{"id":"0749c933-6fa2-40e2-9148-dade0c1af796","arxiv_id":"2509.06197","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A symplectic-geometric formula for the Abel-Jacobi differential on Fano threefolds, yielding a new proof of the tangent bundle theorem for the cubic threefold.","lead":"This paper expresses the differential of Abel-Jacobi maps on Fano threefolds through the symplectic geometry of how curves intersect a fixed anti-canonical surface. It uses this to give a short new proof of a classical result on the Fano surface of a cubic threefold.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem A(ii) rests on the unproved commutativity of diagram (3.4); the manuscript verifies only a small square and leaves the map τ and the ∂–Res compatibility unjustified.","rationale":"The reader's weakest assumption targets the non-degeneracy of ψ, but that condition follows from Theorem A(i) once (i) is established. The derivation of (i) via residues appears plausible. The genuinely load-bearing gap is the commutativity of diagram (3.4), which is the mechanism that produces the formula ψ(α(ξ),β(η)) = ⟨dAJ_X(ξ),η⟩. The manuscript's verification is incomplete: it shows a small square and then asserts the result, leaving τ undefined and the ∂–Res compatibility at the level of H^1(Ω^2_X)|_C unjustified. This is a correctness risk because a sign or missing term in the diagram would invalidate the central claim. The paper also honestly defers a full proof of Theorem B, but Theorem A(ii) is the primary contribution and needs a complete diagram chase. I therefore keep the verdict CONDITIONAL, matching the reader's, but on the more specific ground of the unproved commutativity rather than the non-degeneracy.","tokens_in":8832,"tokens_out":9797,"duration_ms":104921,"concrete_test":"Verify diagram (3.4) by computing both compositions explicitly for Example 1 (cubic threefold, lines): fix a line L, take ξ∈H^0(N_{L/X}) and η∈V=H^1(Ω^2_X). Evaluate the RHS of (4.4) as the period integral ∫_L ξ⌋ω_P, and evaluate the LHS by the residue pairing of the varied line intersection with a K3 surface Y. A disagreement beyond an overall normalization constant would disprove the commutativity. For an analytic check, write the C^∞ representatives in local coordinates and use the Cartan formula to confirm [∂,Res] vanishes on the image of the period map; if a boundary term appears, the diagram must be modified.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's core result, Theorem A(ii), is derived by equating the two paths around diagram (3.4) in Section 3. The proof of commutativity is given only for a proper sub-diagram: the displayed square with maps i*, τ, ρ, and the restriction to C. Missing is a justification that the full diagram (3.4) commutes, specifically: (1) the definition of τ, 'induced by considering a class in H^1(Ω^2_X)|_C as a (1,1) form in A^1(Ω^1_C) whose values are a conormal vector field', is not made precise; (2) the compatibility [∂,Res]=0 is invoked in the C^∞ log complex, but the diagram also involves the connecting map ∇Ψ of the period map and the projection to H^1(Ω^2_X)⊂H^1(Ω^2_X(logY)), and the commutativity of that piece is not shown; (3) the identification of the upper path with ψ(α(ξ),β(η)) and the lower path with ⟨∇_ηΨ, AJ_X(ξ)⟩ depends on the duality between the normal and tangent spaces induced by ψ, which is only stated. If the diagram has a sign error or a missing H^1(Ω^1_C)-valued term, the formula in Theorem A(ii) fails. The non-degeneracy of ψ flagged by the reader would follow from Theorem A(i) once (i) is proved, so it is less load-bearing than the commutativity of (3.4).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a Lagrangian interpretation of the differential of Abel-Jacobi maps for families of curves on smooth Fano threefolds. For a curve C_z in a family Z and a smooth anti-canonical divisor Y, the intersection points define a map f: Z -> Y^{(d)}. Theorem A(i) asserts that f(Z) is a maximal-dimension Lagrangian submanifold of the hyperkähler symmetric product Y^{(d)}; Theorem A(ii) asserts the formula psi(alpha(xi), beta(eta)) = <dAJ_X(xi), eta> for xi in T_z Z and eta in H^1(Omega^2_X). The paper applies this to recover the tangent-bundle theorem for the Fano surface of a cubic threefold and to discuss conics on genus-6 Fano threefolds. Theorem B states a Lagrangian statement for the relative family obtained by deforming X while keeping Y fixed; the paper gives only a dimension count for this theorem.","tokens_in":9169,"tokens_out":6949,"duration_ms":89227,"significance":"If Theorem A(ii) is fully established, the paper gives a genuinely new route to classical Abel-Jacobi statements: the differential of the Abel-Jacobi map is encoded in the symplectic geometry of the intersection cycles C_z · Y. The local-coordinate residue proof of Theorem A(i) is a transparent re-verification of the known Lagrangian statement, and the factorization in (4.3)-(4.5) is illuminating. The authors are explicit about the external framework [DM96, Mar08, IM07] and about the assumptions they make; there is no circularity in the main derivation. However, the central formula in Theorem A(ii) rests on a diagram whose commutativity is not actually proved, and Theorem B is explicitly deferred. The paper reads as a research announcement with the main theorem only partially supported, so the significance is real but conditional on completing the proof.","major_comments":[{"comment":"The proof of Theorem A(ii) reduces to the commutativity of diagram (3.4), but the manuscript only proves a proper square involving i_*, rho and tau. It does not justify the definitions/identifications needed for the full diagram: the map tau is only described as 'induced by considering a class in H^1(Omega^2_X)|_C as a (1,1) form in A^1(Omega^1_C) whose values are a conormal vector field', the compatibility of the connecting map of the period map with the residue/Gysin maps is not shown, and the upper and lower paths are asserted to equal psi(alpha(xi),beta(eta)) and <∇_eta Psi, AJ_X(xi)> without a complete diagram chase. Since this is the central claim of the paper, a full proof of the commutativity of (3.4) is required.","section":"§3, diagram (3.4)"},{"comment":"The paper states 'we crucially use the non-degeneracy in both variables of the symplectically induced bilinear form psi: T_{f(z)}f(Z) ⊗ N_{f(Z)/Y^{(d)}, f(z)} -> C' before proving Theorem A(i). This non-degeneracy is used to identify the normal space with the dual of the tangent space and to define beta. It follows from f(Z) being Lagrangian, which is Theorem A(i), but the logical order and the proof itself are not made explicit. The reader needs either a direct proof of this nondegeneracy or a clear statement that it is a corollary of (i) established immediately after that proof.","section":"§2, non-degeneracy of psi"},{"comment":"Theorem B is stated as a theorem in the introduction, but Section 5 only proves the dimension formula (1.6). The text then says 'the remainder of the argument may be done by an adaptation of the arguments in [DM96]' and 'the complete proof of a more general version of Theorem B will be given elsewhere.' This is not a proof of the Lagrangian statement. If Theorem B is to be included as a result, the Lagrangian claim must be proved; otherwise the statement should be labelled as a conjecture or a program.","section":"§5, Theorem B"},{"comment":"The sentence 'With what was said above this implies that for the Fano surface dAJ_X is an isomorphism' is false as written: dim Z = 2 while dim J(X) = 5, so the differential of Z -> J(X) cannot be an isomorphism. What the subsequent argument needs is that the induced map dAJ_X^*: H^1(Omega^2_X) -> T_z^*Z is surjective (and indeed the next sentence uses exactly this). The wording should be corrected, and the distinction between the differential on Z and the induced Albanese map should be made carefully.","section":"§4, Example 1"}],"minor_comments":[{"comment":"The notation Z is used both for the parameter space and for the universal family in (3.1)-(3.4). The distinction between Z and the sheaf/space \\mathcal Z is not clear in the printed text; please define both symbols explicitly.","section":"§3, notation"},{"comment":"The diagram (3.3) contains question marks and an unexplained '?' label. The connecting maps and the inclusion i_* should be labelled, and the exact sequence whose connecting map gives ∇Psi should be written out.","section":"§3, equation (3.3)"},{"comment":"Reference [IP99] is misspelled 'Iskoviskivh'; should be Iskovskikh.","section":"References"},{"comment":"The geometric proof of surjectivity of beta in Example 1 is only sketched. If kept, it should be expanded, especially the claim that the deformations X' obtained from sections P·Q fill out the normal space.","section":"§4, geometric remark"}],"recommendation":"major_revision","confidential_remarks":"The paper is best read as a research announcement with a promising central idea. The main obstacle is not novelty or external soundness but the incomplete proof of the commutativity of (3.4) and the deferred proof of Theorem B. If the authors can supply a complete proof of the diagram and correct the dimension error in Example 1, the paper could be suitable; as it stands, the central theorem is not fully established."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe thing to know: this is a short note that fits the Donagi–Markman/Markushevich/Iliev–Manivel machinery to a Fano threefold, and the genuinely new output is the formula ψ(α(ξ), β(η)) = ⟨dAJ_X(ξ), η⟩ in Theorem A(ii). I think the formula is probably right, but the proof of Theorem A(ii) is not complete as written, and the paper should not be accepted without a rewritten Section 3.\n\nWhat is good: the authors are clear about what is theirs. Theorem A(i) is credited to Iliev–Manivel and re-proved by a short, neat residue computation. The application to the Fano surface of a cubic threefold gives a quick route to the tangent bundle theorem and to the isomorphism J(X)→Alb(Z), using only the Lagrangian statement plus surjectivity of β. The genus-six prime Fano case is a natural bonus. The dimension count for Theorem B is clean, and the authors honestly say the rest of the proof will appear elsewhere.\n\nThe soft spots, in proportion. The main one is diagram (3.4), which is the load-bearing justification of Theorem A(ii). The paper says 'we will show below that this diagram commutes' and then offers a much smaller square. The map τ is described in words, not defined precisely, and the compatibility [∂,Res]=0 in the C∞ log complex with the period-map connecting map is asserted, not shown. There may well be a sign or a missing term; a referee has to check it line by line. The reader's worry about non-degeneracy of ψ turns out to be less serious: once (i) is proved, maximal isotropy gives the non-degeneracy, so it is an ordering issue rather than a missing hypothesis. Unobstructedness of the curve-family deformations is explicitly assumed, which is fine as a standing hypothesis but should be flagged as such. Theorem B is only a sketch, but that is disclosed.\n\nIf the diagram is fixed, this is a solid, useful note for people working on intermediate Jacobians and Fano threefolds. As it stands, it is a worthwhile preprint but not established. I would send it to a serious referee, with instructions to focus on Section 3 and require a complete proof of the diagram's commutativity before acceptance.","headline":"A plausible and useful formula for dAJ_X via symplectic geometry, but Section 3's key diagram commutativity is under-proved; worth refereeing but not ready to cite.","tokens_in":9695,"tokens_out":5897,"would_cite":false,"duration_ms":64158,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14J45","14K30","14C30","14D07"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a Fano threefold, the pairing between tangent and normal directions of the curve family on an anti-canonical K3 surface reproduces the differential of the Abel-Jacobi map.","keywords":["Fano threefolds","Abel-Jacobi mapping","intermediate Jacobian","Lagrangian submanifolds","anti-canonical divisors","K3 surfaces","symmetric products","cubic threefold"],"falsifier":"Take a smooth cubic threefold and a line L_z. The formula predicts ⟨dAJ_X(ν), ω_P⟩ = ψ(α(ν), β(ω_P)) for every P ∈ L_z^⊥; the left side is the explicit integral ∫_{L_z} ν ⌋ ω_P and the right side is computable from the intersection points L_z∩Y and the deformation of L_z with Y fixed. Checking one line for which both sides are non-zero would confirm or refute the identity.","tokens_in":8706,"feed_emoji":"📐","tokens_out":7552,"duration_ms":81150,"temperature":0.7,"pith_summary":"This paper aims to show that the differential of the Abel-Jacobi map from a family of curves on a smooth Fano threefold to its intermediate Jacobian is encoded in the symplectic geometry of the intersections of those curves with any smooth anti-canonical K3 surface. The central formula, Theorem A(ii), states that the symplectic pairing between the tangent direction α(ξ) and a normal direction β(η) in the symmetric product equals the pairing of the Abel-Jacobi differential applied to ξ with the cohomology class η. If true, this gives a new, largely symplectic route to classical results: the tangent bundle theorem for the Fano surface of a cubic threefold and the isomorphism between its Albanese variety and the intermediate Jacobian. The paper also sketches a relative version, Theorem B, in which deformations of the threefold play a role through a Deligne cohomology fibration.","feed_headline":"Intersections with a K3 divisor determine Abel-Jacobi differentials","feed_subtitle":"The result yields new proofs of the tangent bundle theorem and of Alb(Z) ≅ J(X) for the cubic threefold.","key_machinery":"The load-bearing object is the symmetric product Y^{(d)} of a smooth anti-canonical K3 surface Y, together with its holomorphic symplectic form ψ. The map f sends each curve C_z to the d points C_z ∩ Y. The proof of Theorem A(ii) runs through the relative log complexes of the pair (X, Y) and the differential of the period map; the key mechanism is that the residue of a generator Ψ ∈ H^0(Ω^3_X(log Y)) along Y gives ψ, while the obstruction to lifting Ψ to a section over the deformation space lands in H^1(Ω^2_X), exactly the space of first-order deformations η. The formula ψ(α(ξ), β(η)) = ⟨dAJ_X(ξ), η⟩ is then the statement that the symplectic pairing of tangent and normal directions equals th","core_discovery":"Let X be a smooth Fano threefold, Z the parameter space of a family of degree d curves C_z, and Y ∈ |-K_X| a smooth anti-canonical divisor meeting each curve transversely. Intersection induces a morphism f: Z → Y^{(d)}, and Y^{(d)} is hyperkähler with symplectic form ψ. Theorem A asserts that f(Z) is a maximal-dimensional Lagrangian submanifold of Y^{(d)}, and that for every tangent vector ξ of Z and every first-order deformation class η ∈ H^1(Ω^2_X), the identity ψ(α(ξ), β(η)) = ⟨dAJ_X(ξ), η⟩ holds, where α = f_* and β is a naturally defined map from H^1(Ω^2_X) to the normal space of f(Z) in Y^{(d)}. In words: the infinitesimal Abel-Jacobi map is the symplectic pairing between how the inter","pith_inferences":["The identity suggests that for any Fano threefold, the kernel of dAJ_X is characterized by the directions in which β(η) is symplectically orthogonal to the image of α; this could give a purely geometric test for when Abel-Jacobi maps degenerate.","Because the formula only needs a single smooth anti-canonical divisor, it may extend to singular or movable anti-canonical limits if the residue/logarithmic arguments are replaced by a limiting mixed Hodge structure argument.","The non-degeneracy of ψ, assumed rather than proved, may be equivalent to the unobstructedness of the family; if so, examples with obstructed deformations would be natural testing grounds for the formula.","The cubic-threefold argument that β is surjective uses the fact that H^1(Ω^2_X) ≅ V and that global sections of O_X(1) deform the threefold; the same method could be applied to other Fano threefolds with explicit cohomology to identify which curve families yield isomorphisms."],"forward_implications":["For the Fano surface Z of lines on a smooth cubic threefold, dAJ_X is an isomorphism and the induced map Alb(Z) → J(X) is an isomorphism.","The tangent bundle theorem for the cubic threefold — that T_z Z is the restriction of the universal sub-bundle on the Grassmannian — follows directly from the surjectivity of dAJ_X^* together with the pairing formula.","For general conics on prime Fano threefolds of degree 10, the same argument yields that dAJ_X is an isomorphism.","Theorem A gives a symplecto-geometric description of the full infinitesimal Abel-Jacobi map: it is captured by the normal directions β(η) and the symplectic form ψ.","Theorem B predicts, with a proof sketched, that the relative family F(Z) is maximal Lagrangian in Y^{(d)} × D_λ, exactly compensating for the obstruction to deforming curves as X deforms keeping Y fixed."],"supporting_citations":[{"why":"Provides the general Deligne-cohomology/Lagrangian-fibration framework used to define the relative map F and the symplectic form τ.","marker":"[DM96]"},{"why":"Constructs the Lagrangian submanifold f(Z) ⊂ Y^{(d)} for Fano threefolds and supplies the deformation-theoretic context.","marker":"[Mar08]"},{"why":"Gives the maximal-Lagrangian result and the symplectic-form computations used in the cubic threefold example.","marker":"[IM07]"},{"why":"Establishes the tangent bundle theorem and the isomorphism Alb(Z) ≅ J(X) that the paper reproves via Theorem A.","marker":"[CG72]"},{"why":"Supplies the deformation theory of anti-canonical pairs (X,Y) used in the proof of part (i).","marker":"[Bea04]"},{"why":"Provides the normal-bundle statement for conics in prime Fano threefolds used in Example 2.","marker":"[IP99]"},{"why":"Proves that the variety of conics on the degree-10 Fano threefold has a generically globally generated cotangent bundle, used for the β-surjectivity argument.","marker":"[DIM12]"}],"fun_headline_variants":["Abel-Jacobi differentials via symplectic pairing on Fano threefolds","K3 divisors yield Lagrangian proof of Abel-Jacobi maps","Infinitesimal Abel-Jacobi is a symplectic pairing","Symplectic form links Abel-Jacobi to K3 divisor intersections","Lagrangian geometry explains Abel-Jacobi on Fano threefolds"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that the symplectic form on the symmetric product pairs every tangent direction of the curve family non-trivially with some normal direction; without this non-degeneracy the map β is not well-defined and the central identity collapses, and the paper assumes rather than proves it (it is equivalent to f(Z) being maximal isotropic, the content of Theorem A(i)).","fun_headline_variants_meta":{"raw":{"variants":["Abel-Jacobi differentials via symplectic pairing on Fano threefolds","K3 divisors yield Lagrangian proof of Abel-Jacobi maps","Infinitesimal Abel-Jacobi is a symplectic pairing","Symplectic form links Abel-Jacobi to K3 divisor intersections","Lagrangian geometry explains Abel-Jacobi on Fano threefolds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000623,"raw_usage":{"total_tokens":2677,"prompt_tokens":654,"completion_tokens":2023,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":398,"completion_tokens_details":{"reasoning_tokens":1930}},"tokens_in":398,"tokens_out":2023,"duration_ms":15561,"temperature":1.0,"reasoning_tokens":1930,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T23:57:22.807826+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a smooth cubic threefold and a line L_z. The formula predicts ⟨dAJ_X(ν), ω_P⟩ = ψ(α(ν), β(ω_P)) for every P ∈ L_z^⊥; the left side is the explicit integral ∫_{L_z} ν ⌋ ω_P and the right side is computable from the intersection points L_z∩Y and the deformation of L_z with Y fixed. Checking one line for which both sides are non-zero would confirm or refute the identity.","supporting_citations":[],"review_version":1}