3
$\begingroup$

Let $Dom$ be a uniform space, and $\hspace{.04 in}f$ be a continuous function from $Dom$ to itself satisfying:

  1. For all non-empty open subsets $U$ and $V$ of $Dom$, there exists a natural
    number $n$ and a member $x$ of $U$ such that $f^n(x)$ is a member of $V$.

  2. The periodic points of $f$ are dense in $Dom$.


Does it follow that $f$ satisfies (3)?

$\;\;$3.$\:\:$There exists an entourage $E$ of $Dom$ such that for all members $x$ of
$\quad \;\;$ $Dom$ and all neighborhoods $U$ of $x$, there exists a member $y$ of $U$ and
$\quad \;\;$ a natural number $n$ such that $\:\langle \hspace{.05 in}f^n(x)\hspace{.02 in},\hspace{.03 in}f^n(\hspace{.03 in}y)\rangle\:$ is not a member of $E$.


According to this paper, the implication holds in metric spaces.

$\endgroup$
2

1 Answer 1

2
$\begingroup$

I believe the following works. Follow the proof in the reference supplied by Matthew Daws: choose two distinct periodic orbits and choose a compatible pseudometric $\rho$ such that all points in those orbits are at least $1$ apart under this pseudometric. The proof establishes that the entourage $\lbrace(x,y):\rho(x,y)<\frac18\rbrace$ is as required.

$\endgroup$
1
  • $\begingroup$ I haven't checked the details, but I was wondering if a pseudo-metric approach to uniform spaces might allow one to lift the proof... so I think this is a great idea! $\endgroup$ Commented Sep 14, 2010 at 9:59

You must log in to answer this question.