Monodromy of codimension 1 subfamilies of universal curves (Q428164)

Let \({\mathcal M}_{g,n}[\ell]\) the moduli space of \(n-\)pointed genus \(g\) curves with a level \(\ell\) structure. The main result of the paper is the following: let \(E\) is a smooth variety that dominates a codimension 1 subvariety of \({\mathcal M}_{g,n}[\ell]\) and consider the associated monodromy representation \(\rho: \pi_1(E,e_0)\rightarrow \mathrm{Sp}_g(\hat{\mathbb Z})\). Then, if \(g\geq 3\) and \(\ell\geq 1\), the closure of the image of \(\rho\) has finite index in \(\mathrm{Sp}_g(\hat{\mathbb Z})\). More precisely, the main theorem concerns the \(n\)th power of the universal curve \({\mathcal C}_g^n[\ell]\) over \({\mathcal M}_{g}[\ell]\). It is proven that the same result does not hold for genus 2. A stronger version of the theorem is also provided for a generic linear section of dimension \(\geq 3\) of \({\mathcal C}_g^n[\ell]\), and for codimension 1 families of principally polarized abelian varieties. The proof relies on a technical result, which is defined by the author: a hybrid of a ``non-abelian strictiness theorem'' and a Lefschetz-type theorem. The prototype of \textit{abelian} strictiness theorem is the following (due to Deligne): given a diagram of morphisms of varieties \(Z{\longrightarrow}Y{\longrightarrow}X\) where \(Z\) and \(X\) are smooth and the first morphism is dominant, the image of \(H_1(Z)\rightarrow H_1(X)\) has finite index in the image of \(H_1(Y)\rightarrow H_1(X)\). The idea of a \textit{non-abelian} strictiness theorem would be to replace the first homology group with the fundamental group. This optimistic task can not be achieved, as shown by an example given in the last part of the paper, but a result in this direction is given under some technical assumptions.