Every finite complex has the homology of some \(\mathrm{CAT}(0)\) cubical duality group (Q2349862): Difference between revisions

The starting point of the paper under review is an old result by \textit{D. M. Kan} and \textit{W. P. Thurston} [Topology 15, 253--258 (1976; Zbl 0355.55004)], which associates, to any path-connected space \(X\), a discrete group \(G_X\) and a map from its classifying space \(BG\) to \(X\), inducing an isomorphism on integer homology. (In other words, the Kan-Thurston theorem says that any connected simplicial complex \(X\) has the same homology as the classifying space for some discrete group \(G_X\)). Several authors have generalized this result, among these \textit{J.-C. Hausmann} [Math. Ann. 275, 327--336 (1986; Zbl 0583.55004)], who proved that, whenever \(X\) is finite, the group \(G_X\) may be taken to be a duality group. More recently \textit{I. J. Leary} [J. Topol. 6, No. 1, 251--284 (2013; Zbl 1343.20044)] used metric geometry tools to show that the classifying space \(BG\) of the group \(G_X\) can be realized as a locally CAT(0) cubical complex. In the present paper the author proves that one can actually obtain simultaneously the last two results, showing for instance that every finite connected simplicial complex \(X\) has the same homology as the classifying space for some \(\mathrm{CAT}(0)\) cubical duality group \(G_X\). More precisely, to a given finite simplicial complex \(X\), he associates a finite locally \(\mathrm{CAT}(0)\) cubical complex \(T_X\) and a map \(t_X : T_X \to X\) such that \(\pi_1 (T_X)\) is a duality group and \(t_X\) induces an isomorphism on homology, for any local coefficients on \(X\). Furthermore, for a connected \(X\), the space \(T_X\) is just the classifying space for the group \(\pi_1 (T_X)\).