Critical exponents of boomerang subgroups in the free group
Waltraud Lederle · 2025
Read Critical exponents of boomerang subgroups in the free group by Waltraud Lederle (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-2512-20767.txt.
Introduction
Waltraud Lederle works in group theory, the branch of mathematics that studies symmetry and algebraic structure. Her subject is the free group acting on its Cayley tree, a graph that pictures the group. She studies special subgroups called boomerang subgroups and a growth measure known as the critical exponent.
Lederle constructs boomerang subgroups whose critical exponent comes as close as desired to that of any given finitely generated subgroup. The abstract states this single construction result, so the paper's weight rests on how the subgroups are built. For specialists, the result clarifies how flexible these exponents can be inside the free group.
Catalog record
| CREATOR | Waltraud Lederle |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-20767
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Critical exponents of boomerang subgroups in the free group Waltraud Lederle Abstract We construct, for the free group acting on its Cayley tree, boomerang subgroups whose critical exponent is arbitrarily close to the critical exponent of a given finitely generated subgroup. 1 Introduction Let (X,d_{X}) be a proper geodesic Gromov hyperbolic metric space and \mathrm{Iso}(X) its group of isometries. For a discrete subgroup \Gamma\leq\mathrm{Iso}(X) , the critical exponent of \Gamma is defined by \delta(\Gamma):=\inf\big\{s>0\,\big|\,\sum_{\gamma\in\Gamma}e^{-s\cdot d_{X}(x,\gamma(x))}<\infty\big\}, it is independent of the choice of x\in X . Gekhtman–Levit gave the following lower bound for critical exponents of invariant random subgroups. Theorem 1.1 ( [ 3 , Thm. 1.1] ) . Let X be a proper geodesic Gromov hyperbolic metric space and let G<\mathrm{Iso}(X) be a closed, non-elementary subgroup acting cocompactly on X and admitting a uniform lattice. Let \mu be an invariant random subgroup on G such that \mu -almost every subgroup is discrete and infinite. Then, \mu -almost every H\leq G satisfies: 1. \delta(H)>\frac{1}{2}\dim_{H}(\partial X) , and 2. if H is of divergence type, then \delta(H)=\dim_{H}(\partial X) . In a later paper [ 4 ] they give a similar lower bound on critical exponents of stationary random subgroups. In [ 7 , 6 ] the authors introduced boomerang subgroups of countable groups. For a countable group \Gamma , we denote its set of subgroups by \mathrm{Sub}(\Gamma) and endow it with the subspace topology inherited from the obvious embedding \mathrm{Sub}(\Gamma)\subset\{0,1\}^{\Gamma} . With this topology \mathrm{Sub}(\Gamma) is called the Chabauty space of \Gamma . Note that the set of finitely generated subgroups is dense in the Chabauty space, because for every H\in\mathrm{Sub}(\Gamma) we can enumerate its elements as H=\{h_{1},h_{2},h_{3},\dots\} and then H=\lim_{n\to\infty}\langle h_{1},\dots,h_{n}\rangle . The group \Gamma acts on the compact, metrizable, totally disconnected space \mathrm{Sub}(\Gamma) continuously via conjugation. A boomerang subgroup of \Gamma is a subgroup \Delta\in\mathrm{Sub}(\Gamma) such that, for every g\in\Gamma , there exists a sequence (n_{k})\to\infty with \lim_{k\to\infty}g^{n_{k}}\Delta g^{-n_{k}}=\Delta . An invariant random subgroup of \Gamma is a conjugacy-invariant, Borel probability measure \mu on \mathrm{Sub}(\Gamma) . As a simple consequence of the Poincaré recurrence theorem, given any invariant random subgroup \mu of a countable group \Gamma , \mu -almost every subgroup is a boomerang subgroup of \Gamma . This motivates the question which properties that hold almost surely for invariant random subgroups remain true for boomerang subgroups. In this note, we show that the lower bound on the critical exponent does not. Theorem 1.2 . Let F_{d} be the free group of rank d acting on its Cayley tree. Let \Gamma\leq F_{d} be a finitely generated subgroup, O\subset\mathrm{Sub}(F_{d}) a neighborhood of \Gamma and \varepsilon>0 . Then, there exists a boomerang subgroup \Delta\in\mathrm{Sub}(F_{d}) with \Delta\in O and \delta(\Gamma)\leq\delta(\Delta)\leq\delta(\Gamma)+\varepsilon . The proof uses the visualization of boomerang subgroups via Schreier graphs introduced in [ 6 , Section 2] . The critical exponent is tightly related to the growth rate of a subgroup. Consider again F_{d} , the free group of rank d , acting on its Cayley tree \mathcal{T} . We denote by B_{\mathcal{T}}(1,R) the subgraph that is the closed ball of radius R centered at 1 in \mathcal{T} . For g\in F_{d} we denote by |g| the usual word length. We denote c_{\Gamma}(R):=|\Gamma\cap B_{\mathcal{T}}(1,R)|=\{\gamma\in\Gamma\mid|\gamma|\leq R\}. An alternate description of the critical exponent is \delta(\Gamma)=\lim_{R\to\infty}\frac{\ln(c_{\Gamma}(R))}{R}, where the limit exists by Fekete’s lemma. Exponentiating we get b(\Gamma):=e^{\delta(\Gamma)} , and b(\Gamma) measures the growth rate in the sense that c_{\Gamma}(R)\sim b(\Gamma)^{R} . In a recent sequence of papers [ 11 , 12 ] , Louvaris, Wise and Yehuda showed that every number in [1,2d-1] can be realized as the growth rate of a subgroup of F_{d} , and the set of growth rates of finitely generated subgroups of F_{d} is dense. This now immediately gives the following corollary. Corollary 1.3 . Let \mathrm{Boom}(F_{d})\subset\mathrm{Sub}(F_{d}) denote the set of boomerang subgroups. Then, \delta(\mathrm{Boom}(F_{d})) is dense in [0,\ln(2d-1)] , and b(\mathrm{Boom}(F_{d})) is dense in [1,2d-1] . In some sense this corollary means that there are boomerang subgroups that cannot be “typical instances” of invariant random subgroups, or a given stationary random subgroup. A subgroup \Delta\leq F_{d} is confined if there is no sequence (g_{n}) in F_{d} with \lim_{n\to\infty}g_{n}\Delta g_{n}^{-1}=\{1\} . Corollary 1.4 . The set of critical exponents of non-confined boomerang subgroups is dense in [0,\ln(2d-1)] . Fraczyk [ 2 , Corollary 3.2] showed that subgroups of F_{d} with critical exponent \leq\ln(2d-1)/2 cannot be confined. Lower semicontinuity of the critical exponent (see Proposition 3.1 ), together with Theorem 1.1 now immediately implies the following. Corollary 1.5 . There exist boomerang subgroups \Delta<F_{d} such that the only invariant random subgroup supported on the closure of the conjugacy class of \Delta is the Dirac measure at the trivial group. Remark 1.6 . Using results from [ 4 ] about critical exponents one can derive a similar statement involving stationary random subgroups. But still [ 6 , Theorem 1.8] implies that nontrivial boomerang subgroups of the free group are geometrically dense. Corollary 1.5 is another stark contrast to lattices of higher \mathbb{Q} -rank in simple Lie groups, like \mathrm{SL}_{n}(\mathbb{Z}) with n\geq 3 . The main theorem of [ 6 ] shows that every non-trivial boomerang subgroup there is finite and central or has finite index, in particular has only finitely many conjugates and is in particular confined. An underlying philosophical question that motivated this work is: “How are boomerang subgroups more general than typical instances of invariant random subgroups?” In this work, we look at the question from the geometric point of view of the critical exponent. In joint work with Yair Glasner and Tobias Hartnick [ 5 ] , we consider a completely different angle: non-singular dynamics and orbit equivalence. We start from the observation that whenever \mu is a Borel probability measure on \mathrm{Sub}(\Gamma) such that each \gamma\in\Gamma acts on (\mathrm{Sub}(\Gamma),\mu) as a conservative transformation, then \mu -almost every subgroup is a boomerang subgroup. We then construct uncountably many such measures on \mathrm{Sub}(F_{d}) that are singular with respect to every invariant random subgroup. Whether there is a connection to this work is unclear, see Question 2.9 . Acknowledgements. I thank Tianyi Zheng for asking me whether boomerang subgroups can have small critical exponents. I thank Ilya Gekhtman for helpful clarifications and Yair Glasner and Tobias Hartnick for helpful discussions and feedback. Part of this work was done on the LG&TBQ2 conference in Montreal, my travel there was partially funded by Thomas Koberda’s Shannon Fellowship. 2 Schreier graphs and their cores 2.1 Schreier graphs The subgroups of F_{d}=\langle a_{1},\dots,a_{d}\rangle are in 1-1 correspondence with (isomorphism classes of) directed, labeled, rooted graphs, where each vertex has exactly one ingoing and one outgoing edge labeled a_{i} for each i=1,\dots,d . An edge can be ingoing and outgoing for the same vertex, it is then a loop. We call such a graph a Schreier graph . We can write any g\in F_{d} uniquely as reduced word g=b_{1}\dots b_{k}\in F_{d} , where b_{j}\in\{a_{1},\dots,a_{d},a_{1}^{-1},\dots,a_{d}^{-1}\} . We can interpret g as a path 1 1 1 Our paths and cycles are allowed to visit vertices and edges multiple times, and go against the direction of an edge, but have no backtracking unless explicitly allowed to. \omega_{g} in the Schreier graph starting at the root and