Proper actions on finite products of quasi-trees (Q2077171)

Say that a group \(G\) has property (QT) if it acts isometrically on a finite product of quasi-trees such that orbit maps are quasi-isometric embeddings. Here, a \textit{quasi-tree} is a connected graph whose path metric is quasi-isometric to a tree and the product is given the \(\ell^1\)-metric. In this paper, the authors prove that residually finite Gromov hyperbolic groups have (QT), as do mapping class groups of finite-type surfaces. To build intuition, observe that it is clear that free groups have (QT), and by choosing a filling collection of simple closed curves on a surface and taking the Bass-Serre trees dual to that collection, it is not hard to see that surface groups have (QT); in each of these cases there is in fact a proper action on a finite product of \textit{trees,} a property shared by (subgroups of) Coxeter groups by a result of \textit{A. Dranishnikov} and \textit{T. Januszkiewicz} [Topol. Proc. 24(Spring), 135--141 (1999; Zbl 0973.20029). It follows that \emph{virtually cocompact special} cubulated groups have (QT), a class which includes fundamental groups of closed hyperbolic 3-manifolds. These latter groups act without global fixed point on CAT(0) cube complexes, and thus have the \emph{Haagerup property,} a strong negation of Kazhdan's property (T). On the other hand, uniform lattices in \(\mathrm{Sp}(n,1)\) have Kazhdan's property (T) and are residually finite and Gromov hyperbolic. Therefore the authors' result shows that property (QT) is not strong enough to imply the Haagerup property. Indeed, it is unknown whether mapping class groups of finite-type surfaces have either the Haagerup property or Kazhdan's property (T). Since finite products of (quasi-)trees have finite asymptotic dimension, property (QT) is a stronger version of finiteness of the asymptotic dimension, a result due to Gromov for Gromov hyperbolic groups and to the authors [Publ. Math., Inst. Hautes Étud. Sci. 122, 1--64 (2015; Zbl 1372.20029)] for mapping class groups of finite-type surfaces. Indeed, the proof uses the projection complexes techniques developed in the authors' previous paper. The authors remark that lattices in higher-rank Lie groups have finite asymptotic dimension but cannot have (QT). Thus in this regard the mapping class group of a finite-type surface behaves more like a Gromov hyperbolic group than a (higher-rank) lattice.