1
$\begingroup$

I found the the following claim here:

If $G$ is a group and $G_0 \subset G$ is a subgroup of finite index, then $G$ is hyperbolic if and only if $G_0$ is hyperbolic.

Why is this true? Can anyone provide a proof or reference to a proof? Also, is this a well-known result?

$\endgroup$
  • 2
    $\begingroup$ Try googling quasi-isometry/quasi-isometric. The Cayley graphs of $G$ and $G_0$ with any finite generating sets are quasi-isometric, which means roughly that there is a map between them which is an isometry up to multiplicative and additive constants. You can prove that if two metric spaces are quasi-isometric, then one has uniformly slim/thin triangles if and only if the other one does. This is all straighforward. $\endgroup$ – Derek Holt Feb 14 '17 at 20:01
1
$\begingroup$

Let me reduce this (quite) well known claim to two (very) well known theorems: the Milnor-Svarc lemma; and the theorem that hyperbolicity is a quasi-isometry invariant. Along the way, there are a few simple facts about group actions that are used.

If $\Gamma$ is a Cayley graph of $G$ then $G$ acts properly discontinuously and cocompactly on $\Gamma$, hence $\Gamma$ is quasi-isometric to $G$ (by the Milnor-Svarc lemma).

Now consider the restricted action of the subgroup $G_0$ on $\Gamma$. Proper discontinuity is inherited under restriction of the action of $G$ to any subgroup (pretty obvious), hence the action of $G_0$ on $\Gamma$ is properly discontinuous. Cocompactness is inherited under restriction of the action of $G$ to any finite index subgroup, hence the action of $G_0$ on $\Gamma$ is cocompact (proof: if $K \subset \Gamma$ is a finite subgraph whose $G$-translates cover $\Gamma$, and if $g_1,..,g_N$ are left coset representatives of $G_0$ in $G$, then $(g_1 \cdot K) \cup \cdots \cup (g_N \cdot K)$ is a finite subgraph whose $G_0$-translates cover $\Gamma$).

Again (by the Milnor-Svarc lemma) $\Gamma$ is quasi-isometric to $G_0$.

Since quasi-isometry is an equivalence relation, it follows that $G$ is quasi-isometric to $G_0$. Since hyperbolicity is a quasi-isometry invariant, it follows that $G$ is hyperbolic if and only if $G_0$ is hyperbolic.

$\endgroup$
  • $\begingroup$ Can you elaborate, or spell out, more on showing that the restricted action of a finite index subgroup is cocompact? $\endgroup$ – Al Jebr Mar 30 at 23:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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