2
$\begingroup$

Let $K$ be a field. A polynomial $F \in \mathbb{Q}[X_1, \dots, X_r]$ which is weighted homogeneous of degree $n$ with respect to the grading $\deg(X_k) = k$ is called numerically non-negative for nef vector bundles if for every $n$-dimensional projective $K$-variety $X$ and every nef vector bundle $E$ of rank $r$ on $X$ we have $\int_X F(c_1(E), \dots, c_r(E)) \ge 0$. According to Example 8.3.10 in Lazarsfeld's book Positivity in Algebraic Geometry II a polynomial is numerically non-negative for nef vector bundles if and only if it is a non-negative linear combination of the Schur polynomials. However Lazarsfeld only works over the complex ground field.

Question: Is there a citeable reference where the fact stated above is stated (and proved) over an arbitrary ground field?

Note that the paper Positive Polynomials for Ample Vector Bundles by Fulton and Lazarsfeld allows an arbitrary ground field, but there only positive polyomials are considered which have a similar description.

$\endgroup$
3
  • $\begingroup$ I don't think the proof depends much on the ground field. See Fulton's followup paper, "Positive polynomials for filtered ample vector bundles" (Amer J Math, 1995), which takes the setting to be over a general field, and also Demailly-Peternell-Schneider (J. Alg. Geom, 1994), which proves the result you want over C. $\endgroup$ Commented Nov 14, 2023 at 11:56
  • $\begingroup$ @DaveAnderson Thank you for your comment! Unfortunately, also in Fulton's paper about filtered vector bundles, at least the statement of Theorem' about nef vector bundles refers to the complex ground field (it refers to compact Kähler manifolds). Maybe the proof (in the case of a projective variety) works over an arbitrary ground field, I didn't have time to properly check this yet. $\endgroup$
    – user513306
    Commented Nov 14, 2023 at 14:06
  • $\begingroup$ But even if the proof is alright, it would be nice, to just have the statement over an arbitrary ground field somewhere… $\endgroup$
    – user513306
    Commented Nov 14, 2023 at 14:06

0

You must log in to answer this question.