Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

On polynomials in spectral projections of spin operators

Ood Shabtai

TL;DR

This work analyzes the asymptotic behavior of the operator norm of bivariate polynomials evaluated on spectral projections of spin operators, contrasting structured spin-projection pairs with random projection pairs in the semiclassical limit. By developing a comprehensive two-projections framework, the authors derive explicit norm formulas and connect them to principal angles between subspaces, Jacobi-ensemble spectral limits, and Toeplitz operator spectra. They prove that for half-rank spectral projections of spin operators, the norm converges to a universal bound $M_f$, while random pairs can reach this bound only in certain regimes, revealing a Slepian-type concentration phenomenon. The results are supported by detailed proofs across several settings (finite-dimensional spin representations, $l^2$ and $L^2$ spaces) and culminate in a prolate-theorem describing eigenvalue concentration, along with concluding remarks and a conjecture linking the phenomena to Berezin–Toeplitz quantization. Overall, the paper highlights a fundamental, non-generic discrepancy between structured quantum-projection pairs and random projections, with implications for semiclassical analysis and spectral concentration.

Abstract

We show that the operator norm of an arbitrary bivariate polynomial, evaluated on certain spectral projections of spin operators, converges to the maximal value in the semiclassical limit. We contrast this limiting behavior with that of the polynomial when evaluated on random pairs of projections. The discrepancy is a consequence of a type of Slepian spectral concentration phenomenon, which we prove in some cases.

On polynomials in spectral projections of spin operators

TL;DR

This work analyzes the asymptotic behavior of the operator norm of bivariate polynomials evaluated on spectral projections of spin operators, contrasting structured spin-projection pairs with random projection pairs in the semiclassical limit. By developing a comprehensive two-projections framework, the authors derive explicit norm formulas and connect them to principal angles between subspaces, Jacobi-ensemble spectral limits, and Toeplitz operator spectra. They prove that for half-rank spectral projections of spin operators, the norm converges to a universal bound , while random pairs can reach this bound only in certain regimes, revealing a Slepian-type concentration phenomenon. The results are supported by detailed proofs across several settings (finite-dimensional spin representations, and spaces) and culminate in a prolate-theorem describing eigenvalue concentration, along with concluding remarks and a conjecture linking the phenomena to Berezin–Toeplitz quantization. Overall, the paper highlights a fundamental, non-generic discrepancy between structured quantum-projection pairs and random projections, with implications for semiclassical analysis and spectral concentration.

Abstract

We show that the operator norm of an arbitrary bivariate polynomial, evaluated on certain spectral projections of spin operators, converges to the maximal value in the semiclassical limit. We contrast this limiting behavior with that of the polynomial when evaluated on random pairs of projections. The discrepancy is a consequence of a type of Slepian spectral concentration phenomenon, which we prove in some cases.

Paper Structure

This paper contains 19 sections, 19 theorems, 198 equations, 8 figures.

Table of Contents

  1. Introduction
  2. Main results
  3. Examples
  4. Preliminaries on two orthogonal projections
  5. Polynomials in two non-commuting idempotents
  6. The operator norm of polynomials in two projections
  7. The canonical form and angles between subspaces
  8. Proof of Theorem \ref{['grassmann_theorem']}
  9. Proof of Theorem \ref{['spin_theorem']}
  10. Preliminaries
  11. Matrices of spectral projections of spin operators
  12. The $l^2(\mathbb Z)$ settings
  13. The $L^2(\mathbb T)$ settings
  14. The pair $\Pi_\mathbb T, \mathcal{M}_{\mathbbm 1_{E_{\alpha}}}$
  15. Proof of Theorem \ref{['prolate_theorem']}
  16. ...and 4 more sections

Key Result

Theorem 1.1

Let $0 \ne f \in \mathcal{A}$. There exists a constant $M_f > 0$, depending only on $f$, such that $M_f$ is the universal tight upper bound (universal_upper_bound) for $\Vert f(P,Q)\Vert_{\mathop{\mathrm{op}}\nolimits}$, where $P,Q$ are arbitrary orthogonal projections on a separable Hilbert space.

Figures (8)

  • Figure 1: (Y. Le Floch) $\left \Vert \left[P_{1,n}, P_{3,n} \right] \right \Vert_{\mathop{\mathrm{op}}\nolimits}$ as a function of $n$. The apparent mod $4$ behavior is unproven, except for the case $n = 4d+2$ (L. Polterovich).
  • Figure 2: $\left \Vert \left[P, Q \right] \right \Vert_{\mathop{\mathrm{op}}\nolimits}$ for random projections of rank $\lfloor \frac{n}{2} \rfloor$, as a function of $n$.
  • Figure 3: $\left\Vert \left[P_{\lfloor \alpha n \rfloor}, Q_{\lfloor \alpha n \rfloor} \right] \right\Vert_{\mathop{\mathrm{op}}\nolimits}$ as a function of $n$ for random projections. Here, $\alpha = \frac{1}{20}$. The values concentrate about $2(1-2\alpha)\sqrt{\alpha(1-\alpha)} \approx 0.3923$.
  • Figure 4: $\Vert \left[P_{1, \alpha, n}, P_{3,\alpha,n} \right] \Vert_{\mathop{\mathrm{op}}\nolimits}$ as a function of $n$ for $\alpha = \frac{1}{20}$.
  • Figure 5: The sorted eigenvalues of $P Q P\in \mathop{\mathrm{End}}\nolimits\left(V_{P_0} \right)$, where $P,Q \in \mathrm G_{n}(2n)$ are random and $n = 1000$.
  • ...and 3 more figures

Theorems & Definitions (58)

  • Theorem 1.1
  • Theorem 1.2
  • Theorem 1.4
  • Remark 1.6
  • Theorem 1.7
  • Example 1.9
  • Remark 2.1
  • Theorem 2.2: halmos
  • Lemma 2.3
  • proof
  • ...and 48 more