{"id":"6ea36c79-2d0c-4333-bbae-692926bc2358","arxiv_id":"2507.09203","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"On complete Kähler manifolds with HSC ≥ 2, the first eigenvalue of the Laplacian is at least (320n+256)/(81n+63), which tends to 320/81 as n grows.","lead":"A new Bochner-Kodaira identity proves that every complete Kähler manifold with holomorphic sectional curvature at least 2 has first Laplacian eigenvalue at least (320n+256)/(81n+63). The bound approaches the conjectured sharp constant 4, and the new identity may apply more broadly in spectral geometry.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Inequality (2.29) is load-bearing and rests on sign conventions in Lemma 2.1; the proof also mishandles M0 and leaves n=1 unresolved. An independent algebraic check is needed before acceptance.","rationale":"The paper's central claim is an eigenvalue lower bound for complete Kähler manifolds with HSC ≥ 2, and the proof is a long chain of integral identities culminating in the pointwise inequality (2.29). The reader identified the sign conventions in the ω2 computation as the weakest assumption; I agree that this is the right locus, but my stress-test did not find an outright algebraic error in the displayed chain from (2.37) to (2.43). The discriminant computation with κ0 is correct, and the Cauchy step is valid. However, two proof-hygiene issues reinforce the need for an independent check: the assertion about ∂∂̄f vanishing on M0 is false as stated, and the n=1 case is not covered by the argument. Both are fixable, and neither by itself shows the theorem is false. Because the reader's verdict is already CONDITIONAL and my concerns align with the same region of the proof, I do not recommend changing the verdict. A concrete symbolic/numeric check of (2.29) would settle whether the sign conventions actually break the central inequality.","tokens_in":13063,"tokens_out":45498,"duration_ms":517340,"concrete_test":"Use a computer algebra system to expand inequality (2.29) from equations (2.37)-(2.43) for a symbolic Hermitian n×n matrix A with κ0 as in (2.39), and test random numeric A for n = 2,...,10. If the inequality fails for any A, the constant in Theorem 1.4 is wrong. Separately verify Lemma 2.1's identity ω2 = I_V∂∂̄f for f = |z|² with the paper's convention V = g^{i\\bar k}φ_k ∂_{z_i}, tracking whether the coefficient of aₙₙ changes sign.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The numerical constant in Theorem 1.4 is entirely carried by the pointwise inequality (2.29). Its proof, equations (2.37)-(2.43), depends on the identification ω2 = I_V∂∂̄f and on the convention Δ∂f = -∑ᵢ aᵢᵢ. A single sign error in either identification would change the coefficient -1/2∑ aᵢᵢ aₙₙ to +1/2∑ aᵢᵢ aₙₙ, altering the final constant from 320/81 to a different value. I checked the later algebra: the Cauchy step, the discriminant completion with κ0 from (2.39), and the positivity of the quadratic form are consistent. The weak spot is the unverified sign convention itself, not the subsequent algebra. In addition, equation (2.31) is asserted with an incorrect justification: ∂∂̄f need not vanish a.e. on M0 (for f = |z|², the critical point has nonzero Hessian), so the passage from M\\M0 to M requires a separate proof that M0 has measure zero, which is not given. Finally, the sentence 'When n=1 and assume n≥2' leaves the n=1 case unproved, since (2.29) divides by n-1. These are fixable gaps, but they make the central inequality not fully established as written.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a lower bound for the first positive eigenvalue of the Laplacian on a complete Kähler manifold whose holomorphic sectional curvature satisfies HSC ≥ 2. The main result, Theorem 1.4, states that λ1 ≥ (320(n−1)+576)/(81(n−1)+144), a constant that approaches 320/81 ≈ 3.95 as the complex dimension tends to infinity. The proof introduces a Bochner–Kodaira type identity for the (1,0)-gradient of an eigenfunction, reduces the global estimate to a pointwise algebraic inequality, and then verifies that inequality by a discriminant computation. The paper also states a general Bochner–Kodaira formula (Theorem 1.5), a corollary λ1 ≥ 2, and a product example showing the optimal constant is at most 4, together with a conjecture that the sharp lower bound is 4 with equality only on CP^1.","tokens_in":13300,"tokens_out":29518,"duration_ms":315304,"significance":"If the main theorem is correct, it gives a uniform spectral gap under a positive holomorphic sectional curvature lower bound, a qualitatively new phenomenon in Kähler spectral geometry: the lower bound does not grow with dimension, unlike the Ricci-curvature analogue. The proof is essentially self-contained and the algebraic core is explicit and checkable; the constant κ0 in (2.39) is obtained by solving a discriminant equation rather than by fitting, and the paper includes examples and a sharpness conjecture. The derivation from (2.26) to (2.28) via the combinatorial inequality is coherent, and the subsequent quadratic-form argument is valid provided the stated sign conventions hold. The main limitations are local proof gaps involving the n=1 case and the treatment of the zero set of φ, as detailed below.","major_comments":[{"comment":"The sentence 'When n=1 and assume n≥2' leaves the n=1 case unproved. The algebraic proof of (2.29) divides by n−1 in (2.38) and (2.41), so the argument only covers n≥2. Since Theorem 1.4 is stated for all n and the constant at n=1 equals 4, a separate argument is required; for example, in complex dimension one HSC ≥ 2 implies the real sectional curvature is ≥ 2 and Lichnerowicz's theorem gives λ1 ≥ 4, or an independent direct proof should be supplied. As written, the full statement of Theorem 1.4 is not established.","section":"§2, proof of Theorem 1.4, after 'We set λ = λ1/2'"},{"comment":"The assertion '∂∂̄ f = −∂φ = 0, a.e. on M0' is not correct as stated. For f = |z|², the point where ∂f = 0 has nonzero ∂∂̄ f, and ∂φ = ∂∂f = 0 identically for every smooth f, so the equality conflates ∂∂̄ f with ∂∂ f. This matters because (2.31) is used to pass the integrals in (2.32) from M\\M0 to all of M. The intended fact is that a nonconstant real-analytic eigenfunction has ∂f not identically zero, hence M0 has measure zero; this should be stated and proved, or replaced by a correct citation. Without this justification, the reduction from (2.30) to (2.35) is incomplete.","section":"§2, equation (2.31)"},{"comment":"The pointwise inequality (2.29) is load-bearing, and its verification depends on the sign convention Δ∂ f = −Σ_i a_{ii} and on the identification ω2 = I_V ∂∂̄ f. These conventions are nowhere stated explicitly; a reader using the opposite sign for Δ∂ would obtain +1/2 Σ a_ii a_nn instead of −1/2 Σ a_ii a_nn in (2.37), changing the final constant. The authors should state their conventions for Δ∂, ∂∂̄, and the Hermitian pairing on forms before (2.37), and should also justify the combinatorial inequality (2.38) using a_{ij} = overline{a_{ji}} and a_{ii} ∈ R. This is a documentation gap in a central computation, not a demonstrated algebraic error.","section":"§2, equations (2.37)–(2.43)"}],"minor_comments":[{"comment":"The denominator in the second term is written as (|φ|+ε)^2; the subsequent use of the lemma in §2 has the correct denominator (|φ|²+ε)^2. The displayed formula should be corrected.","section":"Lemma 2.2, equations (2.14)–(2.15)"},{"comment":"The notation ∂∂̄ f = Σ a_{ij} e_i ∧ e_j is ambiguous for a (1,1)-form; it should be written in a unitary frame as Σ a_{ij} e^i ∧ \\bar e^j, with a_{ij} = f_{i\\bar j}.","section":"§2, equation (2.37)"},{"comment":"The inequality is asserted without proof; it follows from the Hermitian symmetry a_{ij} = overline{a_{ji}} and the reality of the diagonal entries, but this should be stated explicitly since the inequality is not obvious from the displayed expression alone.","section":"§2, inequality (2.38)"},{"comment":"The reference [Eva2010, p.310] does not support the asserted measure-zero statement; if the intended fact is that the zero set of a nonconstant real-analytic 1-form has measure zero, a correct reference or proof should be given.","section":"§2, equation (2.31)"},{"comment":"The compactness assertion under HSC ≥ 2 is attributed to [Tsu57] and [XY24+]; [XY24+] is a preprint and its precise role should be clarified or removed from this citation.","section":"§2, start of proof of Theorem 1.4"},{"comment":"The statement that |V|, |ω1|, |ω2| are uniformly bounded by C|φ| is plausible because ω1 and ω2 are linear in φ, but this should be made precise, especially near M0, to justify the dominated convergence argument leading to (2.25).","section":"§2, before (2.25)"}],"recommendation":"major_revision","confidential_remarks":"The core algebraic structure of the proof appears sound: the reduction from (2.26) to (2.28) is coherent, the discriminant choice of κ0 is legitimate, and the final constant is derived without free parameters. The main risks are the n=1 gap, the incorrect justification of the measure-zero passage in (2.31)–(2.32), and the unstated sign conventions in the key pointwise inequality. These are fixable within the scope of the paper, so major revision rather than rejection seems appropriate. An independent computational check of (2.37)–(2.43) with an explicit sign convention would strengthen the paper and help the referee and readers."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read. The paper proves a new lower bound λ1 ≥ (320n+256)/(81n+63) for complete Kähler manifolds with HSC ≥ 2, approaching 320/81 ≈ 3.95, close to the conjectured 4. The Bochner-Kodaira identity (1.3) and the method are genuinely new, and the product example correctly shows the optimal bound is at most 4 and dimension-independent. The derivation is self-contained; no fitted constants, no circularity. I traced the main reduction from (2.26) to (2.28), and the combinatorial positivity argument is coherent.\n\nThe soft spots are real but local. First, (2.31) is wrong as stated: '∂∂̄f = -∂φ = 0 a.e. on M0' does not follow from φ=0 on M0—take f=|z|^2, at the critical point the Hessian is nonzero. What is needed is a proof that the critical set of a nonconstant eigenfunction has measure zero, which is true for elliptic solutions but not cited or shown. The passage from integrals over M\\M0 to M rests on this. Second, the sentence 'When n=1 and assume n≥2' leaves the n=1 case unproved; the proof of (2.29) divides by n-1. This is a fixable gap, but it is a gap. Third, the sign of the term -1/2 Σ a_ii a_nn in the proof of (2.29) depends on conventions in Lemma 2.1; if the sign flips, the constant changes. The later algebra is consistent, but I can't certify this from the text.\n\nNone of these look fatal. The n=1 case can be handled separately, the measure-zero issue is standard, and the sign can be checked. The core idea is new and worth publishing once the proof is tightened. I'd send this to a good referee. For my own use, I'd cite it once the identity is verified; the constant is close enough to 4 to be interesting, and the Bochner-Kodaira route may be useful elsewhere.","headline":"New Bochner-Kodaira eigenvalue bound under HSC≥2 is plausible and close to sharp, but the proof has two fixable gaps and a sign convention that needs checking.","tokens_in":13913,"tokens_out":10935,"would_cite":false,"duration_ms":113987,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C55","58J50"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves a dimension-dependent lower bound, tending to 4, for the first eigenvalue of any complete Kähler manifold with holomorphic sectional curvature at least 2.","keywords":["first eigenvalue","Kähler manifold","holomorphic sectional curvature","Bochner-Kodaira identity","spectral gap","Laplacian","eigenvalue lower bound","complete Kähler manifold"],"falsifier":"Check the algebraic core directly: for any $n$, take a Hermitian matrix $A = (a_{ij})$ representing $\\partial\\partial\\bar f$ in an orthonormal frame, compute the left side $-(9/8)\\sum |a_{ni}|^2 + (7/8)|a_{nn}|^2 - (1/2)\\sum a_{ii} a_{nn}$ and compare it with the right side $-(9/16)\\sum |a_{ij}|^2 - \\kappa_0 (\\sum a_{ii})^2$ for $\\kappa_0 = \\frac{16(n-1)+27}{80(n-1)+144}$. A single Hermitian matrix for which the inequality fails would invalidate the theorem; a numerical search over random matrices for small $n$ would settle the key step.","tokens_in":12804,"feed_emoji":"📐","tokens_out":14326,"duration_ms":145162,"temperature":0.7,"pith_summary":"This paper proves a quantitative spectral gap for all complete Kähler manifolds whose holomorphic sectional curvature is bounded below by 2. The first nonzero eigenvalue of the Laplace operator is shown to be at least $\\frac{320(n-1)+576}{81(n-1)+144}$, a dimension-dependent constant that always exceeds $320/81 \\approx 3.95$ and increases to 4 as the complex dimension $n$ grows. The proof is built on a new Bochner-Kodaira identity that connects the curvature term to a Hodge-theoretic decomposition of the eigenfunction differential, and the sharp numerical constant comes from a pointwise algebraic inequality. A product example shows the optimal universal lower bound, if it exists, cannot exceed 4, and the paper conjectures the sharp value is 4, attained only by $\\mathbb{CP}^1$ with the Fubini-Study metric.","feed_headline":"First eigenvalue on curved Kähler manifolds is at least 3.95","feed_subtitle":"A new curvature identity gives a dimension-dependent bound that approaches 4 for large n.","key_machinery":"The machinery is a Bochner-Kodaira-type identity for the $(1,0)$-gradient $\\varphi = \\partial\\bar f$ of an eigenfunction, together with the two 1-forms $\\omega_1 = \\{\\partial_E \\varphi, \\varphi\\}$ and $\\omega_2 = \\{\\bar\\partial_E \\varphi, \\varphi\\}$ defined through the Chern connection on the holomorphic cotangent bundle. The identity shows that $2\\lambda\\int |\\varphi|^4$ equals the sum of the holomorphic sectional curvature term $\\int R(V,V,V,V)$, a positive Hessian term $\\int |\\varphi|^2|\\partial\\partial\\bar f|^2$, and two nonnegative $L^2$ norms. The proof of the main estimate is a logarithmic modification of this identity: applying it to the regularized form $(\\varphi\\wedge\\varphi)/(|\\varphi|^2+\\varepsilon)$ and taking $\\varepsilon\\to 0$ produces a pointwise algebraic inequality (2.29), which is exactly where the dimension-dependent constant arises. This inequality is obtained by writing $\\partial\\partial\\bar f$ at a point in an orthonormal frame adapted to $V$ and optimizing a quadratic form in the matrix entries.","core_discovery":"The central claim is that on a complete Kähler manifold of complex dimension $n$ with $\\mathrm{HSC} \\ge 2$, the first eigenvalue $\\lambda_1$ of the Laplacian satisfies $\\lambda_1 \\ge \\frac{320(n-1)+576}{81(n-1)+144}$. Equivalently, the Poincaré inequality (1.2) holds with that constant. The proof establishes a new identity (Theorem 1.5): for an eigenfunction $f$ with $\\partial\\bar f = \\varphi$ and dual vector field $V$, $2\\lambda\\int |\\varphi|^4 = \\int(R(V,V,V,V)+|\\varphi|^2|\\partial\\partial\\bar f|^2) + \\|\\omega_1 - \\lambda f \\varphi\\|^2 + \\|\\omega_1\\|^2$, where $\\omega_1$ and $\\omega_2$ are the $(1,0)$ and $(0,1)$ parts of the Chern connection acting on $\\varphi$. This identity is then used in a limiting procedure with the regularized form $|\\varphi|^2+\\varepsilon$, and the key pointwise inequality (2.29) converts the curvature information into the final constant. The paper also shows that products of complex projective spaces with appropriate weights have $\\lambda_1 \\ge 4$, so the theorem's constant cannot be pushed above 4 in general.","pith_inferences":["If the conjecture holds, the extremal manifold would be two-dimensional, in contrast to the Ricci-curvature case where $\\mathbb{CP}^n$ is extremal in every dimension; the dimension-independent nature of the HSC bound would be a genuinely new phenomenon.","The pointwise inequality (2.29) is a purely algebraic statement about Hermitian matrices; testing it numerically for small $n$ would be a quick check of the proof's core, and it might be sharpened to improve the constant toward 4.","The same Bochner-Kodaira identity may apply to other geometric quantities, such as higher eigenvalues or the bottom of the spectrum on noncompact manifolds, since it does not rely on compactness beyond the completeness assumption.","Constructing manifolds with $\\lambda_1$ close to $320/81$ would require a family of examples quite different from products of projective spaces; the product formula gives $\\lambda_1 \\ge 4$, so the region between 3.95 and 4 is currently unexplored."],"forward_implications":["Any complete Kähler manifold with $\\mathrm{HSC} \\ge 2$ has a spectral gap of at least about 3.95, so the low end of the spectrum cannot be crowded near zero regardless of dimension.","The Poincaré inequality (1.2) gives an explicit constant for functions of zero mean, a quantitative control usable in Sobolev-type arguments.","The new identity (1.3) immediately yields the weaker bound $\\lambda_1 \\ge 2$ (Corollary 3.1) and provides a template for other curvature-eigenvalue estimates.","The product of complex projective spaces in Example 1.6 realizes $\\lambda_1$ values at least 4, showing the optimal universal constant is no larger than 4.","The paper proposes Conjecture 1.7 that the sharp bound is $\\lambda_1 \\ge 4$ with equality characterizing $\\mathbb{CP}^1$."],"supporting_citations":[{"why":"Establishes that a complete Kähler manifold with positive holomorphic sectional curvature is compact, the step that lets the proof work on a closed manifold.","marker":"[Tsu57]"},{"why":"Cited alongside [Tsu57] as a modern source for the same compactness fact used to begin the proof of Theorem 1.4.","marker":"[XY24+]"},{"why":"Supplies the PDE fact that $\\partial\\partial\\bar f = -\\partial\\varphi = 0$ almost everywhere on the zero set of $\\varphi$, which justifies passing the regularized integrals to the limit over $M\\setminus M_0$.","marker":"[Eva2010]"}],"fun_headline_variants":["Kähler eigenvalue bound approaches 4 as dimension grows","New identity yields eigenvalue lower bound ~3.95 on Kähler manifolds","Holomorphic curvature condition sets eigenvalue floor near 4","First eigenvalue estimate on Kähler manifolds improves to near 4","Dimension-dependent eigenvalue bound from Bochner-Kodaira identity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole argument hinges on the pointwise algebraic inequality (2.29) and the sign convention under which it is derived; if a single Hermitian matrix violates that inequality, the constant $\\frac{320(n-1)+576}{81(n-1)+144}$ would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Kähler eigenvalue bound approaches 4 as dimension grows","New identity yields eigenvalue lower bound ~3.95 on Kähler manifolds","Holomorphic curvature condition sets eigenvalue floor near 4","First eigenvalue estimate on Kähler manifolds improves to near 4","Dimension-dependent eigenvalue bound from Bochner-Kodaira identity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000874,"raw_usage":{"total_tokens":3766,"prompt_tokens":916,"completion_tokens":2850,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":532,"completion_tokens_details":{"reasoning_tokens":2762}},"tokens_in":532,"tokens_out":2850,"duration_ms":22840,"temperature":1.0,"reasoning_tokens":2762,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:04:48.374311+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check the algebraic core directly: for any $n$, take a Hermitian matrix $A = (a_{ij})$ representing $\\partial\\partial\\bar f$ in an orthonormal frame, compute the left side $-(9/8)\\sum |a_{ni}|^2 + (7/8)|a_{nn}|^2 - (1/2)\\sum a_{ii} a_{nn}$ and compare it with the right side $-(9/16)\\sum |a_{ij}|^2 - \\kappa_0 (\\sum a_{ii})^2$ for $\\kappa_0 = \\frac{16(n-1)+27}{80(n-1)+144}$. A single Hermitian matrix for which the inequality fails would invalidate the theorem; a numerical search over random matrices for small $n$ would settle the key step.","supporting_citations":[],"review_version":1}