2
$\begingroup$

I have a question about open books and Lefschetz fibrations over the 2-disk $D^2$. Please let me set it up first, before going on.

Setup: Say we have a Lefschetz fibration $f: W^4 \rightarrow D^2 $ , i.e., over the 2-disk $D^2 $ . Then f restricted to the boundary of $W^4 $ is an open book $( \Sigma, \Phi)$ for $\partial W^4$, where $\Phi$ is the monodromy and $\Sigma$ is the fiber surface. .

Now, I want to go in the opposite direction and "embed" a given open book $(\Sigma, \Phi)$ in a Lefschetz fibration with singularities, say $x_1,..,x_n$ , i.e., given this open book, I want to find a Lefschetz fibration whose boundary is $( \Sigma, \Phi)$ . A necessary ( and I think sufficient) condition is that the total monodromy of the fibration, given as the composition of Dehn twists $D_1 \circ \ D_2 \circ ...\circ D_n$ about vanishing cycles $\gamma_i$ in the critical surfaces $f^{-1}(x_i)$ must agree with the monodromy $\Phi $ of the open book.

Question ( Phew): Is it always possible to do this, i.e., given $\Phi$ in MCG( $\Sigma $) , is it always possible to express $\Phi $ as the composition of Dehn twists about vanishing cycles?

Thanks in Advance.

$\endgroup$
2
$\begingroup$

Indeed, it is necessary and sufficient that the monodromy be a product of (positive) symplectic Dehn twists.

However, a general mapping class $\Phi\in\operatorname{MCG}(\Sigma)$ need not be expressible as a product of positive Dehn twists. There are probably more elementary ways of seeing this, but here is one: every contact three-manifold has an open book (by Giroux) and if its monodromy is expressible as a product of positive Dehn twists, then it is Stein fillable (by the existence of a Lefschetz fibration), however there certainly exist contact three-manifolds which are not Stein fillable (e.g. anything overtwisted).

You may find this note of Auroux interesting. He mainly discusses Lefschetz fibrations over $S^2$, but of course much of what he says is applicable over $D^2$ as well.

$\endgroup$
1
  • $\begingroup$ Thanks; I guess non-Anosov maps cannot be represented as a composition of positive ( or , in this case, not even negative) Dehn twists? $\endgroup$ – user61837 Nov 16 '14 at 6:01

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.