If $G$ and $H$ are finitely generated residually finite groups such that $G\leq H\leq \hat G$, where $\hat G$ denotes the profinite completion of $G$, does it follow that $$\hat H \cong \hat G$$



By the universal property of profinite completions, we have to show that every homomorphism of groups $G \to K$, where $K$ is a profinite group, extends uniquely to a homomorphism of groups $H \to K$.

Well, we have only one choice: $G \to K$ corresponds to a homomorphism of profinite groups $\hat{G} \to K$, and this can be restricted to $H \leq \hat{G}$. This is clearly an extension of $G \to K$.

Now, assume that two homomorphism $f,g : H \to K$ agree on $G$. Then $\hat{f},\hat{g} : \hat{H} \to K$ are homomorphisms of profinite groups which agree on $\hat{G}$. Then they also agree on $H \leq \hat{G}$, i.e. $f=g$.

  • $\begingroup$ How is $\hat H$ inside $\hat G$? $\endgroup$ – Zeppelin Mar 19 '13 at 13:45
  • $\begingroup$ You need the inclusion map from $G$ to $\hat H$ be continuous? $\endgroup$ – Zeppelin Mar 19 '13 at 13:52
  • $\begingroup$ I am sorry I do not understand your proof. $\endgroup$ – Zeppelin Mar 19 '13 at 13:57
  • $\begingroup$ $G$ is discrete. And I've corrected the proof. Is it OK now? $\endgroup$ – Martin Brandenburg Mar 19 '13 at 14:20
  • $\begingroup$ G is not discrete, it has the profinite topology. By definition of profinite completion, you need the inclusion map from $G$ to $\hat H$ to be continuous. But I do not think it is the case always. $\endgroup$ – Zeppelin Mar 19 '13 at 14:32

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.