Non-hyperreflexive reflexive spaces of operators

Roman Bessonov, Janko Bračič, Michał ZającStudia 2011 · DOI

TL;DR

This paper explores the boundary between reflexivity and hyperreflexivity in operator theory. The authors successfully construct a $C_0$ contraction and a Jordan block operator that are reflexive but fail to be hyperreflexive, demonstrating that even operators that behave "almost" like finite-dimensional ones can lack this quantitative stability property.

What Problem Does This Paper Address?

In the study of bounded linear operators on a Banach space, two concepts describe how an algebra of operators relates to its invariant subspaces:

  1. Reflexivity: A qualitative property. An operator (or space of operators) is reflexive if it is completely determined by its collection of invariant subspaces.
  2. Hyperreflexivity: A quantitative, "stronger" version of reflexivity. It requires that the distance from an arbitrary operator to the space is proportional to the "Arveson distance" (a measure based on how the operator acts on individual vectors).

While it was known since 1985 that reflexivity does not imply hyperreflexivity, examples of the discrepancy are often complex. Furthermore, in finite-dimensional spaces, the two properties are identical. This paper investigates whether this equivalence holds for $C_0$ contractions—a class of operators on infinite-dimensional Hilbert spaces that share many structural similarities with finite-dimensional operators. The authors aim to find out if "nice" operators like these can be reflexive without being hyperreflexive.

Methods

The authors employ several sophisticated techniques from functional analysis and operator theory:

  • Similarity Transformations: They extend existing results to show that similarity preserves hyperreflexivity in general Banach spaces, not just Hilbert spaces.
  • Geometric Estimates: They use the "angle" between subspaces to estimate the hyperreflexivity constant ($\kappa$), which measures the ratio between the two distance metrics mentioned above.
  • Matrix Analysis: They construct specific sequences of $2 \times 2$ matrices where the hyperreflexivity constant grows to infinity, even as the matrices themselves remain bounded.
  • Complex Function Theory: To move into infinite dimensions, they use Blaschke products (analytic functions on the unit disk) and Jordan blocks. They leverage deep results regarding Riesz bases (specifically Vasyunin’s theorems) to prove that their constructed infinite-dimensional operator remains reflexive while its hyperreflexivity constant "blows up."

Datasets

Not discussed in the paper. As a theoretical mathematics paper, it relies on formal proofs and constructions rather than empirical data.

Key Insights and Ideas

  • The "Quantitative Gap": The core insight is that reflexivity is about the existence of a relationship with invariant subspaces, while hyperreflexivity is about the stability of that relationship. By making the relationship increasingly "unstable" across an infinite sequence of subspaces, one can break hyperreflexivity without losing reflexivity.
  • Finite vs. Infinite: In 2D, an operator is reflexive (and thus hyperreflexive) if and only if it has two distinct eigenvalues. The authors show that by carefully picking these eigenvalues and the "angle" between eigenvectors, they can force the hyperreflexivity constant to grow arbitrarily large.
  • The Role of the Carleson Condition: The authors identify that if a Blaschke product satisfies the Carleson condition (a standard condition in interpolation theory), the resulting Jordan block is hyperreflexive. This highlights that non-hyperreflexivity arises specifically when the zeroes of the minimal function "clump together" in a way that violates this condition.

Key Results

  • Proposition 1.1: Proves that if two spaces of operators are similar via invertible operators $A$ and $B$, hyperreflexivity is preserved, and provides bounds for the new hyperreflexivity constant.
  • Proposition 1.12: Demonstrates the existence of a sequence of $2 \times 2$ matrices $\{A_n\}$ such that their direct sum $T = \bigoplus A_n$ is reflexive but not hyperreflexive.
  • Theorem 2.8 (Main Result): Establishes the existence of a Jordan block operator $S_\theta$ (associated with a specific Blaschke product $\theta$) that is reflexive but not hyperreflexive. This provides a definitive "no" to the question of whether all reflexive $C_0$ contractions are hyperreflexive.

Limitations and Future Work

The paper focuses on providing a counterexample to establish that reflexivity $\neq$ hyperreflexivity for $C_0$ contractions. However:

  • Characterization: While they provide a sufficient condition for hyperreflexivity (the Carleson condition), they do not provide a necessary and sufficient condition.
  • General Banach Spaces: Much of the specific construction for $C_0$ contractions relies on the Hilbert space structure (Hardy spaces, orthogonal sums). Extending these specific counterexamples to more general Banach spaces remains an open area.
  • Constants: The paper provides estimates for the hyperreflexivity constant $\kappa$, but calculating exact values for specific classes of operators remains difficult.

Summary updated

Abstract

We study operators whose commutant is reflexive but not hyperreflexive. We construct a $C_0$ contraction and a Jordan block operator $S_B$ associated with a Blaschke product $B$ which have the above mentioned property. A sufficient condition for hyperrefl
BibTeX
@inproceedings{bessonov2011non,
  author = {Roman Bessonov and Janko Bračič and Michał Zając},
  title = {{Non-hyperreflexive reflexive spaces of operators}},
  booktitle = {Studia},
  year = {2011},
  doi = {10.4064/sm202-1-4},
}
Figure