Hilbert curves of polarized varieties (Q2655018)

Let \(X\) be a complex projective irreducible variety, Pic\((X)\) its Picard group and Pic\(_0(X) \subset \)Pic\((X)\) the subgroup of topologically trivial line bundles. For any line bundle \(L\) on \(X\) one can consider the Hilbert polynomial \(\chi(L)\). This provides a polynomial function \(\chi\) from \(N(X)\) to the complex numbers, being \(N(X)\) the affine space of dimension \(\rho\) (the Picard number of \(X\)) defined as Pic\((X)/\)Pic\(_0(X)\) tensored with the complex field. Take a base of \(N(X)\), say \(\{L_1, \dots,L_\rho\}\), then \(\chi(x_1, \dots,x_\rho)=\chi(x_1 L_1+\dots+x_\rho L_\rho)=0\) defines an affine hypersurface \(V_X\), called the Hilbert variety associated to \(X\). Similarly one can choose an ample line bundle \(L\) on \(X\) and the canonical bundle \(K_X\) (\(X\) normal and Gorenstein) and define the Hilbert curve of \((X,L)\) by the equation \(\chi(x_1 L +x_2 K_X)=0\). By Serre duality it is possible to define an involution in \(N(X)\) fixing \(V_X\). The paper under review explores the behavior of the Hilbert variety and the Hilbert curve in several directions: with respect to the Serre involution; when \(\dim(X)=3\), so that the Hilbert curve is a cubic; when \(X\) admits some fibration relevant from the point of view of adjunction theory and so on. Further questions on the subject are posed (see for example Problem 6.6).