On the second cohomology of Kähler groups (Q540011)

Carlson and Toledo conjectured that if \(\Gamma\) is an infinite Kähler group (i.e., the fundamental group of a compact Kähler manifold), then virtually \(H^2(\Gamma, \mathbb R) \not= 0.\) In the paper under review the authors prove the following: Let \(M\) be a compact Kähler manifold. Assume that there exists an irreducible rigid unbounded reductive representation \(\rho: \pi_1(M)\rightarrow \text{GL}(n,\mathbb C)\) for some \(n>0\) and that the variation of the Hodge structure \((E,\theta)=(\bigoplus_{i=0}^{k}E^{i}, \bigoplus_{i=0}^{k-1}\theta_{i})\) over \(M\) which induces \(\rho\) satisfies \(\operatorname{rk} E^{i} =1\) for some \(0<i<k\). Then \({H^{2}(\pi_{1}(M),\mathbb R) \neq 0}\). This conjecture is related to the question of whether the period map kills the second homotopy group or not. Using Gromov's \(h\)-principle, the authors also obtain the result: Let \((E,\theta)\) be a complex variation of the Hodge structure over a Kähler manifold and let \(D\) be the \(PU(p,q)\) period domain associated to \(E\). If \((E, \theta)\) does not satisfy the above hypotheses, then \(\pi_2(D)\) can be generated by superhorizontal 2-spheres.