The swap of the Peterfalvi (13.1) hypothesis
DefinitionOddOrder_Peterfalvi_S15_SAndT_Setup_HypothesisSwapfeit-thompsonfinite-groupsgroup-theoryishida-odd-order
Formalizes the swap of the roles of and in Hypothesis (13.1). Peterfalvi proves the asymmetric facts (13.18), (13.19.c) once and deduces the dual statements by re-instantiating the section with the roles swapped.
Main declarations: NuGridSupplyData (The ν-side §4/§6 grid supply); Hypothesis.swap (The S↔T swap of the (13.1) hypothesis).
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/S15_SAndT_Setup/HypothesisSwap.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_S15_SAndT_Setup_SubcoherenceInputs
namespace OddOrder.Peterfalvi.S15
open OddOrder.GroupTheory
open OddOrder.RepresentationTheory
variable {G : Type*} [Group G]
open scoped FiniteInduce in
/-- **The ν-side §4/§6 grid supply** (the (13.19)-dual pins): the `T`-side certain-type facts
for the `ν`-grid, mirroring the `μ`-side grounding fields of `Hypothesis` under the role swap
`S ↔ T`, `W₁ ↔ W₂`, `q ↔ p`, `δ ↔ δ'`. Each field is the exact swap-image of the
corresponding `μ`-field, so `Hypothesis.swap` can consume them verbatim:
* `nu_irreducible`/`nu_row_injective`/`nu_orthonormal` ↔ `mu_irreducible`/`mu_col_injective`/
`mu_orthonormal` — Peterfalvi (13.1.e)/(4.3.b) at `T`;
* `nu_degree_modEq_deltaPrime` ↔ `mu_degree_modEq_delta` — (4.3.d) at `T` (`Res_{W₂}` value
identity, modulus `p = |W₂|`);
* `deltaPrime_zero_eq_one` ↔ `delta_zero_eq_one` — the (4.4) trivial-row anchor `ν_{00} = 1_T`;
* `nu_rowSum_eq_induce` ↔ `mu_colSum_eq_induce` — (4.5.a) at `T`;
* `nu_diff_support` ↔ `mu_diff_support` — (4.8) at `T` (Coq `prDade_sub_TIirr_on`), stated for
any `TypePData T` reconciled to the abstract `V`/`W₂`/`W₁` (the swap's `Sdata` is the
`reconciled_typePData_T` witness, which is exactly such a datum);
* `nu_apply_of_not_mem_W1` ↔ `mu_apply_of_not_mem_W2` — (4.3.c) at `T` (Coq `prTIirr_id`);
* `nu_conj` ↔ `mu_conj` — (4.9.a) at `T`.
The bundle is **purely grid-theoretic** (issue 9096 API audit): every field is a property of
the canonical certain-type character grid, dischargeable at the FT-layer construction site
with no (14.9) input. The structural fact `IsMulCommutative ↥V` (Peterfalvi (13.2.a) at `T`)
is deliberately *not* a field and is supplied separately to `Hypothesis.swap`; its canonical
producer is unconditional from the symmetric non-type-I setup. -/
structure NuGridSupplyData [Finite G] (hyp : Hypothesis (G := G)) : Prop where
/-- **Peterfalvi (13.1.e) at `T`**: each `ν_{ij}` is an irreducible character of `T`. -/
nu_irreducible : ∀ (i : Fin hyp.q) (j : Fin hyp.p),
OddOrder.RepresentationTheory.IsIrreducibleCharacter (hyp.nu i j)
/-- **Peterfalvi (13.1.e) at `T`, row distinctness**: within a row `i` the `ν_{ij}` are
pairwise distinct. -/
nu_row_injective : ∀ i : Fin hyp.q, Function.Injective (fun j : Fin hyp.p => hyp.nu i j)
/-- **Peterfalvi (4.3.b) at `T`, full-grid orthonormality**:
`⟨ν_{ij}, ν_{kl}⟩ = [(i,j) = (k,l)]`. -/
nu_orthonormal : ∀ (i k : Fin hyp.q) (j l : Fin hyp.p),
OddOrder.RepresentationTheory.ClassFunction.inner (hyp.nu i j) (hyp.nu k l)
= if i = k ∧ j = l then 1 else 0
/-- **Peterfalvi (4.3.d) at `T`**: the degree congruence `ν_{ij}(1) ≡ δ'_i (mod p)`. -/
nu_degree_modEq_deltaPrime : ∀ (i : Fin hyp.q) (j : Fin hyp.p), ∃ a : ℤ,
hyp.nu i j 1 = (hyp.deltaPrime i : ℂ) + (hyp.p : ℂ) * (a : ℂ)
/-- **Peterfalvi (4.4), the `T`-side base sign**: `δ'_0 = 1` (`ν_{00} = 1_T`, the trivial
row's anchor). -/
deltaPrime_zero_eq_one : hyp.deltaPrime ⟨0, hyp.q_prime.pos⟩ = 1
/-- **Peterfalvi (4.5.a) for the `T`-grid**: each `ν`-row sum is induced from an irreducible
character of `T' = [T,T]`, nontrivial on `W₁` off the anchor row. -/
nu_rowSum_eq_induce : ∀ i : Fin hyp.q,
∃ ψ : ClassFunction ↥((derivedInG hyp.T).subgroupOf hyp.T) ℂ,
OddOrder.RepresentationTheory.IsIrreducibleCharacter ψ ∧
(∑ j : Fin hyp.p, hyp.nu i j)
= ClassFunction.induce ((derivedInG hyp.T).subgroupOf hyp.T) ψ ∧
(i ≠ ⟨0, hyp.q_prime.pos⟩ →
¬ (((hyp.W1.subgroupOf hyp.T).subgroupOf ((derivedInG hyp.T).subgroupOf hyp.T) :
Set ↥((derivedInG hyp.T).subgroupOf hyp.T)) ⊆
OddOrder.Peterfalvi.S03.characterKernel ψ))
/-- **Peterfalvi (9.8)/(9.11) reverse dichotomy at `T`** (issue 9092, the ν-side swap-image of
`mu_reducible_dichotomy`): a reducible member of the `T`-side kernel-filter family `S(X)` over
`T' = derivedInG T` is a nonzero `ν`-row sum `∑_j ν_{ij}`, `i ≠ 0`. Under the swap `S ↔ T`,
`μ ↔ νᵀ`, this is exactly the swapped instance's `mu_reducible_dichotomy` (the roles of row/column
transpose), so `Hypothesis.swap` consumes it verbatim. -/
nu_reducible_dichotomy : ∀ {X : Subgroup ↥hyp.T} {ψ : ClassFunction ↥hyp.T ℂ},
ψ ∈ OddOrder.Peterfalvi.S08.inducedKernelFamily ((derivedInG hyp.T).subgroupOf hyp.T) X →
¬ OddOrder.RepresentationTheory.IsIrreducibleCharacter ψ →
∃ i : Fin hyp.q, i ≠ ⟨0, hyp.q_prime.pos⟩ ∧ ψ = ∑ j : Fin hyp.p, hyp.nu i j
/-- **Peterfalvi (4.8) at `T`, `ν`-row-difference support** (Coq `prDade_sub_TIirr_on`): for
nontrivial equal-degree rows `i, k ≠ 0`, the difference `ν_{ij} − ν_{kj}` is supported in
`A₀(T) = A(T) ∪ (V_T)^T`, for any `TypePData T` reconciled to the abstract `V`/`W₂`/`W₁`. -/
nu_diff_support : ∀ (Tdata : TypePData hyp.T), Tdata.U = hyp.V → Tdata.W1 = hyp.W2 →
Tdata.W2 = hyp.W1 → ∀ (j : Fin hyp.p) {i k : Fin hyp.q},
i ≠ ⟨0, hyp.q_prime.pos⟩ → k ≠ ⟨0, hyp.q_prime.pos⟩ →
hyp.nu i j 1 = hyp.nu k j 1 →
(hyp.nu i j - hyp.nu k j).support ⊆
OddOrder.Peterfalvi.S04.supportInSubgroup (S10.typePACore0 hyp.T Tdata) hyp.T
/-- **Peterfalvi (4.3.c) at `T`, value identity** (Coq `prTIirr_id`): on `W ∖ W₁` the
`ν`-grid is the signed `ω`-grid, `ν_{ij}(w) = δ'_i·ω_{ij}(w)`. -/
nu_apply_of_not_mem_W1 : ∀ (i : Fin hyp.q) (j : Fin hyp.p) (w : G) (hwW : w ∈ hyp.W)
(hwT : w ∈ hyp.T), w ∉ (hyp.W1 : Set G) →
hyp.nu i j ⟨w, hwT⟩ = (hyp.deltaPrime i : ℂ) * hyp.omega i j ⟨w, hwW⟩
/-- **Peterfalvi (4.9.a) at `T`**: CF-level conjugation symmetry `ν̄_{ij} = ν_{−i,−j}`. -/
nu_conj : ∀ (i : Fin hyp.q) (j : Fin hyp.p),
(hyp.nu i j).conj = hyp.nu (finNeg hyp.q_prime.pos i) (finNeg hyp.p_prime.pos j)
open scoped FiniteInduce in
/-- **The S↔T swap of the (13.1) hypothesis** (the Coq re-instantiation of
`Thirteen_17_to_19` with the pair roles interchanged). Exchanges `S ↔ T`, `P ↔ Q`, `U ↔ V`,
`C ↔ D`, `W₁ ↔ W₂`, `q ↔ p`, `u ↔ v`, `c ↔ d`, `Sset ↔ Tset`, `τ_S ↔ τ_T`, `δ ↔ δ'`,
`μ ↔ ν` and transposes the `ω`/`η`-grids. Inputs:
* `hV` — `IsMulCommutative ↥V`, Peterfalvi (13.2.a) at `T` (BG Lemma 15.1(b) for the
`(κ∪σ)'`-Hall complement of `T`), the swap's `S_U_commutative`. It is available
unconditionally from `T_nonI` and is not part of the grid bundle `pins` (issue 9096);
* `Tdata` with the `reconciled_typePData_T` reconciliations — the swap's `Sdata`;
* `pins : NuGridSupplyData hyp` — the ν-side §4/§6 grid facts (the swap's `μ`-side fields).
Everything else is filled from `hyp`'s own fields: symmetric pairs exchange, the shared-`W`
facts commute (`inf_comm`/`sup_comm`/`Set.union_comm`), the S-side duals of the T-side
structural fields are read off `hyp.Sdata` (`fitting_inf_U_eq_bot` for `P ⊓ U = ⊥`,
`M_complement` for `S = S' ⋊ W₁`), and the transposed grid facts re-index (with the
column-axis Galois orbit supplying the transposed row-vanishing transport).
Every `∀ hyp`-theorem of the S-side (13.18)/(13.19) layer then holds at `hyp.swap …`,
giving the `T`-side duals (`typeIBetaL_eta_col_constant`, `typeI_caseC_dual_dichotomy`)
with no further proof beyond index transport. -/
noncomputable def Hypothesis.swap [Finite G] (hyp : Hypothesis (G := G))
(hV : IsMulCommutative ↥hyp.V)
(Tdata : TypePData hyp.T) (hU : Tdata.U = hyp.V)
(hW1 : Tdata.W1 = hyp.W2) (hW2 : Tdata.W2 = hyp.W1)
(pins : NuGridSupplyData hyp) : Hypothesis (G := G) where
finiteG := hyp.finiteG
S := hyp.T
T := hyp.S
W1 := hyp.W2
W2 := hyp.W1
W := hyp.W
P := hyp.Q
Q := hyp.P
U := hyp.V
V := hyp.U
C := hyp.D
D := hyp.C
S_maximal := hyp.T_maximal
T_maximal := hyp.S_maximal
S_ne_T := hyp.S_ne_T.symm
S_nonI := hyp.T_nonI
T_nonI := hyp.S_nonI
one_typeII := hyp.one_typeII.symm
theorem88_caseB := fun M hM => (hyp.theorem88_caseB M hM).imp_right Or.symm
W_eq_inter := by rw [hyp.W_eq_inter, inf_comm]
W_eq_join := by rw [hyp.W_eq_join, sup_comm]
W1_inf_W2_eq_bot := by rw [inf_comm]; exact hyp.W1_inf_W2_eq_bot
W1_commutes_W2 := fun x hx y hy => (hyp.W1_commutes_W2 y hy x hx).symm
W_cyclic := hyp.W_cyclic
P_eq_SF := hyp.Q_eq_TF
Q_eq_TF := hyp.P_eq_SF
S_deriv_eq_PU := hyp.T_deriv_eq_QV
T_deriv_eq_QV := hyp.S_deriv_eq_PU
Q_inf_V_eq_bot := by
have h := hyp.Sdata.fitting_inf_U_eq_bot
rwa [hyp.Sdata_U_eq, ← hyp.P_eq_SF] at h
W2_isComplement_T_deriv := by
have h := hyp.Sdata.M_complement
rwa [hyp.Sdata_W1_eq] at h
C_eq := hyp.D_eq
D_eq := hyp.C_eq
W1_normalizes_U := hyp.W2_normalizes_V
W2_normalizes_V := hyp.W1_normalizes_U
q := hyp.p
p := hyp.q
q_prime := hyp.p_prime
p_prime := hyp.q_prime
q_odd := hyp.p_odd
p_odd := hyp.q_odd
q_eq_card_W1 := hyp.p_eq_card_W2
p_eq_card_W2 := hyp.q_eq_card_W1
u := hyp.v
v := hyp.u
c := hyp.d
d := hyp.c
c_eq_card_C := hyp.d_eq_card_D
d_eq_card_D := hyp.c_eq_card_C
card_U_eq_uc := hyp.card_V_eq_vd
card_V_eq_vd := hyp.card_U_eq_uc
Sset := hyp.Tset
Tset := hyp.Sset
A0S := hyp.A0T
A0T := hyp.A0S
tauS := hyp.tauT
tauT := hyp.tauS
omega := fun i j => hyp.omega j i
eta := fun i j => hyp.eta j i
mu := fun i j => hyp.nu j i
nu := fun i j => hyp.mu j i
delta := hyp.deltaPrime
deltaPrime := hyp.delta
delta_pm_one := ⟨hyp.delta_pm_one.2, hyp.delta_pm_one.1⟩
mu_degree_modEq_delta := fun i j => pins.nu_degree_modEq_deltaPrime j i
delta_zero_eq_one := pins.deltaPrime_zero_eq_one
tau3 := hyp.tau3
eta_eq_tau_omega := fun i j => hyp.eta_eq_tau_omega j i
mu_definition := fun i j => hyp.nu_definition j i
mu_irreducible := fun i j => pins.nu_irreducible j i
mu_col_injective := pins.nu_row_injective
mu_orthonormal := fun i k j l => by
rw [pins.nu_orthonormal j l i k]
by_cases h : j = l ∧ i = k
· rw [if_pos h, if_pos ⟨h.2, h.1⟩]
· rw [if_neg h, if_neg fun h' => h ⟨h'.2, h'.1⟩]
mu_colSum_eq_induce := pins.nu_rowSum_eq_induce
mu_reducible_dichotomy := pins.nu_reducible_dichotomy
nu_definition := fun i j => hyp.mu_definition j i
m := 1 - 1 / ((hyp.p : ℚ) - 1) - ((hyp.p : ℚ) - 1) / (hyp.p : ℚ) ^ hyp.q +
1 / (((hyp.p : ℚ) - 1) * (hyp.p : ℚ) ^ hyp.q)
m_eq := rfl
Sdata := Tdata
Sdata_U_eq := hU
Sdata_W1_eq := hW1
S_U_commutative := hV
Sdata_W2_eq := hW2
mu_diff_support := fun i _ _ hj hk hdeg =>
pins.nu_diff_support Tdata hU hW1 hW2 i hj hk hdeg
mu_apply_of_not_mem_W2 := fun i j w hwW hwS hw =>
pins.nu_apply_of_not_mem_W1 j i w hwW hwS hw
mu_conj := fun i j => pins.nu_conj j i
eta_conj := fun i j => hyp.eta_conj j i
tau3_isometry := hyp.tau3_isometry
tau3_trivial := hyp.tau3_trivial
tau3_apply_of_regular := fun α w hwW hnot =>
hyp.tau3_apply_of_regular α w hwW (by rwa [Set.union_comm] at hnot)
tau3_mem_ZIrr := hyp.tau3_mem_ZIrr
omega_orthonormal := fun i k j l => by
rw [hyp.omega_orthonormal j l i k]
by_cases h : j = l ∧ i = k
· rw [if_pos h, if_pos ⟨h.2, h.1⟩]
· rw [if_neg h, if_neg fun h' => h ⟨h'.2, h'.1⟩]
omega_apply_one := fun i j => hyp.omega_apply_one j i
omega_mem_ZIrr := fun i j => hyp.omega_mem_ZIrr j i
omega_mul := fun i j w w' => hyp.omega_mul j i w w'
omega_col_zero_apply_of_mem_W2 := fun i w hw =>
hyp.omega_row_zero_apply_of_mem_W1 i w hw
omega_row_zero_apply_of_mem_W1 := fun j w hw =>
hyp.omega_col_zero_apply_of_mem_W2 j w hw
omega_pow_q_of_mem_W1 := fun i j w hw => hyp.omega_pow_p_of_mem_W2 j i w hw
omega_pow_p_of_mem_W2 := fun i j w hw => hyp.omega_pow_q_of_mem_W1 j i w hw
eta_complete_vanish := fun χ horth w hwW hnot =>
hyp.eta_complete_vanish χ (fun i j => horth j i) w hwW
(by rwa [Set.union_comm] at hnot)
eta_fourcorner_vanish := fun i j hi hj x hx => by
rw [Set.union_comm] at hx
have h := hyp.eta_fourcorner_vanish j i hj hi x hx
linear_combination h
eta_row_vanish_of_one_zero := fun x h0 i hi => by
obtain ⟨u, hu⟩ := hyp.eta_column_galois_orbit i hi
rw [← hu, ClassFunction.mapRingEquiv_apply, h0, map_zero]
eta_row_galois_orbit := fun i hi => hyp.eta_column_galois_orbit i hi
eta_column_galois_orbit := fun j hj => hyp.eta_row_galois_orbit j hj
eta_intCast_of_coprime := fun g hg i j =>
hyp.eta_intCast_of_coprime g (by rwa [Nat.mul_comm] at hg) j i
eta_pair_of_coprime := fun g hg i j =>
hyp.eta_pair_of_coprime g (by rwa [Nat.mul_comm] at hg) j i
eta_principal_of_coprime := fun g hg =>
hyp.eta_principal_of_coprime g (by rwa [Nat.mul_comm] at hg)
end OddOrder.Peterfalvi.S15
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/S15_SAndT_Setup/HypothesisSwap.lean