{"id":"150dd08f-9172-4881-92b0-de8d9b542312","arxiv_id":"2602.13951","paper_version":2,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":2,"one_line_summary":"Sections of Hodge bundles built from Beltrami differentials give explicit deformations of (p,p)-classes, Kähler cones, algebraic approximation, and Hodge loci for compact Kähler manifolds.","lead":"This paper develops a Hodge-theoretic machine using Beltrami differentials and period matrices to track (p,p)-cohomology classes across deformations of compact Kähler manifolds, claiming upper-semicontinuity of Kähler cones and criteria for algebraic approximation and the variational Hodge conjecture. A generalist might read it because it promises explicit, checkable controls on when Kähler properties and Hodge classes survive deformation, a recurring need in complex geometry","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.4's proof uses an invalid type comparison: equation (52) does not follow from (49)–(51), leaving the central Kähler-cone containment unproven.","rationale":"The paper has a plausible framework: the Hodge map construction in Theorem 2.4 via the implicit function theorem is coherent, and the explicit formulas for period matrices and extended classes are valuable. However, the proof of the main Kähler-cone containment contains a concrete unjustified step. The comparison of types in the proof of Theorem 3.4 ignores that the base-harmonic forms eη^(0), eη^(2) are not of pure type in the nearby complex structure; they contribute to the (1,1) part through the Beltrami differential. This is not a matter of external consensus or a missing reference — it is an internal gap in the derivation. The reader's weakest assumption, identity (22), is also load-bearing because the period-map blocks are imported from the companion paper, but even granting that identity, the positivity conclusion does not follow from the written argument. A repair might be possible (e.g., by a more careful computation of the (1,1)-component including α terms, or by working with the X_t-harmonic representative), but the paper as stated does not supply it. Thus the reader's REJECT verdict is supported, with a slightly different emphasis.","tokens_in":35315,"tokens_out":8987,"duration_ms":79915,"concrete_test":"On a compact Kähler manifold with h^{2,0}>0 (e.g., a non-algebraic K3 or a two-dimensional complex torus), compute the X_t-(1,1) component of eH^ω(σ,t) to order t^2 using the explicit α^(0)(σ,t) determined by equation (45). Expand eη^(0)_ω in the X_t-coframe and check whether the coefficient of dz^i∧dzbar^j equals (I−S)^{-1}(ω_{ij}) or differs by the term α^(0)(t)(φ(t)⌟eη^(0)_ω). If the difference is nonzero, equation (52) is false and the proof of Theorem 3.4 fails as written.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 3.4 — the core of Theorem 0.3(1) — derives equation (52) by “comparison of types” between the ω-harmonic representative eH^ω(σ,t) in (49) and its coordinate expansion in the X_t-frame in (50)–(51). This comparison is not legitimate. Equation (49) is a decomposition according to the Hodge types of the central fiber X: eH^ω = α^(0)eη^(0)_ω + ω + \\overline{α^(0)}eη^(2)_ω. But in the X_t-coframe (dz^i + φ dzbar^i) ∧ (dzbar^j + \\overline{φ} dz^j), the forms eη^(0)_ω and eη^(2)_ω are not pure (2,0) and (0,2); they acquire nontrivial (1,1)-components, e.g. φ⌟eη^(0)_ω. Therefore the (1,1)-component of eH^ω on X_t is not ω but ω plus contributions from α^(0) and its conjugate. No argument shows these contributions vanish or are absorbed into the matrix g_{ij}(z,t). The claimed equality (52) is thus an assumption, not a consequence. The subsequent positivity argument — that (I−S)^{-1}(g_{ij}) is positive definite whenever ∥φ∥_E<1 — establishes positivity of a matrix that need not be the X_t-(1,1) component of eH^ω. Consequently, the containment K^{∇1,1}_{t0,t} ⊂ K_t is not established. The same flaw propagates into Proposition 0.2 and Theorem 4.2. The reader's concern about identity (22) being imported from [20] is valid but secondary; the (52) gap is internal to this paper.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops an explicit 'Hodge map' parametrizing nearby (p,p)-classes on deformations of a compact Kähler manifold, using period-matrix blocks and Beltrami differentials from the authors' companion preprint [20]. It defines ∇^{1,1}-flat extensions of Kähler cones, claims upper semicontinuity (K^{∇1,1}_{t0,t} ⊂ K_t) with explicit positive representatives, and a large-scale Kähler stability theorem. It further claims generalizations of Green's density criterion, approximation of real (p,p)-classes by Hodge classes, and a Beltrami-differential criterion for the variational Hodge conjecture. The main theorems are stated for possibly singular Kuranishi bases and obstructed deformations.","tokens_in":35942,"tokens_out":2969,"duration_ms":27321,"significance":"If the central constructions and theorems were correct, the paper would provide a substantial new tool: explicit, higher-order Beltrami expressions for the deformation of (p,p)-classes and Kähler cones, with applications to algebraic approximation and the variational Hodge conjecture. The explicit positivity formula in Proposition 0.2 and the uniform large-scale statement in Theorem 4.2 would go beyond Demailly–Paun's results. However, the paper's key technical steps are not established: the foundational identity (22) is imported from [20], and the proof of the main cone-containment theorem contains a type-comparison gap that appears to invalidate the argument. The later sections repeatedly replace precise local equivalence by 'close to' or '≈' arguments, which are not justified for openness or zero-locus statements. Thus the significance is currently conditional on substantial repair.","major_comments":[{"comment":"The proof of the central containment K^{∇1,1}_{t0,t} ⊂ K_t derives (52) by 'comparison of types' between the ω-harmonic representative (49) and its expansion in the X_t-coframe (50)–(51). This comparison is invalid: the forms eη_ω^(0) and eη_ω^(2) are harmonic with respect to ω on the central fiber, but in the X_t-coframe (dz^i + φ dz̄^i) ∧ (dz̄^j + φ̄ dz^j) they acquire nontrivial (1,1)-components such as φ⌟eη_ω^(0). The displayed (1,1)-component of eH^ω is therefore not simply ω; it contains additional terms from α^(0) and its conjugate. No argument shows these terms vanish or are absorbed into g_{ij}(z,t). Consequently (52) is an assumption, not a consequence, and the subsequent positivity argument does not prove that H(σ,t) is a positive (1,1)-form on X_t. This gap undermines Theorem 0.3(1), Proposition 0.2, and Theorem 4.2.","section":"Theorem 3.4, equations (49)–(52)"},{"comment":"The equivalence of the period-matrix sections (14) and the Beltrami-defined sections (21) is stated as Theorem 1.2 with 'by comparing constant terms', but no proof is given; the identity (22) is imported from the companion preprint [20]. This identity is used throughout: it defines the quasi-period maps in (41), (56), (74), underlies the Hodge map equations (31)–(32), the Hodge locus formula (87), and the variational Hodge criterion. If (22) fails, the period-matrix blocks no longer describe harmonic projections of iφ^k(I+T iφ)^{-1} eη, and all subsequent statements lose their foundation. The paper must either prove (22) or state it with a precise theorem and proof in [20] that is accessible to the reader.","section":"Theorem 1.2, identity (22)"},{"comment":"The proof reduces the openness of the Hodge map to the openness of α^(0)(α^0_(1),·): B→H^{0,2}, asserting that because Φ^{0,2} − Φ^{0,1}Φ^{1,2} = o(Φ^{1,2}), the openness of the former is 'equivalent' to the openness of α^0_(1)Φ^{1,2}. This equivalence is not justified: openness of a map is not invariant under addition of a term that is o of the leading term, without uniform control. Similarly, the step replacing H(iφ(·)(I+T iφ(·))^{-1}ω0) by H(iφ(·)ω0) uses only that the operator norm of (I+T iφ)^{-1} − I tends to 0; small perturbations do not in general preserve openness. The same '≈ implies equivalence' pattern recurs in Theorem 6.3 and Theorem 7.5, where it is load-bearing for the Hodge-locus identification.","section":"Theorem 5.3, proof around (69)"},{"comment":"The implication (91) is derived from Theorem 7.2 and the statement that H(iφ(t)(I+T iφ(t))^{-1}eσ_Z) 'is close to' H(iφ(t)eσ_Z) for small t, so the vanishing of one is equivalent to the vanishing of the other. This is not valid: closeness of functions does not imply equality of their zero loci. The proof needs an exact identity or a precise argument that the zero set is unchanged under the operator (I+T iφ)^{-1}. Without it, the necessary-and-sufficient criterion for the variational Hodge conjecture is unproven. The citation to [6] for the obstruction term H_{N_{Z|X}}(φ(t)|_{N_{Z|X}}) is also vague; the precise definition and deformation-theoretic statement should be included.","section":"Theorem 7.5, proof of (91)"}],"minor_comments":[{"comment":"The notation K^{∇1,1}_{t0,t} is defined only in Definition 0.1 after being used in the abstract; please reorder or add a forward reference.","section":"Throughout"},{"comment":"The constant c0 = min(c1,c2) is not explicit; it depends on the choice of finite cover (53) and on the implicit function theorem radius. The paper should clarify whether c0 is uniform in the initial Kähler form or only in the Beltrami differential.","section":"Section 4, Theorem 4.2"},{"comment":"Reference [24] is listed as 'Rao, Wan, and Zhao' with an apparent typo in the title; please check 'Nagoya Mathematical Journal, 246'. Also [20] is a companion preprint and should include a precise statement of the results used here.","section":"References"},{"comment":"The supremum norm ∥φ∥_E is defined via local charts but the maximum over the cover requires a choice of refinements; the dependence on this choice should be stated explicitly, even if the norm is equivalent to ∥φ∥_ω.","section":"Equation (18)"}],"recommendation":"reject","confidential_remarks":"The paper has a promising main idea and the implicit-function construction of the Hodge map in Theorem 2.4 is clean, but the central containment theorem rests on an invalid type-comparison step and an unproved imported identity from the authors' companion preprint. The later sections rely on repeated 'close to' reductions that do not preserve openness or zero loci. These are load-bearing and not local presentation issues. I recommend rejection, although a substantially revised version with a correct proof of (52) and a complete treatment of the imported identity could be reconsidered."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is worth knowing about even though its main theorem does not hold as written. It constructs an explicit Hodge map for deforming (p,p)-classes via Beltrami differentials and period-matrix blocks, and uses it to define ∇^{1,1}-flat extensions of Kähler cones with explicit positive representatives. That is a genuinely new and potentially useful handle on Kähler cone variation. The proof of the Hodge map (Theorem 2.4) is a clean implicit-function argument, and the idea of controlling nearby complex structures entirely from the central fiber's Hodge theory is attractive.\n\nThe problem is in the central containment statement, Theorem 0.3(1), proved in Theorem 3.4. Equation (52) does not follow from the preceding type comparison. The harmonic representative eH^ω(σ,t) is a sum of the base-harmonic forms eη^ω_(0), ω, eη^ω_(2). Written in the X_t-coframe, eη^ω_(0) and eη^ω_(2) are not pure (2,0) and (0,2); they acquire (1,1) components via contraction with φ. So the (1,1)-component of eH^ω on X_t is not ω, and the matrix g(t) whose positivity is proven need not be the X_t-(1,1)-part of the class. The gap propagates to Proposition 0.2 and Theorem 4.2.\n\nThe reader's other complaint is valid too: the foundational identity (22) is imported from the companion preprint [20] without proof. And the criteria in Sections 5–7 replace exact equivalences with first-order or approximate ones, so those applications are conditional.\n\nThe paper is not incoherent, and the Hodge map may well be salvageable. But as written, the central claims are not established. I would send it to a serious referee: the idea is substantive, the mistakes are identifiable, and a revision could make it a useful contribution. I would not cite it in its current form. For a reading group, it could be a good case study in how type-comparison arguments can go wrong, but it is not a stable reference.","headline":"Fresh and promising framework for deforming (p,p)-classes, but the main cone-containment proof rests on a type-comparison that doesn't survive inspection.","tokens_in":36291,"tokens_out":6085,"would_cite":false,"duration_ms":50814,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14D07","32G05","32Q15","14C30"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that the deformation of (p,p)-classes and of Kähler cones in a family of compact Kähler manifolds is fully controlled by the Beltrami differential on the central fiber through explicit Hodge-bundle sections, yielding upper","keywords":["Beltrami differential","Hodge bundles","Kähler cone","(p,p)-classes","deformation theory","Hodge locus","algebraic approximation","variational Hodge conjecture"],"falsifier":"For a concrete family with computable periods, such as a two-parameter deformation of a complex torus, pick a harmonic (n-p,p)-form η and compute the period block Φ^{(p,p+1)}(t) by classical period theory; then compare its linear coefficient at t = 0 with H(i_{φ_1} η), where φ_1 = Σ θ_i t_i is the first-order Beltrami term. Any mismatch falsifies identity (22), and with it the Hodge map and cone-extension claims.","tokens_in":35217,"feed_emoji":"🔷","tokens_out":6817,"duration_ms":58968,"temperature":0.7,"pith_summary":"The paper aims to show that, for any family of compact Kähler manifolds over an analytic base, the Kähler cones of nearby fibers can be reconstructed from the central fiber alone: one adds to each Kähler class on the central fiber explicit correction terms built from harmonic projections and the Beltrami differential of the deformation. This produces a Hodge map whose values are (p,p)-classes on nearby fibers, and ∇^{1,1}-flat extensions of the Kähler cone that are always contained in the true Kähler cone, equaling it except on a countable union of analytic loci where analytic cycles fail to extend. If correct, the construction gives a uniform mechanism for Kähler stability, valid for obstructed deformations and over large regions of the base, for algebraic approximation criteria for (p,p)-classes, and for an intrinsic description of Hodge loci. The paper matters because it replaces first-order, unobstructed reasoning with explicit all-orders formulas in the Beltrami differential.","feed_headline":"Beltrami data moves Kähler cones across deformations","feed_subtitle":"A central-fiber Hodge map gives explicit nearby Kähler classes except on a rare locus where cycles fail to deform.","key_machinery":"The central object is the Beltrami differential φ(t) of the deformation—an (0,1)-form with values in the holomorphic tangent bundle that records how the complex structure twists—together with the contraction exponential e^{i_φ} = Σ (1/k!) i_φ^k and the harmonic-theoretic operator (I+T i_φ)^{-1} built from T = ∂*G∂ on the central fiber. The key identity (22) equates the period-matrix blocks Φ^{(p,p+k)}(t) with the harmonic projections of (1/k!) i_φ^k (I+T i_φ)^{-1} eη^{(p)}; this identity is what turns abstract period variation into explicit sections of Hodge bundles and makes the Hodge map computable from central-fiber data alone.","core_discovery":"The central claim is that the deformation of (p,p)-classes along a family of compact Kähler manifolds is governed by explicit sections of Hodge bundles of the form H(e^{i_φ(t)}(I+T i_φ(t))^{-1}eη), where φ(t) is the Beltrami differential realizing the nearby complex structures, T = ∂*G∂ is the Green-operator contraction, and H is harmonic projection on the central fiber. These sections coincide with the period-matrix blocks and give, via an implicit-function argument, a real-analytic Hodge map H(σ,t) that sends any Kähler class σ on the central fiber to a (1,1)-class on X_t with a positive definite representative. From this the authors derive upper semicontinuity of Kähler cones, equality of","pith_inferences":["If the foundational identity withstands scrutiny, the framework converts period-map computations on nearby fibers into central-fiber harmonic analysis, making Kähler and Hodge-theoretic questions potentially accessible to explicit computation in examples such as tori or Calabi-Yau families where Beltrami differentials can be written down.","The all-orders formulation suggests the classical obstruction to algebraic approximation for (p,p)-classes with p ≥ 2—failure of the first-order density criterion—may be overcome by higher-order terms; a natural next step is to find concrete classes satisfying the new openness condition but not the old one.","The large-scale stability result hints at a new route to global Kähler rigidity: families whose Beltrami pseudo-distance remains below the threshold c0 can be shown Kähler without elliptic-operator regularity arguments, potentially yielding new proofs that degenerate central fibers in such families are Kähler.","The Hodge-locus formula may make the variational Hodge conjecture computationally approachable: one only needs to compare the vanishing locus of the Beltrami normal-bundle obstruction with the vanishing locus of H(i_φ eσ_Z)."],"forward_implications":["Kähler cones are upper semicontinuous under parallel transport of (1,1)-classes, with explicit positive representatives for every class in the extension.","Away from a countable union of analytic subsets—where some analytic cycle fails to deform—the ∇^{1,1}-flat extension equals the entire Kähler cone of the nearby fiber.","All nearby fibers remain Kähler, and the extension exists, on the whole region of the base where the Beltrami differential has operator norm below an explicit constant, with no unobstructedness assumption.","Strong algebraic approximation follows from an openness condition on a single higher-order Beltrami map; for (p,p)-classes, the full chain of contractions with φ, not just the first-order term, is the right criterion.","The Hodge locus of a rational (p,p)-class is described intrinsically by the vanishing of H(i_φ(t)(I+T i_φ(t))^{-1}eσ), and the variational Hodge conjecture for a smooth subvariety is equivalent to a normal-bundle obstruction statement."],"fun_headline_variants":["Beltrami differentials steer Kähler cones across families","Explicit Hodge sections track Kähler class deformations","Period map and Beltrami data describe Kähler cones","Kähler cones deform via Beltrami-driven Hodge maps","Moving Kähler cones with Beltrami differentials"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is identity (22), imported without proof from the authors' companion preprint: the period-matrix blocks equal the harmonic projections of i_φ^k (I+T i_φ)^{-1} applied to harmonic representatives, and if this equality fails, the Hodge map, the Kähler-cone extensions, and every subsequent application lose their foundation.","fun_headline_variants_meta":{"raw":{"variants":["Beltrami differentials steer Kähler cones across families","Explicit Hodge sections track Kähler class deformations","Period map and Beltrami data describe Kähler cones","Kähler cones deform via Beltrami-driven Hodge maps","Moving Kähler cones with Beltrami differentials"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000531,"raw_usage":{"total_tokens":2418,"prompt_tokens":795,"completion_tokens":1623,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":539,"completion_tokens_details":{"reasoning_tokens":1540}},"tokens_in":539,"tokens_out":1623,"duration_ms":11379,"temperature":1.0,"reasoning_tokens":1540,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T23:20:39.426228+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a concrete family with computable periods, such as a two-parameter deformation of a complex torus, pick a harmonic (n-p,p)-form η and compute the period block Φ^{(p,p+1)}(t) by classical period theory; then compare its linear coefficient at t = 0 with H(i_{φ_1} η), where φ_1 = Σ θ_i t_i is the first-order Beltrami term. Any mismatch falsifies identity (22), and with it the Hodge map and cone-extension claims.","supporting_citations":[],"review_version":1}