Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Peterfalvi §12: the type II cross-isometry pair, opening layer

Definition
OddOrder_Peterfalvi_S12_CrossIsometrySetup

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

feit-thompsonfinite-groupsgroup-theoryishida-odd-order

The opening layer of the type II cross-isometry development: the (5.3.b) base orthogonality on the SSS-side and (5.5) for the T2T_2T2​-coherence.

Main declarations: typeII_tau_diff_inner_chiFam_eq_zero (The (5.3.b) base-orthogonality kernel, S-side); typeII_R_mem_inner_chiFam_eq_zero (The (5.3.b) base orthogonality, per R-element); typeII_T2_extension_lam_mem_span_RFamily (Peterfalvi (5.5) for the T2-coherence); typeII_T2_extension_lam_inner_chiFam_eq_zero (The (5.3.b) S-side grid orthogonality of the coherent image).

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_CrossIsometrySetup.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

namespace OddOrder.Peterfalvi.S12

open OddOrder.RepresentationTheory

open OddOrder.GroupTheory

variable {G : Type*} [Group G]

open scoped Classical FiniteInduce in
/-- **The (5.3.b) base-orthogonality kernel, `S`-side** (the honest content of Coq
`FTtypeP_base_ortho` at an irreducible family member): the `τ_S`-image of the conjugate
difference `λ − λ̄` is orthogonal to every (3.5) grid vector `χ_P`.

`λ − λ̄` is supported on `A(S)` (member supports lie in `A(S) ∪ {1}` and the degrees at `1`
agree), so its Dade image **vanishes on the exceptional set `V`**
(`typeII_tau_apply_eq_zero_of_mem_ticVdiffV` — `V ⊆ A₀(S)` consists of Dade base points
whose value is `(λ − λ̄)(v) = 0` off `S'`).  It is a norm-`2` virtual character (Dade
isometry on the supported pair of distinct orthonormal irreducibles), so the (3.7)/(3.8)
counting kills every σ-coefficient (`sigmaCoeff_eq_zero_of_vanishOnV`) — and the
σ-coefficients **are** the inner products with the grid. -/
theorem typeII_tau_diff_inner_chiFam_eq_zero [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {S : Subgroup G}
    (hSmax : S ∈ maximalSubgroups G) (hSII : IsTypeII S)
    (data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup S)
    [NeZero (Nat.card (typeIIHypothesis46 hG hSmax hSII data.typeP).W1)]
    {Y : Subgroup G} {lam : ClassFunction ↥S ℂ}
    (hlam_mem : lam ∈ OddOrder.Peterfalvi.S11.sOf data Y)
    (hlam_irr : IsIrreducibleCharacter lam)
    (P : ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W1.subgroupOf
        (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W →* ℂˣ) ×
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W2.subgroupOf
        (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W →* ℂˣ)) :
    ClassFunction.inner
      (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0
        (typeIIHypothesis46 hG hSmax hSII data.typeP).tau (lam - lam.conj))
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P) = 0 := by
  classical
  have hModd : Odd (Nat.card ↥S) := hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card S)
  -- family memberships of `λ` and `λ̄` (through the world bridge)
  have hlamIKF := typeII_sOf_subset_inducedKernelFamily data Y hlam_mem
  have hlamBarIKF := OddOrder.Peterfalvi.S08.inducedKernelFamily_closedUnderConjugate
    _ hlamIKF
  -- `λ ≠ λ̄` (no real characters in the family, `|S|` odd)
  have hr : ¬ ClassFunction.IsReal lam :=
    OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters hModd
      (Y.subgroupOf S) hlamIKF
  have hlamne : lam ≠ lam.conj := fun h => hr h.symm
  -- support of the difference: in `A(S)` (degrees at `1` agree)
  have hsupp : (lam - lam.conj).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma S))
        (derivedInG S)) S := by
    have hmem := typeII_T2_member_support hG hSmax hSII 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]
  -- the `A₀`-supported form (the Dade hypothesis' set is `A(S) ∪ V^S`)
  have hsuppA0 : (lam - lam.conj).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma S)) (derivedInG S)
        ∪ conjClassSetIn S (typePV S data.typeP)) S :=
    hsupp.trans (OddOrder.Peterfalvi.S04.supportInSubgroup_mono Set.subset_union_left)
  -- `ψ ∈ ZIrr G`
  have hψZ : OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
      (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0
      (typeIIHypothesis46 hG hSmax hSII data.typeP).tau (lam - lam.conj) ∈ ZIrr G :=
    OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_mem_ZIrr_of_supported
      (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0
      hsuppA0
      (Submodule.sub_mem _
        (OddOrder.Peterfalvi.S08.inducedKernelFamily_mem_ZIrr hlamIKF)
        (OddOrder.Peterfalvi.S08.inducedKernelFamily_mem_ZIrr hlamBarIKF))
  -- `‖ψ‖² = 2` (Dade isometry + distinct orthonormal irreducible pair)
  have hcross : ClassFunction.inner lam lam.conj = 0 := by
    have h := irreducibleCharacter_inner_eq_ite ⟨lam, hlam_irr⟩ ⟨lam.conj, hlam_irr.conj⟩
    rwa [if_neg (fun heq => hlamne (congrArg Subtype.val heq))] at h
  have hcross' : ClassFunction.inner lam.conj lam = 0 := by
    rw [OddOrder.RepresentationTheory.inner_conj_symm, hcross, star_zero]
  have hiso : ClassFunction.inner
      (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0
        (typeIIHypothesis46 hG hSmax hSII data.typeP).tau (lam - lam.conj))
      (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0
        (typeIIHypothesis46 hG hSmax hSII data.typeP).tau (lam - lam.conj))
      = ClassFunction.inner (lam - lam.conj) (lam - lam.conj) :=
    OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_inner_eq_of_supported
      (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0 hsuppA0 hsuppA0
  have hψ2 : ClassFunction.inner
      (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0
        (typeIIHypothesis46 hG hSmax hSII data.typeP).tau (lam - lam.conj))
      (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0
        (typeIIHypothesis46 hG hSmax hSII data.typeP).tau (lam - lam.conj)) = 2 := by
    rw [hiso, ClassFunction.inner_sub_left, ClassFunction.inner_sub_right,
      ClassFunction.inner_sub_right, hlam_irr.inner_self_eq_one,
      hlam_irr.conj.inner_self_eq_one, hcross, hcross']
    norm_num
  -- `ψ` vanishes on `V` (the anchor)
  have hψV : ∀ v ∈ (OddOrder.Peterfalvi.S06.ticVdiff
      (typeIIHypothesis46 hG hSmax hSII data.typeP)).V,
      OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap
        (typeIIHypothesis46 hG hSmax hSII data.typeP).dade0
        (typeIIHypothesis46 hG hSmax hSII data.typeP).tau (lam - lam.conj) v = 0 :=
    fun v hv => typeII_tau_apply_eq_zero_of_mem_ticVdiffV hG hSmax hSII data.typeP hsupp hv
  -- the (3.7)/(3.8) counting kills every σ-coefficient
  have hall := (OddOrder.Peterfalvi.S06.ticVdiff
      (typeIIHypothesis46 hG hSmax hSII data.typeP)).sigmaCoeff_eq_zero_of_vanishOnV rfl
    (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
      (typeIIHypothesis46 hG hSmax hSII data.typeP)) hψZ hψ2 hψV
  exact hall P

open scoped Classical FiniteInduce in
/-- **The (5.3.b) base orthogonality, per `R`-element** (Coq `FTtypeP_base_ortho`): every
member of the dispatched `R(λ)`-family is orthogonal to every (3.5) grid vector `χ_P`.

The constituent trick over the kernel `typeII_tau_diff_inner_chiFam_eq_zero`: an
`R`-element `α` with `⟨α, χ_P⟩ ≠ 0` would satisfy `χ_P = ±α` (both are norm-one virtual
characters, so the integer inner product is `±1` and the difference/sum has norm `0`);
but then the flat expansion `τ_S(λ − λ̄) = ∑_{β ∈ R(λ)} β` (`image_eq`) pairs with `χ_P`
to `⟨α, χ_P⟩ = ±1 ≠ 0` (all other terms die by orthonormality of `R(λ)`), contradicting
the kernel. -/
theorem typeII_R_mem_inner_chiFam_eq_zero [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {S : Subgroup G}
    (hSmax : S ∈ maximalSubgroups G) (hSII : IsTypeII S)
    (data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup S)
    [NeZero (Nat.card (typeIIHypothesis46 hG hSmax hSII data.typeP).W1)]
    {Y : Subgroup G} {lam nu : ClassFunction ↥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)
    {α : ClassFunction G ℂ}
    (hα : α ∈ (typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
      (Set.mem_insert _ _)).imageSet)
    (P : ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W1.subgroupOf
        (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W →* ℂˣ) ×
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W2.subgroupOf
        (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W →* ℂˣ)) :
    ClassFunction.inner α
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P) = 0 := by
  classical
  by_contra hne
  -- both sides are norm-one virtual characters
  have hαZ : α ∈ ZIrr G :=
    (typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
      (Set.mem_insert _ _)).mem_ZIrr α hα
  have hα1 : ClassFunction.inner α α = 1 :=
    (typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
      (Set.mem_insert _ _)).inner_self_of_mem hα
  have hPZ : (OddOrder.Peterfalvi.S06.ticVdiff
      (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
      (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
        (typeIIHypothesis46 hG hSmax hSII data.typeP)) P ∈ ZIrr G :=
    ((OddOrder.Peterfalvi.S06.ticVdiff
      (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam_spec rfl
      (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
        (typeIIHypothesis46 hG hSmax hSII data.typeP))).2.1 P
  have hP1 : ClassFunction.inner
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P)
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P) = 1 := by
    rw [((OddOrder.Peterfalvi.S06.ticVdiff
      (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam_spec rfl
      (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
        (typeIIHypothesis46 hG hSmax hSII data.typeP))).2.2.1, if_pos rfl]
  -- the integer inner product is `±1`
  obtain ⟨c, hc⟩ := inner_intCast_of_mem_ZIrr hαZ hPZ
  have hcne : c ≠ 0 := by
    intro h0
    rw [h0] at hc
    exact hne (by rw [hc]; norm_num)
  have hcstar : ClassFunction.inner
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P) α = (c : ℂ) := by
    rw [OddOrder.RepresentationTheory.inner_conj_symm, hc, star_intCast]
  have hsub : ClassFunction.inner
      (α - (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P)
      (α - (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P) = ((2 - 2 * c : ℤ) : ℂ) := by
    rw [ClassFunction.inner_sub_left, ClassFunction.inner_sub_right,
      ClassFunction.inner_sub_right, hα1, hP1, hc, hcstar]
    push_cast; ring
  have hadd : ClassFunction.inner
      (α + (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P)
      (α + (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P) = ((2 + 2 * c : ℤ) : ℂ) := by
    rw [ClassFunction.inner_add_left, ClassFunction.inner_add_right,
      ClassFunction.inner_add_right, hα1, hP1, hc, hcstar]
    push_cast; ring
  have hle : c ≤ 1 := by
    have h0 := inner_self_re_nonneg (α - (OddOrder.Peterfalvi.S06.ticVdiff
        (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
      (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
        (typeIIHypothesis46 hG hSmax hSII data.typeP)) P)
    rw [hsub, Complex.intCast_re] at h0
    have : (0 : ℤ) ≤ 2 - 2 * c := by exact_mod_cast h0
    omega
  have hge : -1 ≤ c := by
    have h0 := inner_self_re_nonneg (α + (OddOrder.Peterfalvi.S06.ticVdiff
        (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
      (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
        (typeIIHypothesis46 hG hSmax hSII data.typeP)) P)
    rw [hadd, Complex.intCast_re] at h0
    have : (0 : ℤ) ≤ 2 + 2 * c := by exact_mod_cast h0
    omega
  -- `χ_P = ±α`
  have hPeq : (OddOrder.Peterfalvi.S06.ticVdiff
      (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
      (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
        (typeIIHypothesis46 hG hSmax hSII data.typeP)) P = (c : ℂ) • α := by
    interval_cases c
    · -- `c = −1`: `χ_P = −α`
      have hzero : α + (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
          (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
            (typeIIHypothesis46 hG hSmax hSII data.typeP)) P = 0 := by
        refine eq_zero_of_inner_self_re_eq_zero ?_
        rw [hadd]
        norm_num
      rw [show (((-1 : ℤ) : ℂ)) = -1 from by norm_num, neg_one_smul]
      exact eq_neg_of_add_eq_zero_right hzero
    · exact absurd rfl hcne
    · -- `c = 1`: `χ_P = α`
      have hzero : α - (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
          (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
            (typeIIHypothesis46 hG hSmax hSII data.typeP)) P = 0 := by
        refine eq_zero_of_inner_self_re_eq_zero ?_
        rw [hsub]
        norm_num
      rw [show (((1 : ℤ) : ℂ)) = 1 from by norm_num, one_smul]
      exact (sub_eq_zero.mp hzero).symm
  -- the flat `R`-expansion pairs to `±1 ≠ 0`, contradicting the kernel
  have hk1 := typeII_tau_diff_inner_chiFam_eq_zero hG hSmax hSII data hlam_mem hlam_irr P
  rw [(typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
      (Set.mem_insert _ _)).image_eq,
    OddOrder.RepresentationTheory.inner_sum_left] at hk1
  have hsingle : ∀ β ∈ (typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
      (Set.mem_insert _ _)).imageSet, β ≠ α →
      ClassFunction.inner β
        ((OddOrder.Peterfalvi.S06.ticVdiff
            (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
          (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
            (typeIIHypothesis46 hG hSmax hSII data.typeP)) P) = 0 := by
    intro β hβ hβne
    rw [hPeq, OddOrder.RepresentationTheory.inner_smul_right,
      (typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
        (Set.mem_insert _ _)).orthonormal β hβ α hα, if_neg hβne]
    simp
  rw [Finset.sum_eq_single α hsingle (fun h => absurd hα h)] at hk1
  rw [hPeq, OddOrder.RepresentationTheory.inner_smul_right, hα1, mul_one] at hk1
  have : (c : ℂ) = 0 := star_eq_zero.mp hk1
  exact hcne (by exact_mod_cast this)

open scoped Classical FiniteInduce in
/-- **Peterfalvi (5.5) for the `T2`-coherence** (the lane-b template
`coherent_extension_mem_span_Rset_of_mem` re-enacted at the type-II member family): the
coherent image of the irreducible member `λ` lies in `ℤ[R(λ)]`.

The `ofProjection (ψ := 0)` engine against the dispatched `R(λ)`-family
(`typeII_T2_memberRFamily`): the (5.4) decomposition `λ^{τ₂} = X − Y` has `‖Y‖² ≥ 0 = ‖ψ‖²`
automatically, so (5.4.b) forces `Y = 0` and `λ^{τ₂} = X ∈ ℤ[R(λ)]`
(`eq_sum_of_psi_eq_zero`).  The isometry/agreement inputs are the coherence's own fields on
the sponsoring lattice `ℤ[λ, λ̄]  ⊆ ℤ[T2]`; the support input is the `A₀(S)`-support of
`λ − λ̄` (as in the kernel `typeII_tau_diff_inner_chiFam_eq_zero`). -/
theorem typeII_T2_extension_lam_mem_span_RFamily [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {S : Subgroup G}
    (hSmax : S ∈ maximalSubgroups G) (hSII : IsTypeII S)
    (data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup S)
    [NeZero (Nat.card (typeIIHypothesis46 hG hSmax hSII data.typeP).W1)]
    {Y : Subgroup G} {lam nu : ClassFunction ↥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 hSmax hSII data.typeP).dade0
        (typeIIHypothesis46 hG hSmax hSII data.typeP).tau)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥S ℂ))
      (OddOrder.Peterfalvi.S04.supportInSubgroup
        (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma S))
            (derivedInG S)
          ∪ conjClassSetIn S (typePV S data.typeP)) S)) :
    c.extension lam ∈ Submodule.span ℤ
      ((typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
        (Set.mem_insert _ _)).imageSet : Set (ClassFunction G ℂ)) := by
  classical
  have hModd : Odd (Nat.card ↥S) := hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card S)
  -- `λ ≠ λ̄`, and the `A₀(S)`-support of the difference (as in the kernel)
  have hlamIKF := typeII_sOf_subset_inducedKernelFamily data Y hlam_mem
  have hr : ¬ ClassFunction.IsReal lam :=
    OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters hModd
      (Y.subgroupOf S) hlamIKF
  have hlamne : lam ≠ lam.conj := fun h => hr h.symm
  have hsupp : (lam - lam.conj).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
      (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma S))
        (derivedInG S)) S := by
    have hmem := typeII_T2_member_support hG hSmax hSII 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 S)) (derivedInG S)
        ∪ conjClassSetIn S (typePV S data.typeP)) S :=
    hsupp.trans (OddOrder.Peterfalvi.S04.supportInSubgroup_mono Set.subset_union_left)
  -- `⟨λ, λ̄⟩ = 0`
  have hcross : ClassFunction.inner lam lam.conj = 0 := by
    have h := irreducibleCharacter_inner_eq_ite ⟨lam, hlam_irr⟩ ⟨lam.conj, hlam_irr.conj⟩
    rwa [if_neg (fun heq => hlamne (congrArg Subtype.val heq))] at h
  -- lattice memberships
  have hchi_zSpan : lam ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥S)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥S ℂ)) :=
    Submodule.subset_span (Set.mem_insert _ _)
  have hchibar_zSpan : lam.conj ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥S)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥S ℂ)) :=
    Submodule.subset_span (Set.mem_insert_of_mem _ (Set.mem_insert _ _))
  have hdiff_zSpan : lam - lam.conj ∈ OddOrder.Peterfalvi.S07.zSpan (L := ↥S)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥S ℂ)) :=
    Submodule.sub_mem _ hchi_zSpan hchibar_zSpan
  have hsub : OddOrder.Peterfalvi.S07.zSpan (L := ↥S)
      ({lam - lam.conj, lam - 0} : Set (ClassFunction ↥S ℂ)) ≤
      OddOrder.Peterfalvi.S07.zSpan (L := ↥S)
        ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥S ℂ)) := by
    change Submodule.span ℤ _ ≤ Submodule.span ℤ _
    rw [Submodule.span_le]
    intro x hx
    simp only [Set.mem_insert_iff, Set.mem_singleton_iff] at hx
    rcases hx with rfl | rfl
    · exact hdiff_zSpan
    · rw [sub_zero]; exact hchi_zSpan
  -- the (5.4)/(5.5) engine at `ψ = 0`
  obtain ⟨-, hτ1χ, E, hEsub, hXsum, -⟩ :=
    OddOrder.Peterfalvi.S07.CharacterPsiDecomposition.eq_sum_of_psi_eq_zero
      (OddOrder.Peterfalvi.S07.CharacterPsiDecomposition.ofProjection (ψ := 0)
        (typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
          (Set.mem_insert _ _)) c.extension
        (fun φ ζ hφ hζ => c.extension_inner_eq φ ζ (hsub hφ) (hsub hζ))
        (c.extends_on_supported (lam - lam.conj) ⟨hdiff_zSpan, hsuppA0⟩)
        (by rw [sub_zero]; exact c.extension_mem_ZIrr _ hchi_zSpan)
        (ClassFunction.inner_zero_right _)
        (ClassFunction.inner_zero_right _)
        hcross)
  have hgoal : c.extension lam = ∑ α ∈ E, α := hτ1χ.trans hXsum
  rw [hgoal]
  exact Submodule.sum_mem _ fun α hα =>
    Submodule.subset_span (Finset.mem_coe.mpr (hEsub hα))

open scoped Classical FiniteInduce in
/-- **The (5.3.b) `S`-side grid orthogonality of the coherent image** (Coq
`coherent_ortho_cycTIiso` at the irreducible member): `⟨χ_P, λ^{τ₂}⟩ = 0` for every (3.5)
grid vector.  `λ^{τ₂} ∈ ℤ[R(λ)]` by (5.5) (`typeII_T2_extension_lam_mem_span_RFamily`) and
every `R(λ)`-element is orthogonal to `χ_P` (`typeII_R_mem_inner_chiFam_eq_zero`), so the
inner product dies by ℤ-linearity of the right slot. -/
theorem typeII_T2_extension_lam_inner_chiFam_eq_zero [Finite G]
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) {S : Subgroup G}
    (hSmax : S ∈ maximalSubgroups G) (hSII : IsTypeII S)
    (data : OddOrder.Peterfalvi.S11.TypesIIIIIIVSetup S)
    [NeZero (Nat.card (typeIIHypothesis46 hG hSmax hSII data.typeP).W1)]
    {Y : Subgroup G} {lam nu : ClassFunction ↥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 hSmax hSII data.typeP).dade0
        (typeIIHypothesis46 hG hSmax hSII data.typeP).tau)
      ({lam, lam.conj, nu, nu.conj} : Set (ClassFunction ↥S ℂ))
      (OddOrder.Peterfalvi.S04.supportInSubgroup
        (centralizerSupport (sharpSubgroup (OddOrder.BG.Ch3.S10.Msigma S))
            (derivedInG S)
          ∪ conjClassSetIn S (typePV S data.typeP)) S))
    (P : ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W1.subgroupOf
        (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W →* ℂˣ) ×
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W2.subgroupOf
        (OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).W →* ℂˣ)) :
    ClassFunction.inner
      ((OddOrder.Peterfalvi.S06.ticVdiff
          (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
        (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
          (typeIIHypothesis46 hG hSmax hSII data.typeP)) P)
      (c.extension lam) = 0 := by
  classical
  have key : ∀ x ∈ Submodule.span ℤ
      ((typeII_T2_memberRFamily hG hSmax hSII data hlam_mem hnu_mem hdeg
        (Set.mem_insert _ _)).imageSet : Set (ClassFunction G ℂ)),
      ClassFunction.inner
        ((OddOrder.Peterfalvi.S06.ticVdiff
            (typeIIHypothesis46 hG hSmax hSII data.typeP)).chiFam rfl
          (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
            (typeIIHypothesis46 hG hSmax hSII data.typeP)) P) x = 0 := by
    intro x hx
    induction hx using Submodule.span_induction with
    | mem y hy =>
        have h0 := typeII_R_mem_inner_chiFam_eq_zero hG hSmax hSII data
          hlam_mem hlam_irr hnu_mem hdeg (Finset.mem_coe.mp hy) P
        rw [OddOrder.RepresentationTheory.inner_conj_symm, h0, star_zero]
    | zero => exact ClassFunction.inner_zero_right _
    | add y z _ _ ihy ihz => rw [ClassFunction.inner_add_right, ihy, ihz, add_zero]
    | smul a y _ ih =>
        rw [← Int.cast_smul_eq_zsmul ℂ a y,
          OddOrder.RepresentationTheory.inner_smul_right, ih, mul_zero]
  exact key _ (typeII_T2_extension_lam_mem_span_RFamily hG hSmax hSII data
    hlam_mem hlam_irr hnu_mem hdeg c)

open scoped Classical FiniteInduce in
/-- **The (5.3.b) `λ`-orthogonality against the aligned `M`-grid at the canonical pair**
(the `lam_ortho_grid` production; Coq `coherent_ortho_cycTIiso` consumed at `etaC`): the
`T2`-coherent image of the irreducible member `λ` is orthogonal to every `M`-side aligned
σ-grid vector.

Chain: the aligned grid vector is an `M`-side `χ`-family member
(`exists_alignedOmegaSigmaGrid_chiFam_family`); the per-index pair transpose reads it as an
`S`-side grid vector (`section16_pair_chiFam_transpose_T`); and the `S`-side grid is
orthogonal to `λ^{τ₂}` by (5.5) + the constituent trick
(`typeII_T2_extension_lam_inner_chiFam_eq_zero`). -/
theorem Hypothesis.extension_lam_inner_alignedGrid_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)]
    {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))
    (i : Fin hyp.w1) (j : Fin hyp.w2) :
    ClassFunction.inner (hyp.alignedOmegaSigmaGrid hG hG.odd i j) (c.extension lam) = 0 := by
  classical
  subst hT
  -- the aligned vector is an `M`-side `χ`-family member
  obtain ⟨P, -, hP⟩ := hyp.exists_alignedOmegaSigmaGrid_chiFam_family hG hG.odd i
  rw [hP j]
  -- transpose to an `S`-side grid vector, then apply the `S`-side orthogonality
  rw [section16_pair_chiFam_transpose_T hG hG.odd hSW1 hSW2 hyp.typeP hKstar.symm
    (OddOrder.Peterfalvi.S06.ticVdiffFullDadeApplication
      (typeIIHypothesis46 hG mp.S_maximal (section16_S_isTypeII hG mp) data.typeP))
    (hyp.canonicalFullDadeApp hG hG.odd) (P j)]
  exact typeII_T2_extension_lam_inner_chiFam_eq_zero hG mp.S_maximal
    (section16_S_isTypeII hG mp) data hlam_mem hlam_irr hnu_mem hdeg c _

open scoped Classical FiniteInduce in
/-- **Peterfalvi (5.3.b), `M`-side: `ζ^{τ₁}` is orthogonal to the aligned σ-grid** (the
`zeta_ortho_grid` production — the full-`𝒮`-coherence form of the grid-orthogonality
intermediate of `tau1_zeta_vanishes_on_typePV`; cf. its `S(HC)`-port
`SHC_extension_inner_alignedOmegaSigma_eq_zero`).

`(ζ − ζ̄)^τ = ζ^{τ₁} − ζ̄^{τ₁}` (`tau_zeta_sub_conj_eq_tau1`) vanishes on `V`
(`tau_zeta_sub_conj_vanishes_on_typePV`) and has at most `2 < min(w₁, w₂)` nonzero
σ-coefficients (each `τ₁`-image is norm-`1` with at most one,
`ncard_inner_chiFam_ne_zero_le_one`), so every σ-coefficient dies
(`sigmaCoeff_eq_zero_of_sigmaNC_lt`); the norm-`1` projection
(`inner_left_eq_zero_of_inner_sub_eq_zero`) upgrades the difference orthogonality to
`⟨ζ^{τ₁}, χ_P⟩ = 0`, and the aligned grid vector **is** a `χ_P`
(`exists_alignedOmegaSigmaGrid_chiFam_family`). -/
theorem Hypothesis.tau1_zeta_inner_alignedGrid_eq_zero [Finite G] {M : Subgroup G}
    (hG : OddOrder.BG.IsMinimalSimpleOdd G) (hyp : Hypothesis M) (hodd : Odd (Nat.card G))
    {params : CharacterParameters hyp} (coh : CoherentHypothesis hyp params)
    {ζ : ClassFunction ↥M ℂ} (hζS : ζ ∈ inducedFamily M) (hζirr : IsIrreducibleCharacter ζ)
    (i : Fin hyp.w1) (j : Fin hyp.w2) :
    ClassFunction.inner (coh.tau1 ζ) (hyp.alignedOmegaSigmaGrid hG hodd i j) = 0 := by
  have := hyp.finiteG
  classical
  have hModd : Odd (Nat.card ↥M) := hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card M)
  have hζne : ζ.conj ≠ ζ := inducedFamily_hasNoRealCharacters hModd hζS
  let tic := typePData_toTICyclicHypothesis hyp.typeP hodd
  have : NeZero (Nat.card ↥tic.W1) := ⟨Nat.card_pos.ne'⟩
  have : NeZero (Nat.card ↥tic.W2) := ⟨Nat.card_pos.ne'⟩
  let app := hyp.canonicalFullDadeApp hG hodd
  have hVeq : tic.V = tic.Vdiff := rfl
  obtain ⟨P, -, hPeq⟩ := hyp.exists_alignedOmegaSigmaGrid_chiFam_family hG hodd i
  rw [hPeq j]
  have hζcS : ζ.conj ∈ inducedFamily M := inducedFamily_closedUnderConjugate M hζS
  have hζcirr : IsIrreducibleCharacter ζ.conj := hζirr.conj
  have haZ : coh.tau1 ζ ∈ ZIrr G :=
    coh.coherent.extension_mem_ZIrr ζ (Submodule.subset_span hζS)
  have hbZ : coh.tau1 ζ.conj ∈ ZIrr G :=
    coh.coherent.extension_mem_ZIrr ζ.conj (Submodule.subset_span hζcS)
  have ha1 : ClassFunction.inner (coh.tau1 ζ) (coh.tau1 ζ) = 1 :=
    hyp.zeta_tau1_inner_self hG hodd coh hζS hζirr
  have hb1 : ClassFunction.inner (coh.tau1 ζ.conj) (coh.tau1 ζ.conj) = 1 :=
    hyp.zeta_tau1_inner_self hG hodd coh hζcS hζcirr
  have hab : ClassFunction.inner (coh.tau1 ζ) (coh.tau1 ζ.conj) = 0 := by
    change ClassFunction.inner (coh.coherent.extension ζ) (coh.coherent.extension ζ.conj) = 0
    rw [coh.coherent.extension_inner_eq _ _ (Submodule.subset_span hζS)
        (Submodule.subset_span hζcS),
      OddOrder.RepresentationTheory.irr_cf_inner hζirr hζcirr, if_neg (fun h => hζne h.symm)]
  -- `(ζ − ζ̄)^τ` vanishes on `V`, with `NC ≤ 2 < min(w₁, w₂)`
  have hvanish : ∀ w ∈ tic.V, hyp.tau (ζ - ζ.conj) w = 0 := fun w hw =>
    hyp.tau_zeta_sub_conj_vanishes_on_typePV hG hodd hζS hζirr hw
  have hNC : tic.sigmaNC hVeq app (hyp.tau (ζ - ζ.conj))
      < min (Nat.card ↥tic.W1) (Nat.card ↥tic.W2) := by
    have hbound : tic.sigmaNC hVeq app (hyp.tau (ζ - ζ.conj)) ≤ 2 := by
      have hsub : {pq | tic.sigmaCoeff hVeq app (hyp.tau (ζ - ζ.conj)) pq ≠ 0} ⊆
          {pq | ClassFunction.inner (coh.tau1 ζ) (tic.chiFam hVeq app pq) ≠ 0} ∪
          {pq | ClassFunction.inner (coh.tau1 ζ.conj) (tic.chiFam hVeq app pq) ≠ 0} := by
        intro pq hpq
        by_contra hcon
        simp only [Set.mem_union, Set.mem_ofPred_eq, not_or, not_not] at hcon
        apply hpq
        change ClassFunction.inner (hyp.tau (ζ - ζ.conj)) (tic.chiFam hVeq app pq) = 0
        rw [hyp.tau_zeta_sub_conj_eq_tau1 hG hodd coh hζS hζirr,
          ClassFunction.inner_sub_left, hcon.1, hcon.2, sub_zero]
      calc tic.sigmaNC hVeq app (hyp.tau (ζ - ζ.conj))
          = {pq | tic.sigmaCoeff hVeq app (hyp.tau (ζ - ζ.conj)) pq ≠ 0}.ncard := rfl
        _ ≤ ({pq | ClassFunction.inner (coh.tau1 ζ) (tic.chiFam hVeq app pq) ≠ 0} ∪
              {pq | ClassFunction.inner (coh.tau1 ζ.conj)
                (tic.chiFam hVeq app pq) ≠ 0}).ncard :=
            Set.ncard_le_ncard hsub (Set.toFinite _)
        _ ≤ {pq | ClassFunction.inner (coh.tau1 ζ) (tic.chiFam hVeq app pq) ≠ 0}.ncard +
              {pq | ClassFunction.inner (coh.tau1 ζ.conj)
                (tic.chiFam hVeq app pq) ≠ 0}.ncard :=
            Set.ncard_union_le _ _
        _ ≤ 1 + 1 := by
            gcongr
            · exact tic.ncard_inner_chiFam_ne_zero_le_one hVeq app haZ ha1
            · exact tic.ncard_inner_chiFam_ne_zero_le_one hVeq app hbZ hb1
        _ = 2 := rfl
    have h3a := tic.three_le_card_W1
    have h3b := tic.three_le_card_W2
    omega
  -- σ-coefficient of the difference at `P j` dies; project to the `ζ^{τ₁}`-slot
  have hL3 : tic.sigmaCoeff hVeq app (hyp.tau (ζ - ζ.conj)) (P j) = 0 :=
    tic.sigmaCoeff_eq_zero_of_sigmaNC_lt hVeq app hvanish hNC (P j)
  have hdiff : ClassFunction.inner (coh.tau1 ζ - coh.tau1 ζ.conj)
      (tic.chiFam hVeq app (P j)) = 0 := by
    rw [← hyp.tau_zeta_sub_conj_eq_tau1 hG hodd coh hζS hζirr]; exact hL3
  have hsZ : tic.chiFam hVeq app (P j) ∈ ZIrr G := (tic.chiFam_spec hVeq app).2.1 (P j)
  have hs1 : ClassFunction.inner (tic.chiFam hVeq app (P j))
      (tic.chiFam hVeq app (P j)) = 1 := by
    rw [(tic.chiFam_spec hVeq app).2.2.1, if_pos rfl]
  exact inner_left_eq_zero_of_inner_sub_eq_zero haZ hsZ ha1 hb1 hs1 hab hdiff

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_CrossIsometrySetup.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