A Characterization of Borel Measures which Induce Lipschitz-Free Space Elements

Basic Information

• Paper ID: 2412.13319
• Title: A Characterization of Borel Measures which Induce Lipschitz-Free Space Elements
• Author: Lucas Maciel Raad (Federal University of São Paulo, Brazil)
• Classification: math.FA (Functional Analysis), Secondary: 03E55 (Set Theory)
• Publication Date: December 2024 (arXiv v4: November 23, 2025)
• Paper Link: https://arxiv.org/abs/2412.13319

Abstract

This paper resolves a fundamental question posed by Aliaga and Pernecká concerning Lipschitz-free spaces: For a Borel measure $\mu$ on a complete metric space $M$ satisfying $\int d(m,0)d|\mu|(m) < \infty$, does $\mu$ necessarily induce a weak* continuous functional $L_\mu \in \mathcal{F}(M)$ via the mapping $L_\mu(f) = \int f d\mu$?

The author provides a complete characterization: $L_\mu \in \mathcal{F}(M)$ if and only if $\int d(m,0)d|\mu|(m) < \infty$ and $\mu$ is concentrated on a separable subset of $M$. For complete metric spaces, this indeed implies inner regularity. Furthermore, the author proves that the existence of counterexamples is equivalent to the existence of real-valued measurable cardinals, and therefore counterexamples cannot be proven to exist within the ZFC axiom system.

Research Background and Motivation

Problem Background

Lipschitz-free spaces $\mathcal{F}(M)$ are Banach spaces that encode the metric structure of a metric space $M$, with important applications across multiple mathematical fields:

• Metric geometry
• Harmonic analysis
• Machine learning
• Optimal transport theory

One construction method involves the closed linear span of Dirac measures, which suggests a deep connection between measure theory and Lipschitz-free spaces.

Core Problem

From results by Aliaga and Pernecká 3, it is known that:

• Sufficient condition (Proposition 4.3): If $\mu$ is a Borel measure with $\int d(m,0)d|\mu|(m) < \infty$, then $L_\mu \in \mathcal{F}(M)^{**}$
• Stronger sufficient condition (Proposition 4.4): If $\mu$ additionally satisfies inner regularity or $M$ is separable, then $L_\mu \in \mathcal{F}(M)$

Problem 2 in 3: Is the inner regularity assumption necessary for $L_\mu \in \mathcal{F}(M)$?

Research Motivation

1. Theoretical completeness: Characterize which Borel measures induce Lipschitz-free space elements, analogous to the Riesz-Markov-Kakutani representation theorem
2. Regularity theory: Explore the relationship between measure regularity conditions and functional analytic properties
3. Set-theoretic foundations: Reveal the deep connection between this problem and large cardinal axioms (real-valued measurable cardinals)
4. Related problems: Partially resolve Aliaga's question of whether sequentially normal functionals must be normal functionals

Core Contributions

The main contributions of this paper include:

1. Complete Characterization Theorem (Theorem 3.6): For a pointed metric space $M$ and Borel measure $\mu$, $L_\mu \in \mathcal{F}(M) \iff \int d(m,0)d|\mu|(m) < \infty \text{ and } \mu\text{ is concentrated on a separable subset}$
2. Inner Regularity Implication (Corollary): For complete metric spaces, "concentration on a separable subset" is equivalent to inner regularity, thus completely answering Problem 2 in 3
3. Set-theoretic Equivalence (Corollary 3.7): The following are equivalent:
   • No nontrivial measure exists on weight $w(M)$
   • Every $\sigma$-finite Borel measure is concentrated on a separable subset
   • Every measure $\mu$ with $\int \rho d|\mu| < \infty$ satisfies $L_\mu \in \mathcal{F}(M)$
4. ZFC Independence Result: The existence of counterexamples is equivalent to the existence of real-valued measurable cardinals, which cannot be proven in ZFC (assuming ZFC consistency)
5. Partial Resolution of Normality Problem: Under the assumption of real-valued measurable cardinals, the author constructs a sequentially normal but non-weak* continuous functional, providing a conditional counterexample to 1 Problem 3.2

Methodology Details

Task Definition

Input:

• Pointed metric space $(M, 0)$ where $0$ is the basepoint
• Borel measure $\mu$ satisfying $\int d(m,0)d|\mu|(m) < \infty$

Output Decision: Whether the functional $L_\mu: \text{Lip}_0(M) \to \mathbb{R}$ defined by $L_\mu(f) = \int f d\mu$ belongs to $\mathcal{F}(M)$ (i.e., is weak* continuous)

Key Concepts:

• $\text{Lip}_0(M)$: Space of Lipschitz functions satisfying $f(0)=0$
• $\mathcal{F}(M) = \overline{\text{span}}(\delta[M])$: Closed linear span of Dirac measures
• $\mathcal{F}(M)^* \cong \text{Lip}_0(M)$ (dual space isomorphism)

Core Technical Approach

1. Necessity Proof ($L_\mu \in \mathcal{F}(M) \Rightarrow \mu$ concentrated on separable subset)

Key Construction: Using approximating function families to prove $\mu$ vanishes on $M \setminus M'$, where $M' = \{m_n : n \in \mathbb{N}^2\} \cup \{0\}$ is countable.

Detailed Steps:

(a) Represent $L_\mu$ as limit of Dirac measures: $L_\mu = \lim_{n_1 \to \infty} \sum_{n_2 \in \mathbb{N}} a_{(n_1,n_2)} \delta(m_{(n_1,n_2)})$

(b) Define approximating functions: For $M_i = \rho^{-1}[[2^{-i}, \infty)]$, construct three functions:

• $f^u_{i,k}(m) = \max(0, 1 - k2^{i+1}d(m, M_i))$: equals 1 near $M_i$
• $f^l_k(m) = kd(m, M')$: equals 0 on $M'$
• $f_{i,k}(m) = \min(f^u_{i,k}(m), f^l_k(m))$: combines both

(c) Key Calculation: $\mu(M_i \setminus M') = \int \chi_{M_i \setminus M'} d\mu = \lim_{k \to \infty} \langle L_\mu, f_{i,k} \rangle = 0$ since $f_{i,k}|_{M'} = 0$, so $\langle L_\mu, f_{i,k} \rangle = 0$

(d) Extension to total variation: Using Hahn decomposition $M = A^+ \cup A^-$, for $B^\pm = A^\pm \cap (M \setminus M')$:

• Use regularity to approximate with closed sets $F^\pm \subset B^\pm$
• Construct Lipschitz function $g$ equal to 1 on $F^+$, -1 on $F^-$
• Prove $|\mu|(M \setminus M') \leq 4\epsilon + \int g d\mu = 4\epsilon$ (letting $\epsilon \to 0$)

2. Sufficiency Proof (Separability $\Rightarrow L_\mu \in \mathcal{F}(M)$)

Direct application of 3 Proposition 4.4: If $\mu$ is concentrated on a separable subset, then $L_\mu$ is weak* continuous.

3. Set-theoretic Connection (Corollary 2.14)

Constructing Counterexamples from Real-valued Measurable Cardinals:

If $w(M) \geq \kappa$ (the smallest real-valued measurable cardinal), then:

• There exists a uniformly discrete set $M' \subset M$ with $|M'| = \kappa$
• There exists a nontrivial measure $\mu'$ on $M'$ (with $\mu'(\{m\}) = 0$ for all $m$)
• Define $\mu = \frac{1}{\rho}d(\mu'|_M)$, then $\int \rho d|\mu| < \infty$ but $\mu$ is not concentrated on a separable subset

Key Properties:

• Real-valued measurable cardinal $\kappa$ satisfies: uncountable, no predecessor, regular
• These properties cannot be proven in ZFC (assuming ZFC consistency)

Technical Innovations

1. Function Approximation Technique: The clever design of $f_{i,k} = \min(f^u_{i,k}, f^l_k)$ simultaneously controls:
   • Approximation to characteristic function on target set $M_i$
   • Vanishing on countable set $M'$ (using representation of $L_\mu$)
2. Measure Decomposition: Decomposing signed measure $\mu$ into positive and negative parts, precisely controlling total variation via Hahn decomposition
3. Topological-Measure Duality: Exploiting the deep connection between weight of metric space (topological concept) and separability of measures (measure-theoretic concept)
4. Large Cardinal Application: Reducing pure analysis problem (weak* continuity of functionals) to set-theoretic problem (existence of real-valued measurable cardinals)

Experimental Setup

Note: This is pure mathematical theory research with no experiments or numerical computations. All results are rigorous mathematical proofs.

Theoretical Verification Methods

The author verifies theoretical correctness through:

1. Constructive Examples (Example 3.5):
   • Space: $M = \{0\} \cup (1,2]$ (non-complete)
   • Conclusion: $\delta_1 \in \mathcal{F}(\overline{M})$ but no Borel measure $\mu$ on $M$ satisfies $L_\mu = \delta_1$
   • Significance: Demonstrates necessity of completeness assumption
2. Logical Consistency Analysis:
   • Prove counterexample existence $\iff$ real-valued measurable cardinal existence
   • The latter is independent of ZFC (Jech 9)
   • Therefore main theorem is optimal within ZFC framework

Experimental Results

Main Theoretical Results

Result 1: Complete Characterization Theorem (Theorem 3.6)

Statement: For a pointed metric space $M$ and Borel measure $\mu$, $L_\mu \in \mathcal{F}(M) \iff \int d(m,0)d|\mu|(m) < \infty \text{ and } \mu\text{ is concentrated on a separable subset}$

Significance:

• Completely answers 3 Problem 2
• For complete spaces, "separability" $\iff$ "inner regularity" (Bogachev 5)
• Provides operational decision criterion

Result 2: Set-theoretic Equivalence (Corollary 3.7)

The following three conditions are equivalent:

1. No nontrivial measure exists on $w(M)$
2. Every $\sigma$-finite Borel measure is concentrated on a separable subset
3. Every $\mu$ with $\int \rho d|\mu| < \infty$ satisfies $L_\mu \in \mathcal{F}(M)$

Implications:

• If $w(M) < \kappa$ (smallest real-valued measurable cardinal), then (3) holds
• If $w(M) \geq \kappa$, counterexamples exist
• Counterexample existence cannot be proven in ZFC (assuming ZFC consistency)

Result 3: Normality Problem (Section 4)

Definition Recall:

• Normal functional: Preserves limits of monotone bounded nets
• Sequentially normal: Holds only for sequences
• Known: Normal $\iff$ weak* continuous (Aliaga-Pernecká 4)

This Paper's Contribution:

• If $w(M) \geq \kappa$, there exists $\mu$ such that:
  • $L_\mu$ is sequentially normal (by Lebesgue dominated convergence)
  • $L_\mu \notin \mathcal{F}(M)$ (not weak* continuous)
  • Therefore $L_\mu$ is not normal

Conclusion: Under "existence of real-valued measurable cardinal" assumption, 1 Problem 3.2 has negative answer

Case Analysis

Example 3.5: Pathological Behavior in Non-complete Spaces

Setup: $M = \{0\} \cup (1,2]$, $\overline{M} = [0,2]$

Observation:

• $\delta_1 \in \mathcal{F}(\overline{M})$ (point 1 in completion)
• But no Borel measure $\mu$ on $M$ satisfies $L_\mu = \delta_1$

Proof Sketch:

• Assume such $\mu$ exists; can assume $\mu$ concentrated on $(1,2]$
• For closed $F \subset (1,2]$, let $k = d(1,F)^{-1}$
• Define $f_n(m) = \max(0, 1 - k2^n d(m,F))$
• Then $\mu(F) = \lim_{n \to \infty} \int f_n d\mu = 0$ (since $L_\mu = \delta_1$)
• By regularity $\mu = 0$, contradiction

Significance: Completeness assumption cannot be omitted

Related Work

Lipschitz-Free Space Theory

1. Godefroy-Kalton 8 (2003):
   • First systematic study of Lipschitz-free spaces
   • Established foundational theoretical framework
2. Weaver 11 (2018):
   • Comprehensive monograph Lipschitz Algebras
   • Standard reference used in this paper
3. Aliaga-Pernecká 3 (2021):
   • Study integral representations of functionals on Lipschitz spaces
   • Posed Problem 2 solved in this paper
   • Proved $\int \rho d|\mu| < \infty \Rightarrow L_\mu \in \mathcal{F}(M)^{**}$
4. Aliaga-Pernecká 4 (2021):
   • Proved normal functional $\iff$ weak* continuous
   • Provides theoretical foundation for Theorem 4.2

Measure Theory and Large Cardinals

1. Bogachev 5 (2007):
   • Standard measure theory reference
   • Source of Theorem 2.16 on measure regularity
2. Jech 9 (2003):
   • Classical set theory textbook
   • Theory of real-valued measurable cardinals (Theorems 2.12, 2.13)

Application Areas

• Metric geometry: Aliaga et al. 2 on purely 1-unrectifiable spaces
• Harmonic analysis: Doucha-Kaufmann 6 on approximation properties on groups
• Machine learning: von Luxburg-Bousquet 10 on distance-based classification

Conclusions and Discussion

Main Conclusions

1. Complete Characterization: $L_\mu \in \mathcal{F}(M)$ if and only if $\int \rho d|\mu| < \infty$ and $\mu$ is concentrated on a separable subset
2. Inner Regularity: For complete spaces, above condition is equivalent to $\mu$ being inner regular
3. ZFC Independence: Counterexample existence is equivalent to real-valued measurable cardinal existence, which is independent of ZFC
4. Normality Problem: Under large cardinal assumption, sequential normality does not imply normality

Limitations

1. Set-theoretic Assumptions:
   • Counterexample construction depends on real-valued measurable cardinal existence
   • Cannot provide concrete counterexamples in standard ZFC
   • "Constructivity" of theoretical results is limited
2. Completeness Requirement:
   • Example 3.5 shows non-complete spaces have complex behavior
   • While Proposition 3.2 allows reduction to complete case, theoretical statements require care
3. Partial Resolution of Normality Problem:
   • 1 Problem 3.2 answered only under large cardinal assumption
   • Problem remains unresolved in ZFC

Future Directions

Open problems posed by the author:

Problem 4.3: Can one prove in ZFC the existence of a metric space $M$ and a sequentially normal but non-normal functional $\phi \in \mathcal{F}(M)^{**} \setminus \mathcal{F}(M)$?

Possible Research Directions:

1. Weakening Assumptions:
   • Find counterexamples not depending on large cardinals
   • Or prove large cardinal assumption necessary
2. Generalization to Other Spaces:
   • Study non-metric spaces (e.g., uniform spaces)
   • Consider weighted Lipschitz spaces
3. Computational Complexity:
   • Algorithmic complexity of deciding $L_\mu \in \mathcal{F}(M)$
   • Effective construction of separable subsets
4. Application Exploration:
   • Applications in optimal transport
   • Kernel methods in machine learning

In-Depth Evaluation

Strengths

1. Theoretical Depth

• Complete resolution of open problem: Thoroughly answers 3 Problem 2
• Interdisciplinary synthesis: Skillfully combines functional analysis, measure theory, and set theory
• Optimality: Proves result is optimal within ZFC framework

2. Technical Innovation

• Function construction technique: Design of $f_{i,k} = \min(f^u_{i,k}, f^l_k)$ is elegant
• Measure decomposition method: Combining Hahn decomposition with regularity
• Reduction technique: Reducing analysis problem to set-theoretic problem

3. Logical Rigor

• Complete and rigorous proofs
• Clear counterexample construction (Example 3.5)
• Logical chain: separability $\Leftrightarrow$ inner regularity (complete case) $\Leftrightarrow$ weak* continuity

4. Writing Quality

• Clear structure: preliminaries → main results → applications
• Consistent notation
• Comprehensive and accurate citations

Weaknesses

1. Constructivity Limitations

• Main counterexamples depend on unprovable-in-ZFC large cardinals
• Cannot provide "computable" counterexamples for practical applications
• This limitation is inherent to the problem, not author's fault

2. Completeness Assumption Handling

• While Proposition 3.2 allows reduction:
  • Example 3.5 shows subtlety of non-complete case
  • Some statements require extra care (e.g., Remark 2.9)

3. Partial Resolution of Normality Problem

• 1 Problem 3.2 remains unresolved in ZFC
• Problem 4.3 posed as new open question

4. Limited Application Discussion

• Focuses primarily on theoretical results
• Limited discussion of concrete applications to specific metric spaces (function spaces, graph spaces)

Impact

Contribution to Field

1. Lipschitz-free space theory:
   • Perfects measure-functional correspondence theory
   • Provides standard characterization for future research
2. Measure theory:
   • Reveals essential importance of separability condition
   • Connects metric space weight to measure regularity
3. Set theory applications:
   • Demonstrates natural application of large cardinals in analysis
   • Provides new example of set-theoretic independence

Practical Value

• Decision criterion: Theorem 3.6 provides operational method
• Theoretical foundation: Supports applications in optimal transport and machine learning
• Cautionary example: Reminds researchers of necessity of separability assumption

Reproducibility

• Fully verifiable: All proofs can be checked step-by-step
• No computation needed: Pure theoretical results
• Standard tools: Uses standard techniques from functional and measure theory

Applicable Scenarios

1. Theoretical Research:
   • Structural theory of Lipschitz-free spaces
   • Generalizations of measure representation theorems
   • Applications of large cardinals in analysis
2. Specific Spaces:
   • Separable metric spaces: all measures with $\int \rho d|\mu| < \infty$ apply
   • Complete non-separable spaces: must verify separability condition
   • Non-complete spaces: requires special care (see Example 3.5)
3. Related Problems:
   • Dual problems in optimal transport
   • Feature maps in kernel methods
   • Linearization of metric spaces

References

Key references cited in this paper:

1. 3 Aliaga & Pernecká (2021): Integral representation and supports of functionals on Lipschitz spaces, IMRN - Poses core problem solved in this paper
2. 4 Aliaga & Pernecká (2021): Normal functionals on Lipschitz spaces are weak continuous*, J. Inst. Math. Jussieu - Normality characterization theorem
3. 5 Bogachev (2007): Measure Theory - Standard measure theory reference, regularity theory
4. 9 Jech (2003): Set Theory: The Third Millennium Edition - Real-valued measurable cardinal theory
5. 11 Weaver (2018): Lipschitz Algebras (2nd ed.) - Comprehensive Lipschitz-free space reference

Summary

This paper is an important contribution to Lipschitz-free space theory, completely resolving the open problem posed by Aliaga-Pernecká. Main achievements include:

1. Providing complete characterization of Borel measures inducing Lipschitz-free space elements (separability condition)
2. Revealing deep connection to real-valued measurable cardinals
3. Proving counterexample existence is unprovable in ZFC

Technically, the author skillfully combines tools from functional analysis (weak* topology), measure theory (regularity, Hahn decomposition), and set theory (large cardinals), demonstrating the power of interdisciplinary approaches in modern analysis.

Although main counterexamples depend on assumptions beyond ZFC, this precisely demonstrates the problem's depth and result's optimality. For practical applications, Theorem 3.6 provides clear decision criterion: check whether measure is concentrated on separable subset.

This paper provides a complete theoretical framework for measure-theoretic foundations of Lipschitz-free spaces, marking an important milestone in the field.