{"id":"fcac7c6a-6141-4dc0-87ce-58f1532ada27","arxiv_id":"2504.19439","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The local perverse filtration on fibers of a Hilbert scheme map is multiplicative on all fibers exactly when the underlying surface fibration is elliptic.","lead":"This paper studies cohomology rings of Hilbert schemes of points on a surface fibered over a curve, focusing on a natural grading called the local perverse filtration. It shows this filtration is multiplicative on every fiber if and only if the original surface fibration is elliptic.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.3 is the load-bearing step: the Kähler/Douady extension of [18, Thm 4.9, 4.17] is asserted, not proved, and Proposition 4.2 requires both directions of it.","rationale":"The central claim is plausible and the global-to-local reduction is natural, but the argument depends at a single point on an unproved analytic extension of previous algebraic results. The reader's weakest-assumption analysis correctly identifies Theorem 3.3 as the load-bearing step. I do not see an internal contradiction in the paper; the issue is missing support for a cited theorem in the new setting. Since the theorem is used in both directions of the main equivalence, a failure would directly invalidate Theorem 4.4, but no evidence of actual failure is present. Therefore the appropriate verdict is CONDITIONAL rather than REJECT or UNVERDICTED: the author should supply a complete proof or a precise reference for Theorem 3.3 in the Kähler/Douady setting. The proposed concrete check would test the key local model and would reveal whether the analytic extension is genuine or hides an algebraic assumption.","tokens_in":12710,"tokens_out":15171,"duration_ms":179434,"concrete_test":"Re-run the proof of Theorem 3.3 for the analytic local model f: S = E x Δ -> Δ with n = 2, where E is an elliptic curve and Δ is a contractible disk. Using Proposition 3.2 and the Li-Qin-Wang ring formula for Douady spaces, write down an explicit filtered basis of H^*(S[2]) and verify by direct cup-product computation that P_1H^1(S[2]) · P_1H^1(S[2]) is contained in P_2H^2(S[2]). Repeat the same computation with E replaced by a genus-2 curve D; the filtration should fail multiplicativity. If either computation requires invoking projectivity of S or an algebraic compactification of E x Δ, then Theorem 3.3 is not an analytic theorem and Proposition 4.2 is unsupported.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 4.4 is proved through Proposition 4.2, which identifies multiplicativity of the local perverse filtration at the diagonal point nx with multiplicativity of the global perverse filtration associated with the restricted fibration f_V: T -> V, and then invokes Theorem 3.3 to conclude that this is equivalent to f_V being an elliptic fibration. Theorem 3.3 is stated for proper holomorphic fibrations of smooth Kähler surfaces and Douady spaces, but its proof is not supplied. The text says only that the arguments of [18, Section 4.3] 'do not depend on whether C is algebraic or not' and that Nakajima and Virasoro operators for Douady spaces satisfy the same relations as for Hilbert schemes, citing [13]. This conceals two nontrivial analytic extensions: (a) the ring-structure theorem of [11] for Hilbert schemes holds verbatim for Douady spaces of arbitrary compact Kähler surfaces, and (b) the non-multiplicativity criterion for non-elliptic fibrations in [18, Theorem 4.9 or 4.17] survives without projectivity of S or C. Proposition 4.2 uses both directions of Theorem 3.3: the direct direction to get ellipticity from local multiplicativity, and the converse for local models to get local multiplicativity from ellipticity. If either analytic extension fails, the equivalence in Theorem 4.4 collapses. In particular, the 'only if' direction (elliptic implies local multiplicative) depends on the converse of Theorem 3.3 for the local model f_V, where V is contractible and H^*(T) -> H^*(F) is an isomorphism; the paper gives no proof that the converse holds in this analytic setting.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies local perverse filtrations on the fibers of the morphism f^{[n]}: S^{[n]} -> C^{(n)} induced by a proper surjective morphism f: S -> C from a smooth Kähler surface to a smooth curve. The main results, Theorems 4.3 and 4.4, assert that the local perverse filtration is multiplicative at a diagonal point of C^{(n)} if and only if f is an elliptic fibration, and is multiplicative at all points if and only if f is elliptic. The proof strategy is to decompose the fibers over a partition into products of fibers of local models, prove a Künneth formula for local perverse filtrations, and then invoke an analytic extension of the author's earlier algebraic multiplicativity criterion for Hilbert schemes of fibered surfaces.","tokens_in":13104,"tokens_out":15574,"duration_ms":158764,"significance":"If correct, the main theorem gives a clean geometric characterization of local multiplicativity of perverse filtrations for Hilbert schemes of fibered surfaces: the local perverse filtration is multiplicative on every fiber precisely when the general fiber has genus one. The paper also provides a useful reduction of Conjecture 1.1 in this class of examples, showing that a single diagonal point controls the global property when the fibration is an elliptic fibration. However, the proof relies on a substantial unproved analytic extension (Theorem 3.3) and on a neighborhood argument (Proposition 2.6) whose proof is not valid as written. Because these points are load-bearing, the main theorem is not yet established in the present form.","major_comments":[{"comment":"Theorem 3.3 asserts that the main criteria of [18, Theorems 4.9 and 4.17] hold verbatim for proper holomorphic fibrations of Kähler surfaces and Douady spaces, but no proof is supplied. The sentence 'It is straightforward to check the arguments [18, Section 4.3] does not depend on whether C is algebraic or not' is insufficient, because the argument requires the ring-structure theorem of [11] for the cohomology of Douady spaces and the Nakajima/Virasoro operator relations, and it is not automatic that these extend from projective Hilbert schemes to arbitrary compact Kähler surfaces. Since Proposition 4.2 invokes both directions of Theorem 3.3, and Theorem 4.4 rests on Proposition 4.2, this gap is load-bearing.","section":"Section 3, Theorem 3.3"},{"comment":"The proof of Proposition 2.6 is not correct as written. The equality (R^d f_* Q_X)_p = H^d(F) only says the stalk of the direct image is the fiber cohomology; it does not imply that for a fixed neighborhood U the restriction map H^d(f^{-1}(U)) -> H^d(F) is an isomorphism. The additional assertion that the maps H^*(U, P_i) -> H^*(p, i^* P_i) are isomorphisms for each perverse summand P_i in (14) is unjustified and fails in general, for instance for perverse sheaves with monodromy on U\\{p}. Because Proposition 4.2 uses Proposition 2.6 to identify the local perverse filtration on F^{[n]}(nx) with the perverse filtration on H^*(T^{[n]}), the main theorem depends on this statement. Please provide a correct proof, for example by using that f^{-1}(U) deformation retracts to F for a proper surface fibration over a curve, or state and prove the precise conditions under which the claimed isomorphism holds.","section":"Section 2.3, Proposition 2.6"},{"comment":"The proof of Theorem 4.3(1) cites [17, Proposition 4.17] for the multiplicativity of the perverse filtration on an arbitrary fiber of a surface fibration, but [17] is written in the algebraic setting. Please supply the Kähler analogue or explain explicitly why the cited result applies to proper holomorphic fibrations of Kähler surfaces. As written, this step inherits the same type of analytic-extension gap as Theorem 3.3.","section":"Section 4, Theorem 4.3(1)"},{"comment":"In the proof of Proposition 4.2, the argument that the adjunction morphism Rf_{V*} Q_T -> p_* p^* Rf_{V*} Q_T induces an isomorphism H^*(T) -> H^*(F) is terse and conflates Proposition 2.6 with proper base change. Please spell out the isomorphism (20) and its compatibility with the perverse decompositions, since the subsequent claim that ι^{[n]*}: H^*(T^{[n]}) -> H^*(F^{[n]}(nx)) is an isomorphism is central to the proof.","section":"Section 4, Proposition 4.2"}],"minor_comments":[{"comment":"The paper contains numerous typos and encoding artifacts, such as 'mulitplicativity' in the title, 'proper holomophic' in the proof of Proposition 4.2, and 'K¨ahler' with an encoding error. These should be corrected before publication.","section":"Throughout"},{"comment":"The phrase 'Douady space of n points on surface S' should be 'on a surface S' for grammatical correctness.","section":"Introduction, abstract"},{"comment":"The displayed diagram contains a garbled arrow artifact 'axisshort/axisshort/arrowaxisright' that should be fixed.","section":"Section 4, proof of Proposition 4.1"},{"comment":"The justification 'follows from [17, Lemma 2.9] by taking the stalk at ny' needs a brief explanation that the symmetric-power construction and perverse filtration arguments from the algebraic setting carry over to Kähler manifolds without change.","section":"Section 2.4, proof of Proposition 2.8(1)"},{"comment":"Reference [15] contains a doubled word 'and and' and should be corrected.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is built on the author's prior work [18] and asserts, without a detailed proof, that the analytic/Douady extension holds. If the editor seeks a quick decision, I would recommend asking for a full proof of Theorem 3.3 before further consideration. The gap in Proposition 2.6 is also serious and affects the main theorem. The paper's main idea is sound and the reduction to local models is elegant, but the current version is not yet publishable because the principal load-bearing steps are either unproved or incorrectly justified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper proves a genuinely nice statement: for a fibration from a Kähler surface to a curve, the local perverse filtration of the induced Douady-space map is multiplicative on every fiber exactly when the fibration is elliptic. The key move is to define local perverse filtrations on individual fibers and show that checking multiplicativity on a single diagonal point of the symmetric product suffices. That is a real refinement of the global criteria in your earlier work, and the product decomposition in Propositions 2.8 and 4.1 makes the reduction clean. The writing is clear, and the logical skeleton is coherent.\n\nThe soft spot is Theorem 3.3. It states that the main criteria from [18] hold in the Kähler analytic setting for Douady spaces, and the justification is essentially one sentence: the arguments do not depend on algebraicity, and Nakajima/Virasoro operators for Douady spaces satisfy the same relations. That conceals two nontrivial steps: transferring the Li-Qin-Wang ring-structure theorem from Hilbert schemes to Douady spaces of arbitrary compact Kähler surfaces, and extending the non-multiplicativity criterion for non-elliptic fibrations beyond the projective setting. Proposition 4.2 uses both directions of Theorem 3.3, so the equivalence in Theorem 4.4 collapses if either extension fails. This is a genuine gap, not a manufactured concern. It may be fixable, and the author may well be right, but as written the paper does not carry that weight.\n\nMinor: Proposition 2.6's neighborhood argument is sketched and invokes a perverse decomposition on an open analytic set without a reference. That is acceptable but could be tightened.\n\nAll that said, the paper is honest on its own terms: it flags the analytic extension as straightforward, which is optimistic but not misleading. No circularity, no invented entities. The connection to Boalch's conjecture gives useful context.\n\nWho is this for? Anyone working on P=W, perverse filtrations, or Hitchin-type morphisms. It deserves serious peer review, but the referee should push for a complete proof or a precise reference for Theorem 3.3. I would not desk-reject it; I would send it out with a request to substantiate that step.","headline":"A clean local criterion for elliptic fibrations via perverse filtrations, but the proof rests on an analytic extension of prior work that is asserted rather than demonstrated.","tokens_in":13578,"tokens_out":2163,"would_cite":true,"duration_ms":23067,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14C05","14D06","14F08"],"pacs":[],"model":"deepseek-v4-flash","headline":"Multiplicative local filtrations occur exactly for elliptic fibrations.","keywords":["perverse filtration","local perverse filtration","multiplicativity","Douady space","Hilbert scheme of points","elliptic fibration","Kähler surface","P=W phenomenon"],"falsifier":"For a non-elliptic example, say a fibration whose general fiber is a genus-2 curve, compute explicitly the local perverse filtration on $F^{[2]}(2p)$ for some $p\\in C$ and check the cup product $P_1H^1 \\times P_1H^1 \\to H^2$; the theorem predicts the image lies outside $P_2H^2$, so finding it inside $P_2H^2$ would refute the claim.","tokens_in":12522,"feed_emoji":"📐","tokens_out":9007,"duration_ms":82118,"temperature":0.7,"pith_summary":"The paper studies the local perverse filtration on the fibers of the map $f^{[n]}: S^{[n]}\\to C^{(n)}$ induced by a proper holomorphic fibration $f:S\\to C$ from a smooth Kähler surface to a smooth curve, where $S^{[n]}$ is the Douady space of length-$n$ zero-dimensional subspaces of $S$. Its central claim is a dichotomy: at a point of $C^{(n)}$ where the $n$ points are distinct, the local perverse filtration is always multiplicative, while at a point where at least two points collide, it is multiplicative if and only if $f$ is an elliptic fibration. Hence the local perverse filtration is multiplicative on every fiber exactly when the general fiber of $f$ has genus 1. The author presents the result as a local, fiberwise analogue of known global multiplicativity criteria, with consequences for the topology of fibers in Higgs-bundle moduli problems.","feed_headline":"Multiplicative local filtrations occur exactly for elliptic fibrations","feed_subtitle":"Douady-space fibers carry multiplicative perverse filtrations precisely when general fibers have genus 1.","key_machinery":"The load-bearing machinery is the local perverse filtration on a fiber, defined as the image of perverse truncation of the direct image $Rf_*\\mathbb{Q}_X$ after restriction to the point, together with the Douady-Barlet decomposition of $R\\pi_*\\mathbb{Q}_{S^{[n]}}$ into shifted symmetric-power pieces. Proposition 4.1 decomposes a fiber over $x=\\sum \\nu_i x_i$ as a product of fibers of smaller Douady maps, compatibly with the filtration by the Künneth formula. Proposition 4.2 then reduces the diagonal case to the central fiber over $n x$, where multiplicativity is equivalent to multiplicativity of the global perverse filtration over a contractible neighborhood; Theorem 3.3, an analytic counterpart of a global criterion for Hilbert schemes of fibered surfaces, converts this into ellipticity of $f$.","core_discovery":"The central discovery is Theorem 4.4: for $n\\ge 2$, the following are equivalent: $f$ is an elliptic fibration; the local perverse filtration at some point $x\\in \\Delta$ is multiplicative; the local perverse filtration at all points of $C^{(n)}$ is multiplicative. The point is that a global geometric condition on $f$ — that its general fibers are genus-1 curves — is detected by the cup-product behavior of the cohomology of a single special fiber of the induced Douady-space morphism. Along the way the paper establishes the decomposition of any such fiber into a product of central fibers over points $n x$, with the local perverse filtration obeying the Künneth formula.","pith_inferences":["If the analytic transfer in Theorem 3.3 holds, the same fiberwise criterion should govern every Douady-space map $S^{[n]}\\to C^{(n)}$: elliptic fibrations are the unique case where cup products respect the local perverse grading at collisions.","Because a single diagonal fiber detects ellipticity, the theorem offers a computational test: checking multiplicativity on one fiber would certify a global geometric property of $f$.","The paper shows local-to-global equivalence only for local models; whether fiberwise local multiplicativity always implies global multiplicativity is a testable next step, and the paper's Conjecture 1.1 predicts it does.","In weight-filtration matching contexts, the result suggests that failure of local multiplicativity at diagonal fibers may be a precise obstruction preventing such an identity."],"forward_implications":["At points of $C^{(n)}$ where the $n$ points are distinct, the local perverse filtration is multiplicative for every fibration $f:S\\to C$, regardless of the fibers.","At a collision point with a repeated point of multiplicity at least 2, local multiplicativity is equivalent to $f$ being an elliptic fibration.","Local multiplicativity at one diagonal point forces local multiplicativity at all points of $C^{(n)}$, and conversely.","If the global perverse filtration associated with $f^{[n]}$ is multiplicative, then all local perverse filtrations are multiplicative; for local models, the converse also holds.","The dichotomy gives a concrete obstruction in Higgs-bundle moduli settings: non-elliptic fibrations must violate local multiplicativity at diagonal points."],"supporting_citations":[{"why":"Proves the global multiplicativity criterion for Hilbert schemes of fibered surfaces that Theorem 3.3 extends to the Kähler Douady setting.","marker":"[18]"},{"why":"Supplies the perverse decompositions, the cohomology ring structure, and the fiberwise multiplicativity results for Hilbert schemes of fibered surfaces that the paper quotes.","marker":"[17]"},{"why":"Gives the canonical decomposition of the Douady-Barlet pushforward used to compute cohomology and filtrations on Douady spaces.","marker":"[4]"},{"why":"Provides the definition of perverse filtration and motivates the multiplicativity question through the weight-filtration matching phenomenon.","marker":"[3]"},{"why":"Describes the cohomology ring of Hilbert schemes of points on surfaces, which underlies the ring arguments carried over to Douady spaces.","marker":"[11]"},{"why":"Shows that the Nakajima operators for Douady spaces satisfy the same relations as for Hilbert schemes, used to identify the cohomology ring structure.","marker":"[13]"}],"fun_headline_variants":["Elliptic fibrations exactly match multiplicative local perverse filtrations","Local perverse multiplicativity: a precise elliptic fibration test","Multiplicative local filtrations characterize elliptic fibrations","Elliptic fibration iff local perverse filtration multiplicative","Genus-one fibers alone give multiplicative local perverse filtrations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the global multiplicativity criterion proved for algebraic Hilbert schemes still holds for Kähler Douady spaces; the paper states this transfer is straightforward but supplies only a brief justification, and the 'only if' direction depends on it.","fun_headline_variants_meta":{"raw":{"variants":["Elliptic fibrations exactly match multiplicative local perverse filtrations","Local perverse multiplicativity: a precise elliptic fibration test","Multiplicative local filtrations characterize elliptic fibrations","Elliptic fibration iff local perverse filtration multiplicative","Genus-one fibers alone give multiplicative local perverse filtrations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000749,"raw_usage":{"total_tokens":3229,"prompt_tokens":729,"completion_tokens":2500,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":345,"completion_tokens_details":{"reasoning_tokens":2419}},"tokens_in":345,"tokens_out":2500,"duration_ms":19395,"temperature":1.0,"reasoning_tokens":2419,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:52:44.913376+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a non-elliptic example, say a fibration whose general fiber is a genus-2 curve, compute explicitly the local perverse filtration on $F^{[2]}(2p)$ for some $p\\in C$ and check the cup product $P_1H^1 \\times P_1H^1 \\to H^2$; the theorem predicts the image lies outside $P_2H^2$, so finding it inside $P_2H^2$ would refute the claim.","supporting_citations":[{"cited_title":"Zhang, Multiplicativity of perverse filtration for Hilbert scheme s of fibered surfaces, II , Trans","cited_arxiv_id":null,"evidence_quote":"Proves the global multiplicativity criterion for Hilbert schemes of fibered surfaces that Theorem 3.3 extends to the Kähler Douady setting."},{"cited_title":"Zhang, Multiplicativity of perverse filtration for Hilbert scheme s of fibered surfaces , Adv","cited_arxiv_id":null,"evidence_quote":"Supplies the perverse decompositions, the cohomology ring structure, and the fiberwise multiplicativity results for Hilbert schemes of fibered surfaces that the paper quotes."},{"cited_title":"de Cataldo, L","cited_arxiv_id":null,"evidence_quote":"Gives the canonical decomposition of the Douady-Barlet pushforward used to compute cohomology and filtrations on Douady spaces."},{"cited_title":"de Cataldo, T","cited_arxiv_id":null,"evidence_quote":"Provides the definition of perverse filtration and motivates the multiplicativity question through the weight-filtration matching phenomenon."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the cohomology ring of Hilbert schemes of points on surfaces, which underlies the ring arguments carried over to Douady spaces."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows that the Nakajima operators for Douady spaces satisfy the same relations as for Hilbert schemes, used to identify the cohomology ring structure."}],"review_version":1}