I found a proof in the book Fourier Analysis On Number Fields that the closure of any subgroup is a subgroup, using the continuity argument along with the nets.
Nevertheless, the following proof seems also plausible, but is it valid?

Statement: If $H$ is a subgroup of a topological group $G$, then so is its closure in $G$.
Denote the closure of $H$ by $H^{closure}$. Let h,h' be two elements in $H^{closure}$, and let $U$ be a neighbourhood of the unity in $G$. Since h and h' lie in $H^{closure}$, there are elements a, a' in $U$ such that
ha$\in H$,
h'a'$\in H$.
Then, as the congugation function is continuous, the left and the right toplogies are the same, and hence there are b, b' $\in U$ such that
hah'a'= bhh'a'= bh"=h"b' $\in H$, where h" is in $H^{closure}H$ $\subseteq H^{closure}$.
As to $H^{closure}H$ $\subseteq H^{closure}$, take two elements e $\in H^{closure}$, f $\in H$, and $v$ in $V$, an arbitrary neighbourhood of unity in $G$. Then there is an element $g$ in $V$ such that
$efv=(eg)f \in HH \in H$.
So $H^{closure}H$ $\subseteq H^{closure}$.

The proof for the inversion is similar; is there any problem in the argument? Thank a lot here.

  • $\begingroup$ P.S. C.Q.F.D. $\endgroup$ – awllower Nov 23 '11 at 8:11
  • $\begingroup$ Are you not supposed to show $H^{closure}H^{closure} \subseteq H^{closure}$ and ${H^{closure}}^{-1} \subseteq H^{closure}$ ? $\endgroup$ – Matthias Klupsch Nov 23 '11 at 8:15
  • $\begingroup$ I showed the first, by means of showing that $H^{closure}H \subseteq H^{closure}$, but then was tired, and the proof is similar, so I will edit again, if needed. $\endgroup$ – awllower Nov 23 '11 at 8:21

Sorry... this is not the answer to your question, but notice that the product with inversion $f: G \times G \to G$ defined by $f(x,y) = xy^{-1}$ is continuous.

Therefore, $f^{-1}(\overline{H})$ is closed. Now, notice that $H \times H \subset f^{-1}(\overline{H})$. So, taking closures, $$ \overline{H \times H} \subset f^{-1}(\overline{H}). $$ Now, you just have to show that $\overline{H} \times \overline{H} \subset \overline{H \times H}$, to conclude that $$ f(\overline{H} \times \overline{H}) \subset \overline{H}. $$

  • 1
    $\begingroup$ It is extraordinary!! I think even though you did not tell me if the above answer is correct, you did indeed tell me that there is a proof without the concept of nets, which was my initial incentive for this question. Thus I am going to accept the answer. Regards. $\endgroup$ – awllower Nov 24 '11 at 2:51
  • $\begingroup$ @awllower: Glad to help! :-) $\endgroup$ – André Caldas Nov 24 '11 at 3:02

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.