Second bounded cohomology and WWPD (Q824276): Difference between revisions

The weak proper discontinuity property (WPD) was introduced by \textit{M. Bestvina} and \textit{K. Fujiwara} [Geom. Topol. 6, 69--89 (2002; Zbl 1021.57001)] in the context of a group with a hyperbolic action. Qualitatively, a loxodromic element \(h\) is WPD if only finitely many elements approximately fix a long segment of the axis of \(h\). The original motivation was the study of mapping class groups of a closed surfaces. Subsequently, Bestvina, Bromberg, and Fujiwara introduced the notion of WWPD, a yet weaker, and more technical, version of WPD [\textit{M. Bestvina} et al., Ann. Inst. Fourier 66, No. 3, 871--898 (2016; Zbl 1397.20050)]. Qualitatively, a loxodromic element is WWPD if in the definition of WPD we replace ``finitely many elements'' with ``finitely many left cosets of the stabilizer of the limit points of \(h\)''. We remark that this paper itself is a great place to learn why this definition is useful and what it is related to. Both weak and weaker proper discontinuity give control over the bounded cohomology, quasimorphisms, and stable commutator length in the group. Generally, if a group has a hyperbolic action and some (W)WPD-ness, its second bounded cohomology has uncountable dimension. This paper grew out of an effort to study the second bounded cohomology of the group \(\mathrm{Out}(F)\) of outer automorphisms of a free group \(F\). As part of that process, the authors organized and clarified several notions of WWPD and its relationship to WPD. This paper is the extraction of that organization as a self-contained unit dealing only with (W)WPD. And it is quite a useful extraction! The authors first prove the equivalence of 4 (really 5) different notions of WWPD, and along the way provide a number of useful lemmas and corollaries. They then weaken the ``global WPD hypothesis'' of Bestvina-Fujiwara to a ``global WWPD hypothesis'', and they prove that if a group \(\Gamma\) has a finite index normal subgroup \(N\) with a hyperbolic action satisfying the global WWPD hypothesis, then the second bounded cohomology of \(\Gamma\) has uncountable dimension. In the special case that \(\Gamma=N\), the full global WWPD hypothesis is not needed: if the group has just a single WWPD element, then the same conclusion follows. This paper is very well written and contains an extremely useful collection of results about WPD and WWPD. Anyone interested in geometric group theory should read it.