Retraction Theorems for Group Compactifications
Yair Hartman1, Aranka Hrušková2, Mehrdad Kalantar3, Tomer Zimhoni4 · 2025
Read Retraction Theorems for Group Compactifications by Yair Hartman1, Aranka Hrušková2, Mehrdad Kalantar3, Tomer Zimhoni4 (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-2509-10072.txt.
Introduction
Yair Hartman and 3 co-authors characterise when a group compactification admits a symmetry-respecting way to fold the whole space back onto its edge in a 2025 mathematics paper. A retraction that respects the group's own action is a strong demand, and the authors prove an equivariant cousin of Brouwer's classic no-retraction theorem. Large families of compactifications are covered, including the way hyperbolic groups act on their Gromov boundary, where no such folding can exist.
Catalog record
| CREATOR | Yair Hartman1, Aranka Hrušková2, Mehrdad Kalantar3, Tomer Zimhoni4 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-10072
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Retraction Theorems for Group Compactifications Yair Hartman1, Aranka Hrušková2, Mehrdad Kalantar3, Tomer Zimhoni4 Abstract We characterize group compactifications of discrete groups for which there exists an equivariant retraction onto the boundary. In particular, we prove an equivariant analogue of Brouwer’s No-Retraction theorem for large classes of group compactifications, which includes actions of hyperbolic groups on their Gromov boundary. 1 Introduction One of the original inspirations for the modern theory of boundary actions of topological groups was the case of the compact unit disc \mathbb{D} , where the harmonic functions are in correspondence with functions on the boundary \partial\mathbb{D} . This inspired Furstenberg [ Fur63 , Fur73 ] to introduce far-reaching generalizations of compactifications of topological groups, whose resulting boundary theories have proven to be a powerful tool in the study of non-amenable groups and various structures corresponding to them, most prominently in rigidity theory of semisimple groups and their lattices [ Fur73 , Mar78 , SZ94 , Bad+22 ] . Returning to the classical case of the compact disc \mathbb{D} , one of the most celebrated results in the context is Brouwer’s No-Retraction theorem, which states that there is no retraction from \mathbb{D} onto its boundary \partial\mathbb{D} . The main goal of this paper is to investigate analogues of Brouwer’s theorem in the context of group compactifications considered by Furstenberg [ Fur71 ] . Given a countable discrete group \Gamma , by a \Gamma -compactification we mean a compactification \overline{\Gamma} such that the action \Gamma\curvearrowright\Gamma by left multiplication extends to a continuous action \Gamma\curvearrowright\overline{\Gamma} . We focus particularly on the case where the resulting action \Gamma\curvearrowright\partial\overline{\Gamma} of \Gamma on the boundary \partial\overline{\Gamma}:=\overline{\Gamma}\backslash\Gamma is minimal, that is, has no proper closed \Gamma -invariant subspace. Among the most natural examples of this type of compactifications is the Gromov compactification of the free group \mathbb{F}_{n} (or more generally, hyperbolic groups). We prove an equivariant analogue of Brouwer’s No-Retraction theorem in this case. Theorem 1.1 . Let \Gamma be a non-elementary hyperbolic group, and denote by \partial\Gamma its Gromov boundary and by \overline{\Gamma} the Gromov compactification of \Gamma . Then there exists no \Gamma -equivariant retraction \overline{\Gamma}\to\partial\Gamma . Some remarks are in order. First, we note that the condition of equivariance is necessary in this context. Indeed, some retraction \Gamma\cup\partial\Gamma\to\partial\Gamma always exists owing to \Gamma being discrete. Second, we will see that the “no-retraction” property for a \Gamma -compactification is essentially related to the manner of “gluing” of the group \Gamma to the boundary of the compactification. Concretely, we will see that there is an \mathbb{F}_{2} -compactification \overline{\mathbb{F}_{2}} whose boundary \partial\overline{\mathbb{F}_{2}} is the Gromov boundary \partial\mathbb{F}_{2} , and such that there exists a retraction \overline{\mathbb{F}_{2}}\to\partial\overline{\mathbb{F}_{2}} . Let us recall the construction of the \Gamma -compactifications studied by Furstenberg in [ Fur71 ] . Let X be a compact metrizable minimal \Gamma -space, and let \nu\in{\rm Prob}(X) be a contractive measure (i.e., the weak* closure of the orbit \{\gamma\nu:\gamma\in\Gamma\} contains Dirac measures). Define the topology \tau_{\nu} on the disjoint union \Gamma\cup X to be the one that on X restricts to the original topology on X , on \Gamma to the discrete topology on \Gamma , and such that for an unbounded sequence (\gamma_{n}) in \Gamma and x\in X , \gamma_{n}\to x if and only if \gamma_{n}\nu\to\delta_{x} in the weak* topology on {\rm Prob}(X) . In Lemma 3.1 we show that there indeed exists a unique topology on \Gamma\cup X that satisfies the above properties, and it does define a \Gamma -compactification. We refer to this class of \Gamma -compactifications as orbital compactifications, and if \nu=\delta_{x_{0}} is a Dirac measure at some point x_{0}\in X , we call it a point-orbital compactification. Examples 3.10 and 3.11 show that not all \Gamma -compactifications are orbital, however, we prove the following. Theorem 1.2 . Let \overline{\Gamma} be a \Gamma -compactification with minimal boundary \partial\overline{\Gamma} . Then there exists a retraction \overline{\Gamma}\to\partial\overline{\Gamma} if and only if \overline{\Gamma} is point-orbital. 2 Preliminaries Let \Gamma denote a countable discrete group. We denote by \ell^{\infty}(\Gamma) the algebra of all complex-valued bounded functions from \Gamma and by c_{0}(\Gamma) the ideal of all functions “vanishing at infinity”, namely, the space of all those functions f\in\ell^{\infty}(\Gamma) such that for every \varepsilon>0 , the set \{\gamma\in\Gamma:\left|f\left(\gamma\right)\right|>\varepsilon\} is finite. An action \Gamma\curvearrowright X of \Gamma on a compact Hausdorff space X by homeomorphisms induces further canonical actions of \Gamma . First, the action on the compact convex space {\rm Prob}(X) of all Borel probability measures on X by affine homeomorphisms given by (g\cdot\nu)(E)=\nu(g^{-1}E) for \nu\in{\rm Prob}(X) , g\in\Gamma . Second the action on C(X) by *-automorphisms, defined by (g\cdot f)(x)=f(gx) for g\in\Gamma , f\in C(X) . Given a measure \nu\in{\rm Prob}(X) , the corresponding Poisson transform P_{\nu}\colon C(X)\longrightarrow\ell^{\infty}(\Gamma) is defined by P_{\nu}(f)(\gamma)=\int_{X}f\,d\gamma\nu . The transform P_{\nu} is unital, positive and \Gamma -equivariant. Conversely, every unital positive \Gamma -equivariant map \varphi\colon C\left(X\right)\to\ell^{\infty}\left(\Gamma\right) is a Poisson transform, namely, \varphi=P_{\nu} for \nu\in{\rm Prob}(X) which is defined by \nu(f)=[\varphi(f)](e) for every f\in C\left(X\right) . The Poisson transform is multiplicative (hence a *-homomorphism) if and only if \nu is a Dirac measure. By Gelfand duality, a unital \Gamma -invariant C^{*} -subalgebra \mathcal{A} of \ell^{\infty}\left(\Gamma\right) is of the form C(Z) for some compact \Gamma -space Z . Furthermore, if c_{0}\left(\Gamma\right)\subset\mathcal{A} , then \text{spec}\left(\mathcal{A}\right) contains a copy of \Gamma . 2.1 Group compactifications As mentioned in the introduction, in this paper, we restrict our attention to those group compactifications whose boundary is minimal as a \Gamma -space. Definition 2.1 ( \Gamma -compactification) . Let X be a compact minimal \Gamma -space. A topology \tau on the set \Gamma\cup X is called a \Gamma -compactification if 1. the space (\Gamma\cup X,\tau) is compact and Hausdorff, \Gamma is dense in \Gamma\cup X , and the induced topologies \tau|_{\Gamma} and \tau|_{X} are the original topologies on \Gamma and X respectively, 2. the actions \Gamma\curvearrowright\Gamma (by left multiplication) and \Gamma\curvearrowright X together define a continuous action of \Gamma on the union \Gamma\cup X . Our main focus will be on the case where X is metrizable, and for the rest of the paper, unless otherwise stated, X is assumed to be metrizable . Note that in this case, any compactification \Gamma\cup X is also metrizable since by countability of \Gamma , one can choose a countable separating family of continuous functions on \Gamma\cup X . Definition 2.1 is, however, valid for general compact \Gamma -spaces X , and indeed there are important examples of \Gamma -compactifications in which X is non-metrizable (e.g. the Furstenberg boundary compactification). The restriction to the metrizable case simplifies our constructions, however, all our results hold with modified arguments, replacing sequences with nets. When a compact metrizable minimal \Gamma -space X is given, in order to define a concrete topology \tau on the disjoint union \Gamma\cup X that turns it into a \Gamma -compactification, our task, informally speaking, is to describe the “gluing of \Gamma to X ” in a “legal” way: declaring which sequences in \Gamma converge to which points in X in a way that (i) every sequence has a convergent subsequence (compactness); (ii) if a sequence converges to some x\in X , then so do all of its subsequences (well-definedness); (iii) if \gamma_{n}\underset{n\to\infty}{\longrightarrow}x , then for every g\in\Gamma , g\gamma_{n}\underset{n\to\infty}{\longrightarrow}gx