A configuration space model for algebraic function spaces
Oishee Banerjee
TL;DR
The paper develops a configuration-space–type model for the moduli space $\mathrm{Mor}_{\mathbf{d}}(X,Y)$ of algebraic morphisms by constructing a natural compactification and a proper hypercover augmented over a discriminant locus. It leverages $\Delta S$-sheaves and cohomological descent to produce a first-quadrant spectral sequence computing $H_c^*(\mathrm{Mor}_{\mathbf{d}}(X,Y);\mathbf{Q})$, with $E_1$-terms given by a twisted configuration-space cohomology factor, the Picard variety, and the cohomology of auxiliary fibers $Y(D_p)$. In the special case $Y=\mathbb{P}^N$, the spectral sequence collapses at $E_2$ in a stability range $p\le r(\mathbf{d})+1$, informed by purity and Lefschetz hyperplane arguments, yielding concrete bounds on the growth of $\mathrm{Mor}_{\mathbf{d}}(X,Y)$. The approach connects algebro-geometric function spaces to topological configuration-space phenomena via a Koszul-type complex arising from descent, while preserving Galois/mixed Hodge structures on both sides and emphasizing intersection-theoretic data of $X$.
Abstract
We prove that the space of algebraic maps between two smooth projective varieties, under certain conditions, admit a configuration space model, thereby obtaining an algebro-geometric analogue of Bendersky-Gitler's result on topological function spaces. Our result is a natural higher dimensional counterpart of \cite[Theorem 3]{Ban24}.
