Double coset decompositions and intertwining (Q5955030): Difference between revisions

Let \(F\) be a nonarchimedean local field of odd residual characteristic. Let \(G\) denote a symplectic, orthogonal, or unitary group, realized in the usual way as the group of self-adjoint elements of some general linear group \(\widetilde{G}=\text{GL}_n(F)\). \textit{C. Bushnell} and \textit{P. Kutzko} [Proc. Lond. Math. Soc., III. Ser. 77, No. 3, 582-634 (1998; Zbl 0911.22014), The admissible dual of \(\text{GL}(n)\) via compact open subgroups, Ann. Math. Stud. 129, Princeton Univ. Press, Princeton, N. J. (1993; Zbl 0787.22016), Compos. Math. 119, No. 1, 53-97 (1999; Zbl 0933.22027)] have described the smooth representation theory of \(\widetilde{G}\) in terms of ``\(\widetilde{G}\)-types'', i.e., certain pairs that consist of a compact subgroup \(K\) of \(\widetilde{G}\) and an irreducible representation of \(K\). The present paper is a step toward a similar description for \(G\). (When \(G\) is a unitary group in three variables, this task was recently completed by \textit{L. Blasco} [Manuscr. Math. 107, No. 2, 151-186 (2002)].) Since \(G\)-types should be difficult to construct intrinsically, the author uses the explicit embedding of \(G\) in \(\widetilde{G}\), and transfers some of the results of Bushnell and Kutzko. This allows him to define ``semisimple strata'', and to compute their intertwining. As a consequence, one can show that any irreducible representation of \(G\) that contains a stratum satisfying certain regularity and ellipticity conditions must be supercuspidal. The author addresses elsewhere [Proc. Lond. Math. Soc., III. Ser. 83, No. 1, 120-140 (2001; Zbl 1017.22012)] the problem of constructing inducing data for these representations. The construction here is not yet general enough to produce all supercuspidal representations; it only produces ``very'' supercuspidal representations, roughly analogous to those constructed by \textit{H. Carayol} [Ann. Sci. Éc. Norm. Supér., IV. Sér. 17, 191-225 (1984; Zbl 0549.22009)] for \(\widetilde{G}\), and \textit{L. Morris} [Proc. Lond. Math. Soc., III. Ser. 57, No. 2, 329-356 (1988; Zbl 0663.22010)] and the reviewer [Pac. J. Math. 185, No. 1, 1-32 (1998; Zbl 0924.22015)] for more general groups. However, as in the first two of these constructions, some of the representations obtained cannot arise from any of the several constructions that use only tamely ramified data (the most general of which is due to \textit{J.-K. Yu} [J. Am. Math. Soc. 14, No. 3, 579-622 (2001; Zbl 0971.22012]).