Galois representations, Mumford-Tate groups and good reduction of Abelian varieties (Q1890252)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Galois representations, Mumford-Tate groups and good reduction of Abelian varieties
scientific article

    Statements

    Galois representations, Mumford-Tate groups and good reduction of Abelian varieties (English)
    0 references
    0 references
    29 December 2004
    0 references
    Let \(A\) be an abelian variety defined over a number field \(E\). The paper under review is related to the following conjecture due to \textit{Y. Morita} [J. Fac. Sci., Univ. Tokyo, Sect. I A 22, 437--447 (1975; Zbl 0322.14015)]: if the Mumford-Tate group of \(A\) contains no non trivial unipotent \(\mathbb{Q}\)-rational point then \(A\) has potentially good reduction at every prime of \(E\). This is related to Shimura varieties in the following sense. Let \(\text{MT}(A)\) be the Mumford-Tate group of \(A\), \(\mathbb{S}=\text{Res}_{\mathbb{C}| \mathbb{R}}(\mathbb{G}_{m,\mathbb{C}})\) and \(h:\mathbb{S}\to\text{MT}(A)_{\mathbb{R}}\) the usual morphism. A Shimura datum is the couple \((\text{MT}(A),h)\). This allows one to construct the canonical model of a Shimura variety Sh over a number field. This variety is a moduli space for polarized abelian varieties with a level structure and an additional structure. Roughly speaking, it is a moduli space for abelian varieties whose Mumford-Tate group is contained in MT\((A)\). The author shows that the condition on unipotents in Morita's conjecture is equivalent to the compactness of Sh\({}_{\mathbb{C}}\). In the PEL case, partially treated by Morita, the additional structure that defines the moduli problem is given by endomorphisms of the abelian variety. \textit{R. Noot} [J. Reine Angew. Math. 519, 155--169 (2000; Zbl 1042.14014)] proved the conjecture in the case of Mumford-Shimura curves that are not PEL Shimura varieties. In this case, the additional structure used to define the moduli problem for Sh is not given by endomorphisms of abelian varieties, but rather by absolute Hodge cycles. In the paper under review the author proves the following. Let \(V=H_1(A_{\mathbb{C}},\mathbb{Q})\) be the first singular homology group and \(\rho:\text{MT}(A)\to\text{GL}(V)\) the natural representation. First he shows that if the image of \(\rho_{\mathbb{Q}_{\ell}}\) contains no non trivial unipotent element over \(\mathbb{Q}_{\ell}\) of index 2, then \(A\) has potentially good reduction at every discrete place of \(E\). Next he classifies representations satisfying the hypothesis of the criterion in terms of the combinatorics of the Galois action of their tensorial and irreducible components. Third, using the machinery of Shimura varieties he produces examples of abelian varieties satisfying the criterion. Finally, he shows that the method applies to abelian varieties with same adjoing Mumford-Tate groups as abelian varieties satisfying the criterion. In all cases examples are produced.
    0 references
    0 references
    Galois representations
    0 references
    Mumford-Tate group
    0 references
    abelian varieties
    0 references
    0 references
    0 references