Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Peterfalvi (10.7): the cross-isometry package at the canonical pair

Definition
OddOrder_Peterfalvi_S12_TypeIICrossIsometryPair

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

feit-thompsonfinite-groupsgroup-theoryishida-odd-order

Formalizes the cross-isometry package of Peterfalvi (10.7) at the canonical pair (Peterfalvi, Character Theory for the Odd Order Theorem, §10), including support facts for the induced family and the core-order coprimality for a nonconjugate pair.

Main declarations: Hypothesis.mem_zSpan_inducedFamily_support_sharp_derived (ℤ[𝒮]-combinations vanishing at 1 are supported in (M′)^#); Hypothesis.supportInSubgroup_sharp_derived_subset_A0 ((M′)^# ⊆ A₀(M)); typeP_core_order_coprime (The core-order coprimality for a nonconjugate pair); typeP_escaping_centralizer_not_le_typeII (Peterfalvi (8.13.b/c4) at the canonical pair, escape-landing exclusion).

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/S12_TypeIICrossIsometryPair.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_S12_TypeIIGridTranspose
import Definitions.Def_OddOrder_Peterfalvi_S09_CrossOrthogonality
import Definitions.Def_OddOrder_Peterfalvi_S12_TypeIIFrobenius
import Definitions.Def_OddOrder_Isaacs_Ch06_FrobeniusActions_FrobeniusGroupQuotient
import Definitions.Def_OddOrder_Peterfalvi_S12_CrossIsometrySetup

namespace OddOrder.Peterfalvi.S12

open OddOrder.RepresentationTheory

open OddOrder.GroupTheory

variable {G : Type*} [Group G]

open scoped FiniteInduce in
/-- **`ℤ[𝒮]`-combinations vanishing at `1` are supported in `(M′)^#`** (the `zchar_on`/`defA1`
step of Coq `Frob_der1_type2`): every member of `inducedFamily M` is induced from the normal
`M′`, hence vanishes off `M′` (a ℤ-linear property, so every lattice element does); vanishing
at `1` then pins the support inside the sharp set.  This is the `M`-side support refinement
`supp ⊆ A₁(M) = (M′)^#` feeding the (8.18.b) disjointness (`A₀`-support alone would not do:
the pair's `A₀`-sets share the conjugates of `V`). -/
theorem Hypothesis.mem_zSpan_inducedFamily_support_sharp_derived [Finite G] {M : Subgroup G}
    (hyp : Hypothesis M) {φ : ClassFunction ↥M ℂ}
    (hφ : φ ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥M) (inducedFamily M))
    (hφ1 : φ 1 = 0) :
    φ.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (sharpSubgroup (derivedInG M)) M := by
  have := hyp.finiteG
  classical
  have hKcomm : (derivedInG M).subgroupOf M = commutator ↥M := by
    rw [derivedInG, Subgroup.subgroupOf,
      Subgroup.comap_map_eq_self_of_injective M.subtype_injective]
  have hKnormal : ((derivedInG M).subgroupOf M).Normal := by rw [hKcomm]; infer_instance
  -- every lattice element vanishes off `M′` (span induction, the property is ℤ-linear)
  have hvanish : ∀ χ ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥M) (inducedFamily M),
      ∀ w : ↥M, w ∉ (derivedInG M).subgroupOf M → χ w = 0 := by
    intro χ hχ
    induction hχ using Submodule.span_induction with
    | mem x hx =>
        intro w hw
        obtain ⟨θ, -, hxeq⟩ := hx
        rw [hxeq]
        exact ClassFunction.induce_eq_zero_of_not_mem_normal _ hw
    | zero => intro w _; rfl
    | add x y _ _ ihx ihy =>
        intro w hw
        rw [ClassFunction.add_apply, ihx w hw, ihy w hw, add_zero]
    | smul a x _ ih =>
        intro w hw
        rw [← Int.cast_smul_eq_zsmul ℂ a x, ClassFunction.smul_apply, ih w hw, mul_zero]
  intro z hz
  rw [ClassFunction.mem_support] at hz
  have hzK : z ∈ (derivedInG M).subgroupOf M := by
    by_contra hzK
    exact hz (hvanish φ hφ z hzK)
  have hz1 : z ≠ 1 := by rintro rfl; exact hz hφ1
  rw [OddOrder.Peterfalvi.S04.mem_supportInSubgroup]
  exact ⟨Subgroup.mem_subgroupOf.mp hzK,
    fun h0 => hz1 (Subtype.ext (Set.mem_singleton_iff.mp h0))⟩

open scoped FiniteInduce in
/-- **`(M′)^# ⊆ A₀(M)`**: every nonidentity element of the derived subgroup centralizes
itself (a nonidentity element of `M`), so it lies in the `A(M)`-disjunct of `typePA0` —
the same witness as `inducedFamily_sub_support`'s tail.  Composes the sharp-support
refinement back into the `A₀`-supported lattice (`extends_on_supported`). -/
theorem Hypothesis.supportInSubgroup_sharp_derived_subset_A0 [Finite G] {M : Subgroup G}
    (hyp : Hypothesis M) :
    OddOrder.Peterfalvi.S04.supportInSubgroup (sharpSubgroup (derivedInG M)) M
      ⊆ hyp.A0 := by
  intro z hz
  rw [OddOrder.Peterfalvi.S04.mem_supportInSubgroup] at hz
  obtain ⟨hzM', hz1⟩ := hz
  change (z : G) ∈ typePA0 M hyp.typeP
  unfold typePA0
  rw [Set.mem_union]
  left
  exact ⟨hzM', hz1, (z : G), ⟨z.2, hz1⟩, Subgroup.mem_centralizer_singleton_iff.mpr rfl⟩

open scoped Classical FiniteInduce Pointwise in
/-- **The core-order coprimality for a nonconjugate pair** (the `coxTs`-step of Coq
`FT_Dade_support_disjoint` `part_a2`): the order of an `(M′)^#`-point of the type-`P₁` `M` is
coprime to `|S_F| = |M_σ(S)|` of any maximal `S` not conjugate to `M`.

No (8.17.a) partition is needed: the type-`P₁` structure collapses the FTcore to the σ-Hall —
`M′ = M_σ(M)` (Coq `typePfacts`, `isTypeP1_derivedInG_eq_Msigma`) — so `|a|` is a `σ(M)`-number,
and `σ(M) ∩ σ(S) = ∅` for a nonconjugate pair (BG Theorem 13.9,
`sigma_disjoint_of_nonconjugate`); both `M_σ`'s are Hall, so a common prime divisor would lie in
the empty intersection.

⚠ Stated for an arbitrary nonconjugate maximal `S`, not for the canonical `Section16MaximalPairCore`
(issue 1044 着手順 5).  Neither `S`'s type nor its cyclic data enters: the old signature carried
`data`/`hSW1`/`hSW2`/`hKstar`/`ha0` and used none of them. -/
theorem typeP_core_order_coprime [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M S : Subgroup G}
    (hyp : Hypothesis M) (hSmax : S ∈ maximalSubgroups G)
    (hnc : ¬ ∃ g : G, MulAut.conj g • M = S)
    {a : G} (haM' : a ∈ sharpSubgroup (derivedInG M)) :
    Nat.Coprime (orderOf a) (Nat.card (OddOrder.BG.Ch3.S10.Msigma S)) := by
  classical
  -- the type-`P₁` `M` has `M′ = M_σ(M)` (BG `typePfacts`)
  have hP : OddOrder.BG.Ch4.S14.IsTypeP M :=
    OddOrder.BG.Ch4.S16.isTypeP_of_isTypeNonI hG hyp.maximal
      (Or.inr hyp.type_alt)
  have hP1 : OddOrder.BG.Ch4.S14.IsTypeP1 M := by
    refine ⟨hP, ?_⟩
    by_contra hne
    exact not_isTypeP2_of_isTypeIII_or_IV_or_V hG hyp.maximal hyp.type_alt ⟨hP, hne⟩
  have hM'σ : derivedInG M = OddOrder.BG.Ch3.S10.Msigma M :=
    OddOrder.BG.Ch4.S16.isTypeP1_derivedInG_eq_Msigma hG hyp.maximal hP1
  -- `σ`-disjointness of the nonconjugate pair (BG Theorem 13.9)
  have hdisj : Disjoint (OddOrder.BG.Ch3.S10.sigma M) (OddOrder.BG.Ch3.S10.sigma S) :=
    OddOrder.BG.Ch3.S13.sigma_disjoint_of_nonconjugate hG hyp.maximal hSmax hnc
  -- a common prime would lie in `σ(M) ∩ σ(S) = ∅`
  by_contra hne
  obtain ⟨q, hqp, hqdvd⟩ := Nat.exists_prime_and_dvd hne
  have hqa : q ∣ Nat.card (OddOrder.BG.Ch3.S10.Msigma M) :=
    (hqdvd.trans (Nat.gcd_dvd_left _ _)).trans
      (Subgroup.orderOf_dvd_natCard _ (hM'σ ▸ haM'.1))
  have hqσM : q ∈ OddOrder.BG.Ch3.S10.sigma M :=
    (OddOrder.BG.Ch3.S10.Msigma_isHall hG hyp.maximal).1 q
      (Nat.mem_primeFactors.mpr ⟨hqp, hqa, Nat.card_pos.ne'⟩)
  have hqb : q ∣ Nat.card (OddOrder.BG.Ch3.S10.Msigma S) :=
    hqdvd.trans (Nat.gcd_dvd_right _ _)
  have hqσS : q ∈ OddOrder.BG.Ch3.S10.sigma S :=
    (OddOrder.BG.Ch3.S10.Msigma_isHall hG hSmax).1 q
      (Nat.mem_primeFactors.mpr ⟨hqp, hqb, Nat.card_pos.ne'⟩)
  exact (Set.disjoint_left.mp hdisj hqσM) hqσS

set_option linter.unusedVariables false in
open scoped Classical FiniteInduce Pointwise in
/-- **Peterfalvi (8.13.b/c4) at the canonical pair, escape-landing exclusion**: an
`(M′)^#`-point of the type-`P₁` `M` whose centralizer **escapes `M`** cannot land in a
conjugate of any type-II maximal `S`.

Coq `FTsupport_facts` (b)+(c4) + the (10.7) consumer's `notFrobM`: the escaping `A₀`-point
has a *unique* supporting maximal `N[a]` (8.13.b), here `= S^g` by `C_G(a) ≤ S^g`; and
(8.13.c4) says a type-II supporter forces `M` to be a Frobenius group with kernel `M_F` —
impossible for the type-`P₁` `M`.

Proven from **BG Theorem D(4)** (`theoremD_msigma_conjugacy_and_centralizers`): the
type-`P₁` collapse `M′ = M_σ` puts `a ∈ M_σ(M)^#`, D(4) attaches to the escaping `a` a
supporting maximal `N₀` with the package clause `IsTypeP2 N₀ → IsTypeF M ∧ …`; the given
uniqueness pins `N₀ = S^g`, a conjugate of the type-II `S`, hence type `P₂` (Proposition
16.1(b), `proposition_type_classification`); the fired clause makes `M` type `F` = type I —
contradicting type III/IV/V (`not_isTypeI_of_isTypeNonI`).

⚠ Stated for an arbitrary type-II maximal `S`, not for the canonical `Section16MaximalPairCore`
(issue 1044 着手順 5): the old signature carried `mp`/`data`/`hSW1`/`hSW2`/`hKstar`/`ha0` and used
none of them beyond `S`'s type. -/
theorem typeP_escaping_centralizer_not_le_typeII [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M S : Subgroup G}
    (hyp : Hypothesis M) (hSII : IsTypeII S)
    {a : G} (haM' : a ∈ sharpSubgroup (derivedInG M))
    (hesc : ¬ Subgroup.centralizer ({a} : Set G) ≤ M)
    {g : G} (huniq : ∀ N ∈ maximalSubgroupsContaining
      (Subgroup.centralizer ({a} : Set G)), N = MulAut.conj g • S) :
    False := by
  classical
  -- the type-`P₁` collapse puts `a` in `M_σ(M)^#`
  have hP : OddOrder.BG.Ch4.S14.IsTypeP M :=
    OddOrder.BG.Ch4.S16.isTypeP_of_isTypeNonI hG hyp.maximal (Or.inr hyp.type_alt)
  have hP1 : OddOrder.BG.Ch4.S14.IsTypeP1 M := by
    refine ⟨hP, ?_⟩
    by_contra hne
    exact not_isTypeP2_of_isTypeIII_or_IV_or_V hG hyp.maximal hyp.type_alt ⟨hP, hne⟩
  have hM'σ : derivedInG M = OddOrder.BG.Ch3.S10.Msigma M :=
    OddOrder.BG.Ch4.S16.isTypeP1_derivedInG_eq_Msigma hG hyp.maximal hP1
  have haσ : a ∈ OddOrder.BG.Ch4.S14.sigmaSharp M := by
    change a ∈ sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma M)
    rw [← hM'σ]
    exact haM'
  -- BG Theorem D(4): the unique supporting maximal with the type-`F`/`P₂` package
  obtain ⟨-, -, -, hD4⟩ :=
    OddOrder.BG.Ch4.S16.theoremD_msigma_conjugacy_and_centralizers hG hyp.maximal
  obtain ⟨R, -, N₀, hQ, -⟩ := hD4 a haσ hesc
  obtain ⟨hN₀mem, -, -, -, -, -, hP2clause⟩ := hQ
  have hN₀eq : N₀ = MulAut.conj g • S := huniq N₀ hN₀mem
  have hN₀max : N₀ ∈ maximalSubgroups G := (mem_maximalSubgroupsContaining.mp hN₀mem).1
  -- `N₀` is a conjugate of the type-II `S`, hence type `P₂` (Proposition 16.1(b))
  have hN₀II : IsTypeII N₀ := by
    rw [hN₀eq]
    exact isTypeII_pointwise_smul (MulAut.conj g) hSII
  obtain ⟨-, hIIiff, -⟩ :=
    OddOrder.BG.Ch4.S16.proposition_type_classification hG hN₀max
  -- fire the (c4)-clause: `M` is type `F` = type I — contradicting type III/IV/V
  obtain ⟨hFM, -, -⟩ := hP2clause (hIIiff.mp hN₀II)
  obtain ⟨hIiffM, -, -⟩ :=
    OddOrder.BG.Ch4.S16.proposition_type_classification hG hyp.maximal
  exact OddOrder.BG.Ch4.S16.not_isTypeI_of_isTypeNonI hG hyp.maximal
    (Or.inr hyp.type_alt) (hIiffM.mpr hFM)

open scoped Classical FiniteInduce Pointwise in
/-- **Peterfalvi (8.18.b), bare base-point disjointness at the canonical pair — the
landing analysis** (Coq `part_a2` of `FT_Dade_support_disjoint`): no `(M′)^#`-point of the
type-`P₁` `M = mp.T` is conjugate to an `A(S)`-point of the type-II member `mp.S`.

Suppose `c·a·c⁻¹ = b ∈ A(S)`.  Then `b` is a `(κ(S) ∪ σ(S))′`-element of `S`: its `σ`-part
dies by the (8.17.a) core-order coprimality (`typeP_core_order_coprime`), its
`κ`-part because `b ∈ S′` and `S′` complements the cyclic `κ(S)`-Hall `W₁`
((8.4)/`typeP_derivedInG_isComplement_kappaHall`).  The `A(S)`-witness `u ∈ M_σ(S)^#`
commutes with `b`, so BG Lemma 15.1(c) (`uniqueMaximal_of_kappaSigmaCompl_element`, the
type-generic (8.12)) pins `ℳ(C_G(b)) = {S}`.  If `C_G(a) ≤ M`, then
`C_G(b) = C_G(a)^c ≤ M^c` forces `M^c = S` — contradicting the pair's nonconjugacy
(`mp.S_T_not_conj`).  Otherwise `a` escapes `M` while `C_G(a) ≤ S^{c⁻¹}` — killed by the
(8.13.c4) exclusion (`typeP_escaping_centralizer_not_le_typeII`). -/
theorem typeP_pair_base_bare_not_isConj [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G}
    (hyp : Hypothesis M) {mp : Section16MaximalPairCore G}
    (hT : mp.T = M) (_hKstar : mp.Kstar = hyp.typeP.W1)
    {data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup mp.S}
    (_hSW1 : data.typeP.W1 = mp.K) (_hSW2 : data.typeP.W2 = mp.Kstar)
    {a : G} (haM' : a ∈ sharpSubgroup (derivedInG M)) (_ha0 : a ∈ typePA0 M hyp.typeP)
    {b : G} (hbS : b ∈ centralizerSupport
      (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S)) (derivedInG mp.S)) :
    ¬ IsConj a b := by
  classical
  intro hconj
  have : Group.IsSolvable ↥mp.S := hG.isSolvable_of_mem_maximalSubgroups mp.S_maximal
  obtain ⟨c, hc⟩ := isConj_iff.mp hconj
  obtain ⟨hbM', hb1, u, huS, hbu⟩ := hbS
  have ha1 : a ≠ 1 := fun h0 => haM'.2 (Set.mem_singleton_iff.mpr h0)
  have hbne : b ≠ 1 := fun h0 => hb1 h0
  have hordeq : orderOf b = orderOf a := by
    rw [← hc]
    exact orderOf_injective (MulAut.conj c).toMonoidHom (MulAut.conj c).injective a
  -- `b` is a `(κ(S) ∪ σ(S))′`-element
  have hcopMσ := typeP_core_order_coprime hG hyp mp.S_maximal
    (by
      rintro ⟨g, hg⟩
      exact mp.S_T_not_conj ⟨g⁻¹, by
        rw [← hg, ← mul_smul, ← map_mul, inv_mul_cancel, map_one, one_smul, hT]⟩)
    haM'
  have hSP : OddOrder.BG.Ch4.S14.IsTypeP mp.S :=
    OddOrder.BG.Ch4.S14.isTypeP_of_isTypeP2 mp.S_typeP2
  have hW1hall := typePData_W1_isHallSubgroup_kappa hG mp.S_maximal hSP data.typeP
  have : IsCyclic ↥data.typeP.W1 := data.typeP.W1_cyclic
  have hcompl := OddOrder.BG.Ch4.S14.typeP_derivedInG_isComplement_kappaHall hG
    mp.S_maximal hSP data.typeP.W1_le hW1hall
  have hcardS' : Nat.card ((derivedInG mp.S).subgroupOf mp.S) = Nat.card (derivedInG mp.S) :=
    Nat.card_congr (Subgroup.subgroupOfEquivOfLe
      (OddOrder.Peterfalvi.S11.derivedInG_le_self mp.S)).toEquiv
  have hyπ : OddOrder.GroupTheory.IsPiElement
      ((OddOrder.BG.Ch4.S14.kappa mp.S ∪ OddOrder.BG.Ch3.S10.sigma mp.S)ᶜ) b := by
    intro q hq
    obtain ⟨hqp, hqdvd, -⟩ := Nat.mem_primeFactors.mp hq
    simp only [Set.mem_compl_iff, Set.mem_union, not_or]
    constructor
    · -- `q ∉ κ(S)`: `q ∣ |b| ∣ |S′| = index(W₁)`, and the Hall `W₁` has `κ`-free index
      intro hqκ
      have hbord : orderOf b ∣ Nat.card (derivedInG mp.S) :=
        Subgroup.orderOf_dvd_natCard _ hbM'
      have hqidx : q ∣ ((data.typeP.W1).subgroupOf mp.S).index := by
        rw [hcompl.index_eq_card, hcardS']
        exact hqdvd.trans hbord
      exact hW1hall.2 q (Nat.mem_primeFactors.mpr
        ⟨hqp, hqidx, Subgroup.index_ne_zero_of_finite⟩) hqκ
    · -- `q ∉ σ(S)`: `σ`-primes divide `|M_σ(S)|`, coprime to `|a| = |b|` by (8.17.a)
      intro hqσ
      have hHall := OddOrder.BG.Ch3.S10.Msigma_isHall hG mp.S_maximal
      have hqnidx : ¬ q ∣ (OddOrder.BG.Ch3.S10.Msigma mp.S).index := fun hdvd =>
        hHall.2 q (Nat.mem_primeFactors.mpr
          ⟨hqp, hdvd, Subgroup.index_ne_zero_of_finite⟩) hqσ
      have hqG : q ∣ Nat.card G := hqdvd.trans (orderOf_dvd_natCard b)
      have hqMσ : q ∣ Nat.card (OddOrder.BG.Ch3.S10.Msigma mp.S) := by
        rcases (Nat.Prime.dvd_mul hqp).mp
          ((Subgroup.card_mul_index (OddOrder.BG.Ch3.S10.Msigma mp.S)) ▸ hqG) with h | h
        · exact h
        · exact absurd h hqnidx
      have hgcd : q ∣ Nat.gcd (orderOf a) (Nat.card (OddOrder.BG.Ch3.S10.Msigma mp.S)) :=
        Nat.dvd_gcd (hordeq ▸ hqdvd) hqMσ
      rw [Nat.Coprime.gcd_eq_one hcopMσ] at hgcd
      exact hqp.one_lt.ne' (Nat.dvd_one.mp hgcd)
  -- the `(κ∪σ)′`-Hall of `S` (Hall E in the solvable `S`)
  obtain ⟨U', hU'⟩ := OddOrder.Isaacs.Ch03.hall_E_exists (G := ↥mp.S)
    ((OddOrder.BG.Ch4.S14.kappa mp.S ∪ OddOrder.BG.Ch3.S10.sigma mp.S)ᶜ)
  have hUeq : (U'.map mp.S.subtype).subgroupOf mp.S = U' :=
    Subgroup.comap_map_eq_self_of_injective mp.S.subtype_injective U'
  have hUhall : OddOrder.Isaacs.Ch03.IsHallSubgroup
      ((OddOrder.BG.Ch4.S14.kappa mp.S ∪ OddOrder.BG.Ch3.S10.sigma mp.S)ᶜ)
      ((U'.map mp.S.subtype).subgroupOf mp.S) := by
    rw [hUeq]
    exact hU'
  -- BG 15.1(c): `ℳ(C_G(b)) = {S}`
  have hbmem : b ∈ mp.S := OddOrder.Peterfalvi.S11.derivedInG_le_self mp.S hbM'
  have hCne : OddOrder.BG.Ch3.S10.Msigma mp.S ⊓
      Subgroup.centralizer ({b} : Set G) ≠ ⊥ := by
    intro hbot
    have humem : u ∈ OddOrder.BG.Ch3.S10.Msigma mp.S ⊓
        Subgroup.centralizer ({b} : Set G) := by
      refine Subgroup.mem_inf.mpr ⟨huS.1, Subgroup.mem_centralizer_iff.mpr fun z hz => ?_⟩
      rw [Set.mem_singleton_iff] at hz
      subst hz
      exact Subgroup.mem_centralizer_singleton_iff.mp hbu
    rw [hbot] at humem
    exact huS.2 (Set.mem_singleton_iff.mpr (Subgroup.mem_bot.mp humem))
  have hB := OddOrder.BG.Ch4.S16.uniqueMaximal_of_kappaSigmaCompl_element hG mp.S_maximal
    (Subgroup.map_subtype_le U') hUhall hbmem hbne hyπ hCne
  -- transport: `C_G(b) = C_G(a)^c`
  have hCconj : Subgroup.centralizer ({b} : Set G)
      = MulAut.conj c • Subgroup.centralizer ({a} : Set G) := by
    rw [← hc]
    exact (OddOrder.Peterfalvi.S10.conj_smul_centralizer_singleton' c a).symm
  by_cases hCa : Subgroup.centralizer ({a} : Set G) ≤ M
  · -- `C_G(b) ≤ M^c`, so `M^c = S` — contradicting the pair's nonconjugacy
    have hMc : MulAut.conj c • M ∈ maximalSubgroupsContaining
        (Subgroup.centralizer ({b} : Set G)) := by
      rw [mem_maximalSubgroupsContaining]
      refine ⟨OddOrder.BG.Ch3.S12.isCoatom_conj_smul
        (mem_maximalSubgroups.mp (hT ▸ mp.T_maximal)), ?_⟩
      rw [hCconj]
      exact Subgroup.pointwise_smul_le_pointwise_smul_iff.mpr hCa
    rw [hB, Set.mem_singleton_iff] at hMc
    exact mp.S_T_not_conj ⟨c⁻¹, by
      rw [← hMc, ← mul_smul, ← map_mul, inv_mul_cancel, map_one, one_smul, hT]⟩
  · -- `a` escapes `M` while `C_G(a) ≤ S^{c⁻¹}` — the (8.13.c4) exclusion
    have hSmem : mp.S ∈ maximalSubgroupsContaining
        (Subgroup.centralizer ({b} : Set G)) := by
      rw [hB]
      rfl
    have hCbS : Subgroup.centralizer ({b} : Set G) ≤ mp.S :=
      (mem_maximalSubgroupsContaining.mp hSmem).2
    have huniq : ∀ N ∈ maximalSubgroupsContaining
        (Subgroup.centralizer ({a} : Set G)), N = MulAut.conj c⁻¹ • mp.S := by
      intro N hN
      rw [mem_maximalSubgroupsContaining] at hN
      have hcN : MulAut.conj c • N ∈ maximalSubgroupsContaining
          (Subgroup.centralizer ({b} : Set G)) := by
        rw [mem_maximalSubgroupsContaining]
        refine ⟨OddOrder.BG.Ch3.S12.isCoatom_conj_smul (mem_maximalSubgroups.mp hN.1), ?_⟩
        rw [hCconj]
        exact Subgroup.pointwise_smul_le_pointwise_smul_iff.mpr hN.2
      rw [hB, Set.mem_singleton_iff] at hcN
      rw [← hcN, ← mul_smul, ← map_mul, inv_mul_cancel, map_one, one_smul]
    exact typeP_escaping_centralizer_not_le_typeII hG hyp (section16_S_isTypeII hG mp)
      haM' hCa huniq

open scoped Classical FiniteInduce in
/-- **Peterfalvi (8.18.b), base-point disjointness at the canonical pair** (the
support-geometry core of Coq `oST`): no `H(a)`-thickened `A₁(M)`-point of the type-`P₁`
`M = mp.T` is conjugate to an `A(S)`-point of the type-II member `mp.S`.

**The thickening reduces to the bare case by a power trick**: `h ∈ H(a)` commutes with `a`
((2.2) `H(a) ≤ C_G(a)`) with coprime order (`|H(a)| ⊥ |C_M(a)|` and `|a| ∣ |C_M(a)|`), so
CRT provides `k` with `(a·h)^k = a`; conjugating, `a` is conjugate to `b^k`, which is again
an `A(S)`-point (powers stay in `S′` and keep the centralizer witness).  The bare case is
`typeP_pair_base_bare_not_isConj`. -/
theorem typeP_pair_base_not_isConj [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G}
    (hyp : Hypothesis M) {mp : Section16MaximalPairCore G}
    (hT : mp.T = M) (hKstar : mp.Kstar = hyp.typeP.W1)
    {data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup mp.S}
    (hSW1 : data.typeP.W1 = mp.K) (hSW2 : data.typeP.W2 = mp.Kstar)
    {a : G} (haM' : a ∈ sharpSubgroup (derivedInG M)) (ha0 : a ∈ typePA0 M hyp.typeP)
    {h : G} (hh : h ∈ hyp.dadeData.dade.H ⟨a, ha0⟩)
    {b : G} (hbS : b ∈ centralizerSupport
      (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S)) (derivedInG mp.S)) :
    ¬ IsConj (a * h) b := by
  classical
  -- `a` and `h` commute with coprime orders ((2.2): `H(a) ≤ C_G(a)`, `|H(a)| ⊥ |C_M(a)|`),
  -- so `a` is a *power* of `a·h` (CRT), hence of `b` after conjugation — reducing to the
  -- bare case `a ~ b^k ∈ A(S)`.
  intro hconj
  have ha1 : a ≠ 1 := fun h0 => haM'.2 (Set.mem_singleton_iff.mpr h0)
  have hHle : hyp.dadeData.dade.H ⟨a, ha0⟩ ≤ Subgroup.centralizer ({a} : Set G) := by
    rw [hyp.dadeData.dade.centralizer_eq_sup ⟨a, ha0⟩]
    exact le_sup_left
  have hcomm : Commute a h := (Subgroup.mem_centralizer_singleton_iff.mp (hHle hh)).symm
  -- coprime orders
  have hha : orderOf h ∣ Nat.card (hyp.dadeData.dade.H ⟨a, ha0⟩) :=
    Subgroup.orderOf_dvd_natCard _ hh
  have haa : orderOf a ∣ Nat.card
      (OddOrder.Peterfalvi.S04.centralizerIn M a) :=
    Subgroup.orderOf_dvd_natCard _
      (OddOrder.Peterfalvi.S04.mem_centralizerIn.mpr
        ⟨hyp.dadeData.dade.mem_L ha0, rfl⟩)
  have hcopHC := hyp.dadeData.dade.centralizer_coprime ⟨a, ha0⟩ ⟨a, ha0⟩
  have hcop : Nat.Coprime (orderOf a) (orderOf h) :=
    (hcopHC.coprime_dvd_left hha).symm.coprime_dvd_left haa
  -- CRT: `(a·h)^k = a` for `k ≡ 1 (mod |a|)`, `k ≡ 0 (mod |h|)`
  obtain ⟨k, hk1, hk0⟩ := Nat.chineseRemainder hcop 1 0
  have hpow : (a * h) ^ k = a := by
    rw [hcomm.mul_pow]
    have h1 : a ^ k = a ^ 1 := pow_eq_pow_iff_modEq.mpr hk1
    have h2 : h ^ k = h ^ 0 := pow_eq_pow_iff_modEq.mpr hk0
    rw [h1, h2, pow_one, pow_zero, mul_one]
  -- transport along the conjugacy: `b^k` is a conjugate of `a`
  obtain ⟨c, hc⟩ := isConj_iff.mp hconj
  have hbk : b ^ k = c * a * c⁻¹ := by
    rw [← hc, conj_pow, hpow]
  -- `b^k` is again an `A(S)`-point (a power stays in `S′` and keeps the centralizer witness)
  obtain ⟨hbM', hb1, u, huS, hbu⟩ := hbS
  have hbkS : b ^ k ∈ centralizerSupport
      (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S)) (derivedInG mp.S) := by
    refine ⟨Subgroup.pow_mem _ hbM' k, ?_, u, huS,
      Subgroup.pow_mem (Subgroup.centralizer _) hbu k⟩
    intro h0
    apply ha1
    have h1 : c * a * c⁻¹ = 1 := by rw [← hbk, h0]
    have h2 : a = c⁻¹ * (c * a * c⁻¹) * c := by group
    rw [h2, h1]
    group
  -- the bare case kills `a ~ b^k`
  exact typeP_pair_base_bare_not_isConj hG hyp hT hKstar hSW1 hSW2 haM' ha0 hbkS
    (isConj_iff.mpr ⟨c, hbk.symm⟩)

open scoped Classical FiniteInduce in
/-- **Peterfalvi (8.18.b), cross-Dade orthogonality at the canonical pair** (Coq `oST` of
`Frob_der1_type2`, `PFsection10.v:577-590`): for `φ` a `(M′)^#`-supported class function of
the type-`P₁` `M = mp.T` and `ψ` an `A(S)`-supported class function of the type-II member
`mp.S`, the two Dade images are orthogonal, `⟨φ^{τ_M}, ψ^{τ_S}⟩ = 0`.

The image supports are the **restricted** thickenings
(`IsDadeMap.exists_base_of_map_apply_ne_zero`): a nonvanishing point of `φ^{τ_M}` is
conjugate to `a·h` with `a ∈ Supp(φ) ⊆ A₁(M) = (M′)^#` and `h ∈ H(a)`, and a nonvanishing
point of `ψ^{τ_S}` is conjugate to a bare `b ∈ Supp(ψ) ⊆ A(S)` (the (8.16) TI-route has
trivial signalizers).  A common point would make `a·h ~ b` — impossible by the (8.18.b)
base-point disjointness (`typeP_pair_base_not_isConj`), so the supports are disjoint and
the inner product vanishes (`inner_eq_zero_of_disjoint_support`). -/
theorem Hypothesis.cross_dade_inner_eq_zero_at_pair [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G}
    (hyp : Hypothesis M) {mp : Section16MaximalPairCore G}
    (hT : mp.T = M) (hKstar : mp.Kstar = hyp.typeP.W1)
    {data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup mp.S}
    (hSW1 : data.typeP.W1 = mp.K) (hSW2 : data.typeP.W2 = mp.Kstar)
    [NeZero (Nat.card (typeIIHypothesis46 hG mp.S_maximal
      (section16_S_isTypeII hG mp) data.typeP).W1)]
    {φ : ClassFunction ↥M ℂ} {ψ : ClassFunction ↥mp.S ℂ}
    (hφsupp : φ.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (sharpSubgroup (derivedInG M)) M)
    (hψsupp : ψ.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S))
        (derivedInG mp.S)) mp.S) :
    ClassFunction.inner (hyp.tau φ)
      (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).dade0
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).tau ψ)
      = 0 := by
  classical
  have hφA0 : φ.support ⊆ hyp.A0 :=
    hφsupp.trans hyp.supportInSubgroup_sharp_derived_subset_A0
  have hψA0 : ψ.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S))
          (derivedInG mp.S)
        ∪ conjClassSetIn mp.S (typePV mp.S data.typeP)) mp.S :=
    hψsupp.trans (OddOrder.Peterfalvi.S04.supportInSubgroup_mono Set.subset_union_left)
  refine ClassFunction.inner_eq_zero_of_disjoint_support ?_
  rw [Set.disjoint_left]
  intro g hg1 hg2
  rw [ClassFunction.mem_support] at hg1 hg2
  -- `M`-side base point: `g ~ a·h`, `a ∈ Supp(φ) ⊆ (M′)^#`, `h ∈ H(a)`
  have hg1' : (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap hyp.dadeData.dade
      (hyp.dadeData.dade.fullDadeIsometryData) φ) g ≠ 0 := hg1
  rw [OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_apply_of_support
    hyp.dadeData.dade _ hφA0] at hg1'
  obtain ⟨a, h, hh, hconja, hvala⟩ :=
    OddOrder.Peterfalvi.S04.IsDadeMap.exists_base_of_map_apply_ne_zero
      (hyp.dadeData.dade.isDadeMap_dadeMap (k := ℂ)) _ hg1'
  have haM' : a.1 ∈ sharpSubgroup (derivedInG M) := by
    have := hφsupp (ClassFunction.mem_support.mpr hvala)
    rwa [OddOrder.Peterfalvi.S04.mem_supportInSubgroup] at this
  -- `S`-side base point: `g ~ b·k` with `k ∈ ⊥`, so `g ~ b`, `b ∈ Supp(ψ) ⊆ A(S)`
  rw [OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_apply_of_support
    (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).dade0
    _ hψA0] at hg2
  obtain ⟨b, k, hk, hconjb, hvalb⟩ :=
    OddOrder.Peterfalvi.S04.IsDadeMap.exists_base_of_map_apply_ne_zero
      ((typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp)
        data.typeP).dade0.isDadeMap_dadeMap (k := ℂ)) _ hg2
  have hbS : b.1 ∈ centralizerSupport
      (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S)) (derivedInG mp.S) := by
    have := hψsupp (ClassFunction.mem_support.mpr hvalb)
    rwa [OddOrder.Peterfalvi.S04.mem_supportInSubgroup] at this
  -- the TI-route has trivial signalizers: `k = 1`
  have hk1 : k = 1 := by
    have hbot : (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp)
        data.typeP).dade0.H b = ⊥ := rfl
    rw [hbot, Subgroup.mem_bot] at hk
    exact hk
  -- combine the conjugacies: `a·h ~ g ~ b`
  exact typeP_pair_base_not_isConj hG hyp hT hKstar hSW1 hSW2 haM' a.2 hh hbS
    (hconja.trans (by rw [hk1, mul_one] at hconjb; exact hconjb.symm))

open scoped Classical FiniteInduce in
/-- **The (10.7) cross-side orthogonality `⟨ζ^{τ₁}, λ^{τ₂}⟩ = 0`** (the `zeta_lam_ortho`
production; Coq `Frob_der1_type2`'s `orthonormal_vchar_diff_ortho` step): the `M`-side
coherent image of `ζ` is orthogonal to the `S`-side coherent image of `λ`.

The conjugate-pair difference trick: `{ζ^{τ₁}, ζ̄^{τ₁}}` and `{λ^{τ₂}, λ̄^{τ₂}}` are
orthonormal pairs of virtual characters (the coherence isometries on distinct
irreducibles), the differences are the Dade images `(ζ − ζ̄)^{τ_M}` and `(λ − λ̄)^{τ_S}`
(coherence agreement on the supported lattice) — orthogonal by the (8.18.b) support
disjointness (`cross_dade_inner_eq_zero_at_pair`) and both vanishing at `1`
(`dadeIntegralCharacterMap_apply_one_eq_zero`), so the cross-orthogonality primitive
`orthonormal_vchar_diff_ortho` applies. -/
theorem Hypothesis.tau1_zeta_inner_extension_lam_eq_zero_at_pair [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G}
    (hyp : Hypothesis M) {mp : Section16MaximalPairCore G}
    (hT : mp.T = M) (hKstar : mp.Kstar = hyp.typeP.W1)
    {data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup mp.S}
    (hSW1 : data.typeP.W1 = mp.K) (hSW2 : data.typeP.W2 = mp.Kstar)
    [NeZero (Nat.card (typeIIHypothesis46 hG mp.S_maximal
      (section16_S_isTypeII hG mp) data.typeP).W1)]
    {params : CharacterParameters hyp} (coh : CoherentHypothesis hyp params)
    {Y : Subgroup G} {lam nu : ClassFunction ↥mp.S ℂ}
    (hlam_mem : lam ∈ OddOrder.Peterfalvi.S11.sOf data Y)
    (hlam_irr : IsIrreducibleCharacter lam)
    (_hnu_mem : nu ∈ OddOrder.Peterfalvi.S11.sOf data Y)
    (_hdeg : lam 1 = nu 1)
    (c : OddOrder.Peterfalvi.S07.IsCoherent
      (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).dade0
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).tau)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥mp.S ℂ))
      (OddOrder.Peterfalvi.S04.supportInSubgroup
        (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S))
            (derivedInG mp.S)
          ∪ conjClassSetIn mp.S (typePV mp.S data.typeP)) mp.S)) :
    ClassFunction.inner (coh.tau1 params.zeta) (c.extension lam) = 0 := by
  classical
  subst hT
  -- `M`-side pieces: `ζ`, `ζ̄`, their coherent images
  have hζS : params.zeta ∈ inducedFamily mp.T := params.zeta_mem_S
  have hζirr := params.zeta_irreducible
  have hModd : Odd (Nat.card ↥mp.T) :=
    hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card mp.T)
  have hζne : params.zeta.conj ≠ params.zeta := inducedFamily_hasNoRealCharacters hModd hζS
  have hζcS : params.zeta.conj ∈ inducedFamily mp.T :=
    inducedFamily_closedUnderConjugate mp.T hζS
  have haZ : coh.tau1 params.zeta ∈ ZIrr G :=
    coh.coherent.extension_mem_ZIrr _ (Submodule.subset_span hζS)
  have hbZ : coh.tau1 params.zeta.conj ∈ ZIrr G :=
    coh.coherent.extension_mem_ZIrr _ (Submodule.subset_span hζcS)
  have ha1 := hyp.zeta_tau1_inner_self hG hG.odd coh hζS hζirr
  have hb1 := hyp.zeta_tau1_inner_self hG hG.odd coh hζcS hζirr.conj
  have hab : ClassFunction.inner (coh.tau1 params.zeta) (coh.tau1 params.zeta.conj) = 0 := by
    change ClassFunction.inner (coh.coherent.extension _) (coh.coherent.extension _) = 0
    rw [coh.coherent.extension_inner_eq _ _ (Submodule.subset_span hζS)
        (Submodule.subset_span hζcS),
      OddOrder.RepresentationTheory.irr_cf_inner hζirr hζirr.conj,
      if_neg (fun h => hζne h.symm)]
  -- `S`-side pieces: `λ`, `λ̄`, their coherent images (as in the (5.5) span lemma)
  have hSmodd : Odd (Nat.card ↥mp.S) :=
    hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card mp.S)
  have hlamIKF := typeII_sOf_subset_inducedKernelFamily data Y hlam_mem
  have hr : ¬ ClassFunction.IsReal lam :=
    OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters hSmodd
      (Y.subgroupOf mp.S) hlamIKF
  have hlamne : lam ≠ lam.conj := fun h => hr h.symm
  have hlam_zspan : lam ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥mp.S)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥mp.S ℂ)) :=
    Submodule.subset_span (Set.mem_insert _ _)
  have hlamc_zspan : lam.conj ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥mp.S)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥mp.S ℂ)) :=
    Submodule.subset_span (Set.mem_insert_of_mem _ (Set.mem_insert _ _))
  have hcZ : c.extension lam ∈ ZIrr G := c.extension_mem_ZIrr _ hlam_zspan
  have hdZ : c.extension lam.conj ∈ ZIrr G := c.extension_mem_ZIrr _ hlamc_zspan
  have hc1 : ClassFunction.inner (c.extension lam) (c.extension lam) = 1 := by
    rw [c.extension_inner_eq _ _ hlam_zspan hlam_zspan]
    exact hlam_irr.inner_self_eq_one
  have hd1 : ClassFunction.inner (c.extension lam.conj) (c.extension lam.conj) = 1 := by
    rw [c.extension_inner_eq _ _ hlamc_zspan hlamc_zspan]
    exact hlam_irr.conj.inner_self_eq_one
  have hcd : ClassFunction.inner (c.extension lam) (c.extension lam.conj) = 0 := by
    rw [c.extension_inner_eq _ _ hlam_zspan hlamc_zspan,
      OddOrder.RepresentationTheory.irr_cf_inner hlam_irr hlam_irr.conj,
      if_neg (fun h => hlamne h)]
  -- supports of the two differences
  have hsupp : (lam - lam.conj).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S))
        (derivedInG mp.S)) mp.S := by
    have hmem := typeII_T2_member_support hG mp.S_maximal (section16_S_isTypeII hG mp) data
      hlam_mem hlam_mem
    refine diff_support_subset_of_support_subset_union_one
      (hmem lam (by simp)) (hmem lam.conj (by simp)) ?_
    obtain ⟨n, -, hn⟩ := typeII_sOf_apply_one_eq_pos_natCast data hlam_mem
    rw [ClassFunction.conj_apply, hn, star_natCast]
  have hsuppA0 : (lam - lam.conj).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S))
          (derivedInG mp.S)
        ∪ conjClassSetIn mp.S (typePV mp.S data.typeP)) mp.S :=
    hsupp.trans (OddOrder.Peterfalvi.S04.supportInSubgroup_mono Set.subset_union_left)
  have hζ1conj : params.zeta.conj 1 = params.zeta 1 := by
    obtain ⟨nn, -, hn, -⟩ := hζirr.exists_natDegree_charValue_one_dvd_card
    simp only [ClassFunction.conj_apply, hn, star_natCast]
  have hφsupp : (params.zeta - params.zeta.conj).support
      ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
        (sharpSubgroup (derivedInG mp.T)) mp.T :=
    hyp.mem_zSpan_inducedFamily_support_sharp_derived
      (Submodule.sub_mem _ (Submodule.subset_span hζS) (Submodule.subset_span hζcS))
      (by rw [ClassFunction.sub_apply, hζ1conj, sub_self])
  -- the two difference identities: coherent images = Dade images
  have hMdiff : coh.tau1 params.zeta - coh.tau1 params.zeta.conj
      = hyp.tau (params.zeta - params.zeta.conj) :=
    (hyp.tau_zeta_sub_conj_eq_tau1 hG hG.odd coh hζS hζirr).symm
  have hSdiff : c.extension lam - c.extension lam.conj
      = OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).dade0
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).tau
        (lam - lam.conj) := by
    rw [← c.extends_on_supported _
      ⟨Submodule.sub_mem _ hlam_zspan hlamc_zspan, hsuppA0⟩, map_sub]
  -- (8.18.b): the differences are orthogonal
  have hdiff : ClassFunction.inner
      (coh.tau1 params.zeta - coh.tau1 params.zeta.conj)
      (c.extension lam - c.extension lam.conj) = 0 := by
    rw [hMdiff, hSdiff]
    exact hyp.cross_dade_inner_eq_zero_at_pair hG rfl hKstar hSW1 hSW2 hφsupp hsupp
  -- both differences vanish at `1`
  have hab1 : (coh.tau1 params.zeta) 1 = (coh.tau1 params.zeta.conj) 1 := by
    have hz : (coh.tau1 params.zeta - coh.tau1 params.zeta.conj) (1 : G) = 0 := by
      rw [hMdiff]
      exact OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_apply_one_eq_zero
        hyp.dadeData.dade (hyp.zeta_sub_conj_support hG hG.odd hζS hζirr)
    rw [ClassFunction.sub_apply] at hz
    exact sub_eq_zero.mp hz
  have hcd1 : (c.extension lam) 1 = (c.extension lam.conj) 1 := by
    have hz : (c.extension lam - c.extension lam.conj) (1 : G) = 0 := by
      rw [hSdiff]
      exact OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_apply_one_eq_zero
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).dade0 hsuppA0
    rw [ClassFunction.sub_apply] at hz
    exact sub_eq_zero.mp hz
  exact orthonormal_vchar_diff_ortho haZ hbZ hcZ hdZ ha1 hb1 hc1 hd1 hab hcd hdiff hab1 hcd1

open scoped Classical FiniteInduce in
/-- **The (10.7) `cross_zero` production** (Coq `Frob_der1_type2`'s
`'[alpha^\tau, beta^{tau2}] = 0` step): for a nontrivial column `s`, the `τ₁`-image of
`α = μ_s − d·ζ` is orthogonal to `ν^{τ₂} − λ^{τ₂}`.

Both sides are Dade images of supported lattice elements: `μ_s = ∑_i μ_{is} ∈ 𝒮`
(`muGrid_column_sum_mem_inducedFamily`) and `ζ ∈ 𝒮` with `α(1) = w₁·d − d·w₁ = 0`
((10.3) degree independence and `ζ(1) = w₁`), so `α` is `(M′)^#`-supported
(`mem_zSpan_inducedFamily_support_sharp_derived`) and `τ₁α = α^{τ_M}`
(`extends_on_supported` through `(M′)^# ⊆ A₀`); `ν − λ` is `A(S)`-supported with
`τ₂(ν − λ) = (ν − λ)^{τ_S}`.  The (8.18.b) support disjointness
(`cross_dade_inner_eq_zero_at_pair`) kills the cross inner product. -/
theorem Hypothesis.tau1_muColumn_sub_zeta_inner_extension_diff_eq_zero_at_pair [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G}
    (hyp : Hypothesis M) {mp : Section16MaximalPairCore G}
    (hT : mp.T = M) (hKstar : mp.Kstar = hyp.typeP.W1)
    {data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup mp.S}
    (hSW1 : data.typeP.W1 = mp.K) (hSW2 : data.typeP.W2 = mp.Kstar)
    [NeZero (Nat.card (typeIIHypothesis46 hG mp.S_maximal
      (section16_S_isTypeII hG mp) data.typeP).W1)]
    {params : CharacterParameters hyp} (coh : CoherentHypothesis hyp params)
    (hζ1 : params.zeta 1 = (hyp.w1 : ℂ))
    (hμd : ∀ (i : Fin hyp.w1) (j : Fin hyp.w2), j ≠ 0 →
      hyp.muGrid hG hG.odd i j 1 = (params.d : ℂ))
    {Y : Subgroup G} {lam nu : ClassFunction ↥mp.S ℂ}
    (hlam_mem : lam ∈ OddOrder.Peterfalvi.S11.sOf data Y)
    (_hlam_irr : IsIrreducibleCharacter lam)
    (hnu_mem : nu ∈ OddOrder.Peterfalvi.S11.sOf data Y)
    (hdeg : lam 1 = nu 1)
    (c : OddOrder.Peterfalvi.S07.IsCoherent
      (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).dade0
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).tau)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥mp.S ℂ))
      (OddOrder.Peterfalvi.S04.supportInSubgroup
        (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S))
            (derivedInG mp.S)
          ∪ conjClassSetIn mp.S (typePV mp.S data.typeP)) mp.S))
    (s : Fin hyp.w2) (hs : s ≠ 0) :
    ClassFunction.inner
      (coh.tau1 ((∑ i : Fin hyp.w1, hyp.muGrid hG hG.odd i s)
        - (params.d : ℂ) • params.zeta))
      (c.extension nu - c.extension lam) = 0 := by
  classical
  subst hT
  -- `M`-side: `α = μ_s − d·ζ` is a `1`-vanishing lattice element, `(M′)^#`-supported
  have hd01 : hyp.muGrid hG hG.odd 0 s 1 ≠ 1 := by
    rw [hμd 0 s hs]
    intro h
    have hd : params.d = 1 := by exact_mod_cast h
    have := params.d_gt_one
    omega
  have hμS : (∑ i : Fin hyp.w1, hyp.muGrid hG hG.odd i s) ∈ inducedFamily mp.T :=
    hyp.muGrid_column_sum_mem_inducedFamily hG hG.odd s hd01
  have hdsmul : (params.d : ℂ) • params.zeta = (params.d : ℤ) • params.zeta := by
    rw [← Int.cast_smul_eq_zsmul ℂ (params.d : ℤ) params.zeta]
    norm_num
  have hαspan : (∑ i : Fin hyp.w1, hyp.muGrid hG hG.odd i s)
      - (params.d : ℂ) • params.zeta
      ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥mp.T) (inducedFamily mp.T) := by
    refine Submodule.sub_mem _ (Submodule.subset_span hμS) ?_
    rw [hdsmul]
    exact Submodule.smul_mem _ _ (Submodule.subset_span params.zeta_mem_S)
  have hα1 : ((∑ i : Fin hyp.w1, hyp.muGrid hG hG.odd i s)
      - (params.d : ℂ) • params.zeta) 1 = 0 := by
    rw [ClassFunction.sub_apply, ClassFunction.smul_apply, hζ1,
      ClassFunction.finset_sum_apply,
      Finset.sum_congr rfl (fun i _ => hμd i s hs),
      Finset.sum_const, Finset.card_univ, Fintype.card_fin, nsmul_eq_mul]
    ring
  have hαsharp := hyp.mem_zSpan_inducedFamily_support_sharp_derived hαspan hα1
  have hM : coh.tau1 ((∑ i : Fin hyp.w1, hyp.muGrid hG hG.odd i s)
      - (params.d : ℂ) • params.zeta)
      = hyp.tau ((∑ i : Fin hyp.w1, hyp.muGrid hG hG.odd i s)
        - (params.d : ℂ) • params.zeta) :=
    coh.coherent.extends_on_supported _
      ⟨hαspan, hαsharp.trans hyp.supportInSubgroup_sharp_derived_subset_A0⟩
  -- `S`-side: `ν − λ` is `A(S)`-supported
  have hψsupp : (nu - lam).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S))
        (derivedInG mp.S)) mp.S := by
    have hmem := typeII_T2_member_support hG mp.S_maximal (section16_S_isTypeII hG mp) data
      hlam_mem hnu_mem
    exact diff_support_subset_of_support_subset_union_one
      (hmem nu (by simp)) (hmem lam (by simp)) hdeg.symm
  have hψA0 : (nu - lam).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma mp.S))
          (derivedInG mp.S)
        ∪ conjClassSetIn mp.S (typePV mp.S data.typeP)) mp.S :=
    hψsupp.trans (OddOrder.Peterfalvi.S04.supportInSubgroup_mono Set.subset_union_left)
  have hnu_zspan : nu ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥mp.S)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥mp.S ℂ)) :=
    Submodule.subset_span (Set.mem_insert_of_mem _
      (Set.mem_insert_of_mem _ (Set.mem_insert _ _)))
  have hlam_zspan : lam ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥mp.S)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥mp.S ℂ)) :=
    Submodule.subset_span (Set.mem_insert _ _)
  have hS : c.extension nu - c.extension lam
      = OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).dade0
        (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP).tau
        (nu - lam) := by
    rw [← c.extends_on_supported _ ⟨Submodule.sub_mem _ hnu_zspan hlam_zspan, hψA0⟩, map_sub]
  rw [hM, hS]
  exact hyp.cross_dade_inner_eq_zero_at_pair hG rfl hKstar hSW1 hSW2 hαsharp hψsupp

set_option linter.unusedVariables false in
open scoped Classical FiniteInduce in
/-- **The (10.7) cross-isometry package at the canonical pair** (Coq `Frob_der1_type2`,
production step): for the `M`-seeded pair (`T = M`, `Kstar = hyp.typeP.W1`) and a
`(K, K*)`-reconciled §9 setup on the type-II member `mp.S`, the package exists with the
grid fields **honestly produced**: `tau2` is the (5.7) `T2`-coherent extension
(`typeII_T2_coherent`), and `nu_tau2_eq` is the (5.8) row pin
(`exists_nu_extension_eq_alignedRow_at_pair`).

The four support-geometry fields ((5.3.b) grid orthogonality and the (8.18.b)
cross-support vanishing) are the obligation-3 content, all discharged in this file
(see the module docstring). -/
theorem exists_typeIICrossIsometryData_at_pair [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G}
    {hyp : Hypothesis M} {params : CharacterParameters hyp}
    (coh : CoherentHypothesis hyp params)
    (hζ1 : params.zeta 1 = (hyp.w1 : ℂ))
    (hμd : ∀ (i : Fin hyp.w1) (j : Fin hyp.w2), j ≠ 0 →
      hyp.muGrid hG hG.odd i j 1 = (params.d : ℂ))
    {mp : Section16MaximalPairCore G}
    (hT : mp.T = M) (hKstar : mp.Kstar = hyp.typeP.W1)
    {data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup mp.S}
    (hSW1 : data.typeP.W1 = mp.K) (hSW2 : data.typeP.W2 = mp.Kstar)
    [NeZero (Nat.card (typeIIHypothesis46 hG mp.S_maximal
      (section16_S_isTypeII hG mp) data.typeP).W1)]
    {Y : Subgroup G} {lam nu : ClassFunction ↥mp.S ℂ}
    (hlam_mem : lam ∈ OddOrder.Peterfalvi.S11.sOf data Y)
    (hlam_irr : IsIrreducibleCharacter lam)
    (hnu_mem : nu ∈ OddOrder.Peterfalvi.S11.sOf data Y)
    (hnu_red : ¬ IsIrreducibleCharacter nu)
    (hdeg : lam 1 = nu 1) :
    Nonempty (TypeIICrossIsometryData hG coh lam nu) := by
  classical
  -- the (5.7) `T2`-coherence on `mp.S`
  obtain ⟨c⟩ := typeII_T2_coherent hG mp.S_maximal (section16_S_isTypeII hG mp) data
    hlam_mem hlam_irr hnu_mem hnu_red hdeg
  -- the (5.8) row pin (the assembled grid transpose)
  obtain ⟨r', delta', hpm, hrow⟩ := hyp.exists_nu_extension_eq_alignedRow_at_pair hG
    hT hKstar hSW1 hSW2 hlam_mem hlam_irr hnu_mem hnu_red hdeg c
  exact ⟨{ tau2 := c.extension
           r' := r'
           delta' := delta'
           delta'_pm := hpm
           nu_tau2_eq := hrow
           lam_ortho_grid := fun i j =>
             hyp.extension_lam_inner_alignedGrid_eq_zero_at_pair hG hT hKstar hSW1 hSW2
               hlam_mem hlam_irr hnu_mem hdeg c i j
           zeta_ortho_grid := fun i j =>
             hyp.tau1_zeta_inner_alignedGrid_eq_zero hG hG.odd coh params.zeta_mem_S
               params.zeta_irreducible i j
           zeta_lam_ortho :=
             hyp.tau1_zeta_inner_extension_lam_eq_zero_at_pair hG hT hKstar hSW1 hSW2
               coh hlam_mem hlam_irr hnu_mem hdeg c
           cross_zero := fun s hs =>
             hyp.tau1_muColumn_sub_zeta_inner_extension_diff_eq_zero_at_pair hG hT hKstar
               hSW1 hSW2 coh hζ1 hμd hlam_mem hlam_irr hnu_mem hdeg c s hs }⟩

open scoped Classical FiniteInduce in
/-- **Peterfalvi (10.7) at the canonical pair member** (Coq `Frob_der1_type2`, assembled):
under Hypothesis (10.4) for `M`, the `(K, K*)`-reconciled Types-II/III/IV setup on the
pair's type-II member `mp.S` has `[S,S] = S_F ⋊ U` Frobenius with kernel `S_F`.

The dichotomy assembly of the retired `typeII_HU_frobenius_of_coherent_aux`, with its sorried gate
replaced by the **honest pair-witness producer** (`exists_typeIICrossIsometryData_at_pair`,
axiom-clean): the §9 Clifford dichotomy's contradiction branches (Case A and the
non-exceptional Case B) produce an equal-degree irreducible/reducible pair in `𝒮(H₀)`,
whose (10.7) cross-isometry package is contradictory (`TypeIICrossIsometryData.elim`
against the (10.2)/(10.3) canonical parameters, `exists_charParameters_full`); the
exceptional Case B yields the Frobenius structure directly (Peterfalvi (9.10)). -/
theorem typeII_HU_frobenius_of_coherent_at_pair [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G}
    {hyp : Hypothesis M} {params : CharacterParameters hyp}
    (coh : CoherentHypothesis hyp params)
    {mp : Section16MaximalPairCore G}
    (hT : mp.T = M) (hKstar : mp.Kstar = hyp.typeP.W1)
    {data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup mp.S}
    (hSW1 : data.typeP.W1 = mp.K) (hSW2 : data.typeP.W2 = mp.Kstar) :
    OddOrder.Isaacs.Ch06.IsFrobeniusGroup ↥(derivedInG mp.S)
      (data.typeP.H.subgroupOf (derivedInG mp.S))
      (data.typeP.U.subgroupOf (derivedInG mp.S)) := by
  have := hyp.finiteG
  classical
  have : NeZero (Nat.card (typeIIHypothesis46 hG mp.S_maximal
      (section16_S_isTypeII hG mp) data.typeP).W1) := ⟨Nat.card_pos.ne'⟩
  -- the (10.3) canonical parameters carrying the grid/`ζ` pins
  obtain ⟨params', hmu, hos, hzS, hz1, hzconj, hδpm, hδj⟩ := hyp.exists_charParameters_full hG
  let coh' : CoherentHypothesis hyp params' := ⟨coh.coherent⟩
  have hμd : ∀ (i : Fin hyp.w1) (j : Fin hyp.w2), j ≠ 0 →
      hyp.muGrid hG hG.odd i j 1 = (params'.d : ℂ) := fun i j hj => by
    rw [← hmu]
    exact params'.degree_independent i j hj
  obtain ⟨chief, -⟩ := OddOrder.Peterfalvi.S11.exists_chiefFactorData hG data
  -- §9 character data (only the genuine `u`/`u_eq` pair is consumed by the counts)
  let chars : OddOrder.Peterfalvi.S11.Section11CharacterData data chief :=
    { u := Nat.card ↥(((OddOrder.Isaacs.Ch03.IsAInvariant.quotientMulAutHom
          (N := chief.N) chief.N_aInvariant).comp
          (data.typeP.U.subgroupOf (data.typeP.U ⊔ data.typeP.W1)).subtype).range)
      u_eq_card_quotient := rfl
      H0CprimeSupport := ∅
      tau := 0
      quotientSemidirectFrobenius := True }
  -- the reducible `ν ∈ 𝒮(H₀)`: the (9.8.a)/(9.9.b) count `p − 1 ≥ 1`
  have hred_ne : {φ ∈ OddOrder.Peterfalvi.S11.sOf data chief.H0 |
      ¬ IsIrreducibleCharacter φ}.Nonempty := by
    apply Set.nonempty_of_ncard_ne_zero
    rw [OddOrder.Peterfalvi.S11.reducible_count_sOf_H0 hG chief]
    have := chief.p_prime.two_le
    omega
  obtain ⟨nu, hnu_mem, hnu_red⟩ := hred_ne
  -- the left-branch refutation, now through the honest pair-witness producer
  have hleft : ∀ lam : ClassFunction ↥mp.S ℂ,
      lam ∈ OddOrder.Peterfalvi.S11.sOf data chief.H0 → IsIrreducibleCharacter lam →
      lam 1 = nu 1 → False := fun lam hlam_mem hlam_irr hdeg =>
    (exists_typeIICrossIsometryData_at_pair hG coh' hz1 hμd hT hKstar hSW1 hSW2
      hlam_mem hlam_irr hnu_mem hnu_red hdeg).elim fun pkg =>
      pkg.elim hG hmu hos hzS hz1 hzconj hδpm hδj
  -- Split on the exceptional condition first: in the exceptional case (9.10) now selects the
  -- Clifford branch itself (`..._of_trigger`), so no case carrier is needed here.
  by_cases hex : ∃ χ ∈ chars.SOf (chief.H0 ⊔ chars.Cprime),
      IsIrreducibleCharacter χ ∧ χ 1 = ((data.q * chars.u : ℕ) : ℂ)
  · -- non-exceptional: the degree-`q·u` irreducible exists — contradiction.  The reducible
    -- degree fact for `nu` is the one place the (9.7) dichotomy is still needed.
    exfalso
    rcases OddOrder.Peterfalvi.S11.clifford_dichotomy hG chars with hA | hB
    · obtain ⟨caseA⟩ := hA
      obtain ⟨-, hbred, ⟨lam, hlam_mem, hlam_irr, hlam_deg⟩, -⟩ :=
        OddOrder.Peterfalvi.S11.caseA_character_counts hG chars caseA
      have hnu_deg := (hbred nu hnu_mem hnu_red).1
      exact hleft lam
        (OddOrder.Peterfalvi.S11.sOf_antitone data le_sup_left hlam_mem)
        hlam_irr (by rw [hlam_deg, hnu_deg])
    · obtain ⟨caseB⟩ := hB
      obtain ⟨lam, hlam_mem, hlam_irr, hlam_deg⟩ := hex
      obtain ⟨-, -, hbred, -⟩ :=
        OddOrder.Peterfalvi.S11.caseB_character_counts hG chars caseB
      have hnu_deg := (hbred nu hnu_mem hnu_red).1
      exact hleft lam
        (OddOrder.Peterfalvi.S11.sOf_antitone data le_sup_left hlam_mem)
        hlam_irr (by rw [hlam_deg, hnu_deg])
  · -- exceptional: (9.10) gives the `H ⊔ U` Frobenius on the `derivedInG mp.S` carrier
    have hfrobHU := (OddOrder.Peterfalvi.S11.exceptional_case_frobenius_realization_of_trigger
      hG chars hex).2.2.2 (section16_S_isTypeII hG mp)
    have hM'eq : derivedInG mp.S = data.typeP.H ⊔ data.typeP.U := by
      rw [data.typeP.derivedInG_eq_fitting_sup_U, ← data.typeP.H_eq]
    rw [hM'eq]
    exact hfrobHU

open scoped Classical FiniteInduce Pointwise in
/-- **Peterfalvi (10.7)** (setup form, pair-witness route; Coq `Frob_der1_type2`): under
Hypothesis (10.4) for `M`, every Type-II maximal `S` with a Types-II/III/IV setup has
`[S,S] = S_F ⋊ U` Frobenius with kernel `S_F`, on the given datum's factors.

The honest replacement of `typeII_HU_frobenius_of_coherent_aux` (whose left branch cited
the sorried gate `exists_typeIICrossIsometryData`): the WLOG engine
`exists_reconciled_conj_typePData_S` transports the datum onto the canonical pair member
`mp.S` with `H`/`U` carried along as conjugates; the pair-member theorem
`typeII_HU_frobenius_of_coherent_at_pair` (fully proven) gives the Frobenius structure
there; and `IsFrobeniusGroup.mapEquiv` along the inverse conjugation returns it on the
original factors verbatim. -/
theorem typeII_HU_frobenius_of_coherent' [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {M : Subgroup G}
    {hyp : Hypothesis M} {params : CharacterParameters hyp}
    (coh : CoherentHypothesis hyp params)
    {S : Subgroup G} (data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup S)
    (hSII : IsTypeII S) :
    OddOrder.Isaacs.Ch06.IsFrobeniusGroup ↥(derivedInG S)
      (data.typeP.H.subgroupOf (derivedInG S))
      (data.typeP.U.subgroupOf (derivedInG S)) := by
  have := hyp.finiteG
  classical
  -- WLOG: transport the datum onto the canonical pair member, `H`/`U` tracked
  obtain ⟨mp, u, dataS, hT, hKstar, huS, hSW1, hSW2, hH, hU⟩ :=
    exists_reconciled_conj_typePData_S hG hyp data.maximal hSII data.typeP
  obtain ⟨dII⟩ := section16_S_isTypeII hG mp
  -- (10.7) at the pair member, on the reconciled setup
  have hfrob := typeII_HU_frobenius_of_coherent_at_pair hG coh hT hKstar
    (data := { maximal := mp.S_maximal
               typeP := dataS
               nontrivial := dII.common.transfer dataS
               type_alt := Or.inl (section16_S_isTypeII hG mp) }) hSW1 hSW2
  have hfrob2 : OddOrder.Isaacs.Ch06.IsFrobeniusGroup ↥(derivedInG mp.S)
      ((MulAut.conj u • data.typeP.H).subgroupOf (derivedInG mp.S))
      ((MulAut.conj u • data.typeP.U).subgroupOf (derivedInG mp.S)) := by
    rw [← hH, ← hU]
    exact hfrob
  -- the two factors sit inside `[S,S]`
  have hM'eq : derivedInG S = data.typeP.H ⊔ data.typeP.U := by
    rw [data.typeP.derivedInG_eq_fitting_sup_U, ← data.typeP.H_eq]
  have hHle : data.typeP.H ≤ derivedInG S := by rw [hM'eq]; exact le_sup_left
  have hUle : data.typeP.U ≤ derivedInG S := by rw [hM'eq]; exact le_sup_right
  -- the conjugation isomorphism `[S,S] ≃* [mp.S, mp.S]`
  have hD : (derivedInG S).map (MulAut.conj u).toMonoidHom = derivedInG mp.S := by
    have h1 : MulAut.conj u • derivedInG S = derivedInG mp.S := by
      rw [derivedInG_pointwise_smul, huS]
    rw [← h1]
    rfl
  set e : ↥(derivedInG S) ≃* ↥(derivedInG mp.S) :=
    ((MulAut.conj u).subgroupMap (derivedInG S)).trans
      (MulEquiv.subgroupCongr hD) with he
  -- carrier bridge: `map e.symm ((u • X).subgroupOf [mp.S,mp.S]) = X.subgroupOf [S,S]`
  have hbridge : ∀ X : Subgroup G,
      ((MulAut.conj u • X).subgroupOf (derivedInG mp.S)).map e.symm.toMonoidHom
        = X.subgroupOf (derivedInG S) := by
    intro X
    have hcoe : ∀ z : ↥(derivedInG S), ((e z : ↥(derivedInG mp.S)) : G)
        = u * (z : G) * u⁻¹ := fun z => rfl
    ext z
    simp only [Subgroup.mem_map, Subgroup.mem_subgroupOf]
    constructor
    · rintro ⟨w, hw, rfl⟩
      change ((e.symm w : ↥(derivedInG S)) : G) ∈ X
      have hz : ((e.symm w : ↥(derivedInG S)) : G)
          = u⁻¹ * ((w : ↥(derivedInG mp.S)) : G) * u := by
        have := hcoe (e.symm w)
        rw [MulEquiv.apply_symm_apply] at this
        have h2 := congrArg (fun x => u⁻¹ * x * u) this
        simpa [mul_assoc] using h2.symm
      rw [hz]
      rw [Subgroup.mem_pointwise_smul_iff_inv_smul_mem] at hw
      have hinv : (MulAut.conj u)⁻¹ • ((w : ↥(derivedInG mp.S)) : G)
          = u⁻¹ * ((w : ↥(derivedInG mp.S)) : G) * u := by
        rw [← map_inv, MulAut.smul_def, MulAut.conj_apply, inv_inv]
      rwa [hinv] at hw
    · intro hz
      refine ⟨e z, ?_, by simp⟩
      show ((e z : ↥(derivedInG mp.S)) : G) ∈ MulAut.conj u • X
      rw [hcoe z]
      have : u * (z : G) * u⁻¹ = (MulAut.conj u) • (z : G) := by
        rw [MulAut.smul_def, MulAut.conj_apply]
      rw [this]
      exact Subgroup.smul_mem_pointwise_smul _ _ _ hz
  -- transport back
  have hfrob' := hfrob2.mapEquiv e.symm
  rw [hbridge data.typeP.H, hbridge data.typeP.U] at hfrob'
  exact hfrob'

end OddOrder.Peterfalvi.S12
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/S12_TypeIICrossIsometryPair.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