3
$\begingroup$

Let $\mathcal{I}$ be a proper ideal on $\omega$. If $\mathcal{I}$ is Borel as a subset of $2^\omega$, does it follow that $\mathcal{I}$ is meager?

Edit: What if $\mathcal{I}$ contains all finite subsets of $\omega$?

$\endgroup$
2
  • 1
    $\begingroup$ All sets not containing $0$ ... that's an ideal, right? And Borel. In fact clopen. So not meager. $\endgroup$ May 6, 2013 at 13:06
  • $\begingroup$ Right. What if we assume that all finite subsets are in $\mathcal{I}$? $\endgroup$
    – jack
    May 6, 2013 at 14:00

1 Answer 1

2
$\begingroup$

Let us consider a prime ideal J containing I. Suppose I is non meager. Then J would have non empty interior (modulo meager). Since complementing a set of integers is a meager preserving operation so the dual ultrafilter U of J has the same property. But both J and U are closed under rational translations and they both have interiors (modulo meager) so they are both comeager which is impossible.

$\endgroup$
7
  • $\begingroup$ Much as I like ultrafilters, I have to point out that they're not really needed here. If I is a proper, non-meager, Borel ideal that contains the finite sets, and if F is the dual filter, then, being closed under finite modifications (a.k.a. rational translations), both I and F are comeager, which is absurd because they are disjoint. $\endgroup$ May 6, 2013 at 14:27
  • $\begingroup$ Hi. I'm new to descriptive set theory, so could you use layman's language? I don't understand the terms "prime ideal", "modulo meager", and "rational translations". $\endgroup$
    – jack
    May 6, 2013 at 14:34
  • $\begingroup$ After googling, I found what "rational translations" means. But, Andreas how does that imply that $I$ and $F$ are comeager? Where did you use the fact that $I$ is Borel? $\endgroup$
    – jack
    May 6, 2013 at 14:51
  • 1
    $\begingroup$ @jack: The assumption that I and therefore F are Borel is used to ensure that they have the Baire property, which in turn implies that, if they're non-meager, then they are comeager in some open sets. That plus closure under finite modifications makes them comeager in all of $2^\omega$. $\endgroup$ May 6, 2013 at 15:21
  • $\begingroup$ I see. Why is $F$ non-meager? $\endgroup$
    – jack
    May 6, 2013 at 15:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.