1
$\begingroup$

This example is confusing me.

Is $\mathbb{Z}_3 = \langle a\vert a^3\rangle$ $\operatorname{CAT}(0)$ and/or (Gromov) $\delta$-hyperbolic?

The Cayley graph clearly has bounded diameter, therfore it is (Gromov) $\delta$-hyperbolic.

But triangles are more "fat" than comparison triangles in $\mathbb{R}^2$. Which would mean that it is not CAT(0).

I also read that it is an open question if all Gromov hyperbolic groups are also CAT(0). Therefore for this simple example I think it is more likely that the group is in fact CAT(0).

I can't see where I make a mistake.

$\endgroup$
3
$\begingroup$

Edit: As pointed out by Lee Mosher in comments, it is sufficient to say that finite groups act geometrically on a point: of course, a point is hyperbolic and CAT(0)...


Finite groups are clearly hyperbolic, therefore $\mathbb{Z}_3$ so is.

Then, whether a group is CAT(0) is not readable on its Cayley graph: we want to know if it acts geometrically on a CAT(0) space. In the case of $\mathbb{Z}_3$, you can make it act on a disk $D^2$ thanks to a rotation of order 3: $D^2$ is CAT(0) and the action is geometric, so $\mathbb{Z}_3$ is a CAT(0) group.

More generally, it is not difficult to prove that any finite group $F$ is CAT(0): let $T$ be the graph whose vertices are $F \cup \{ o \}$ and where $o$ is linked by an edge with any vertex of $F$ - therefore, $T$ looks like a star. Now, let $F$ act on $T$ fixing the vertex $o$. Clearly, $T$ is CAT(0) - it is a tree - and the action $F \curvearrowright T$ is geometric.

$\endgroup$
  • $\begingroup$ What rotation of order 3 acts on the square $[0,1]^2$? $\endgroup$ – Lee Mosher Feb 11 '15 at 15:37
  • 1
    $\begingroup$ Also, every finite group acts geometrically on a point. $\endgroup$ – Lee Mosher Feb 11 '15 at 15:38
  • $\begingroup$ @LeeMosher: You are completely right, I searched something really too complicated! I added a remark in my answer, thank you for your comment. $\endgroup$ – Seirios Feb 11 '15 at 18:55
  • $\begingroup$ Also, every finite graph quasi-isometrizes to a point, so they are automatically $\delta$-hyperbolic. $\endgroup$ – Balarka Sen Feb 11 '15 at 19:12

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.