MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In a paper I am reading at the moment (Hrushovski Martin, elimination of imaginaries in $Q_p$), in some proof they use the following fact (at least this would be enough to get their proof going, but maybe we have more hypothesis) :

If we have a group $G$ acting on a set $X$ containing elements $(e_i\mid i<\omega)$ such that all $g\in G$ fixes all the $e_i$ but a finite number, is it true that all but a finite number of $e_i$ are fixed by $G$?

The authors say it is a consquence of Neumann's lemma (my guess is the one saying that a group covered by a finite number of cosets of subgroups is covered by the ones of finite index) but I have been trying to figure it out for some time already and cannot. If anyone has any idea of how that might work, a little help would be very welcome.

Some days later : Thanks a lot for the counter-examples, they really helped me understand.

If anyone is interested, what I really needed was that there are at most a finite number of $e_i$ with an infinite orbit under $G$. This is false in general (see the counter-example given in the answer) but Newmann's lemma does imply that if there are an infinity of elements with an infinite orbit then for all $N$, we can find a $g$ and $N$ elements that are not fixed by $g$. And as it happens that my $G$ is in fact the automorphism group of a very saturated model (sorry, some model theory had to barge in) by compacity I can find an automorphism that moves an infinity of $e_i$.

I hope any of this made sense.

share|cite|improve this question
How about $\oplus_{n=1}^\infty(C_2)$, $C_2$ cyclic of order 2, acting on $X=\{1,2,3,4,...\}$ via the generator of the $n$th $C_2$ switching $2n-1$ and $2n$, and leaving the rest alone? This is perhaps a counterexample if it all adds up... – Kevin Buzzard Jun 4 '11 at 14:04
It seems to work indeed... I'll go back to the paper to see if there are some hypothesis I forgot... – Silvain Rideau Jun 4 '11 at 14:41
A "brute force" counterexample: Let $X$ be any infinite set and let $G$ consist of those permutations $g$ of $X$ such that $\{x\in X: g(x)\neq x\}$ is finite. (It's trivial to check that $G$ is a group.) Then take the $e_i$ to be any countable sequence of distinct elements of $X$. – Andreas Blass Jun 5 '11 at 2:44
up vote 0 down vote accepted

Let $S_\infty=\bigcup_{n=1}^\infty S_n$ be the infinite symmetric group consisting of permutations of $\mathbb{N}$ fixing all but finitely many elements. Then $\{e_i=i\}$ satisfy your conditions but no element $i\in \mathbb{N}$ is fixed by $S_\infty$.

(Sorry, I see that Andreas Blass already gave this counterexample. I've made this answer community wiki.)

share|cite|improve this answer

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.