Abstract
We prove that any surjective isometry between unit spheres of the $\ell^\infty$-sum of strictly convex normed spaces can be extended to a linear isometry on the whole space, and we solve the isometric extension problem affirmatively in this case.
Introduction
In the study of functional analysis and the geometry of Banach spaces, a recurring question is the relationship between local geometric properties and global structural characteristics. One of the most intriguing problems in this domain is the isometric extension problem, originally proposed by D. Tingley in 1987. The problem asks: if a surjective isometry exists between the unit spheres of two real normed spaces, must that mapping be the restriction of a linear isometry acting on the entire space?
While the Mazur-Ulam theorem guarantees that any surjective isometry between real normed spaces is affine, Tingley’s problem focuses specifically on mappings restricted to the boundary—the unit sphere. This paper provides an affirmative answer to this problem for a specific and significant class of spaces: the $\ell^\infty$-sum of strictly convex normed spaces.
The Mathematical Context
Tingley's problem is notoriously difficult because the unit sphere lacks the linear structure of the full space. To prove an extension exists, one must demonstrate that the rigid distance-preserving constraints on the sphere's surface are sufficient to dictate the behavior of the mapping throughout the interior and exterior of the sphere.
Prior research has solved this problem for various classical Banach spaces, such as $L^p$ spaces or $C(K)$ spaces. However, general solutions for arbitrary normed spaces remain elusive. This research advances the field by examining spaces constructed via the $\ell^\infty$-sum. In such a space, the norm of an element is defined as the maximum (supremum) of the norms of its components in the underlying strictly convex spaces.
Methodology and Proof Strategy
The authors employ a rigorous logical framework to bridge the gap between the spherical isometry and the global linear isometry. The proof is structured around several key milestones:
1. Antipodal Preservation
A fundamental hurdle in Tingley’s problem is proving that the isometry $V$ preserves antipodal points—that is, $V(-u) = -V(u)$ for all $u$ on the unit sphere. The authors establish this symmetry (Proposition 2.1) for the $\ell^\infty$-sum of strictly convex spaces, which serves as the foundation for the subsequent linear extension.
2. Support Analysis and Component Mapping
The authors analyze the "support" of elements within the sum. In an $\ell^\infty$-sum, an element's support consists of the indices where its components are non-zero. A critical step in the proof involves showing that the isometry $V$ preserves "single-point supports." If an element has only one non-zero component in the domain space, its image under $V$ must also have exactly one non-zero component in the codomain space.
3. Establishing a Bijection
By analyzing these supports, the authors establish a one-to-one correspondence (a bijection) between the index sets of the domain and codomain spaces. This allows the global isometry to be decomposed into a collection of local mappings acting on the individual strictly convex component spaces.
4. Localization and Linear Extension
With the mapping decomposed, the problem is "localized." The authors demonstrate that the isometry acts component-wise. By applying the Mazur-Ulam theorem to these individual components and leveraging the properties of strictly convex norms, they prove that each local mapping is linear. Finally, these local linearities are synthesized to show that the global mapping $V$ is indeed the restriction of a linear isometry.
Key Insights: The Role of Strict Convexity
The success of this proof relies heavily on the property of strict convexity in the component spaces. In a strictly convex space, the unit sphere contains no line segments; the triangle inequality $\|x + y\| = \|x\| + \|y\|$ holds only when $x$ and $y$ are linearly dependent.
The authors exploit this lack of "flat" regions on the component spheres to force algebraic contradictions if the isometry were to deviate from linear behavior. By combining the "flatness" of the $\ell^\infty$ norm (which uses the maximum) with the "roundness" of the strictly convex components, the authors identify a structural rigidity that necessitates a linear extension.
Conclusion
The main result of the paper, Theorem 2.6, confirms that for the $\ell^\infty$-sum of strictly convex normed spaces, every surjective isometry between unit spheres can be extended to a linear isometry on the whole space.
While this represents a significant step forward, the authors acknowledge that the general isometric extension problem remains one of the most challenging open questions in the geometry of Banach spaces. The techniques developed here—specifically the support analysis and the exploitation of strict convexity—provide a potential roadmap for tackling other types of space sums, though the $\ell^\infty$ structure remains a unique and specialized case.
Summary updated