Weyl group multiple Dirichlet series. II: The stable case (Q2501077): Difference between revisions

Recall the basic situation of Part I [Proc. Symp. Pure Math. 75, 91--114 (2006; Zbl 1112.11025)]: Given a reduced root system \(\Phi\) of rank \(r\) with Weyl group \(W\) and an algebraic number field \(F\) containing the group \(\mu_n\) the authors introduced a family of Dirichlet series in \(r\) complex variables, to be called Weyl group multiple Dirichlet series. The coefficients of these series can be given (basically) in terms of Gauss sums. These coefficients show a ``twisted'' multiplicative behaviour that reduces their determination to those that are powers of a single prime \(p\); there is no Euler product (in general). For each prime \(p\), the number of nonzero such coefficients is equal to the order of \(W\). The main Theorem of the present work roughly states: The Weyl group multiple Dirichlet series, which by definition initially converges only in the intersection of \(r\) half-spaces, has a meromorphic continuation to the entire space. After multiplying the series by a normalizing factor consisting of some gamma and zeta functions, the normalized function satisfies a group of functional equations which is (conjecturally) isomorphic to \(W\). The precise version of this result is too involved to be given here. The proof assumes a stability hypothesis which states that \(n\) is sufficiently large, depending on \(\Phi\). (The authors expect that without this assumption the main result will hold similarly.) The root system plays a fundamental role in the highly complicated combinatorics of the proof of the main Theorem, and the result is finally reduced to a functional equation for Kubota's Dirichlet series. The authors conjecture that the Weyl group multiple Dirichlet series can be understood in the context of automorphic forms as Whittaker coefficients of an Eisenstein series on the \(n\)-fold metaplectic cover of a simply connected algebraic group over \(F\) whose root system is the dual of \(\Phi\). This conjecture would imply a less precise version of the main Theorem. More evidence for this conjecture is given in the present paper and in Part III: \textit{B. Brubaker}, \textit{D. Bump}, \textit{S. Friedberg} and \textit{J. Hoffstein}: Weyl group multiple Dirichlet series. III: Eisenstein series and twisted unstable \(A_r\). Ann. Math. (2) 166, No. 1, 293--316 (2007; Zbl 1154.11016).