Let $S_\omega$ be the group of bijections of the countable ordinal $\omega:=\{0,1,2,\dots\}$ and $Alt_\omega$ be the subgroup of $S_\omega$ consisting of even permutations of $\omega$ (i.e., the compositions of even number of transpositions). By Onofri-Schreier-Ulam Theorem, $Alt_\omega$ is the smallest non-trivial normal subgroup in the permutation group $S_\omega$.

It is known that

$\bullet$ the group $Alt_\omega$ is not isomorphic to a subgroup of a compact topological group;

$\bullet$ the group $S_\omega$ is not isomorphic to a dense subgroup of a non-discrete locally compact topological groups.

These two results motivate the following

Problem 1. Is the group $Alt_\omega$ isomorphic to a dense subgroup of a non-discrete locally compact topological group?

Since the group $Alt_\omega$ is simple, this problem can be reformulated in the following form.

Problem 2. Let $h:Alt_\omega\to G$ be a homomorphism of $Alt_\omega$ to a locally compact topological group $G$. Is the image $h(Alt_\omega)$ discerete?

Remark. By the answer of @YCor to this question, for every homomorphism $h:S_\omega\to G$ to a locally compact topological group $G$ the image $h(S_\omega)$ is discrete.

  • $\begingroup$ What is the definition of $Alt_\omega$, there is a subgroup of even permutations of finite support, is that what you mean? Why don't you ask it for $S^{fin}_\omega$ first? $\endgroup$ Feb 23, 2018 at 13:52
  • 1
    $\begingroup$ @AndreasThom You are right: $Alt_\omega$ is the subgroup of even permutations (with automatically finite support). I wrote the definition in the third line of the question: the composition of even number of pemutations. Why $Alt_\omega$? Because it is a smallest normal subgroup of $S_\omega$ and the answer for $Alt_\omega$ will imply the answer for all larger subgroups of $S_\omega$, including $S_\omega^{fin}$. I think that the questions for $Alt_\omega$ and $S_\omega^{fin}$ could be equivalent (as $Alt_\omega$ has index 2 in $S_\omega^{fin}$). $\endgroup$ Feb 23, 2018 at 14:00
  • 3
    $\begingroup$ I cannot understand what is wrong with this question that it has got two downvotes. A question as many other questions. At least it has a natural motivation. What is wrong? Istead of demonstrating the power in downvoting, better suggest a solution! $\endgroup$ Feb 23, 2018 at 14:20
  • 1
    $\begingroup$ @TarasBanakh the result that $Alt_\omega$ is the minimal nontrivial normal subgroup in $Sym(\omega)$ was first proved by Onofri in 1929, reproved 4 years later by Schreier-Ulam, and owes nothing to Baer (who generalized to uncountable cardinals). math.stackexchange.com/a/2645097/35400 $\endgroup$
    – YCor
    Feb 27, 2018 at 22:40
  • $\begingroup$ @YCor Thank you for the info. I made the corresponding changes in the text of OP. $\endgroup$ Feb 27, 2018 at 23:01

1 Answer 1


Corollary 1.5 of this Vissarion Belyaev's paper says that

A group containing an infinite inert residually finite subgroup embeds into a non-discrete locally compact group.

Here, a subgroup $H$ of a group $G$ is called inert if $H\cap g^{-1}Hg$ has a finite index in $H$ for any $g\in G$.

So, in the alternating group $Alt_\omega$, for instance, the (abelian) subgroup generated by the 3-cycles $(1\;2\;3),(4\;5\;6),(7\;8\;9),\dots$ is inert. infinite, and residually finite .

Thus, the answer to Question 1 is yes (and to Question 2 is no).

  • $\begingroup$ Anton, thank you for the answer! Truly speaking, I expected that it will be opposite. $\endgroup$ Feb 27, 2018 at 19:30
  • 1
    $\begingroup$ The emphasized statement is (essentially) originally due to Schlichting, 1980. At least Schlichting's construction applies here. On the other hand, I think the first explicit dense embeddings of the alternating group into a locally compact group are in Willis' J. Algebra 2007 paper. $\endgroup$
    – YCor
    Feb 27, 2018 at 22:37
  • 1
    $\begingroup$ Yves, it seems that Belyaev was unaware of Schlichting's work. But this is not just "a construction". This is actually a criterion: a group $G$ is isomorphic to a dense subgroup of a totally disconnected non-discrete locally compact group if and only if $G$ contains an infinite inert residually finite subgroup (I am citing Belyaev's paper).. $\endgroup$ Feb 28, 2018 at 13:02
  • $\begingroup$ Taras, you are welcome. The symmetric group and finitary symmetric group are quite different. $\endgroup$ Feb 28, 2018 at 13:03

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.