A characterisation of probabilistic metrizability for approach spaces

E. Colebunders, R. Lowen · 2026

Read A characterisation of probabilistic metrizability for approach spaces by E. Colebunders, R. Lowen (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-2601-06269.txt.

Introduction

Some spaces track how near things are by degrees, not by a yes or no. Eva Colebunders and 1 co-author ask when such a space can be built from chance-based gaps in a 2026 math paper. Chance-based gaps give a spread of lengths, not one fixed length, for each pair of points.

The first part of the paper solves an open case for these approach spaces. It tells which ones come from chance-based gaps through a known bridge map. The next part does the same job for a twin class, gauge spaces. The last part gives a clean build of the whole class of chance-gap spaces. Old tries used sets of plain gaps and held just in small cases. The new build uses gaps that skip the old triangle rule but meet a mixed form of it.

Catalog record

CREATORE. Colebunders, R. Lowen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE41 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-06269

Browse the full library · About the Tower of Babel

Full text

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

A characterisation of probabilistic metrizability for approach spaces E. Colebunders, R. Lowen Abstract Abstract Characterisations of metrizable topological spaces or metrizable uniform spaces are well known. A natural counterpart to being metrizable for topological spaces can be expressed in terms of probabilistic metrizability for approach spaces. The notion of a probabilistic metrizable approach space is based on a well known concrete functor \Gamma, as introduced in [ 9 ] , from the category of probabilistic metric spaces with respect to a continuous arbitrary t-norm to the category of approach spaces. A characterization of those probabilistic metrizable approach spaces is still missing and in the first part of this paper we solve this problem. A natural counterpart to being metrizable for uniform spaces can be expressed in terms of probabilistic metrizability for uniform gauge spaces. In the second part of the paper we start from another concrete functor \Lambda, as described in [ 7 ] , on the category of probabilistic metric spaces with respect to a continuous t-norm to the category of uniform gauge spaces. In a similar way as for the functor \Gamma we obtain a characterisation of probabilistic metrizability of uniform gauge spaces. The last section of the paper is devoted to an isomorphic description of the category of probabilistic metric spaces. This problem is not new. Previous attempts in providing isomorphic descriptions of the category of probabilistic metric spaces worked with collections of (pseudo)metrics. These attempts were only formulated in restricted cases. Our isomorphic description is in terms of objects that are sets endowed with a collection of distances, where the distances involved do not satisfy the triangle inequality but fulfil a mixed triangle condition instead. Keywords: Probabilistic metric space; Approach space; Local system; Uniform gauge space; Uniform system; Probabilistic metrizability; Category of probabilistic metric spaces; Mixed triangle inequality Mathematics Subject Classification: 54E70; 54E35; 54A05; 54E15. 1. Introduction For many applications the context of approach spaces with contractions is quite suitable as was shown in the context of probability theory [ 1 ] and [ 2 ] , hyperspaces [ 12 ] , functional analysis [ 17 ] or complexity analysis [ 5 ] and [ 6 ] . A first lax-algebraic description of approach spaces was established by Clementino and Hofmann in [ 4 ] . The isomorphic description of \mathsf{App} that was obtained there, became an important example in the development of monoidal topology [ 8 ] , [ 3 ] . In [ 9 ] and [ 7 ] it became clear that linking a probabilistic metric space to an associated approach space is also quite fruitful. In this paper we pursue this path. Recently the authors of [ 9 ] described a concrete functor \Gamma:\mathsf{ProbMet}\rightarrow\mathsf{App}, from probabilistic metric spaces with non-expansive maps, to the category of approach spaces with contractions [ 11 ] , [ 10 ] . From [ 9 ] it is clear that being probabilistic metrizable is a natural counterpart to being metrizable in topological spaces. As in [ 9 ] our setting will be probabilistic metric spaces with respect to some continuous t-norm * on [0,1]. The image through the functor \Gamma consists of those approach spaces that are probabilistic metrizable. Using an expression from [ 9 ] the objects in the image of \mathsf{ProbMet} by \Gamma will be described as sets endowed with a local system. In the first part of this paper we characterise the image of \Gamma in terms of local distances that do not satisfy a triangle inequality but fulfil a mixed triangle condition instead. Every metric space has an underlying uniform structure and the associated theory of metrizable uniform spaces is well known. As a counterpart we will show that there is a natural concrete functor \Lambda:\mathsf{ProbMet}\rightarrow\mathsf{UG} from probabilistic metric spaces with respect to some continuous t-norm * on [0,1] with non-expansive maps to the category \mathsf{UG} with objects defined as sets endowed with uniform systems and with uniform contractions as morphisms. The objects in the image of \mathsf{ProbMet} through \Lambda consists of those spaces that are probabilistic metrizable and will be described as sets endowed with a basis for the uniform system consisting of distances. As in the local case, the distances involved do not satisfy the triangle inequality but fulfil a mixed triangle condition instead. The last section of the paper is devoted to an isomorphic description of the category of probabilistic metric spaces. This problem is not new. Previous attempts in providing isomorphic descriptions of the category of probabilistic metric spaces worked with collections of (pseudo)metrics. These attempts were only formulated in restricted cases, for the minimum t-norm M, or for a weaker form of probabilistic metric spaces, see for instance [ 14 ] or [ 15 ] . Our isomorphic desription is in terms of objects that are sets endowed with a collection of distances labeled by the unit interval ]0,1] and by morphisms that are levelwise non-expansive. Again the distances involved do not satisfy the triangle inequality but fulfil a mixed triangle condition instead. 2. Preliminaries For more information on probabilistic metric spaces we refer to [ 15 ] and for the link with approach spaces to [ 9 ] and [ 7 ] . A function d:X\times X\rightarrow[0,\infty] that is zero on the diagonal is called a distance. This notion is not to be confused with distances between points and sets which also play an important role in approach theory but will not be encountered here. A distance that is symmetric is called a symmetric distance. A generalised metric (quasi-metric in [ 11 ] ) on a set X is a map d:X\times X\rightarrow[0,\infty] which is zero on the diagonal and satisfies the triangle inequality. If moreover d is symmetric we call it a symmetric generalised metric . We recall some terminology from [ 15 ] and [ 9 ] . A function \varphi:[0,\infty]\rightarrow[0,1] is called a distance distribution if \varphi is monotone, \varphi(0)=0,\varphi(\infty)=1 and \varphi is left-continuous on ]0,\infty[ . A binary operation * on the interval [0,1] is a continuous t-norm if ([0,1],*,1) is a commutative monoid, which is continuous as a function on [0,1]^{2} to [0,1] with respect to the usual topologies and satisfies p*q\leq p^{\prime}*q^{\prime} whenever p\leq p^{\prime} and q\leq q^{\prime} in [0,1]. A probabilistic metric space is a set X endowed with a map (1) \alpha:X\times X\times[0,\infty]\rightarrow[0,1] and a continuous t-norm * such that (P1) \alpha(x,y,-):[0,\infty]\rightarrow[0,1] is a distance distribution, (P2) \alpha(x,x,-)=\varepsilon_{0}, the largest distance distribution, (P3) \alpha(x,y,r)=\alpha(y,x,r), (P4) \alpha(x,y,-)=\varepsilon_{0}\Rightarrow x=y, (P5) \alpha(y,z,r)*\alpha(x,y,s)\leq\alpha(x,z,r+s), for all x,y,z\in X,r,s\in[0,\infty]. A map f:(X,\alpha)\rightarrow(Y,\beta) between probabilistic metric spaces is non-expansive if (2) \alpha(x,x^{\prime},t)\leq\beta(f(x),f(x^{\prime}),t), for all x,x^{\prime}\in X,t\in[0,\infty]. In this paper we will always assume that the probabilistic metric spaces are defined with respect to some continuous t-norm which we denote by * . The category of probabilistic metric spaces with non-expansive maps is denoted by \mathsf{ProbMet}. Next we recall some notions from approach theory. For more details on concepts and results on approach spaces we refer to [ 11 ] or [ 10 ] . We recall terminology and basic results that will be needed in this paper. 2.1 Definition . A collection of ideals (\mathcal{A}(x))_{x\in X} in [0,\infty]^{X} indexed by the points of X is a

Read the complete text