Asymptotic independence for random permutations from surface groups
Yotam Maoz
TL;DR
This work analyzes random n-sheeted covers of a fixed hyperbolic surface X through Hom(Γ,S_n). It proves that, for distinct primitive conjugacy classes, products of fixed-point counts F_n(γ^a) become asymptotically independent as n grows, and characterizes their limiting distributions in terms of independent Poisson sums. The authors develop a robust framework of subcovers, resolutions, and Euler-analytic tools to decompose lift counts and relate them to BR/SBR subcovers, enabling precise asymptotic factorization via a key lemma. They further extend the analysis to short cycle statistics and establish joint convergence to Poisson-derived limits, with a multivariate method of moments ensuring distributional convergence. The results deepen the connection between random covers, fixed-point statistics in random permutations, and combinatorial-geometric structures on surface groups, with implications for high-energy spectral statistics in random surface models.
Abstract
Let $X$ be an orientable hyperbolic surface of genus $g\geq 2$ with a marked point $o$, and let $Γ$ be an orientable hyperbolic surface group isomorphic to $π_{1}(X,o)$. Consider the space $\text{Hom}(Γ,S_{n})$ which corresponds to $n$-sheeted covers of $X$ with labeled fiber. Given $γ\inΓ$ and a uniformly random $φ\in\text{Hom}(Γ,S_{n})$, what is the expected number of fixed points of $φ(γ)$? Formally, let $F_{n}(γ)$ denote the number of fixed points of $φ(γ)$ for a uniformly random $φ\in\text{Hom}(Γ,S_{n})$. We think of $F_{n}(γ)$ as a random variable on the space $\text{Hom}(Γ,S_{n})$. We show that an arbitrary fixed number of products of the variables $F_{n}(γ)$ are asymptotically independent as $n\to\infty$ when there are no obvious obstructions. We also determine the limiting distribution of such products. Additionally, we examine short cycle statistics in random permutations of the form $φ(γ)$ for a uniformly random $φ\in\text{Hom}(Γ,S_{n})$. We show a similar asymptotic independence result and determine the limiting distribution.
