Projective structures and Hodge theory (Q7007618)

The paper under review is the written version of a talk given by G. P. Pirola at the XXII Congress of the \textit{Unione Matematica Italia} in Pisa in 2023. It is therefore expository in nature. It retraces several results, mostly due to Pirola and his collaborators over the last 25 years, on the differential geometry of the \textit{period map} \(j: \mathcal{M}_g \to \mathcal{A}_g\) that assigns to a compact Riemann surface \(X\) of genus \(g\) its Jacobian. The paper concerns \textit{projective structures} on compact Riemann surfaces. A projective structure on a Riemann surface is a complex atlas with linear fractional transformations as transition functions. The authors briefly review the main result of \textit{R. C. Gunning} [Math. Ann. 170, 67--86 (1967; Zbl 0144.33501)] according to which the set of projective structures on a compact Riemann surface \(X\) of genus \(g \geq 2\) is an affine space with associated vector space \(H^0 (X, 2 K_X)\); the latter is the space of quadratic differentials on \(X\), which can be interpreted as the cotangent space of the moduli space \(\mathcal{M}_g\) at the point \([X]\). It follows that there is an affine bundle \(\mathcal{P} \to \mathcal{M}_g\) whose fiber over \([X]\) is the set of projective structures on \(X\). The classical uniformization theorem provides a natural section \(p_u\) of this affine bundle, hence a projective structure on each Riemann surface \(X\).\N\NThe authors review the construction in [\textit{I. Biswas} et al., J. Math. Pures Appl. (9) 149, 1--27 (2021; Zbl 1476.14056)] of another such section \(p_h\) through the study of the second fundamental form of the period map. The paper concludes explaining why the two sections \(p_u\) and \(p_h\) do not coincide. The core of the argument is that such sections are in \(1\) to \(1\) correspondence with \(\bar{\partial}\)-closed forms of type \((1,1)\) having Dolbeault cohomology class equal to the Hodge class \(\lambda\); on the one hand, it is known that \(p_u\) corresponds to the Kähler form of the Weil-Petersson metric, which is everywhere positive definite, while \(p_h\) corresponds to the pull-back of the Siegel metric via the period map, which is not positive definite on the hyperelliptic locus.