Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Peterfalvi (8.15) claim 3: Hypothesis (5.2) for the type PPP induced family

Definition
OddOrder_Peterfalvi_S10_SubcoherentTypeP

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

feit-thompsonfinite-groupsgroup-theoryishida-odd-order

Formalizes claim 3 of Peterfalvi (8.15) (Peterfalvi, Character Theory for the Odd Order Theorem, §8, pp. 47-48): Hypothesis (5.2) for the induced family {IndKLθ∣θ∈Irr K,H⊈ker⁡θ}\{\mathrm{Ind}_K^L\theta\mid\theta\in\mathrm{Irr}\,K, H\not\subseteq\ker\theta\}{IndKL​θ∣θ∈IrrK,H⊆kerθ} when MMM is of type PPP.

Main declarations: inducedNonKernelFamily (Peterfalvi (5.3.b)/(8.15.3)'s family); inducedNonKernelFamily_apply_eq_zero (Peterfalvi (4.7) on the (8.15.3) family); inducedNonKernelFamily_conjDiff_support (The hconjsupp input of (5.2) for the (8.15.3) family); inducedNonKernelFamily_diff_support (Member differences of the (8.15.3) family are A-supported — the pair version).

The bundle also proves 4 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/S10_SubcoherentTypeP.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_S10_MinimalSimpleBasic
import Definitions.Def_OddOrder_Peterfalvi_S07_Subcoherent
import Definitions.Def_OddOrder_Peterfalvi_S08_SixTwoGeneral
import Definitions.Def_OddOrder_Peterfalvi_S10_TypePSupportA0
import Definitions.Def_OddOrder_Peterfalvi_S10_StructureSetup
import Definitions.Def_OddOrder_Peterfalvi_S12_MaximalIII_IV_V_Core_Hypothesis
import Definitions.Def_OddOrder_BG_Ch4_FamilyOfMaximal_S15_MF_SetupLemma151

namespace OddOrder.Peterfalvi.S10

open OddOrder.GroupTheory

open OddOrder.RepresentationTheory

variable {G : Type*} [Group G]

section CoreNontrivialFamily

variable {M : Subgroup G} [Fintype ↥M]

variable (K : Subgroup ↥M) [Invertible (Nat.card ↥K : ℂ)]

/-- **Peterfalvi (5.3.b)/(8.15.3)'s family** `{Ind_K^L θ | θ ∈ Irr K, H ⊄ Ker θ}`.

The `H`-nontrivial analogue of `S08.inducedKernelFamily` (which filters by `θ ≠ 1` together with a
*positive* kernel condition `X ⊆ Ker θ`).

At `H = K` the two filters coincide — for irreducible `θ ∈ Irr K`, `K ⊆ Ker θ` iff `θ = 1` — which
is exactly the type-`P₁` coincidence `M_s = M'`.  ⚠ Only the direction used downstream is
formalized: `inducedNonKernelFamily_subset_inducedKernelFamily_bot` proves `H ⊄ Ker θ ⟹ θ ≠ 1`
(hence `⊆`).  The converse `θ ≠ 1 ⟹ K ⊄ Ker θ` — an irreducible character with full kernel is
trivial, via `IsIrreducibleCharacter.inner_self_eq_one` forcing degree 1 — is standard, and is
stated here as motivation, **not** as a formalized lemma. -/
def inducedNonKernelFamily (H : Subgroup ↥M) : Set (ClassFunction ↥M ℂ) :=
  {φ | ∃ θ : IrreducibleCharacter ↥K,
    ¬ ((H.subgroupOf K : Set ↥K) ⊆
        OddOrder.Peterfalvi.S03.characterKernel (θ : ClassFunction ↥K ℂ)) ∧
    φ = ClassFunction.induce K (θ : ClassFunction ↥K ℂ)}

variable {K}

/-- **The (8.15.3) family is monotone in the kernel-test subgroup**: `H₁ ≤ H₂` makes the filter
`H ⊄ Ker θ` easier to pass, so `inducedNonKernelFamily K H₁ ⊆ inducedNonKernelFamily K H₂`.

This is what lets a Hypothesis (4.6) whose `H` is *larger* than `M_σ` — e.g. the §10/§13 packaging,
where (4.6.c) is instantiated at `H = K = M'` — consume the `M_σ`-indexed containment
`S11.sOf_subset_inducedNonKernelFamily`.  Note the covering (4.6.d) is a hypothesis *of* the
ambient `h46`, so the support estimate `inducedNonKernelFamily_conjDiff_support` is unaffected by
which `H` the family was named with (issue 1045). -/
theorem inducedNonKernelFamily_mono {H₁ H₂ : Subgroup ↥M} (h : H₁ ≤ H₂) :
    inducedNonKernelFamily K H₁ ⊆ inducedNonKernelFamily K H₂ := by
  rintro φ ⟨θ, hθ, rfl⟩
  exact ⟨θ, fun hsub => hθ (fun x hx => hsub (Subgroup.subgroupOf_mono K h hx)), rfl⟩

/-- **The book's filter is finer than the repo's**: `{θ | H ⊄ Ker θ} ⊆ {θ | θ ≠ 1}`, so the
(8.15.3) family sits inside `S08.inducedKernelFamily K ⊥`.

The trivial character has every element in its kernel, so `H ⊄ Ker θ` forces `θ ≠ 1`; and the
`⊥`-kernel condition of `inducedKernelFamily` is vacuous.  This inclusion is what lets the whole
`S08.inducedKernelFamily_*` suite (non-reality, pairwise orthogonality, finiteness, conjugation
closure) apply verbatim to the finer family — only the *support* estimate has to be redone, and
that is `inducedNonKernelFamily_conjDiff_support`. -/
theorem inducedNonKernelFamily_subset_inducedKernelFamily_bot {H : Subgroup ↥M} :
    inducedNonKernelFamily K H ⊆ OddOrder.Peterfalvi.S08.inducedKernelFamily K ⊥ := by
  rintro φ ⟨θ, hker, rfl⟩
  refine ⟨θ, ?_, ?_, rfl⟩
  · rintro rfl
    exact hker fun x _ => by
      simp [IrreducibleCharacter.coe_trivialIrreducibleCharacter]
  · intro x hx
    rw [SetLike.mem_coe, Subgroup.mem_subgroupOf, Subgroup.mem_bot] at hx
    rw [show x = 1 from Subtype.ext hx]
    rfl

/-- **Member degrees of the (8.15.3) family are natural numbers** (hence real): `φ(1) = |L:K|·θ(1)`
with `θ(1)` a positive natural.  This is what kills the `{1}` left over by (4.7) when passing from
`Supp φ ⊆ A ∪ {1}` to `Supp (φ − φ̄) ⊆ A`. -/
theorem inducedNonKernelFamily_apply_one_eq_natCast {H : Subgroup ↥M}
    {φ : ClassFunction ↥M ℂ} (hφ : φ ∈ inducedNonKernelFamily K H) :
    ∃ n : ℕ, φ 1 = (n : ℂ) := by
  obtain ⟨θ, -, rfl⟩ := hφ
  obtain ⟨d, -, hd⟩ := irreducibleCharacter_apply_one_eq_pos_natCast θ
  exact ⟨K.index * d, by rw [ClassFunction.induce_apply_one, hd]; push_cast; ring⟩

end CoreNontrivialFamily

section CoreNontrivialSupport

variable [Fintype G] [Invertible (Nat.card G : ℂ)]

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

omit [Invertible (Nat.card G : ℂ)] [Invertible (Nat.card ↥M : ℂ)] in
/-- **Peterfalvi (4.7) on the (8.15.3) family**: every member vanishes outside `A ∪ {1}`.

Direct application of `S06.induce_apply_eq_zero_of_not_mem_union_of_not_subset_characterKernel`,
whose ambient set `A` is already a parameter — so it applies verbatim at `A = typePACore M`
(the book-literal `A(M)`), not only at the `P₁`-regime `typePA M`. -/
theorem inducedNonKernelFamily_apply_eq_zero {A : Set G}
    (h : OddOrder.Peterfalvi.S06.Hypothesis46Core A M) [Invertible (Nat.card ↥h.K : ℂ)]
    {φ : ClassFunction ↥M ℂ} (hφ : φ ∈ inducedNonKernelFamily h.K h.subH)
    {z : ↥M} (hz : (z : G) ∉ A ∪ ({1} : Set G)) :
    φ z = 0 := by
  obtain ⟨θ, hker, rfl⟩ := hφ
  exact OddOrder.Peterfalvi.S06.induce_apply_eq_zero_of_not_mem_union_of_not_subset_characterKernel
    h θ hker hz

omit [Invertible (Nat.card G : ℂ)] [Invertible (Nat.card ↥M : ℂ)] in
/-- **The `hconjsupp` input of (5.2) for the (8.15.3) family**: `Supp (φ − φ̄) ⊆ A`.

This is the type-uniform replacement for `S08.inducedKernelFamily_conjDiff_support`, which needs
`(M')^# ⊆ A` and therefore only reads correctly in the `P₁` regime.  Here (4.7) gives
`Supp φ ⊆ A ∪ {1}` outright, and the identity is removed by the degree being a natural number
(`φ(1) = φ̄(1)`), so no containment `(M')^# ⊆ A` is required. -/
theorem inducedNonKernelFamily_conjDiff_support {A : Set G}
    (h : OddOrder.Peterfalvi.S06.Hypothesis46Core A M) [Invertible (Nat.card ↥h.K : ℂ)]
    {φ : ClassFunction ↥M ℂ} (hφ : φ ∈ inducedNonKernelFamily h.K h.subH) :
    ((φ - φ.conj : ClassFunction ↥M ℂ)).support
      ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup A M := by
  intro z hz
  rw [OddOrder.Peterfalvi.S04.mem_supportInSubgroup]
  by_contra hzA
  refine hz ?_
  rcases eq_or_ne ((z : G)) 1 with hz1 | hz1
  · -- at the identity the two degrees agree (both the same natural number)
    have hzone : z = 1 := Subtype.ext hz1
    obtain ⟨n, hn⟩ := inducedNonKernelFamily_apply_one_eq_natCast (K := h.K) (H := h.subH) hφ
    rw [ClassFunction.sub_apply, hzone, ClassFunction.conj_apply, hn, star_natCast, sub_self]
  · -- off `A ∪ {1}` the member itself vanishes, hence so does its conjugate
    have hmem : (z : G) ∉ A ∪ ({1} : Set G) := by
      rw [Set.mem_union, Set.mem_singleton_iff, not_or]
      exact ⟨hzA, hz1⟩
    have h0 : φ z = 0 := inducedNonKernelFamily_apply_eq_zero h hφ hmem
    rw [ClassFunction.sub_apply, ClassFunction.conj_apply, h0, star_zero, sub_self]

omit [Invertible (Nat.card G : ℂ)] [Invertible (Nat.card ↥M : ℂ)] in
/-- **Member differences of the (8.15.3) family are `A`-supported — the pair version**:
for `φ, ψ` in the family of the *same degree*, `Supp (φ − ψ) ⊆ A`.

Same (4.7) argument as `inducedNonKernelFamily_conjDiff_support`: off `A ∪ {1}` both members
vanish outright, and at the identity the equal degrees cancel.  The conjugate version is the
special case `ψ = φ̄` (where the degrees agree automatically, being the same natural number).

This is the `hsuppdiff` input of the (5.7) uniform-degree coherence producer
`S07.coherent_subset_of_constant_degree`, and hence of the §9-level route to the (9.11) base
coherence (issue 1045).  Compare the `S`-side analogue
`S15.Hypothesis.sSetIrrDeg_member_diff_supported`. -/
theorem inducedNonKernelFamily_diff_support {A : Set G}
    (h : OddOrder.Peterfalvi.S06.Hypothesis46Core A M) [Invertible (Nat.card ↥h.K : ℂ)]
    {φ ψ : ClassFunction ↥M ℂ}
    (hφ : φ ∈ inducedNonKernelFamily h.K h.subH)
    (hψ : ψ ∈ inducedNonKernelFamily h.K h.subH)
    (hdeg : φ 1 = ψ 1) :
    ((φ - ψ : ClassFunction ↥M ℂ)).support
      ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup A M := by
  intro z hz
  rw [OddOrder.Peterfalvi.S04.mem_supportInSubgroup]
  by_contra hzA
  refine hz ?_
  rcases eq_or_ne ((z : G)) 1 with hz1 | hz1
  · -- at the identity the equal degrees cancel
    have hzone : z = 1 := Subtype.ext hz1
    rw [ClassFunction.sub_apply, hzone, hdeg, sub_self]
  · -- off `A ∪ {1}` both members vanish
    have hmem : (z : G) ∉ A ∪ ({1} : Set G) := by
      rw [Set.mem_union, Set.mem_singleton_iff, not_or]
      exact ⟨hzA, hz1⟩
    rw [ClassFunction.sub_apply, inducedNonKernelFamily_apply_eq_zero h hφ hmem,
      inducedNonKernelFamily_apply_eq_zero h hψ hmem, sub_self]

/-- **Peterfalvi (8.15), claim 3 — type-uniform**, on the book's own family
`𝒮 ⊆ {Ind_K^L θ | θ ∈ Irr K, H ⊄ Ker θ}` and an arbitrary ambient Dade support `A`.

This is (5.3.b) verbatim: assume Hypothesis (4.6) (here its Dade-free core `h46` together with the
`A`-Dade datum `d`), (5.2.a) (`hconjS`), and that `𝒮` sits inside the `H`-nontrivial induced
family; then Hypothesis (5.2) holds for `L = M`.

Unlike `inducedKernelFamily_subcoherent` (§8D) this needs **no containment `(M')^# ⊆ A`**, so it is
not confined to the type-`P₁` regime: the book's `H ⊄ Ker θ` filter buys the support estimate
through (4.7) instead (`inducedNonKernelFamily_conjDiff_support`).  Instantiating
`A = typePACore M` with `H = M_s = M_σ` (`typePACore_toHypothesis46_core`) gives the book-literal
(8.15.3) for **every** type `𝒫`, types II and V included.

Everything except the support estimate is inherited from the coarser `S08` family along
`inducedNonKernelFamily_subset_inducedKernelFamily_bot`. -/
noncomputable def inducedNonKernelFamily_subcoherent {A : Set G}
    (hodd : Odd (Nat.card ↥M))
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46Core A M) [Invertible (Nat.card ↥h46.K : ℂ)]
    (d : DadeSupportHypothesisData M A)
    {S : Set (ClassFunction ↥M ℂ)}
    (hsub : S ⊆ inducedNonKernelFamily h46.K h46.subH)
    (hirr : ∀ χ ∈ S, IsIrreducibleCharacter χ)
    (hconjS : OddOrder.Peterfalvi.S03.ClosedUnderConjugate S) :
    OddOrder.Peterfalvi.S07.Hypothesis (L := ↥M) (G := G) S
      (OddOrder.Peterfalvi.S04.supportInSubgroup A M) := by
  classical
  -- the coarser `S08` family, for the non-support inputs
  have hbot : ∀ ⦃χ : ClassFunction ↥M ℂ⦄, χ ∈ S →
      χ ∈ OddOrder.Peterfalvi.S08.inducedKernelFamily h46.K ⊥ := fun _ hχ =>
    inducedNonKernelFamily_subset_inducedKernelFamily_bot (hsub hχ)
  refine OddOrder.Peterfalvi.S07.irrSubcoherent
    (OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap d.dade
      (d.dade.fullDadeIsometryData))
    (OddOrder.Peterfalvi.S04.supportInSubgroup A M)
    (fun χ hχ => ?_) hconjS ?_ ?_ ?_ ?_
  · -- `Rdatum`: the (5.2.d) datum, with the **(4.7)** support estimate in place of `(M')^# ⊆ A`
    exact OddOrder.Peterfalvi.S07.dadeCharacterDifferenceImageOfDiff d.dade
      ⟨χ, hirr χ hχ⟩
      (OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters hodd ⊥ (hbot hχ))
      (by
        have h := inducedNonKernelFamily_conjDiff_support h46 (hsub hχ)
        intro x hx
        refine h ?_
        rw [ClassFunction.mem_support] at hx ⊢
        intro h0
        refine hx ?_
        rw [ClassFunction.sub_apply] at h0 ⊢
        rw [sub_eq_zero] at h0 ⊢
        exact h0.symm)
  · exact fun χ hχ =>
      OddOrder.Peterfalvi.S08.inducedKernelFamily_hasNoRealCharacters hodd ⊥ (hbot hχ)
  · exact fun χ ψ hχ hψ hne =>
      OddOrder.Peterfalvi.S08.inducedKernelFamily_pairwise_orthogonal (hbot hχ) (hbot hψ) hne
  · exact fun χ hχ => inducedNonKernelFamily_conjDiff_support h46 (hsub hχ)
  · exact fun φ ψ hφ hψ =>
      OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_inner_eq_of_supported
        d.dade hφ.2 hψ.2

/-- **Peterfalvi (5.7) on the (8.15.3) family**: a uniform-degree subfamily is coherent.

The (5.7)∘(5.3.b) companion of `inducedNonKernelFamily_subcoherent`, stated at the same
generality — arbitrary ambient support `A`, with Hypothesis (4.6)'s core `h46` and the `A`-Dade
datum `d` as parameters.  Given a conjugation-closed set `S` of *irreducible* members of the
family, all of the same nonzero degree, `S` is coherent for the Dade map.

This is the book's route to the (9.11) base coherence: (8.15.3) supplies Hypothesis (5.2) and
(5.7) turns a constant-degree subfamily into a coherent one.  The repo previously obtained the
same conclusion from the §10 μ-grid engine
(`S12.Hypothesis.inducedFamily_degreeSubfamily_isCoherent`), which is what tied it to types
III/IV; nothing here carries a type hypothesis.

⚠ `h2 : 2 ≤ S.ncard` is **exposed as a parameter** rather than derived.  This follows the
`S`-side precedent `S15.Hypothesis.sSetIrrDeg_coherent`: the §9 degree count
((9.8.d), `caseA_character_count_exact`) yields an *existence* statement `∃ ζ`, not two distinct
members, so deriving `2 ≤ ncard` here would need genuine upstream counting content.  Exposing it
keeps this producer sorry-free and leaves the count to the caller.

Obligation sources: `hirr` = `IsIrreducibleCharacter.inner_self_eq_one`;
`hZIrr` = `S07.dadeIntegralCharacterMap_mem_ZIrr_of_supported` fed by `hsuppdiff`;
`hsuppdiff` = `inducedNonKernelFamily_diff_support`. -/
theorem inducedNonKernelFamily_degreeSubfamily_coherent {A : Set G}
    (hodd : Odd (Nat.card ↥M))
    (h46 : OddOrder.Peterfalvi.S06.Hypothesis46Core A M) [Invertible (Nat.card ↥h46.K : ℂ)]
    (d : DadeSupportHypothesisData M A)
    {S : Set (ClassFunction ↥M ℂ)}
    (hsub : S ⊆ inducedNonKernelFamily h46.K h46.subH)
    (hirrS : ∀ χ ∈ S, IsIrreducibleCharacter χ)
    (hconjS : OddOrder.Peterfalvi.S03.ClosedUnderConjugate S)
    (hSfin : S.Finite) (h2 : 2 ≤ S.ncard)
    (deg : ℂ) (hconst : ∀ χ ∈ S, ((χ : ClassFunction ↥M ℂ) : ↥M → ℂ) 1 = deg) (hdeg0 : deg ≠ 0)
    (h1A : (1 : ↥M) ∉ OddOrder.Peterfalvi.S04.supportInSubgroup A M) :
    Nonempty (OddOrder.Peterfalvi.S07.IsCoherent
      (inducedNonKernelFamily_subcoherent hodd h46 d hsub hirrS hconjS).tau S
      (OddOrder.Peterfalvi.S04.supportInSubgroup A M)) := by
  classical
  set hyp := inducedNonKernelFamily_subcoherent hodd h46 d hsub hirrS hconjS with hhyp
  -- `hsuppdiff`: equal degrees, so the (4.7) estimate applies to every member difference.
  have hsuppdiff : ∀ a ∈ S, ∀ b ∈ S,
      ((a - b : ClassFunction ↥M ℂ)).support
        ⊆ OddOrder.Peterfalvi.S04.supportInSubgroup A M := fun a ha b hb =>
    inducedNonKernelFamily_diff_support h46 (hsub ha) (hsub hb)
      (by rw [hconst a ha, hconst b hb])
  refine OddOrder.Peterfalvi.S07.coherent_subset_of_constant_degree hyp (subset_refl _)
    hconjS hSfin h2 (fun ζ hζ => (hirrS ζ hζ).inner_self_eq_one) ?_
    (fun a ha b hb => by rw [hconst a ha, hconst b hb]) (fun a ha => by rw [hconst a ha]; exact
      hdeg0) h1A hsuppdiff
  -- `hZIrr`: the Dade map carries `A`-supported virtual characters into `ℤ[Irr G]`.
  intro a ha b hb
  exact OddOrder.Peterfalvi.S07.dadeIntegralCharacterMap_mem_ZIrr_of_supported d.dade
    (hsuppdiff a ha b hb)
    (Submodule.sub_mem _ (hirrS a ha).mem_ZIrr (hirrS b hb).mem_ZIrr)

end CoreNontrivialSupport

end OddOrder.Peterfalvi.S10
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/S10_SubcoherentTypeP.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