Rank deficiency at a point divides the determinant of a polynomial matrix
ProvedLatticePrepro.pow_X_sub_C_dvd_detLet be a field, a finite index set, and a square matrix of polynomials. For , write for the matrix obtained by evaluating every entry at . Then the rank deficiency of bounds the multiplicity of as a root of :
where is the number of rows and the rank is taken over .
This is the divisibility step in the proof of Lemma A.6 (as in Fancsali–Sziklai). It is applied to a nonsingular square submatrix of the polynomial matrix of an evaluation table; since the factors for distinct are coprime, the total rank deficiency over all sample points is at most the degree of the determinant, which yields the weight bound of Lemma A.6.
Formalization Note: is M.map (eval s) and is X - C s. The exponent uses natural-number subtraction, which never truncates because .
import Mathlib
open Polynomial Matrix
variable {𝕊 : Type*} [Field 𝕊] {n : Type*} [Fintype n] [DecidableEq n]theorem LatticePrepro.pow_X_sub_C_dvd_det (M : Matrix n n 𝕊[X]) (s : 𝕊) :
(X - C s) ^ (Fintype.card n - (M.map (eval s)).rank) ∣ M.det := by sorry