REVIEW 4 major objections 5 minor 6 references
Growth estimates for Nevanlinna matrices of order larger than one half
T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash
Pith's one-line read For simply critical Jacobi matrices, the Nevanlinna order is exactly 1/(2(β−1)) when 3/2 < β < 2.
desk verdict A genuinely new lower bound for monotone-angle Hamiltonians and a plausible order computation for critical Jacobi matrices, but the Jacobi theorem rests on a lemma whose proof is omitted. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The mechanism is geometric: the monodromy of the canonical system is controlled by how much the Hamiltonian directions rotate. The proof of the lower bound uses the identity $\det\Omega(x_m,x_n)=\frac12\sum_{j,k=m+1}^{n} l_j l_k \sin^2(\varphi_j-\varphi_k)$ together with a lower-bound lemma that converts many disjoint time intervals with $\det\Omega(s_{j-1},s_j)\geq c/r^2$ into the estimate $\log|w_{H,22}(ir)|\gtrsim k(r)$. When the angles are monotone, one can choose a permutation $\sigma$ pairing the index block $I_j^+$ into $I_j^-$ so that $|\sin(\varphi_k-\varphi_{\sigma(k)})|\asymp 1$, making the determinant large on roughly $k(r)$ intervals. On the Jacobi side, the load is carried by a recessive solution $(f_n)$ of the three-term recurrence with $|f_n|^2\sim c\sqrt{\gamma_n}/b_n$ and $\arg f_n-\arg f_{n-1}\to0$; via the formulas (3.1)--(3.4) this yields lengths $l_{n+1}\asymp\sqrt{\gamma_n}/b_n$ and angle increments $\asymp1/\sqrt{\gamma_n}$, and a sign analysis of $\cos(\varphi_{n+1}-\varphi_n)$ produces the monotone angle representation required by Theorem 1.2(iii).
What would settle it
Take a concrete Jacobi matrix in the theorem's range, say $\beta=7/4$, $x_0=1$, $y_0=2x_0$, $\sigma=3$, with $\chi_n=\mu_n=0$, and compute the associated canonical system: if the normalized angle increments $\sqrt{n}(\varphi_{n+1}-\varphi_n)$ do not converge to a nonzero constant of one sign, or if $\log|w_{H,22}(ir)|/r^{1/(2(\beta-1))}$ does not stay bounded between two positive constants as $r\to\infty$, then Theorem 1.4 fails.
Extended reading notes
Core claim
The paper's central claim is an exact growth formula in the regime where the order of the Nevanlinna matrix exceeds one half. Theorem 1.1 states that whenever a Hamburger Hamiltonian in limit circle case has regularly varying lengths and angle steps, $l_j \gtrsim d_l(j)$ and $|\varphi_{j+1}-\varphi_j|\asymp d_\varphi(j)$ with indices $-\delta_l$ and $-\delta_\varphi$, $\delta_\varphi\in(0,1)$, and the angles are eventually monotone, then $\log|w_{H,22}(ir)| \gtrsim r \int_{[d_\varphi/d_l]^-(r)}^{\infty} d_l(x)\,dx$. Theorem 1.2(iii) upgrades this to $\max_{|z|=r}\log\|W_H(z)\| \asymp m(r)$, with order $(1-\delta_\varphi)/(\delta_l-\delta_\varphi)$, when $1<\delta_l+\delta_\varphi<2$. In Jacobi terms, Theorem 1.4 asserts that for parameters satisfying (1.11) in the simply critical case with $3/2<\beta<\min\{2,\sigma\}$ and $\sum_{n=1}^{\infty}\sqrt{n}(|\chi_n|+|\mu_n|)<\infty$, the operator is limit circle and $\max_{|z|=r}\log\|W(z)\| \asymp r^{1/(2(\beta-1))}$, so the order is exactly $1/(2(\beta-1))$. The proof also yields the exceptional case $\delta_l=\delta_\varphi=1$ with $l_j\asymp j^{-1}(\log j)^{-\nu}$ and angle increments $\asymp j^{-1}$, where the order is $1/\nu$.
Load-bearing premise
The decisive premise is that the special small-at-infinity solution of the Jacobi recurrence has squared modulus comparable to $\sqrt{\gamma_n}/b_n$ and phase steps tending to zero, and that the power-asymptotic parameters satisfy the auxiliary regularity condition (asserted as Lemma 3.5, without proof) that makes this asymptotic available; if either part fails, the monotone-angle picture and the final order formula collapse.
Editorial extensions
If this is right
- For Hamburger Hamiltonians satisfying (1.8) with eventually monotone angles, $\delta_\varphi>0$ and $1<\delta_l+\delta_\varphi<2$, the exact growth $\max_{|z|=r}\log\|W_H(z)\|\asymp m(r)$ and order $(1-\delta_\varphi)/(\delta_l-\delta_\varphi)$ now follow.
- For Jacobi matrices with power asymptotics in the simply critical case and $3/2<\beta<\min\{2,\sigma\}$, the Nevanlinna order is exactly $1/(2(\beta-1))$, not merely bracketed by $1/\beta$ and that value.
- With the earlier analyses, Theorem 1.5 gives a complete classification: limit circle occurs exactly in the listed small-diagonal, simply critical, and doubly critical cases, and the order is $1/\beta_1$ except in the simply critical range $3/2<\beta_1<2$, where it is $1/(2(\beta_1-1))$.
- The exceptional Proposition 2.5 shows that at the parameter boundary $\delta_l=\delta_\varphi=1$, monotone angles with logarithmic corrections still force order $1/\nu>1/2$ for $\nu\in(1,2)$.
- According to Remark 1.6, these are the first limit-circle examples with a computed Nevanlinna order different from the convergence exponent of $(b_n)$.
Reading between the lines
- Editorial inference: if the sign-assignment construction sketched in Remark 1.3 works, then any regularly varying $f$ with $k\lesssim f\lesssim m$ is realized as the growth of some Hamiltonian, so the exact growth would encode the averaged rotation direction rather than just the order.
- Editorial inference: the interval-pairing lower bound is a phase-coherence argument that could be tested on other one-dimensional discrete systems, such as Schrödinger operators with monotone phase shifts; a numerical check of the determinant lower bound in those models would show whether monotonicity alone drives the same $\asymp$ growth.
- Editorial inference: the proof of Theorem 3.2 rests on the recessive-solution asymptotic; supplying the omitted $\gamma$-tempered verification (Lemma 3.5) for wider perturbation classes would automatically extend the exact growth formula to slowly varying or periodically modulated coefficients whenever the angle sequence is eventually monotone.
- Editorial inference: the boundary case $\beta=2$ is left with order $1/2$ but without an exact $\asymp$; the method suggests a logarithmic correction interpolates between the two growth regimes, and a direct calculation there would settle its form.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the growth (order) of the Nevanlinna matrix of an indeterminate Hamburger moment problem, equivalently of the monodromy matrix of the associated two-dimensional canonical system, in the regime where the order is larger than one half. The main abstract result is Theorem 1.1, a new lower bound for max|z|=r log ||W_H(z)|| under the assumptions that the Hamiltonian lengths are comparable to a regularly varying function, the angle increments are comparable to another such function, and the angles are eventually monotone. Theorem 1.2 combines this lower bound with upper bounds from PRW23 to give a classification of exact growth in the range 1 < δ_l + δ_φ < 2. The paper then specializes to Jacobi matrices with two-term power asymptotics in the simply critical case. Theorem 1.4 states that for 3/2 < β < min{2,σ}, the Nevanlinna matrix satisfies max|z|=r log ||W(z)|| ≍ r^{1/(2(β−1))}, determining the order in the regime where Pru20 only gave 1/β ≤ ρ ≤ 1/(2(β−1)). The proof passes from Jacobi parameters to a Hamburger Hamiltonian and invokes the γ-tempered framework of Świderski–Trojan, ultimately applying Theorem 1.2(iii).
Significance. If the proof can be completed, Theorem 1.4 is a substantial contribution: it gives the first computed examples of limit-circle Jacobi matrices whose Nevanlinna order differs from the convergence exponent of b_n, and it closes the gap left by Pru20 in the simply critical case. The lower-bound method for canonical systems, based on a determinant comparison lemma (Lemma 2.3), is novel and is developed in considerable detail in Section 2; that part of the paper appears sound and is a genuine methodological advance. The authors are also careful to separate the new lower bound from previously known upper bounds, and the reliance on the preprint LRW24 for Lemma 2.1 is not circular because that lemma has an independent proof in a different paper. However, the Jacobi part of the paper currently rests on at least one unproved algebraic verification, Lemma 3.5, which is the entry condition for the entire γ-tempered framework in the power-asymptotics setting. Until that calculation is supplied, the headline order computation is conditional rather than proven.
major comments (4)
- [Section 3.3, Lemma 3.5] Lemma 3.5 is load-bearing but its proof is omitted. The lemma asserts that the unperturbed parameters (3.11) are γ-tempered with γ_n = n and that τ_0 = β − σ. These are precisely the hypotheses that allow Theorem 3.2 to be applied in the proof of Theorem 3.4, and hence Theorem 1.4. The text says only that the lemma follows by "somewhat lengthy but simple calculations, which we omit." This is not acceptable for a central verification: the six bounded-variation conditions in Definition 3.1 must be checked, and the limit defining τ_0 must be computed. Please include the full calculation, at least in an appendix.
- [Section 3.1, equations (3.1)–(3.3)] A footnote at this point reads "Some technical details are left out." The formulas (3.1)–(3.3) are the bridge between Jacobi parameters and the Hamiltonian data, and they are used directly in Step 1 of the proof of Theorem 3.2 to obtain (3.20)–(3.21). Since the proof of the paper's main Jacobi result depends on these formulas, the omitted details should be supplied or the exact derivation from Kac99 should be reproduced, rather than left to a footnote.
- [Section 3.2, Lemma 3.3] The proof of Lemma 3.3 is a sketch rather than a complete proof. In particular, it invokes "Claim 5.2" of Świderski–Trojan to obtain |f_n|^2 ∼ c√(γ_n/b_n), and it invokes Theorem 4.4 of ŚT22 for the perturbation step, but the exact statements are not quoted and the way they apply here is not fully documented. The linear-independence argument is also compressed: the assertion that the right-hand side is divergent uses lim √γ_n/n = 0 together with (3.18), but the convergence of the relevant series is not shown. Since the asymptotic (3.15) and the angle convergence are the keys to (3.20)–(3.21), this lemma needs a more self-contained proof or, at minimum, precise references to the claims used.
- [Section 2.2, Step 3] The construction of the permutation σ with σ(I_j^+) ⊆ I_j^- is asserted immediately after the cardinality comparison |I_j^+| < |I_j^-|. This does not by itself justify the existence of a global permutation on the union of the intervals: one must also check that the target intervals I_j^- are pairwise disjoint (and similarly the source intervals I_j^+), so that the piecewise injections combine into a permutation. This disjointness does follow from regular variation, but it is not stated or proved. Please add the short verification.
minor comments (5)
- [Throughout] The notation "/greaterorsimilar" for ≳ is nonstandard and appears in several displayed formulas; please use standard symbols such as ≳ and ≲ consistently.
- [Section 1.1, equation (1.7)] The lower bound in Theorem 1.1 is stated as log |w_H,22(ir)| ≳ r ∫ ...; the dimensions of the right-hand side are not transparent. A brief explanation of why the expression has the expected order would help the reader.
- [Section 3.2, Lemma 3.3] The conclusion "lim_{n→∞}(arg f_n − arg f_{n−1}) exists modulo π and is equal to 0" is phrased awkwardly. Since the limit is asserted to equal 0 modulo π, it would be clearer to say that arg f_n − arg f_{n−1} → 0 after choosing representatives, or to define precisely what "modulo π" means here.
- [Remark 1.6] The claim that these are "to the best of our knowledge" the first examples where the order differs from the convergence exponent of b_n is appropriately hedged, but it would be useful to state explicitly what previous examples are known to be excluded by the cited literature.
- [References] The paper relies heavily on the unpublished preprint LRW24 and on the preprint PRW23. If the journal policy allows citing preprints, please ensure the versions are clearly identified; it would also be helpful to state which parts of LRW24 are used beyond Lemma 2.1.
Circularity Check
No significant circularity: the central growth estimate is supported by independent determinant and upper-bound lemmas; the omitted γ-temperedness verification in Lemma 3.5 is a proof gap, not a self-referential reduction.
full rationale
The derivation chain for Theorem 1.4 runs Theorem 1.1 -> Theorem 1.2(iii) -> Theorem 3.2 -> Theorem 3.4 -> Theorem 1.4. The lower bound is proved from Lemma 2.1 (a determinant criterion taken from LRW24) plus Lemma 2.3, which is proved in the paper; the upper bound comes from PRW23. Both LRW24 and PRW23 are parameter-free external results whose stated assumptions do not include the target growth formula, so under the review rules they count as independent support rather than circular self-citation. Lemma 3.3, which supplies the recessive-solution asymptotics |f_n|^2 ~ c sqrt(gamma_n)/b_n and the angle increment control, is proved in the paper using the external framework of [ST23]. No fitted constants or data are involved, and no equation is shown to be equivalent to its own input by construction. The one load-bearing assertion that is explicitly left unproved is Section 3.3, Lemma 3.5: 'The following lemma is proved by somewhat lengthy but simple calculations, which we omit,' stating that the unperturbed power-asymptotic parameters are gamma-tempered with gamma_n = n and tau_0 = beta - sigma. This is an omitted verification of an entry condition to Theorem 3.2, so it is a completeness/correctness risk, but it is not a circular step: nothing in the paper defines gamma-temperedness or tau_0 in terms of the claimed growth estimate. The final formula r^{1/(2(beta-1))} is obtained by evaluating the general m(r) expression with dl(j) = sqrt(gamma_j)/b_j and dphi(j) = 1/sqrt(gamma_j); that evaluation is a direct computation, not a reintroduction of the conclusion. Hence the paper does not exhibit a circular derivation, and the only qualification is the unproved but plausibly routine Lemma 3.5.
Assumptions & free parameters
assumptions (6)
- standard math Regular variation theory, including Karamata's theorem and asymptotic inverse relations ([BGT89]).
- domain assumption Lemma 2.1 from [LRW24] bounding log|wH,22(ir)| below in terms of determinants det Ω(s_{j-1},s_j) ≥ c/r².
- domain assumption Upper bounds from [PRW23, Corollary 4.7, Corollary 5.2, Theorem 5.3] for max log||WH|| in terms of k(r) and m(r).
- domain assumption Jacobi-to-Hamiltonian conversion formulas (3.1)-(3.4), attributed to [Kac99].
- domain assumption γ-tempered perturbation theory from [ST23, Theorems 9.1, 9.2, Theorem A] and [ST22, Theorem 4.4].
- domain assumption The asymptotic form of f_n in Lemma 3.3, namely |f_n|² ∼ c√γ_n/b_n with argument increments tending to 0.
Cite this review
Pith. "Pith review of Growth estimates for Nevanlinna matrices of order larger than one half." pith.science (2026). https://pith.science/paper/37DTAOWL
@misc{pith2026250111400,
author = {Pith},
title = {Pith review of: Growth estimates for Nevanlinna matrices of order larger than one half},
year = {2026},
howpublished = {\url{https://pith.science/paper/37DTAOWL}},
note = {Machine review of arXiv:2501.11400}
}
read the original abstract
Our objects of study are two-dimensional canonical systems that arise from indeterminate Hamburger moment problems and associated half-line Jacobi operators in limit circle case. The monodromy matrix of such a system coincides, up to a permutation of its entries, with the Nevanlinna matrix of the associated moment problem. Its growth relates to the density of eigenvalues of self-adjoint realisations of the system and the Jacobi operator, respectively. The order of the Nevanlinna matrix is known to be at most 1. In the case of "large" order, meaning order greater than one half, determining the growth of the monodromy matrix is known to be harder than for "small" order, i.e., order less than one half. As our main result, we establish an explicit new lower bound for data featuring a certain kind of monotonicity, which correctly describes the growth in the case of large order. Moreover, we compute the order of the Nevanlinna matrix of a limit circle Jacobi matrix with two-term power asymptotics in a critical case.
Reference graph
Works this paper leans on
-
[1]
[Akh61] N.I. Akhiezer, Klassiqeskaꦪ problema momentov i nekotorye .ۈRussian, English translation: The classical moment problem and some related questions in a na- lysis, Oliver & Boyd, Edinburgh, 1965, Gosudarstv. Izdat. Fiz.-Mat. Lit., Moscow,
work page 1965
-
[52]
Romanov, Order problem for canonical systems and a co njecture of Valent, Trans
[Rom17] R. Romanov, Order problem for canonical systems and a co njecture of Valent, Trans. Amer. Math. Soc. 369 (2017), 1061–1078. [´ST22] Grzegorz ´Swiderski and Bartosz Trojan, About essential spectra of unbounded Jacobi matrices, Journal of Approximation Theory 278 (2022), 105746. [´ST23] Grzegorz ´Swiderski and Bartosz Trojan, Orthogonal polynomials ...
work page 2017
-
[234]
Eigenvalue distribution of canonical systems: trace class and sparse spectrum
[Kac02] I.S. Kac, Linear relations generated by a canonical differen tial equa- tion of dimension 2, and eigenfunction expansions, Algebra i Analiz 14 (2002), 86–120. [Kre51] M.G. Krein, On the theory of entire matrix functions of expo nential type, Ukrain. Mat. ˇZurnal 3 (1951), English translation in Oper. Theory Adv. Appl. 95 (1997), 361–371, 164–173. [...
work page Pith review arXiv 2002
-
[1961]
[Ber56] Yu.M. Berezanski ˘ ı, Expansion according to eigenfunction of a partial difference equation of order two, Trudy Moskov. Mat. Obˇ sˇ c.5 (1956), 203–268. [BS14] C. Berg and R. Szwarc, On the order of indeterminate momen t prob- lems, Adv. Math. 250 (2014), 105–143. [BGT89] N.H. Bingham, C.M. Goldie and J.L. Teugels, Regular variation , vol. 27, Ency...
work page 1956
-
[1989]
de Branges, Some Hilbert spaces of entire functions
[Bra61] L. de Branges, Some Hilbert spaces of entire functions. II , Trans. Amer. Math. Soc. 99 (1961), 118–152. [Kac99] I.S. Kac, Inclusion of the Hamburger power moment problem in the spectral theory of canonical systems, Russian, Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. (POMI) 262 (1999), English translation: J. Math. Sci. (New York) ...
work page 1961
-
[2023]
Riesz, Sur le probl` eme des moments
[Rie23] M. Riesz, Sur le probl` eme des moments. III. French, Ark. Mat. As- tron. Fys. 17 (1923),
work page 1923
Reviewed August 10, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.