Homological dimension of discrete subgroups in higher rank Lie groups

Chris Connell, D. B. McReynolds, Shi Wang · 2025

Read Homological dimension of discrete subgroups in higher rank Lie groups by Chris Connell, D. B. McReynolds, Shi Wang (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-2504-18923.txt.

Introduction

Chris Connell, McReynolds, and Shi Wang bound the homological size of discrete subgroups inside large symmetry groups, in a paper from 2025. The head title adds the word simple to the catalog title. The team uses a flow method they introduced earlier to probe how such subgroups sit inside the group.

New estimates on a critical index give an upper bound built from the space dimension and the group rank. For subgroups whose limit shapes are regular, the bound tightens further. The results cover groups that are dense in a strong algebraic sense but not lattices. The authors add guesses on which sizes can occur and open questions for the wider family of such groups.

Catalog record

CREATORChris Connell, D. B. McReynolds, Shi Wang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-18923

Browse the full library · About the Tower of Babel

Full text

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

Homological dimension of discrete subgroups in higher rank Lie groups Chris Connell, D. B. McReynolds, Shi Wang Abstract In this short note, we improve on a recent result by the authors. We show that infinite volume torsion free discrete subgroups of higher rank Lie groups have homological dimension gap at least one-eighth of the real rank, provided the injectivity radius of the quotient manifold is uniformly bounded from zero. 1 Introduction Using Patterson-Sullivan theory, the authors [ CMW23 ] introduced the natural flow for general nonpositively curved manifolds. As an application of the k –volume contracting property of the flow, the following result was obtained. Theorem 1.1 . [ CMW23 ] If X is a Hadamard space with bounded sectional curvatures, \Gamma<\mathrm{Isom}(X) is a discrete, torsion-free subgroup with M=X/\Gamma , and \mathbb{V} is a flat bundle over M , then for every \epsilon>0 , the homomorphism i_{k}:H_{k}(M_{\varepsilon};\mathbb{V})\rightarrow H_{k}(M;\mathbb{V}) induced by inclusion is surjective whenever k\geq j_{X}(\Gamma) . Here M_{\varepsilon} is the \varepsilon -thin part of M , and j_{X}(\Gamma) is the critical index of \Gamma , given by the following. Definition 1.2 . Let X be an n –dimensional Hadamard manifold, \Gamma<\operatorname{Isom}(X) a torsion-free discrete subgroup, and \mu_{x} be any conformal density of dimension \delta_{\mu} . We define the critical index of \Gamma associated to \mu to be j_{X}(\Gamma,\mu)=j_{X}(\mu):=\min\left(\{k\in\mathbb{N}:\inf_{(x,\theta)\in X\times\partial_{\mu}X}\operatorname{tr}_{k}(\nabla dB_{(x,\theta)})>\delta_{\mu}\}\cup\left\{n+1\right\}\right). and the critical index of \Gamma j_{X}(\Gamma)=\inf\left\{j_{X}(\mu):\mu\text{ is a }\delta_{\mu}-\text{conformal density}\right\}. We now restrict our context and assume X=G/K be an irreducible, higher rank symmetric space of non-compact type and \Gamma<G be a torsion free discrete subgroup. Using the general scheme of [ CMW23 , Theorem 1.11] and [ FL24 , Proposition 5.27] , we give a sharpened estimate on j_{X}(\Gamma) for all irreducible higher rank symmetric spaces. We will refer to [ CMW23 ] for further notation. Theorem 1.3 . Let G be a non-compact, simple real Lie group of rank r , n be the dimension of the associated symmetric space G/K , and \Gamma<G be a Zariski-dense, discrete, torsion-free subgroup which is not a lattice. Then we have the following upper bound on j_{X}(\Gamma) according to the type of restricted root system of G . 1. If G is of Type A_{r} , then j_{X}(\Gamma)<n+1-\frac{r}{8} . 2. If G is of Type B_{r} or D_{r} , then j_{X}(\Gamma)<n+1-\frac{3r}{16} . 3. If G is of Type C_{r} or (BC)_{r} , then j_{X}(\Gamma)<n+1-\frac{r}{4} . Let R be a commutative ring with unit and \Gamma be a discrete group. We denote \hd_{R}(\Gamma) the homological dimension of \Gamma associated to ring R , given by the following \operatorname{hd}_{R}(\Gamma):=\inf\{s\in\mathbb{Z}_{+}\,:\,H_{i}(\Gamma;V)=0\text{ for any }i>s\text{ and any $R\Gamma$--module $V$}\}. Thus, in view of Theorem 1.1 and Theorem 1.3 , we have the following corollaries. Corollary 1.4 . Let X,\Gamma,n,r be as in Theorem 1.3 . Suppose the injectivity radius of \Gamma\backslash X is uniformly bounded from zero. 1. If G=\SL_{r+1}(\mathbb{R}),\SL_{r+1}(\mathbb{C})\textrm{ or }\SU^{*}_{2r+2} , then \hd_{R}(\Gamma)<n-\frac{r}{8} . 2. If G=\SO_{r,s}(s\geq r),\SO_{2r}(\mathbb{C})\textrm{ or }\SO_{2r+1}(\mathbb{C}) , then \hd_{R}(\Gamma)<n-\frac{3r}{16} . 3. If G=\Sp_{r}(\mathbb{R}),\Sp_{r}(\mathbb{C}),\SU_{r,s}(s\geq r),\Sp_{r,s}(s\geq r),\SO^{*}_{4r}\textrm{ or }\SO^{*}_{4r+2} , then \hd_{R}(\Gamma)<n-\frac{r}{4} . Corollary 1.5 . Let X,\Gamma,n,r be as in Theorem 1.3 . Suppose the injectivity radius of \Gamma\backslash X is uniformly bounded from below. Then \hd_{R}(\Gamma)<n-\frac{r}{8}. Proof. When G is of type A_{r},B_{r},C_{r},D_{r},(BC)_{r} , this follows already from Corollary 1.4 . In the remaining exceptional cases, the real rank satisfies r\leq 8 . If r<8 , the inequality holds automatically since \hd_{R}(\Gamma)\leq\dim(\Gamma\backslash X)=n . If r=8 (i.e. G is of type E_{8} ), then it follows from [ CMW23 , Corollary 1.12] . ∎ Remark 1 . We believe that the sharp bound is exactly \hd_{R}(\Gamma)\leq n-r , as suggested in the following example. Take G=\SL_{r+1}(\mathbb{R}) and \SL_{r}(\mathbb{R})\times\mathbb{R}<G via the block diagonal embedding. Take \Gamma_{0}<\SL_{r}(\mathbb{R}) and \mathbb{Z}<\mathbb{R} two cocompact lattices, and take \Gamma=\Gamma_{0}\times\mathbb{Z} . Then \Gamma<\SL_{r}(\mathbb{R})\times\mathbb{R}<G is an infinite volume discrete subgroup in G . It follows that \hd_{\mathbb{Z}}(\Gamma)=n-r , where n=\dim(\SL_{r+1}(\mathbb{R})/\SO_{r+1}) . This example shows that the sharpest bound cannot supersede \hd_{R}(\Gamma)\leq n-r . Remark 2 . We also believe that the condition on the injectivity radius can be removed. This amounts to saying that the homological dimension contributed by the “zero thin at infinity” is bounded by n-r , which requires further investigation on the structure of the thin part for general infinite volume locally symmetric spaces. Acknowledgements. This work was partially supported by grants Simons Foundation/SFARI-965245 (CC) and NSF DMS-1812153 (DBM) and NSFC No.12301085 (SW). We would like to especially thank Ben Lowe for pointing out the possible improvement of our previous work, which led to the main result of this paper. 2 Proof of Theorem 1.3 –Real split cases The proof strategy follows the general scheme of [ CMW23 , Theorem 1.11] , together with an improved argument suggested by [ FL24 , Proposition 5.27] , and in fact the case of \SL_{r+1}(\mathbb{R}) is already proved in [ FL24 , Proposition 5.27] . By [ CMW23 , Theorem 4.10] , j_{X}(\Gamma)\leq\min\{k:\operatorname{tr}_{k}L_{z^{\prime}}-\delta_{\phi}(\Gamma)>0\}, for any choice of \phi\in\mathfrak{a}^{*} which extremizes z\in\mathcal{C}_{\Gamma} . Now since \psi_{\Gamma}\leq 2\rho-\Theta according to [ LO22 , Thm 7.1] , we can choose \phi parallel to 2\rho-\Theta , and that there exists 0<\lambda\leq 1 such that \phi:=\lambda(2\rho-\Theta) extremizes some z\in\mathcal{C}_{\Gamma} . Then it follows from [ CMW23 , Proposition 4.7] that z^{\prime}=\phi^{*}/{|\phi^{*}|}=(2\rho-\Theta)/|2\rho-\Theta| . On the other hand, by the same proposition, we have \delta_{\phi}(\Gamma)=\frac{\psi_{\Gamma}(z)}{\langle z^{\prime},z\rangle}\leq\frac{(2\rho-\Theta)(z)}{\frac{2\rho-\Theta}{|2\rho-\Theta|}(z)}=|2\rho-\Theta|. Thus, finding the optimal k such that \operatorname{tr}_{k}L_{z^{\prime}}>\delta_{\phi}(\Gamma) amounts to solve the inequality that \alpha_{r+1}(z^{\prime})+\dots+\alpha_{k}(z^{\prime})>|2\rho-\Theta| where \{\alpha_{i}|r+1\leq i\leq n\} is the set of all positive roots in the order \alpha_{r+1}(z^{\prime})\leq\alpha_{r+2}(z^{\prime})\leq\dots\leq\alpha_{n}(z^{\prime}) . Note that \alpha_{r+1}+\dots+\alpha_{n}=2\rho , so the inequality \alpha_{r+1}(z^{\prime})+\dots+\alpha_{k}(z^{\prime})>|2\rho-\Theta| is equivalent to \langle\alpha_{k+1}+\dots+\alpha_{n},2\rho-\Theta\rangle<\langle\Theta,2\rho-\Theta\rangle. Now if we denote \ell=\max_{r+1\leq i\leq n}\langle\alpha_{i},2\rho-\Theta\rangle , then the minimal k such that \operatorname{tr}_{k}L_{z^{\prime}}>\delta_{\phi}(\Gamma) satisfies k\leq n+1-\frac{\langle\Theta,2\rho-\Theta\rangle}{\ell}. Therefore, it follows immediately from the definition that j_{X}(\Gamma)\leq n+1-\frac{\langle\Theta,2\rho-\Theta\rangle}{\ell}. (1) The rest of the proof is to compute by cases the quantity \frac{\langle\Theta,2\rho-\Theta\rangle}{\ell} . However, for readability concern, here we only present the proof in the case when G is real split. These are exactly the cases where all roots have multiplicity one, but the computations are representable enough so that we can simply leave the remaining cases to Section 3 . According to [ Kna02 , Appendix C] and [ Oh02 , Appendix] , the root data (including the set of simple roots \{\alpha_{1},\dots,\alpha_{r}\} , the set of positive roots \Delta^{+} , the sum of all positive roots 2\rho ), and the half sum of the roots in a maximal strongly orthogonal system \Theta is listed in Table 1 . (\mathfrak{g},\mathfrak{a}) A_{r} B_{r} \{\alpha_{1},\dots,\alpha_{r}\} \{e_{1}-e_{2},e_{2}-e_{3},\dots,e_{r}-e_{r+1}\} \{e_{1}-e_{2},e_{2}-e_{3},\dots,e_{r-1}-e_{r},e_{r}\} \Delta^{+} \{e_{i}-e_{j}|i<j\} \{e_{i}\pm e_{j}|i<j\}\cup\{e_{i}\} 2\rho re_{1}+(r-2)e_{2}+\dots+(-r)e_{r+1} (2r-1)e_{1}+(2r-3)e_{2}+\dots+e_{r} \Theta if r is odd \sum_{i=1}^{(r-1)/2}\frac{i}{2}\alpha_{i}+\sum_{i=(r+1)/2}^{r}\frac{r-i+1}{2}\alpha_{i} \sum_{i=1}^{(r-1)/2}i\alpha_{i}+\sum_{i=(r+1)/2}^{r}\frac{r}{2}\alpha_{i} \Theta if r is even \sum_{i=1}^{r/2}\frac{i}{2}\alpha_{i}+\frac{r}{4}\alpha_{r/2+1}+\sum_{i=r/2+2}^{r}\frac{r-i+1}{2}\alpha_{i} \sum_{i=1}^{r/2}i\alpha_{i}+\sum_{i=r/2+1}^{r}\frac{r}{2}\alpha_{i} (\mathfrak{g},\mathfrak{a}) C_{r} D_{r} \{\alpha_{1},\dots,\alpha_{r}\} \{e_{1}-e_{2},e_{2}-e_{3},\dots,e_{r-1}-e_{r},2e_{r}\} \{e_{1}-e_{2},e_{2}-e_{3},\dots,e_{r-1}-e_{r},e_{r-1}+e_{r}\} \Delta^{+} \{e_{i}\pm e_{j}|i<j\}\cup\{2e_{i}\} \{e_{i}\pm e_{j}|i<j\} 2\rho (2r)e_{1}+(2r-2)e_{2}+\dots+2e_{r} (2r-2)e_{1}+(2r-4)e_{2}+\dots+2e_{r-1} \Theta if r is odd \sum_{i=1}^{r-1}i\alpha_{i}+\frac{r}{2}\alpha_{r} \sum_{i=1}^{(r-1)/2}i\alpha_{i}+\sum_{i=(r+1)/2}^{r-2}\frac{r-1}{2}\alpha_{i}+\frac{r-1}{4}(\alpha_{r-1}+\alpha_{r}) \Theta if r is even \sum_{i=1}^{r-1}i\alpha_{i}+\frac{r}{2}\alpha_{r} \sum_{i=1}^{r/2}i\alpha_{i}+\sum_{i=r/2+1}^{r-2}\frac{r}{2}\alpha_{i}+\frac{r}{4}(\alpha_{r-1}+\alpha_{r}) Table 1: Root data of G 1. Type A_{r} : the real split Lie group is G=\SL_{r+1}(\mathbb{R}) . This is already proved in [ FL24 , Proposition 5.27] . To fit in our language and also for the completeness, we include it here. When r=2m+1 is odd, we compute from Table 1 that \Theta=\frac{1}{2}e_{1}+\dots+\frac{1}{2}e_{m+1}-\frac{1}{2}e_{m+2}-\dots-\frac{1}{2}e_{r+1} and 2\rho-\Theta=\frac{4m+1}{2}e_{1}+\frac{4m-3}{2}e_{2}+\dots+\frac{1}{2}e_{m+1}-\frac{1}{2}e_{m+2}-\dots-\frac{4m-3}{2}e_{r}-\frac{4m+1}{2}e_{r+1}. Thus we have \displaystyle\langle\Theta,2\rho-\Theta\rangle \displaystyle=\left(\frac{4m+1}{4}+\frac{4m-3}{4}+\dots+\frac{1}{4}\right)\times 2=\frac{r(r+1)}{4}, and \ell=\langle e_{1}-e_{r+1},2\rho-\Theta\rangle=2r-1 . Therefore, by (

Read the complete text