Peterfalvi (9.11.3): sum-of-squares and kernel stability (continuation)
DefinitionOddOrder_Peterfalvi_S11_NineElevenCoherencefeit-thompsonfinite-groupsgroup-theoryishida-odd-order
A continuation of the (9.11) coherence argument with the sum of squares for of (9.11.3), kernel conditions stable under ambient conjugation, and the conjugation-stability of realized -normal subgroups.
Main declarations: sum_xiOf_H0C_degreeSq (Peterfalvi (9.11.3), the 𝒳(H₀C) character sum-of-squares (fully resolved: p^q·u − u)); subsetCharacterKernel_conjBy_iff_of_invariant (Kernel conditions are stable under ambient conjugation); realized_conjByMulEquiv_mem (A realized M-normal subgroup of HU is conjByMulEquiv-stable); nineElevenThree_orbit_split (Peterfalvi (9.11.3), the W₁-orbit count split).
The bundle also proves 12 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_NineElevenCoherence.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_S11_SingleFactorCentralizer
import Definitions.Def_OddOrder_Peterfalvi_S11_NineElevenSetup
namespace OddOrder.Peterfalvi.S11
open OddOrder.RepresentationTheory
open OddOrder.Isaacs.Ch03.IsAInvariant (quotientMulAutHom)
open scoped Pointwise
variable {G : Type*} [Group G]
open OddOrder.Isaacs.Ch03 (IsAInvariant isAInvariant_iff_smul_mem)
section UniformSubfamily
variable [Finite G] {M : Subgroup G}
/-- **The uniform sub-family `sumnS` value.** For a finite family `Si` of irreducible characters of
common degree `d`, `sumnS Si = |Si|·d²`: each member is norm-one (`inner_self_eq_one`, so `Snorm =
degree²`) of degree `d`. The (9.11.5) left endpoint (`sumnS_of_norm_one_constant_degree` fed the
two pointwise facts from irreducibility + the degree). -/
theorem sumnS_irreducible_constant_degree [Fintype ↥M] [Invertible (Nat.card ↥M : ℂ)]
(Si : Finset (ClassFunction ↥M ℂ)) {d : ℕ}
(hirr : ∀ ψ ∈ Si, IsIrreducibleCharacter ψ) (hdeg : ∀ ψ ∈ Si, ψ 1 = (d : ℂ)) :
OddOrder.Peterfalvi.S07.sumnS Si = (Si.card : ℝ) * (d : ℝ) ^ 2 := by
refine OddOrder.Peterfalvi.S07.sumnS_of_norm_one_constant_degree
(fun ψ hψ => ?_) (fun ψ hψ => ?_)
· have h : ClassFunction.inner ψ ψ = 1 := by
simpa using OddOrder.RepresentationTheory.irreducibleCharacter_inner_eq_ite
(⟨ψ, hirr ψ hψ⟩ : IrreducibleCharacter ↥M) ⟨ψ, hirr ψ hψ⟩
rw [h, Complex.one_re]
· rw [hdeg ψ hψ, Complex.natCast_re]
end UniformSubfamily
section FiveArithmetic
/-- **`q + 2 < 2^q` for `q ≥ 3`** (the (9.11.5) endgame `2^x > x+2`, `x ≥ 3`), by induction:
base `3 + 2 = 5 < 8`, step `2·2^q ≥ 2(q+3) ≥ (q+1)+3`. -/
theorem add_two_lt_two_pow {q : ℕ} (hq : 3 ≤ q) : q + 2 < 2 ^ q := by
induction q with
| zero => omega
| succ n ih =>
rcases Nat.lt_or_ge n 3 with hn | hn
· interval_cases n <;> simp_all
· have := ih hn
rw [pow_succ]
omega
/-- **`2·(q.choose 2) = q·(q−1)`** (the `k = 2` binomial coefficient in exact form): `q.choose 2 =
q(q−1)/2` and `q(q−1)` is even (consecutive product). -/
theorem two_mul_choose_two (q : ℕ) : 2 * q.choose 2 = q * (q - 1) := by
rw [Nat.choose_two_right]
rcases Nat.eq_zero_or_pos q with hq | hq
· simp [hq]
· have heven : 2 ∣ q * (q - 1) := by
rcases Nat.even_or_odd q with h | h
· exact Dvd.dvd.mul_right (even_iff_two_dvd.mp h) _
· exact Dvd.dvd.mul_left (even_iff_two_dvd.mp (Nat.Odd.sub_odd h odd_one)) _
exact Nat.mul_div_cancel' heven
/-- **The (9.11.5) binomial lower bound** `2^q·a^q + 2q(q−1)·a² + 2q·a + 1 ≤ (2a+1)^q` (`q ≥ 3`):
the sum of the `k ∈ {0,1,2,q}` terms of `(2a+1)^q = ∑ (2a)^k·C(q,k)`, all others being
nonnegative. -/
theorem binomial_lower_bound {q a : ℕ} (hq : 3 ≤ q) :
2 ^ q * a ^ q + 2 * q * (q - 1) * a ^ 2 + 2 * q * a + 1 ≤ (2 * a + 1) ^ q := by
have hexp : (2 * a + 1) ^ q = ∑ k ∈ Finset.range (q + 1), (2 * a) ^ k * q.choose k := by
rw [add_pow]
exact Finset.sum_congr rfl fun k _ => by rw [one_pow, mul_one, Nat.cast_id]
rw [hexp]
have hsub : ({0, 1, 2, q} : Finset ℕ) ⊆ Finset.range (q + 1) := by
intro x hx
simp only [Finset.mem_insert, Finset.mem_singleton] at hx
rw [Finset.mem_range]; omega
refine le_trans ?_ (Finset.sum_le_sum_of_subset hsub)
-- The subset sum over `{0,1,2,q}` equals the four extracted terms.
have h01 : (0 : ℕ) ∉ ({1, 2, q} : Finset ℕ) := by simp; omega
have h12 : (1 : ℕ) ∉ ({2, q} : Finset ℕ) := by simp; omega
have h2q : (2 : ℕ) ∉ ({q} : Finset ℕ) := by simp; omega
rw [show ({0, 1, 2, q} : Finset ℕ) = insert 0 (insert 1 (insert 2 {q})) from rfl,
Finset.sum_insert h01, Finset.sum_insert h12, Finset.sum_insert h2q, Finset.sum_singleton]
-- Evaluate each term.
have e0 : (2 * a) ^ 0 * q.choose 0 = 1 := by simp
have e1 : (2 * a) ^ 1 * q.choose 1 = 2 * q * a := by
rw [pow_one, Nat.choose_one_right]; ring
have e2 : (2 * a) ^ 2 * q.choose 2 = 2 * q * (q - 1) * a ^ 2 := by
rw [mul_pow, show (2 : ℕ) ^ 2 = 4 from rfl, mul_comm ((4 : ℕ) * a ^ 2) (q.choose 2),
← mul_assoc, show (q.choose 2) * 4 = 2 * (2 * q.choose 2) by ring, two_mul_choose_two]
ring
have eq : (2 * a) ^ q * q.choose q = 2 ^ q * a ^ q := by
rw [Nat.choose_self, mul_one, mul_pow]
rw [e0, e1, e2, eq]; omega
/-- **Peterfalvi (9.11.5), the arithmetic contradiction.** For `q ≥ 3` and `a ≥ 1`, the polynomial
bound `(2a+1)^q − 1 ≤ (q+2)a³ + q²a² + 2qa` (from `|𝒮₄| ≤ ‖α‖²` via (9.11.2)-(9.11.4), with
`p = 2a+1`) is impossible: the binomial lower bound forces `2^q·a^q ≤ (q+2)·a³` (the `a²` terms
cancel since `2q(q−1) ≥ q²` for `q ≥ 2`), hence `2^q ≤ q+2` (as `a^q ≥ a³`), contradicting
`add_two_lt_two_pow`. -/
theorem nineElevenFive_arithmetic_contradiction {q a : ℕ} (hq : 3 ≤ q) (ha : 1 ≤ a)
(hbound : (2 * a + 1) ^ q - 1 ≤ (q + 2) * a ^ 3 + q ^ 2 * a ^ 2 + 2 * q * a) : False := by
have hlb := binomial_lower_bound (a := a) hq
-- `2^q a^q + 2q(q−1)a² + 2qa ≤ (2a+1)^q − 1 ≤ (q+2)a³ + q²a² + 2qa`, cancel `2qa`.
have hchain : 2 ^ q * a ^ q + 2 * q * (q - 1) * a ^ 2
≤ (q + 2) * a ^ 3 + q ^ 2 * a ^ 2 := by
have h1 : 2 ^ q * a ^ q + 2 * q * (q - 1) * a ^ 2 + 2 * q * a
≤ (2 * a + 1) ^ q - 1 := by omega
omega
-- `q²a² ≤ 2q(q−1)a²` (since `2q(q−1) ≥ q²` for `q ≥ 2`), so `2^q a^q ≤ (q+2)a³`.
have hq2 : q ^ 2 ≤ 2 * q * (q - 1) := by nlinarith [Nat.sub_add_cancel (by omega : 1 ≤ q)]
have hqaq : 2 ^ q * a ^ q ≤ (q + 2) * a ^ 3 := by nlinarith [Nat.mul_le_mul_right (a ^ 2) hq2]
-- `a^q ≥ a^3` (as `q ≥ 3`, `a ≥ 1`), so `2^q a³ ≤ (q+2)a³`, giving `2^q ≤ q+2`.
have haq : a ^ 3 ≤ a ^ q := Nat.pow_le_pow_right ha hq
have hfinal : 2 ^ q ≤ q + 2 := by
have ha3 : 0 < a ^ 3 := pow_pos (by omega : (0 : ℕ) < a) 3
have : 2 ^ q * a ^ 3 ≤ (q + 2) * a ^ 3 :=
le_trans (Nat.mul_le_mul_left (2 ^ q) haq) hqaq
exact Nat.le_of_mul_le_mul_right this ha3
exact absurd hfinal (Nat.not_le.mpr (add_two_lt_two_pow hq))
/-- **Peterfalvi (9.11.3), the `|𝒮₄|` count** (arithmetic core, denominator-cleared).
Book (9.11.3): the sum-of-squares class equation on the quotient `HŪ/(H₀C)` — `n` characters of
degree `u` and `q(p−1)u/a²` of degree `a`, plus the `u` linear characters of `Ū` — gives
`p^q·u = u + n·u² + q(p−1)·u` (`hclass`). Cancelling one `u` yields `n·u + q(p−1) + 1 = p^q`; with
the conjugate/reducible split `n = q·|𝒮₄| + (p−1)` (`hn`: `(9.8.b)` gives `p−1` of the `n` inducing
reducibly, the rest falling into `W₁`-orbits of size `q`), this rearranges to the cleared count
`|𝒮₄|·qu + (p−1)u + (p−1)q + 1 = p^q` — the `hcount` input of `nineElevenFive_refutation` (with
`p = 2a+1`). The two group inputs `hclass` (character sum-of-squares) and `hn` (W₁-orbit count) are
the deep (9.11.3) content, supplied separately. -/
theorem nineElevenThree_count {p q u n S4 : ℕ} (hu : 1 ≤ u)
(hclass : u + n * u ^ 2 + q * (p - 1) * u = p ^ q * u)
(hn : n = S4 * q + (p - 1)) :
S4 * (q * u) + (p - 1) * u + (p - 1) * q + 1 = p ^ q := by
-- Cancel one `u` from the class equation: `n·u + q(p−1) + 1 = p^q`.
have key : (n * u + q * (p - 1) + 1) * u = p ^ q * u := by
have hexp : (n * u + q * (p - 1) + 1) * u = u + n * u ^ 2 + q * (p - 1) * u := by ring
rw [hexp]; exact hclass
have hnu : n * u + q * (p - 1) + 1 = p ^ q :=
Nat.eq_of_mul_eq_mul_right (by omega) key
-- Substitute the `W₁`-orbit split and reassociate.
rw [hn] at hnu
have hre : (S4 * q + (p - 1)) * u + q * (p - 1) + 1
= S4 * (q * u) + (p - 1) * u + (p - 1) * q + 1 := by ring
rw [← hre]; exact hnu
/-- **Peterfalvi (9.11.5), the full refutation** (`|𝒮₄| ≤ ‖α‖²` is impossible).
The (9.11.5) argument in denominator-cleared `ℕ` form. Inputs (all `ℕ`, integrality already used):
* `hcount` — (9.11.3) cleared: `|𝒮₄|·qu + (p−1)q + (p−1)u + 1 = p^q`, i.e. with `p = 2a+1`,
`S₄·qu + 2aq + 2au + 1 = (2a+1)^q`. (Book: `|𝒮₄| = (p^q−1)/(qu) − (p−1)/u − (p−1)/q`.)
* `hnorm` — (9.11.4) cleared: `‖α‖²·u = (a+1)u + (q−1)a²`. (Book: `‖α‖² = a+1+(q−1)a²/u`;
`‖α‖² ∈ ℤ` as a virtual-character norm.)
* `hua2` — (9.11.2): `u ≤ a²`.
* `hle` — the contradiction hypothesis `|𝒮₄| ≤ ‖α‖²`.
From `hle`: `S₄·qu ≤ q·(‖α‖²·u) = (a+1)qu + q(q−1)a²`; substituting into `hcount` and using `u ≤ a²`
to bound the `u`-linear part `(aq+q+2a)u ≤ (aq+q+2a)a²` gives `(2a+1)^q − 1 ≤ (q+2)a³ + q²a² + 2qa`,
refuted by `nineElevenFive_arithmetic_contradiction`. This is what the (9.11.6)–(9.11.8) coherence
contradiction (or directly the maximality refutation) feeds. -/
theorem nineElevenFive_refutation {q a u S4 N : ℕ} (hq : 3 ≤ q) (ha : 1 ≤ a) (_hu : 1 ≤ u)
(hua2 : u ≤ a * a)
(hcount : S4 * (q * u) + 2 * a * q + 2 * a * u + 1 = (2 * a + 1) ^ q)
(hnorm : N * u = (a + 1) * u + (q - 1) * a ^ 2)
(hle : S4 ≤ N) : False := by
refine nineElevenFive_arithmetic_contradiction hq ha ?_
rw [← hcount, Nat.add_sub_cancel]
-- `S₄·(q·u) ≤ (a+1)·q·u + q·(q−1)·a²`.
have key1 : S4 * (q * u) ≤ (a + 1) * (q * u) + q * ((q - 1) * a ^ 2) := by
calc S4 * (q * u) ≤ N * (q * u) := Nat.mul_le_mul_right _ hle
_ = q * (N * u) := by ring
_ = q * ((a + 1) * u + (q - 1) * a ^ 2) := by rw [hnorm]
_ = (a + 1) * (q * u) + q * ((q - 1) * a ^ 2) := by ring
-- `(a·q + q + 2·a)·u ≤ (a·q + q + 2·a)·a²` from `u ≤ a²`.
have key2 : (a * q + q + 2 * a) * u ≤ (a * q + q + 2 * a) * (a * a) :=
Nat.mul_le_mul_left _ hua2
-- Clear the `q − 1` subtraction (`q ≥ 3`), then combine as a polynomial inequality.
obtain ⟨q', rfl⟩ : ∃ q', q = q' + 1 := ⟨q - 1, by omega⟩
simp only [Nat.add_sub_cancel] at key1 ⊢
nlinarith [key1, key2, ha, _hu, sq_nonneg a, Nat.zero_le q']
end FiveArithmetic
section CharacterSumOfSquares
variable [Finite G] {M : Subgroup G}
/-- **Peterfalvi (9.11.3) quotient order**: the realized `H₀C` in `HU` has index `p^q·u`.
`[HU : H₀C] = [HU : HC]·[HC : H₀C] = u·p^q`, where `HC = H·C = hInHu ⊔ cInHu`:
* `[HU : HC] = u` (`index_hcInHu_eq_relindex_cInHu` ∘ `index_cInHu_subgroupOf_uInHu_eq_u`);
* `[HC : H₀C] = p^q` by the second isomorphism `HC/H₀C ≅ H/(H ⊓ H₀C) = H/H₀ = H̄`
(`relIndex_sup_right` with `HC = H ⊔ H₀C`, and `H ⊓ H₀C = H₀` via
`hInHu_inf_realizedH0supC_eq_realizedH0`), so `[HC : H₀C] = [H : H₀] = |H|/|H₀| = p^q`
(`chief.quotient_order`).
This is the first term of the (9.11.3) class equation, folded into `sum_xiOf_H0C_degreeSq`. -/
theorem index_realizedH0supC_eq (data : TypesIIIIIIVSetup M) (chief : ChiefFactorData data)
(chars : Section11CharacterData data chief) :
Nat.card (↥(huSub data) ⧸
((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data))
= chief.p ^ data.q * chars.u := by
have : (((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data)).Normal :=
(chiefFactor_H0supC_subgroupOf_normal chief).subgroupOf (huSub data)
rw [← Subgroup.index_eq_card]
have hHC : hInHu data ⊔ ((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data)
= hInHu data ⊔ cInHu data chief := hInHu_sup_realizedH0supC chief
have hle : ((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data)
≤ hInHu data ⊔ cInHu data chief := by rw [← hHC]; exact le_sup_right
have hu : (hInHu data ⊔ cInHu data chief).index = chars.u :=
(index_hcInHu_eq_relindex_cInHu data chief).trans
(index_cInHu_subgroupOf_uInHu_eq_u data chief chars)
-- `[HC : H₀C] = [H : H₀] = p^q` via the second isomorphism theorem.
have hrel : (((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data)).relIndex
(hInHu data ⊔ cInHu data chief) = chief.p ^ data.q := by
rw [← hHC, Subgroup.relIndex_sup_right, ← Subgroup.inf_relIndex_left,
hInHu_inf_realizedH0supC_eq_realizedH0 chief]
-- `[H : H₀] = |H|/|H₀| = p^q`.
have hH0le : (chief.H0.subgroupOf M).subgroupOf (huSub data) ≤ hInHu data :=
Subgroup.subgroupOf_mono (huSub data) (Subgroup.subgroupOf_mono M chief.H0_lt_H.le)
have hcard_H0r : Nat.card ↥((chief.H0.subgroupOf M).subgroupOf (huSub data))
= Nat.card ↥chief.H0 := by
have hH0M : chief.H0 ≤ M := chief.H0_lt_H.le.trans (H_le_M data)
have hH0subM : chief.H0.subgroupOf M ≤ huSub data :=
Subgroup.subgroupOf_mono M (chief.H0_lt_H.le.trans le_sup_left)
calc Nat.card ↥((chief.H0.subgroupOf M).subgroupOf (huSub data))
= Nat.card ↥(chief.H0.subgroupOf M) :=
Nat.card_congr (Subgroup.subgroupOfEquivOfLe hH0subM).toEquiv
_ = Nat.card ↥chief.H0 := Nat.card_congr (Subgroup.subgroupOfEquivOfLe hH0M).toEquiv
have hlag := Subgroup.index_mul_card
(((chief.H0.subgroupOf M).subgroupOf (huSub data)).subgroupOf (hInHu data))
rw [Nat.card_congr (Subgroup.subgroupOfEquivOfLe hH0le).toEquiv, hcard_H0r,
Nat.card_congr (hInHuEquivH data).toEquiv, chief.quotient_order] at hlag
exact Nat.eq_of_mul_eq_mul_right Nat.card_pos hlag
rw [← Subgroup.relIndex_mul_index hle, hrel, hu]
open scoped Classical in
/-- **Peterfalvi (9.11.3), the `𝒳(H₀C)` character sum-of-squares (fully resolved: `p^q·u − u`).**
The kernel-interval degree-square sum (`sumDegreeSq_kernelInterval`) instantiated at the realized
`N = H₀C`, `K = H` inside `HU`: over `𝒳(H₀C) = {χ ∈ Irr(HU) | H₀C ⊆ ker, H ⊄ ker}`,
`∑ χ(1)² = |HU/(H₀C)| − |HU/(H ⊔ H₀C)|`, and `H ⊔ H₀C = HC` (`hInHu_sup_realizedH0supC`) with
`|HU/HC| = [HU:HC] = u` (`index_hcInHu`) and `|HU/(H₀C)| = p^q·u` (`index_realizedH0supC_eq`).
So `∑ χ(1)² = p^q·u − u`, the (9.11.3) class equation `|Ū| = |HU/H₀C| − ∑ χ(1)²` value fed to
`nineElevenThree_count`'s `hclass`. -/
theorem sum_xiOf_H0C_degreeSq (data : TypesIIIIIIVSetup M) (chief : ChiefFactorData data)
(chars : Section11CharacterData data chief) :
∑ χ ∈ Finset.univ.filter (fun χ : IrreducibleCharacter ↥(huSub data) =>
(((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data) :
Set ↥(huSub data)) ⊆
OddOrder.Peterfalvi.S03.characterKernel (χ : ClassFunction ↥(huSub data) ℂ) ∧
¬ ((hInHu data : Set ↥(huSub data)) ⊆
OddOrder.Peterfalvi.S03.characterKernel (χ : ClassFunction ↥(huSub data) ℂ))),
((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2
= (chief.p ^ data.q * chars.u : ℂ) - (chars.u : ℂ) := by
have : (((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data)).Normal :=
(chiefFactor_H0supC_subgroupOf_normal chief).subgroupOf (huSub data)
have : (hInHu data).Normal := hInHu_normal data
rw [OddOrder.RepresentationTheory.sumDegreeSq_kernelInterval
(((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data)) (hInHu data)]
have hHC : hInHu data ⊔ ((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data)
= hInHu data ⊔ cInHu data chief := hInHu_sup_realizedH0supC chief
have hu : (hInHu data ⊔ cInHu data chief).index = chars.u :=
(index_hcInHu_eq_relindex_cInHu data chief).trans
(index_cInHu_subgroupOf_uInHu_eq_u data chief chars)
-- First term `|HU/(H₀C)| = p^q·u` (`index_realizedH0supC_eq`); second `|HU/HC| = u`.
congr 1
· exact_mod_cast index_realizedH0supC_eq data chief chars
· rw [hHC]; exact_mod_cast hu
end CharacterSumOfSquares
section NineElevenFour
/-- **Peterfalvi (9.11.4), the norm reduction `‖γ − ψ‖² = ‖γ‖² + 1`** (`cfnormBd`).
For a norm-one irreducible character `ψ` orthogonal to a class function `γ` (`⟨γ, ψ⟩ = 0`), the
difference `α = γ − ψ` has `⟨α, α⟩ = ⟨γ, γ⟩ + 1`: sesquilinearity expands
`⟨γ−ψ, γ−ψ⟩ = ⟨γ,γ⟩ − ⟨γ,ψ⟩ − ⟨ψ,γ⟩ + ⟨ψ,ψ⟩`, with `⟨γ,ψ⟩ = 0` (hypothesis),
`⟨ψ,γ⟩ = conj⟨γ,ψ⟩ = 0` (`inner_conj_symm`), and `⟨ψ,ψ⟩ = 1`
(`IsIrreducibleCharacter.inner_self_eq_one`). This is the character-side step of (9.11.4): with
`γ = Ind_{HU₁}^M 1` and `ψ = ψ₁ ∈ 𝒮₁` (orthogonal because `H ⊆ ker` of every constituent of `γ`
but not of `ψ₁`), it gives `‖α‖² = ‖γ‖² + 1`; the remaining `‖γ‖²` is the (9.11.2)-gated Mackey
count. Stated for an arbitrary character `γ` (no coherence/induction hypothesis), so it is reusable
wherever a norm-one irreducible is subtracted off orthogonally. -/
theorem cfnorm_sub_irreducible_orthogonal {Γ : Type*} [Group Γ] [Fintype Γ]
[Invertible (Nat.card Γ : ℂ)] {γ ψ : ClassFunction Γ ℂ}
(hψ : IsIrreducibleCharacter ψ) (hortho : ClassFunction.inner γ ψ = 0) :
ClassFunction.inner (γ - ψ) (γ - ψ) = ClassFunction.inner γ γ + 1 := by
have hψγ : ClassFunction.inner ψ γ = 0 := by
rw [inner_conj_symm γ ψ, hortho, star_zero]
have hψψ : ClassFunction.inner ψ ψ = 1 := by
simpa using irreducibleCharacter_inner_eq_ite
(⟨ψ, hψ⟩ : IrreducibleCharacter Γ) ⟨ψ, hψ⟩
rw [ClassFunction.inner_sub_left, ClassFunction.inner_sub_right, ClassFunction.inner_sub_right,
hortho, hψγ, hψψ]
ring
end NineElevenFour
/-- **Peterfalvi (9.11.2)–(9.11.5), the equality-branch refutation.**
In the (9.11.1) equality configuration `p = 2a+1` (`hpeq`), the three deep inputs — the (9.11.2)
inertia identity (`hK₁`/`hK₂`/`hCinf`), the (9.11.3) class equation and `W₁`-orbit split
(`hclass`/`hn`), and the (9.11.4) Mackey norm (`hnorm`) — combine with the coherence bound
`|𝒮₄| ≤ ‖α‖²` (`hle`) to a contradiction. This is the equality branch of the (9.11) caseA refuter:
`nineElevenOne_configuration` produces `hpeq` (and `C = U′`, `χdeg = u`) from the (9.11.1) squeeze,
and here (9.11.2)/(9.11.3)/(9.11.4) are chained through to `nineElevenFive_refutation`. Only the
three named group/character inputs remain honest content; the arithmetic is fully discharged. -/
theorem nineElevenCaseA_equality_refutation [Finite G] {M : Subgroup G}
{data : TypesIIIIIIVSetup M} {chief : ChiefFactorData data}
{chars : Section11CharacterData data chief} (caseA : CliffordCaseAData chars)
(hq3 : 3 ≤ data.q) (hu : 1 ≤ chars.u) (hpeq : chief.p = 2 * caseA.a + 1)
{K₁ K₂ : Subgroup G} (hK₁ : K₁.relIndex data.U = caseA.a)
(hK₂ : K₂.relIndex data.U = caseA.a) (hCinf : chars.C = K₁ ⊓ K₂)
{n S4 : ℕ}
(hclass : chars.u + n * chars.u ^ 2 + data.q * (chief.p - 1) * chars.u
= chief.p ^ data.q * chars.u)
(hn : n = S4 * data.q + (chief.p - 1))
{N : ℕ} (hnorm : N * chars.u = (caseA.a + 1) * chars.u + (data.q - 1) * caseA.a ^ 2)
(hle : S4 ≤ N) : False := by
have ha : 1 ≤ caseA.a := caseA.a_pos
-- (9.11.2): `u ≤ a²` from the inertia identity `C = K₁ ⊓ K₂`.
have hua2 : chars.u ≤ caseA.a * caseA.a := nineElevenTwo_u_le_a_sq caseA hK₁ hK₂ hCinf
-- (9.11.3): the cleared count `|𝒮₄|·qu + (p−1)u + (p−1)q + 1 = p^q`.
have hcount0 : S4 * (data.q * chars.u) + (chief.p - 1) * chars.u
+ (chief.p - 1) * data.q + 1 = chief.p ^ data.q := nineElevenThree_count hu hclass hn
-- Substitute `p = 2a+1` to match `nineElevenFive_refutation`'s `(2a+1)^q` form.
have hp1 : chief.p - 1 = 2 * caseA.a := by omega
have hcount : S4 * (data.q * chars.u) + 2 * caseA.a * data.q + 2 * caseA.a * chars.u + 1
= (2 * caseA.a + 1) ^ data.q := by
rw [hp1, hpeq] at hcount0; omega
-- (9.11.5): the exponential-beats-polynomial contradiction.
exact nineElevenFive_refutation hq3 ha hu hua2 hcount hnorm hle
section NineElevenThreeOrbitSplit
variable [Finite G] {M : Subgroup G}
omit [Finite G] in
/-- **Kernel conditions are stable under ambient conjugation** (`cfker_conjg` at an invariant
subgroup). For `N ≤ HU` whose underlying set is carried into itself by every `M`-conjugation
(`conjByMulEquiv`), the kernel condition `N ⊆ Ker χ` is invariant under `χ ↦ χ^g` (`g ∈ M`):
both directions of `subsetCharacterKernel_conjBy_iff`, transported along the invariance. This is
what makes the `𝒳(H₀C)` filter conjugation-closed — the `hT` input of the (9.11.3) orbit count. -/
theorem subsetCharacterKernel_conjBy_iff_of_invariant (data : TypesIIIIIIVSetup M)
(N : Subgroup ↥(huSub data))
(hN : ∀ (w : ↥M) (x : ↥(huSub data)), x ∈ N →
ClassFunction.conjByMulEquiv (G := ↥M) (H := huSub data) w x ∈ N)
(g : ↥M) (χ : ClassFunction ↥(huSub data) ℂ) :
((N : Set ↥(huSub data)) ⊆ OddOrder.Peterfalvi.S03.characterKernel
(ClassFunction.conjBy (G := ↥M) (H := huSub data) g χ))
↔ (N : Set ↥(huSub data)) ⊆ OddOrder.Peterfalvi.S03.characterKernel χ := by
rw [subsetCharacterKernel_conjBy_iff]
constructor
· intro h x hx
have h1 := h _ (hN g⁻¹ x hx)
have h2 : ClassFunction.conjByMulEquiv (G := ↥M) (H := huSub data) g
(ClassFunction.conjByMulEquiv (G := ↥M) (H := huSub data) g⁻¹ x) = x := by
apply Subtype.ext
rw [ClassFunction.conjByMulEquiv_apply, ClassFunction.conjByMulEquiv_apply]
group
rwa [h2] at h1
· intro h x hx
exact h (hN g x hx)
omit [Finite G] in
/-- **A realized `M`-normal subgroup of `HU` is `conjByMulEquiv`-stable**: for `Y ≤ G` with
`(Y.subgroupOf M) ◁ M`, the realization `(Y.subgroupOf M).subgroupOf HU` is carried into itself by
every `M`-conjugation of `HU`. Supplies the invariance input of
`subsetCharacterKernel_conjBy_iff_of_invariant` at `N = H₀C`-realized
(`chiefFactor_H0supC_subgroupOf_normal`) and `N = H`-realized (`hSubgroupOfM_normal`). -/
theorem realized_conjByMulEquiv_mem (data : TypesIIIIIIVSetup M) {Y : Subgroup G}
(hY : (Y.subgroupOf M).Normal) (w : ↥M) (x : ↥(huSub data))
(hx : x ∈ (Y.subgroupOf M).subgroupOf (huSub data)) :
ClassFunction.conjByMulEquiv (G := ↥M) (H := huSub data) w x
∈ (Y.subgroupOf M).subgroupOf (huSub data) := by
rw [Subgroup.mem_subgroupOf] at hx ⊢
have hval : ((ClassFunction.conjByMulEquiv (G := ↥M) (H := huSub data) w x :
↥(huSub data)) : ↥M) = w * ((x : ↥(huSub data)) : ↥M) * w⁻¹ :=
ClassFunction.conjByMulEquiv_apply (G := ↥M) (H := huSub data) w x
rw [hval]
exact hY.conj_mem _ hx w
/-- **A source inducing irreducibly has inertia index `q`** (the free `W₁`-orbit of (9.11.3)).
If `Ind_{HU}^M χ` is irreducible then `‖Ind χ‖² = 1`, so `|I_M(χ)| = |HU|`
(`card_mul_inner_self_induce_eq_card_inertia`), and `[M : I_M(χ)]·|I_M(χ)| = |M| = q·|HU|` gives
`[M : I_M(χ)] = q` — the source has exactly `q` distinct `M`-conjugates ("each of the others has
`q` conjugates under `W₁`"). -/
theorem inertia_index_eq_q_of_induce_irreducible (data : TypesIIIIIIVSetup M)
[Fintype ↥M] [Invertible (Nat.card ↥M : ℂ)]
[Invertible (Nat.card ↥(huSub data) : ℂ)]
(θ : IrreducibleCharacter ↥(huSub data))
(hirr : IsIrreducibleCharacter (ClassFunction.induce (huSub data) θ.toClassFunction)) :
(IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ).index = data.q := by
let : Fintype ↥(huSub data) := Fintype.ofFinite _
have hone : ClassFunction.inner (ClassFunction.induce (huSub data) θ.toClassFunction)
(ClassFunction.induce (huSub data) θ.toClassFunction) = 1 := by
have h := irreducibleCharacter_inner_eq_ite
(⟨_, hirr⟩ : IrreducibleCharacter ↥M) ⟨_, hirr⟩
rwa [if_pos rfl] at h
have hcard := card_mul_inner_self_induce_eq_card_inertia (G := ↥M) (H := huSub data) θ
rw [hone, mul_one] at hcard
have hcardN : Nat.card ↥(huSub data)
= Nat.card ↥(IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ) :=
Nat.cast_inj.mp hcard
have hchain : (IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ).index
* Nat.card ↥(huSub data) = data.q * Nat.card ↥(huSub data) := by
calc (IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ).index
* Nat.card ↥(huSub data)
= (IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ).index
* Nat.card ↥(IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ) := by
rw [← hcardN]
_ = Nat.card ↥M := Subgroup.index_mul_card _
_ = (huSub data).index * Nat.card ↥(huSub data) := (Subgroup.index_mul_card _).symm
_ = data.q * Nat.card ↥(huSub data) := by rw [huSub_index_eq_q]
exact Nat.eq_of_mul_eq_mul_right Nat.card_pos hchain
/-- **A source inducing reducibly has inertia index `1`** (the `W₁`-invariant sources of
(9.11.3), "`p−1` of these characters induce reducible characters of `M`" — each reducible member
has a *single* source). `HU ≤ I_M(χ)` always (`subgroup_le_inertia`), so `[M : I_M(χ)]` divides
`[M : HU] = q`; `q` prime leaves `1` or `q`, and index `q` would force `I_M(χ) = HU`, making
`Ind_{HU}^M χ` irreducible (`isIrreducibleCharacter_induce_of_inertia_eq`) — contradiction. -/
theorem inertia_index_eq_one_of_induce_reducible (data : TypesIIIIIIVSetup M)
[Fintype ↥M] [Invertible (Nat.card ↥M : ℂ)]
[Invertible (Nat.card ↥(huSub data) : ℂ)]
(θ : IrreducibleCharacter ↥(huSub data))
(hred : ¬ IsIrreducibleCharacter (ClassFunction.induce (huSub data) θ.toClassFunction)) :
(IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ).index = 1 := by
let : Fintype ↥(huSub data) := Fintype.ofFinite _
have hqp : (data.q).Prime := data.nontrivial.2.1
have hle : huSub data ≤ IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ :=
IrreducibleCharacter.subgroup_le_inertia θ
have hdvd : (IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ).index
∣ (huSub data).index := Subgroup.index_dvd_of_le hle
rw [huSub_index_eq_q] at hdvd
rcases hqp.eq_one_or_self_of_dvd _ hdvd with h1 | hq'
· exact h1
· exfalso
have hri := Subgroup.relIndex_mul_index hle
rw [hq', huSub_index_eq_q] at hri
have h1' : (huSub data).relIndex
(IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ) = 1 :=
Nat.eq_of_mul_eq_mul_right hqp.pos (by rw [hri, one_mul])
have heq : IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ = huSub data :=
le_antisymm (Subgroup.relIndex_eq_one.mp h1') hle
exact hred (isIrreducibleCharacter_induce_of_inertia_eq θ heq)
open scoped Classical in
/-- **Peterfalvi (9.11.3), the `W₁`-orbit count split** (issue 9083 Phase C).
In the (9.11.1) equality configuration — the degree-`qa` subfamily of `𝒮(H₀U′)` inside `𝒮₂`
(`hS₁'sub`), every `𝒮(H₀C)`-member outside `𝒮₂` of degree `qu` (`hS3deg`), every `𝒮₂`-member of
degree `qa` (`hS2deg`, the Phase-E `𝒮₂ = 𝒮₁` extraction), `C = U′` (`hCU`), and the (9.8.d) count
equality (`hcount`) — the `𝒳(H₀C)` class equation holds with the character count `n` already
split into `W₁`-orbits:
`u + (|𝒮₄|·q + (p−1))·u² + q·(p−1)·u = p^q·u`,
where `𝒮₄ = {φ ∈ 𝒮(H₀C) | φ irreducible, φ ∉ 𝒮₂}` (the book's irreducible members of
`𝒮₃ ∩ 𝒮(H₀C)`; `𝒮(H₀C) ⊆ 𝒮(H₀C′)` makes "`∈ 𝒮₃`" the same as "`∉ 𝒮₂`"). This is
`nineElevenCaseA_equality_refutation`'s `hclass` instantiated at `n = |𝒮₄|·q + (p−1)`, so its
`hn` is definitional.
Proof: fibre `𝒳(H₀C)` over `𝒮(H₀C)` under `Ind_{HU}^M`. Fibres are full `M`-conjugation orbits
(`card_filter_induce_eq_index_inertia`; the filter is conjugation-closed since `H₀C`-realized and
`H`-realized are `M`-normal), of size `1` over reducible members and `q` over irreducible ones
(`inertia_index_eq_one_of_induce_reducible` / `inertia_index_eq_q_of_induce_irreducible`). The
three member classes have `p − 1` (`reducible_count_sOf_H0C`), `|𝒮₁′|`, and `|𝒮₄|` members with
source degrees `u` ((9.8.b) `caseA_reducible_induceHU_apply_one_eq_qu`), `a`, and `u`
respectively; `|𝒮₁′|·a² = (p−1)·[U:U′] = (p−1)·u` (`hcount`, `relIndex_cSub_U_eq_u` along
`C = U′`); and `∑_{𝒳(H₀C)} χ(1)² = p^q·u − u` (`sum_xiOf_H0C_degreeSq`) closes the ledger. -/
theorem nineElevenThree_orbit_split (hG : OddOrder.BG.IsMinimalSimpleOdd G)
{data : TypesIIIIIIVSetup M} {chief : ChiefFactorData data}
{chars : Section11CharacterData data chief} (caseA : CliffordCaseAData chars)
{S₂ : Set (ClassFunction ↥M ℂ)}
(hS₁'sub : {χ ∈ sOf data (chief.H0 ⊔ uprimeSub data) | IsIrreducibleCharacter χ ∧
χ 1 = ((data.q * caseA.a : ℕ) : ℂ)} ⊆ S₂)
(hS3deg : ∀ χ ∈ sOf data (chief.H0 ⊔ cSub data chief), χ ∉ S₂ →
χ 1 = ((data.q * chars.u : ℕ) : ℂ))
(hS2deg : ∀ χ ∈ S₂, χ 1 = ((data.q * caseA.a : ℕ) : ℂ))
(hCU : cSub data chief = uprimeSub data)
(hcount : {χ ∈ sOf data (chief.H0 ⊔ uprimeSub data) | IsIrreducibleCharacter χ ∧
χ 1 = ((data.q * caseA.a : ℕ) : ℂ)}.ncard * (caseA.a * caseA.a)
= (chief.p - 1) * ((uprimeSub data).relIndex data.U)) :
chars.u + ({φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∉ S₂}.ncard * data.q + (chief.p - 1)) * chars.u ^ 2
+ data.q * (chief.p - 1) * chars.u
= chief.p ^ data.q * chars.u := by
have := huSub_normal data
let : Fintype ↥M := Fintype.ofFinite _
let : Fintype ↥(huSub data) := Fintype.ofFinite _
let : Invertible (Nat.card ↥M : ℂ) :=
invertibleOfNonzero (Nat.cast_ne_zero.mpr Nat.card_pos.ne')
let : Invertible (Nat.card ↥(huSub data) : ℂ) :=
invertibleOfNonzero (Nat.cast_ne_zero.mpr Nat.card_pos.ne')
have hqp : (data.q).Prime := data.nontrivial.2.1
have hq0 : (data.q : ℂ) ≠ 0 := Nat.cast_ne_zero.mpr hqp.pos.ne'
-- ── the source family `T = 𝒳(H₀C)`, as the `sum_xiOf_H0C_degreeSq` filter
set T : Finset (IrreducibleCharacter ↥(huSub data)) :=
Finset.univ.filter (fun χ : IrreducibleCharacter ↥(huSub data) =>
(((chief.H0 ⊔ cSub data chief).subgroupOf M).subgroupOf (huSub data) :
Set ↥(huSub data)) ⊆
OddOrder.Peterfalvi.S03.characterKernel (χ : ClassFunction ↥(huSub data) ℂ) ∧
¬ ((hInHu data : Set ↥(huSub data)) ⊆
OddOrder.Peterfalvi.S03.characterKernel (χ : ClassFunction ↥(huSub data) ℂ)))
with hT_def
have hT_mem : ∀ θ : IrreducibleCharacter ↥(huSub data),
θ ∈ T ↔ θ ∈ xiOf data (chief.H0 ⊔ cSub data chief) := by
intro θ
rw [hT_def, Finset.mem_filter]
constructor
· rintro ⟨-, hker, hH⟩
exact ⟨hH, hker⟩
· rintro ⟨hH, hker⟩
exact ⟨Finset.mem_univ _, hker, hH⟩
-- ── `T` is `M`-conjugation-closed (both kernel conditions are `M`-normal)
have hH0Cnorm : ((chief.H0 ⊔ cSub data chief).subgroupOf M).Normal :=
chiefFactor_H0supC_subgroupOf_normal chief
have hTconj : ∀ θ ∈ T, ∀ g : ↥M,
IrreducibleCharacter.conjBy (G := ↥M) (H := huSub data) g θ ∈ T := by
intro θ hθ g
rw [hT_def, Finset.mem_filter] at hθ ⊢
obtain ⟨-, hker, hH⟩ := hθ
refine ⟨Finset.mem_univ _, ?_, ?_⟩
· rw [IrreducibleCharacter.coe_conjBy]
exact (subsetCharacterKernel_conjBy_iff_of_invariant data _
(realized_conjByMulEquiv_mem data hH0Cnorm) g _).mpr hker
· rw [IrreducibleCharacter.coe_conjBy]
intro hcon
exact hH ((subsetCharacterKernel_conjBy_iff_of_invariant data _
(realized_conjByMulEquiv_mem data (hSubgroupOfM_normal data)) g _).mp hcon)
-- ── membership dictionary: `T`-sources induce into `𝒮(H₀C)`
have hmem_sOf : ∀ θ ∈ T, ClassFunction.induce (huSub data) θ.toClassFunction
∈ sOf data (chief.H0 ⊔ cSub data chief) := fun θ hθ =>
⟨θ, (hT_mem θ).mp hθ, (induceHU_eq_induce data _).symm⟩
-- ── generic fibre count: a fibre-closed part of `T` with constant inertia index `c`
-- has cardinality `|image|·c`
have hfib_card : ∀ (T' : Finset (IrreducibleCharacter ↥(huSub data))) (c : ℕ),
T' ⊆ T →
(∀ θ θ₀ : IrreducibleCharacter ↥(huSub data),
ClassFunction.induce (huSub data) θ.toClassFunction
= ClassFunction.induce (huSub data) θ₀.toClassFunction →
θ ∈ T → θ₀ ∈ T' → θ ∈ T') →
(∀ θ₀ ∈ T', (IrreducibleCharacter.inertia (G := ↥M) (H := huSub data) θ₀).index = c) →
T'.card = (T'.image
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction)).card * c := by
intro T' c hsub hfibclosed hc
calc T'.card
= ∑ φ ∈ T'.image (fun θ => ClassFunction.induce (huSub data) θ.toClassFunction),
(T'.filter (fun θ =>
ClassFunction.induce (huSub data) θ.toClassFunction = φ)).card :=
Finset.card_eq_sum_card_image _ _
_ = ∑ _φ ∈ T'.image (fun θ => ClassFunction.induce (huSub data) θ.toClassFunction),
c := by
refine Finset.sum_congr rfl fun φ hφ => ?_
obtain ⟨θ₀, hθ₀T', hθ₀eq⟩ := Finset.mem_image.mp hφ
have hfib : T'.filter (fun θ =>
ClassFunction.induce (huSub data) θ.toClassFunction = φ)
= T.filter (fun θ => ClassFunction.induce (huSub data) θ.toClassFunction
= ClassFunction.induce (huSub data) θ₀.toClassFunction) := by
subst hθ₀eq
ext θ
rw [Finset.mem_filter, Finset.mem_filter]
constructor
· rintro ⟨hθT', heq⟩
exact ⟨hsub hθT', heq⟩
· rintro ⟨hθT, heq⟩
exact ⟨hfibclosed θ θ₀ heq hθT hθ₀T', heq⟩
rw [hfib, card_filter_induce_eq_index_inertia T hTconj θ₀ (hsub hθ₀T'),
hc θ₀ hθ₀T']
_ = (T'.image
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction)).card * c := by
rw [Finset.sum_const, smul_eq_mul]
-- ── generic image dictionary: the members of a part cut out by a member predicate `Q`
have himg : ∀ (T' : Finset (IrreducibleCharacter ↥(huSub data)))
(Q : ClassFunction ↥M ℂ → Prop),
T' ⊆ T →
(∀ θ ∈ T, (θ ∈ T' ↔ Q (ClassFunction.induce (huSub data) θ.toClassFunction))) →
((T'.image (fun θ => ClassFunction.induce (huSub data) θ.toClassFunction) :
Finset (ClassFunction ↥M ℂ)) : Set (ClassFunction ↥M ℂ))
= {φ ∈ sOf data (chief.H0 ⊔ cSub data chief) | Q φ} := by
intro T' Q hsub hQmem
ext φ
constructor
· intro hφ
rw [Finset.mem_coe, Finset.mem_image] at hφ
obtain ⟨θ, hθT', rfl⟩ := hφ
exact ⟨hmem_sOf θ (hsub hθT'), (hQmem θ (hsub hθT')).mp hθT'⟩
· rintro ⟨⟨θ, hθxi, hφeq⟩, hQ⟩
rw [induceHU_eq_induce data] at hφeq
subst hφeq
rw [Finset.mem_coe, Finset.mem_image]
have hθT : θ ∈ T := (hT_mem θ).mpr hθxi
exact ⟨θ, (hQmem θ hθT).mpr hQ, rfl⟩
-- ── source degrees from member degrees (`q` cancels)
have hdeg_of_member : ∀ (θ : IrreducibleCharacter ↥(huSub data)) (d : ℕ),
ClassFunction.induce (huSub data) θ.toClassFunction 1 = ((data.q * d : ℕ) : ℂ) →
(θ : ClassFunction ↥(huSub data) ℂ) 1 = (d : ℂ) := by
intro θ d hmem
rw [← induceHU_eq_induce data, induceHU_apply_one_eq_q_mul] at hmem
apply mul_left_cancel₀ hq0
rw [hmem]
push_cast
ring
-- ── the three parts of `T`
have hred_of : ∀ θ ∈ T, ¬ IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction) →
(θ : ClassFunction ↥(huSub data) ℂ) 1 = (chars.u : ℂ) := by
intro θ hθ hred
refine hdeg_of_member θ chars.u ?_
exact caseA_reducible_induceHU_apply_one_eq_qu caseA hG _
(sOf_antitone data le_sup_left (hmem_sOf θ hθ)) hred
have hi4_of : ∀ θ ∈ T, ClassFunction.induce (huSub data) θ.toClassFunction ∉ S₂ →
(θ : ClassFunction ↥(huSub data) ℂ) 1 = (chars.u : ℂ) := by
intro θ hθ hnot
exact hdeg_of_member θ chars.u (hS3deg _ (hmem_sOf θ hθ) hnot)
have hi2_of : ∀ (θ : IrreducibleCharacter ↥(huSub data)),
ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂ →
(θ : ClassFunction ↥(huSub data) ℂ) 1 = (caseA.a : ℂ) := by
intro θ hin
exact hdeg_of_member θ caseA.a (hS2deg _ hin)
-- ── the class equation, folded to `T`
have hsum := sum_xiOf_H0C_degreeSq data chief chars
rw [← hT_def] at hsum
-- split `T` by member irreducibility, then the irreducible part by `𝒮₂`-membership
have hA := Finset.sum_filter_add_sum_filter_not T
(fun θ => IsIrreducibleCharacter (ClassFunction.induce (huSub data) θ.toClassFunction))
(fun χ => ((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2)
have hB := Finset.sum_filter_add_sum_filter_not
(T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction)))
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂)
(fun χ => ((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2)
-- ── part `𝒳red`: the `p − 1` sources of the reducible members, degree `u`, fibre `1`
have hcard_red : (T.filter (fun θ => ¬ IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).card = chief.p - 1 := by
have h1 := hfib_card (T.filter (fun θ => ¬ IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))) 1
(Finset.filter_subset _ _)
(fun θ θ₀ heq hθT hθ₀ => by
rw [Finset.mem_filter] at hθ₀ ⊢
exact ⟨hθT, by rw [heq]; exact hθ₀.2⟩)
(fun θ₀ hθ₀ => inertia_index_eq_one_of_induce_reducible data θ₀
(Finset.mem_filter.mp hθ₀).2)
rw [mul_one] at h1
rw [h1, ← Set.ncard_coe_finset,
himg _ (fun φ => ¬ IsIrreducibleCharacter φ) (Finset.filter_subset _ _)
(fun θ hθ => by rw [Finset.mem_filter]; exact ⟨fun h => h.2, fun h => ⟨hθ, h⟩⟩)]
have hcnt := reducible_count_sOf_H0C hG chars
rw [show chars.C = cSub data chief from rfl] at hcnt
exact hcnt
have hvred : ∑ χ ∈ T.filter (fun θ => ¬ IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction)),
((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2
= ((chief.p - 1 : ℕ) : ℂ) * ((chars.u : ℕ) : ℂ) ^ 2 := by
rw [show ∑ χ ∈ T.filter (fun θ => ¬ IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction)),
((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2
= ∑ _χ ∈ T.filter (fun θ => ¬ IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction)),
((chars.u : ℕ) : ℂ) ^ 2 from
Finset.sum_congr rfl fun θ hθ => by
rw [Finset.mem_filter] at hθ
rw [hred_of θ hθ.1 hθ.2]]
rw [Finset.sum_const, nsmul_eq_mul, hcard_red]
-- ── part `𝒳₂`: sources of the irreducible `𝒮₂`-members, degree `a`, fibre `q`
have hcard_i2 : ((T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂)).card
= {φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∈ S₂}.ncard * data.q := by
have hsub : (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂) ⊆ T :=
fun θ hθ => Finset.filter_subset _ _ ((Finset.filter_subset _ _) hθ)
have hmem' : ∀ θ ∈ T, (θ ∈ (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂)
↔ (IsIrreducibleCharacter (ClassFunction.induce (huSub data) θ.toClassFunction) ∧
ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂)) := by
intro θ hθ
rw [Finset.mem_filter, Finset.mem_filter]
exact ⟨fun h => ⟨h.1.2, h.2⟩, fun h => ⟨⟨hθ, h.1⟩, h.2⟩⟩
have h1 := hfib_card _ data.q hsub
(fun θ θ₀ heq hθT hθ₀ => by
rw [hmem' θ hθT]
have h₀ := (hmem' θ₀ (hsub hθ₀)).mp hθ₀
rw [heq]
exact h₀)
(fun θ₀ hθ₀ => inertia_index_eq_q_of_induce_irreducible data θ₀
((hmem' θ₀ (hsub hθ₀)).mp hθ₀).1)
rw [h1, ← Set.ncard_coe_finset,
himg _ (fun φ => IsIrreducibleCharacter φ ∧ φ ∈ S₂) hsub hmem']
have hvi2 : ∑ χ ∈ (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂),
((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2
= (({φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∈ S₂}.ncard * data.q : ℕ) : ℂ)
* ((caseA.a : ℕ) : ℂ) ^ 2 := by
rw [show ∑ χ ∈ (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂),
((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2
= ∑ _χ ∈ (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂),
((caseA.a : ℕ) : ℂ) ^ 2 from
Finset.sum_congr rfl fun θ hθ => by
rw [Finset.mem_filter] at hθ
rw [hi2_of θ hθ.2]]
rw [Finset.sum_const, nsmul_eq_mul, hcard_i2]
-- ── part `𝒳₄`: sources of the `𝒮₄`-members, degree `u`, fibre `q`
have hcard_i4 : ((T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ¬ ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂)).card
= {φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∉ S₂}.ncard * data.q := by
have hsub : (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ¬ ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂) ⊆ T :=
fun θ hθ => Finset.filter_subset _ _ ((Finset.filter_subset _ _) hθ)
have hmem' : ∀ θ ∈ T, (θ ∈ (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ¬ ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂)
↔ (IsIrreducibleCharacter (ClassFunction.induce (huSub data) θ.toClassFunction) ∧
ClassFunction.induce (huSub data) θ.toClassFunction ∉ S₂)) := by
intro θ hθ
rw [Finset.mem_filter, Finset.mem_filter]
exact ⟨fun h => ⟨h.1.2, h.2⟩, fun h => ⟨⟨hθ, h.1⟩, h.2⟩⟩
have h1 := hfib_card _ data.q hsub
(fun θ θ₀ heq hθT hθ₀ => by
rw [hmem' θ hθT]
have h₀ := (hmem' θ₀ (hsub hθ₀)).mp hθ₀
rw [heq]
exact h₀)
(fun θ₀ hθ₀ => inertia_index_eq_q_of_induce_irreducible data θ₀
((hmem' θ₀ (hsub hθ₀)).mp hθ₀).1)
rw [h1, ← Set.ncard_coe_finset,
himg _ (fun φ => IsIrreducibleCharacter φ ∧ φ ∉ S₂) hsub hmem']
have hvi4 : ∑ χ ∈ (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ¬ ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂),
((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2
= (({φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∉ S₂}.ncard * data.q : ℕ) : ℂ)
* ((chars.u : ℕ) : ℂ) ^ 2 := by
rw [show ∑ χ ∈ (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ¬ ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂),
((χ : ClassFunction ↥(huSub data) ℂ) 1) ^ 2
= ∑ _χ ∈ (T.filter (fun θ => IsIrreducibleCharacter
(ClassFunction.induce (huSub data) θ.toClassFunction))).filter
(fun θ => ¬ ClassFunction.induce (huSub data) θ.toClassFunction ∈ S₂),
((chars.u : ℕ) : ℂ) ^ 2 from
Finset.sum_congr rfl fun θ hθ => by
rw [Finset.mem_filter] at hθ
rw [hi4_of θ (Finset.filter_subset _ _ hθ.1) hθ.2]]
rw [Finset.sum_const, nsmul_eq_mul, hcard_i4]
-- ── the (9.8.d) count equality at `C = U′`: `|𝒮₁′|·a² = (p−1)·u`
have hcount' : {φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∈ S₂}.ncard * (caseA.a * caseA.a)
= (chief.p - 1) * chars.u := by
have hfam : sOf data (chief.H0 ⊔ cSub data chief)
= sOf data (chief.H0 ⊔ uprimeSub data) := by rw [hCU]
have hset : {φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∈ S₂}
= {χ ∈ sOf data (chief.H0 ⊔ uprimeSub data) |
IsIrreducibleCharacter χ ∧ χ 1 = ((data.q * caseA.a : ℕ) : ℂ)} := by
ext φ
constructor
· rintro ⟨hs, hirr, hS⟩
exact ⟨hfam ▸ hs, hirr, hS2deg φ hS⟩
· rintro ⟨hs, hirr, hdeg⟩
exact ⟨hfam.symm ▸ hs, hirr, hS₁'sub ⟨hs, hirr, hdeg⟩⟩
rw [hset, hcount, ← hCU, relIndex_cSub_U_eq_u chars]
-- ── assemble in `ℂ`, then descend to `ℕ`
have hcC : (({φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∈ S₂}.ncard * (caseA.a * caseA.a) : ℕ) : ℂ)
= (((chief.p - 1) * chars.u : ℕ) : ℂ) := by
exact_mod_cast congrArg (Nat.cast (R := ℂ)) hcount'
have hgoalC : ((chars.u + ({φ ∈ sOf data (chief.H0 ⊔ cSub data chief) |
IsIrreducibleCharacter φ ∧ φ ∉ S₂}.ncard * data.q + (chief.p - 1)) * chars.u ^ 2
+ data.q * (chief.p - 1) * chars.u : ℕ) : ℂ)
= ((chief.p ^ data.q * chars.u : ℕ) : ℂ) := by
push_cast at hsum hA hB hvred hvi2 hvi4 hcC ⊢
linear_combination hsum + hA + hB - hvi2 - hvi4 - hvred - (data.q : ℂ) * hcC
exact Nat.cast_inj.mp hgoalC
end NineElevenThreeOrbitSplit
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_NineElevenCoherence.lean