Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have this assertion which looks rather easy (or as always I am missing something): We have $G$ topological group which is zero dimensional, i.e it admits a basis for a topology which consists of clopen sets, then every open nbhd that contains the identity element of G also contains a clopen subgroup.

I naively thought that if I take $\{e\}$, i.e the trivial subgroup, it's obviously closed, so it's also open in this topology, i.e clopen, and it's contained in every nbhd that contains $e$, but isn't it then too trivial.

Missing something right? :-)

share|cite|improve this question
In every topological group, an open subgroup is also closed. But if in a Hausdorff topological group every closed subgroup is open, then the subgroup $\{e\}$ is open and therefore all singleton sets are open and $G$ is discrete. Just because you have a base consisting of clopen sets doesn't mean that every closed set is open! So that explains your mistake...are you also asking for a correct proof? – Pete L. Clark Nov 25 '11 at 19:47
Regarding Mike's answer, you probably want your group to be locally compact as well as zero-dimensional. – Matt E Nov 25 '11 at 21:57

This is certainly false. Take $G=\mathbb{Q}$ with the standard topology and additive group structure. The topology is zero-dimensional since intersecting with $\mathbb{Q}$ the countably many open intervals whose endpoints are rational translates of $\sqrt 2$ gives a clopen basis for the topology on $G$. The trivial subgroup $\{0\}$ is certainly not open since the standard topology on $\mathbb{Q}$ is not discrete. So any clopen subgroup $H \subset G$ contains a nonzero rational number $q$ and so also $2q,3q,4q$ and so on. This shows $H$ is unbounded so, for example, the open neighbourhood $(-1,1) \cap \mathbb{Q}$ of $0 \in G$ contains no nontrival subgroup.

share|cite|improve this answer
Probably the group is supposed to be locally compact as well. (This is a pretty standard assumption in topological group theory.) – Matt E Nov 25 '11 at 21:58

First of all, in a general topological space, a single point need not be closed. But even if you are assuming that points are closed (e.g., this is a Hausdorff space), it does not necessarily follow from the zero-dimensional assumption that $\{e\}$ is open:

$\{e\}$ closed $\implies$ complement of $\{e\}$ open $\implies$ complement of $\{e\}$ is the union of clopen sets

But you cannot conclude from here (as you may want to) that the complement of $\{e\}$ is closed, because there may be infinitely many clopen sets involved in the union, and the infinite union of closed sets need not be closed.

share|cite|improve this answer
Ah, I thought that it meant that the topology is composed of clopen sets. Anyway, it doesn't say that $G$ is locally compact, but if there's no other way to prove it, I guess I would like to see how to show this with the added assumption of local compactness. Thanks. – MathematicalPhysicist Nov 26 '11 at 7:32

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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