Date: Tuesday, April 17, 2018
Title: Hodge Theory and ominimal geometry
Abstract: (joint w. Ben Bakker) Hodge theory studies algebraic varieties by studying the periods of its global differential forms. It gives a way to assign to every algebraic variety X a linear algebraic object called a "Hodge Structure", and the famous Hodge conjecture states that one can understand much about the geometry of X by studying the associated hodge structures. One fruitful way of understanding hodge structures is by looking at their moduli space M, which can naturally be given the structure of a complex orbifold. In the case of weight 1 structures, M parametrizes abelian varieties and so is naturally an algebraic variety. However, in the general case it is known that M does not admit an algebraic structure. This creates a difficult situation, since families of algebraic varieties over an algebraic base B give holomorphic maps (known as period mappings) B>M, but holomorphic maps can behave very badly in general (for instance, their asymptotics can be quite unwieldy, as opposed to algebraic maps).
We explain how to provide a partial substitute for the lack of an algebraic structure by equipping M with an ominimal structure, and show that the period mappings are "definable" with respect to this structure. It turns out that ominimality gives an extremely useful notion of "tameness"; for instance, a very powerful theorem of PeterzilStarchenko says that holomorphic maps which are ominimal have to be algebraic in a wide variety of circumstances. As a consequence of this work, we give an easy proof of a result of CattaniDeligneKaplan giving evidence towards the Hodge conjecture.
The proof of our main theorem relies heavily on work of Kashiwara, Schmid and CattaniKaplanSchmid on asymptotics of Period mappings.
Speaker: Jacob Tsimerman
Institution: University of Toronto
