{"id":"19f350fd-a984-4c46-8faf-67adb6afead8","arxiv_id":"2607.19797","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"New coupled difference equations for the three-term recurrence coefficients of monic orthogonal polynomials for oscillatory Gegenbauer and Jacobi-type weights are derived, with conjectured symbolic forms.","lead":"This paper derives simpler recurrence equations for the coefficients of orthogonal polynomials tied to two oscillatory Jacobi-type weights, useful for numerical quadrature of oscillatory integrals. It also conjectures closed-form symbolic patterns for the coefficients based on computed tables.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Ladder-operator extension to complex, sign-changing weights (especially w2) is asserted without proof; the central difference equations inherit this unverified foundation.","rationale":"The strongest claim is that the coupled difference equations (2.3)-(2.4) and (3.19)-(3.20) allow iterative computation of all recurrence coefficients from initial values. The single most load-bearing condition for this claim is the validity of Proposition 1.1 for the oscillatory Jacobi weights. All subsequent lemmas and theorems are algebraic consequences of the ladder operators and compatibility conditions; if Proposition 1.1 does not apply, the derived equations are unsupported. The reader identified exactly this as the weakest assumption, and I agree. The paper's justification is one sentence asserting that the positivity condition was not used, without proof. This is particularly concerning for w2, where the corresponding w0 is not positive and has an interior zero, creating a pole in v'. The proposed numerical test would provide direct evidence on whether the ladder identities actually hold for these weights. The reader's CONDITIONAL verdict remains appropriate: the derivation is credible and backed by numerical agreement in Remark 3.12, but it is not fully rigorous until the ladder-operator extension is either proved or independently verified. The missing Section 4 is noted but is not what moves the verdict.","tokens_in":26944,"tokens_out":11516,"duration_ms":106806,"concrete_test":"Take w2(x)=x(1-x^2)^{-1/2} e^{iηx} with η≈2.4048255577 (first positive zero of J0). Using adaptive complex quadrature on [-1,1], compute moments μ_k=∫ x^k w2(x) dx for k=0..5 and perform Gram-Schmidt to obtain P_0,P_1,P_2 and norms h_0,h_1,h_2 (so β_1=h_1/h_0, β_2=h_2/h_1). Then evaluate A_n(z),B_n(z) from their defining integrals (1.13)-(1.14) for n=1,2 at z=0.3,0.7,2.0, and check numerically whether P_n'(z)=-B_n(z)P_n(z)+β_n A_n(z)P_{n-1}(z) holds to relative error <1e-8. A failure would invalidate the extension of Proposition 1.1; a pass would support the paper's assertion.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's main results (Lemmas 2.1 and 3.1; Theorems 2.2 and 3.4) are built entirely on Proposition 1.1, quoted from [12] under the hypothesis w0(x)>0. The authors assert in Section 1 that 'w0(x)>0 ... was not used in the derivation' and therefore Proposition 1.1 'also applies' to the complex weights (1.1)-(1.2). This is not demonstrated. The extension is especially delicate for w2(x)=x(1-x^2)^{γ-1/2} e^{iηx}: here w0(x)=x e^{iηx} is non-positive and vanishes at x=0, and (3.2) shows v'(x) has a simple pole at 0. While the A_n,B_n expressions in Lemma 3.1 formally cancel this pole in the difference quotient, the integration-by-parts and boundary-term structure in [12] was developed for smooth positive w0; whether positivity of h_n or continuity of v' is needed is not addressed. Since (2.3)-(2.4) and (3.19)-(3.20) are algebraic consequences of these ladder operators, an unproven extension leaves the central claim without a foundation. Additionally, the introduction promises conclusions in Section 4, but no Section 4 appears; this is a separate editorial flaw, not the main scientific risk.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies two classes of oscillatory Jacobi-type weights: w1(x)=(1-x^2)^{λ-1/2}e^{iζx} and w2(x)=x(1-x^2)^{γ-1/2}e^{iηx}. For each, the authors compute the ladder-operator quantities A_n(z), B_n(z) explicitly in terms of the recurrence coefficients {α_n, β_n} and auxiliary quantities, then apply the compatibility conditions (S1), (S2), (S'_2) to derive coupled difference equations for α_n and β_n. The main results are Theorem 2.2, giving (2.3)-(2.4) for w1 with explicit initial values (2.5), and Theorem 3.4, giving (3.19)-(3.20) for w2 after eliminating p(n) from the system (3.16)-(3.18). Tables 1-4 list computed coefficients for special cases, and several conjectured symbolic forms are proposed in Conjectures 2.5, 2.7, 3.7, and 3.9. The paper claims that these difference equations are structurally simpler and of lower order than existing ones and that, once initial values are known, the recurrence coefficients can be iteratively computed.","tokens_in":27343,"tokens_out":9149,"duration_ms":84241,"significance":"If the central derivations are valid, the paper delivers explicit, low-order coupled difference equations for the recurrence coefficients of two nontrivial non-Hermitian orthogonal-polynomial systems. This would be a useful improvement over the existing literature: for w1 the equations are indeed simpler than those in [15], and for w2 the paper provides explicit iterable equations not previously written down. The extensive tables, appendices with hand derivations of initial values, and Mathematica code are valuable and make the algebraic steps reproducible. The conjectures are clearly labeled as conjectures and do not affect the correctness of the difference equations. However, the entire derivation rests on the unproved extension of Proposition 1.1 from positive w0 to complex-valued, sign-changing w0, and the final difference equations contain divisions by quantities whose nonvanishing is not established. These are load-bearing issues, so the central claim is only conditionally established.","major_comments":[{"comment":"The manuscript's central tool is Proposition 1.1, restated from [12] under the hypothesis w0(x)>0. The sentence 'the condition w0(x)>0 ... was not used in the derivation' is an assertion, not a proof. This matters especially for w2(x)=x(1-x^2)^{γ-1/2}e^{iηx}: here w0(x)=x e^{iηx} is not positive and, as shown in (3.2), v'(x) has a simple pole at x=0. The formal cancellation of that pole in the difference quotient leading to Lemma 3.1 does not by itself justify the integration-by-parts identities and boundary-term structure used in [12]. Since Lemmas 2.1 and 3.1 and hence Theorems 2.2 and 3.4 are algebraic consequences of Proposition 1.1, the unproved extension is load-bearing. The authors should either prove that the ladder-operator theorem holds for the complex, sign-changing weights (1.1)-(1.2), or give a precise reference where it is proved under hypotheses covering these cases.","section":"Section 1, Proposition 1.1"},{"comment":"The proof of Theorem 3.4 solves (3.18) and (3.17) for p(n) and p(n-1), respectively. This introduces explicit divisions by α_{n-1}-α_n and β_n-β_{n-1}; moreover both (3.19) and (3.20) contain β_{n+1} in the denominator, and (3.20) also contains β_{n-1}-β_n in the denominator. The paper nowhere proves that these factors are nonzero in the parameter regimes considered. Consequently the iterative claim in the paragraph after (3.19)-(3.20) is only formal unless a genericity assumption is added. The same remark applies to Theorem 2.2, where (2.3) contains β_{n+1} in a denominator, and to the initial-value formulas (2.5), which require ζ^2≠2λ. If nonvanishing follows from known existence theorems or from the explicit forms in Tables 1-4, that argument should be supplied; otherwise the statements should be qualified.","section":"Theorem 3.4, Eqs. (3.19)-(3.20)"}],"minor_comments":[{"comment":"The text says 'Our conclusions are presented in Section 4,' but the manuscript ends with Section 3 and appendices. Either add a conclusions section or correct the cross-reference.","section":"Introduction, final paragraph"},{"comment":"The sentence 'When λ=0 and ζ is a positive zero of the Bessel function J0, or λ=1/2 and ζ=mπ' uses the wrong symbols; it should read 'γ=0 and η...' and 'γ=1/2 and η=mπ'.","section":"Section 3, paragraph after (3.18)"},{"comment":"The auxiliary quantities R_n(ζ) and r_n(ζ) are defined in (3.5)-(3.6) with exp(iηx), so the argument should be η, not ζ. The notation is inconsistent.","section":"Lemma 3.1, Eqs. (3.3)-(3.4)"},{"comment":"Typographical errors: 'Corollay' in Corollary 2.4, 3.5, and 3.8; 'Frourier-type' in reference [9]; 'G V. Milovanović' in reference [14] should be 'G. V. Milovanović'.","section":"Throughout"},{"comment":"The remark reports last-digit discrepancies with [14, Table 2] without comment. A brief explanation of rounding precision would be helpful, since a reader may otherwise wonder whether the difference equations or the code are at fault.","section":"Remark 3.12"}],"recommendation":"major_revision","confidential_remarks":"The paper is a serious technical contribution with reproducible computations, but the two major comments above block acceptance. The first is a genuine foundational gap: the extension of the ladder-operator framework to the complex weight w2 is nontrivial because v' has a pole and w0 is not positive. The second is a technical gap in the iteration argument. Both are fixable within the manuscript's scope, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does something genuinely new. For the oscillatory Gegenbauer weight w1 it reduces the known second-order difference equation to a first-order one (2.3), and for w2 = x(1-x^2)^{γ-1/2} e^{iηx} it gives the first derived difference equations for the recurrence coefficients. The symbolic conjectures are new and clearly labelled as conjectures. The derivations are mostly transparent, the initial values are computed in detail (including hand derivations in appendices), and the tables agree with earlier results where they overlap. The Mathematica code and the numerical check against [14] give real evidence that the equations are correct.\n\nThe soft spot is the foundation. Proposition 1.1 is quoted from [12] under w0(x)>0. The paper asserts in one sentence that positivity was not used in the derivation and therefore the proposition applies to the complex weights (1.1)-(1.2). No proof is offered. This matters most for w2, where w0 = x e^{iηx} is sign-changing and v'(x) has a simple pole at 0. The A_n, B_n expressions in Lemma 3.1 formally cancel the pole in the difference quotient, and the numerical checks pass, so the gap is probably fixable rather than fatal. But the central theorems inherit an unproven hypothesis, and a referee should ask for either a direct proof of the ladder operators and compatibility conditions for these weights or a clear reference where the extension is established. Theorem 3.4 also divides by β_n−β_{n−1} and α_{n−2}−α_{n−1} without stating that they are nonzero; the special cases in the tables have nonzero denominators, but the theorem as stated needs a nonvanishing assumption or a separate argument. Those are the two substantive issues.\n\nMinor: the introduction promises conclusions in Section 4, and no Section 4 exists. That is an editorial slip, easily fixed.\n\nBottom line: the algebra and the numerical evidence are strong enough that I would not desk-reject this. It deserves a serious referee, with the ladder-operator extension and denominator issues flagged. For someone working on oscillatory weights or Gaussian quadrature, the paper is useful and citeable. I would probably bring it to a reading group, mostly to see how the ladder-operator gap gets resolved.","headline":"Solid, useful difference-equation work with one load-bearing yet fixable gap: the ladder-operator framework is quoted under positivity and extended to complex weights without proof; plus a missing conclusions section.","tokens_in":27744,"tokens_out":2597,"would_cite":true,"duration_ms":26704,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["33C45","42C05","33C10"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper derives two coupled difference equations that determine all three-term recurrence coefficients for monic orthogonal polynomials with oscillatory Jacobi-type weights w1(x)=(1−x^2)^{λ−1/2}e^{iζx} and w2(x)=x(1−x^2)^{γ−1/2}e^{iηx} o","keywords":["oscillatory Gegenbauer weight","Jacobi-type weight","three-term recurrence coefficients","ladder operators","compatibility conditions","difference equations","orthogonal polynomials","Bessel functions"],"falsifier":"Take λ=0 and ζ the first positive zero of J_{−1}; compute α_4 and β_4 directly from the inner products h_n using high-precision numerical quadrature of (1−x^2)^{−1/2}e^{iζx}, and compare with the values from iterating (2.3)–(2.4) from (2.5). A difference beyond integration round-off would refute Theorem 2.2. Similarly, take γ=0 and η a positive zero of J_0; direct quadrature of x(1−x^2)^{−1/2}e^{iηx} against iteration of (3.19)–(3.20) would decide Theorem 3.4.","tokens_in":26862,"feed_emoji":"🧮","tokens_out":6638,"duration_ms":57531,"temperature":0.7,"pith_summary":"The paper sets out to show that the three-term recurrence coefficients of orthogonal polynomials for two oscillatory Jacobi-type weights are governed by simple coupled difference equations. For the oscillatory Gegenbauer weight (1−x^2)^{λ−1/2}e^{iζx}, it proves equations (2.3)–(2.4) that need only four initial values; for the weight x(1−x^2)^{γ−1/2}e^{iηx}, it proves equations (3.19)–(3.20). If these hold, all coefficients can be generated iteratively, and the equations are structurally simpler and lower-order than previously known ones. The paper also uses the generated tables to conjecture exact rational symbolic forms for α_n and β_n, offering an explicit road from the weight to the recurrence data needed for Gaussian quadrature of oscillatory integrals.","feed_headline":"Two coupled difference equations pin down every recurrence coefficient","feed_subtitle":"For two oscillatory Jacobi weights, a few initial values generate the whole sequence—and suggest exact rational formulas.","key_machinery":"The central object is the pair of three-term recurrence coefficients α_n, β_n, together with the ladder operators (1.11)–(1.12) and the three compatibility conditions (S1), (S2), and (S′2) from Proposition 1.1. These conditions relate the raising-and-lowering functions A_n(z) and B_n(z) through identities that must hold for every n; substituting explicit rational expressions for A_n and B_n computed from the oscillatory weights and eliminating auxiliary quantities (p(n), R_n, r_n) yields the coupled difference equations. A second load-bearing ingredient is the Bessel-function condition on ζ and η — zeros of J_{λ−1} or J_0, or η=mπ — which guarantees existence of the orthogonal polynomials an","core_discovery":"The central claim is Theorem 2.2 for w1 and Theorem 3.4 for w2: the monic orthogonal polynomials' recurrence coefficients satisfy coupled difference equations, so once a small set of initial coefficients is fixed, all later coefficients are determined. The derivations substitute explicit formulas for the ladder quantities A_n(z) and B_n(z) into the three compatibility conditions and eliminate auxiliary integrals; the initial values themselves come from known Bessel-function integral identities. The paper further observes that the iterated coefficients are real for w1 and rational in the special w2 cases, and formulates Conjectures 2.5–3.9 giving closed rational expressions in ζ^2 or η^2 with","pith_inferences":["The same ladder-operator route may apply to other oscillatory Jacobi-type weights whose log-derivative is a rational function of x, such as (1−x)^α(1+x)^β e^{iζx} with α≠β; a direct test would be to derive analogous difference equations in that case.","The conjectured symbolic forms, with monic polynomials indexed by k^2 and k(k±1), suggest these polynomials are Hankel determinants or related combinatorial objects; one testable route is to verify they satisfy a finite Toda-type recurrence in k.","If the conjectures hold, the explicit rational expressions could be expanded to give direct large-n asymptotics for α_n and β_n without solving the difference equations numerically."],"forward_implications":["For w1, the four numbers α0, α1, β1, and β2 determine every α_n and β_n through (2.3)–(2.4), independently of β0.","For w2, starting from α0,...,α3 and β1,...,β4, equations (3.19)–(3.20) generate all later coefficients.","The equations are structurally simpler than existing ones: for w1, equation (2.3) is first-order in β, replacing a second-order equation in prior literature.","Iterating reproduces known numerical tables, including the γ=1/2, η=100π case previously computed by Gaussian quadrature methods, confirming internal consistency.","The pattern of computed coefficients motivates explicit rational symbolic conjectures for all α_n and β_n in terms of monic polynomials in ζ^2 or η^2."],"fun_headline_variants":["Coupled difference equations fix all recurrence coefficients","From few initials to all coefficients: oscillatory Jacobi solved","Recurrence coefficients via two coupled difference equations","Simpler coupled equations pin down Jacobi recurrence coefficients"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The whole derivation rests on the assertion that the ladder operators and compatibility conditions, originally proved for a positive base weight w0(x)>0, also hold for the complex-valued oscillatory factors e^{iζx} and x e^{iηx}; the paper states this follows because the positivity condition was not used in the original proof, but does not give a proof of that extension.","fun_headline_variants_meta":{"raw":{"variants":["Coupled difference equations fix all recurrence coefficients","From few initials to all coefficients: oscillatory Jacobi solved","Recurrence coefficients via two coupled difference equations","Simpler coupled equations pin down Jacobi recurrence coefficients"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000657,"raw_usage":{"total_tokens":2810,"prompt_tokens":679,"completion_tokens":2131,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":423,"completion_tokens_details":{"reasoning_tokens":2069}},"tokens_in":423,"tokens_out":2131,"duration_ms":15834,"temperature":1.0,"reasoning_tokens":2069,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T11:41:24.324452+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take λ=0 and ζ the first positive zero of J_{−1}; compute α_4 and β_4 directly from the inner products h_n using high-precision numerical quadrature of (1−x^2)^{−1/2}e^{iζx}, and compare with the values from iterating (2.3)–(2.4) from (2.5). A difference beyond integration round-off would refute Theorem 2.2. Similarly, take γ=0 and η a positive zero of J_0; direct quadrature of x(1−x^2)^{−1/2}e^{iηx} against iteration of (3.19)–(3.20) would decide Theorem 3.4.","supporting_citations":[],"review_version":1}