Suppose $H$ is some group with finite isomorphic subgroups $A$ and $B$, and isomorphism $\psi: A\rightarrow B$. Suppose that the HNN-extension $G=\langle H, t\mid a^t=\psi(b)~\forall~a\in A\rangle$ is hyperbolic. Is $H$ hyperbolic?

My thoughts on this are rather sparse. I know that the opposite is true - that if $H$ is hyperbolic and $A$ and $B$ are finite then $G$ is hyperbolic. Clearly $H$ needs to be a finitely presented non-hyperbolic subgroup of a hyperbolic group. I know that such subgroups exist (discovered by Noel Brady), but that their construction is a serious result.

  • 3
    $\begingroup$ It is enough to show that $H$ embeds quasi-isometrically into $G$ which, since $A$ and $B$ are finite seems clear. In fact if you include elements of $A$ in the generating set of $G$ then any element of $H$ has a geodesic representative in the generators of $H$. $\endgroup$ – Derek Holt Dec 12 '16 at 12:24

As Derek Holt indicates in comments, this is standard. You should do the following easy exercise.

If $G = A*_CB$ or $G=A*_C$ is hyperbolic and $C$ is quasiconvex, then so is $A$ (and $B$).

Since finite subgroups are trivially quasiconvex, and quasiconvex subgroups are themselves hyperbolic, the result follows.

| cite | improve this answer | |
  • $\begingroup$ Okay, thanks. I'll do the exercise :-). $\endgroup$ – user1729 Dec 12 '16 at 13:00

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.