Commensurating actions of birational groups and groups of pseudo-automorphisms (Q2326242)

Pseudo-automorphisms on a variety \(X\) are birational transformations acting as regular automorphisms in codimension 1. They form a group \(\mathrm{PsAut}(X)\) with \(\mathrm{Aut}(X)\subset\mathrm{PsAut}(X)\subset\mathrm{Bir}(X)\). For smooth projective surfaces, the first inclusion is an equality; for rational varieties of dimension at least 3, both inclusions are strict; for Calabi-Yau manifolds, the second inclusion is an equality, and there are examples of such varieties such that \(\mathrm{PsAut}(X)\) is infinite while \(\mathrm{Aut}(X)\) is trivial [\textit{S. Cantat} and \textit{K. Oguiso}, Am. J. Math. 137, No. 4, 1013--1044 (2015; Zbl 1386.14147)]. A group \(G\subset \mathrm{Bir}(X)\) is (pseudo-)regularizable if it is conjugate via a birational map to a group of (pseudo-)automorphisms. In the present paper, the authors prove the following: Theorem 1. Let \(X\) be a projective variety over an algebraically closed field. Let \(G\) be a subgroup of \(\mathrm{Bir}(X)\). If \(G\) has Property (FW), then \(G\) is pseudo-regularizable. Property (FW) is a fixed point property for groups acting on CAT(0) cubical complexes. Equivalently, a group satisfies Property (FW) if all commensurated subsets are transfixed for any set on which the group acts. This property is satisfied for example by discrete countable groups with Kazhdan Property (T). As the title of the article suggests, the authors look at Property (FW) from the point of view of commensurated subsets. Furthermore, they apply Theorem 1 to classify infinite groups of birational transformations of (smooth, projective, irreducible) surfaces satisfying Property (FW): Any such group is birationally conjugate to a subgroup of the automorphism group of the projective plane, of a Hirzebruch surface \(\mathbb{F}_m\) with \(m\geq1\), or of a product of a curve with \(\mathbb{P}^1\). Moreover, in positive characteristic only the case of the projective plane can occur. This is Theorem 2. The paper ends with four open questions. Question 10.1 asks whether every birational action of \(G\) is regularizable if \(G\) satisfies Property (FW), and it was affirmatively settled by the second named author [\textit{Y. Cornulier}, Confluentes Math. 12, No. 2, 3--10 (2021; Zbl 1460.14034)], and also by [\textit{C. Urech} and \textit{A. Lonjou}, ``Actions of Cremona groups on CAT(O) cube complexes'', Duke Math. J. (to appear)]. The other three questions still seem to be open.