2
$\begingroup$

Consider $A := \mathrm{Aut}(\mathbb{Q},\leq)$, the group of order-automorphisms of $(\mathbb{Q},\leq)$. Call a subgroup $U$ highly order-transitive if for any two finite ordered sequences $s_1$ and $s_2$ of the same size, there is an order-automorphism of $U$ which sends $s_1$ to $s_2$, and note that $A$ has this property.

It is well known that $A$ contains countable subgroups which are also highly order-transitive (whereas $A$ is not countable). For instance, the free groups $F_\eta$ with $\eta \geq 2$ and finite can be embedded in $A$ with this property.

My question is: do there exist countable, highly order-transitive subgroups $U$ of $A$ with the following additional property:

  • There exists an element $u$ in $U$ which fixes every element of $\mathbb{Q}$ outside some open interval $(a,b)$ with $a \ne b$ and $a, b \in \mathbb{Q}$, and no element in $(a,b)$ itself.
$\endgroup$
3
  • 3
    $\begingroup$ If you know there exists a countable highly order-transitive (“h.o.t.”) subgroup $F$ of $A$, and clearly there exists $u\in A$ with the property in your final point, why can't you just take the subgroup of $A$ generated by $F\cup\{u\}$? It seems to be countable, h.o.t. and satisfy the final point. Did I misunderstand the question? $\endgroup$
    – Gro-Tsen
    Commented May 10, 2023 at 12:10
  • 4
    $\begingroup$ Can you not just take the group of strictly increasing piecewise-linear maps with rational breakpoints and rational slopes? Or the subgroup of that consisting of maps that are equal to the identity outside some finite interval? $\endgroup$ Commented May 10, 2023 at 12:48
  • 2
    $\begingroup$ About the bare existence: it is immediate that every highly transitive subgroup contains a countable highly transitive subgroup (just pick one element for each pair of finite subsets of $\mathbf{Q}$ of the same cardinal). $\endgroup$
    – YCor
    Commented May 10, 2023 at 12:57

0

You must log in to answer this question.

Browse other questions tagged .