Twisted conjugacy classes of the unit element. (Q359322)

Given a group \(G\) and an endomorphism \(\varphi\colon G\to G\) let us consider the naive problem: \textit{when the subset of \(G\) given by \(\{g^{-1}\varphi(g)\mid g\in G\}\), denoted by \([e]_\varphi\), is a group?} The main stream of the paper under review is to explore this problem and related questions. The study of the question above fits well in the framework of the study of twisted conjugacy classes or Reidemeister classes. The authors consider mainly symmetric groups, nilpotent groups and solvable groups. To give a sample of the results, they show: Corollary 5. For every IA-automorphism \(\varphi\) of the free \(2\)-step nilpotent group \(N_{n,2}\) the class \([e]_\varphi\) is a subgroup. and Proposition 10. The group \(A_n\) for \(n\geq 4\) contains an element \(h\) for which the class \([e]_h\) is not a subgroup. Here \([e]_h\) means the twisted class which contains the identity with respect to the inner automorphism defined by \(h\). Several examples are provided to illustrate various aspects and as well as several questions. In particular they state the following conjecture: If in some group \(G\) the twisted \(\varphi\)-conjugacy class \([e]_\varphi\) of the unit element is a subgroup for every automorphism \(\varphi\in\Aut G\) then \(G\) is a nilpotent group. Furthermore, if \(G\) is finitely generated then \(G\) is Abelian. The tools used are mainly from combinatorial group theory.