The essential spectrum of canonical systems

Christian Remling; Kyle Scarbrough

REVIEW 2 major objections 4 minor 9 references

The essential spectrum of canonical systems

T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Essential spectrum minimum pinned within factor 2.62 of diagonal case

desk verdict A genuinely quantitative upgrade of Romanov-Woracek with clean oscillation proofs; the main caveat is a borrowed lemma from a companion paper, not a fundamental gap. read the letter →

arxiv 1908.02266 v1 pith:K5CRBSBE submitted 2019-08-06 math.SP math-phmath.MP

Christian Remling , Kyle Scarbrough This is my paper · ORCID

classification math.SPmath-phmath.MP MSC 34C1034L4047A06

keywords canonicalsystemessentialspectrumoscillationtheoryPrüferanglediagonalSchrödingeroperatordiscreteRiccatiequation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper studies the smallest absolute value $M(H)$ of the essential spectrum of a half-line canonical system $Ju'=-zHu$. Under the normalization that $\sin\varphi$ is square integrable, it proves that $M(H)$ lies between one half and about $2.62$ times $M(H_d)$, the corresponding diagonal system obtained by deleting the off-diagonal entries. This quantifies a recent qualitative result asserting that a canonical system has purely discrete spectrum exactly when its diagonal part does. For diagonal systems, the paper computes $M(H_d)$ to within a factor of two from the limsup $A$ of $x$ times the tail integral of $\sin^2\varphi$, and shows equality with the threshold where the Schrödinger operators $-d^2/dx^2 - t^2\sin^2\varphi$ have infinitely many negative eigenvalues. A constant-potential example shows the lower factor one half cannot be improved.

What carries the argument

The key tool is oscillation theory for the Prüfer angle $\theta(x)$, which solves $\theta' = t e_\theta^* H e_\theta$. The bottom of the essential spectrum is characterized as $M_+(H)=\inf\{t>0 : \text{the equation is oscillatory}\}$, meaning $\theta(x)\to\pm\infty$; non-oscillatory solutions have finite limits, and under the $\sin\varphi\in L^2$ normalization these limits must be integer multiples of $\pi$. The comparison results follow by bounding the quadratic form in $\theta$ against the corresponding diagonal-system form, with the constant $2/(3-\sqrt5)$ emerging from a quadratic inequality whose optimality condition is $c^2-3c+1=0$. For diagonal systems, the Prüfer equation is transformed into a Riccati equation and then into the Schrödinger equation $-u''-t^2\sin^2\varphi\,u=0$, whose threshold for infinitely many negative eigenvalues is shown to equal $M(H_d)$.

What would settle it

For a fixed $B>0$, set $t^2B=0.24$ and numerically solve $\alpha_1' = (\alpha_1 - 0.24/x)^2$ from a negative initial value at some large $a$; if the solution blows up in finite time, the quoted lemma from [5] and the lower bound of Theorem 1.3 collapse.

Watch

Extended reading notes

Core claim

The central claim is a quantitative comparison theorem: whenever $\sin\varphi \in L^2(0,\infty)$, the essential-spectrum minimum of the original system and of its diagonal part satisfy $\frac12 M(H_d) \le M(H) \le \frac{2}{3-\sqrt5} M(H_d)$. The first inequality is optimal, and the upper constant is conjectured to be replaceable by $1$. In the diagonal case, the paper pins $M(H_d)$ between $1/(2\sqrt{A})$ and $1/\sqrt{A}$, where $A=\limsup_{x\to\infty} x \int_x^\infty \sin^2\varphi(t)\,dt$; this recovers the pure-discrete-spectrum criterion $A=0$ as a special case. The proof converts spectral non-oscillation into the question of whether Prüfer angles of solutions to the canonical system tend to finite limits, and estimates the quadratic form that governs those angles.

Load-bearing premise

The load-bearing premise is a lemma quoted from a companion paper that was still to appear: for the comparison Riccati equation $\alpha' = (\alpha - t^2B/x)^2$, if $t^2B < 1/4$ then a global solution with $\alpha\le0$ exists on a half-line, and the bound $M(H_d)\ge1/(2\sqrt{A})$ collapses if that lemma fails.

Editorial extensions

If this is right

For any canonical system satisfying the $\sin\varphi\in L^2$ normalization, the essential spectrum is absent exactly when it is absent for the diagonal part, and when both are present the two minima differ by a universal factor not exceeding $2.62$.
If $A=\limsup_{x\to\infty} x\int_x^\infty \sin^2\varphi(t)\,dt$ is zero, the diagonal system has no essential spectrum, recovering the known discreteness criterion as a special case; for positive finite $A$, $M(H_d)$ lies between $1/(2\sqrt{A})$ and $1/\sqrt{A}$.
Theorem 1.2 gives a concrete route to computing $M(H_d)$: locate the threshold $S$ where the negative spectrum of $-u''-t^2\sin^2\varphi\,u$ changes from finite to infinite, and $M(H_d)=S$.
The lower comparison factor $1/2$ is sharp, as shown by an explicit constant-potential example with $M(H)=1/4$ and $M(H_d)=1/2$.
One-sided comparison statements fail; both signs of $t$ must be treated together because diagonal systems have spectra symmetric about zero while general systems need not.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the authors' conjecture that the upper constant can be lowered to $1$ is correct, the off-diagonal coefficient $g$ would have no effect on the minimum of $|\sigma_{\mathrm{ess}}|$ at all, making the independence theorem exact in this quantitative sense.
The Schrödinger threshold description suggests a practical numerical method for $M(H_d)$: compute the bottom of the spectrum of $-d^2/dx^2 - t^2\sin^2\varphi$ for increasing $t$ and find where the negative eigenvalues accumulate at zero; this avoids integrating the canonical system directly.
The worst-case constant $2/(3-\sqrt5)$ likely reflects the geometry of two nearly opposite Prüfer angles; testing families of off-diagonal functions $g$ designed to maximize $\theta_+-\theta_-$ growth could show whether constant $1$ fails or holds.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Dear colleague,

The short version: this paper earns its keep. It takes the Romanov-Woracek qualitative criterion for purely discrete spectrum of canonical systems and makes it two-sided and quantitative: M(H) is pinned between 1/2 and about 2.62 times M(H_d), with the lower constant shown optimal, and M(H_d) is bounded by 1/(2√A) and 1/√A where A is a limsup integral. Those statements are new, and the proofs are oscillation-theoretic and mostly very readable.

The central estimates check out. The comparison f(H) ≤ 2f(H_d) is a clean AM-GM step, the quadratic reduction that produces c=(3−√5)/2 is correct, and Theorem 1.3 recovers (1.4) when A=0. The Schrödinger characterization in Theorem 1.2 is a nice bridge between canonical systems and classical oscillation theory, and the Riccati manipulations in Theorems 1.3 and 1.4 are sound. No fitted constants, no circularity.

Now the soft spots, in proportion. The one real issue: the lower bound in Theorem 1.3 leans on a lemma from the companion paper [5], which was still 'to appear' at submission. The lemma is not proved here, and the bound rises or falls with it. A referee should insist on a proof in the paper or an appendix, or at least a precise statement with the lemma included.

Two smaller blemishes. First, the proof of Theorem 1.2 contains a displayed identity that is false: cos²φ sin²θ + sin²φ cos²θ is not equal to sin²θ + cos²θ sin²φ. The inequality that follows it is still true because the error term has the right sign, so this is a typo, but it should be fixed. Second, the sharpness example in Section 5 does not satisfy sinφ∈L², so it doesn't fall under the stated hypothesis of Theorem 1.1. The fix is easy: the first inequality's proof does not use sinφ∈L², so the theorem can be split or the example can be framed as sharpness for the more general inequality. The authors should sort that out.

I also saw a stress-test note claiming that the Prüfer limit assertion in Section 3 is false, citing H=P_0. That concern is wrong. For φ=0, (2.1) is θ'=t sin²θ, and the solutions tend to 0 mod π, not π/2. The paper's one-sentence justification is terse, but a short L¹ argument backs it up: if the limit were not 0 mod π, the integral of f would diverge because ∫cos²φ diverges.

Net: this is a serious paper with real new content. I'd send it to a good referee. The main revision is to make the dependency on [5] explicit and prove the missing lemma, plus the small cleanups.

Referee Report

2 major / 4 minor

Summary. The paper studies the bottom of the essential spectrum, M(H) = min{|t| : t ∈ σ_ess(H)}, for half-line canonical systems J u' = -z H u with trace-one coefficient H. Under the normalization sin φ ∈ L², it proves three quantitative results: Theorem 1.1 compares M(H) with the corresponding quantity M(H_d) for the diagonal system, giving (1/2)M(H_d) ≤ M(H) ≤ (2/(3−√5))M(H_d), with the first inequality shown sharp in Section 5. Theorem 1.2 identifies M(H_d) with the critical coupling S for the Schrödinger operator L(t) = −d²/dx² − t² sin²φ. Theorems 1.3 and 1.4 give explicit bounds on M(H_d) in terms of A = limsup x∫_x^∞ sin²φ and B = liminf x∫_x^∞ sin²φ, recovering the Romanov–Woracek discreteness criterion as the case A = 0.

Significance. The results are a substantial quantitative sharpening of the Romanov–Woracek comparison theory. If the proof points flagged below are addressed, the paper establishes explicit, falsifiable two-sided bounds on the bottom of the essential spectrum, recovers known criteria as special cases, and provides a sharpness example for the comparison constant. The proofs are mostly elementary and transparent, and the comparison theorem with explicit constants is new. I especially credit the example in Section 5 showing optimality of the lower constant, and the clean reduction of Theorem 1.4 to a Riccati equation.

major comments (2)

[Section 4 (proof of Theorem 1.3, lower bound)] The lower bound M(H_d) ≥ 1/(2√A) rests on the assertion, imported from [5, eqn. (4.3)], that for t²B < 1/4 the comparison Riccati equation (4.2), namely α₁' = (α₁ − t²B/x)², has a global solution α₁ ≤ 0 on a tail. At submission [5] is listed as 'to appear', so the present manuscript does not demonstrate this load-bearing step. The lemma is true and short: with c = t²B, choose u(x) = x^r, r = (1−√(1−4c))/2, so that u'' = −(c/x²)u, and set α₁ = −u'/u + c/x = (c−r)/x; then α₁ ≤ 0 and one checks directly that α₁ solves (4.2). Please include this argument, or cite the published version of [5] with the precise lemma, so that Theorem 1.3 is self-contained.
[Section 3 (proof of Theorem 1.1, second inequality)] The step 'We assumed that sin φ ∈ L², so these limits must be ≡ 0 mod π' is load-bearing and is currently justified in one sentence. The assertion is true, not false: if a nonoscillatory solution had θ(x) → L with sin L ≠ 0, then the average of θ'(x) over a long interval [a,b] would tend t sin² L, because the sin²φ and cross terms in e_L^T H e_L have vanishing averages when ∫ sin²φ < ∞; this contradicts convergence of θ. However, the text does not provide this argument, and it also passes without comment from limits to the existence of θ± with strict inequalities on a tail. Please expand this into a short lemma, since (3.2) and the constant c = (3−√5)/2 depend on it.

minor comments (4)

[Section 5 (example)] The example uses the coefficient H = [[e^x,1],[1,e^{-x}]], which is not trace-normed, although the paper's general framework assumes tr H = 1. Please state explicitly that M(H) and M(H_d) are invariant under the trace-normalizing change of independent variable, or normalize H before presenting the example.
[Section 2 (oscillation theory)] The characterization (2.2) via nonoscillatory solutions presupposes 0 ∉ σ_ess. The theorem statements do not explicitly state this; the proofs use it implicitly when M(·) > 0, and the cases M(·) = 0 are trivial. Please add a sentence clarifying this point.
[Section 3 (proof of Theorem 1.2)] The global comparison of the diagonal Prüfer equation with θ' = t(θ² + sin²φ) uses the inequalities sin²φ cos²θ ≤ sin²φ and cos²φ sin²θ ≤ θ², which hold because sin²θ ≤ θ² for all real θ. Please state this explicitly, as it is not obvious from the display.
[Reference [5]] Please update the publication status of [5] and, if the lemma used in Section 4 appears there in a different form, give a precise statement or theorem number in the text rather than referring only to equation (4.3) of [5].

Circularity Check

0 steps flagged · score 2.0 of 10

No circular reduction: target quantities are outputs, constants come from comparison estimates; minor self-citation to companion paper [5] does not make the derivation circular.

full rationale

The derivation chain is open. M(H) and M(Hd) are defined as spectral minima and never appear as fit parameters; A and B are data-defined limsup/liminf integrals, and the constants 1/2, 2/(3−√5), 1/(2√A), and 1/√A arise from comparison principles, a quadratic optimization in Section 3, tent-function estimates, and the Euler-equation threshold t²C = 1/4. No 'prediction' is equivalent by construction to an input. The only self-citation is the companion-paper lemma [5] used in the proof of Theorem 1.3: 'We established in [5] that if t²B < 1/4, then (4.2) will have a global solution α₁(x) ≤ 0...' This lemma is parameter-free and its assumptions do not include the target bound on M(Hd), so under the stated rules it is real evidence rather than a circular reduction. Two non-circular issues should be noted separately: the Section 3 assertion that sin φ ∈ L² forces non-oscillatory Prüfer limits ≡ 0 mod π is not proved and is false for H = P₀, and the [5] lemma is not demonstrated in this paper. These are correctness/verification gaps, not input-output circularity, and they do not raise the circularity score beyond the low range.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters: A and B are defined from the coefficient (limsup/liminf of x∫_x^∞ sin²φ), not fitted, and the constants 1/2, 2/(3-√5), 1/√A, 1/(2√B) are outputs of explicit estimates and optimization, not inputs. The load-bearing inputs are background facts from [4], the oscillation toolkit from the authors' companion paper [5] (in press at submission), and standard Schrödinger oscillation theory. No invented entities: the auxiliary functions α, β, W are analytic devices within the proofs. The main external dependency is the [5] comparison lemma used in Theorem 1.3; the main internal unproved claim is the 0-mod-π limit statement in Theorem 1.1's proof.

assumptions (5)

domain assumption Oscillation criterion (2.2): M_+(H) = inf{t > 0 : (2.1) is oscillatory}, valid when 0 ∉ σ_ess.
Quoted as equation (2.2) from the authors' companion paper [5]. It is the bridge that converts the spectral quantity M(H) into the dynamics of the Prüfer angle, and every theorem in the paper relies on it.
domain assumption Comparison lemma quoted from [5] (eqn. (4.3) there): if t²B < 1/4, then the Riccati equation (4.2) has a global solution α₁(x) ≤ 0 on a half-line x ≥ a.
Used in the proof of Theorem 1.3 (Section 4) to get the lower bound M(H_d) ≥ 1/(2√A). It is not proved here and [5] was in press at submission, so this is a real external dependency.
domain assumption For non-oscillatory solutions at t and -t, the Prüfer angles converge to limits ≡ 0 mod π when sin φ ∈ L².
One-sentence assertion in the proof of Theorem 1.1 (Section 3). It is true (a finite angle limit forces sin²θ∞ · cos²φ integrable, and ∫cos²φ = ∞, hence sinθ∞ = 0), but the paper gives no proof.
domain assumption Canonical-system background from [4]: σ_ess(R*HR) = σ_ess(H) and diagonal canonical systems have spectra symmetric about 0.
Cited in Section 1 (Theorem 3.20) for the WLOG reduction to sin φ ∈ L², and in the proof of Theorem 1.2 (Theorem 6.13) to avoid a separate treatment of negative t.
standard math Classical Schrödinger oscillation theory: zero-free solutions imply absence of negative spectrum on a tail; Euler equation -y'' - t²C x^{-2} y = 0 is oscillatory iff t²C > 1/4; integrable non-positive potentials have finite negative spectrum.
Used in Theorems 1.2, 1.3, 1.4 (Sections 3-4); standard results, cited to [7] and classical literature.

how reviews work

0 comments

Cite this review

Pith. "Pith review of The essential spectrum of canonical systems." pith.science (2026). https://pith.science/paper/K5CRBSBE

@misc{pith2026190802266,
  author       = {Pith},
  title        = {Pith review of: The essential spectrum of canonical systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/K5CRBSBE}},
  note         = {Machine review of arXiv:1908.02266}
}

abstract

We study the minimum of the essential spectrum of canonical systems $Ju'=-zHu$. Our results can be described as a generalized and more quantitative version of the characterization of systems with purely discrete spectrum, which was recently obtained by Romanov and Woracek [6]. Our key tool is oscillation theory.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

9 extracted references · 9 canonical work pages

[5]

Remling and K

C. Remling and K. Scarbrough, Oscillation theory and semibounded canonical systems, to appear in J. Spectral Theory

work page
[1]

Hartman, Ordinary Diﬀerential Equations, Classics in Applied Ma themat- ics 38, SIAM, Philadelphia, 2002

P. Hartman, Ordinary Diﬀerential Equations, Classics in Applied Ma themat- ics 38, SIAM, Philadelphia, 2002

work page 2002
[2]

Kac, A criterion for the discreteness of a singular canonical system, Funct

I.S. Kac, A criterion for the discreteness of a singular canonical system, Funct. Anal. Appl. 29 (1995), 75–78

work page 1995
[3]

Kac and M.G

I.S. Kac and M.G. Krein, Criteria for the discreteness of the spec trum of a singular string, Izv. Vyss. Ucebn. Zaved. Mat. 3 (1958), 136–153

work page 1958
[4]

Remling, Spectral Theory of Canonical Systems, de Gruyter Studies in Mathematics 70, Berlin/Boston, 2018

C. Remling, Spectral Theory of Canonical Systems, de Gruyter Studies in Mathematics 70, Berlin/Boston, 2018

work page 2018
[6]

Canonical systems with discrete spectrum

R. Romanov and H. Woracek, Canonical systems with discrete sp ectrum, preprint, arxiv.org/abs/1904.03662

work page Pith review arXiv 1904
[7]

Weidmann, Spectral Theory of Ordinary Diﬀerential Operato rs, Springer Lecture Notes 1258, Springer, Berlin, 1987

J. Weidmann, Spectral Theory of Ordinary Diﬀerential Operato rs, Springer Lecture Notes 1258, Springer, Berlin, 1987. 14 CHRISTIAN REMLING AND KYLE SCARBROUGH

work page 1987
[8]

Winkler, Canonical systems with a semibounded spectrum, Op

H. Winkler, Canonical systems with a semibounded spectrum, Op. Theory Adv. Appl. 106 (1998), 397-417

work page 1998

Show all 9 references

[9]

Winkler and H

H. Winkler and H. Woracek, On semibounded canonical systems, Linear Al- gebra Appl. 429 (2008), 1082-1092. Department of Mathematics, University of Oklahoma, Norman , OK 73019 E-mail address : christian.remling@ou.edu URL: www.math.ou.edu/∼cremling Department of Mathematics, U...

2008

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [5]

Remling and K

C. Remling and K. Scarbrough, Oscillation theory and semibounded canonical systems, to appear in J. Spectral Theory

work page

[2] [1]

Hartman, Ordinary Diﬀerential Equations, Classics in Applied Ma themat- ics 38, SIAM, Philadelphia, 2002

P. Hartman, Ordinary Diﬀerential Equations, Classics in Applied Ma themat- ics 38, SIAM, Philadelphia, 2002

work page 2002

[3] [2]

Kac, A criterion for the discreteness of a singular canonical system, Funct

I.S. Kac, A criterion for the discreteness of a singular canonical system, Funct. Anal. Appl. 29 (1995), 75–78

work page 1995

[4] [3]

Kac and M.G

I.S. Kac and M.G. Krein, Criteria for the discreteness of the spec trum of a singular string, Izv. Vyss. Ucebn. Zaved. Mat. 3 (1958), 136–153

work page 1958

[5] [4]

Remling, Spectral Theory of Canonical Systems, de Gruyter Studies in Mathematics 70, Berlin/Boston, 2018

C. Remling, Spectral Theory of Canonical Systems, de Gruyter Studies in Mathematics 70, Berlin/Boston, 2018

work page 2018

[6] [6]

Canonical systems with discrete spectrum

R. Romanov and H. Woracek, Canonical systems with discrete sp ectrum, preprint, arxiv.org/abs/1904.03662

work page Pith review arXiv 1904

[7] [7]

Weidmann, Spectral Theory of Ordinary Diﬀerential Operato rs, Springer Lecture Notes 1258, Springer, Berlin, 1987

J. Weidmann, Spectral Theory of Ordinary Diﬀerential Operato rs, Springer Lecture Notes 1258, Springer, Berlin, 1987. 14 CHRISTIAN REMLING AND KYLE SCARBROUGH

work page 1987

[8] [8]

Winkler, Canonical systems with a semibounded spectrum, Op

H. Winkler, Canonical systems with a semibounded spectrum, Op. Theory Adv. Appl. 106 (1998), 397-417

work page 1998