On cohomology rings of infinite groups. (Q860442)

Let \(R\) be any ring, \(G\) a group and \(RG\) the corresponding group ring. Moreover let \(\text{Ext}^*_{RG}(M,M)\) be the cohomology ring (with Yoneda product) associated with the \(RG\)-module \(M\) and let \(H\) be a subgroup of finite index in \(G\). The main theorem states that if \(H\) and \(G\) satisfy suitable conditions (in particular, if the profinite completion of \(G\) is torsion free), then an element \(\zeta\in\text{Ext}^*_{RG}(M,M)\) is nilpotent if and only if its restriction to \(\text{Ext}^*_{RH}(M,M)\) is nilpotent. An interesting example of a group which satisfies the hypotheses of this theorem is the Thompson group. If one drops the assumption on the profinite completion of \(G\), then the analogue of the previous theorem is false in general. An important tool in the proof is the crossed product construction.