Sign up ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.


Let $G$ be an infinite finitely generated discrete group. I call an infinite set $S$ irregular iff for every $g\in G$, $g\neq 1$, we have that $S\cap gS$ is finite. For example $\{z^3|z\in\mathbb{Z}\}$ is irregular in $\mathbb{Z}$. Now my easy to state question: Does every free ultrafilter on $G$ contain at least one irregular subset?

The following is true and might probably be useful: $G$ acts freely on the space of all ultrafilters (the Stone-Cech compactification of $G$ as a discrete space).

free ultrafilters:

share|cite|improve this question
In the definition of irregular, I guess you want $g$ to be non-identity? – Joel David Hamkins Nov 17 '12 at 11:31
To save people searching through the wikipedia article: free = non-principal. – HJRW Nov 17 '12 at 12:03
@Joel: Yeah, I'll correct that, thx. – Werner Thumann Nov 17 '12 at 14:05

1 Answer 1

up vote 4 down vote accepted

No, a free ultrafilter on the additive group of integers need not contain an irregular set. The Galvin-Glazer proof of Hindman's theorem (which is nowadays the standard proof of that theorem) begins by showing the existence of idempotent ultrafilters $U$ on $\mathbb N$. I won't bother to define "idempotent" here, since what I need is not the definition but the following consequence of it. If $X$ is any set in $U$, then there is an infinite set $\{x_0<x_1<\dots\}$ of positive integers such that any finite sum of distinct $x_i$'s is in $X$. In particular, the intersection of $X$ and $x_0+X$ contains $x_0+x_i$ for all $i>0$, and therefore $X$ is not irregular.

share|cite|improve this answer
Thx for the answer, but are you sure this argument carries over to the group Z? Does Elli-Numakura really give an idempotent besides the trivial one around 0? Im a layman in this field, so plz dont blame me if this is a stupid question. ;) – Werner Thumann Nov 18 '12 at 18:32
Any idempotent ultrafilter on $\mathbb N$ generates (by closing it under supersets) an idempotent ultrafilter on $\mathbb Z$. More generally, the canonical embedding $\beta\mathbb N\to\beta\mathbb Z$ is a homomorphism of semigroups. – Andreas Blass Nov 19 '12 at 1:37

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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