Base change for tempered irreducible representations of \(\mathrm{GL}(n,\mathbb R)\) (Q1144149)

To each irreducible representation \(\pi\) of \(\mathrm{GL}(n, \mathbb R)\) is associated its ``base change lifting'', an irreducible representation \(\Pi\) of \(\mathrm{GL}(n, \mathbb C)\). It is expected that the characters of these two representations are related in a certain way, at least if \(\pi\) is tempered, and this relation has in fact been proved for \(\mathrm{GL}(2, \mathbb R)\) by \textit{T. Shintani} [J. Math. Soc. Japan 29, 165--188 (1977; Zbl 0342.20021)], and for representations of \(\mathrm{GL}(n, \mathbb R)\) induced from unramified quasicharacters of a minimal parabolic subgroup by \textit{L. Clozel} [Lect. Notes Math. 728, 17--41 (1979; Zbl 0495.20020)], In this paper, we prove the relation for arbitrary tempered irreducible representations of \(\mathrm{GL}(n, \mathbb R)\). The proof involves computations not unlike those used to calculate the character of an induced representation. The representations in question are all induced from parabolic subgroups whose Levi components are products of copies of \(\mathrm{GL}(2)\) and \(\mathrm{GL}(1)\), so we are able to use Shintani's results for \(\mathrm{GL}(2)\) as a starting point. In a subsequent paper we shall prove a similar ``inductive step'' for general quasi-split connected real reductive groups, which gives the general case from Clozel's work on discrete series.