In solving some exercises in Sageev's notes on $CAT(0)$ cube complexes I've needed to use the following fact that I would like to know how to prove.

A group $G$ is residually finite if and only if the profinite topology on $G$ is Hausdorff.

Here the working definition of residually finite is that the trivial subgroup is separable; where a subgroup $H < G$ is separable if for all $g \in G-H$ there is a finite index $K < G$ with $H<K$ and $g \not\in K$.

The profinite topology is generated by the basis of cosets of finite index subgroups of $G$. It can also be shown that you could instead use cosets of finite index normal subgroups of $G$.

Any hints are appreciated, and I actually only need to understand the forward direction.

  • $\begingroup$ I guess that if you take classical definition of residual finiteness (every $g \in G$ is nontrivial in some finite quotient of $G$) and observe that canonical map $G \to \hat G$ is mono iff $G$ residually finite, you can easily deduce your statement. $\endgroup$ – xsnl Apr 6 '17 at 14:18

Hint: if $x,y \in G$ with $x \neq y$, then $xy^{-1} \neq 1$, hence there is an open normal subgroup $N$ (of finite index) with $xy^ {-1} \notin N$. But then $xN \cap yN = \emptyset$.


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.