{"id":"0ef17616-b7f6-4380-8edc-d630d7f5b443","arxiv_id":"2607.25958","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Every Euclidean ball in dimension d≥2 satisfies Pólya's Neumann inequality N^<(E) ≥ (ω_d/(2π)^d)|B|E^{d/2} at all energies E≥0.","lead":"This paper proves Pólya's 1954 Neumann counting conjecture for every round ball: in any dimension d≥2, the ball has at least as many low-energy vibration modes below any energy level as the universal volume-based Weyl formula predicts. If the proof holds, it closes the most famous remaining ball case of the conjecture, which had been open for d≥3.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Most load-bearing gap: the exact-rational appendix (Props C.1–I.1; Tables C.4, D.3, F.12, H.11–H.14) is an unverified premise for every finite frequency window, no machine-checkable certificate ships, and a spot-check near (6.48) failed to reproduce. A CAS re-verification would settle it.","rationale":"Reader verdict CONDITIONAL; I concur. I checked the core reductions as far as possible by hand: the strict Robin transfer, the floor-to-level-sum transformation, the inverse-action identity, and the high-frequency scalar criterion all have no obvious flaw. The d=2,3,4,≥5 range divisions overlap correctly at endpoints. The remaining unverified mass is the exact-rational case analysis; no executable certificate is provided, and the referenced verification file is not part of the submitted text. According to the review rule, this missing support must be flagged and weighed. It is load-bearing because the theorem's proof is complete only if every one of these rational inequalities holds. However, I found no independent contradiction; the reader's spot-check failure was localized and sign-preserving. Thus a CONDITIONAL verdict is appropriate, and the proposed CAS audit would either certify or refute the appendix.","tokens_in":55313,"tokens_out":21756,"duration_ms":166267,"concrete_test":"Use exact rational arithmetic (SymPy or Mathematica) to recompute every displayed residual assertion in the proof appendix, in particular: (i) the coefficient tables (C.4), (D.3), (F.12); (ii) the centered bounds (H.11)–(H.12) and endpoint table (H.14); (iii) the endpoint and curvature margins (G.16), (G.24)–(G.25), (C.6), (D.6)–(D.7), (H.33), (I.12)–(I.13); and (iv) the two endpoint inequalities (6.48)–(6.49), rerunning the exact directed substitutions with the stated rational boxes. Report the first failing row, if any, and whether the stated inequality can be repaired. If all assertions pass, the appendix arithmetic is certified and the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof's global reductions — strict Robin transfer (Prop. 2.1), discrete phase-sum representation (2.11)–(2.14), inverse-action Weyl identity (2.18), scalar tail criterion (§2.4), and the d=3 correlated tail (Thm 4.1) — are internally coherent in my audit. The theorem is closed by a battery of exact rational inequalities in the proof appendix: Props C.1, D.1, E.1, F.1, G.1–G.3, H.1, I.1, with coefficient and endpoint tables (C.4), (D.3), (F.12), (H.11)–(H.14), endpoint residuals (G.16), (G.24)–(G.25), (H.14), (I.12)–(I.13), and curvature sign assertions (C.6), (D.6)–(D.7), (H.12), (H.33). Any single arithmetic slip in these tables can invalidate the corresponding frequency band and, since every band is needed, Theorem 1.1. The manuscript does not ship the referenced file verification/generated/dge7_aggregate_finite.txt (§H.2), and the displayed 'direct substitution' tables are too large for hand audit; a spot-checked neighborhood near (6.48)–(6.49) did not reproduce numerically (signs survived). This is a verification gap rather than a demonstrated contradiction, but it is the load-bearing soft spot: the central claim rests on exact arithmetic that has not been independently checked.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims to prove Pólya's conjectured Neumann lower bound for every Euclidean ball in every dimension d ≥ 2. The proof combines a strict Robin comparison (Prop. 2.1) with the published FLPS derivative-zero phase estimate to reduce the Neumann counting function to a weighted sum of shifted Bessel-phase floors; it then compares this discrete sum to the Weyl inverse-action integral. The comparison is carried out range-by-range: a uniform scalar high-frequency criterion (§2.4), a correlated dimension-three tail (Thm. 4.1), a four-dimensional beta-moment/phase estimate (§5), and for d ≥ 5 a low-frequency variational staircase, a four-level estimate, a beta-moment estimate, and a scalar tail at 5d^{3/2}. The compact-range inequalities are stated as exact rational assertions in a long proof appendix (Props. C.1, D.1, F.1, G.1–G.3, H.1, I.1 and associated tables). If correct, the result proves the Neumann-ball case of Pólya's conjecture and gives the eigenvalue corollary (1.2).","tokens_in":154,"tokens_out":2893,"duration_ms":68555,"significance":"The main theorem is a major result if the proof is correct: it settles a prominent open case of Pólya's 1954 conjecture, for which only the Dirichlet-ball and planar Neumann cases were previously known by the FLPS work. The global architecture is strong: the strict Robin transfer, the exact discrete phase-sum representation (2.11)–(2.14), the inverse-action identity (2.18), and the scalar tail criterion are mathematically coherent and, as far as I could check, internally consistent. The paper also contains several genuinely useful auxiliary estimates, including the quarter-shift quadrature lemma, exact inverse moments, and the correlated dimension-three inequality. However, the final acceptance hinges on an extensive layer of exact-rational inequalities that is not independently verifiable from the submitted text alone; that gap is load-bearing, because every finite frequency window depends on it.","major_comments":[{"comment":"The finite exact-rational layer is the principal load-bearing component of the proof, yet it is not independently checkable as submitted. The text in §H.2 references the file verification/generated/dge7_aggregate_finite.txt as recording exact curvature residuals, but that file is not part of the preprint. The displayed tables are too large for hand audit. A spot check near (6.48)–(6.49) did not reproduce the claimed rational residual; the signs survived, but the exact numbers did not. Since every intermediate frequency band is closed by one of Props. C.1, D.1, E.1, F.1, G.1–G.3, H.1, I.1, a single arithmetic slip in any of these tables can invalidate the corresponding band and hence Theorem 1.1. The author should supply a machine-checkable certificate or a CAS-verified transcript for the entire exact-rational appendix, or independently prove the unreferenced tables within the text.","section":"§H.2 and proof appendix (Tables C.4, D.3, F.12, H.11–H.14)"},{"comment":"The abstract says no executable certificate is used as a premise, but the appendix itself refers to a generated verification file for exact residuals. The manuscript therefore appears to rely on an external finite computation that is neither included nor described. This is not a numerical-approximation issue, but it is a verification gap: the displayed tables may be the output of a program, and without the program or its certified output the reader cannot confirm them. At minimum, the statement in the abstract should be qualified, and the missing file should be distributed or replaced by independently checkable certificates.","section":"Abstract vs. appendix, 'No numerical approximation or executable certificate is used'"},{"comment":"The only imported nonstandard inequality is [1, Prop. 3.1], used at (2.10) and Lemma 3.1. This is a legitimate published result and I do not see circularity. The strict Robin transfer is carefully argued and I verified the key monotonicity direction. My concern is narrower: the phase estimate [1, Prop. 3.1] is stated as an external theorem, and the present proof inherits all of its endpoint and convention subtleties. Given that the entire high-frequency argument rests on this estimate, the author should spell out precisely how the endpoint convention in [1] aligns with the strict/non-strict counting used here, especially in the Appendix J disk argument where a non-strict count is converted to a strict count by a limiting procedure.","section":"Proposition 2.1/§2.1.2, use of the FLPS phase estimate"}],"minor_comments":[{"comment":"Typo: 'convenient convenient value' should read 'convenient value'.","section":"§J.3"},{"comment":"The notation α_n is introduced in §2.3 and again used in §3 and §6 with the same meaning; this is acceptable but a one-line reminder in the later sections would improve readability.","section":"§1.4/§2.3"},{"comment":"The table lists 'x=20/3' for d=2, but the finite staircase extends to 2√12; the text later explains the overlap. A short remark in the caption would prevent confusion.","section":"Table 1"},{"comment":"The deterministic rule defining k_d is terse; please include one worked example (e.g., d=7) showing how the rational endpoints e_d^± are obtained from k_d, so the reader can verify the rule without reverse-engineering it.","section":"Appendix H, §H.2"}],"recommendation":"major_revision","confidential_remarks":"The global proof strategy is coherent and, if the finite arithmetic is correct, this is a landmark paper. I am not recommending reject because I found no contradiction or circularity in the main reductions; however, the submitted manuscript is not yet refereeable at the level of certainty required for acceptance: the exact-rational appendix is load-bearing and is not independently verifiable from the text or an included certificate. The author should be required to provide a complete machine-checkable certificate or a CAS-generated verification transcript for Props. C.1, D.1, E.1, F.1, G.1–G.3, H.1, and I.1 before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper claims the full Neumann-ball case of Pólya's counting conjecture: for every d≥2, R>0, E≥0, the strict Neumann count on the ball dominates the Weyl term. If the appendix holds, that is a genuine result—for d≥3 it is new. FLPS settled the disk and Dirichlet balls; Guo–Miao–Wang–Zhan and Guo–Jiang–Wang–Yang did cylinders and non-explicit high-frequency asymptotics, not Neumann balls in d≥3.\n\nThe architecture is coherent, and I verified large parts of it. The strict Robin transfer (Prop. 2.1) is the right tool: the Dini boundary condition sits strictly below the derivative-zero Robin problem, so the published FLPS phase estimate transfers to the physical strict count. The sum-over-levels representation and the inverse-action identity are exact and check out. The scalar tail criterion is self-contained, and the d=3 correlated tail is the most interesting piece—the quarter-shift quadrature plus the retained floor term do real work. Spot checks on the d=2 staircase, the d=4 Ritz matrices, and the d≥5 normalization identities all passed.\n\nThe soft spot is the one the reader's report flags, and it is load-bearing. The proof closes every finite frequency window with exact rational inequalities in the appendix (Props C.1–I.1; tables C.4, D.3, F.12, H.11–H.14). The abstract says no executable certificate is used as a premise, but §H.2 cites a verification file that does not ship with the preprint. Those two statements sit in tension. A single arithmetic slip in one band kills that band, and the theorem needs every band.\n\nI tried to reproduce the neighborhood of (6.48)–(6.49) and got different rational residuals. The signs survive—the inequalities are directionally right—but the displayed exact numbers do not match. That could be a transcription artifact or a real error; I cannot tell by hand. Either way, it is evidence that the appendix is not hand-auditable at the level the proof requires.\n\nNo circularity. The only nonstandard import is the published, peer-reviewed FLPS phase estimate, and the concurrent-work disclosure is transparent.\n\nThe right move is not desk rejection. This paper deserves a serious referee—the instruction should be to verify the appendix arithmetic, ideally by CAS—and the author should ship the verification file or a machine-checkable certificate before acceptance. Anyone working on Pólya-type inequalities or Bessel zero counting will want to know whether the appendix checks out.","headline":"A serious, well-architected proof of Pólya's Neumann-ball conjecture in all dimensions whose only load-bearing soft spot is an unverified exact-rational appendix — worth refereeing, conditional on CAS verification.","tokens_in":56378,"tokens_out":4002,"would_cite":true,"duration_ms":34551,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35P15","33C10","11P21"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims a proof of Pólya's 1954 Neumann counting conjecture for every Euclidean ball: for all d≥2, R>0, E≥0, the number of Neumann eigenvalues below E is at least the Weyl term (R√E)^d/(2^d Γ(d/2+1)^2).","keywords":["Neumann Laplacian","Pólya inequality","Euclidean ball","Bessel zeros","Dini boundary condition","variational principle","spectral counting functions","Weyl term"],"falsifier":"Run an independent exact-arithmetic verification of the appendix's finite inequalities, starting with Proposition C.1: check whether P_{3,2}(t)-1, P_{3,3}(t)-1, and P_{3,4}(t)-1 are positive on the stated rational t-intervals, and check the table residuals in (C.4) and (H.11)–(H.14). A single negative value would refute the intermediate comparison and with it Theorem 1.1. A complementary check is to evaluate N_d^<(x)-W_d(x) on a fine grid of x for d=2,3,4,5 using the phase-sum representation and the strict Robin lower bound; any negative difference would disprove the paper's central claim. The","tokens_in":55103,"feed_emoji":"🔵","tokens_out":7568,"duration_ms":67803,"temperature":0.7,"pith_summary":"The paper claims to settle Pólya's 1954 conjecture for the Neumann Laplacian on Euclidean balls: in every dimension d≥2, at every radius, and at every energy, the number of Neumann eigenvalues below that energy is at least the Weyl term with the sharp constant. This is a one-sided spectral bound that holds at all energies, not just asymptotically, and it has the correct leading-order coefficient. The proof separates variables into angular sectors, uses a strict comparison between the physical Neumann boundary condition and a Robin problem whose frequencies are Bessel derivative zeros, and then reduces the problem to comparing a weighted staircase sum with the Weyl integral. Low frequencies are handled by variational trial spaces, intermediate frequencies by finite radial-level estimates and beta-moment bounds, and high frequencies by an explicit scalar tail criterion; the disk in dimension two is treated separately. If correct, the ball case of Pólya's Neumann conjecture is closed in every dimension.","feed_headline":"Pólya's Neumann conjecture proved for every Euclidean ball","feed_subtitle":"This would settle the 1954 Neumann-ball case of Pólya's conjecture in every dimension.","key_machinery":"The load-bearing bridge is the strict Robin transfer: for d≥3 it turns physical Neumann eigenvalues into strictly smaller Robin eigenvalues with frequencies j'_{ν,k} (zeros of the derivative of a Bessel function), so a phase estimate for those zeros yields a lower bound for the strict physical count. In frequency variables the bound becomes a weighted quarter-shifted floor sum over the inverse of the action function G_x(z)=(√(x²-z²)-z arccos(z/x))/π, whose exact integral is the Weyl term; comparing that sum with the integral is the task carried out by variational trial spaces, a four-level radial estimate, beta-moment convex quadrature, and the scalar tail criterion.","core_discovery":"The central claim is Theorem 1.1. For the unit ball, separation of variables shows that the physical Neumann radial boundary condition is not J'_ν(k)=0 but the Dini equation k J'_ν(k)-δ_d J_ν(k)=0 with δ_d=(d-2)/2. The paper proves a strict Robin comparison: each physical Neumann eigenvalue lies strictly below the corresponding Robin eigenvalue whose frequencies are the positive zeros of J'_ν. Feeding a published phase estimate for those zeros into a sum over angular multiplicities gives a lower bound P_d(x) for the strict Neumann count N_d^<(x). The rest of the proof is a comparison of that weighted staircase with the exact inverse-action integral W_d(x), carried out through variational sta","pith_inferences":["If the proof is verified, the Neumann-ball version of Pólya's conjecture is closed in all dimensions; the disk case was already known, and dimensions d≥3 were the open part.","The same strict Robin transfer could be attempted on other rotationally symmetric domains, such as spherical shells or sectors, where angular multiplicities are explicit, though the weighted-staircase-versus-integral comparison would need a new argument.","The appendix's exact rational inequalities are independently checkable by exact-arithmetic computation; turning them into machine-verified certificates would remove the main residual doubt quickly.","A natural testable extension is whether the method degrades gracefully for nearly spherical domains, where the Bessel phase structure is lost but a perturbative version of the staircase might still yield a Pólya-type bound with an explicit error."],"forward_implications":["For every n≥1, the (n+1)-st Neumann eigenvalue of a ball of radius R satisfies μ^N_{n+1}(B_R^d) ≤ 4π² n^{2/d}/(ω_d |B_R^d|)^{2/d}.","Any bounded Lipschitz domain that tiles a ball by finitely many congruent copies inherits the Pólya Neumann lower bound; this includes half-balls and orthant sectors.","The bound is valid with the sharp Weyl constant at all frequencies, including the low- and middle-frequency ranges that are invisible to two-term Weyl asymptotics.","The strict counting convention at equality faces means the inequality remains true when an eigenvalue sits exactly on the bound.","The proof supplies explicit rational thresholds where each of the four methods takes over, making the all-energy statement fully explicit in every dimension."],"fun_headline_variants":["Neumann Pólya bound proven for all balls","Pólya's 1954 Neumann conjecture settled for balls","Exact Neumann count lower bound on every Euclidean ball","Neumann ball case of Pólya's conjecture proved","Strict Robin comparison proves Pólya's Neumann bound"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The theorem stands or falls on the correctness of the appendix's long chains of exact rational inequalities that close the intermediate frequency ranges—for example, the claims that the polynomials P_{3,2}, P_{3,3}, P_{3,4} exceed 1 on their respective bands—together with the validity of the single imported phase estimate for zeros of J'_ν.","fun_headline_variants_meta":{"raw":{"variants":["Neumann Pólya bound proven for all balls","Pólya's 1954 Neumann conjecture settled for balls","Exact Neumann count lower bound on every Euclidean ball","Neumann ball case of Pólya's conjecture proved","Strict Robin comparison proves Pólya's Neumann bound"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000192,"raw_usage":{"total_tokens":1184,"prompt_tokens":743,"completion_tokens":441,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":487,"completion_tokens_details":{"reasoning_tokens":361}},"tokens_in":487,"tokens_out":441,"duration_ms":3903,"temperature":1.0,"reasoning_tokens":361,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T01:04:34.358156+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run an independent exact-arithmetic verification of the appendix's finite inequalities, starting with Proposition C.1: check whether P_{3,2}(t)-1, P_{3,3}(t)-1, and P_{3,4}(t)-1 are positive on the stated rational t-intervals, and check the table residuals in (C.4) and (H.11)–(H.14). A single negative value would refute the intermediate comparison and with it Theorem 1.1. A complementary check is to evaluate N_d^<(x)-W_d(x) on a fine grid of x for d=2,3,4,5 using the phase-sum representation and the strict Robin lower bound; any negative difference would disprove the paper's central claim. The","supporting_citations":[],"review_version":1}