Geometric analysis on small unitary representations of GL(N, R)

DOI10.1016/J.JFA.2010.12.008MaRDI QIDQ630780zbMATH OpenOpenAlexFDO

Authors Toshiyuki Kobayashi, Michael Pevzner, Bent Ørsted

Publication date 21 March 2011

Published in Journal of Functional Analysis (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1002.3006

branching law degenerate principal series representations phase space representation reductive group small representation symmetric pair symplectic group

Mathematics Subject Classification ID

Representation theory for linear algebraic groups (20G05) Unitary representations of locally compact groups (22D10) Analysis on real and complex Lie groups (22E30)

Abstract: The most degenerate unitary principal series representations {pi}_{i{lambda},{delta}} (with {lambda} in R, delta in Z/2Z) of G = GL(N,R) attain the minimum of the Gelfand-Kirillov dimension among all irreducible unitary representations of G. This article gives an explicit formula of the irreducible decomposition of the restriction {pi}_{i{lambda},{delta}}|_H (branching law) with respect to all symmetric pairs (G,H). For N=2n with n geq 2, the restriction {pi}_{i{lambda},{delta}}|_H remains irreducible for H=Sp(n,R) if {lambda} eq0 and splits into two irreducible representations if {lambda}=0. The branching law of the restriction {pi}_{i{lambda},{delta}}|_H is purely discrete for H = GL(n,C), consists only of continuous spectrum for H = GL(p,R) imes GL(q,R) (p+q=N), and contains both discrete and continuous spectra for H=O(p,q) (p>qgeq1). Our emphasis is laid on geometric analysis, which arises from the restriction of `small representations' to various subgroups.

Review text

In general, it is very hard to have a complete spectral decomposition of an irreducible unitary representation of a real reductive group, when restricted to a symmetric subgroup. In the paper under review, the authors achieve such decompositions for a family of small unitary representations of the general linear group \(\text{GL}(N,\mathbb R)\), when restricted to all symmetric subgroups. This is a nice result of representation theory, and is also interesting from the point of view of geometric analysis. To be more precise, for every purely imaginary number \(i\lambda\), and every \(\delta\in \mathbb Z/2\mathbb Z\), denote by \(\pi_{i\lambda, \delta}\) the normalized parabolic induction \(\text{Ind}_{(\mathbb R^\times \times \text{GL}(N-1,\mathbb R))\ltimes\mathbb R^{N-1}}^{\text{GL}(N,\mathbb R)}\chi_{i\lambda,\delta}\otimes 1\otimes 1\), where \(\chi_{i\lambda,\delta}\) is the character \(a\mapsto |a|^{i\lambda}\operatorname{sign}(a)^\delta\) of \(\mathbb R^\times\). This is an irreducible unitary representation of \(\text{GL}(N,\mathbb R)\) of minimal Gelfand-Kirillov dimension. The authors obtain explicit decompositions of \(\pi_{i\lambda, \delta}|_{\text{Sp}(2n,\mathbb R)}\), \(\pi_{i\lambda, \delta}|_{\text{GL}(n,\mathbb C)}\) (when \(N=2n\) is even), \(\pi_{i\lambda, \delta}|_{\text{GL}(p,\mathbb R)\times \text{GL}(q,\mathbb R)}\) and \(\pi_{i\lambda, \delta}|_{O(p,q)}\) (\(p,q>0\), \(p+q=N\)). In particular, they show that \(\pi_{i\lambda, \delta}|_{\text{Sp}(2n,\mathbb R)}\) is either irreducible or the direct sum of two irreducible subrepresentations; \(\pi_{i\lambda, \delta}|_{\text{GL}(n,\mathbb C)}\) is a countable sum of irreducible unitary representations; \(\pi_{i\lambda, \delta}|_{\text{GL}(p,\mathbb R)\times \text{GL}(q,\mathbb R)}\) contains only a continuous spectrum, and \(\pi_{i\lambda, \delta}|_{O(p,q)}\) contains both a discrete and a continuous spectrum.

Recommendations

Cites work

Cited in

(23)

This page was built for publication: Geometric analysis on small unitary representations of \(\text{GL}(N, \mathbb R)\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q630780)