The p-adic topology on a free group: A counterexample (Q799799)

In [Math. Z. 184, 1-17 (1983; Zbl 0496.57006)] \textit{J. Stallings} made several conjectures related to the Kervaire conjecture. In the paper under review some of these conjectures are disproved. Let F be a finitely generated free group, and let p be a prime (or 0). Define \({}_ pF_ i\) inductively by \({}_ pF_ 1=F\), \(_ pF_{i+1}=sgp\{u^ p[v,w]:\quad u,v\in_ pF_ i,\quad w\in F\}.\) Taking these subgroups as a basis for 1 gives rise to the p-adic topology on F, and to the p-adic completion \({}_ p\hat F\) of F. Stallings made the following \(conjectures:\) \(A_ n(p):\) for any n elements \(u_ 1,u_ 2,...,u_ n\in F\) the set \(\{w:\quad w=\prod^{n}_{i=1}z_ i^{-1}u_ iz_ i\quad\exists z_ 1,z_ 2,...,z_ n\in F\}\) is closed in the p-adic topology. C(p): for any normal subgroup N of F the intersection of F and the normal closure of N in \({}_ p\hat F\) is equal to N. Stallings showed that if \(A_ n(p)\) holds for all n then C(p) holds. He also showed that \(A_ 1(p)\) holds. In the present paper the author gives an example showing that C(p) and \(A_ 2(p)\) (and hence \(A_ n(p)\) for \(n>2)\) are false. Stallings made several other conjectures in his paper. Related conjectures have been made by \textit{S. M. Gersten} [''Non-singular equations of small weight over groups'', Proc. Alta Conf. Theory Groups, Ann. Math. Stud. (to appear)] and \textit{J. Howie} [Math. Proc. Camb. Philos. Soc. 96, 255-268 (1984; Zbl 0542.20013)]. Recently, \textit{S. Gersten} [''Products of conjugacy classes in a free group: a counterexample'' (preprint)] has disproved these conjectures (except for the Kervaire conjecture itself).