8
$\begingroup$

Is there a nontrivial profinite group $G$ with a binary transitive relation $<$ such that

  1. $x<y$ implies $x\neq y$, and for any different $x,y \in G$ either $x < y$ or $y < x$ and such that for any $x,y,z \in G$ we have that $x < y$ implies that $zx < zy$ (i.e., $<$ defines a left-invariant strict total order)
  2. $\{(x,y):x<y\}$ is open?
$\endgroup$
7
  • 2
    $\begingroup$ My guess would be that only $G = \{1\}$ works (finite groups are profinite). I don't have any proofs though. $\endgroup$
    – jmc
    Commented Dec 15, 2014 at 13:08
  • 3
    $\begingroup$ @jmc maybe continuous orders should be considered since by a result of Levi every abelian group with no torsion is orderable and this works for the additive profinite p-adic groups. $\endgroup$
    – Pablo
    Commented Dec 15, 2014 at 14:20
  • $\begingroup$ Ok, yes, a relation with the topology is very reasonable. (I kind of assumed it, but indeed, you did not state it.) $\endgroup$
    – jmc
    Commented Dec 15, 2014 at 14:26
  • 2
    $\begingroup$ If the order is closed in the topology on the direct product the answer is no by reduction to the finite case I believe. $\endgroup$ Commented Dec 15, 2014 at 15:25
  • $\begingroup$ Benjamin, I do not understand your comment. The profinite group is not necessarily a product of finite groups. $\endgroup$
    – Pablo
    Commented Dec 15, 2014 at 16:24

1 Answer 1

7
$\begingroup$

There is no such ordering, which is compatible with the profinite topology in the following way: If $x<y$, then there are small neighbourhoods $U, V$ of $x, y$, such that $u<v$ for all $u\in U, v\in V$.

To see this note that If $x>1$, then $x^n>1$ for all $n>0$ and $x^n<1$ for all $n<0$. But in the pro-finite topology, the sequence $x^{n!-1}$ converges to $x^{-1}$, so the set $\{x:x\geq 1\}$ is not closed. But this contradicts our assumption that the ordering is compatible with the topology.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .