Cubic threefolds, Fano surfaces and the monodromy of the Gauss map (Q5965193)

Consider a complex abelian variety \(A\) and a closed subvariety \(Z\). The author continues his investigations on the geometrical nature of a certain semisimple complex algebraic group \(G_Z\) which he and \textit{R. Weissauer} attached to the couple \((A,Z)\), cf. [J. Algebr. Geom. 24, No. 3, 531--568 (2015; Zbl 1338.14023)]. The construction was done by means of Tannakian techniques. These questions are asked here: In all previously known examples they are classical groups, and they are simply connected unless \(Z\) is a sum of two positive-dimensional subvarieties of \(A\) \( \dots\) (1) Are there examples where \(G_Z\) is an exceptional group? (2) If \(G_Z\) is not simply connected, then is its simply connected cover always realized as the Tannaka group of some other subvariety? (3) What sort of geometric information is encoded in these groups? By construction \(G_Z\) comes with a defining representation \(\omega_Z\) whose dimension is the Euler characteristic \(\chi\) of \( \delta_Z \), the perverse intersection cohomology sheaf supported on \(Z\). When \(A= J(V)\) is the intermediate jacobian of smooth cubic threefolds \(V\) the geometry of its theta divisor \(\Theta\) is well understood, and it is proved here that \(\chi ( \delta_\Theta )= 78\). Moreover \(\Theta\) is the image of the product \(S\times S\) by the difference map, where \(S \subset J(V) \) is the Fano surface of lines, and it is classically known \(\chi ( \delta_S) = 27\). Building on such data, the author proves that some facts from representation theory together with the Tannakian nature of the constructions yield then the following result: (i) \(G_\Theta \; \cong \; E_6(\mathbb C)/Z,\) where \(Z\) is the center of the simply connected complex algebraic group \(E_6(\mathbb C)\) (ii) \(G_S \; \cong \; E_6(\mathbb C)\). Now the Weyl group \(W(E_6)\) is known to be the symmetry group of the \(27\) lines on a cubic surface, and those lines are also fibres for the Gauss map of the Fano surface \(S\). It is indeed the case that the monodromy of the Gauss map from \(S\) coincides with \(W(E_6)\). This situation prompts the author to formulate a conjecture to the effect that in general for any smooth subvariety \(Z\hookrightarrow A\) the Weyl group \(W(G_Z)\) should contain the monodromy group of the natural Gauss map from \(Z\) as a subgroup of small index. This is checked to be the case for the situation of a curve in its jacobian and for the theta divisor in a general principally polarized abelian variety.