University of Georgia

Mathematics Department Colloquium

Chris Skinner

Institute for Advanced Study

January 21 1999

Recent results on modularity of two-dimensional Galois representations


Abstract: Two-dimensional $p$-adic representations of $G_Q=Gal(\overline Q/Q)$, the absolute Galois group of the field of rational numbers $Q$, (i.e., homomorphisms $G_Q \rightarrow GL_2(R)$, $R$ the ring of integers of a finite extension of the $p$-adics $Q_p$) occur often in number theory. For example, the Tate-module at a prime $p$ of an elliptic curve over $Q$ gives rise to such a representation, and in a similar manner such representations appear in the Tate-modules of Abelian varieties with "lots of real multiplications." Similarly, a holomorphic modular form that is an eigenform for the Hecke operators has associated to it such a two-dimensional representation for each prime $p$. In all of these cases the representations encode much of the arithmetic information about the objects (e.g., the number of points over finite fields, the $L$-series, the Hecke eigenvalues). The celebrated conjecture of Shimura and Taniyama asserts that the representations in the first two examples also belong to the third example. This has been subsumed in a conjecture of Mazur and Fontaine which purports to give criteria for any two-dimensional $p$-adic representation of $G_Q$ to be a representation associated to an eigenform. Such a representation is said to be ``modular''. As is well-known, Wiles' proof of Fermat's Last Theorem included the first significant results in the direction of these conjectures. In this talk, I will recall the above notions, discuss the conjectures, and talk about some recent work extending/complementing the earlier work of Wiles. In particular I will discuss some recent joint work of Wiles and myself on the modularity of representations that are ordinary at $p$. This work entails not just establishing modularity of two-dimensional representations of $G_Q$ but also suitable generalizations to totally real fields.