One-dimensional representations of \({U(p,q)}\) and the Howe correspondence (Q1865296): Difference between revisions

In the present paper, the authors consider the Howe correspondence \[ \theta:\text{Irr}(\widetilde{\text{U}}(p,q))\rightarrow \text{Irr}(\widetilde{\text{U}}(r,s))\cup \{0\} \] corresponding to the reductive dual pair \((\text{U} (p,q),\text{U} (r,s))\). They give an explicit description of \(\theta (\chi)\), in terms of Langlands parameters, for any one-dimensional representation \(\chi \) of \(\widetilde{\text{U}}(p,q)\). Also, the authors completely determine the unitarity of \(\theta (\lambda)\) by showing that \(\theta (\chi)\) is unitary if and only if \(\theta (\chi)\) is a weakly fair \(A_{\mathfrak q}(\lambda)\) module. Further, for \(\chi\) lifting to unitary representations \(A _{\mathfrak q}(\lambda)\), both \(\mathfrak q\) and \(\lambda\) are explicitly computable. There are a number of interesting bi-products of these results. For example, D. Vogan has conjectured that any unitary representation of \(\text{U}(p,q)\) with integral infinitesimal character is a weakly fair \(A_{\mathfrak q}(\lambda)\) module. This conjecture leads to the prediction that a stable range theta lift of a weakly fair \(A_{\mathfrak q}(\lambda)\) module is again a weakly fair \(A_{\mathfrak q}(\lambda)\) module. The authors point out that the results of the present paper verify this prediction for the one-dimensional \(A_{\mathfrak q}(\lambda)\) modules.