Factorization of point configurations, cyclic covers and conformal blocks

TL;DR

This paper explores the relationship between point configurations in projective space, cyclic covers, and conformal blocks, providing a geometric framework that unifies and extends recent results in the theory of line bundles on moduli spaces.

Contribution

It introduces a new geometric approach linking GIT invariants, cyclic covers, and conformal blocks, and defines divisorial factorization to classify associated line bundles.

Findings

GIT polarizations and Hodge classes induce line bundles on _{0,n} with boundary restrictions.

A divisorial factorization property uniquely determines these line bundles.

Provides a unified proof of recent results on conformal block bundles.

Abstract

We describe a relation between the invariants of $n$ ordered points in $P^d$ and of points contained in a union of linear subspaces $P^{d1}\cup P^{d2} \subset P^d$. This yields an attaching map for GIT quotients parameterizing point configurations in these spaces, and we show that it respects the Segre product of the natural GIT polarizations. Associated to a configuration supported on a rational normal curve is a cyclic cover, and we show that if the branch points are weighted by the GIT linearization and the rational normal curve degenerates, then the admissible covers limit is a cyclic cover with weights as in this attaching map. We find that both GIT polarizations and the Hodge class for families of cyclic covers yield line bundles on $\bar{M}_{0,n}$ with functorial restriction to the boundary. We introduce a notion of divisorial factorization, abstracting an axiom from rational conformal field theory, to encode this property and show that it determines the isomorphism class of these line bundles. As an application, we obtain a unified, geometric proof of two recent results on conformal block bundles, one by Fedorchuk and one by Gibney and the second author.