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

Given a geodesic metric space $X$ together with a choice of midpoints $m:X\times X\rightarrow X$ (i.e. $d(m(x,y),x)=d(m(x,y),y)=d(x,y)/2$). Assume furthermore, that the following nonpositive curvature condition is satisfied:

$d(m(x,y),m(x,z))\le \frac{d(y,z)}{2}$ for all $x,y,z\in X$ . This is just a special case of the CAT(0) inequality for the "triangle" $x,y,z$. Lets call such a space a M-space.

Such a space needn't be CAT(0), as the example $(\mathbb{R}^n,d^1)$ shows, where $d^1$ is the $l^1$ metric. The choice of midpoints is given by $m(x,y)=\frac{x+y}{2}$. It also needn't be unique geodesic.

But this space can be equipped with another metric, that makes it a CAT(0) space.

So my question is: Is every group, that acts properly, isometrically and cocompactly on a M-space already a CAT(0)-group?

share|cite|improve this question
You might find this paper and references therein helpful: See also: – Ian Agol Sep 7 '10 at 15:47
Tiny nitpick: I would call this a 'non-positive curvature' condition, rather than a 'negative curvature' condition. – HJRW Sep 7 '10 at 20:25
Ian, median spaces are not the same as M-spaces. Otherwise the answer would be positive (median spaces admitting co-compact proper isometric actions should be closely related to CAT(0)-cubical complexes). I think that an M-space does not necessarily have quadratic "isoperimetric" function, hence the answer should be in fact negative. – Mark Sapir Sep 8 '10 at 1:25

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.