The isomorphism problem for profinite completions of finitely presented, residually finite groups. (Q473140)

A Grothendieck pair is a monomorphism \(u\colon P\to\Gamma\) of finitely presented, residually finite groups such that the associated map of profinite completions \(\widehat u\colon\widehat P\to\widehat\Gamma\) is an isomorphism but \(P\) is not isomorphic to \(\Gamma\). The existence of Grothendieck pairs was established in a joint paper by \textit{F. J. Grunewald} and the first author [Ann. Math. (2) 160, No. 1, 359-373 (2004; Zbl 1083.20023)], raising the following recognition problem: is there an algorithm that, given a monomorphism of finitely presented, residually finite groups \(u\colon P\to\Gamma\), can determine whether or not \(\widehat u\) is an isomorphism? The authors show that the answer is negative (nor do there exist algorithms that can decide whether \(\widehat u_n\) is surjective, or whether \(\widehat P\) is isomorphic to \(\widehat\Gamma\)) by proving the following theorem. There are recursive sequences of finite presentations for residually finite groups \(P_n\) and \(\Gamma_n\) together with explicit monomorphisms \(u_n\colon P_n\to\Gamma_n\), such that: \(\widehat P_n\cong\widehat\Gamma_n\) if and only if the induced map of profinite completions \(\widehat u_n\) is an isomorphism; \(\widehat u_n\) is an isomorphism if and only if \(\widehat u_n\) is surjective; and the set \(\{n\in\mathbb N\mid\widehat P_n\not\cong\widehat\Gamma_n\}\) is recursively enumerable but not recursive.