Consider a permutation group $G$ acting on an infinite set $X$. Assume $G$ has finitely many orbits, and every point stabiliser $G_x$ has finite orbits. Now consider a permutation $\tau\in\operatorname{Sym}(X)$ of finite order, and let $H=\langle G,\tau\rangle$. Is it necessarily true that every point stabiliser $H_x$ has finite orbits?

The situation I'm particularly interested in is when $\tau$ is a cycle with one element in each orbit of $G$, such that $H$ is transitive.


1 Answer 1


No. Consider $G=\mathbf{Z}$ acting on itself by translation. Let $\tau$ be the transposition $(0,1)$. Then $H$ is the group of permutations of $\mathbf{Z}$ coinciding to translations at infinity; in particular it contains all finitely supported permutations; thus the stabilizer $H_0$ acts transitively on the complement of $\{0\}$.

It remains no in your more restricted setting (one cycle meeting each orbit once). Consider $G$ acting on $\mathbf{Z}$, generated by $\alpha:n\mapsto -n$ and $\beta: n\mapsto 2-n$. Thus $G$ is infinite dihedral, and has 2 orbits (odd and even numbers). Let $t$ be the transposition $(0,1)$ and $H=\langle G,t\rangle$. Then $\beta t\beta^{-1}$ is the transposition $(1,2)$ and hence also belongs to $H$. Given that $\beta\alpha$ is the translation $n\mapsto n+2$, we deduce that all transpositions $(n,n+1)$ belong to $H$ and hence $H$ contains all finitely supported permutations, and hence has infinite point stabilizers.

Note that the answer is yes when $G$ acts freely, and $t$ is an $n$-cycle meeting once each orbit. Indeed, this case, the $G$-conjugates of $t$ pairwise have disjoint support, and generate a group preserving the partition by these supports, which contains all point stabilizers.

  • $\begingroup$ Thank you for the illuminating counterexamples! I was too optimistic thinking any $\tau$ would work. Might the title question still be salvable in the following way: for general $G$ with finitely many orbits and finite suborbits, can one always find an element $\tau$ (not necessarily of finite order) such that $\langle G,\tau\rangle$ is transitive with finite suborbits? $\endgroup$ Sep 28, 2018 at 11:40
  • $\begingroup$ You mean $\tau\notin G$? or some additional requirement? Also, do you still assume that $G$ has finite stabilizers? $\endgroup$
    – YCor
    Sep 28, 2018 at 12:22
  • $\begingroup$ Yes, any $\tau\in\operatorname{Sym}(X)$ but $\tau\notin G$ (otherwise $\langle G,\tau\rangle=G$ will not be transitive). The only assumptions on $G$ are (1) finitely many orbits and (2) finite suborbits (orbits of point stabilisers). $\endgroup$ Sep 28, 2018 at 15:29
  • $\begingroup$ Ok, this is a reasonable question but should be asked as a separate post. $\endgroup$
    – YCor
    Sep 28, 2018 at 15:31
  • $\begingroup$ Follow-up: mathoverflow.net/questions/311637 $\endgroup$
    – YCor
    Oct 30, 2018 at 15:52

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.