Let $G$ be a finitely generated, residually finite group and $\widehat{G}$ its profinite completion. Must the natural image of $G$ inside $\widehat{G}$ be a normal subgroup of $\widehat{G}$? Must it be characteristic?


No. Suppose $G$ is residually finite and is normal in $\hat{G}$.

Any element of $\hat{G}$ that commutes with all elements of $G$ commutes with all elements of $\hat{G}$, the closure of $G$. So $C_{\hat{G}}(G)=Z(\hat{G})$. Also an element $g\in G$ is in the centre of $\hat{G}$ if and only if it's in the centre of $G$, so we have injective homomorphisms $$G/Z(G)\hookrightarrow\hat{G}/Z(\hat{G})\hookrightarrow\operatorname{Aut}(G),$$ the second one induced by the action of $\hat{G}$ on $G$ by conjugation.

If $G$ is finitely generated then $\operatorname{Aut}(G)$ is countable. But $\hat{G}/Z(\hat{G})$ is a profinite group and so must be finite or uncountable.

So if $G$ is a finitely generated residually finite group with $|G:Z(G)|$ infinite (for example, a free group on two generators) then $G$ is never normal in $\hat{G}$.

  • $\begingroup$ If $G/Z(G)$ is infinite, how do we know that $[\widehat{G}:Z(\widehat{G})] = [G:Z(G)]$? Is that a standard fact? $\endgroup$ – oxeimon Apr 21 '15 at 23:41
  • $\begingroup$ @oxeimon I've edited my answer to add some details. $\endgroup$ – Jeremy Rickard Apr 22 '15 at 11:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.