Teichmüller discs with completely degenerate Kontsevich-Zorich spectrum (Q894375)

Let \(\mathcal M_g\) denote the moduli space of pairs \((X, \omega)\), where \(X\) is a compact genus-\(g\) complex curve and \(\omega\) is a holomorphic 1-form on \(X\). A point in \((X, \omega) \in \mathcal M_g\) yields (and in fact, is determined by) an atlas of charts to \(\mathbb C\) away from the zeros of \(\omega\), where the transition maps are translations, and hence, these objects are known as \textit{translation surfaces}. There is an \(\mathrm{SL}(2, \mathbb R)\)-action on this space, given by linear post composition with charts. The diagonal flow associated to this action induces a cocycle on the bundle with fibers \(H^1(X, \mathbb C)\) (to be precise, we consider the product bundle on the space of marked complex structures, and then quotient by the mapping class group). To each translation surface is associated a family of dynamical systems, the directional flows on the surface. The study of deviation exponents for this flow is related to the study of the exponents of growth of the cocycle, an observation due to Zorich, and made more precise by Forni. With respect to the natural \(\mathrm{SL}(2, \mathbb R)\) measure, Avila-Viana showed that the spectrum of exponents is simple (up to obvious symmetries coming from the symplectic structure). An open question is to understand the opposite situation: when is the spectrum completely degenerate? That is, when are all possible exponents zero? The paper under review studies closed \(\mathrm{SL}(2, \mathbb R)\) orbits (following McMullen, we call these closed complex geodesics) which have a more refined property: Let \(\mathcal D_g(1)\) denote the locus in \(\mathcal M_g\) where the derivative of the period matrix has rank one. This is a sufficient condition for degeneracy of the spectrum. Using a careful study of surfaces as they degenerate toward the boundary of \(\mathcal M_g\), the paper proves the following elegant (partially conditional) classification result: there are no complex closed geodesics in \(\mathcal D_2(1)\), unique complex closed geodesics in \(D_3(1)\) and \(D_4(1)\), and if there are no complex closed geodesics in \(D_5(1)\), then there are none in \(D_6(1)\).