The Ax-Schanuel conjecture for variations of Hodge structures (Q2419715)

The paper is an interesting sequel to the article [\textit{N. Mok}, \textit{J. Pila}, \textit{J. Tsimerman}, Ann. Math. (2) 189, No. 3, 945--978 (2019; Zbl 1481.14048)]. The Ax-Schanuel property is proven here for all smooth algebraic varieties \(X\) supporting a pure polarizable integral variation of Hodge structures. Consider the period domain \(D\) which parametrizes the corresponding Hodge structures, it sits as an open set in its compact dual \(\check{D}\) which has the structure of a projective variety. The Griffiths period map \(\phi: X \to \Gamma\backslash D\) lands in a quotient of \(D\) by the action of an arithmetic group. The version of the Ax-Schanuel conjecture, given here, states that atypical algebraic relations between \(X\) and \(D\) are controlled by the Hodge locus, this is a countable union of closed algebraic subvarieties in \(X\) by [\textit{E. Cattani} et al., J. Am. Math. Soc. 8, No. 2, 483--506 (1995; Zbl 0851.14004)]. To simplify matters, we assume that the monodromy mapping from \(\pi_1(X)\) to the authomorphisms of the Hodge structure has image whose Zariski closure contains the relevant full special orthogonal or symplectic group, but the proof holds for arbitrary Mumford-Tate domain as target. The Ax-Schanuel property says: Suppose that \(V \subset X \times \check{D}\) is algebraic. Write \(W\) for the image of \(\tilde{X}\) in \(X \times \check{D}\). If \(U \subset V \cap W\) is an irreducible analytic component such that \(\operatorname{codim}_{X \times \check{D}}(U) < \operatorname{codim}_{X \times \check{D}}(V) + \operatorname{codim}_{X \times \check{D}}(W),\) then the projection of \(U\) to \(X\) is contained in a proper ``weak Mumford-Tate'' algebraic subvariety. A consequence is the following transcendence property of period mappings: notation as above and let \(\Phi: \widetilde{X} \to \check{D}\) be the induced map, suppose that \(Z \subset \check{D}\) is an algebraic subvariety, if \( \operatorname{codim}_{\check{D}}(Z) \geqslant \dim(X) \) then any irreducible component of \(\Phi^{-1}(Z)\) is contained inside the preimage, in \(\widetilde{X}\), of the complex points of a proper subvariety of \(X\). In [``Diophantine problems and \(p\)-adic period mappings'', \url{arXiv:1807.02721}], \textit{B. Lawrence} and \textit{A. Venkatesh} remark that these statements provide one crucial ingredient for their arithmetical results. The authors note that their proof, based on o-minimal techniques, follows closely the strategy implemented in [loc. cit.]. Still new results are required here to cover the more general situation. First, a suitable fundamental domain in \(D\) is shown to exist, such as to to be definable in the relevant o-minimal structure and to satisfy convenient growth property. The given proof depends on Schmid's theorems on degenerations of Hodge structures. Second a volume bond on Griffiths transverse subvarieties is produced, the method of proof relies on the differential geometric properties of the period map.