset hyp (hnonlin i) hη₁).symm,
    inner_Yset_irr_eq_zero hyp hconj h1 (caseB_induce_ne_Yset hyp (hnonlin i) hconj).symm,
    inner_Yset_irr_eq_zero hyp hconj h1.conj
      (caseB_induce_conj_ne_Yset hyp (hnonlin i) hconj).symm⟩

/-- The `tau1` field of a (5.4) decomposition is unchanged when its `χ`-index is transported along
an equality `χ = χ'` (the field type `IntegralCharacterMap ↥L G` does not mention `χ`).  Used to
read off `tau1 = hyp.tau` through the column-branch index cast of the per-constituent dispatch. -/
theorem charPsiDecomp_eqRec_tau1
    {τ : OddOrder.Peterfalvi.S07.IntegralCharacterMap ↥L G}
    {χ χ' ψ : ClassFunction ↥L ℂ} (h : χ = χ')
    (D : OddOrder.Peterfalvi.S07.CharacterPsiDecomposition τ χ ψ) :
    (h ▸ D : OddOrder.Peterfalvi.S07.CharacterPsiDecomposition τ χ' ψ).tau1 = D.tau1 := by
  cases h; rfl

/-- The `X` image side of a (5.4) decomposition is unchanged when its `χ`-index is transported along
`χ = χ'` (`X : ClassFunction G ℂ` does not mention `χ`).  Used to read the column-branch `X` through
the dispatch index cast for the seam-1 orthogonality `⟨X, cY.ext η₁⟩ = 0`. -/
theorem charPsiDecomp_eqRec_X
    {τ : OddOrder.Peterfalvi.S07.IntegralCharacterMap ↥L G}
    {χ χ' ψ : ClassFunction ↥L ℂ} (h : χ = χ')
    (D : OddOrder.Peterfalvi.S07.CharacterPsiDecomposition τ χ ψ) :
    (h ▸ D : OddOrder.Peterfalvi.S07.CharacterPsiDecomposition τ χ' ψ).X = D.X := by
  cases h; rfl

/-- **(6.8.2.3) integrality of the orthogonal side against a `ZIrr` anchor.**  For any (5.4)
decomposition `D` with image side `D.X ⊥ Y₀` (`hX`), `τ₁`-image a virtual character
`D.tau1 (χ − ψ) ∈ ZIrr G` (`hτmem`) and a virtual-character anchor `Y₀ ∈ ZIrr G` (`hY₀`), the
orthogonal residual pairs integrally with the anchor:
`⟨D.Y, Y₀⟩ = ⟨D.X, Y₀⟩ − ⟨D.tau1 (χ−ψ), Y₀⟩ = −⟨D.tau1 (χ−ψ), Y₀⟩ ∈ ℤ`.
This is the route-independent `hbi` of `per_phi_anchored_image` (`Y = X − τ₁(χ−ψ)` from
`tau1_image`;
`⟨τ₁(χ−ψ), Y₀⟩ ∈ ℤ` by `inner_mem_ZIrr_int`). -/
theorem psiDecomp_Y_inner_int {L' G' : Type*} [Group L'] [Group G'] [Fintype L'] [Fintype G']
    [Invertible (Nat.card L' : ℂ)] [Invertible (Nat.card G' : ℂ)]
    {τ : OddOrder.Peterfalvi.S07.IntegralCharacterMap L' G'} {χ ψ : ClassFunction L' ℂ}
    (D : OddOrder.Peterfalvi.S07.CharacterPsiDecomposition τ χ ψ)
    {Y₀ : ClassFunction G' ℂ}
    (hX : ClassFunction.inner D.X Y₀ = 0)
    (hτmem : D.tau1 (χ - ψ) ∈ ZIrr G')
    (hY₀ : Y₀ ∈ ZIrr G') :
    ∃ n : ℤ, ClassFunction.inner D.Y Y₀ = (n : ℂ) := by
  obtain ⟨n, hn⟩ := ClassFunction.inner_mem_ZIrr_int hτmem hY₀
  refine ⟨-n, ?_⟩
  have hYeq : D.Y = D.X - D.tau1 (χ - ψ) := by rw [D.tau1_image]; abel
  rw [hYeq, ClassFunction.inner_sub_left, hX, hn]
  push_cast; ring

/-- The per-`θ` **column-branch bundle** (mixed dispatch).  For each nontrivial column character
`χ₂` whose certain-type column `μ_j = columnSum h46 χ₂` equals `Ind^L_H θ`, the structural inputs of
`columnDecompositionTau`: equal column degrees, the `τ`-image agreement on `μ_j − μ̄_j`, the two
`H^#`-supports, the `ZIrr`-membership of `(μ_j − a·η₁)^{hyp.tau}`, and the three (5.2.c)
orthogonality
scalars.  Extracted as an abbreviation so the dispatch lemmas share one bundle type with
`caseB_constituentDecomposition` (definitionally the same `hcol` Prop). -/
abbrev CaseBColBundle (hyp : SibleyDadeHypothesis G L H)
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    [Fintype (OddOrder.Peterfalvi.S06.ticVdiff h46).W]
    [Invertible (Nat.card (OddOrder.Peterfalvi.S06.ticVdiff h46).W : ℂ)]
    (θ : IrreducibleCharacter ↥H) (η₁ : ClassFunction ↥L ℂ) (a : ℕ) : Prop :=
  ∀ χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ, χ₂ ≠ 1 →
    OddOrder.Peterfalvi.S06.columnSum h46 χ₂ = ClassFunction.induce H (θ : ClassFunction ↥H ℂ) →
    (∑ i, ((h46.columnFamily χ₂).mu i : ClassFunction ↥L ℂ) 1
        = ∑ i, ((h46.columnFamily χ₂⁻¹).mu i : ClassFunction ↥L ℂ) 1)
    ∧ (hyp.tau (OddOrder.Peterfalvi.S06.columnSum h46 χ₂
          - (OddOrder.Peterfalvi.S06.columnSum h46 χ₂).conj)
        = OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau
          (OddOrder.Peterfalvi.S06.columnSum h46 χ₂
            - (OddOrder.Peterfalvi.S06.columnSum h46 χ₂).conj))
    ∧ (∀ s ∈ ({OddOrder.Peterfalvi.S06.columnSum h46 χ₂
          - (OddOrder.Peterfalvi.S06.columnSum h46 χ₂).conj,
          OddOrder.Peterfalvi.S06.columnSum h46 χ₂ - a • η₁} : Set (ClassFunction ↥L ℂ)),
        s.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup (sharpImage H) L)
    ∧ (hyp.tau (OddOrder.Peterfalvi.S06.columnSum h46 χ₂ - a • η₁) ∈ ZIrr G)
    ∧ (ClassFunction.inner (OddOrder.Peterfalvi.S06.columnSum h46 χ₂)
        (a • η₁ : ClassFunction ↥L ℂ) = 0)
    ∧ (ClassFunction.inner (OddOrder.Peterfalvi.S06.columnSum h46 χ₂).conj
        (a • η₁ : ClassFunction ↥L ℂ) = 0)

/-- The per-`θ` **irreducible-branch bundle** (mixed dispatch).  When no nontrivial column equals
`Ind^L_H θ`, the structural inputs of `irreducibleDecompositionTau` (=
`decompositionDaFromDadeOfDiff
hyp.dade`): irreducibility, non-realness, the two `H^#`-supports, the `ZIrr`-membership of
`(Ind^L_H θ − a·η₁)^{hyp.tau}`, and the (5.2.c) orthogonality scalars.  Shared with
`caseB_constituentDecomposition` (definitionally the same `hirr` Prop). -/
abbrev CaseBIrrBundle (hyp : SibleyDadeHypothesis G L H)
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    [Fintype (OddOrder.Peterfalvi.S06.ticVdiff h46).W]
    [Invertible (Nat.card (OddOrder.Peterfalvi.S06.ticVdiff h46).W : ℂ)]
    (θ : IrreducibleCharacter ↥H) (η₁ : ClassFunction ↥L ℂ) (a : ℕ) : Prop :=
  (∀ χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ, χ₂ ≠ 1 →
      OddOrder.Peterfalvi.S06.columnSum h46 χ₂
        ≠ ClassFunction.induce H (θ : ClassFunction ↥H ℂ)) →
    IsIrreducibleCharacter (ClassFunction.induce H (θ : ClassFunction ↥H ℂ))
    ∧ (¬ ClassFunction.IsReal (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)))
    ∧ (((ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj
          - ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).support ⊆
        OddOrder.Peterfalvi.S04.supportInSubgroup (sharpImage H) L)
    ∧ ((ClassFunction.induce H (θ : ClassFunction ↥H ℂ) - a • η₁).support ⊆
        OddOrder.Peterfalvi.S04.supportInSubgroup (sharpImage H) L)
    ∧ (hyp.tau (ClassFunction.induce H (θ : ClassFunction ↥H ℂ) - a • η₁) ∈ ZIrr G)
    ∧ (ClassFunction.inner (ClassFunction.induce H (θ : ClassFunction ↥H ℂ))
        (a • η₁ : ClassFunction ↥L ℂ) = 0)
    ∧ (ClassFunction.inner (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj
        (a • η₁ : ClassFunction ↥L ℂ) = 0)
    ∧ (ClassFunction.inner (ClassFunction.induce H (θ : ClassFunction ↥H ℂ))
        (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj = 0)

/-- **(6.8.2.3) full column-branch bundle.**  Assembles the entire `CaseBColBundle` for the dispatch
weight `aθ = constituentWeight hφ' θ`: all six conjuncts hold. The two degree-coupled ones
(`hSdiff`,
`htau1_mema`) are supplied with the degree match `caseB_column_degree_match` (needs the source to
occur in `θ`, `0 < constituentWeight`); the rest are unconditional.  This is the column-branch input
`hcol` of the mixed per-`φ` dispatch (`caseB_per_phi_anchored_fromYset`). -/
theorem caseB_column_bundle
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    (hW1 : h46.W1 = hyp.W1)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    [Fintype (OddOrder.Peterfalvi.S06.ticVdiff h46).W]
    [Invertible (Nat.card (OddOrder.Peterfalvi.S06.ticVdiff h46).W : ℂ)]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {θ : IrreducibleCharacter ↥H} (hweight : 0 < constituentWeight hφ' θ)
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset) :
    CaseBColBundle hyp h46 θ η₁ (constituentWeight hφ' θ) := by
  have : Fintype ↥H := Fintype.ofFinite _
  intro χ₂ hχ₂ hcoleq
  have h1 := caseB_column_degree_match hyp h46 hW2H hcen hφ' hweight hη₁ hcoleq
  exact ⟨(OddOrder.Peterfalvi.S06.columnSum_inv_apply_one h46 χ₂).symm,
    caseB_column_mapagree hyp h46 hχ₂,
    caseB_column_hSdiff hyp h46 hHK hη₁ hχ₂ (constituentWeight hφ' θ) h1,
    caseB_column_htau1_mema hyp h46 hHK hη₁ χ₂ (constituentWeight hφ' θ) h1,
    caseB_column_orthogonal_Yset hyp h46 hW1 hη₁ χ₂ (constituentWeight hφ' θ),
    caseB_column_conj_orthogonal_Yset hyp h46 hW1 hη₁ χ₂ (constituentWeight hφ' θ)⟩

/-- **(6.8.2.3) full irreducible-branch bundle.**  Assembles the entire `CaseBIrrBundle` for the
dispatch weight `aθ = constituentWeight hφ' θ`.  Given the structural inputs `θ ≠ 1` (for
irreducibility) and the `X ⊥ Y` distinctnesses of the induced character and its conjugate from the
anchor, all eight conjuncts are discharged (irreducibility, non-realness, the two supports, `ZIrr`,
the two anchor orthogonalities, self-conjugate orthogonality).  This is the irreducible-branch input
`hirr` of the mixed per-`φ` dispatch. -/
theorem caseB_irr_bundle
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    [Fintype (OddOrder.Peterfalvi.S06.ticVdiff h46).W]
    [Invertible (Nat.card (OddOrder.Peterfalvi.S06.ticVdiff h46).W : ℂ)]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {θ : IrreducibleCharacter ↥H} (hweight : 0 < constituentWeight hφ' θ)
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (hθne : (θ : ClassFunction ↥H ℂ) ≠ trivialClassFunction ↥H)
    (hneη : ClassFunction.induce H (θ : ClassFunction ↥H ℂ) ≠ η₁)
    (hneη' : (ClassFunction.induce H (θ : ClassFunction ↥H ℂ)).conj ≠ η₁) :
    CaseBIrrBundle hyp h46 θ η₁ (constituentWeight hφ' θ) := by
  have : Fintype ↥H := Fintype.ofFinite _
  intro hnotcol
  have h1 := caseB_irr_induce_isIrreducible h46 hHK hθne hnotcol
  exact ⟨h1, caseB_irr_nonreal hyp h1, caseB_irr_conj_diff_support hyp θ,
    caseB_irr_sub_smul_support hyp hW2H hcen hφ' hweight hη₁,
    caseB_irr_htau1_mema hyp hW2H hcen hφ' hweight hη₁,
    caseB_irr_orthogonal_Yset hyp h1 hη₁ hneη (constituentWeight hφ' θ),
    caseB_irr_conj_orthogonal_Yset hyp h1 hη₁ hneη' (constituentWeight hφ' θ),
    caseB_irr_conj_inner hyp h1⟩

/-- **(6.8.2.3) column dispatch bundle over the whole positive-weight family** (`hcol` input of
`caseB_per_phi_anchored_fromYset`).  Maps `caseB_column_bundle` over every positive-weight
constituent `i`; the column branch needs no non-linearity hypothesis (it is gated on the column
witness equation `columnSum χ₂ = Ind^L_H θ`). -/
theorem caseB_hcol
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    (hW1 : h46.W1 = hyp.W1)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    [Fintype (OddOrder.Peterfalvi.S06.ticVdiff h46).W]
    [Invertible (Nat.card (OddOrder.Peterfalvi.S06.ticVdiff h46).W : ℂ)]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset) :
    ∀ i : {θ : IrreducibleCharacter ↥H // 0 < constituentWeight hφ' θ},
      CaseBColBundle hyp h46 i.val η₁ (constituentWeight hφ' i.val) :=
  fun i => caseB_column_bundle hyp h46 hHK hW1 hW2H hcen hφ' i.property hη₁

/-- **(6.8.2.3) irreducible dispatch bundle over the whole positive-weight family** (`hirr` input of
`caseB_per_phi_anchored_fromYset`).  Maps `caseB_irr_bundle` over every positive-weight constituent
`i`, supplying the three structural inputs `θ ≠ 1`, `Ind θ ≠ η₁`, `(Ind θ)‾ ≠ η₁` from the single
non-linearity hypothesis `hnonlin` (`caseB_hnonlin`): `θ ≠ 1` since `θ(1) ≠ 1`, and the two `≠ η₁`
distinctnesses by `caseB_induce_ne_Yset` / `caseB_induce_conj_ne_Yset` (degree mismatch). -/
theorem caseB_hirr
    (hyp : SibleyDadeHypothesis G L H) [H.Normal] [Finite ↥H]
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 (sharpImage H) L) (hHK : h46.K = H)
    [NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥h46.K : ℂ)]
    [Fintype ↥(h46.W1 ⊔ h46.W2)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
    [Fintype (OddOrder.Peterfalvi.S06.ticVdiff h46).W]
    [Invertible (Nat.card (OddOrder.Peterfalvi.S06.ticVdiff h46).W : ℂ)]
    {W2 : Subgroup ↥L} (hW2H : W2 ≤ H) [Fintype ↥(W2.subgroupOf H)]
    [Invertible (Nat.card ↥(W2.subgroupOf H) : ℂ)]
    (hcen : W2.subgroupOf H ≤ Subgroup.center ↥H)
    {φ : ClassFunction ↥W2 ℂ}
    (hφ' : IsIrreducibleCharacter
      (ClassFunction.compHom (Subgroup.subgroupOfEquivOfLe hW2H).toMonoidHom φ))
    {η₁ : ClassFunction ↥L ℂ} (hη₁ : η₁ ∈ hyp.Yset)
    (hnonlin : ∀ i : {θ : IrreducibleCharacter ↥H // 0 < constituentWeight hφ' θ},
      (i.val : ClassFunction ↥H ℂ) 1 ≠ 1) :
    ∀ i : {θ : IrreducibleCharacter ↥H // 0 < constituentWeight hφ' θ},
      CaseBIrrBundle hyp h46 i.val η₁ (constituentWeight hφ' i.val) :=
  fun i => caseB_irr_bundle hyp h46 hHK hW2H hcen hφ' i.property hη₁
    (fun heq => hnonlin i (by rw [heq, trivialClassFunction_apply]))
    (caseB_induce_ne_Yset hyp (hnonlin i) hη₁)
    (caseB_induce_conj_ne_Yset hyp (hnonlin i) hη₁)

end OddOrder.Peterfalvi.S08
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_CaseBAssembly/BranchBundles.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