On the tangent space to the Hilbert scheme of points in P3

TL;DR

This paper investigates the tangent space structure of the Hilbert scheme of points in P3, proving a key conjecture approximately, improving bounds, and providing counterexamples to related conjectures.

Contribution

It introduces a decomposition of the tangent space for monomial subschemes, proves a major conjecture up to a factor, and constructs counterexamples to existing conjectures.

Findings

Proved the Briançon-Iarrobino conjecture up to a 4/3 factor.

Improved the asymptotic bound on the dimension of Hilb^d P3.

Constructed infinitely many counterexamples to the second Briançon-Iarrobino conjecture.

Abstract

In this paper we study the tangent space to the Hilbert scheme $\mathrm{Hilb}^d \mathbf{P}^3$, motivated by Haiman's work on $\mathrm{Hilb}^d \mathbf{P}^2$ and by a long-standing conjecture of Brian\c{c}on and Iarrobino on the most singular point in $\mathrm{Hilb}^d \mathbf{P}^n$. For points parametrizing monomial subschemes, we consider a decomposition of the tangent space into six distinguished subspaces, and show that a fat point exhibits an extremal behavior in this respect. This decomposition is also used to characterize smooth monomial points on the Hilbert scheme. We prove the first Brian\c{c}on-Iarrobino conjecture up to a factor of 4/3, and improve the known asymptotic bound on the dimension of $\mathrm{Hilb}^d \mathbf{P}^3$. Furthermore, we construct infinitely many counterexamples to the second Brian\c{c}on-Iarrobino conjecture, and we also settle a weaker conjecture of Sturmfels in the negative.