The Tarski numbers of groups (Q499286): Difference between revisions

A group \(G\) admits a paradoxical decomposition if there exist disjoint subsets \(P_1,\dots, P_m, Q_1,\dots,Q_n\) of \(G\) and elements \(g_1,\dots,g_m, h_1,\dots h_n\) of \(G\) such that \(G=\bigcup P_ig_i=\bigcup Q_jh_j\). The non-amenability of the group \(G\) is equivalent to the existence of paradoxical decompositions. The minimal possible value \(m+n\) in a paradoxical decomposition of \(G\) is the Tarski number of this group and denoted by \(\tau(G)\). In the paper under review, the authors obtain brilliant results concerning the Tarski numbers of groups. Showing that \(\text{Fin}_m\), the class of all groups where all \(m\)-generated subgroups are finite, contains non-amenable groups for every \(m\geq 1\), they prove that there exist non-amenable groups with arbitrary large Tarski numbers. They also prove that the set of Tarski numbers of 2-generated non-amenable groups is infinite. This way, they show that the minimal number of generators of a non-amenable group has nothing to do with its Tarski number. If \(H\) is a non-amenable group which is either a subgroup or a quotient of a group \(G\), it is well-known that \(\tau(G)\leq \tau(H)\). In this paper, among other results of this type, a lower bound for \(\tau(H)\) in terms of \(\tau(G)\) is also found where \(H\) has finite index in \(G\). Indeed, \(\tau(H)-2\leq [G:H](\tau(G)-2)\). Finally, using the relation between the Caylay graph of \(G\) and the notion of cost of group action, the authors provide the first examples of groups with Tarski numbers 5 and 6. Regardless of the groups with free subgroups these are the first examples of non-amenable groups whose Tarski numbers are precisely computed.