Does there exist a quasi-isometric embedding $$MCG(S) \to (\mathrm{Teich}(S), d)$$ for $d$ any "known" distance on the Teichmuller space (i.e. Teichmuller, Weil-Petersson, Thurston...) ?

  • $\begingroup$ This is not exactly the answer you want but you can find some related results here <arxiv.org/abs/math/0701719> $\endgroup$
    – Cusp
    Oct 13, 2013 at 3:52
  • 1
    $\begingroup$ I believe the answer is no for Teichmuller and Weil-Petersson: I think this follows from work of Minsky, Rafi, and Brock, who found combinatorial models for measuring distance in the mapping class group, Teichmuller metric, and Weil-Petersson metric respectively. $\endgroup$
    – Ian Agol
    Oct 13, 2013 at 4:14
  • $\begingroup$ @Ian: I am not sure, since a qi embedding might be unrelated to the standard action of the mcg. $\endgroup$
    – Misha
    Oct 13, 2013 at 14:29
  • $\begingroup$ Ok, good point Misha! I implicitly assumed he was looking for an equivariant quasi-isometry, so whether the orbit of a point under the mcg is a q.i. embedding. $\endgroup$
    – Ian Agol
    Oct 13, 2013 at 16:48

1 Answer 1


A result of Behrstock and Minsky (cf. Hamenstadt too) implies that the rank of mapping class groups is the maximal rank of abelian subgroups, which is $3g+p-3$ for a connected hyperbolic surface of genus $g$ with $p$ boundary components. The rank of Teichmuller space with respect to the Weil-Petersson metric they show is $\lfloor \frac{3g+p-2}{2}\rfloor < 3g-3+p$, so there is no q.i. embedding of the mapping class group to Teichmuller space with the Weil-Petersson metric.

On the other hand, Eskin-Masur-Rafi show that the rank of the Teichmuller metric is equal to that of the mapping class group. They point out that the orbit of a point under the mapping class group is not a quasi-isometric embedding. But this leaves open whether there might be a non-equivariant q.i. embedding.


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.