{"id":"dc681112-3643-4eed-862b-26ca28899ee8","arxiv_id":"1908.02266","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For canonical systems, the start of the essential spectrum is at least half and at most about 2.62 times that of the diagonal system, with the 1/2 factor optimal; the diagonal value itself is pinned by limsup/liminf integrals of sin²φ.","lead":"For half-line canonical systems, this paper proves explicit two-sided bounds on the lowest point of the essential spectrum: at least half, and at most about 2.62 times, the corresponding value for the diagonal system, with the lower constant shown optimal. It also pins the diagonal value between 1/(2√A) and 1/√A, where A is a single integral of the coefficient matrix.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section 3's claim that nonoscillatory Prüfer limits are 0 mod π is false; the proof of Theorem 1.1's upper bound needs a new argument.","rationale":"The reader's weakest assumption flagged the terse Prüfer-limit premise but treated it as verified; my stress test shows it is in fact false. This is an internal and demonstrable error in the proof of the paper's headline comparison theorem, so it is more immediate than the external [5]-lemma dependency. The reader's conditional verdict remains appropriate, but the conditions should explicitly require repairing the Prüfer-limit argument, not merely fixing the displayed identity in Theorem 1.2 and the Section 5 example. Because the theorem is likely salvageable, rejection is not warranted.","tokens_in":9737,"tokens_out":44757,"duration_ms":488509,"concrete_test":"Verify the disputed claim by solving (2.1) for H=P_0 at t=1: dθ/dx=cos²θ, θ=arctan(x+C) → π/2 as x→∞. This contradicts the assertion that sinφ∈L² forces Prüfer limits ≡0 mod π. Then check whether the second inequality of Theorem 1.1 can be proved for the family H=P_φ (e.g. φ=e^{-x}) by the Section 3 argument; if θ± cannot be chosen small there, the missing case is confirmed.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The most load-bearing gap is in the proof of Theorem 1.1, second inequality. After choosing t∈(0,M(H)), the authors state that because sinφ∈L², the nonoscillatory solutions at ±t have Prüfer limits ≡0 mod π, so θ± can be made small. This is false. For H=P_0 (φ≡0, sinφ=0∈L²), eq. (2.1) at t is θ'=t cos²θ, whose solutions θ=arctan(tx+C) all tend π/2; no solution has limit 0 mod π. The same limiting behaviour persists for perturbations such as H=P_φ with φ=e^{-x}∈L², φ→0, where the asymptotic coefficient is a projection whose kernel direction is not e1. Consequently the smallness of θ± used to prove (3.2) is not justified. Since (3.2) is the only mechanism producing the constant c=(3−√5)/2, the proof of the upper bound in Theorem 1.1 is incomplete as written. The theorem may still be true, but this step must be replaced, e.g. by a two-case analysis of limits 0 and π/2 mod π or by a local argument around the actual limit.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":9803,"tokens_out":25495,"duration_ms":258876,"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":[{"comment":"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":"Section 4 (proof of Theorem 1.3, lower bound)"},{"comment":"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.","section":"Section 3 (proof of Theorem 1.1, second inequality)"}],"minor_comments":[{"comment":"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":"Section 5 (example)"},{"comment":"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":"Section 2 (oscillation theory)"},{"comment":"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.","section":"Section 3 (proof of Theorem 1.2)"},{"comment":"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].","section":"Reference [5]"}],"recommendation":"major_revision","confidential_remarks":"The central mathematical claims appear sound; my recommendation is driven by self-containedness rather than by any error I could locate. In particular, the worry that nonoscillatory Prüfer limits need not be 0 mod π under sin φ ∈ L² does not survive contact with the actual equations: for φ ≡ 0, for instance, (2.1) reads θ' = t sin²θ, whose limits are 0 mod π. The requested addition of the short lemma from [5] is easy and will make the paper self-contained; if the companion paper has appeared and the reference is updated, the revision could be minor."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe 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.\n\nThe 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.\n\nNow 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.\n\nTwo 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.\n\nI 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.\n\nNet: 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.","headline":"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.","tokens_in":10658,"tokens_out":8087,"would_cite":true,"duration_ms":80356,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["34C10","34L40","47A06"],"pacs":[],"model":"deepseek-v4-flash","headline":"Essential spectrum minimum pinned within factor 2.62 of diagonal case","keywords":["canonical system","essential spectrum","oscillation theory","Prüfer angle","diagonal system","Schrödinger operator","discrete spectrum","Riccati equation"],"falsifier":"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.","tokens_in":9294,"feed_emoji":"📐","tokens_out":9858,"duration_ms":87907,"temperature":0.7,"pith_summary":"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.","feed_headline":"Essential spectrum minimum pinned within factor 2.62 of diagonal case","feed_subtitle":"Two-sided comparison plus explicit bounds in terms of the tail of sin² φ, with a sharpness example.","key_machinery":"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 $M(H_d)$.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Supplies the characterization of purely discrete spectrum and the independence/comparison result that Theorem 1.1 quantifies.","marker":"[6]"},{"why":"Supplies the oscillation-theoretic framework and the Riccati lemma (eqn. (4.3)) used in the proof of Theorem 1.3.","marker":"[5]"},{"why":"Supplies the basic theory of canonical systems, including spectral symmetry for diagonal systems and transformations to Dirac and Schrödinger operators used in the example.","marker":"[4]"},{"why":"Supplies the comparison principle for first-order ODEs used to transfer non-oscillation between the full and diagonal systems.","marker":"[1]"},{"why":"Supplies the classical oscillation theory for Schrödinger operators used in the proof of Theorem 1.2.","marker":"[7]"}],"fun_headline_variants":["Sharp factor 2.62 bound for essential spectrum","Essential spectrum minimum: optimal lower, 2.62 upper","Oscillation theory pins essential spectrum bounds","Canonical systems: quantitative spectral gap control","Discrete spectrum criterion extended with explicit ratios"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Sharp factor 2.62 bound for essential spectrum","Essential spectrum minimum: optimal lower, 2.62 upper","Oscillation theory pins essential spectrum bounds","Canonical systems: quantitative spectral gap control","Discrete spectrum criterion extended with explicit ratios"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000371,"raw_usage":{"total_tokens":1896,"prompt_tokens":766,"completion_tokens":1130,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":382,"completion_tokens_details":{"reasoning_tokens":1057}},"tokens_in":382,"tokens_out":1130,"duration_ms":11446,"temperature":1.0,"reasoning_tokens":1057,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:55:31.268028+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[{"cited_title":"Canonical systems with discrete spectrum","cited_arxiv_id":"1904.03662","evidence_quote":"Supplies the characterization of purely discrete spectrum and the independence/comparison result that Theorem 1.1 quantifies."},{"cited_title":"Remling and K","cited_arxiv_id":null,"evidence_quote":"Supplies the oscillation-theoretic framework and the Riccati lemma (eqn. (4.3)) used in the proof of Theorem 1.3."},{"cited_title":"Remling, Spectral Theory of Canonical Systems, de Gruyter Studies in Mathematics 70, Berlin/Boston, 2018","cited_arxiv_id":null,"evidence_quote":"Supplies the basic theory of canonical systems, including spectral symmetry for diagonal systems and transformations to Dirac and Schrödinger operators used in the example."},{"cited_title":"Hartman, Ordinary Differential Equations, Classics in Applied Ma themat- ics 38, SIAM, Philadelphia, 2002","cited_arxiv_id":null,"evidence_quote":"Supplies the comparison principle for first-order ODEs used to transfer non-oscillation between the full and diagonal systems."},{"cited_title":"Weidmann, Spectral Theory of Ordinary Differential Operato rs, Springer Lecture Notes 1258, Springer, Berlin, 1987","cited_arxiv_id":null,"evidence_quote":"Supplies the classical oscillation theory for Schrödinger operators used in the proof of Theorem 1.2."}],"review_version":1}