1
$\begingroup$

Let $X$ be a compact Haussdorf space, let $\mu$ be a Borel measure on $X$ with $\mathrm{supp}(\mu)=X$ and let $(\phi_s)_{s\in\mathbb R}$ be a one-parameter group of homeomorphisms which is

  • continuous in the sense that $\mathbb R\times X \ni (s,r) \mapsto \sigma_s(x) \in X$ is a continuous map, and
  • ergodic in the sense that $\phi_s(\mu)$, $s\in\mathbb R$, is a quasi-equivalent family of measures with the property that every $\phi$-invariant measurable subset is a null or a co-null set.

The ergodic system is called transitive if, for almost all $x\in X$, the orbit $O_x= \{ \phi_s(x) : s\in\mathbb R\}$ is a co-null set and non-transitive (or properly ergodic) if, for almost all $x\in X$, the orbit $O_x= \{ \phi_s(x) : s\in\mathbb R\}$ is a null set.

My question is the following: Denote by $p(x) = \min\{s>0 : \sigma_s(x) = x\}$ the minimal period of a point $x$. In the transitive case, almost all points have the same period. What can we say about $p(x)$ in the non-transitive case? Do almost all points have $p(x)=\infty$? If not, do points $x$ with arbitrarily large periods exist?

$\endgroup$
6
  • 1
    $\begingroup$ The question is too vague. Is the measure mu finite? Do you understand that ergodicity implies transitivity in restriction to the support of the measure if the measure is finite? The answer is no in general to your two questions but it is not clear exactly what kind of systems you are interested in. $\endgroup$
    – coudy
    Commented Nov 17, 2023 at 13:23
  • $\begingroup$ Lets say the measure $\mu$ is finite. I forgot to say that the measure $\mu$ is supposed to have full support, I'll add that. Thanks! As I understand it, $\mu$ being finite or not does not affect the question which only depends on the equivalence class of the measure $\mu$ (equivalence as in: two measures on $X$ are equivalent if they have the same null sets). $\endgroup$
    – Lau
    Commented Nov 17, 2023 at 13:43
  • $\begingroup$ @coudy Ok, I see. The point is that not every equivalence class of Borel measures contains a finite measure. Then I don't want to assume finiteness of $\mu$. So you are saying that we cannot have non-transitivity if $\mu$ is finite. Can you point me to a reference? I'd still be interested in the non-finite case though. $\endgroup$
    – Lau
    Commented Nov 17, 2023 at 14:19
  • $\begingroup$ @coudy Could you explain why the answer is no in general? Also, can you point me to a reference for (or explain why) $\mu$ cannot be finite if $\phi$ is non-transitive and ergodic. Thanks! $\endgroup$
    – Lau
    Commented Nov 18, 2023 at 13:31
  • 1
    $\begingroup$ Reading your question again, I see I misinterpreted the meaning of transitivity in your post. The standard definition of transitivity is the existence of a dense orbit, but this is not what you are interested in. So my previous comment does not apply. $\endgroup$
    – coudy
    Commented Nov 18, 2023 at 18:29

0

You must log in to answer this question.

Browse other questions tagged .