Exotic proper actions on homogeneous spaces via convex cocompact representations

Maciej Bocheński, Yosuke Morita · 2025

Read Exotic proper actions on homogeneous spaces via convex cocompact representations by Maciej Bocheński, Yosuke Morita (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-14274.txt.

Introduction

Maciej Bocheński and Yosuke Morita construct, in 2025, homogeneous spaces with a surprising split personality. A homogeneous space can host well-behaved group actions, and earlier work doubted whether discrete subgroups could act properly where no large semisimple subgroup can. For some spaces, that was unknown even in the smallest case.

Building on work about convex cocompact groups and right-angled Coxeter groups, the authors produce spaces admitting proper actions by discrete groups that mimic hyperbolic lattices, in dimensions two, three, and four. A separate argument, using nilpotent orbits and counting, bars any such action by a noncompact semisimple subgroup. The examples settle the existence question.

Catalog record

CREATORMaciej Bocheński, Yosuke Morita
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-14274

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Exotic proper actions on homogeneous spaces via convex cocompact representations Maciej Bocheński, Yosuke Morita Abstract We construct a series of homogeneous spaces G / H G/H of reductive type which admit proper actions of discrete subgroups of G G isomorphic to cocompact lattices of O ⁡ ( n , 1 ) \OO(n,1) ( n = 2 , 3 , 4 n=2,3,4 ) but do not admit proper actions of non-compact semisimple subgroups of G G . The existence of such homogeneous spaces was previously not known even for n = 2 n=2 . Our construction of proper actions of discrete subgroups is based on Guéritaud–Kassel’s work on convex cocompact subgroups of O ⁡ ( n , 1 ) \OO(n,1) and Danciger–Guéritaud–Kassel’s work on right-angled Coxeter groups. On the other hand, the non-existence of proper actions of non-compact semisimple subgroups is proved by the theory of nilpotent orbits and elementary combinatorics. 1 Introduction 1.1 The problem and the main result Let G be a linear reductive Lie group and H\subset G a reductive subgroup. If a discrete subgroup \Gamma\subset G acts properly on the homogeneous space G/H , then the double coset space \Gamma\backslash G/H is called a Clifford-Klein form . We also say that \Gamma is a discontinuous group for G/H . We are interested in the non-Riemannian case, i.e. the case where H is non-compact. In this situation, not every discrete subgroup of G acts properly on G/H . Indeed, lattices of G never act properly on G/H by, for instance, Moore’s ergodicity theorem. Fortunately, there is a convenient way of obtaining non-trivial examples of Clifford–Klein forms: if a reductive non-compact subgroup L of G acts properly on G/H , then any closed subgroup of L (and in particular, any discrete subgroup of L ) also acts properly on G/H . Proper actions of discrete subgroups obtained in this way, as well as the associated Clifford–Klein forms, are called standard . For some G/H , this construction even yields cocompact proper actions: Example 1.1 (Kobayashi [ Kob89 , Prop. 4.9] ; see also Kulkarni [ Kul81 , Thm. 6.1] ) . Let G=\OO(2n,2) , H=\UU(n,1) , and L=\OO(2n,1) . Then, we have G=LH , and furthermore, H\cap L=\UU(n) is compact. Thus, L acts properly and transitively on L/(H\cap L)\cong G/H . Taking \Gamma to be a cocompact lattice of L=\OO(2n,1) , we obtain a compact Clifford–Klein form of G/H=\OO(2n,2)/\UU(n,1) . Of course, not all proper actions are standard. In fact, it is now known that some compact Clifford–Klein forms are non-standard in a very strong sense: Example 1.2 (Lee–Marquis [ LM19 , Thm. A] + Guéritaud–Guichard–Kassel–Wienhard [ GGKW17 , Cor. 1.9] for n=2,3,4 ; Monclair–Schlenker–Tholozan [ MST23+ , Cor. 1.9] for general n ) . For any n\geqslant 2 , there exists a compact Clifford–Klein form of \OO(2n,2)/\UU(n,1) whose discontinuous group \Gamma is not commensurable to any lattice of any linear reductive Lie group. In general, a proper action which is not even a deformation of a standard proper action is called exotic (see Salein [ Sal00 ] and Lakeland–Leininger [ LL17 ] for other examples of exotic compact Clifford–Klein forms). One of the main themes in the theory of Clifford–Klein forms is the following general question, with varying meanings of the word ‘large’: To what extent the existence of a ‘large’ discontinuous group \Gamma for G/H implies the existence of a ‘large’ reductive subgroup L acting properly on G/H ? One instance of this question, where ‘large’ means ‘cocompact’, is Toshiyuki Kobayashi’s conjecture which asserts that a homogeneous space of reductive type which admits a compact Clifford–Klein form also admits a standard compact Clifford–Klein form ( [ Kob97 , §6, (vi)] , [ Kob01 , Conj. 4.3] ). It is recognized as one of the biggest open problems in this area, and its solution seems to be still out of reach. Note that the conjecture does not say that every compact Clifford–Klein form is standard. In particular, Example 1.2 is not a counterexample to this conjecture. Another important topic in this area to study the above question for a fixed class of discrete groups \Gamma . In this paper, we study this problem in the case where \Gamma is isomorphic to a cocompact lattice of \OO(n,1) for n=2,3 , and 4 . A simplified version of our main result is the following: Theorem 1.3 (cf. Theorem 1.12 ) . There is a homogeneous space G/H of reductive type with the following two properties: • For n=2,3 , and 4 , there exists a discrete subgroup of G which is isomorphic to a cocompact lattice of \OO(n,1) and acts properly on G/H . • There does not exist a non-compact semisimple subgroup of G which acts properly on G/H . In other words, an analogue of Kobayashi’s conjecture is false in this setting: the above G/H admits a proper action of a cocompact lattice of \OO(n,1)\ (n=2,3,4) , whereas it does not admit a standard one. 1.2 Comparison to previous results When \Gamma\cong\mathbb{Z} and L\cong\mathbb{R} , the problem was completely settled by Kobayashi in his Ph.D. thesis. In this case, the existence of proper actions is equivalent to the existence of standard ones: Fact 1.4 (Kobayashi [ Kob89 , §4] ) . For a homogeneous space G/H of reductive type, the following three conditions are equivalent: • (P-inf) — G/H admits a proper action of an infinite discrete subgroup of G . • (P-Z) — G/H admits a proper action of a discrete subgroup of G isomorphic to \mathbb{Z} . • (P-R) — G/H admits a proper action of a reductive subgroup of G isomorphic to \mathbb{R} . Furthermore, these conditions are also equivalent to \rank_{\mathbb{R}}G>\rank_{\mathbb{R}}H , where \rank_{\mathbb{R}} stands for the real rank of reductive Lie groups. Fact 1.4 is called the Calabi–Markus phenomenon because this kind of phenomenon was first found in their work on Lorentzian geometry [ CM62 ] . Remark 1.5 . The above terminologies (P-inf), (P-Z) and (P-R) are not conventional at all. We just introduced them in order to facilitate the comparison of various results in this area. The same applies to the other conditions which we shall introduce below. The ‘second smallest case’, namely, the case where L is locally isomorphic to \SL(2,\mathbb{R}) (or equivalently, to \OO(2,1) ), has been studied actively in the last three decades. In this case, the natural classes of \Gamma to be considered are surface groups and free groups, as they are respectively the classes of cocompact and non-cocompact lattices of \SL(2,\mathbb{R}) , up to commensurability. Thus, let us introduce the following three conditions on a homogeneous space G/H of reductive type: • (P-sl 2 R) — G/H admits a proper action of a closed subgroup of G locally isomorphic to \SL(2,\mathbb{R}) . • (P-surf) — G/H admits a proper action of a discrete subgroup of G isomorphic to a surface group of genus \geqslant 2 . • (P-free) — G/H admits a proper action of a discrete subgroup of G isomorphic to a non-abelian free group. We also consider: • (P-nva) — G/H admits a proper action of a non-virtually abelian discrete subgroup of G . An easy observation is that the following implications hold: (\text{P-sl}_{2}\text{R})\

Read the complete text