2505.07667
Subgroup mixing in Baumslag-Solitar groups
Sasha Bontemps
correcthigh confidence
- Category
- Not specified
- Journal tier
- Specialist/Solid
- Processed
- Sep 28, 2025, 12:56 AM
- arXiv Links
- Abstract ↗PDF ↗
Audit review
The paper proves two main theorems: (i) for any symmetric, generating measure with bounded support on BS(m,n), the conjugation action is topologically μ-mixing on K∞, and on every KN when |m|=|n|; (ii) if |m|≠|n|, there exists a finitely supported, symmetric, generating μ for which the action is not topologically μ-mixing on any finite phenotype piece KN. These are stated as Theorems 1.2 and 1.1, respectively, and proved via ends convergence on the Bass–Serre tree (Cartwright–Soardi, with the amenability hypothesis verified for BS(m,n)) and a key deterministic “gluing” lemma (Lemma 3.9) in the mixing direction, and via a biased t vs t^{-1} random walk plus a p-adic growth estimate (Proposition 3.1 and Theorem 3.3) in the non-mixing direction . The model’s Part (2) departs crucially from the paper: it posits zero-mean increments E[Y]=0 yet claims “drift to +∞ along a typical sample path” and uniform-in-time failure of mixing, whereas the paper requires a bias μ(t)≠μ(t^{-1}) to obtain a positive-probability event on which partial sums stay positive for all times, yielding uniform non-mixing at each large time step (Theorem 3.3) . In Part (1), the model also relies on an unjustified “any fixed reduced subword appears many times” claim; the paper instead uses ends convergence, escape of compacts, independence, and Lemma 3.9 to guarantee the needed concatenations with high probability (Equations (8)-(9) and the final gluing step) .
Referee report (LaTeX)
\textbf{Recommendation:} no revision \textbf{Journal Tier:} specialist/solid \textbf{Justification:} The manuscript cleanly establishes a sharp dichotomy for topological μ-mixing in the perfect kernel of BS(m,n): positive results on K∞ (and on each KN in the unimodular case) and a uniform negative result on all finite phenotype pieces in the non-unimodular case. The techniques are standard yet deftly combined: ends convergence on trees, phenotype-based decomposition of (m,n)-graphs, and a robust gluing lemma. The exposition is clear and the results refine the literature on subgroup dynamics beyond acylindrical actions.