2
$\begingroup$

Assume $V$ be a genus larger than 1 handlebody, $S=\partial_{+} V$.

Denote $N$ be the normal closure of $MCG(V)$ in $MCG(S)$.

Is there any material related to the quotient group $MCG(S)/N$ ? Thanks!

$\endgroup$
2
  • $\begingroup$ @l.Mosher:sorry I have it edited. $\endgroup$
    – yanqing
    Jun 28, 2012 at 13:26
  • $\begingroup$ @Lee: He probably means normal closure. $\endgroup$
    – Misha
    Jun 28, 2012 at 13:27

1 Answer 1

6
$\begingroup$

The normal closure of $MCG(V)$ in $MCG(S)$ is all of $MCG(S)$. To see why, we know that $MCG(S)$ is generated by Dehn twists, so it suffices to prove that each Dehn twist is in the normal closure. Consider the Dehn twist $\tau_c$ about an essential simple closed curve $c$. There exists a homeomorphism $f : S \to S$ such that $c' = f(c)$ bounds a properly embedded disc $D \subset V$, and $\tau_{c'}$ extends to a twist about the disc $D$, so $\tau_c = f^{-1} \circ \tau_{c'} \circ f$ is in the normal closure of $MCG(V)$. To find this homeomorphism, observe that the there is one $MCG(S)$-orbit of nonseparating curves, and for each partition $genus(S)=m+n$ there is one $MCG(S)$ orbit of separating curves whose complements have genus $m,n$. Each such orbit clearly has a representative bounding a disc in $V$.

$\endgroup$
7
  • $\begingroup$ @L.Mosher: Assign $MCG(S)$ a clay graph $G$. Obviously there is a subgraph $G^{*}$ with vertices corresponding to $MCG(V)$. Is there a vertex in $G$ which is arbitrarily far from $G^{*}$ ? $\endgroup$
    – yanqing
    Jun 28, 2012 at 14:00
  • $\begingroup$ @yanqing: I don't know what a clay graph is. $\endgroup$
    – Lee Mosher
    Jun 28, 2012 at 14:38
  • $\begingroup$ Cayley graph. And the answer to your question is "Yes." Said another way, the inclusion of $MCG(V)$ into $MCG(S)$ is not "coarsely dense." I don't know a slick proof of this. $\endgroup$
    – Sam Nead
    Jun 28, 2012 at 17:38
  • $\begingroup$ This follows from the fact that $MCG(V)$ has infinite index in $MCG(S)$. For any subgroup of any finitely generated group $G$, if the subgroup is coarsely dense then it has finite index in $G$. There might be a very slick proof somewhere, but there is a pretty quick proof of a more general result in my paper with Sageev and Whyte "Quasi-actions on trees I", which says that two subgroups of $G$ at finite Hausdorff distance from each other are commensurable in $G$. $\endgroup$
    – Lee Mosher
    Jun 28, 2012 at 17:56
  • $\begingroup$ @Lee: This is very simple: If $|G:H|=\infty$ then $H\backslash G$ is infinite, so $X=H\backslash \Gamma_G$ has infinite diameter, where $\Gamma_G$ is a Cayley graph of $G$. Thus, you have geodesic paths $p_i\subset X$ of arbitrary large length. Lifting these paths to $\Gamma_G$ proves failure of coarse denseness of $H$ in $G$. The same proof gives you the reverse implication. $\endgroup$
    – Misha
    Jun 29, 2012 at 0:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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