Distinction by the quasi-split unitary group (Q607836): Difference between revisions

The article under review proves a refinement of the following result of \textit{H. Jacquet} [Ann. Sci. Éc. Norm. Supér. (4) 38, No. 4, 609--669 (2005; Zbl 1111.11028)]. For a quadratic extension of number fields \(E/F\), let \(\Pi\) be a cuspidal automorphic representation of \(\mathrm{GL}(n,E_{\mathbb{A}})\) that is the base change of an automorphic representation \(\pi\) of \(\mathrm{GL}(n,F_{\mathbb{A}})\). In the paper cited above, the author proved that \(\Pi\) is distinguished by some unitary group \(H\). The refinement proved in this paper is that one can take \(H\) to be quasi-split. The author notes that the main result has already been proved under the assumption that every real place of \(F\) splits in \(E\), so the results of this paper would remove this assumption on the real places of \(F\). The proof of the main theorem is reduced proving a statement regarding the local Bessel distribution \(\mathcal{B}_{\pi_v}\) attached to the local representation \(\pi_v\) where \(v\) is a place of \(F\) that is inert in \(E\). In particular, it is shown that \(\mathcal{B}_{\pi_v}\) does not vanish on a certain open set. While this statement about the local Bessel distribution attached to \(\pi_v\) has previously been proved for nonarchimedean places \(v\), this paper reproves it in the nonarchimedean case using a new method. This method of proof is very similar to the one used to prove the analogous results about the local Bessel distribution attached to \(\pi_v\) at an archimeadean place \(v\). Section 2 of the paper proves the statement regarding the local Bessel distribution attached to \(\pi_v\) at a nonarchimedean place. Sections 3-5 are concerned with the analogous statement attached to \(\pi_v\) at real places. The multitude of sections covering the archimedean cases partially stem from the need to define and prove some intermediate results in the archimedean case that were known results which were simply cited in the nonarchimedean case. For instance, the main result of Section 3.2 (Proposition 4) is an analog to a lemma of \textit{I. M. Gelfand} and \textit{D. A. Kazhdan} [Lie Groups Represent., Proc. Summer Sch. Bolyai János Math. Soc., Budapest 1971, 95--118 (1975; Zbl 0348.22011)] which is cited as Lemma 1 in the paper under consideration. Finally, Section 6 uses the various results regarding the non-vanishing of the local Bessel distributions attached to the various \(\pi_v\) to prove the distinction of \(\Pi\) by the quasi-split unitary group \(H\). The argument follows from some application of a relative trace formula where the contribution of \(\pi\) to a spectral expression is described using a distribution that factors locally as a product of Bessel distributions attached to \(\pi_v\).