The Berezin liminf criterion fails for radial Toeplitz operators

This paper disproves the Perälä–Virtanen conjecture by constructing explicit radial symbols on Bergman and Fock spaces where a positive limit inferior of the Berezin transform fails to guarantee essential positivity, demonstrating that the Berezin liminf criterion is insufficient for radial Toeplitz operators in all complex dimensions.

The Gist

Imagine you are a detective trying to figure out if a machine is "safe" (mathematically speaking, "positive"). The machine is a complex mathematical engine called a Toeplitz operator, and it lives inside a special world of functions called a Bergman space or a Fock space.

For a long time, mathematicians believed they had a perfect "smoke detector" to check if this machine was safe. This detector was called the Berezin Transform.

The Old Belief (The Conjecture)

The rule was simple:

"If you look at the machine's output as it gets closer to the edge of the world (or infinity), and the signal stays positive, then the machine is safe."

Mathematicians Perälä and Virtanen proposed this as a universal law for a specific type of machine: one that is radial.

• What does "radial" mean? Imagine a target with concentric circles. A radial machine behaves the same way no matter which direction you look; it only cares about how far you are from the center. It's like a lighthouse beam that gets brighter or dimmer based on distance, but not on the angle.

The conjecture said: If the lighthouse beam looks bright (positive) as you get far away, the whole machine is safe.

The Plot Twist (The Disproof)

Sam Looi, the author of this paper, says: "Not so fast."

Looi built a very specific, tricky machine that breaks this rule. He created a "radial" machine that looks perfectly safe from a distance, but is actually dangerous up close.

The Analogy: The Flickering Lighthouse

Imagine a lighthouse that is supposed to shine a steady, bright white light (representing safety).

1. The Berezin Transform (The Distant Observer): You are standing miles away on a boat. You look at the lighthouse. Because of the way the light waves travel and average out over the distance, the light looks steady and bright to you. Your detector says, "All clear! The light is positive!"
2. The Eigenvalues (The Close-Up Inspector): Now, imagine you are a tiny insect landing on the glass of the lighthouse bulb. You see the truth: the light is actually flickering violently. It's flashing bright white, then dim, then dark, then bright again. In fact, it flashes into total darkness (negative values) so fast that from far away, your eye just sees the average brightness.

The Catch: The "Distant Observer" (Berezin Transform) and the "Close-Up Inspector" (Eigenvalues) are looking at the same flickering light, but they are averaging the flickers differently.

• The distant observer averages the flickers over a wide area, smoothing out the darkness.
• The close-up inspector catches the exact moments of darkness.

Looi proved that for these specific radial machines, the "Distant Observer" can be fooled into thinking the light is safe, while the "Close-Up Inspector" sees that the machine is actually broken (it has negative parts in its essential spectrum).

The "Oscillatory" Secret

How did Looi build this trick? He used a cosine wave (a smooth, repeating wave like a sine wave) as the machine's setting.

• He set the machine to oscillate (flicker) at a very specific speed.
• In the Fock space (a world of infinite expansion), he made the light flicker based on the distance from the center ($|z|$).
• In the Bergman space (a world inside a circle), he made the light flicker based on how close you are to the edge ($1 - |z|^2$).

Because the "Distant Observer" and the "Close-Up Inspector" measure distance in slightly different ways, the flickering gets "attenuated" (dampened) differently for each.

• The Inspector sees the full depth of the dark flashes.
• The Observer sees a dimmed version of the flashes, where the dark parts aren't dark enough to trigger the alarm.

Why This Matters

This paper is a big deal because:

1. It kills a popular theory: It proves that the "Berezin liminf criterion" (the smoke detector rule) is false for these radial machines. You cannot just look at the edge of the world to judge the safety of the machine.
2. It works everywhere: Looi showed this isn't just a fluke in one dimension. It happens in 1D, 2D, 3D, and even higher dimensions.
3. It explains why: It's not just a random failure; it's a fundamental mismatch in how different mathematical tools "average" oscillating signals.

The Takeaway

In the world of complex math, what you see from a distance isn't always the whole truth. Just because a signal looks positive on average at the edge of the universe doesn't mean the machine isn't secretly oscillating into danger right under the hood. Looi found the perfect "flickering lighthouse" to prove that the old rule was wrong.

Technical Summary

Here is a detailed technical summary of the paper "The Berezin liminf criterion fails for radial Toeplitz operators" by Sam Looi.

1. Problem Statement and Context

The paper addresses a fundamental question in the spectral theory of Toeplitz operators on complex function spaces: Can the essential positivity of a Toeplitz operator be determined solely by the asymptotic behavior of its Berezin transform?

• Definitions:
  • Essential Positivity: A self-adjoint operator $T$ is essentially positive if its essential spectrum $\sigma_{ess}(T)$ is contained in $[0, \infty)$.
  • Berezin Transform: For a symbol $f$, the Berezin transform $\tilde{f}(z)$ is the expectation value $\langle T_f k_z, k_z \rangle$, where $k_z$ is the normalized reproducing kernel.
• The Conjecture: Perälä and Virtanen [13] conjectured that for a bounded, real-valued, radial symbol $f$ on the Bergman space $A^2(\mathbb{D})$ (or Fock space $\mathcal{F}^2(\mathbb{C})$), the operator $T_f$ is essentially positive if and only if:
  $$\liminf_{|z| \to \text{boundary}} \tilde{f}(z) \geq 0$$
  (Specifically, $|z| \to 1^-$ for Bergman and $|z| \to \infty$ for Fock).
• The Gap: While previous results established this for symbols with actual boundary limits (using Tauberian arguments), the conjecture for the weaker $\liminf$ condition remained open, particularly whether the radial structure imposes enough regularity to make the criterion hold.

2. Methodology

The author employs a direct, constructive approach using explicit counterexamples rather than indirect operator-theoretic obstructions. The core methodology relies on the specific spectral properties of radial Toeplitz operators:

1. Diagonalization: On both Bergman and Fock spaces, radial symbols $f(z) = \phi(|z|)$ result in Toeplitz operators that are diagonal with respect to the standard monomial basis. The eigenvalues $\lambda_n$ depend only on the total degree of the monomials.
2. Asymptotic Mismatch: The paper identifies a crucial "scale mismatch" between how the eigenvalues and the Berezin transform average the oscillatory symbol:
   • Eigenvalues ($\lambda_n$): Average the symbol against a weight concentrated at a specific radius determined by the degree $n$ (e.g., $r \approx \sqrt{n}$ in Fock, $r \approx 1 - 1/n$ in Bergman).
   • Berezin Transform ($\tilde{f}(z)$): Averages the symbol against a Gaussian (Fock) or Poisson-like (Bergman) kernel centered at $z$.
3. Oscillatory Analysis: The author constructs symbols with high-frequency oscillations. Due to the different asymptotic scales, the oscillations are attenuated by different factors in the eigenvalue sequence versus the Berezin transform. This allows the Berezin transform to remain positive (due to damping) while the eigenvalues dip into negative territory.
4. Reduction to 1D: Despite working in arbitrary complex dimension $d \geq 1$, the radial symmetry reduces the problem to explicit asymptotic analysis of one-dimensional oscillatory integrals.

3. Key Contributions and Results

A. Disproof of the Conjecture in All Dimensions

The paper proves that the Perälä–Virtanen conjecture is false for radial symbols in all complex dimensions $d \geq 1$ for both Bergman and Fock spaces.

1. Fock Space ($\mathcal{F}^2(\mathbb{C}^d)$):

• Counterexample Symbol: $f(z) = \frac{1}{2} + \cos(2|z|)$.
• Result:
  • The Berezin transform satisfies $\liminf_{|z| \to \infty} \tilde{f}(z) = \frac{1}{2} - e^{-1} > 0$.
  • The eigenvalues satisfy $\liminf_{n \to \infty} \lambda_n = \frac{1}{2} - e^{-1/2} < 0$.
• Significance: This counterexample is dimension-free; it works for all $d \geq 1$.

2. Bergman Space ($A^2(\mathbb{B}^d)$):

• Counterexample Symbol: $f(z) = c_d + \cos\left(\log \frac{1}{1-|z|^2}\right)$.
  • Here, the oscillation occurs at the hyperbolic boundary scale $1-|z|^2$.
• Result:
  • The Berezin transform satisfies $\liminf_{|z| \to 1^-} \tilde{f}(z) = c_d - |\beta_d| > 0$.
  • The eigenvalues satisfy $\liminf_{n \to \infty} \lambda_n = c_d - |\alpha| < 0$.
• Dimensional Nuance: Unlike the Fock case, the constant $c_d$ must be adjusted based on the dimension.
  • For $1 \leq d \leq 11$, the constant$c = 1/2$ works.
  • For $d \geq 12$, the offset $c_d$ must be increased to ensure the Berezin transform stays positive while the eigenvalues go negative.
• Mechanism: The eigenvalues average the oscillation at a depth of $\sim 1/n$, while the Berezin transform averages at a depth of $\sim 1-|z|^2$. The specific oscillatory profile $\cos(\log(1/(1-|z|^2)))$ interacts differently with these scales, leading to different attenuation factors ($|\alpha|$ vs $|\beta_d|$).

B. Explicit Asymptotic Formulas

The paper provides precise asymptotic expansions for the eigenvalues and Berezin transforms of these oscillatory symbols, involving Gamma functions and specific attenuation factors:

• Fock: Attenuation factors are $e^{-\beta^2/8}$ (eigenvalues) and $e^{-\beta^2/4}$ (Berezin).
• Bergman: Attenuation factors involve ratios of Gamma functions, specifically $|\alpha| = \sqrt{\pi/\sinh \pi}$ and $|\beta_d| = \sqrt{2}\pi^{-1} \sinh \pi \cdot \dots$ (derived via Euler products).

4. Significance

• Resolution of Open Problems: The paper definitively answers "No" to both Conjecture 1.1 and Question 1.2 posed by Perälä and Virtanen, showing that the $\liminf$ criterion fails even in the highly structured radial setting.
• Clarification of Spectral Mechanisms: It elucidates that the failure is not due to a lack of regularity in the symbol, but rather a fundamental asymptotic scale mismatch. The Berezin transform and the eigenvalue sequence are essentially different "averages" of the same oscillatory data.
• Generality: By providing explicit counterexamples for all dimensions $d \geq 1$, the paper demonstrates that the failure is a robust phenomenon inherent to the geometry of these spaces, not an artifact of low-dimensional behavior.
• Methodological Impact: The reduction of high-dimensional operator problems to explicit 1D oscillatory integrals via radial symmetry offers a powerful template for analyzing spectral properties of Toeplitz operators.

In summary, Sam Looi demonstrates that for radial Toeplitz operators, a positive limit inferior of the Berezin transform is insufficient to guarantee essential positivity, as the essential spectrum can still contain negative points due to the differential attenuation of oscillations in the eigenvalue sequence versus the Berezin transform.