Peterfalvi (9.9.b): the -family dispatch and case (9.7.b), §9 level
DefinitionOddOrder_Peterfalvi_S11_NineElevenRFamilyfeit-thompsonfinite-groupsgroup-theoryishida-odd-order
Formalizes the (9.9.b) dichotomy, the per-member orthonormal -family dispatch, and case (9.7.b) at the §9 level (Peterfalvi, Character Theory for the Odd Order Theorem, §8 (8.15.3), §9 (9.5), pp. 47-51).
Main declarations: sOf_columnSum_of_not_irreducible (Peterfalvi (9.9.b) at §9 level); sOf_memberRFamily (Per-member orthonormal R-family over 𝒮(Y), at §9 level); sOf_memberRFamily_imageSet_of_irr (sOf_memberRFamily reduction, irreducible case); sOf_memberRFamily_imageSet_of_col (sOf_memberRFamily reduction, column case).
The bundle also proves 2 supporting lemmas used by later statements.
Definition code
/-
Copyright (c) 2026 Yawara Ishida. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Yawara Ishida
-/
-- From github.com/yawara/odd-order @ 20a6fccbb63c (Apache-2.0; LICENSE at that commit):
-- OddOrder/Peterfalvi/S11_NineElevenRFamily.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_S06_CertainTypeSubcoherent
import Definitions.Def_OddOrder_Peterfalvi_S07_PivotCoherence
import Definitions.Def_OddOrder_Peterfalvi_S11_NineElevenBridgeBase
import Definitions.Def_OddOrder_Peterfalvi_S10_SubcoherentTypeP
import Definitions.Def_OddOrder_Peterfalvi_S11_MaximalII_III_IV_ThetaCountAssembly
import Definitions.Def_OddOrder_Peterfalvi_S11_NineElevenCoherence
import Definitions.Def_OddOrder_Peterfalvi_S11_SingleFactorCentralizer
namespace OddOrder.Peterfalvi.S11
open OddOrder.GroupTheory
open OddOrder.RepresentationTheory
variable {G : Type*} [Group G]
/-- **Peterfalvi (9.9.b), member form** (§6 level, no transport): a reducible induced character
from `h46.K` is a certain-type column sum.
This is the content the §9 caseB `R`-family dispatch needs, isolated so that **no coercion between
`↥h46.K` and `↥(huSub data)` appears** — everything stays inside `h46.K`. An earlier attempt stated
it over `sOf data Y` directly and tried to move the source character across
`h46.K = huSub data`; that rewrite is not type-correct (the motive mentions
`IrreducibleCharacter ↥_a` *and* `chiRestrict χ₂ = χ`, both dependent on the subgroup). Splitting
it this way keeps the dependent step transport-free and leaves the §9 identification to the
Set-valued level, where the `▸` idiom does work (cf. `sOf_subset_inducedNonKernelFamily`).
The §13 analogue is `S13.caseB_sOf_member_dichotomy`, whose conclusion is phrased in the §10 μ-grid;
the book builds these members from (4.7) and Theorem (4.5), both §6, so the `columnSum` form here is
the faithful one. -/
theorem induce_columnSum_of_not_irreducible {M : Subgroup G} {A : Set G} [Finite G] [Fintype G]
[Fintype ↥M]
[Invertible (Nat.card G : ℂ)] [Invertible (Nat.card ↥M : ℂ)]
(h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
[Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)] [Invertible (Nat.card ↥h46.K : ℂ)]
(χ : OddOrder.RepresentationTheory.IrreducibleCharacter ↥h46.K)
(hχne : χ ≠ OddOrder.RepresentationTheory.trivialIrreducibleCharacter ↥h46.K)
(hred : ¬ IsIrreducibleCharacter
(ClassFunction.induce h46.K (χ : ClassFunction ↥h46.K ℂ))) :
∃ χ₂ ≠ 1, ClassFunction.induce h46.K (χ : ClassFunction ↥h46.K ℂ)
= OddOrder.Peterfalvi.S06.columnSum h46 χ₂ := by
obtain ⟨χ₂, hχ₂⟩ := (h46.induce_not_isIrreducible_iff χ).mp hred
refine ⟨χ₂, ?_, ?_⟩
· -- `χ₂ = 1` would make the source trivial (`chiRestrict_one_eq_trivial`)
rintro rfl
rw [h46.chiRestrict_one_eq_trivial] at hχ₂
exact hχne hχ₂.symm
· rw [← hχ₂, OddOrder.Peterfalvi.S06.columnSum_def]
exact h46.induce_restrict_certainType_eq χ₂
/-- **Transporting an induction source across an equality of induction subgroups.** For `K = K'`
an irreducible `K'`-character has an irreducible `K`-counterpart with the *same* induction, and the
two are trivial together.
The equality is discharged by `subst` (both subgroups are variables here), which is exactly why the
lemma is stated in this generality rather than at `h46.K = huSub data`: there `h46.K` is a
projection, so `subst` does not apply and a direct `▸` on the dependent character type breaks the
motive (cf. the docstring of `induce_columnSum_of_not_irreducible`). The residual `Invertible`
mismatch after `subst` is propositional (`Subsingleton`), the same wrinkle
`induceHU_eq_induce` handles. -/
theorem exists_induce_eq_of_subgroup_eq {L : Type*} [Group L] [Fintype L] {K K' : Subgroup L}
(hKK' : K = K') [Invertible (Nat.card ↥K : ℂ)] [Invertible (Nat.card ↥K' : ℂ)]
(χ : OddOrder.RepresentationTheory.IrreducibleCharacter ↥K') :
∃ χ' : OddOrder.RepresentationTheory.IrreducibleCharacter ↥K,
ClassFunction.induce K (χ' : ClassFunction ↥K ℂ)
= ClassFunction.induce K' (χ : ClassFunction ↥K' ℂ) ∧
(χ' = OddOrder.RepresentationTheory.trivialIrreducibleCharacter ↥K ↔
χ = OddOrder.RepresentationTheory.trivialIrreducibleCharacter ↥K') := by
subst hKK'
refine ⟨χ, ?_, Iff.rfl⟩
convert rfl using 2
exact Subsingleton.elim (α := Invertible (Nat.card ↥K : ℂ)) _ _
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **Peterfalvi (9.9.b) at §9 level**: a *reducible* member of `𝒮(Y)` is a nontrivial certain-type
column sum `μ_{χ₂}` of Hypothesis (4.6).
This is the §9 replacement for `S13.caseB_sOf_member_dichotomy`, whose conclusion is stated in the
§10 μ-grid (`hyp.base.muColumnChar`) and therefore drags the whole §10/§11 packaging — and with it
the type III/IV restriction — into the caseB `R`-family dispatch. The book has no such detour: it
builds the reducible members from (4.7) and Theorem (4.5), both §6 results, which is exactly the
`S06.columnSum` form produced here. (§10 converts §6-columns to μ-grid columns and `certainTypeR`
converts them back, so the packaging route is a round trip; issue 1045.)
The `𝒮`-side source is nontrivial because `𝒳` demands `H ⊄ Ker χ` while the trivial character has
kernel everything, and `hKeq` moves it into the (4.6) world where
`induce_columnSum_of_not_irreducible` applies. Callers get `hKeq` from
`huSub_eq_derivedInG_subgroupOf` together with the `K`-field of the (4.6) producer
(`S10.typePACore_toHypothesis46_core`). -/
theorem sOf_columnSum_of_not_irreducible [Finite G] {M : Subgroup G} {A : Set G}
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
[Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data) {Y : Subgroup G} {φ : ClassFunction ↥M ℂ}
(hφ : φ ∈ sOf data Y) (hred : ¬ IsIrreducibleCharacter φ) :
∃ χ₂ ≠ 1, φ = OddOrder.Peterfalvi.S06.columnSum h46 χ₂ := by
obtain ⟨χ, hχ, rfl⟩ := hφ
-- `χ ≠ 1`: the trivial character's kernel is everything, contradicting `H ⊄ Ker χ` (`𝒳`).
have hχne : χ ≠ OddOrder.RepresentationTheory.trivialIrreducibleCharacter ↥(huSub data) := by
intro htriv
apply hχ.1
rw [htriv]
simp only [OddOrder.RepresentationTheory.IrreducibleCharacter.coe_trivialIrreducibleCharacter,
OddOrder.Peterfalvi.S03.characterKernel_trivialClassFunction]
exact Set.subset_univ _
obtain ⟨χ', hind, htriviff⟩ := exists_induce_eq_of_subgroup_eq hKeq χ
have hred' : ¬ IsIrreducibleCharacter
(ClassFunction.induce h46.K (χ' : ClassFunction ↥h46.K ℂ)) := by
rw [hind, ← induceHU_eq_induce data (χ : ClassFunction ↥(huSub data) ℂ)]
exact hred
obtain ⟨χ₂, hχ₂ne, heq⟩ :=
induce_columnSum_of_not_irreducible h46 χ' (fun h => hχne (htriviff.mp h)) hred'
exact ⟨χ₂, hχ₂ne,
(induceHU_eq_induce data (χ : ClassFunction ↥(huSub data) ℂ)).trans (hind.symm.trans heq)⟩
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **Per-member orthonormal `R`-family over `𝒮(Y)`, at §9 level** — the raw (5.2.d) datum feeding
the norm-general (5.7) engine `S07.uniform_degree_coherence_of_families` in case (9.7.b) of (9.11).
Every member of `𝒮(Y)` is either irreducible or, by (9.9.b)
(`sOf_columnSum_of_not_irreducible`), a nontrivial certain-type column `μ_{χ₂}`; the `R`-family is
dispatched accordingly:
* **irreducible `η`** — the 2-element signed Dade family
`S07.dadeOrthonormalCharacterImageFamilyOfDiff` (`τ(η − η̄) = ε·(μ − ν)`), whose no-realness and
`A₀`-supported difference come from the `⊥`-kernel world-bridge
(`sOf_subset_inducedKernelFamily_bot`) rather than from a `S13.Hypothesis`;
* **column `η = μ_{χ₂}`** — the `2q`-element certain-type family `S06.certainTypeR`.
This is the §9 replacement for `S13.caseB_sOf_memberRFamily`. The two differ only in where the
reducible branch gets its column: the §13 version reads a μ-grid index `k : Fin hyp.base.w2` off the
§10 packaging (which is what confined it to types III/IV), while here it is the §6 column `χ₂` that
`certainTypeR` consumes anyway. **No type hypothesis appears on this route.**
⚠ `τ` is *not* a free parameter here (unlike case (9.7.a)'s `caseA_coherent_sOf_cprime_of_refuter`):
`certainTypeR` produces its family over `dadeIntegralCharacterMap h46.dade0 h46.tau`, so the
conclusion is pinned to it, and the irreducible branch — which lands on
`h46.dade0.fullDadeIsometryData hconj` — is matched to it by `htau`. For the intended producer
`S10.typePACore_toHypothesis46_core`, which stores `tau := dade0.fullDadeIsometryData hconj`
verbatim, `htau` is `rfl`.
⚠ The family is assembled field-by-field rather than by transporting along `htau` / `η = μ_{χ₂}`,
so that `.imageSet` is *definitionally* the underlying constructor's — the form in which the (5.2.e)
cross-orthogonality lemmas consume it (same reason as in the §13 version). -/
noncomputable def sOf_memberRFamily [Finite G] {M : Subgroup G} {A : Set G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G) (hM : M ∈ maximalSubgroups G)
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
[Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data)
(htau : h46.tau = h46.dade0.fullDadeIsometryData)
(hKsupp : ∀ x : ↥M, x ∈ (derivedInG M).subgroupOf M → x ≠ 1 →
x ∈ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
{Y : Subgroup G} {η : ClassFunction ↥M ℂ} (hη : η ∈ sOf data Y) :
OddOrder.Peterfalvi.S07.OrthonormalCharacterImageFamily
(OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau) η := by
classical
-- `M' ⊴ M`; in the §13 version this instance arrived through the packaging's import closure.
haveI := derivedInG_subgroupOf_normal M
have hηIKF0 := sOf_subset_inducedKernelFamily_bot hG hM data Y hη
by_cases hirr : IsIrreducibleCharacter η
· -- irreducible member: the signed Dade image family, matched to `h46.tau` by `htau`
have hreal : ¬ ClassFunction.IsReal (η : ClassFunction ↥M ℂ) :=
OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters
(hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card M)) (⊥ : Subgroup ↥M) hηIKF0
have hdiffsupp := OddOrder.Peterfalvi.S08.inducedKernelFamily_conjDiff_support hKsupp hηIKF0
exact
{ imageSet := (OddOrder.Peterfalvi.S07.dadeOrthonormalCharacterImageFamilyOfDiff
h46.dade0 ⟨η, hirr⟩ hreal hdiffsupp).imageSet
mem_ZIrr := (OddOrder.Peterfalvi.S07.dadeOrthonormalCharacterImageFamilyOfDiff
h46.dade0 ⟨η, hirr⟩ hreal hdiffsupp).mem_ZIrr
orthonormal := (OddOrder.Peterfalvi.S07.dadeOrthonormalCharacterImageFamilyOfDiff
h46.dade0 ⟨η, hirr⟩ hreal hdiffsupp).orthonormal
image_eq := by
rw [htau]
exact (OddOrder.Peterfalvi.S07.dadeOrthonormalCharacterImageFamilyOfDiff
h46.dade0 ⟨η, hirr⟩ hreal hdiffsupp).image_eq }
· -- reducible member: (9.9.b) gives the column `χ₂ ≠ 1`, and `certainTypeR` its `2q`-family
exact OddOrder.Peterfalvi.S06.columnR h46
(sOf_columnSum_of_not_irreducible data h46 hKeq hη hirr).choose_spec.1
(sOf_columnSum_of_not_irreducible data h46 hKeq hη hirr).choose_spec.2
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **`sOf_memberRFamily` reduction, irreducible case**: for an irreducible member the dispatched
family *is* `S07.dadeOrthonormalCharacterImageFamilyOfDiff` (imageSet form). The realness and
support proofs are existential — they are proof-irrelevant inputs to a proof-independent
`imageSet` — so the (5.2.e) lemmas apply after rewriting. -/
theorem sOf_memberRFamily_imageSet_of_irr [Finite G] {M : Subgroup G} {A : Set G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G) (hM : M ∈ maximalSubgroups G)
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
[Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data)
(htau : h46.tau = h46.dade0.fullDadeIsometryData)
(hKsupp : ∀ x : ↥M, x ∈ (derivedInG M).subgroupOf M → x ≠ 1 →
x ∈ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
{Y : Subgroup G} {η : ClassFunction ↥M ℂ} (hη : η ∈ sOf data Y)
(hirr : IsIrreducibleCharacter η) :
∃ (hr : ¬ ClassFunction.IsReal (η : ClassFunction ↥M ℂ))
(hs : ((η : ClassFunction ↥M ℂ).conj - (η : ClassFunction ↥M ℂ)).support ⊆
OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M),
(sOf_memberRFamily hG hM data h46 hKeq htau hKsupp hη).imageSet =
(OddOrder.Peterfalvi.S07.dadeOrthonormalCharacterImageFamilyOfDiff
h46.dade0 ⟨η, hirr⟩ hr hs).imageSet := by
classical
have := derivedInG_subgroupOf_normal M
have hηIKF0 := sOf_subset_inducedKernelFamily_bot hG hM data Y hη
refine ⟨OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters
(hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card M)) (⊥ : Subgroup ↥M) hηIKF0,
OddOrder.Peterfalvi.S08.inducedKernelFamily_conjDiff_support hKsupp hηIKF0, ?_⟩
unfold sOf_memberRFamily
rw [dif_pos hirr]
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **`sOf_memberRFamily` reduction, column case**: for a reducible member the dispatched family
*is* `S06.certainTypeR` at the (9.9.b) column `χ₂` (imageSet form), exposed together with the
membership equation `η = μ_{χ₂}` — which is what supplies the `≠`-side conditions of the μ×μ and
μ×irr cross-orthogonality. -/
theorem sOf_memberRFamily_imageSet_of_col [Finite G] {M : Subgroup G} {A : Set G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G) (hM : M ∈ maximalSubgroups G)
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
[Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data)
(htau : h46.tau = h46.dade0.fullDadeIsometryData)
(hKsupp : ∀ x : ↥M, x ∈ (derivedInG M).subgroupOf M → x ≠ 1 →
x ∈ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
{Y : Subgroup G} {η : ClassFunction ↥M ℂ} (hη : η ∈ sOf data Y)
(hcol : ¬ IsIrreducibleCharacter η) :
∃ (χ₂ : (h46.W2.subgroupOf (h46.W1 ⊔ h46.W2)) →* ℂˣ) (hχ₂ : χ₂ ≠ 1),
η = OddOrder.Peterfalvi.S06.columnSum h46 χ₂ ∧
(sOf_memberRFamily hG hM data h46 hKeq htau hKsupp hη).imageSet =
(OddOrder.Peterfalvi.S06.certainTypeR h46 hχ₂
(OddOrder.Peterfalvi.S06.columnSum_inv_apply_one h46 χ₂).symm).imageSet := by
classical
have hex := sOf_columnSum_of_not_irreducible data h46 hKeq hη hcol
refine ⟨hex.choose, hex.choose_spec.1, hex.choose_spec.2, ?_⟩
unfold sOf_memberRFamily
rw [dif_neg hcol]
rfl
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **The Dade image of an `A₁`-supported function vanishes on the exceptional set `V`**, at
Hypothesis (4.6) generality.
`V ⊆ A₀` — the exceptional elements *are* Dade base points — so the explicit (2.5) evaluation
(`dadeValue_eq` at `a = v`, `h = 1`) reduces `α^τ(v)` to `α(v)`, which vanishes as soon as `α` is
supported on some set `A₁` that `V` avoids.
This is the (4.6)-level form of `S13.tau_apply_eq_zero_of_mem_typePV`, which fixes
`A₁ = A(M) = (M')^#` and reads the avoidance off `typePData_typePV_not_mem_derived`. Keeping `A₁`
separate from the (4.6) ambient `A` matters: the members of `𝒮(Y)` have `(M')^#`-supported
differences, and for a type-uniform `A(M)` (i.e. `typePACore`) that is *strictly larger* than `A`,
so the §13 phrasing would not transfer. -/
theorem dadeICM_apply_eq_zero_of_avoidV [Finite G] {M : Subgroup G} {A : Set G}
(h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
(tau : OddOrder.Peterfalvi.S04.FullDadeIsometryData (G := G) h46.dade0)
{α : ClassFunction ↥M ℂ}
(hαA0 : α.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
{A₁ : Set G} (hαA₁ : α.support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup A₁ M)
{v : G} (hv : v ∈ (OddOrder.Peterfalvi.S06.ticVdiff h46).V) (hvA₁ : v ∉ A₁) :
OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 tau α v = 0 := by
classical
have hvV : v ∈ h46.tic.V := by rw [h46.tic_V]; exact hv
have hvA0 : v ∈ A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V :=
Or.inr ⟨v, hvV, 1, M.one_mem, by group⟩
rw [OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_apply_of_support h46.dade0 _ hαA0,
OddOrder.Peterfalvi.S04.Hypothesis.dadeMap_apply,
h46.dade0.dadeValue_eq _ (a := ⟨v, hvA0⟩) (Subgroup.one_mem _) (by rw [mul_one])]
by_contra hne
exact hvA₁ (OddOrder.Peterfalvi.S04.mem_supportInSubgroup.mp
(hαA₁ (ClassFunction.mem_support.mpr hne)))
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **(5.2.e) cross-orthogonality of the dispatched `R`-families over `𝒮(Y)`, at §9 level** — the
`hRorth` input of the norm-general (5.7) engine.
For members `φ, ξ` with `⟨φ, ξ⟩ = ⟨φ, ξ̄⟩ = 0`, the families `R(φ) ⊥ R(ξ)`: a `2×2` case split on
the member dichotomy.
* **irr × irr** — `S08.dadeOrthonormalCharacterImageFamilyOfDiff_orthogonal`; the two extra scalars
`⟨φ̄, ξ⟩`, `⟨φ̄, ξ̄⟩` are `star`-conjugates of `⟨φ, ξ̄⟩`, `⟨φ, ξ⟩`;
* **irr × column** / **column × irr** —
`S08.certainTypeR_imageSet_orthogonal_dadeOfDiff_of_vanishOnV` (irr-on-left via an
`inner_conj_symm` swap), whose anchor is `dadeICM_apply_eq_zero_of_avoidV` at `A₁ = M'`;
* **column × column** — `S06.certainTypeR_imageSet_orthogonal_certainTypeR`, whose `χ₂ ≠ χ₂'` and
`χ₂ ≠ χ₂'⁻¹` side conditions come from `⟨φ, ξ⟩ = 0` and `⟨φ, ξ̄⟩ = 0`: equality would force
`φ = ξ` (resp. `φ = ξ̄` by `columnSum_conj_eq`) and the self-norm `w₁ ≠ 0`.
The §13 analogue is `S13.caseB_sOf_memberRFamily_orthogonal`. The only genuinely ambient input
here is `hVsub`: the exceptional set `V` avoids `M'` (at the intended instantiation, exactly
`S10.typePData_typePV_not_mem_derived`). **No type hypothesis appears.** -/
theorem sOf_memberRFamily_orthogonal [Finite G] {M : Subgroup G} {A : Set G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G) (hM : M ∈ maximalSubgroups G)
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
[Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data)
(htau : h46.tau = h46.dade0.fullDadeIsometryData)
(hKsupp : ∀ x : ↥M, x ∈ (derivedInG M).subgroupOf M → x ≠ 1 →
x ∈ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
(hVsub : ∀ v ∈ (OddOrder.Peterfalvi.S06.ticVdiff h46).V, v ∉ (derivedInG M : Set G))
{Y : Subgroup G} {φ ξ : ClassFunction ↥M ℂ}
(hφ : φ ∈ sOf data Y) (hξ : ξ ∈ sOf data Y)
(h1 : ClassFunction.inner φ ξ = 0) (h2 : ClassFunction.inner φ ξ.conj = 0) :
(sOf_memberRFamily hG hM data h46 hKeq htau hKsupp hφ).Orthogonal
(sOf_memberRFamily hG hM data h46 hKeq htau hKsupp hξ) := by
classical
have := derivedInG_subgroupOf_normal M
-- member differences are `M'`-supported (the `A₁` of the anchor), on top of being `A₀`-supported
have hMderiv : ∀ {ζ : ClassFunction ↥M ℂ}, ζ ∈ sOf data Y →
((ζ.conj - ζ).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
((derivedInG M : Set G)) M) := fun {ζ} hζ =>
OddOrder.Peterfalvi.S08.inducedKernelFamily_conjDiff_support
(fun _ hx _ => Subgroup.mem_subgroupOf.mp hx)
(sOf_subset_inducedKernelFamily_bot hG hM data Y hζ)
-- the anchor of the mixed stratum: `(ζ − ζ̄)^τ` vanishes on the exceptional `V`
have hanchor : ∀ {ζ : ClassFunction ↥M ℂ}, ζ ∈ sOf data Y →
∀ v ∈ (OddOrder.Peterfalvi.S06.ticVdiff h46).V,
OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau (ζ - ζ.conj) v = 0 := by
intro ζ hζ v hv
have hflip : (ζ - ζ.conj : ClassFunction ↥M ℂ) = -(ζ.conj - ζ) := by abel
refine dadeICM_apply_eq_zero_of_avoidV h46 _ ?_ ?_ hv (hVsub v hv)
· rw [hflip, ClassFunction.support_neg]
exact OddOrder.Peterfalvi.S08.inducedKernelFamily_conjDiff_support hKsupp
(sOf_subset_inducedKernelFamily_bot hG hM data Y hζ)
· rw [hflip, ClassFunction.support_neg]
exact hMderiv hζ
have hw1ne : (Nat.card h46.W1 : ℂ) ≠ 0 := Nat.cast_ne_zero.mpr (NeZero.ne _)
intro α hα β hβ
by_cases hφirr : IsIrreducibleCharacter φ <;> by_cases hξirr : IsIrreducibleCharacter ξ
· -- irr × irr
obtain ⟨hrφ, hsφ, hφeq⟩ :=
sOf_memberRFamily_imageSet_of_irr hG hM data h46 hKeq htau hKsupp hφ hφirr
obtain ⟨hrξ, hsξ, hξeq⟩ :=
sOf_memberRFamily_imageSet_of_irr hG hM data h46 hKeq htau hKsupp hξ hξirr
rw [hφeq] at hα
rw [hξeq] at hβ
have hbarχ : ClassFunction.inner φ.conj ξ = 0 := by
rw [← ClassFunction.conj_conj ξ, inner_conj_conj, h2, star_zero]
have hbarχbar : ClassFunction.inner φ.conj ξ.conj = 0 := by
rw [inner_conj_conj, h1, star_zero]
exact OddOrder.Peterfalvi.S08.dadeOrthonormalCharacterImageFamilyOfDiff_orthogonal
h46.dade0 (x := ⟨φ, hφirr⟩) (χ := ⟨ξ, hξirr⟩) hrφ hsφ hrξ hsξ h1 h2 hbarχ hbarχbar
α hα β hβ
· -- irr × column
obtain ⟨hrφ, hsφ, hφeq⟩ :=
sOf_memberRFamily_imageSet_of_irr hG hM data h46 hKeq htau hKsupp hφ hφirr
obtain ⟨χ₂, hχ₂, -, hξeq⟩ :=
sOf_memberRFamily_imageSet_of_col hG hM data h46 hKeq htau hKsupp hξ hξirr
rw [hφeq] at hα
rw [hξeq] at hβ
rw [inner_conj_symm β α]
rw [OddOrder.Peterfalvi.S08.certainTypeR_imageSet_orthogonal_dadeOfDiff_of_vanishOnV
h46 hχ₂ (OddOrder.Peterfalvi.S06.columnSum_inv_apply_one h46 χ₂).symm
h46.dade0 ⟨φ, hφirr⟩ hrφ hsφ (hanchor hφ) β hβ α hα, star_zero]
· -- column × irr
obtain ⟨χ₂, hχ₂, -, hφeq⟩ :=
sOf_memberRFamily_imageSet_of_col hG hM data h46 hKeq htau hKsupp hφ hφirr
obtain ⟨hrξ, hsξ, hξeq⟩ :=
sOf_memberRFamily_imageSet_of_irr hG hM data h46 hKeq htau hKsupp hξ hξirr
rw [hφeq] at hα
rw [hξeq] at hβ
exact OddOrder.Peterfalvi.S08.certainTypeR_imageSet_orthogonal_dadeOfDiff_of_vanishOnV
h46 hχ₂ (OddOrder.Peterfalvi.S06.columnSum_inv_apply_one h46 χ₂).symm
h46.dade0 ⟨ξ, hξirr⟩ hrξ hsξ (hanchor hξ) α hα β hβ
· -- column × column
obtain ⟨χ₂, hχ₂, hφcol, hφeq⟩ :=
sOf_memberRFamily_imageSet_of_col hG hM data h46 hKeq htau hKsupp hφ hφirr
obtain ⟨χ₂', hχ₂', hξcol, hξeq⟩ :=
sOf_memberRFamily_imageSet_of_col hG hM data h46 hKeq htau hKsupp hξ hξirr
rw [hφeq] at hα
rw [hξeq] at hβ
-- `χ₂ ≠ χ₂'`: else `φ = ξ` and `⟨φ, φ⟩ = w₁ ≠ 0` contradicts `h1`
have hne1 : χ₂ ≠ χ₂' := by
intro heq
rw [hφcol, hξcol, heq, OddOrder.Peterfalvi.S06.columnSum_def,
OddOrder.Peterfalvi.S06.columnFamily_mu_sum_inner, if_pos rfl] at h1
exact hw1ne h1
-- `χ₂ ≠ χ₂'⁻¹`: else `φ = ξ̄` and `⟨φ, ξ̄⟩ = w₁ ≠ 0` contradicts `h2`
have hne2 : χ₂ ≠ χ₂'⁻¹ := by
intro heq
rw [hφcol, hξcol, OddOrder.Peterfalvi.S06.columnSum_conj_eq, heq,
OddOrder.Peterfalvi.S06.columnSum_def,
OddOrder.Peterfalvi.S06.columnFamily_mu_sum_inner, if_pos rfl] at h2
exact hw1ne h2
exact OddOrder.Peterfalvi.S06.certainTypeR_imageSet_orthogonal_certainTypeR h46 hχ₂ hχ₂'
(OddOrder.Peterfalvi.S06.columnSum_inv_apply_one h46 χ₂).symm
(OddOrder.Peterfalvi.S06.columnSum_inv_apply_one h46 χ₂').symm hne1 hne2 α hα β hβ
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **Every member of `𝒮(Y)` has natural-number self-norm** — `1` for an irreducible, `w₁` for a
(9.9.b) column `μ_{χ₂}` (`columnFamily_mu_sum_inner` on the diagonal). This is the `hN` input of
the norm-general (5.7) engine, and the one place where being *norm-general* rather than
all-irreducible is visible. -/
theorem sOf_member_inner_self_natCast [Finite G] {M : Subgroup G} {A : Set G}
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data) {Y : Subgroup G} {η : ClassFunction ↥M ℂ} (hη : η ∈ sOf data Y) :
∃ n : ℕ, ClassFunction.inner η η = (n : ℂ) := by
classical
by_cases hirr : IsIrreducibleCharacter η
· exact ⟨1, by rw [hirr.inner_self_eq_one, Nat.cast_one]⟩
· obtain ⟨χ₂, -, hcol⟩ := sOf_columnSum_of_not_irreducible data h46 hKeq hη hirr
exact ⟨Nat.card h46.W1, by
rw [hcol, OddOrder.Peterfalvi.S06.columnSum_def,
OddOrder.Peterfalvi.S06.columnFamily_mu_sum_inner, if_pos rfl]⟩
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **Peterfalvi (9.11), case (9.7.b), at §9 level**: a uniform-degree `𝒮(Y)` is coherent on
`A₀(M)`.
The book's case (b) is two citations — "By (9.9.a) and (5.7), `𝒮(H₀C′)` is coherent" — and this is
that sentence: (9.9.a) supplies the uniform degree `d = qu` (the parameter `hunif`, discharged by
the type-free `S11.caseB_degree_qu`), and the norm-general (5.7) engine
`S07.uniform_degree_coherence_of_families` does the rest. The engine is the norm-general one
because (9.9.b) puts *reducible* members (the columns `μ_j`) in `𝒮(H₀C′)`, so the all-irreducible
(5.7) wrapper does not apply — this is exactly why the `R`-family dispatch above was needed.
The §13 analogue is `S13.caseB_coherent_sOf_H0Cprime`, which anchors on a §10 μ-grid column
(`hyp.base.muColumnChar`) obtained through (11.7) `H₀ = 1` — available only in types III/IV. Here
the pivot `η₁` is an explicit parameter: the book gets it from (9.9.b)'s count of reducible members,
which is genuine upstream content and is honest to expose rather than to re-derive from packaging
(the same pattern as the `2 ≤ ncard` parameter of `S15.Hypothesis.sSetIrrDeg_coherent`).
**No type hypothesis appears on this route.** -/
theorem sOf_caseB_coherent [Finite G] {M : Subgroup G} {A : Set G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G) (hM : M ∈ maximalSubgroups G)
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data)
(htau : h46.tau = h46.dade0.fullDadeIsometryData)
(hKsupp : ∀ x : ↥M, x ∈ (derivedInG M).subgroupOf M → x ≠ 1 →
x ∈ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
(hVsub : ∀ v ∈ (OddOrder.Peterfalvi.S06.ticVdiff h46).V, v ∉ (derivedInG M : Set G))
{Y : Subgroup G} (d : ℕ)
(hunif : ∀ φ ∈ sOf data Y, ((φ : ClassFunction ↥M ℂ) : ↥M → ℂ) 1 = (d : ℂ)) (hd0 : d ≠ 0)
{η₁ : ClassFunction ↥M ℂ} (hη₁ : η₁ ∈ sOf data Y) :
Nonempty (OddOrder.Peterfalvi.S07.IsCoherent
(OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau) (sOf data Y)
(OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)) := by
classical
have := derivedInG_subgroupOf_normal M
have hModd : Odd (Nat.card ↥M) := hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card M)
have hIKF : ∀ ⦃x : ClassFunction ↥M ℂ⦄, x ∈ sOf data Y →
x ∈ OddOrder.Peterfalvi.S08.inducedKernelFamily
((derivedInG M).subgroupOf M) (⊥ : Subgroup ↥M) := fun {_} hx =>
sOf_subset_inducedKernelFamily_bot hG hM data Y hx
-- every member difference is `A₀`-supported: pass through the pivot `η₁`
have hsuppdiff : ∀ a ∈ sOf data Y, ∀ b ∈ sOf data Y,
((a - b : ClassFunction ↥M ℂ)).support ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M := by
intro a ha b hb
have ha1 := sOf_anchor_diff_support hG hM data Y hKsupp d hunif hη₁ ha
have hb1 := sOf_anchor_diff_support hG hM data Y hKsupp d hunif hη₁ hb
rw [show (a - b : ClassFunction ↥M ℂ) = (a - η₁) - (b - η₁) by abel]
exact (ClassFunction.support_sub_subset _ _).trans (Set.union_subset ha1 hb1)
have hnr : ∀ a ∈ sOf data Y, a ≠ (a : ClassFunction ↥M ℂ).conj := fun a ha h =>
OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters hModd _ (hIKF ha) h.symm
refine OddOrder.Peterfalvi.S07.uniform_degree_coherence_of_families (sOf_finite data Y) hη₁
(fun η hη => sOf_memberRFamily hG hM data h46 hKeq htau hKsupp hη)
(fun a ha b hb hab =>
OddOrder.Peterfalvi.S08.inducedKernelFamily_pairwise_orthogonal (hIKF ha) (hIKF hb) hab)
(sOf_closedUnderConjugate data Y) hnr
(sOf_member_inner_self_natCast data h46 hKeq hη₁)
(fun {φ ψ} hφ hψ => ?_) (fun a ha b hb => ?_) hsuppdiff
(fun {φ ξ} hφ hξ h1 h2 =>
sOf_memberRFamily_orthogonal hG hM data h46 hKeq htau hKsupp hVsub hφ hξ h1 h2)
(fun a ha => (hunif a ha).trans (hunif _ hη₁).symm)
(by rw [hunif _ hη₁]; exact Nat.cast_ne_zero.mpr hd0)
(fun h => absurd (OddOrder.Peterfalvi.S04.mem_supportInSubgroup.mp h) (fun hmem =>
(OddOrder.Peterfalvi.S04.mem_sharp.mp (h46.dade0.subset_sharp hmem)).2 rfl))
(sOf_closedUnderConjugate data Y hη₁) (Ne.symm (hnr _ hη₁))
· -- `hiso`: the Dade lift is an isometry on `A₀`-supported functions
rw [htau]
exact OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_inner_eq_of_supported h46.dade0
(OddOrder.Peterfalvi.S07.support_subset_of_mem_zSupportedSpan hφ)
(OddOrder.Peterfalvi.S07.support_subset_of_mem_zSupportedSpan hψ)
· -- `hZdiff`: supported differences of virtual characters have virtual images
rw [htau]
exact OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_mem_ZIrr_of_supported h46.dade0
(hsuppdiff a ha b hb)
(Submodule.sub_mem _ (OddOrder.Peterfalvi.S08.inducedKernelFamily_mem_ZIrr (hIKF ha))
(OddOrder.Peterfalvi.S08.inducedKernelFamily_mem_ZIrr (hIKF hb)))
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **Peterfalvi (9.11) in case (9.7.b)**: `𝒮(H₀C′)` is coherent, with the uniform degree supplied
by **(9.9.a)** — the book's "By (9.9.a) and (5.7)" verbatim.
Specializes `sOf_caseB_coherent` to the caseB Clifford datum: (9.9.a)
(`S11.caseB_degree_qu`, itself type-free) gives every member degree `qu`, and `qu ≠ 0` since `q` is
a subgroup order and `u` is odd.
Only the pivot `η₁` remains as a parameter. The book takes it from **(9.9.b)** — `𝒮(H₀)` contains
exactly `p−1` reducible members `μ_j`, all lying in `𝒮(H₀C) ⊆ 𝒮(H₀C′)` — which is genuine upstream
content, not something to reconstruct from packaging (the §13 route reaches it only through (11.7)
`H₀ = 1`, i.e. types III/IV). -/
theorem caseB_coherent_sOf_cprime [Finite G] {M : Subgroup G} {A : Set G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G) (hM : M ∈ maximalSubgroups G)
{data : TypesIIIIIIVSetup M} {chief : ChiefFactorData data}
(chars : Section11CharacterData data chief) (caseB : CliffordCaseBData chars)
(h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data)
(htau : h46.tau = h46.dade0.fullDadeIsometryData)
(hKsupp : ∀ x : ↥M, x ∈ (derivedInG M).subgroupOf M → x ≠ 1 →
x ∈ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
(hVsub : ∀ v ∈ (OddOrder.Peterfalvi.S06.ticVdiff h46).V, v ∉ (derivedInG M : Set G))
{η₁ : ClassFunction ↥M ℂ} (hη₁ : η₁ ∈ sOf data (chief.H0 ⊔ chars.Cprime)) :
Nonempty (OddOrder.Peterfalvi.S07.IsCoherent
(OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau)
(sOf data (chief.H0 ⊔ chars.Cprime))
(OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)) :=
sOf_caseB_coherent hG hM data h46 hKeq htau hKsupp hVsub (data.q * chars.u)
(caseB_degree_qu hG chars caseB)
(mul_ne_zero Nat.card_pos.ne' (u_odd hG chars).pos.ne') hη₁
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **`𝒮(H₀C′) ≠ ∅`** — the pivot of the (9.11) case (9.7.b) engine, from **(9.9.b)**.
The book's (9.9.b) says `𝒮(H₀)` *and* `𝒮(H₀C)` each contain exactly `p − 1` reducible characters
`μ_j`; since `C′ = [C,C] ≤ C`, the family `𝒮(H₀C)` sits inside `𝒮(H₀C′)` (`sOf_antitone`), so those
`μ_j` are members here too — and `p − 1 > 0` because `p` is prime.
Note the count is taken at `H₀C`, not `H₀`: the reducibles of `𝒮(H₀)` need not lie in the *smaller*
`𝒮(H₀C′)`, which is exactly why the book states (9.9.b) for both carriers. The §13 route reaches
the same pivot through the §10 μ-grid and (11.7) `H₀ = 1`, hence only in types III/IV
(issue 1045). -/
theorem sOf_cprime_nonempty [Finite G] {M : Subgroup G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G)
{data : TypesIIIIIIVSetup M} {chief : ChiefFactorData data} :
(sOf data (chief.H0 ⊔ cprimeSub data chief)).Nonempty := by
have hne : {φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
¬ IsIrreducibleCharacter φ}.Nonempty := by
refine Set.nonempty_of_ncard_ne_zero ?_
rw [reducible_count_sOf_H0supC hG chief]
have := chief.p_prime.one_lt
omega
obtain ⟨φ, hφ, -⟩ := hne
exact ⟨φ, sOf_antitone data (sup_le_sup_left (cprimeSub_le_C data chief) chief.H0) hφ⟩
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **Peterfalvi (5.5) for a coherent subfamily of `𝒮(Y)`, at §9 level** (Coq
`mem_coherent_sum_subseq`, `PFsection5.v:957`) — the §9 form of
`S13.coherent_extension_eq_sum_memberRFamily`.
A coherent extension of `T ⊆ 𝒮(Y)` evaluates a member `ψ` (whose conjugate is also in `T`) as a
partial sum `∑_{α ∈ E} α` over the dispatched `R(ψ)`-family `sOf_memberRFamily`. This is
`CharacterPsiDecomposition.ofProjection` at `ψ = 0` — the auxiliary isometry is the coherent
extension itself (isometric on `ℤ[T]`, agreeing with `τ` on the supported `ψ − ψ̄`) — closed by
`eq_sum_of_psi_eq_zero`.
The §13 version derives `ψ ≠ ψ̄` and the supportedness of `ψ − ψ̄` from the packaging's
`hyp.base.mderivSharp_subset_A0`; here they come from `hKsupp` and odd order, so **no type
hypothesis appears**. -/
theorem sOf_coherent_extension_eq_sum_memberRFamily [Finite G] {M : Subgroup G} {A : Set G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G) (hM : M ∈ maximalSubgroups G)
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)] [Fintype ↥(h46.W1 ⊔ h46.W2)]
[Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data)
(htau : h46.tau = h46.dade0.fullDadeIsometryData)
(hKsupp : ∀ x : ↥M, x ∈ (derivedInG M).subgroupOf M → x ≠ 1 →
x ∈ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
{Y : Subgroup G} {T : Set (ClassFunction ↥M ℂ)} (hTsub : T ⊆ sOf data Y)
(c' : OddOrder.Peterfalvi.S07.IsCoherent
(OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau) T
(OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M))
{ψ : ClassFunction ↥M ℂ} (hψT : ψ ∈ T) (hψcT : ψ.conj ∈ T) :
∃ E ⊆ (sOf_memberRFamily hG hM data h46 hKeq htau hKsupp (hTsub hψT)).imageSet,
c'.extension ψ = ∑ α ∈ E, α := by
classical
have := derivedInG_subgroupOf_normal M
have hψsOf := hTsub hψT
have hψcsOf := hTsub hψcT
have hψbot := sOf_subset_inducedKernelFamily_bot hG hM data Y hψsOf
have hψcbot := sOf_subset_inducedKernelFamily_bot hG hM data Y hψcsOf
have hModd : Odd (Nat.card ↥M) := hG.odd.of_dvd_nat (Subgroup.card_subgroup_dvd_card M)
have hne : ψ ≠ ψ.conj := fun h =>
OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters hModd
(⊥ : Subgroup ↥M) hψbot h.symm
have hχχbar : ClassFunction.inner ψ ψ.conj = 0 :=
OddOrder.Peterfalvi.S08.inducedKernelFamily_pairwise_orthogonal hψbot hψcbot hne
have hdiffsupp : ((ψ - ψ.conj : ClassFunction ↥M ℂ)).support ⊆
OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M := by
rw [show (ψ - ψ.conj : ClassFunction ↥M ℂ) = -(ψ.conj - ψ) from by abel,
ClassFunction.support_neg]
exact OddOrder.Peterfalvi.S08.inducedKernelFamily_conjDiff_support hKsupp hψbot
have hle : OddOrder.Peterfalvi.S07.zSpan (L := ↥M)
({ψ - ψ.conj, ψ - 0} : Set (ClassFunction ↥M ℂ))
≤ OddOrder.Peterfalvi.S07.zSpan (L := ↥M) T :=
Submodule.span_le.mpr (by
intro x hx
simp only [Set.mem_insert_iff, Set.mem_singleton_iff] at hx
rcases hx with rfl | rfl
· exact Submodule.sub_mem _ (Submodule.subset_span hψT) (Submodule.subset_span hψcT)
· exact Submodule.sub_mem _ (Submodule.subset_span hψT) (Submodule.zero_mem _))
obtain ⟨-, hτ1ψ, E, hEsub, hXsum, -⟩ :=
OddOrder.Peterfalvi.S07.CharacterPsiDecomposition.eq_sum_of_psi_eq_zero
(OddOrder.Peterfalvi.S07.CharacterPsiDecomposition.ofProjection
(sOf_memberRFamily hG hM data h46 hKeq htau hKsupp hψsOf) c'.extension
(fun φ ζ hφ hζ => c'.extension_inner_eq φ ζ (hle hφ) (hle hζ))
(c'.extends_on_supported (ψ - ψ.conj)
⟨Submodule.sub_mem _ (Submodule.subset_span hψT) (Submodule.subset_span hψcT),
hdiffsupp⟩)
(by rw [sub_zero]; exact c'.extension_mem_ZIrr ψ (Submodule.subset_span hψT))
(by rw [ClassFunction.inner_zero_right])
(by rw [ClassFunction.inner_zero_right])
hχχbar)
exact ⟨E, hEsub, hτ1ψ.trans hXsum⟩
open scoped OddOrder.Peterfalvi.S12.FiniteInduce in
/-- **Cross-orthogonality of coherent images, at §9 level** (Coq `coherent_ortho`,
`PFsection5.v:986`) — the §9 form of `S13.coherent_extension_cross_orthogonal`.
For coherent subfamilies `T₁, T₂ ⊆ 𝒮(Y)` and members `ψ ∈ T₁`, `λ ∈ T₂` with `ψ ∉ {λ, λ̄}`
(so `⟨ψ, λ⟩ = ⟨ψ, λ̄⟩ = 0` by the family's pairwise orthogonality), the coherent images are
orthogonal: both are partial `R`-family sums by (5.5), and the `R`-families are cross-orthogonal
by (5.2.e) (`sOf_memberRFamily_orthogonal`). **No type hypothesis appears.** -/
theorem sOf_coherent_extension_cross_orthogonal [Finite G] {M : Subgroup G} {A : Set G}
(hG : OddOrder.BG.IsMinimalSimpleOdd G) (hM : M ∈ maximalSubgroups G)
(data : TypesIIIIIIVSetup M) (h46 : OddOrder.Peterfalvi.S06.Hypothesis46 A M)
[NeZero (Nat.card h46.W1)]
[Invertible (Nat.card ↥(h46.W1 ⊔ h46.W2) : ℂ)]
(hKeq : h46.K = huSub data)
(htau : h46.tau = h46.dade0.fullDadeIsometryData)
(hKsupp : ∀ x : ↥M, x ∈ (derivedInG M).subgroupOf M → x ≠ 1 →
x ∈ OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M)
(hVsub : ∀ v ∈ (OddOrder.Peterfalvi.S06.ticVdiff h46).V, v ∉ (derivedInG M : Set G))
{Y : Subgroup G} {T₁ T₂ : Set (ClassFunction ↥M ℂ)}
(hT₁sub : T₁ ⊆ sOf data Y) (hT₂sub : T₂ ⊆ sOf data Y)
(c₁ : OddOrder.Peterfalvi.S07.IsCoherent
(OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau) T₁
(OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M))
(c₂ : OddOrder.Peterfalvi.S07.IsCoherent
(OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap h46.dade0 h46.tau) T₂
(OddOrder.Peterfalvi.S04.supportInSubgroup
(A ∪ OddOrder.GroupTheory.conjClassSetIn M h46.tic.V) M))
{ψ lam : ClassFunction ↥M ℂ}
(hψT : ψ ∈ T₁) (hψcT : ψ.conj ∈ T₁) (hlamT : lam ∈ T₂) (hlamcT : lam.conj ∈ T₂)
(hne1 : ψ ≠ lam) (hne2 : ψ ≠ lam.conj) :
ClassFunction.inner (c₁.extension ψ) (c₂.extension lam) = 0 := by
classical
have : Fintype ↥(h46.W1 ⊔ h46.W2) := Fintype.ofFinite _
have := derivedInG_subgroupOf_normal M
have h1 : ClassFunction.inner ψ lam = 0 :=
OddOrder.Peterfalvi.S08.inducedKernelFamily_pairwise_orthogonal
(sOf_subset_inducedKernelFamily_bot hG hM data Y (hT₁sub hψT))
(sOf_subset_inducedKernelFamily_bot hG hM data Y (hT₂sub hlamT)) hne1
have h2 : ClassFunction.inner ψ lam.conj = 0 :=
OddOrder.Peterfalvi.S08.inducedKernelFamily_pairwise_orthogonal
(sOf_subset_inducedKernelFamily_bot hG hM data Y (hT₁sub hψT))
(sOf_subset_inducedKernelFamily_bot hG hM data Y (hT₂sub hlamcT)) hne2
obtain ⟨E₁, hE₁sub, hE₁⟩ := sOf_coherent_extension_eq_sum_memberRFamily hG hM data h46
hKeq htau hKsupp hT₁sub c₁ hψT hψcT
obtain ⟨E₂, hE₂sub, hE₂⟩ := sOf_coherent_extension_eq_sum_memberRFamily hG hM data h46
hKeq htau hKsupp hT₂sub c₂ hlamT hlamcT
have horth := sOf_memberRFamily_orthogonal hG hM data h46 hKeq htau hKsupp hVsub
(hT₁sub hψT) (hT₂sub hlamT) h1 h2
rw [hE₁, hE₂, OddOrder.RepresentationTheory.inner_sum_left]
refine Finset.sum_eq_zero fun α hα => ?_
rw [OddOrder.RepresentationTheory.inner_sum_right]
exact Finset.sum_eq_zero fun β hβ => horth α (hE₁sub hα) β (hE₂sub hβ)
end OddOrder.Peterfalvi.S11
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/S11_NineElevenRFamily.lean