\(R\)-groups and parameters (Q429099)

Let \(F\) be a \(p\)-adic field. Let \(G\) be a connected reductive quasisplit algebraic \(F\)-group, and let \(P\) be a parabolic \(F\)-subgroup of \(G\) with Levi component \(M\) defined over \(F\). Let \(\pi\) be an irreducible admissible representation in the discrete series of \(M(F)\), and let \(\mathrm{ind}_{P(F)}^{G(F)}(\pi)\) be the representation of \(G\) obtained from \(\pi\) via normalized parabolic induction. It is well known that the Knapp--Stein \(R\)-group \(R(\pi)\) provides a great deal of information about the structure and constituents of \(\mathrm{ind}_{P(F)}^{G(F)}(\pi)\). In particular, the intertwining algebra of \(\mathrm{ind}_{P(F)}^{G(F)}(\pi)\) is isomorphic to a certain twist of the group algebra of \(R(\pi)\). Hence this twist determines the number of inequivalent irreducible constituents of \(\mathrm{ind}_{P(F)}^{G(F)}(\pi)\), as well as their multiplicities. It also dictates the ellipticity of these constituents. \textit{J. Arthur} [in: Orbites unipotentes et représentations. II: Groupes \(p\)-adiques et réels, Astérisque 171--172, 13--71 (1989; Zbl 0728.22014)] conjectured that \(R(\pi)\) can be interpreted in terms of the Langlands parameter \(\phi\) of the \(L\)-packet of \(M(F)\) containing \(\pi\). In particular, he associates to \(\phi\) an \(R\)-group that is conjecturally isomorphic to \(R(\pi)\). That the Arthur \(R\)-group is indeed isomorphic to the Knapp--Stein \(R\)-group has been established in a number of cases: for the archimedean case by \textit{D. Shelstad} [Math. Ann. 259, 385--430 (1982; Zbl 0506.22014)], for principal series representations by \textit{C. D. Keys} [Ann. Sci. Éc. Norm. Supér. (4) 20, No. 1, 31--64 (1987; Zbl 0634.22014)], for the group \(G = \text{SO}_{2n+1}\) by \textit{D. Ban} and \textit{Y. Zhang} [Compos. Math. 141, No. 2, 323--343 (2005; Zbl 1131.11033)], and for certain supercuspidal representations of the groups \(G = \text{SO}_n, \text{Sp}_{2n}\) by \textit{D. Goldberg} [Clay Mathematics Proceedings 13, 159--185 (2011; Zbl 1280.22020)]. The paper under review gives a verification of the conjecture when \(G = \text{SO}_{2n+1}\), \(\text{Sp}_{2n}\), \(\text{O}_{2n}\), or \(\text{U}_n\). The proof uses the classification of the discrete series for classical groups by \textit{C. Moeglin} [J. Eur. Math. Soc. (JEMS) 4, No. 2, 143--200 (2002; Zbl 1002.22009), Pac. J. Math. 233, No. 1, 159--204 (2007; Zbl 1157.22010)] and \textit{C. Moeglin} and \textit{M. Tadic} [J. Am. Math. Soc. 15, No. 3, 715--786 (2002; Zbl 0992.22015)].