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

Assume that $G$ is a group that fits into a short exact sequence $$1 \longrightarrow F \longrightarrow G \longrightarrow \mathbb{Z}^n \longrightarrow 1$$ with $F$ a finite group. For $f \in F$ not the identity, can we find some finite-index subgroup $A_f \subset G$ such that $f \notin A_f$?

Actually, the answer to this is yes : By assumption we have that $G$ is quasi-isometric to $\mathbb{Z}^n$, and it can easily be derived from Gromov's theorem on groups with polynomial growth that $G$ is virtually abelian. But this is attacking an ant with a sledgehammer. Does anyone know a more elementary proof?

share|cite|improve this question
@Todd Trimble : Yes, of course. I corrected the question. – Curtis Jan 17 '13 at 20:59
I'd call these groups 'finite-by-abelian'. The distinction is quite important in this case, since the abelian-by-finite case is trivial. – HJRW Jan 18 '13 at 10:42
up vote 7 down vote accepted

It follows easily from the result that (for finitely generated $G$) $G'$ finite implies $|G:Z(G)|$ finite.

If that is still too sledgehammer-like, then by replacing $G$ by $C_G(F)$ we can assume $F$ is central and hence $G$ is class 2 nilpotent. Let $|F|=k$, and let $x_1,\ldots,x_n$ generate $G$ modulo $F$. Since the commutator map is bilinear in class 2 nilpotent groups, we have $[x_i^k,x_j] = [x_i,x_j]^k = 1$ for all $i,j$, so the subgroup $\langle x_1^k,\ldots,x_n^k \rangle$ is central and has finite index in $G$.

share|cite|improve this answer
Doesn't this solution assume that $F$ is abelian? I'm not assuming that. – Curtis Jan 17 '13 at 21:40
@ Curtis: there is an action of $G$ on $F$ by conjugation, giving a homomorphism $G \to Aut(F)$. The kernel of this action is the centralizer $C_G(F)$ of $F$, and is clearly finite index. – Ian Agol Jan 17 '13 at 22:35
Assuming $F$ is central so $G$ is 2-step nilpotent, there are homomorphisms $G \to F$ given by $a \mapsto [a,x_i]$, $i=1,\ldots,n$, since $[ab,x]=[a,x][b,x]$. The intersection of the kernels of these maps for all $i$ gives $Z(G)$, and therefore is finite index since $F^n$ is finite. – Ian Agol Jan 17 '13 at 23:03
What is sledgehammerish about the first statement? If $G'$ is finite, then any conjugacy class is finite (the conjugates of $x$ are contained in $xG'$). Thus for every generator $g_i$, its conjugacy class is finite, so its centralizer $C_G(g_i)$ is finite index. The intersection of these is the center $Z(G)$, so as long as $G$ is finitely generated, $Z(G)$ is finite index. When $G$ is not finitely generated, instead you can show that the second center is finite index. – Steve D Jan 17 '13 at 23:11
Great, thanks!! – Curtis Jan 17 '13 at 23:14

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.