\((\text O(V\oplus F),\text O(V))\) is a Gelfand pair for any quadratic space \(V\) over a local field \(F\) (Q1006366): Difference between revisions

Let \(F\) be an arbitrary local field of characteristic different from \(2\). Consider the standard embedding \(\text O(V) \hookrightarrow\text O(W)\) where \((W:=V\oplus F{\cdot}e,Q)\) is a quadratic space over \(F\) with \(Q(e)=1\). In the paper under review it is proved that \((\text O(V),\text O(W))\) is a Gelfand pair. This means that for every irreducible admissible representation \((\pi,E)\) of \(\text O(W)\) the inequality \(\dim\Hom_{\text O(V)}(E,\mathbb{C})\leq 1\) holds. The authors use a technique of Gelfand and Kazhdan to deduce this representation theoretic result from considerations about invariant distributions. They show that any distribution on \(\text O(W)\) which is invariant with respect to the standard two-sided action of \(\text O(V)\times\text O(V)\) is also invariant with respect to transposition. The proof of the latter result forms the main part of this paper. When \(F\) is the field of real or complex numbers or a \(p\)-adic field this has previously been obtained by \textit{G.~van Dijk} [Math. Z. 193, 581--593 (1986; Zbl 0613.43009)], \textit{S.~Aparicio} and \textit{G.~van Dijk} [Complex generalized Gelfand pairs (Tambov University Reports, Tambov) (2006)] and \textit{E. P. H.~Bosman} and \textit{G.~van Dijk} [Geom. Dedicata 50, No. 3, 261--282 (1994; Zbl 0829.22015)], respectively. The authors however give a uniform proof for the archimedean and non-archimedean case except at one point where the archimedean case requires some extra work. Their proof relies on Bruhat filtration, Frobenius reciprocity and Bernstein's localization property. All of these basic results on invariant distributions are recalled for locally compact totally disconnected spaces and for smooth manifolds. Readers needing more background information will find that in an earlier paper of the same authors, where they showed an analogous result about general linear groups [Compos. Math. 144, No. 6, 1504--1524 (2008; Zbl 1157.22004)]. Moreover the main theorem of these two papers stands in connection with similar results for special orthogonal groups and unitary groups that have been obtained previously by \textit{B. H.~Gross} and \textit{D.~Prasad} [Can. J. Math. 44, No. 5, 974--1002 (1992; Zbl 0787.22018)] and \textit{G.~van Dijk} [Math. Z. 261, No. 3, 525--529 (2009; Zbl 1158.22010)], respectively.