Rewriting products of group elements. I (Q1104417)

A group G has Property \(Q_ n\) (n\(\geq 2)\) if for every choice of elements \(x_ 1,...,x_ n\) there exist distinct permutations \(\sigma\) and \(\tau\) on n letters such that \[ x_{\sigma (1)}...x_{\sigma (n)}=x_{\tau (1)}...x_{\tau (n)}. \] If \(\sigma\) can always be chosen to be the identity, then G is said to have property \(P_ n\) [see \textit{M. Curzio}, \textit{P. Langobardi}, \textit{M. Maj} and \textit{D. J. S. Robinson}, Arch. Math. 44, 385-389 (1985; Zbl 0544.20036)]. The following is proved: Theorem. A group has \(Q_ n\) for some n if and only if it is finite-by-abelian-by-finite. Clearly, \(P_ 2=Q_ 2=commutativity\) and \(P_ n\subseteq Q_ n\), the last inclusion being strict for \(n\geq 3\) (Proposition 2.10). By the analogue to the Theorem for \(P_ n\) (loc. cit.), every group G in \(Q_ n\) has \(P_ m\) for some \(m=f(n,G)\). The analogous statement is false for semigroups [see \textit{G. Pirillo}, in: Group Theory, Proc. Conf., Brixen/Italy 1986, Lect. Notes Math. 1281, 118-119 (1987; Zbl 0629.20028)]. It seems to be an open question whether \(Q_ n\subseteq P_{f(n)}\) for some function f (compare with Proposition 2.11).