{"id":"ae245cb0-a0be-47f3-ba15-b6d25dce5eda","arxiv_id":"2608.08714","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Adjacent centered binomial samples of even or odd entire functions weakly interlace after endpoint removal whenever the source zeros lie in a strip, with optimal width sqrt(15/28) for even sources.","lead":"This paper proves that certain polynomials built from a function by sampling at half-integers obey an interlacing order as the sampling window grows. It gives optimal bounds on the allowable width of the function's zeros, with consequences for zeros of Dedekind zeta and modular form L-functions.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the central Theorem 1.1(i) interlacing proof is self-contained. The only external dependency, companion preprint [23], affects the arithmetic applications, not the central claim's validity.","rationale":"The central claim, Theorem 1.1(i), asserts weak interlacing of reduced centered binomial quotients for even entire sources with zeros in the optimal strip of half-width sqrt(15/28). I traced the entire proof: the Jacobi spectral representation (Lemma 2.2 and surrounding identities) is elementary; the variation-diminishing framework (Proposition 3.5) rests on standard TN/checkerboard arguments and the proved Lemmas 3.3 and 3.4; the second-order factors (Lemma 4.1), reflected quartets (Lemma 4.9), and paired outer real factors (Lemma 4.10) are each justified by explicit total-nonnegativity proofs with no unstated assumptions. The unpaired outer real factor is handled directly in Lemma 5.1 with a complete discriminant and root-location analysis. The optimality of the uniform constant is reduced to the n=1 sharpness of Lemma 5.1, which is proven by exhibiting a real-pair polynomial. None of these steps invokes the companion preprint [23]. The reader's identified weakest assumption, the dependence on [23], is real but only affects the fixed-degree support lemma, the odd-endpoint support lemma, and the arithmetic applications in Section 8; it does not touch the central interlacing theorem or its sharp constant. Thus the central argument holds up under scrutiny, and the proper verdict remains CONDITIONAL because the broader paper's applications and derivative-order interlacing conclusions still await independent verification of [23]. I agree with the reader that the central theorem is self-contained while the companion preprint is the principal external risk.","tokens_in":49635,"tokens_out":40845,"duration_ms":362341,"concrete_test":"Independently verify the two theorem-level inputs from the companion preprint [23] used here: the sharp fixed-degree unit-circle support theorem (Theorems 1.1(2)) and the simplicity plus strict cyclic interlacing of consecutive derivative samples (Theorem 1.2), including the nondegeneracy condition T^d_{zeta,delta}H^{(m)} not identically zero. A re-derivation or machine-checked proof of those statements would settle the only external dependency; the central endpoint-pencil interlacing theorem does not require this check.","verdict_should_be":"UNCHANGED","load_bearing_attack":"After reviewing the main proof chain for Theorem 1.1(i) — the Jacobi spectral representation (Section 2), the interval-root monotonicity (Proposition 3.5), the zero-orbit total-nonnegativity lemmas (Lemmas 4.1, 4.9, 4.10), and the unpaired-real-factor endpoint-pencil lemma (Lemma 5.1) — I find no gap that would invalidate the central weak interlacing theorem or its sharp uniform strip constant sqrt(15/28). The proof is self-contained: the companion preprint [23] is invoked only for fixed-degree unit-circle support (Lemma 5.3, Lemma 5.8), for simplicity and strict cyclic interlacing of derivative samples, and in the arithmetic applications (Section 8). These uses do not enter the derivation of Theorem 1.1(i) or its optimality. The paper explicitly limits its dependence on [23] to those auxiliary statements. I therefore identify no load-bearing concern about the central claim. The only risk to the paper's broader conclusions is the unverified status of [23], which the reader already flagged and which warrants the CONDITIONAL verdict.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies centered binomial samples B_{2n+1}[H] of an even or odd real entire function H of order at most one. After removing the forced endpoint at z=±1 and passing to x=z+z^{-1}, the sample becomes a real quotient C_n. The central result, Theorem 1.1, gives uniform strip conditions on the zeros of H under which the adjacent quotients C_n and C_{n+1} generate a real-rooted pencil; in the even case the half-width sqrt(15/28) is optimal uniformly in n, and in the odd case width 1 is sufficient. The proof is organized as preservation of the endpoint pencil y^n(y+t) by Jacobi spectral multipliers M_{F,ν}. It combines Bernstein variation diminution, total nonnegativity of checkerboard Jacobi matrices, direct quadratic analysis of a possible unpaired outer real pair, an exact threshold for the first nontrivial pencil, and a remote-zero-orbit deformation that upgrades weak to strict interlacing. Applications are given to Dedekind zeta derivatives and to nested critical-value blocks of self-dual newforms, yielding strict interlacing and signed resultant inequalities.","tokens_in":49832,"tokens_out":34773,"duration_ms":355413,"significance":"If valid, the results are significant: they establish family-level interlacing across sampling degrees rather than only fixed-degree unit-circle support, identify a sharp uniform strip constant with an explicit counterexample, and connect Jacobi spectral-multiplier theory with arithmetic L-function quotients. The central weak interlacing theorem is largely self-contained and rests on explicit, checkable computations: the checkerboard Bernstein matrices of the shifted Jacobi operator, the shifted-cofactor identities for quartet and paired-outer factors, and the direct discriminant analysis for the unpaired real factor. The exact n=1 threshold and the remote-orbit strictness mechanism are genuine novelties, and the paper supplies falsifiable tests such as Proposition 5.10 and the resultant inequality in Corollary 8.7. The main caveats are a sign error in the strictness lemma and the heavy reliance on the unpublished companion preprint [23] for the arithmetic sections; neither appears to affect the weak interlacing theorem itself, but both affect the full advertised scope.","major_comments":[{"comment":"The final displayed formula of Lemma 7.4 is internally contradictory: the expression (-a(ξ))^{-2} W(p0,q0)(ξ)/|p0(ξ)q0(ξ)| is strictly negative under the Wronskian sign W(p0,q0)<0 established in (35), yet it is asserted to be positive. The preceding root-velocity computation can likely be repaired, because the sign of the oriented gap derivative depends on whether the collision is a right gap or a left gap, but as printed the proof of the sign in this central strict-interlacing lemma is incorrect. This is load-bearing for Theorem 1.3 and for the strict-interlacing applications in Section 8.","section":"Lemma 7.4"},{"comment":"Several theorem-level inputs are imported from the unpublished companion preprint [23] without proof: the fixed-degree unit-circle support theorem (used in Lemma 5.3, Lemma 5.8, and Corollary 5.4) and the simplicity and strict cyclic interlacing of derivative samples (used in Theorem 8.5 and hence in Theorem 8.3, Corollary 8.6, and Corollary 8.12). If [23] is not available to the reader, the arithmetic results in Section 8 are unverified. The central weak interlacing Theorem 1.1(i) is self-contained, but the paper's advertised strict and arithmetic claims are conditional on [23]. The authors should either include complete statements and proofs of the imported theorems or explicitly label the affected results as conditional on [23].","section":"Sections 5.2, 5.3, 8.2, 8.5"}],"minor_comments":[{"comment":"In the paragraph after Theorem 6.1, the polynomial called '2p_{1/2}' is exactly twice the polynomial p_ν defined in (17) specialized to ν=1/2; the factor of 2 should be noted or removed to avoid confusion.","section":"Section 6"},{"comment":"The sentence preceding (63), stating that absolute convergence and the functional equation give Z(F_f)⊆S_{1/2}, is correct but terse; the reader should be reminded that S_{1/2} is the narrow critical strip obtained from the Deligne bound and the functional equation, not a claim of the Lindelöf or Riemann hypothesis.","section":"Section 8.5"},{"comment":"In the orbit-configuration table, the spectral factor 'Jν + a1' should read 'Jν + a I' or 'Jν + a' with the identity operator made explicit; as typeset, 'a1' is ambiguous.","section":"Section 4, table"},{"comment":"The proof of Lemma 8.13 is too terse: the asymptotic (68) for F_f^{(m)} is asserted without displaying the contribution of derivatives of L(f,s), and the reader must rely on [23, Lemma 5.4] to fill the gap.","section":"Appendix A.3"},{"comment":"There are several minor typographical issues, including 'nonforced' in Remark 8.11 and the occasional use of 'row' where 'summand' or 'coefficient index' is meant in Section 8.1; these do not affect the mathematics.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The central weak interlacing theorem is strong and appears sound on a full reading, and the explicit computations make it a worthwhile publication once the strictness lemma and the external-dependency issue are addressed. I would ask the authors to either supply the proofs of the imported statements from [23] or restructure the paper so that the arithmetic claims are clearly marked as conditional; the current presentation makes it difficult for a reader to separate what is proved here from what is imported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the central theorem is real and mostly self-contained; the arithmetic payoff is only as solid as the companion preprint [23].\n\nWhat is genuinely new: the adjacent sampling-degree interlacing question itself. Prior work varied modular weight or level; this paper fixes the source and varies the sampling degree from 2n+1 to 2n+3, then proves a sharp uniform strip width sqrt(15/28) for even sources and width 1 for odd sources under which the reduced quotient pencils are real-rooted. The Jacobi endpoint-pencil framing is a good idea: it reduces the problem to preserving y^n(y+t) under Jacobi spectral multipliers, and the sharp constants come out of zero-orbit total nonnegativity plus a direct discriminant argument for the unpaired outer real pair.\n\nI checked the main proof chain for Theorem 1.1(i) and could not find a load-bearing gap. Lemma 5.1 treats the unpaired outer real factor explicitly, Section 4 handles the reflected nonreal quartets and paired outer factors by total nonnegativity, and the strip constant is derived from the Jacobi spectrum, not borrowed from the conclusion. The proof is self-contained as advertised. The paper is also honest about what remains open: the uniform odd endpoint optimum is not determined, and the capped fixed-n regime is unresolved. Those are quantitative gaps, not flaws.\n\nThe soft spot is real but localized: the reliance on [23]. Lemma 5.3, Lemma 5.8, and Section 8 import theorem-level results on fixed-degree unit-circle support, simplicity, and strict cyclic interlacing of derivative samples. Those are not reproduced here and there is no machine-checked or code-based verification. The central interlacing theorem does not depend on [23]—I agree with the stress-test note on that—but the arithmetic corollaries (Theorem 8.3, Corollaries 8.6 and 8.12) do. If [23] has a gap, the strict interlacing and signed resultants lose their support. That warrants a condition on acceptance, not a rejection.\n\nThe citation pattern is fair: prior period-polynomial interlacing, total positivity, and zero-preserver work is credited, and the companion preprint is flagged as the source of the fixed-degree inputs. No invented entities, no hidden parameters.\n\nWho this is for: people working on zero distributions of period polynomials and L-function values, and anyone interested in Obreschkoff-type preservation under spectral multipliers. It deserves a serious referee. My recommendation: send it to peer review, and ask the referee to scrutinize the imports from [23], or require the author to state explicitly which arithmetic results are conditional until that preprint is independently verified.","headline":"The central adjacent-degree interlacing theorem with sharp Jacobi endpoint-pencil constants is real and self-contained; the arithmetic applications are conditional on an unverified companion preprint and should be reviewed with that in mind.","tokens_in":50342,"tokens_out":1637,"would_cite":true,"duration_ms":20190,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["26C10","15B48","11M26","11F67","11R42","30D15","33C45"],"pacs":[],"model":"deepseek-v4-flash","headline":"A strip half-width of $\\sqrt{15/28}$ is the sharp uniform condition for interlacing of adjacent centered binomial quotients of even entire functions.","keywords":["Jacobi spectral multipliers","endpoint-pencil preservation","real-rooted pencils","weak and strict interlacing","total nonnegativity","centered binomial samples","Dedekind zeta functions","newform critical values"],"falsifier":"Take $F(u)=u^2-15/28$ and compute $C^-_{F,1}$ and $C^-_{F,2}$; if any member of the pencil $C^-_{F,2}+tC^-_{F,1}$ has a nonreal conjugate pair, the strip theorem is false. The boundary is exactly testable: for $F_a(u)=u^2-a^2$, the discriminant of this pencil member as a quadratic in $y$ is $\\Delta_a(t)=(9/4-a^2)^2t^2+(16a^2+60)t+400$, and the theorem predicts $\\Delta_{\\sqrt{15/28}}(t)\\ge0$ for all $t$, while a direct scan at $a=0.74$ produces a negative value for some real $t$, exhibiting the claimed failure outside the strip.","tokens_in":49424,"feed_emoji":"📏","tokens_out":13677,"duration_ms":134675,"temperature":0.7,"pith_summary":"The paper establishes a sharp Sturm-type law for centered binomial samples: if a nonzero even real entire function $F$ of order at most one has all its zeros in the strip $|\\Re u|\\le\\sqrt{15/28}$, then the reduced quotients $C^-_{F,n}$ and $C^-_{F,n+1}$ weakly interlace for every $n\\ge0$, and no larger uniform half-width works. The odd analogue holds with half-width $1$. This is stronger than knowing each sampled polynomial has its zeros on the unit circle, because adjacent samples can both be unit-circle rooted while their reduced quotients fail to interlace. The argument reduces the quotient pair to a Jacobi spectral multiplier acting on the endpoint pencil $y^n(y+t)$, and proves preservation of the one-defect class $E_{n+1}$ through Bernstein variation diminution and total nonnegativity of finite Jacobi matrices. For nonpolynomial sources with simple interior zeros, a remote-zero-orbit deformation turns the weak interlacing into strict interlacing, yielding explicit orderings and signed resultant inequalities for Dedekind zeta derivatives and self-dual newform critical-value blocks.","feed_headline":"Adjacent binomial samples interlace in a sharp 0.732 strip","feed_subtitle":"Even entire functions keep the quotient pencil real-rooted up to this bound; one step beyond, degrees 3 and 5 fail.","key_machinery":"The load-bearing object is the Jacobi spectral multiplier $M_{F,\\nu}$, diagonal in the monic Jacobi basis $P_r^{(\\nu)}$ with eigenvalues $F(r+1/2)$, for $0<\\nu<2$. The endpoint reduction makes the even quotient equal to $M_{F,3/2}(y^n)$ and, after $G=uF$, the odd quotient a multiple of $M_{F,1/2}(y^n)$. The entire interlacing claim is encoded in one pencil: $M_{F,\\nu}(y^n(y+t))\\in E_{n+1}$ for all $t\\in\\mathbb{R}$, where $E_{n+1}$ allows at most one zero outside $[0,4]$, and this implies $M_{F,\\nu}(y^n)\\preceq M_{F,\\nu}(y^{n+1})$. The proof mechanism is Bernstein variation diminution: each zero-orbit factor of $F$ has a checkerboard Bernstein matrix, the Bernstein coefficient matrix with alternating sign conjugation, that is totally nonnegative, so interval-root counts cannot decrease, while a possible unpaired outer real factor is handled by an explicit quadratic discriminant on the pencil. Strict interlacing is obtained by deforming a remote zero orbit to infinity, whose normalized derivative converges on each finite-dimensional space to the Jacobi operator $J_\\nu=y(y-4)D_y^2+(2y-4\\nu)D_y$.","core_discovery":"The central discovery is that the weak interlacing relation between adjacent quotient indices is governed by a structural theorem for Jacobi spectral multipliers. For $0<\\nu<2$ and $n\\ge1$, define $h_{n,\\nu}=\\min\\{1,\\sqrt{1/4+e_{n,\\nu}}\\}$ with $e_{n,\\nu}=n(n+1)(2-\\nu)/(2n+\\nu)$. If $F$ is a real even polynomial or entire function of order at most one with $Z(F)\\subseteq S_{h_{n,\\nu}}$, then every member of the endpoint pencil $M_{F,\\nu}(y^n(y+t))$ lies in the one-defect class $E_{n+1}$, so $M_{F,\\nu}(y^n)\\preceq M_{F,\\nu}(y^{n+1})$. The specializations $\\nu=3/2$ and $\\nu=1/2$ recover the even and odd quotient families, and the strip widths $\\sqrt{15/28}$ and $1$ are sharp where asserted. For nonpolynomial even $E$ with zeros in the same strip and with simple zeros of both images in $(0,4)$, the interlacing is strict, and this is what the arithmetic applications use. The paper also determines the exact optimal strip for the first nontrivial pencil $n=1$: $h^{\\mathrm{opt}}_{1,1/2}=1.065615\\ldots$ and $h^{\\mathrm{opt}}_{1,3/2}=\\sqrt{15/28}$.","pith_inferences":["Because the mechanism is diagonal in an orthogonal-polynomial basis with a totally nonnegative checkerboard matrix, the same endpoint-pencil argument should transfer to other sampling schemes, not only the Jacobi parameters $\\nu=1/2$ and $\\nu=3/2$ used here.","The signed resultant inequality is directly computable for any number field by evaluating $\\Xi_K(1)$, $\\Xi_K(2)$, and $\\Xi_K(3)$; a numerical check of the sign would audit the strict interlacing theorem and localize any missing hypothesis.","The exact first-pencil widths leave a concrete finite optimization problem: the uniform odd-endpoint optimum lies between $1$ and $1.065615\\ldots$, and searching over single-pair versus repeated-pair extremizers might close the gap.","The remote-zero-orbit deformation suggests that strict interlacing is a generic property of nonpolynomial even sources in the admissible strip, independent of arithmetic origin, so arbitrary entire functions with zeros in the strip could be tested numerically to separate the analytic mechanism from the number-theoretic applications."],"forward_implications":["For every even real entire $F\\not\\equiv0$ of order at most one with all zeros in $|\\Re u|\\le\\sqrt{15/28}$, the quotient families satisfy $C^-_{F,n}\\preceq C^-_{F,n+1}$ for all $n\\ge0$, and the half-width cannot be enlarged uniformly.","For every odd real entire $G\\not\\equiv0$ of order at most one with all zeros in $|\\Re u|\\le1$, the anti-reciprocal quotients satisfy $C^+_{G,n}\\preceq C^+_{G,n+1}$ for all $n\\ge0$.","Separate unit-circle rootedness of the two sampled polynomials is insufficient: with $\\sqrt{15/28}