Sobolev-to-Lipschitz property of geodesically complete spaces with curvature bounded from above

Emanuele Caputo, Nicola Cavallucci, Toni Ikonen · 2026

Read Sobolev-to-Lipschitz property of geodesically complete spaces with curvature bounded from above by Emanuele Caputo, Nicola Cavallucci, Toni Ikonen (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-2606-24374.txt.

Introduction

Caputo, Cavallucci, and Ikonen prove that rough maps on curved spaces secretly have well-behaved representatives. The setting is length spaces with curvature bounded from above. This flexible Alexandrov-style class includes many singular geometries. Sobolev maps into such spaces are defined through energy rather than derivatives. The Sobolev-to-Lipschitz question asks whether each such map agrees almost everywhere with a Lipschitz map whose constant matches its energy.

For the exponent-infinity energy the answer is yes. The team proves it in 2026, with the Lipschitz constant equalling the map's infinity energy. The geometric proof perturbs geodesics by arbitrarily small amounts, to dodge the singular set. A comparison curve of zero length is arranged there. Consequences matter for the theory of these spaces. The infinity Poincaré inequality follows, along with an essential assumption behind Lipschitz-volume rigidity results. Caputo's theorem gives analysis on singular curved spaces a firmer foundation.

Catalog record

CREATOREmanuele Caputo, Nicola Cavallucci, Toni Ikonen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.24374

Browse the full library · About the Tower of Babel

Full text

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

Sobolev-to-Lipschitz property of geodesically complete spaces with curvature bounded from above Emanuele Caputo, Nicola Cavallucci, Toni Ikonen Abstract . We prove that every length space with curvature bounded from above that is geodesically complete has the Sobolev-to-Lipschitz property with exponent infinity. That is, every Sobolev map in the W 1 , ∞ W^{1,\infty} -space has a Lipschitz representative so that the Lipschitz constant coincides with the infinity energy of the map. The proof is geometric and relies on arbitrarily small perturbations of geodesics to a curve that has zero length on the singular set. The motivation is to develop the analytic theory of such spaces; in particular, our result implies that GCBA spaces satisfy the infinity Poincaré inequality and an essential assumption in the theory of Lipschitz–Volume rigidity. 2020 Mathematics Subject Classification 46E36, 30L99, 46E35 1. Introduction Metric spaces with curvature bounded from above ( {\sf CBA} ) have been introduced by Alexandrov in [ 1 ] , in terms of comparison of triangles with the simply connected two-dimensional spaces of constant sectional curvature. The flexibility of this class prevents a comprehensive study of their local geometry, see ( [ 24 , Example 4.2] ) and the discussion in their introduction. For these reasons, Lytchak–Nagano studied in [ 24 ] the local geometry of a suitable subclass, namely locally compact, locally geodesically complete spaces with curvature bounded from above ( {\sf GCBA} ). Geodesically complete means that every local geodesic can be extended to a longer local geodesics from both ends. Under these assumptions, for every point x in a {\sf GCBA} space ({\rm X},{\sf d}) , there exists an integer n_{x}\in\mathbb{N} such that the Hausdorff dimension of every small enough ball around x is exactly n_{x} ( [ 24 , Theorem 1.2] ). The quantity n_{x} is called the dimension of x . For k\in\mathbb{N} , we denote by {\rm X}^{k} the set of points of dimension k . Then, for every k\in\mathbb{N} , the set {\rm X}^{k} contains a dense and open subset \mathcal{R}^{k} which is locally biLipschitz to \mathbb{R}^{k} ( [ 24 , Theorem 1.2] ). The singular set is \mathcal{S}:=\bigcup_{k\in\mathbb{N}}{\rm X}^{k}\setminus\mathcal{R}^{k} . With this stratification of {\rm X} , we consider the canonical measure \mu\coloneqq\sum_{i=1}^{\infty}\mathcal{H}^{k}\llcorner{{\rm X}^{k}} that is positive on open sets and locally finite ( [ 24 , Theorem 1.4] ). Moreover, \mu(\mathcal{S})=0 , see [ 24 , Theorem 1.2] . We also note that the measure is infinitesimally doubling : \limsup_{r\rightarrow 0^{+}}\frac{\mu(B(x,2r))}{\mu(B(x,r))}<\infty\quad\text{for almost every $x\in X$.} We emphasize that for a connected {\sf GCBA} space {\rm X} , the measure \mu is (locally) doubling if and only if {\rm X} is purely dimensional; see [ 10 , Theorem 5.2] . The main goal of this paper is to initiate the study of analytical properties of a {\sf GCBA} space as a metric measure space ({\rm X},{\sf d},\mu) . In general, several analytic properties are not well-understood. These include local properties such as local (1,p) -Poincaré inequalities (see [ 20 , Chapters 5-9] ), other potential-theoretic tools such as capacity estimates (see [ 19 , 5 ] ), or other strong connectivity properties of the space (see [ 9 , 8 ] and references therein). Part of the difficulty in developing an analytic toolkit in this context, is the vastness of the class. However, it is clear that e.g. the validity of local Poincaré inequality depends highly on the local geometry of the {\sf GCBA} space of interest. For instance, the wedge sum of n -planes is a {\sf GCBA} space with a local (1,p) -Poincaré inequality valid only for p>n — see [ 21 , Section 6.12] for related discussion. On the other hand, this leads to the natural question whether a local (1,\infty) -Poincaré inequality could be valid on the whole class (as raised in [ 18 , Remark 7.4] ). To get an idea on some basic examples, consider the following. (1) {\sf GCBA} is stable under gluings of {\sf GCBA} spaces along isometric locally complete convex subsets, see [ 6 , Theorem II.11.3] . For example, a gluing of n - and n^{\prime} -dimensional planes along a k -plane is of non-constant dimension if n\neq n^{\prime} and there exists an n -dimensional singular set of Hausdorff dimension k . (2) The spherical suspension of the Poincaré three-sphere is a {\sf GCBA} space. Indeed, the Poincaré sphere has a canonical metric that has curvature bounded from above by one, and dilating the metric suitably and taking the spherical suspension gives the desired example. We also recall that a {\sf CBA} space that is a homology manifold is always {\sf GCBA} , see the proof of [ 6 , Proposition II.5.12] . (3) Every finite simplicial complex has a metric that makes it a \sf{CBA} space, see [ 4 , Theorem 1] . If it has no free faces, it is {\sf GCBA} by [ 6 , Proposition II.5.10] . (4) There is a {\sf GCBA} space that admits no triangulations, see [ 28 , Example 2.7] . To formulate our main result, we recall that a metric measure space ({\rm X},{\sf d},\mu) is an \infty -thick length space if for every x,y\in{\rm X} and N\subseteq{\rm X} with \mu(N)=0 , it holds that {\sf d}(x,y)=\inf\left\{\ell(\gamma)\colon\gamma\textup{ joins }x\textup{ to }y,\,\ell(\gamma\cap N)=0\right\}. Our main theorem shows that a length {\sf GCBA} space is, in fact, an \infty -thick length space. Theorem 1.1 . Let ({\rm X},{\sf d}) be a length {\sf GCBA} space. Then ({\rm X},{\sf d},\mu) is an \infty -thick length space. In particular, the following holds: (i) (Sobolev-to-Lipschitz property) for every u\in W^{1,\infty}({\rm X}) , there exists a Lipschitz representative \tilde{u} such that the global Lipschitz constant \textup{Lip}(\tilde{u}) is equal to the essential supremum \|g_{u}\|_{\infty} where g_{u} is the minimal \infty -weak upper gradient; (ii) (Poincaré inequality) ({\rm X},{\sf d},\mu) satisfies a local (1,\infty) -Poincaré inequality. We refer to Section 2 for the terminology. However, we emphasize that Theorem 1.1 (i) is equivalent to being an \infty -thick length space by [ 18 , Theorem 1.4] while Theorem 1.1 (ii) follows from being an \infty -thick length space by [ 18 , Theorem 1.5] . Hence, being an \infty -thick length space is the main content of Theorem 1.1 . It implies strong connectivity properties for weakly convex subsets of {\sf GCBA} spaces. Corollary 1.2 . Let ({\rm X},{\sf d}) be a {\sf GCBA} space. Then any open and weakly convex subset of {\rm X} is an \infty -thick length space. Here a subset C\subseteq{\rm X} is weakly convex if every x,y\in C can be joined with a geodesic that lies in C and has length equal to that of {\sf d}(x,y) . In particular, the restriction (C,{\sf d}|_{C\times C},\mu\llcorner{C}) is a geodesic length {\sf GCBA} space. Basic examples include small enough balls in any {\sf GCBA} space. It is also clear that the {\sf GCBA} property is stable when passing to the intrinsic distance on an open and connected subset. Thus we have the following corollary. Corollary 1.3 . Let ({\rm X},{\sf d}) be a {\sf GCBA} space. Then any connected and open subset of {\rm X} equipped with the intrinsic distance and the restriction of the reference measure is an \infty

Read the complete text