Weighted dimensions of generalized eigenspaces from a factored characteristic polynomial
ProvedWeierstrassEllipticZeta.weighted_generalized_eigenspace_dimensioncommutative-algebraelliptic-functionstranscendencezero-estimates
Let be a field, , and a finite type. Let be an matrix over , let , and let . Suppose
For , write
Then
and the same formula holds for whenever . The values need not be distinct; every index in the fibre contributes its weight. Zero weights, empty index types and are allowed. No characteristic assumption on is required.
Preamble
import Mathlib.LinearAlgebra.Eigenspace.Zero import Mathlib.Algebra.Polynomial.Roots open scoped Classical
Formal statement
theorem WeierstrassEllipticZeta.weighted_generalized_eigenspace_dimension
(K : Type*) [Field K] (n : ℕ) (V : Type*) [Fintype V]
(A : Matrix (Fin n) (Fin n) K) (w : V → ℕ) (f : V → K)
(hchar : A.charpoly = ∏ v : V, (Polynomial.X - Polynomial.C (f v)) ^ w v) :
∀ z : K,
Module.finrank K (Module.End.maxGenEigenspace A.mulVecLin z) =
∑ v ∈ Finset.univ.filter (fun v : V => f v = z), w v ∧
∀ N : ℕ, n ≤ N →
Module.finrank K (Module.End.genEigenspace A.mulVecLin z N) =
∑ v ∈ Finset.univ.filter (fun v : V => f v = z), w v := by sorry
Source
Derived linear-algebra lemma for the approach associated with Senthil Kumar K (2026), Appendix A and Theorem A.2, https://doi.org/10.1017/S001309152610145X. This lemma is proved here and is not quoted from the article. If a matrix characteristic polynomial factors as a product of powers of X-f(v), its maximal generalized eigenspace at z has dimension equal to the sum of weights over f(v)=z. The same dimension holds at every generalized-eigenspace order at least the matrix size. Repeated values, zero weights and empty index types are allowed. Uses Mathlib root multiplicities and generalized-eigenspace stabilization. No Prove2Me theorem dependencies or new definitions.