First passage percolation preserves sublinearly Morse boundaries

Sagnik Jana, Yulan Qing · 2026

Read First passage percolation preserves sublinearly Morse boundaries by Sagnik Jana, Yulan Qing (2026) 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-2603-24423.txt.

Introduction

Jana and Qing study sublinearly Morse boundaries in proper geodesic spaces. They give a new characterisation of sublinearly Morse geodesic lines based on middle recurrence. They then turn to first passage percolation on proper geodesic graphs of bounded degree. Each edge gets a random passage time drawn from the same distribution. Under suitable conditions on that distribution, they prove the sublinearly Morse boundary does not change under the percolation. The result links a coarse geometric notion to a random growth model.

Catalog record

CREATORSagnik Jana, Yulan Qing
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE62816
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.24423

Browse the full library · About the Tower of Babel

Full text

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

First passage percolation preserves sublinearly Morse boundaries Sagnik Jana, Yulan Qing Abstract . Sublinearly Morse directions in proper geodesic spaces are defined by sublinearly Morse stability. In this paper we offer an alternative characterization for sublinearly Morse geodesic lines via middle recurrence. We then study first passage percolation (FPP) on proper geodesic graphs of bounded degree. We associate an i.i.d. collection of random passage times to each edge. Under suitable conditions on the passage time distribution, we prove that sublinearly Morse boundaries are invariant under first passage percolation. 1. Introduction A sublinearly Morse boundary [ QRT22 , QRT23 ] is a topological space consists of quasi-geodesic rays in a given metric space. It is constructed to extend the Gromov boundary beyond the hyperbolic setting. If the metric space in question, denoted X , is proper and geodesic, then the sublinearly Morse boundary, denoted {\partial_{s}}X , is always quasi-isometrically invariant and metrizable. In many situations {\partial_{s}}X provides a natural group-invariant topological model compatible with probabilistic boundary theory for random walks on X . A closely related topological set, the Morse boundary [ Cor18 , CS15 ] of metric spaces is often non-metrizable and does not always realize the Poisson boundary. These distinctions make the sublinearly Morse boundary particularly well-suited for studying asymptotic geometry alongside stochastic and dynamical phenomena. It is natural to ask how sublinearly Morse directions behave under random perturbations of the geometry. One natural model for such perturbations is first passage percolation (FPP), which systematically introduce randomness to graphs. In this process, one assigns i.i.d. random weights to the edges, and then defines the distance between two vertices as the smallest total weight among all paths joining them. The path that achieves this minimum is called a geodesic (an optimal path)[see section 2.2 ]. For the Cayley graph of \mathbb{Z}^{2} with respect to the standard generators, this model was introduced by Hammersley and Welsh [ HW65 ] . While first passage percolations on \mathbb{Z}^{2} and, more generally, on \mathbb{Z}^{d} has been studied extensively and proven to be hard. For instance, it is still unknown whether bi-infinite geodesics exist (see [ Kes84 ] ) after first passage percolation on \mathbb{Z}^{2} . Much less is known for other metric spaces until recently. Existence of bi-infinite geodesics and other interesting results [ BZ12 , BT17 ] were obtained when the graph is Gromov hyperbolic. Indeed, [ BT17 ] shows that if X contains a bi-infinite Morse geodesic line, then X_{\omega} contains a bi-infinite geodesic line almost surely. However, one surprise is that [ BJQ25 ] demonstrates that FPP does not, in general, preserve the Morseness of geodesic lines (a property all geodesic lines in hyperbolic spaces enjoy). This motivates the authors of this paper to study the problem in a setting that contains more hyperbolic behavior than \mathbb{Z}^{d} but less than that of a hyperbolic graph. Specifically, we are interested in understand what happends to sublinear Morse directions: recently [ JQ25 ] establishes that if X has a bi-infinite sublinearly Morse geodesic line, then X_{\omega} contains a bi-infinite geodesic line almost surely. Therefore, the immediate motivating question for this paper is: Question 1.1 . Are sublinearly Morseness of sets almost surely preserved under first passage percolations? We answer the question positively: Theorem A . Let X be a graph with uniformly bounded degree and \partial_{s}X\neq\emptyset . Assume that the edge length distribution has finite expectation and also \nu(\{0\})=0 . Then let \omega be induced graphs with weight functions. Then for almost every \omega , such the first passage percolation induces a homeomorphism on the sublinearly Morse boundaries: \partial_{s}X\simeq\partial_{s}X_{\omega} . This result opens up the possibility of analysis into the interaction between sublinearly Morse sets and first passage percolations. We give several natural and immediate questions that can be investigated. Since first passage percolation does not preserve quasi-geodesics, this paper depends on characterizations of sublinear Morseness using paths of bounded slope. This notion is established for Morse geodesic rays and is utilized in [ DMS09 ] and improved in [ ADT17 ] . We first develop an proper analogue of such a notion in the context of sublinear Morseness: Definition 1.2 . We say that a path (geodesic line or quasi-geodesic) \gamma is middle third recurrent if for every C\geq 1 there is a constant c (depends on C and \gamma ) such that the following holds: for any path p with endpoints a,b\in\gamma satisfying \ell(p)\leq Cd(a,b) , intersects the c\cdot\kappa -neighborhood of the middle third of \gamma[a,b] for some sublinear function \kappa . That is, p\cap\mathcal{N}_{\kappa}(\gamma_{\frac{1}{3}[a,b]},c)\neq\emptyset. Where, \gamma_{\frac{1}{3}[a,b]}:=\biggl\{x\in\gamma:\min\{d(x,a),d(x,b)\}\geq\frac{1}{3}\cdot d(a,b)\biggr\} . We are able to establish the path characterization of sublinearly Morse sets: Theorem B . A geodesic ray is sublinearly Morse if and only if it has the middle third recurrent property. We remark that it is communicated to the authors during the preparation of this document that one of the directions of this characterization, Proposition 3.4, is also in the process of being established by Raghunath Jaganathan Vennila as part of his research project. History in the boundary construction The sublinearly Morse boundary is a geometric boundary construction generalizing the Gromov boundary. It is always quasi-isometrically invariant, metrizable, and often provides a group-invariant topological model for suitable random walks. By contrast, the Morse boundary, constructed in [ Cor18 , CS15 ] , frequently fails to be metrizable and does not always realize the Poisson boundary. Recent work also highlights strong genericity properties for sublinearly Morse directions. In [ GQR22 ] , genericity with respect to the Patterson–Sullivan measure was established for actions admitting a strongly contracting element. The stationary measure results of [ GQR22 ] were later claimed in a different context by Choi [ Choi22 ] using a pivoting method of Gouëzel [ Gou22 ] . Following [ Yan22 ] , genericity on the horofunction boundary was proved for all proper statistically convex-cocompact actions [ QY24 ] . Further investigations There are several immediate directions of interest as a result of this study. We record them here and offer also existing related results that may be useful in pursuing these questions. (1) Are hyperbolic graphs mapped to hyperbolic graphs by first passage percolation if the support is finite? A recent preprint by Dominic Bair and the authors of this paper [ BJQ25 ] shows that, assuming the passage time distribution have infinite support, then the hyperbolicity of the graphs are almost surely not preserved. We ask further if the infinite support assumption can be removed or weakened. (2) Does first passage percolation induced a Hölder map on the associated sublinearly Morse boundaries? Theorem A establishes that a first passage percolation of suitable assumption induced a homeomorphism between the sublinearly Morse boundaries. Given that sublinearly Morse boundaries are metrizable topological spaces, we ask if stronger observations can be made. It is established in [ DV15 ] that radial maps induced Hölder map on Gromov boundaries and [ BJQ25 ] showed that first passage percolation exhibits certain radial property. Hölder equivalence is a relation on metric spaces and sublinearly Morse boundaries are known to be metrizable via topological arguments. In an upcoming preprint by Manisha Garg and the authors of this paper, we construct a visual metric for the sublinearly Morse boundary.

Read the complete text