Hodge theory and o-minimality at CIRM Gregorio Baldi Abstract . We discuss the relationship between o-minimality and the so called Zilber-Pink conjecture. Since the work of Pila and Zannier, algebraization theorems in o-minimal geometry had profound impacts in Diophantine geometry (most notably on the study of special points in abelian and Shimura varieties). We will first focus on functional transcendence, discussing various recent and spectacular Ax-Schanuel theorems, and the related geometric part of Zilber-Pink. Armed with these tools, we will study the distribution of the Hodge locus of an arbitrary variation of Hodge structures (the typical/atypical dichotomy) and present some recent applications. We will conclude by describing the algebraicity and quasiprojectivity of images of period maps. † † 2020 Mathematics Subject Classification . 14D07, 14C30, 14G35, 22F30, 03C64. † † Key words and phrases . Hodge theory and Mumford–Tate domains, Functional transcendence, Typical and atypical intersections, o-minimality. In memory of Nicola Gatti (1990–2023), whose boundless curiosity and passion for schemes would have made him a treasured reader. Preface 0.1. Hodge theory Compact Kähler manifolds are unique in that they are equipped with several compatible structures: complex, Riemannian, and symplectic. This rich interplay makes them fundamental objects in mathematics. Classical Hodge theory [ 39 ] enhances the classical topological invariants of the underlying space of a Kähler manifold X by incorporating a \mathbb{C}^{*} -action. Building on de Rham’s work, the key insight is that for any given Riemannian metric on X , every cohomology class has a canonical harmonic representative. Since complex differential forms on can be uniquely expressed as sums of forms of type (p,q) , the fact that the (p,q) components of a harmonic form remain harmonic gives rise to the celebrated Hodge decomposition. This decomposition splits the cohomology of X with complex coefficients into a direct sum of complex vector spaces: \qquad\qquad\qquad\qquad H^{n}(X,\mathbb{C})=\bigoplus_{p+q=n}H^{p,q}(X)\qquad\qquad(\textbf{HoDec}). This decomposition is elegantly encoded by a \mathbb{C}^{*} -action, where H^{p,q} is the subspace on which z\in\mathbb{C}^{*} acts as z^{p}\overline{z}^{q} . The symmetry under complex conjugation (namely \overline{H^{p,q}}=H^{q,p} ) reflects the fact that \mathbb{C}^{*} is viewed as a real algebraic torus (and usually denoted by \mathbb{S} ). Algebraic Geometry, on the other hand, studies algebraic varieties over complex numbers–the solution sets of algebraic equations with complex coefficients. The Kodaira embedding theorem characterizes smooth complex projective varieties among all compact Kähler manifolds and underpins the concept of a polarized Hodge structure, which grants favorable properties to the period domains of such objects that play a central role in Griffiths’ approach [ 35 , 34 ] to Hodge theory. Another key distinction between algebraic and general Kähler manifolds is that while general Kähler manifolds may not admit complex submanifolds, algebraic varieties possess them in abundance. For instance, an algebraic subvariety Z\subset X of codimension j (or more generally, a cycle) gives rise to an algebraic cohomology class [Z]\in H^{2k}(X,\mathbb{Q}) , whose image in H^{2k}(X,\mathbb{C}) lies in H^{k,k}(X) relatively to (HoDec). A rational cohomology class lying in some H^{k,k}(X) is called a Hodge class (equivalently those are classes fixed by the action of \mathbb{C}^{*} ). This suggests the tempting belief that the Hodge structure of a compact complex variety is governed by the geometry of its subvarieties, or more precisely, its Chow group. A special case of this belief is indeed one of the Clay millennium problems: Conjecture 0.1 (Hodge) . A class \lambda\in H^{*}(X,\mathbb{Q}) is algebraic if and only if it is Hodge. This conjecture remains a central open problem in mathematics, lying at the intersection of Complex and Algebraic Geometry. It provides a simple criterion to determine which topological cycles of a smooth projective algebraic variety can be represented by algebraic cycles. Finally Deligne [ 27 ] later vastly generalized Hodge’s results, showing that the cohomology of any complex algebraic variety (not necessarily compact) is functorially endowed with a structure called a mixed \mathbb{Z} -Hodge structure. 0.2. Cycles and Hodge classes in families: the Hodge locus Recent years have seen significant activity in understanding how algebraic cycles vary within families, following a path laid in the previous century by figures such as Max Noether, Lefschetz, Weil, Griffiths, Grothendieck, and Deligne, among others. For a smooth projective family of algebraic varieties f:\mathcal{X}\to S , the Hodge structures associated with each fiber \mathcal{X}_{s}:=f^{-1}(s) enrich the local system R^{*}f_{*}\mathbb{Z}_{\operatorname{prim}} of primitive cohomology with the structure of {\mathbb{V}} , a polarized variation of Hodge structures (VHS). Since Hodge theory fundamentally compares two distinct algebraic structures in a transcendental fashion, understanding how algebraic cycles and Hodge classes vary with s\in S(\mathbb{C}) is a challenging problem. This has led to the study of the Hodge locus, a subset of the base where the dimension (over \mathbb{Q} ) of the cohomology group H^{k,k}(\mathcal{X}_{s},\mathbb{Q}) , or possibly some tensor construction thereof, is larger than expected. In 1995, Cattani-Deligne-Kaplan [ 24 ] achieved a breakthrough in Hodge theory by proving that the aforementioned Hodge locus of (S,{\mathbb{V}}) is a countable union of algebraic subvarieties of S . More recently, this result has been established using methods from o-minimal geometry [ 9 ] . The Cattani-Deligne-Kaplan result is regarded by many as the most compelling evidence for the Hodge conjecture’s validity, alongside Deligne’s proof that Hodge classes on abelian varieties are absolute and various special cases such as the Lefschetz (1,1) theorem, and results by Schoen and Markman on Hodge-Weil 4-folds. The Hodge locus also plays a central role in conjectures like André-Oort and Zilber-Pink, which will be discussed in more detail later. A simple yet rich example, still far from being fully understood, involves hypersurfaces. A hypersurface X of degree d and dimension n is defined as the zero set of a homogeneous polynomial F of degree d in n+2 variables. It is smooth if the partial derivatives of F do not vanish simultaneously. For fixed n and d , the parameter space of hypersurfaces in \mathbb{CP}^{n+1} of degree d is simply the complex vector space of homogeneous polynomials of degree d , denoted by V=\mathbb{C}[X_{0},\dots,X_{n+1}]_{d} . However, the parameter space of smooth hypersurfaces U_{n,d}=V-\Delta is more intricate. The relevant VHS enriches the (primitive) cohomology H^{n}(X,\mathbb{Z}) , and the Hodge locus remains far from understood. In the case n=2 and d\geq 4 , this locus contains the Noether-Lefschetz locus , which can be defined purely using algebraic geometry: \NL_{d}:=\{[X]\in U_{2,d}:\operatorname{Pic}(\mathbb{CP}^{3})\to\operatorname{Pic}(X)\text{ is not an isomorphism}\}. For a surface X outside \NL_{d} , every curve on X has the pleasant and useful property that it is the complete intersection of X with another surface in \mathbb{CP}^{3} . This object has been the subject of many beautiful studies by Griffiths, Green, Voisin, Ciliberto, Harris, Miranda and others. The most pressing challenges in the field revolve around two fundamental questions: Question 0.2 . What can we say about about Hodge loci? What do Hodge loci reveal to us? 0.3. Organization of the Lecture Notes This is the second part of a lecture series on Hodge theory and o-minimality. The key object of study is the period map associated with the aforementioned variations of Hodge structures (VHS, from now on). The first part [ 5 ] , delivered by B. Bakker, covered the following topics: (1) Basic definitions and local definability of maps; (2) Definable structures on period spaces; (3) Global definability of period maps and the algebraicity of Hodge loci. The second part, consisting of four 1-hour lectures, discusses: (a) Functional transcendence and the Ax-Schanuel theorem for period maps; (b) Distribution of the Hodge locus; (c) Algebraicity and quasi-projectivity of images of period maps. We will build on and use the language of [ 5 ] , and refer to it whenever possible. Item (c) is a natural continuation of [ 5 ] , whereas (a) and (b) were not covered in op. cit. . There are already many excellent expository notes on this subject, so we have tried to focus on topics not already addressed in the following recent references: • Bakker, Tsimerman, Lectures on the Ax-Schanuel conjecture [ 11 ] ; • Fresán, Hodge theory and o-minimality (after Bakker, Brunebarbe, Klingler, Tsimerman) [ 31 ] ; • Tsimerman, Functional transcendence and arithmetic applications [ 58 ] ; • Klingler, Hodge theory between algebraicity and functional transcendence [ 44 ] . In particular: • Our exposition on functional transcendence follows the most recent advances (albeit this approach no longer relies on o-minimality), with original sources in [ 21 , 12 ] ; • We compile here most of the known results concerning the Zilber-Pink conjecture for variations of Hodge structures; • Each section includes several exercises of varying difficulty, which are collected in the final Section 5. References to Section 5.n.m indicate the m-th exercise in the n-th lecture. Finally, the notes cover roughly four hours of material and we tried to stick to the program exposed in Luminy. In particular the exercises play a fundamental role to complete the discussions and provide many complementary proofs that were not exposed during the lectures. Remark 0.3 . To conclude, we wish to mention also the beautiful monographs written by Zannier [ 61 ] and Pila [ 50 ] . Both discuss related results and focus mainly on the arithmetic side, rather the setting of VHS. Nevertheless, they are highly recommended to anyone who wants to learn more about unlikely intersections! 0.4. Acknowledgements These notes were prepared for a mini-course held during a Research School as part of the CIRM-SMF program (15–19 January, Luminy, France, 2024). We thank the organizers and participants for the opportunity to share this material and for their valuable feedbacks. During the final stages of this work, the author was partially supported by the grant ANR-HoLoDiRibey of the Agence Nationale de la Recherche. 1. Lecture 1: (Historical) Overview The goal of this section is to describe various instances of the Zilber-Pink philosophy (about the finiteness of atypical intersections ) and the so-called Pila-Zannier strategy . Historically, this strategy introduced the combination of o-minimality and functional transcendence that will underpin most of our lectures [ 53 , 49 ] . For additional insights, we refer to Scanlon’s exposition on o-minimality as an approach to the André-Oort conjecture [ 55 ] . As a roadmap, one might conceptualize “three stages” in the theory of atypical (also known as unlikely ) intersections, each of increasing complexity: (1) complex abelian varieties A or, more generally, semi-abelian varieties, (2) Shimura varieties \text{Sh}_{K}(G,X) , and (3) variations of Hodge structures (S,{\mathbb{V}}) . O-minimality plays a role in two critical ways: first, in the arithmetic aspects developed in (1) and extended to (2), and second, in providing a framework where the period map S\to\Gamma\backslash D can naturally exist. Remark 1.1 . The most general setting, which will not be discussed here, involves admissible graded-polarizable integral variations of mixed Hodge structures . For details, see [ 18 , Sec. 4] . The Zilber-Pink conjecture stated in op. cit. indeed implies the other cases discussed herein. 1.1. Some Number Theory In this section we first introduce the Zilber-Pink conjecture (ZP, henceforth) for the moduli space of principally polarized abelian varieties of dimension g , \mathcal{A}_{g} , which is the period space of weight one Hodge structures on the standard symplectic lattice [ 5 ] . ZP is a far-reaching generalization of the André-Oort (AO) and Mordell-Lang conjectures, which we now recall. A special case of AO can be summarized as follows [ 3 ] : Theorem 1.2 (André) . Let V be an irreducible algebraic curve in the complex affine plane, which is neither horizontal nor vertical. Then V is a modular curve Y_{0}(N) for some N>0 if and only if it contains infinitely many points (j^{\prime},j^{\prime\prime})\in\mathbb{C}^{2} such that j^{\prime} and j^{\prime\prime} are j -invariants of elliptic curves with complex multiplication. Here, an elliptic curve E with complex multiplication (CM) is an elliptic curve that has an endomorphism ring larger than the integers. The complex affine plane is the j -plane, Y_{0}(1)\cong\operatorname{SL}_{2}(\mathbb{Z})\backslash\mathbb{H} (cf. [ 5 ] ). The Mordell-Lang conjecture, proven by Faltings, includes the assertion that a smooth projective curve of genus >1 has only finitely many K -points over any number field K , as well as the Manin-Mumford conjecture which states the following [ 54 ] : Theorem 1.3 (Raynaud) . Let A be a complex abelian variety and V\subset A an irreducible subvariety. V is a torsion cosets a+B , where B is an abelian subvariety of A and a\in A is a torsion point if an only if it contains a Zariski dense set of torsion points. Although elementary proofs for these results exist, the one based on o-minimality is the most natural for generalizations to the aforementioned three stages. A key idea, explored in Section 5.1.1 and Section 5.1.2 , links these results to functional transcendence. Specifically, the proofs leverage the following: Theorem 1.4 (Ax-Lindemann, first cases) . We have: (1) Let \pi=(j,j):\mathbb{H}\times\mathbb{H}\to Y_{0}(1)\times Y_{0}(1) be the complex uniformization map. Let Y be an irreducible algebraic subvariety of \mathbb{P}^{1}\times\mathbb{P}^{1} . Each irreducible component of the Zariski closure of \pi(Y\cap\mathbb{H}\times\mathbb{H}) is either a point, Y_{0}(1)^{2} , an horizontal or vertical line, Y_{0}(N) for some N . (2) Let \pi:\mathbb{C}^{n}\to A be the complex uniformization map. Let Y be an irreducible algebraic subvariety of \mathbb{C}^{n} . Then each component of the Zariski closure of \pi(Y) in A is a coset a+B , where B is an abelian subvariety of A and some a\in A(\mathbb{C}) . These examples suggest why Hodge theory enters the picture: the first involves a family of elliptic curves, and the second concerns an abelian variety, corresponding to a polarized Hodge structure of type (h^{1,0}=g,h^{0,1}=g) . One further example brings us closer to the full conjectural framework. First of all, an irreducible component of a Shimura subvariety of \mathcal{A}_{g} , or of its image under a Hecke operator, is called a special subvariety of \mathcal{A}_{g} . Let S be a fixed irreducible subvariety of \mathcal{A}_{g} . An irreducible subvariety Y\subset S is called special (with respect to S ) if it is a component of S\cap Sh , where Sh is a special subvariety of \mathcal{A}_{g} . If Sh_{Y} is the smallest special subvariety of \mathcal{A}_{g} containing Y (the so called “special closure”), we call Y atypical if: \operatorname{codim}_{\mathcal{A}_{g}}(Y)<\operatorname{codim}_{\mathcal{A}_{g}}(S)+\operatorname{codim}_{\mathcal{A}_{g}}(Sh_{Y}). Conjecture 1.5 (Zilber-Pink for \mathcal{A}_{g} ) . S contains only finitely many maximal special atypical subvarieties. The text then continues with deeper explorations into these conjectures, their applications, and broader philosophical principles, particularly in connection to variations of Hodge structures and period maps. 1.2. General Expectations Before dealing with the general case of VHS, it is worth pausing for a second to reflect on the abstract slogan given by 1.2 , 1.3 , and 1.5 that we discussed so far. We are given a manifold M which naturally comes with a distinguished class of submanifolds \{M_{i}\}_{i\in I} exhibiting a special behaviour. Zilber-Pink-type results in such settings suggest distribution behaviour of the intersections between a fixed sub-manifold S\subset M and the M_{i} s can often be understood by simple dimension criteria. In particular, one says that a component Y of S\cap M_{i} is atypical if (1.2.1) \operatorname{codim}_{M}(Y)<\operatorname{codim}_{M}(S)+\operatorname{codim}_{M}(M_{i}), and typical otherwise. One then expects that: (1) S contains only finitely many maximal atypical intersections. (2) The following are equivalent: (i) S contains one typical intersection; (ii) the collection of typical intersections is dense in S ; and (iii) there exists an i\in I such that: \dim S-\operatorname{codim}_{M}(M_{i})\geq 0. Condition (iii) is simply asserting that a typical intersection could exists for combinatorial reasons. 1.3. Hodge theory and period maps There is a version of Zilber-Pink (that recovers 1.5 as a special case) for arbitrary families of varieties/motives. This is given by using the formalism of variation of Hodge structures and their associated period maps, detailed in [ 5 , Section 4] as well as in the monograph [ 34 ] . We briefly recall what we need: • Let (S,{\mathbb{V}}) be a pure polarized integral VHS on a smooth quasi-projective variety S ; • It corresponds to a period map \Phi:S^{\textnormal{an}}\longrightarrow\Gamma\backslash D, in the category of complex analytic manifolds ( \Gamma\backslash D is in general not algebraic); • The period domain D is is canonically an open subset (in the euclidean topology) of an algebraic variety D^{\vee} ; • \Gamma\backslash D comes with a collection of special submanifolds , the so called sub-Mumford-Tate domains; • The Hodge Locus of (S,{\mathbb{V}}) is the subset of S given by the collection of preimages of the sub-Mumford-Tate domains along \Phi . We will go back to this setting in Section 2 , but we can already observe that we are very well in the setting described in the previous section and that Zilber-Pink type conjectures can be at least formulated. (See Section 5.1.4 for a link with the AO problem.) Before continuing on this topic, we would like to conclude this historical overview with a couple of applications of this circle of ideas to related areas. 1.4. Some applications 1.4.1. Integral points The first theorem is due to Lawrence and Venkatesh [ 45 ] and in fact was part of the motivation of Bakker and Tsimerman to prove their Ax-Schanuel conjecture (a broad generalization of 1.4 that will be discussed in details in Section 3 ). As in the Preface, let U_{n,d} denote the parameter space of smooth degree d hypersurfaces in the n+1 -dimensional projective space. Theorem 1.6 ( [ 45 , Thm 10.1, Prop. 10.2] ) . There exist n_{0}\in\mathbb{N}_{\geq 3} and a function d_{0}:\mathbb{N}\to\mathbb{N} such that, (1.4.1) \textnormal{ for every}\;n\geq n_{0}\text{ and }d\geq d_{0}(n), the set U_{n,d}(\mathbb{Z}[S^{-1}]) is not Zariski dense in U_{n,d} , for every finite set of primes S . See also [ 40 ] for an effective version of ( 1.4.1 ). Remark 1.7 . In brief (and ignoring many crucial details), the idea behind the above result is the following (closely related to the Kim-Chabauty method). If we have a (immersive, say) period map \Phi:S\to\Gamma\backslash D , we we want to show that the S -integral points S(\mathbb{Z}[S^{-1}]) are not Zariski dense in S . Working p -adically, there’s a lift of the period map \Phi_{p}:S(\mathbb{Q}_{p})\to D(\mathbb{Q}_{p}) and, using some results from p -adic Hodge theory, the authors can show that \Phi_{p}(S(\mathbb{Z}[S^{-1}])) lies in an algebraic subvariety Y of D^{\vee} . At this point, a p -adic version of Ax-Schanuel implies that \Phi_{p}^{-1}(Y) can not be Zariski dense in S (and in fact it has to lie in the Hodge locus). In particular S(\mathbb{Z}[S^{-1}]) cannot be Zariski dense. Consider f_{n,d}:X_{n,d}\to U_{n,d} the universal family of smooth degree d hypersurfaces in \mathbb{P}^{n+1} . We denote by {\mathbb{V}} the polarized \mathbb{Z} -variation of Hodge structure (R^{n}{f^{\textnormal{an}}_{n,d,\mathbb{C}}}_{*}\mathbb{Z})_{\textnormal{prim}} on U_{n,d,\mathbb{C}} and by \Phi:{U_{n,d,}}^{\textnormal{an}}\to\Gamma\backslash D the associated period map. An irreducible algebraic subvariety Y\subset S is said to be of positive period dimension if \Phi(Y_{\mathbb{C}}^{\textnormal{an}}) has positive dimension. As the above remark outlined, 1.6 is already employing tools that we will discover in our lecture. In fact, these viewpoints are even closer than one might expect. For example, in [ 14 ] , using the Zilber-Pink philosophy, the authors prove: Theorem 1.8 . As long as ( 1.4.1 ) is satisfied, there exists a closed strict subscheme E\subset U_{n,d} such that, for all finite set of primes S , we have \overline{U_{n,d}(\mathbb{Z}[S^{-1}])}_{\textnormal{pos}}\subset E, where \overline{U_{n,d}(\mathbb{Z}[S^{-1}])}_{\textnormal{pos}} denotes the union of the irreducible components of the Zariski closure of U_{n,d}(\mathbb{Z}[S^{-1}]) in U_{n,d} of positive period dimension. That is: the Zariski closure of U_{n,d}(\mathbb{Z}[S^{-1}])-E(\mathbb{Z}[S^{-1}]) has period dimension zero. In a nutshell, the improvement comes from understanding whether the Hodge locus (of positive period dimension) is Zariski closed or not, which brings us back to 0.2 . 1.4.2. Jacobians with prescribed properties We recall the following folklore question. Question 1.9 . Let g\geq 4 , and let A be a generic abelian variety over an algebraically closed field k . What is the lowest integer g^{\prime} such that there exists a Jacobian J of dimension g^{\prime} and a surjection J\to A ? In characteristic zero, the first results were obtained by Chai-Oort, Tsimerman, and after in a more general setting by Masser-Zannier [ 46 ] . For simplicity we just cite the most recent and general one, due to Tsimerman [ 59 ] (see also references therein): Theorem 1.10 . For any two integers g\geq 4 and g^{\prime}\leq 2g-1 , there exist g -dimensional abelian varieties over \overline{\mathbb{Q}} which are not quotients of a Jacobian of dimension g^{\prime} . All known proofs use o-minimality and are inspired once more by the Pila-Zannier strategy (that we discussed in the exercises). On the other hand, Mumford [ 48 ] shows that there exist principally polarized abelian 4folds A with trivial endomorphism ring, that are not Hodge generic in \mathcal{A}_{4} (they have an exceptional Hodge class in H^{4}(A^{2},\mathbb{Z}) ). A question often attributed to Serre is to describe “as explicitly as possible” such abelian varieties of Mumford’s type . The most satisfying way would be to show the existence of a smooth projective curve over \overline{\mathbb{Q}} of genus 4 , whose Jacobian is of Mumford’s type. The Zilber-Pink philosophy is the crucial ingredient behind the following [ 15 , Thm. 3.17] , as we will explore in Section 5.3.9 : Theorem 1.11 . There exists a smooth projective curve C/\overline{\mathbb{Q}} of genus 4 whose Jacobian is of Mumford’s type, i.e. it has Mumford-Tate group isogenous to a \mathbb{Q} -form of the complex group \mathbb{G}_{\textnormal{m}}\times\Sl_{2}\times\Sl_{2}\times\Sl_{2} . In fact such curves are dense in the moduli space \mathcal{M}_{4} . For more about the Hodge locus of \mathcal{M}_{g} and the Zilber-Pink conjecture in this special case, we refer to the forthcoming survey [ 13 ] . 2. Lecture 2: Distribution of the Hodge locus After the historical overview and the motivations discussed from number theoretic problems, we now focus solely on the complex analytic side. We begin with some general notation and conventions. Here an algebraic variety S is a reduced scheme of finite type over the field of complex numbers, but may be reducible. If S is an algebraic (resp. analytic) variety, by a subvariety Y\subset S we always mean a closed algebraic (resp. analytic) subvariety. The smooth locus of Y is denoted by Y^{\operatorname{sm}} . A \mathbb{Q} -Hodge structure of weight n on a finite dimensional \mathbb{Q} -vector space V_{\mathbb{Q}} is a decreasing filtration F^{\bullet} on the complexification V_{\mathbb{C}} such that V_{\mathbb{C}}=\bigoplus_{p\in\mathbb{Z}}F^{p}\oplus\overline{F^{n-p}}. The category of pure \mathbb{Q} -Hodge-structures is Tannakian (and semisimple if we consider polarizable ones, as we will usually do). The Mumford–Tate group \mathbf{MT}(V)\subset\mathbf{GL}(V) of a \mathbb{Q} -Hodge structure V is the Tannakian group of the Tannakian subcategory \langle V\rangle^{\otimes} of \mathbb{Q} -Hodge structures generated by V . Equivalently, \mathbf{MT}(V) is the smallest \mathbb{Q} -algebraic subgroup of \mathbf{GL}(V) whose base-change to \mathbb{R} contains the image of h:\mathbb{S}\to\mathbf{GL}(V_{\mathbb{R}}) . It is also the stabilizer in \mathbf{GL}(V) of the Hodge tensors for V . As V is polarised, this is a reductive algebraic group. Cf. Section 5.2.1 . 2.1. Typical and atypical intersections Recall that, from Section 1.3 and [ 5 ] , to understand a VHS (S,{\mathbb{V}}) , we consider the associated holomorphic period map (2.1.1) \Phi:S^{\textnormal{an}}\longrightarrow\Gamma\backslash D,\ \ s\mapsto[{\mathbb{V}}_{s}] which completely describes {\mathbb{V}} . Here ({\mathbf{G}},D) denotes the generic Hodge datum of {\mathbb{V}} and \Gamma\backslash D is the associated Hodge variety. The Mumford--Tate domain 1 1 1 It is worth elucidating an important and perhaps confusing difference in the literature between period domains and Mumford-Tate domains. In [ 5 ] , the period domains are defined as orbits under the group of automorphism of a polarized Hodge structure (H_{\mathbb{Q}},q_{\mathbb{Q}}) , where q_{\mathbb{Q}} is a (-1)^{\operatorname{weight}(H)} symmetric bilinear form. However, a VHS can actually have extra fixed Hodge tensors. The simplest example is a family of squares of elliptic curves. The term MT domain, refers to the fact that we always normalise this issue by taking as ambient group \mathbf{G} the generic Mumford-Tate group of the VHS (which is nothing but the MT of {\mathbb{V}}_{s} for a very general point s\in S ). In particular \mathbf{G}(\mathbb{R}) can be more general than an orthogonal or a symplectic group. This point will implicitly come back when dealing with weakly special subvarieties , cf. Section 3 D decomposes as a product D_{1}\times\cdots\times D_{k} , according to the decomposition of the adjoint group {\mathbf{G}}^{\textnormal{ad}} into a product {\mathbf{G}}_{1}\times\cdots\times{\mathbf{G}}_{k} of simple factors (notice that some factors {\mathbf{G}}_{i} may be \mathbb{R} -anisotropic). Replacing S by a finite étale cover and reordering the factors if necessary, the lattice \Gamma\subset{\mathbf{G}}^{\textnormal{ad}}(\mathbb{R})^{+} decomposes as a direct product \Gamma\cong\Gamma_{1}\times\cdots\times\Gamma_{r} , r\leq k , where \Gamma_{i}\subset{\mathbf{G}}_{i}(\mathbb{R})^{+} is an arithmetic lattice for each i , and thus is Zariski-dense in {\mathbf{G}}_{i} . Writing D^{\prime}=D_{r+1}\times\cdots\times D_{k} for the product of factors where the monodromy is trivial (in particular, D^{\prime} contains in particular all the factors D_{i} for which {\mathbf{G}}_{i} is \mathbb{R} -anisotropic), the period map is written (2.1.2) \Phi:S^{\textnormal{an}}\longrightarrow\Gamma\backslash D\cong\Gamma_{1}\backslash D_{1}\times\cdots\times\Gamma_{r}\backslash D_{r}\times D^{\prime}\;\;, and the projection of \Phi(S^{\textnormal{an}}) on D^{\prime} has image a point. See also the structure theorem for VHS from [ 34 , (III.A.2)] . We distinguish between special subvarieties Z of zero period dimension (i.e., \Phi(Z^{\textnormal{an}}) is a point of \Gamma\backslash D ), which are geometrically elusive; and those of positive period dimension (i.e., \dim_{\mathbb{C}}\Phi(Z^{\textnormal{an}})>0 ). Definition 2.1 . (1) A subvariety Z of S is said to be of positive period dimension for {\mathbb{V}} if \Phi(Z^{\textnormal{an}}) has positive dimension. (2) The Hodge locus of positive period dimension \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{pos}} is the union of the special subvarieties of S for which {\mathbb{V}} has positive period dimension. The Cattani-Deligne-Kaplan theorem [ 24 ] recalled in the introduction and discussed in [ 5 ] equips the Hodge locus with a structure of a countable (possibly finite) union of subvarieties of S . Using period maps, special subvarieties can also be defined as intersection loci . Indeed, a closed irreducible subvariety Z\subset S is special for {\mathbb{V}} (we will equivalently say that it is special for \Phi ) precisely when Z^{\textnormal{an}} coincides with an analytic irreducible component \Phi^{-1}(\Gamma^{\prime}\backslash D^{\prime})^{0} of \Phi^{-1}(\Gamma^{\prime}\backslash D^{\prime}) , for ({\mathbf{G}}^{\prime},D^{\prime})\subset({\mathbf{G}},D) the generic Hodge sub-datum of Z and \Gamma^{\prime}\backslash D^{\prime}\subset\Gamma\backslash D the associated Mumford-Tate subdomain. Definition 2.2 . Let Z=\Phi^{-1}(\Gamma^{\prime}\backslash D^{\prime})^{0}\subset S be a special subvariety for {\mathbb{V}} with generic Hodge datum ({\mathbf{G}}^{\prime},D^{\prime}) . Then Z is said to be atypical if \Phi(S^{\textnormal{an}}) and \Gamma^{\prime}\backslash D^{\prime} do not intersect generically along \Phi(Z) . That is, Z is atypical when (2.1.3) \operatorname{codim}_{\Gamma\backslash D}\Phi(Z^{\textnormal{an}})<\operatorname{codim}_{\Gamma\backslash D}\Phi(S^{\textnormal{an}})+\operatorname{codim}_{\Gamma\backslash D}\Gamma^{\prime}\backslash D^{\prime}\;\;. Otherwise Z is said to be typical . The atypical Hodge locus \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{atyp}}\subset\textnormal{HL}(S,{\mathbb{V}}^{\otimes}) (resp. the typical Hodge locus \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{typ}}\subset\textnormal{HL}(S,{\mathbb{V}}^{\otimes}) ) is the union of the atypical (resp. strict typical) special subvarieties of S for {\mathbb{V}} . The Hodge locus is the recovered by the union of its typical and atypical parts. 2.2. Conjectures Let {\mathbb{V}} be a polarizable \mathbb{Z} VHS on an irreducible smooth quasi-projective variety S , from [ 15 ] (which refines [ 43 ] , as well as the discussion from Section 1 ) we expect: Conjecture 2.3 (Zilber–Pink conjecture for the atypical Hodge locus, strong version) . The atypical Hodge locus \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{atyp}} is a finite union of atypical special subvarieties of S for {\mathbb{V}} . Remark 2.4 . The above conjecture is a vast generalization of the case of family of abelian varieties that we discussed before ( 1.5 ), as well as the notable study of the Noether-Lefschetz locus (cf. Section 5.2.7 ). It is worth mentioning that these two examples are the cases of level 1 and 2 (a refinement of the weight of the Hodge structure, cf [ 15 , Sec. 4.6] ). In the literature there is also a case of study in level 3, but very little (nothing, to the best of my knowledge) was speculated in level >3. The level 3 case, is intimately related to String theory. Indeed their interest in Calabi-Yau 3-manifolds stems from their connection to conformal field theories (CFTs). Gukov and Vafa [ 37 ] posed a question regarding the existence of infinitely many Calabi-Yau manifolds with complex multiplication of a fixed dimension, motivated by the connection to rational conformal field theories (RCFT), introduced earlier by Friedan-Qiu-Shenker. Around the same time, Moore explored the arithmetic-string theory connection, particularly the role of attractor varieties in black hole constructions within IIB string theory. These investigations revealed connections to atypical intersections that are not necessarily of CM type. For more recent results, see also [ 23 ] . Conjecture 2.5 (Density of the typical Hodge locus) . If \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{typ}} is not empty then it is dense (for the analytic topology) in S . At first sight, these conjectures might look formal and might be hard to grasp their meaning. We refer to Section 5.2 for a collection of exercises on special and concrete cases. 2.3. Results–atypical intersections As above, we let (S,{\mathbb{V}}) be a polarizable \mathbb{Z} VHS on a smooth connected complex quasi-projective variety S , and we denote by ({\mathbf{G}},D) the generic Hodge datum. In Section 3 describe a proof of the following [ 15 , Thm. 3.1] : Theorem 2.6 (Geometric Zilber–Pink) . Let Z be an irreducible component of the Zariski closure of the union of the atypical special subvarieties of positive period dimension in S . Then either (a) Z is a maximal atypical special subvariety, or (b) the adjoint Mumford–Tate group {\mathbf{G}}_{Z}^{\textnormal{ad}} decomposes as a nontrivial product {\mathbf{H}}^{\textnormal{ad}}_{Z}\times{\mathbf{L}}_{Z} , Z contains a Zariski-dense set of fibers of \Phi_{{\mathbf{L}}_{Z}} which are atypical weakly special subvarieties of S for \Phi , where (possibly up to an étale covering) \Phi_{|Z^{\textnormal{an}}}=(\Phi_{{\mathbf{H}}_{Z}},\Phi_{{\mathbf{L}}_{Z}}):Z^{\textnormal{an}}\longrightarrow\Gamma_{{\mathbf{G}}_{Z}}\backslash D_{G_{Z}}=\Gamma_{{\mathbf{H}}_{Z}}\backslash D_{H_{Z}}\times\Gamma_{{\mathbf{L}}_{Z}}\backslash D_{L_{Z}}\subset\Gamma\backslash D, and Z is Hodge generic in a special subvariety \Phi^{-1}(\Gamma_{{\mathbf{G}}_{Z}}\backslash D_{G_{Z}})^{0} of S for \Phi which is monodromically typical (and therefore typical). 2.4. Results–typical intersections We first describe two results that don’t use functional transcendence (whose proofs can be found in [ 15 , Sec. 7 and 10] ). Theorem 2.7 . If the typical Hodge locus \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{typ},\textnormal{pos}} is nonempty then \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{typ},\textnormal{pos}} is analytically (hence Zariski) dense in S . Theorem 2.8 . Let {\mathbb{V}} be a polarizable \mathbb{Z} VHS on a smooth connected complex quasi-projective variety S , with generic Hodge datum ({\mathbf{G}},D) and algebraic monodromy group {\mathbf{H}} . Suppose that {\mathbf{H}}={\mathbf{G}}^{\textnormal{der}} . If {\mathbb{V}} is of level at least 3 then \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{typ}}=\emptyset (and thus \textnormal{HL}(S,{\mathbb{V}}^{\otimes})=\textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{atyp}} ). Remark 2.9 . The above theorem roughly states that outside the case of surfaces and abelian varieties, every special subvariety is an atypical intersection. This is of course an oversimplification of the concept of level. The proof of the theorem rests upon the following representation theoretical result which may be helpful to state, see [ 15 , Prop. 7.5] ). Suppose \mathfrak{g}_{\mathbb{R}} is a simple \mathbb{R} -Hodge-Lie algebra generated in level 1 and of level at least 3 (i.e.. \mathfrak{g}^{2}=[\mathfrak{g}^{1},\mathfrak{g}^{1}] , \mathfrak{g}^{3}=[\mathfrak{g}^{1},\mathfrak{g}^{2}] and so on, and \mathfrak{g}^{3}\neq 0 ). If \mathfrak{g}^{\prime}_{\mathbb{R}}\subset\mathfrak{g}_{\mathbb{R}} is an \mathbb{R} -Hodge-Lie subalgebra satisfying {\mathfrak{g}^{\prime}}^{i}=\mathfrak{g}^{i} for all |i|\geq 2 then \mathfrak{g}^{\prime}=\mathfrak{g} . We record here two results that use once more some functional transcendence. Hodge theory actually gives a simple combinatorial criterion to decide whether \textnormal{HL}(S,{\mathbb{V}}^{\otimes})_{\textnormal{typ}} is empty or not. Indeed see the recent works of Eterovic-Scanlon, Khelifa-Urbanik [ 29 , 42 ] . They defined: Definition 2.10 . A strict Hodge sub-datum (\mathbf{H},D_{H})\subset(\mathbf{G},D_{G}) is said {\mathbb{V}} - admissible if \dim\Phi(S^{\operatorname{an}})+\dim D_{M}\geq\dim D. Theorem 2.11 (Eterovic-Scanlon, Khelifa-Urbanik) . If (\mathbf{H},D_{H})\subset(\mathbf{G},D_{G}) is a {\mathbb{V}} -admissible Hodge sub-datum, then \textnormal{HL}(S,{\mathbb{V}}^{\otimes},\mathbf{H}):=\{s\in S:\exists g\in\mathbf{G}(\mathbb{Q})^{+},\mathbf{MT}({\mathbb{V}}_{s})\subset g\mathbf{H}g^{-1}\} is dense in S^{\operatorname{an}} . (We will come back to this interesting statement in Section 5.3.5 .) On a related direction, there is another result (taken from [ 16 ] ), which builds on the geometric Zilber-Pink to characterise VHS from their Hodge Locus (when possible). Definition 2.12 . Let \mathbb{V}_{1},{\mathbb{V}}_{2}\in\mathbb{Z}\text{VHS}/S . We say that {\mathbb{V}}_{1} and {\mathbb{V}}_{2} are isogenous if there is an equivalence of tensor categories \langle{\mathbb{V}}_{1,\mathbb{Q}}\rangle^{\otimes}\cong\langle{\mathbb{V}}_{2,\mathbb{Q}}\rangle^{\otimes} , where \langle{\mathbb{V}}_{i,\mathbb{Q}}\rangle^{\otimes} denotes the smallest Tannakian subcategory of \mathbb{Q} VHS’s containing {\mathbb{V}}_{i,\mathbb{Q}} . Of course the Tannakian categories \langle{\mathbb{V}}_{i,\mathbb{Q}}\rangle^{\otimes} appearing above are equivalent (as tensor categories) to the category of finite dimensional representations of their generic Mumford-Tate group. (The equivalence is realised by the functor of \otimes -automorphisms of the fiber functor). It can happen that two VHS have isomorphic Mumford-Tate groups, but the isomorphism does not induce an equivalence of tenor categories. cf. Def. 1.10 and Thm. 2.11 in the article of Deligne and Milne [ 28 ] . Remark 2.13 . If two complex principally polarized abelian varieties A,B that are isogenous in the usual sense (either via a polarized or an unpolarized isogeny), then the Hodge structures H^{1}(A,\mathbb{Z}),H^{1}(B,\mathbb{Z}) are also isogenous in the sense of 2.12 (here we take as base S the spectrum of \mathbb{C} ). However the converse is in general not true, for example H^{1}(A,\mathbb{Z}) is isogenous to H^{1}(A\times A,\mathbb{Z}) . However, for two principally polarized g -dimensional abelian varieties with Mumford–Tate group \mathbf{GSp}_{2g} the two notions agree, since this is the case when the Mumford–Tate group is as big as possible. Theorem 2.14 ( [ 16 , Thm. 1.18] ) . Let S be a smooth quasi-projective variety and \mathbb{V}_{1},\mathbb{V}_{2} two pure polarized \mathbb{Z} VHSs on S . Assume that the generic Mumford–Tate groups of {\mathbb{V}}_{1} and {\mathbb{V}}_{2} are \mathbb{Q} -simple. If \textnormal{HL}(S,\mathbb{V}_{1}^{\otimes})_{\textnormal{pos},\textnormal{typ}}=\textnormal{HL}(S,\mathbb{V}_{2}^{\otimes})_{\textnormal{pos},\textnormal{typ}}\neq\emptyset, then \mathbb{V}_{1} is isogenous to \mathbb{V}_{2} . As a consequence, \textnormal{HL}(S,\mathbb{V}_{1}^{\otimes})=\textnormal{HL}(S,\mathbb{V}_{2}^{\otimes}) . 3. Lecture 3: Ax-Schanuel and applications We are finally ready to discuss the topic of functional transcendence and the main theorem of the theory: Ax-Schanuel. It will provide a vast generalization of 1.4 . 3.1. Statements Let S be a smooth complex quasi-projective variety supporting a polarizable \mathbb{Z} VHS \mathbb{V} , with generic Hodge datum ({\mathbf{G}},D) , algebraic monodromy group {\mathbf{H}} , and period map \Phi:S\to\Gamma_{{\mathbf{H}}}\backslash D_{H}\subset\Gamma\backslash D . Without loss of generality we can assume, replacing if necessary S by a finite étale cover, that \Gamma is torsion-free. Let \tilde{\Phi}:\widetilde{S^{\textnormal{an}}}\to D_{H} be the lift of \Phi at the level of universal coverings. As already discussed, the domain D_{H} is canonically embedded as an open complex analytic real semi-algebraic subset in a flag variety D_{H}^{\vee} (called its compact dual ). For example, let \mathcal{L}=\mathcal{L}(V_{\mathbb{Q}},\varphi) denote the symplectic Grassmanian of Lagrangian, i.e. the scheme parametrizing maximal isotropic subspaces of V_{\mathbb{Q}} with respect to the form \varphi . The map \mathbb{H}_{g}\to\mathcal{L}(\mathbb{C}) that sends a Hodge structure y\in\mathbb{H}_{g} to the corresponding Hodge filtration \operatorname{Fil}^{0}\subset V_{\mathbb{C}} is the open immersion (the Borel embedding) of the g -dimensional Siegel space into its compact dual. We define an irreducible algebraic subvariety of D_{H} (resp. S\times D_{H} ) as a complex analytic irreducible component of the intersection of an algebraic subvariety of D_{H}^{\vee} (resp. S\times D_{H}^{\vee} ) with D_{H} (resp. S\times D_{H} ). We have already introduce the special subvarieties of (S,{\mathbb{V}}) in this context but in the sequel, and in general when ignoring the arithmetic side of the story, there is a better suited concept of weakly special subvarieties. Definition 3.1 . The weakly special subvarieties of (S,{\mathbb{V}}) are the closed irreducible algebraic subvarieties Y\subset S maximal among the closed irreducible algebraic subvarieties Z of S whose algebraic monodromy group \mathbf{H}_{Z} with respect to {\mathbb{V}} equals \mathbf{H}_{Y} . Any special subvariety (defined in Section 2 ) is weakly special but the converse does not hold. The reader might wish to consult [ 15 , Sec. 4.4] for more details, and properties, as well as Section 5.3.1 and Section 5.3.2 (which gives the basis to understand the link between the different notions). The following result is the so called Ax-Schanuel Theorem (for pure variations of Hodge structures). It was conjectured by B. Klingler [ 43 , Conj. 7.5] and later proved by Bakker and Tsimerman [ 10 , Thm. 1.1] , generalising the work of Mok, Pila and Tsimerman [ 47 , Thm. 1.1] from level one to arbitrary levels. Theorem 3.2 (Ax-Schanuel for VHS) . Let W\subset S\times D_{H} be an algebraic subvariety. Let U be an irreducible complex analytic component of W\cap S\times_{\Gamma_{\mathbf{H}}\backslash D_{H}}D_{H} such that \operatorname{codim}_{S\times D_{H}}U<\operatorname{codim}_{S\times D_{H}}W+\operatorname{codim}_{S\times D_{H}}(S\times_{\Gamma_{{\mathbf{H}}}\backslash D_{H}}D_{H})\;\;. Then the projection of U to S is contained in a strict weakly special subvariety of S for {\mathbb{V}} . Notice that S\times_{\Gamma_{{\mathbf{H}}}\backslash D_{H}}D_{H} is simply the image of the graph of \tilde{\Phi}:\widetilde{S^{\textnormal{an}}}\to D_{H} under \widetilde{S^{\textnormal{an}}}\times D_{H}\to S\times D_{H} . Remark 3.3 . The intersection W\cap S\times_{\Gamma_{\mathbf{H}}\backslash D_{H}}D_{H} can be identified with the intersection in S\times D_{H} between W and the the image of \widetilde{S^{\textnormal{an}}} in S\times D_{H} along the map (\pi,\tilde{\Phi}) . A special case might be already interesting to observe: Corollary 3.4 (Special case of AS in the period domain, Bakker-Tsimerman) . Let Y\subset D^{\vee} be an algebraic subvariety of codimension \geq\dim S and denote by Y_{\Gamma} the image of Y\cap D in \Gamma\backslash D . Each component of \Phi^{-1}(Y_{\Gamma}) lies in a strict weakly special subvariety of S . In the conclusion, in particular, we observe that \Phi^{-1}(Y_{\Gamma}) is not Zariski dense in S . 3.2. Proof of the geometric Zilber-Pink conjecture All finiteness results described in the lectures come from the following: the only countable and definable sets (in some o-minimal structure over \mathbb{R} ), are finite sets. Behind the proof 2.6 there are two basic observations: (1) Ax-Schanuel 3.2 gives the countability; (2) The theory of degenerations of VHS gives the definability. Remark 3.5 . Different approaches can be found in [ 20 , 25 , 18 ] . They still use (1) from above, but not (2). Nevertheless the key idea is similar and boils down to the fact that a constructible set which is a countable union of algebraic subvarieties can be expressed as a finite union. 3.2.1. Definable fundamental sets, in Hodge theory Let \Phi:S\to\Gamma_{\mathbf{H}}\backslash D_{H} be the period map for {\mathbb{V}} . As recalled, the Mumford-Tate domain D_{H}=H(\mathbb{R})^{+}/M is a real semi-algebraic open subset of its compact dual D_{H}^{\vee} . Recall also that S it is naturally definable in any extension of the o-minimal structure \mathbb{R}_{\operatorname{alg}} , since it is an algebraic variety. From now on, definable will be always understood in the o-minimal structure \mathbb{R}_{\textnormal{an},\exp} . Let us introduce a notion of “definable fundamental set” of S for \Phi , arguing as at the beginning of [ 10 , Section 3] . Let (\overline{S},E) be a log-smooth compactification of S , and choose a definable atlas of S by finitely many polydisks \Delta^{k}\times(\Delta^{*})^{\ell} . Let \exp:\Delta^{k}\times\mathbb{H}^{\ell}\to\Delta^{k}\times(\Delta^{*})^{\ell} be the standard universal cover, and choose \Sigma=[-b,b]\times[1,+\infty[\subset\mathbb{H} such that \Delta^{k}\times\Sigma^{\ell} is a fundamental set for the \mathbb{Z}^{\ell} -action by covering transformation. Let \mathcal{F} be the disjoint union of \Delta^{k}\times\Sigma^{\ell} over all charts and choose lifts \Delta^{k}\times\mathbb{H}^{\ell}\to D of the period map restricted to each chart to obtain a lift \tilde{\Phi}=\tilde{\Phi}_{\mathcal{F}}:\mathcal{F}\to D_{H} . From the Nilpotent Orbit Theorem [ 56 , (4.12)] we obtain the following diagram in the category of definable complex manifolds: \mathcal{F} D_{H} S^{\textnormal{an}} \tilde{\Phi}_{\mathcal{F}} \exp . 3.3. Main steps of the proof Let \Omega be a fixed finite set of representatives for the H(\mathbb{R})^{+} -conjugacy classes of semisimple algebraic subgroups of H_{\mathbb{R}} that are without compact factors. Set: \Pi_{0}:=\{(x,g,M)\in\mathcal{F}\times H(\mathbb{R})^{+}\times\Omega:\tilde{\Phi}(x)\in g\cdot D_{M}\} (the last condition means that the image of the Deligne torus via \tilde{\Phi}(x) lies in gG_{M}g^{-1} ). We remark that all Hodge data we have to consider, belong to the above definable set. Notice that if (x,g,F)\in\Pi_{0} then g\cdot Mg^{-1}\tilde{\Phi}(x)=g\cdot D_{M} is an algebraic subvariety of D . By “dimension” we will always mean the complex dimension (and possibly the local dimension at some point y , which we denote by \dim_{y} ). Consider the functions: d:\Pi_{0}\to\mathbb{R},\qquad(x,g,M)\mapsto d(x,g,F):=\dim_{\tilde{\Phi}(x)}(g\cdot D_{M}). d_{S}:\Pi_{0}\to\mathbb{R},\qquad(x,g,M)\mapsto d(x,g,F):=\dim_{\tilde{\Phi}(x)}\left(g\cdot D_{M}\cap\tilde{\Phi}(\mathcal{F})\right). To ease the notation, from now on, by x we mean \tilde{\Phi}(x) (and we don’t distinguish between groups and their real points). Definition 3.6 . We define two sets: \Pi_{1}:=\{(x,g,M)\in\Pi_{0}:\forall(x,g_{1},M_{1})\in\Pi_{0}:gMg^{-1}x\varsubsetneq g_{1}M_{1}g_{1}^{-1}x\Rightarrow d(x,g,M)-d_{S}(x,g,M)