Let be a real reductive Lie group and a -stable closed reductive subgroup with maximal compact subgroup . For an irreducible unitary representation of , the restriction is called -admissible when it splits into a discrete Hilbert direct sum of irreducible unitary representations of with finite multiplicities. Kobayashi proved (Invent. Math. 1994) that -admissibility of forces -admissibility of , and conjectured the converse (Conjecture D, 2000). The conjecture, shown by Zhu and Liang (2010) to be equivalent to the equivalence of -admissibility, -admissibility, and -admissibility of the underlying Harish-Chandra module, remains open in general. It is known for discrete series representations (the square-integrable irreducible representations; Duflo-Vargas) and under auxiliary hypotheses on infinitesimal characters (Zhu-Liang). An affirmative answer would upgrade Kobayashi's symplectic cone criterion to a necessary and sufficient condition for admissible restrictions.
Background
For a real reductive Lie group (the standard class of Lie groups in Harish-Chandra's representation theory, e.g. , , or the metaplectic group ) with Lie algebra , fix a Cartan involution (an involutive automorphism whose fixed-point set is a maximal compact subgroup) and let be the corresponding maximal compact subgroup. By Harish-Chandra's admissibility theorem, every irreducible unitary representation of is -admissible: its restriction to decomposes as a direct sum of irreducible representations of (called -types), each occurring with finite multiplicity. For a non-compact reductive subgroup no such theorem holds: the restriction decomposes in general as a direct integral over the unitary dual (a branching law), possibly with continuous spectrum and infinite multiplicities. Kobayashi introduced the framework of admissible restrictions: is -admissible if it decomposes as a discrete Hilbert direct sum of irreducible unitary representations of with finite multiplicities . In this case the restriction can be studied algebraically through Harish-Chandra modules (the -module of -finite vectors, which determines ): one says is -discretely decomposable if decomposes into a direct sum of irreducible -modules, and -admissible if moreover all multiplicities in this decomposition are finite, where and is a maximal compact subgroup of when is -stable. Assume throughout that is -stable, closed, and reductive. Three fundamental results organize the subject. First (Kobayashi, Invent. Math. 117, 1994, Theorem 1.2), if is -admissible — every -type occurs with finite multiplicity — then is automatically -admissible, so the restriction has no continuous spectrum and finite multiplicities. Second (Kobayashi, Invent. Math. 131, 1998, Part III of the same series, Theorem 4.2), for the cohomologically induced modules (Zuckerman derived functor modules attached to a -stable parabolic , with in the weakly fair range, a standard cohomological range condition) over reductive symmetric pairs, -discrete decomposability is equivalent to -admissibility. Third, the modern cone criterion (Kobayashi, Pure Appl. Math. Q. 2021, completing a program announced at ICM 2002): an admissible -module of finite length is -admissible if and only if , where is the asymptotic -support (the limit cone of the highest weights of occurring -types) and is the momentum cone of the Hamiltonian -action on the cotangent bundle ; combined with the first result this gives a checkable sufficient condition for -admissibility. What is missing is the converse: nothing known excludes the possibility that a unitary restriction decomposes discretely with finite multiplicities while some -type occurs with infinite multiplicity across the infinitely many -constituents. Kobayashi conjectured in 2000 (Conjecture D in his Adv. Stud. Pure Math. 26 survey) that this cannot happen, and Zhu and Liang (C. R. Math. Acad. Sci. Paris 2010) proved that the conjecture is equivalent to the equality of the three admissibility notions: their Theorem 1.3 shows -admissibility implies -admissibility in general, using Kostant's theorem that the center of the universal enveloping algebra of acts finitely — the infinitesimal character of an irreducible module being the character by which this center acts — so that the infinitesimal characters of the -constituents lie in translates of a fixed infinitesimal character by weights of , together with a Casimir inequality.
Problem Statement
Let be a real reductive Lie group, a Cartan involution, the fixed maximal compact subgroup, and a -stable closed reductive subgroup with maximal compact subgroup . Let be an irreducible unitary representation of whose restriction is -admissible, that is,
with all multiplicities finite. Must be -admissible, i.e. must every irreducible representation of the compact group occur with finite multiplicity, ? Equivalently, are the three notions — -admissibility of the unitary restriction, -admissibility, and -admissibility of the Harish-Chandra module (the -module of -finite vectors of ) as a -module, i.e. decomposability of into a direct sum of irreducible -modules with finite multiplicities — equivalent for every such triple , without any auxiliary hypothesis such as -relatedness of infinitesimal characters (pairwise differences of the occurring infinitesimal characters lying in the lattice of -weights) or reality of the occurring infinitesimal characters?
References
- F. Zhu, K. Liang, On a branching law of unitary representations and a conjecture of Kobayashi, C. R. Math. Acad. Sci. Paris 348 (2010) 959-962, DOI 10.1016/j.crma.2010.09.006, https://doi.org/10.1016/j.crma.2010.09.006
- T. Kobayashi, Discretely decomposable restrictions of unitary representations of reductive Lie groups — examples and conjectures, in Analysis on Homogeneous Spaces and Representation Theory of Lie Groups, Okayama-Kyoto 1997, Advanced Studies in Pure Mathematics 26, Mathematical Society of Japan, 2000, pp. 99-127, DOI 10.2969/aspm/02610099, https://doi.org/10.2969/aspm/02610099
- T. Kobayashi, Discrete decomposability of the restriction of with respect to reductive subgroups and its applications, Inventiones Mathematicae 117 (1994) 181-205, DOI 10.1007/BF01232239, https://doi.org/10.1007/BF01232239
- T. Kobayashi, Discrete decomposability of the restriction of , Part III: Restriction of Harish-Chandra modules and associated varieties, Inventiones Mathematicae 131 (1998) 229-256, DOI 10.1007/s002220050203, https://doi.org/10.1007/s002220050203
- M. Duflo, J. A. Vargas, Branching laws for square integrable representations, Proceedings of the Japan Academy, Series A, Mathematical Sciences 86 (2010) 49-54, DOI 10.3792/pjaa.86.49, https://doi.org/10.3792/pjaa.86.49
- T. Kobayashi, A program for branching problems in the representation theory of real reductive groups, in Representations of Reductive Groups, Progress in Mathematics 312, Birkhäuser/Springer, 2015, pp. 277-322, DOI 10.1007/978-3-319-23443-4_10, arXiv:1509.08861, https://arxiv.org/abs/1509.08861
- T. Kobayashi, Admissible restrictions of irreducible representations of reductive Lie groups: symplectic geometry and discrete decomposability, Pure and Applied Mathematics Quarterly 17 (2021) 1321-1343, DOI 10.4310/PAMQ.2021.v17.n4.a5, arXiv:1907.12964, https://arxiv.org/abs/1907.12964
- T. Kobayashi, Recent advances in branching problems of representations, Sugaku Expositions 37 (2024) 129-177, DOI 10.1090/suga/485, arXiv:2112.00642, https://arxiv.org/abs/2112.00642
- T. Kobayashi, Harish-Chandra's admissibility theorem and beyond, 2024, arXiv:2412.20387, https://arxiv.org/abs/2412.20387
- M. Kitagawa, Kobayashi's Conjectures on the Discrete Decomposability, in Symmetry in Geometry and Analysis, Volume 3: Festschrift in Honor of Toshiyuki Kobayashi, Progress in Mathematics 359, Birkhäuser/Springer, 2025, pp. 31-99, DOI 10.1007/978-981-97-7666-5_3, https://doi.org/10.1007/978-981-97-7666-5_3