Normalizers in groups and in their profinite completions. (Q2449576)

In the paper under review the authors are concerned with questions of conjugacy separability, such as the following one. Given a group \(R\) that is an extension of a free group by a finite soluble one, is it true that two elements of \(R\) are conjugate in \(R\), provided their images are conjugate in each finite quotient of \(R\)? To answer these questions, one is lead to investigate the following situation. Let \(R\) be a residually finite group, which thus embeds in its profinite completion \(\widehat R\). Let \(H\) be a finitely generated subgroup of \(R\). Is it true that the normalizer \(N_{\widehat R}(\overline H)\) of the closure of \(H\) in \(\widehat R\) coincides with the closure of \(N_R(H)\)? A positive answer to this question for the case when \(R\) is a finite extension of a polycyclic group was given by \textit{L. Ribes} et al. [J. Lond. Math. Soc., II. Ser. 57, No. 3, 609-628 (1998; Zbl 0922.20031)], using number theoretic arguments. In the paper under review the authors consider the case when \(R\) is the extension of a free group by a group in an extension-closed variety \(\mathcal C\) of finite groups. Examples of such varieties comprise the class of all finite groups, the class of all finite soluble groups, and the class of all finite \(p\)-groups, for a given prime \(p\). The profinite completion is replaced here by the pro-\(\mathcal C\) completion. The authors obtain a positive answer for the normalizer problem in this context, and derive the corresponding conjugacy separability results, as in the case of polycyclic groups. The methods, however, are based here on the analogue for profinite groups of the Bass-Serre theory of groups acting on trees, and are of independent interest.