Cohomological characterization of relative hyperbolicity and combination theorem. (Q2271359)

\textit{S. M. Gersten}, [in the paper Topology 37, No. 5, 1031-1072 (1998; Zbl 0933.20026)], gives a cohomological characterization of the Gromov hyperbolicity which was up to then a purely geometrical notion. The purpose of the paper under review is to adopt and extend the above result to the setting of the strong relative hyperbolicity, i.e., relative hyperbolicity in Gromov's sense. The authors introduce some notion of the \(l_\infty\)-cohomology of a group \(G\) relative to a family of subgroups \(\mathcal H\). In order to define the relative cohomology of a pair \((G,\mathcal H)\), they follow the classical approach of \textit{R. Bieri} and \textit{B. Eckmann} [J. Pure Appl. Algebra 13, 277-319 (1978; Zbl 0392.20032)]. The main result is the following: The second relative \(l_\infty\)-cohomology of a group \(G\) relative to a family of subgroups \(\mathcal H\), denoted by \(H^2_{(\infty)}(G,\mathcal H)\), is well defined as soon as \(G\) admits a finite presentation relative to \(\mathcal H\). If \(G\) is strongly hyperbolic relative to \(\mathcal H\), then \(H^2_{(\infty)}(G,\mathcal H)\) strongly vanishes. Then the authors give an example which says that the above theorem is false in case of weak relative hyperbolicity, i.e., relative hyperbolicity in Farb's sense. In the paper are also given some applications in very specific situations of the class of groups which are semidirect products of relative hyperbolic groups with the free group. We want to mention that in another still unpublished paper I. Mineyev and A. Yaman are dealing with a similar subject.