Let $G$ be an infinite locally finite (non-solvable) group, let $\{H_i\}_{i\in\mathbb{N}}$ be a strictly totally ordered family of subgroups of $G$ and let $H$ be the intersection of that family. If $H$ is normal in $G$, it is clear that $G/H$ is an infinite locally finite group and hence there is no bound on the cardinality of the finite subgroups $F$ such that $\langle H,F\rangle$ is a proper subgroup of $G$. The same holds if $H$ is finite. Is this true for a not necessarily normal (or finite) subgroup $H$, intersection of such a family of proper subgroups of $G$? Under which conditions could this hold?

(Clearly, the only case to consider is that in which the family is ordered as the negative integers with $<$)


I think the following example works. Let $X$ be any finite group and let $H = \times_{k \in{\mathbb N}} X_i$ be the (restricted) direct product of countably many copes of $X$. Let $G = H \wr C_2$.

So $G$ is an extension of the base group $H_1 \times H_2$ with $H_i \cong H$ by a group $F = \langle t \rangle$ of order $2$, where $H_1^t=H_2$.

Then $G = \langle H_1,F \rangle$, and it is straightforward to write $H_1$ as the intersection of an infinite strictly descending chain of subgroups.

  • $\begingroup$ I'm now thinking that the title may be misleading. The question, posed in terms of chains, should be the following: "Is $H$, intersection of an infinite strictly descending chain of subgroups of a locally finite group $G$, contained in an infinite strictly ascending chain of subgroups of $G$ (just as in the cases in which $H$ is normal or finite)? Are there additional conditions for this to hold?" $\endgroup$ – Alex Doe Aug 17 '18 at 8:17
  • $\begingroup$ Which is clearly the case of your example, since $H_1$ is contained in the base group. But is it true in general? $\endgroup$ – Alex Doe Aug 17 '18 at 8:24
  • $\begingroup$ The example shows that the answer to the question in the title is yes. But I will think about your more detailed question. $\endgroup$ – Derek Holt Aug 17 '18 at 9:18
  • $\begingroup$ I know, but, trying to simplify my question for the title, I overdid it getting to a misleading title. The true question is in the text, as it should be. Sorry for the bungle! Maybe I should change the title. $\endgroup$ – Alex Doe Aug 17 '18 at 9:47
  • $\begingroup$ I've just changed the title, which now is general and hopefully not misleading! $\endgroup$ – Alex Doe Aug 20 '18 at 8:09

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.