Ubiquity of complete intersection liaison classes (Q887279)

Liaison theory branches into two important sub-theories: complete intersection (CI) liaison and Gorenstein liaison. This paper was inspired by a paper of the reviewer and \textit{U. Nagel} [Compos. Math. 149, No. 9, 1583--1591 (2013; Zbl 1284.13020)] which showed that any reduced ACM subscheme of \(\mathbb P^n\) is in the Gorenstein liaison class of a complete intersection (i.e. is glicci) provided one is allowed to link in the next higher projective space. This paper further clarifies how far this is from being true for the complete intersection liaison class of an ACM subscheme of \(\mathbb P^n\), and how much more complicated the picture is for CI liaison (in some sense). The authors study the join of two ideals in a regular local ring. When some of the individual ideals are themselves not licci, they show that distinct such join ideals must lie in distinct liaison classes. As a consequence, they show that there are at least as many CI-liaison classes of codimension \(c+3\) ACM subschemes in \(\mathbb P^{n+5}\) as there are generic complete intersection ACM subschemes of codimension \(c\) in \(\mathbb P^n\). The authors introduce an invariant, motivated by work of \textit{C. Polini} and \textit{B. Ulrich} [Math. Ann. 310, No. 4, 631--651 (1998; Zbl 0919.13013)], which they apply to their joins. As a consequence they obtain obstructions for join ideals to lie in the same liaison class, and they provide a simple way to construct new ideals that are maximal in their liaison class from old ones. They also show that if \(I\) is an unmixed ideal, and \(X\) and \(Y\) two new variables, then \((I,X)\) and \((I,Y)\) lie in the same liaison class if and only if \(I\) is licci.