An odd power distinguishes the two directed Cayley traces
ProvedCouret.CayleyDoublet.abelianThree_odd_trace_witness_of_odd_quotientcouret-programmecouret-unification-20261108
Let be a finite abelian group and . Suppose , , the image of has additive order exactly three in , the subgroup generated by is all of , and is odd. Let be the rational directed Cayley adjacency matrices for supports and . Then
This gives an odd positive discriminating degree; it does not assert that the degree is the first discrepancy. Directed arcs and any loops retain their original meaning.
Preamble
import Definitions.Def_Couret_CayleyTwoColourSpectral_cff7591a9c5d
import Mathlib
import Theorems.Thm_Couret_CayleyDoublet_abelianThree_odd_quotient_return
import Theorems.Thm_Couret_CayleyDoublet_abelianThree_odd_trace_ne_iff
import Definitions.Def_Couret_CayleyTwoColourSpectral_cff7591a9c5d
/-!
COU-20 — research branch only, spectral necessity work in progress.
This file is deliberately separate from the certified Core/Family/Abelian/
Spectral/Parity modules. It does not claim a new Lean or GitHub Actions
certificate until it has been compiled against the pinned toolchain.
The cyclic lemma below produces an odd *discriminating* exponent. It is
slightly more general than the canonical candidate |Q|/3: its exponent
need not be the least such one. Directed adjacency and loops are untouched.
The generic rational charpoly-to-all-power-traces theorem is NOT assumed as
proved. The final assembly is explicitly conditional on that missing bridge.
-/
set_option autoImplicit false
noncomputable section
open Matrix
namespace Couret.CayleyNecessity
end Couret.CayleyNecessity
namespace Couret.CayleyDoublet
open Couret.CayleyNecessity
variable {G : Type*} [AddCommGroup G] [DecidableEq G] [Fintype G]
Formal statement
theorem abelianThree_odd_trace_witness_of_odd_quotient (a d h : G)
(hhalf : h + h = 0) (hne : h ≠ 0)
(horder : addOrderOf (d : G ⧸ AddSubgroup.zmultiples h) = 3)
(hcyc : AddSubgroup.closure ({a, h} : Set G) = ⊤)
(hodd : Odd (Nat.card (G ⧸ AddSubgroup.zmultiples h))) :
∃ n : ℕ,
Matrix.trace (abelianCayleyMatrix (AbelianSource a d h) ^ (2 * n + 1)) ≠
Matrix.trace (abelianCayleyMatrix (AbelianTarget a d h) ^ (2 * n + 1)) := by sorry
end Couret.CayleyDoublet
Source
Alexandre Couret, https://github.com/alexcour/couret-finite-verifications/blob/65b828d9ffd016ce2a1b05403f303fa5f948c10f/research/correspondants_20261013/lean/CayleyTwoColourNecessity.lean#L105-L119, lines 105–119; source file SHA-256 7675ca2a1f08e8c6a20617d181810b5c2b6a4f0cffcf96b93678cb0bc9c7869d; theorem Couret.CayleyDoublet.abelianThree_odd_trace_witness_of_odd_quotient