Ball Separation Characterization of Ball Dentability and Related Properties

Basic Information

• Paper ID: 2307.03621
• Title: Ball separation characterization of ball dentability and related properties
• Authors: Sudeshna Basu (Loyola University, Baltimore), Susmita Seal (NISER Bhubaneswar)
• Classification: math.FA (Functional Analysis)
• Publication Date: July 2023 (arXiv v4: November 11, 2025)
• Paper Link: https://arxiv.org/abs/2307.03621

Abstract

In Euclidean space, every closed bounded convex set can be characterized through two equivalent separation properties. However, this equivalence does not hold in general Banach spaces. This paper provides ball separation characterizations for Banach spaces whose unit balls are dentable and explores related properties.

Research Background and Motivation

Core Problem

In functional analysis, the classical separation theorem guarantees that points outside closed bounded convex sets can be separated by hyperplanes. The ball separation property considers whether points and closed bounded convex sets can be separated using closed balls (rather than half-spaces).

Problem Significance

1. Dimensional Differences: In finite-dimensional spaces, hyperplane separation and ball separation are equivalent; however, in infinite-dimensional Banach spaces, this equivalence fails, revealing profound duality relationships.
2. Geometric Properties: Ball separation properties are closely related to important geometric properties of Banach spaces (such as dentability and the Mazur intersection property).
3. Theoretical Value: Connects local geometric structures of Banach spaces with global separation properties.

Limitations of Existing Research

1. Mazur Intersection Property (MIP): Mazur first studied whether closed bounded convex sets can be represented as intersections of closed balls. Giles et al. proved that MIP is equivalent to w*-denting points on the dual unit ball being norm-dense on the dual unit sphere.
2. Chen-Lin's Work: Established connections between the existence of individual w*-denting or w*-PC points and weak ball separation properties, introducing Property (II).
3. Research Gaps:
   • Lack of systematic ball separation characterizations for the large class of Banach spaces with dentable unit balls
   • Pointwise variants of these properties have not been systematically studied within the ball separation framework

Research Motivation

To fill the above gaps by providing complete ball separation characterizations for Banach spaces with dentable unit balls and extending results to broader classes of geometric properties.

Core Contributions

The main contributions of this paper include:

1. Ball Separation Characterizations of Small Diameter Properties (Section 3): Provides necessary and sufficient conditions for Banach spaces to possess BDP, BHP, BSCSP (and their w*-versions) through the existence of specific functional elements and ball covering properties.
2. Introduction and Characterization of Pointwise Properties (Section 4):
   • Introduces the concepts of semi PC and semi SCS points (generalizing semi denting points)
   • Provides ball separation characterizations of these pointwise properties
   • Proves that Property (II) is equivalent to every point on the dual unit sphere being a semi w*-PC point
3. Study of A-SCS Points (Section 5):
   • Defines A-Small Combination of Slice points, generalizing w*-SCS points
   • Establishes necessary ball separation conditions for the existence of A-SCS points
   • Proves that when the linear span of A-SCS points is dense in the τ_A topology, every closed bounded convex set in the compatible collection A is ball-generated
4. Unified Framework: Integrates multiple results from Chen-Lin, Giles, and others within a unified geometric framework.

Methodology Details

Core Technical Approach

The paper employs a duality method, establishing ball separation properties through correspondence between geometric properties in the original space and functional properties in the dual space.

Main Definitions

1. Small Diameter Properties

• BDP (Ball Dentable Property): The unit ball B_X has slices of arbitrarily small diameter
• BHP (Ball Huskable Property): B_X has non-empty relatively weakly open subsets of arbitrarily small diameter
• BSCSP (Ball Small Combination of Slice Property): B_X has convex combinations of slices of arbitrarily small diameter

2. Pointwise Properties

For x ∈ B_X:

• semi denting point: For any ε > 0, there exists a slice S of B_X such that S ⊂ B(x, ε)
• semi PC point: For any ε > 0, there exists a weakly open set V in B_X such that V ⊂ B(x, ε)
• semi SCS point: For any ε > 0, there exists a convex combination of slices ∑λᵢSᵢ such that ∑λᵢSᵢ ⊂ B(x, ε)

Note: These definitions do not require x to be in the corresponding set, only that the set is contained in a neighborhood of x.

Key Lemmas

Lemma 3.2: If f, g ∈ S_X*, A = {x ∈ B_X : f(x) > ε/2}, and inf g(A) > 0, then ‖f - g‖ < ε

Lemma 3.3: If diam(S(B_X*, x, δ)) ≤ ε, then $\sup_{y \in B_X} \frac{\|x + \frac{\delta}{2}y\| + \|x - \frac{\delta}{2}y\| - 2}{\delta/2} \leq \varepsilon$

These lemmas establish quantitative relationships between slice diameter and norm geometry.

Core Theorem Framework

Theorem 3.4 (Ball Separation Characterization of BDP)

For a Banach space X, the following are equivalent:

1. X* has w*-BDP
2. For ε > 0, there exists x₀* ∈ S_X* such that: for subsets C of B_X satisfying inf x₀*(C) > ε, there exists a closed ball B such that C ⊂ B and inf x₀*(B) > 0
3. (Bidual version)

Proof Strategy:

• (i)⇒(ii): Utilizing the small diameter property of w*-slices, constructing a ball containing C via Lemma 3.3, and proving C is indeed in the ball using proof by contradiction
• (ii)⇒(i): For given ε, constructing the set K_δ = {x ∈ B_X : x₀*(x) > ε+δ}, obtaining ball B from the condition, then constructing w*-slice S and estimating its diameter using Lemma 3.2

Theorem 4.5 (Characterization of semi w*-denting Points)

For x₀* ∈ S_X* and H = {x ∈ X : x₀*(x) = 0}, the following are equivalent:

1. x₀* is a semi w*-denting point of B_X*
2. For bounded sets C ⊂ X** satisfying inf x₀*(C) > 0, there exists a closed ball centered in X such that C ⊂ B and 0 ∉ B
3. For bounded sets C ⊂ X satisfying inf x₀*(C) > 0, there exists a family of balls {Bᵢ} such that C ⊂ ∩Bᵢ and (∩Bᵢ) ∩ H = ∅

This theorem provides three equivalent ball separation characterizations, with the third condition being particularly geometrically intuitive.

Technical Innovations

1. Unified Proof Technique: Systematically transforms properties in the original space to properties in the dual space through relationships between w*-topology and norm topology.
2. Pointwise Method: Introduces "semi" versions of properties, relaxing the requirement that points be in the set, achieving finer characterizations.
3. Compatible Collection Framework: Generalizes ball separation properties to more general families of bounded sets through compatible conditions.
4. New Characterization of Property (II): Proves that Property (II) is equivalent to "every point on the dual unit sphere is a semi w*-PC point" (Theorem 4.14), which is stronger than Chen-Lin's "w*-PC points are norm-dense".

Experimental Setup

As a pure mathematical theory paper, this work does not involve numerical experiments but rather proceeds through:

• Theorem Proofs: Rigorous mathematical proofs
• Counterexample Construction: Example 4.23 provides concrete spaces distinguishing different concepts

Example Analysis (Example 4.23)

1. Space l₁ⁿ:
   • w*-PC(B_l₁ⁿ) = PC(B_l₁ⁿ) = B_l₁ⁿ (all points are PC points)
   • semi denting(B_l₁ⁿ) = {±eᵢ : 1 ≤ i ≤ n} (only standard basis vectors)
   • Demonstrates: existence of semi PC but non-semi denting points
2. Space l₁:
   • 0 is a semi SCS and semi w*-SCS point
   • But 0 cannot be a semi PC or semi w*-PC point
   • Demonstrates: semi SCS does not imply semi PC
3. Space C0,1:
   • The dual space C0,1* has w*-strong diameter 2 property
   • B_C0,1* contains no semi w*-SCS points
   • But has denting points (such as χ_(0,1])
   • Demonstrates: properties in original and dual spaces can differ

Experimental Results

Main Theoretical Results

1. Complete Characterization of Small Diameter Properties

Theorems 3.4, 3.7, 3.11 respectively provide ball separation characterizations of BDP, BHP, BSCSP with unified form:

2. Hierarchical Structure of Pointwise Properties

Establishes the following implication relationships (Example 4.23 proves converses fail):

semi denting ⟹ semi PC ⟹ semi SCS
     ⇕                ⇕            ⇕
semi w*-denting ⟹ semi w*-PC ⟹ semi w*-SCS

3. New Characterization of Property (II)

Theorem 4.14: X has Property (II) ⟺ every point in S_X* is a semi w*-PC point of B_X*

This improves Chen-Lin's result (w*-PC points are norm-dense), providing a stronger equivalent condition.

4. Application of A-SCS Points

Theorem 5.7: If A is a compatible collection and the linear span of A-SCS points is dense in X* under the τ_A topology, then every closed bounded convex set in A is ball-generated.

In particular, taking A to be all bounded sets recovers the result from 2, Proposition 2.5.

Key Findings

1. Fine Structure of Duality: Small diameter properties in the original space correspond precisely to the existence of specific points in the dual space.
2. Power of Pointwise Characterization: "Semi" versions of properties provide, in some cases (such as Property (II)), stronger characterizations than density results.
3. Open Problems:
   • Remark 3.12: The implication (ii)⇒(i) in Theorem 3.11 is unknown
   • Remark 4.2: Does BSCSP guarantee the existence of non-zero SCS or semi SCS points?
   • Remark 4.22: Does every point being semi w*-SCS imply w*-SCS points are dense?

Related Work

Historical Context

1. Mazur (1933): First studied the Mazur intersection property (MIP)
2. Phelps (1960): Established duality between ball separation and dentability in dual spaces
3. Giles-Gregory-Sims (1978): Proved MIP is equivalent to w*-denting points being norm-dense
4. Chen-Lin (1996, 1998):
   • Introduced Property (II)
   • Established connections between existence of individual w*-PC points and weak ball separation properties
5. Giles (2006): Studied the Mazur intersection problem, providing ball separation characterizations of semi denting points

Positioning of This Work

Building on the work of Chen-Lin and Giles, this paper:

1. Systematizes: Provides complete ball separation characterizations of BDP, BHP, BSCSP
2. Pointwise-ifies: Introduces and studies semi PC and semi SCS points
3. Generalizes: Extends to more general families of sets through the A-SCS point framework
4. Unifies: Integrates multiple existing results within a unified framework

Comparison of Related Concepts

Conclusions and Discussion

Main Conclusions

1. Complete Characterizations: Provides necessary and sufficient conditions for ball separation of BDP, BHP, BSCSP and their pointwise versions.
2. New Perspective on Property (II): Offers stronger equivalent characterizations through semi w*-PC points.
3. Generalization Framework: The A-SCS point theory provides tools for studying more general families of sets.

Limitations

1. Incomplete Implication Relationships:
   • The reverse implications in Theorems 3.11 and 4.16 (ii)⇒(i) are unknown
   • The relationship between BSCSP and existence of non-zero SCS points is unclear
2. Restrictions on Compatible Collections: Theorem 5.7 requires A to be compatible, which may be too strong in some applications.
3. Density of w-SCS Points*: Remark 4.22 indicates that whether "all points are semi w*-SCS" implies "w*-SCS points are dense" remains open.

Future Directions

Implicit research directions from the paper:

1. Refine Ball Separation Characterizations: Resolve open problems in Remarks 3.12 and 4.17
2. Relax Compatibility Conditions: Study versions of Theorem 5.7 under weaker conditions
3. Application to Concrete Spaces: Apply abstract results to classical Banach spaces (such as L_p spaces, operator spaces)
4. Quantitative Versions: Study quantitative versions of ball separation properties, such as relationships between the number of balls and ε

In-Depth Evaluation

Strengths

1. Theoretical Depth:
   • Systematically resolves the long-standing problem of ball separation characterization for spaces with dentable unit balls
   • The introduced "semi" concept reveals fine hierarchical structures of properties
   • The new characterization of Property (II) (Theorem 4.14) is an important theoretical contribution
2. Technical Rigor:
   • Proofs are detailed and complete with clear lemma-theorem structure
   • Appropriate use of proof by contradiction, duality arguments, and other techniques
   • Honestly marks open problems (Remarks 3.12, 4.17, 4.22)
3. Unification:
   • Treats BDP, BHP, BSCSP within a unified framework
   • Develops results in original and dual spaces in parallel
   • Generalizes multiple results from Chen-Lin and Giles
4. Rich Examples:
   • Example 4.23 uses concrete spaces to illustrate strict distinctions between concepts
   • Counterexample constructions help understand theoretical boundaries

Weaknesses

1. Readability:
   • Many symbols (BDP, BHP, BSCSP, w*-versions, semi versions, etc.)
   • Motivation section could be more intuitive for non-specialists
   • Lacks geometric illustrations to aid understanding
2. Completeness:
   • Some reverse implications in theorems remain unresolved (e.g., Theorem 3.11)
   • Pointwise characterization of BSCSP is less complete than BDP and BHP
   • Sufficient conditions for A-SCS points are only given for compatible collections
3. Applicability:
   • No application examples to concrete Banach spaces
   • Connections to classical properties (Radon-Nikodým property, Asplund spaces) are insufficiently discussed
4. Novelty:
   • Main techniques (duality method, slice diameter estimation) derive from Chen-Lin and Giles
   • Some results (e.g., Corollary 4.6) are reformulations of known results

Impact

1. Theoretical Contribution:
   • Fills important gaps in ball separation theory
   • The introduction of semi properties provides new tools for subsequent research
   • The new characterization of Property (II) may simplify related proofs
2. Potential Applications:
   • Applicable to Banach space geometric theory
   • Potential connections to optimization theory (convex analysis)
   • May provide insights for operator space theory
3. Reproducibility:
   • As a pure theoretical work, results are highly verifiable
   • Proof techniques can be learned and applied by other researchers

Applicable Scenarios

1. Theoretical Research:
   • Study of geometric properties of Banach spaces
   • Duality theory and separation theorems
   • Convex analysis and nonsmooth analysis
2. Related Fields:
   • Operator theory (geometric properties of operator spaces)
   • Optimization theory (ball representation of feasible sets)
   • Measure theory (geometric characterization of Radon-Nikodým property)
3. Educational Value:
   • Demonstrates the power of duality methods in functional analysis
   • Illustrates essential differences between infinite and finite-dimensional spaces

References

The paper cites 26 references, with key works including:

1. 17 Giles, Gregory, Sims (1978): Established equivalence between MIP and norm-density of w*-denting points
2. 11 Chen, Lin (1998): Systematic study of ball separation properties, introduction of Property (II)
3. 16 Giles (2006): Mazur intersection problem, ball separation characterization of semi denting points
4. 7 Basu, Seal (2022): Authors' prior work on stability of small diameter properties
5. 21 Phelps (1960): Representation theorems for bounded convex sets, early work on ball separation

Overall Assessment

This is a high-quality theoretical paper in functional analysis that systematically resolves several important problems in ball separation characterization. Its main strengths are theoretical depth, technical rigor, and unification; main weaknesses are incomplete results in some areas and lack of concrete applications. The paper makes substantial contributions to Banach space geometric theory, particularly the new characterization of Property (II) (Theorem 4.14) and the A-SCS point framework (Section 5) have considerable theoretical value. Recommended future work should focus on resolving open problems and applying the theory to concrete spaces.