Diffusion on homogeneous ultrametric spaces: the contributions of Alessandro Figà-Talamanca

Wolfgang Woess · 2026

Read Diffusion on homogeneous ultrametric spaces: the contributions of Alessandro Figà-Talamanca by Wolfgang Woess (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-2601-02809.txt.

Introduction

Wolfgang Woess surveys the later work of Alessandro Figà-Talamanca, the Italian mathematician who led his country's harmonic analysis school for many years and died in 2023. From the late nineteen seventies his research centered on harmonic analysis on free groups and trees.

Later he turned to diffusion processes on homogeneous ultrametric spaces, such as local fields and totally disconnected abelian groups, spaces closely tied to trees and their boundaries. In particular he built such diffusion processes from discrete time random walks on trees. Woess gives detailed comments on that part of his work and its abundant related literature, written for a volume of Figà-Talamanca's selected papers.

Catalog record

CREATORWolfgang Woess
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE74 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-02809

Browse the full library · About the Tower of Babel

Full text

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

Diffusion on homogeneous ultrametric spaces: the contributions of Alessandro Figà-Talamanca Wolfgang Woess Abstract . Alessandro Figà-Talamanca (1938-2023) was an influential Italian mathematician, scientific leader of the Italian group of harmonic analysis for many years. Since the late 1970ies, his interest focussed on harmonic analysis on free groups and trees. In the later years of his scientific work he became also interested in diffusion processes on homogeneous ultrametric spaces such as local fields and totally disconnected Abelian groups. This is related with the close connection of those spaces with trees and their boundaries and concerns, in particular, the construction of such processes via discrete-time walks on trees. The present notes provide rather detailed comments on this part of his work and the related, quite abundant literature. This is intended to become part of a volume of selected papers by Figà-Talamanca, accompanied by comments such as the present text. 2020 Mathematics Subject Classification 60B15; 35S05, 43A90, 60J60 1. Introduction These notes provide comments on the following articles of A. Figà-Talamanca and coauthors plus related literature. The late work of Alessandro Figà-Talamanca concerns diffusion processes on ultrametric spaces and their description via random walks on trees. It started with \ocite FT-41, \ycite FT-44 for compact spaces. The main contents of these two papers coincide. Here, we refer mainly to the first one. 1 1 1 The second one was written shortly before the first one, as the notes of a minicourse at a conference, but its proceedings volume appeared much later. The first part of \ocite FT-43 applies similar methods with a slightly different goal in a non-compact setting. The two main papers are \ocite FT-46, \ycite FT-47. There is an ample and diverse literature on the subject of diffusion and Laplace-type operators on ultrametric spaces, with considerably varying notation and many mutual citations which are not necessarily based on mutual reading. As the author of the present notes who was only relatively briefly involved in this fascinating field of work via \ocite Bendikov-et-al2014 and \ocite Bendikov-et-al2019, I will try here to elaborate a common picture, connecting the mentioned papers of Figà-Talamanca with a selection from the large body of other important work and translating different notation and terminolgy into one setting. Since it is not expected that every potential reader is acquainted with the relation between ultrametric spaces and trees, the exposition is quite detailed, starting with the basics. 2. Ultrametric spaces and trees An ultrametric space is a metric space (\mathcal{X},\rho) which in place of the triangle equality satisfies the stronger utrametric inequality \rho(x,z)\leq\max\{\rho(x,w),\rho(w,z)\}. As a consequence, if y\in B(x,r) , the closed ball with centre x and radius r , then B(x,r)=B(y,r) . Closed balls are either disjoint or one contains the other. We always assume that the space is separable and that closed balls are compact. This implies that for any x\in\mathcal{X} , the set \Lambda(x)=\{\rho(x,y):y\neq x\} does not accumulate in (0\,,\,\infty) , whence \Lambda=\bigcup_{x\in\mathcal{X}}\Lambda(x) is countable. Here we also assume that \mathcal{X} has no isolated points. The classical example is the field of \mathsf{p} -adic numbers, where \mathsf{p} is a prime. Before further examples, we explain the well-known realisation of ultrametric spaces as boundaries of trees. There are many references for this basic equivalence, e.g. \ocite Rammal, also \ocite FT-42. Let T be an infinite, locally finite tree with set of non-oriented edges E=E_{T} . We write u,v,\dots for its vertices and [u,v]\equiv[v,u] for edges, in which we also write u\sim v , the neighbourhood relation. The degree \deg(u) of a vertex is the number of its neighbours. The graph distance d(u,v) is the length (number of edges) of the unique minimal path (geodesic arc) \pi(u,v) between the vertices u and v . A geodesic ray in T is a one-sided infinite path [u_{0}\,,u_{1}\,,u_{2}\,,\dots] where u_{i}\sim u_{i-1} and no vertex appears twice. Two geodesic rays are called equivalent, if their symmetric difference is finite, i.e., they differ by finite initial pieces. An end of T is an equivalence class of rays, and the set of all ends of T is the boundary \partial T , later to be denoted \mathcal{X} . We write x,y,z etc for the elements of \partial T . For every u\in T and x\in\partial T there is precisely one geodesic ray \pi(u,x) which starts at u and represents x . A) Compact case. We choose a reference vertex (origin) o\in T . For any vertex u we write |u| for its graph distance from o . If u\neq o then the predecessor u^{\prime} of u is the neighbour of u on \pi(o,u) . If x,y\in T\cup\partial T are distinct, then x\wedge y denotes the common vertex on \pi(o,x)\cap\pi(o,y) which is furthest from o , that is, \pi(o,x)\cap\pi(o,y)=\pi(o,x\wedge y). We now choose any real number q>1 and define \rho(x,y)=\begin{cases}0\,,&\text{if}\;x=y\\ q^{-|x\wedge y|}\,,&\text{if}\;x\neq y.\end{cases} This makes \mathcal{X}=\partial T into a compact, separable ultrametric space. Closed (open-compact) balls in this metric are as follows: let u\in T , and let \partial T_{u}=\{x\in\partial T:\pi(o,x)\;\text{passes through}\;u\}. This is the closed ball B(x,q^{-n}) , where n=|x| . In particular, the induced topology does not depend on the choice of q , and there are various other ways to define a topologically equivalent ultrametric with the same collection of compact-open balls. We also see the following: if there is u\in T\setminus\{o\} with \deg(u)=2 then \partial T_{u}=\partial T_{u^{\prime}} , so that this ball of radius q^{-|u|} also has radius q^{-|u|+1} . We avoid this: for the compact case, we assume that \deg(u)\geq 3 for all u\in T\setminus\{o\} . Conversely, start with a compact ultrametric space \mathcal{X} without isolated points. Consider the set T=\{u=B(x,r):x\in\mathcal{X},r\in\Lambda(x)\}. It is countable. Let r_{0}=\max\Lambda=\max\Lambda(x) , so that o:=B(x,r_{0})=\mathcal{X} for every x . Recall that balls are open-compact. Each ball u=B(x,r)\in T\setminus\{o\} has a unique predecessor ball u^{\prime}=B(x,r^{\prime}) where r^{\prime}=\min\Lambda(x)\cap(r\,,\,\infty) , and the number of successors of u , i.e., the balls whose predecessor is u , is finite. It is always \geq 2 , since we assume that \mathcal{X} has no isolated points. This gives T the structure of an infinite tree where \deg(u)\geq 3 for all vertices u\neq o . Any end of T is described by an infinite, decreasing sequence of balls. Its intersection is a unique x\in\mathcal{X} which is a centre of each of those balls. If x,y\in\mathcal{X} are distinct, then in T their confluent u=x\wedge y is u=B(x,r)=B(y,r) , where r=\rho(x,y) (not necessarily of the form q^{-|x\wedge y|} , according to the choice of the ultrametric). For later use, we select one elemement of \mathcal{X}\equiv\partial T which we denote \mathit{0} (later on to become a neutral element) and write \pi(o,\mathit{0})=[o=o_{0}\,,o_{1}\,,o_{2}\,,\dots] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Read the complete text