{"id":"129f66ea-6f47-4e3e-8869-83982e310a49","arxiv_id":"1908.05027","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For elliptic surfaces with two multiple fibers of multiplicities 2 and m at least 3 and with chi equal to 1, every singular point of the moduli space M(c2) is canonical, and for compact M(c2) the Kodaira dimension is (dim M(c2)+1)/2.","lead":"This paper studies the singularities of moduli spaces of rank-two stable sheaves on elliptic surfaces with Kodaira dimension one, and shows that under certain conditions the singularities are canonical. In a special two-multiple-fiber case it computes the Kodaira dimension of the moduli space, settling a case left open by Friedman.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 5.11, the key local estimate for Theorem 6.1, leaves two load-bearing cases to the reader; without completing them the canonical-singularity and Kodaira-dimension conclusions are not proven.","rationale":"The paper's central claim is a chain: rank bound for H^1(ad(f)) → canonical singularities → Kodaira dimension. The weakest point is the rank bound, because its proof delegates two case checks to the reader. This is a genuine missing proof, not a disagreement with the literature. The surrounding material is plausible: Fact 5.14 is standard, and Claims 5.12-5.15 are detailed; the omitted cases may well be fillable by symmetry. But in a theorem whose conclusion is a numerical inequality, a single unverified case can shift the bound. I also note a likely typo in the proof of Theorem 7.1(i): after Lemma 7.3 shows Ēη decomposes, the text says 'E corresponds to Case II', whereas Fact 2.11 says Case II has Ēη indecomposable; this appears to be a mislabeling, not the main obstruction. The conditional verdict is appropriate: the argument should be accepted only after the missing cases are supplied.","tokens_in":45471,"tokens_out":20903,"duration_ms":214366,"concrete_test":"Provide a complete proof of the two omitted cases in Proposition 5.11. Concretely, for a component B0=nF with L|F=O(-sF), 0 X. The main theorems claim: in Case I, every singular point is canonical when the number of multiple fibers is small relative to d (Theorems 1.3 and 6.1); in the special case d = 1 with two multiple fibers of multiplicities 2 and m >= 3, all singular points are in Case I and hence canonical, and the Kodaira dimension is (dim M(c2) + 1)/2 (Theorem 7.1); and in Case II, for d = 1 and two multiple fibers, every locally free obstructed sheaf is a hypersurface singularity with R >= 1, with existence results for R = 1 (Theorem 1.4). The strategy combines Laudal's presentation of the completion ring, Friedman's morphism for K-dimension, and detailed local analysis of multiple fibers.","tokens_in":1686,"tokens_out":1933,"duration_ms":61220,"significance":"The paper addresses a genuinely difficult problem: the birational type of moduli of stable sheaves on elliptic surfaces with c1(E) = 0, where the restriction to the generic fiber is only strictly semistable. The proposed route, via canonical singularities and Friedman's map, is coherent and promising. If the missing arguments are completed, the results would be a substantial advance: they would give the Kodaira dimension of these moduli spaces for a nonempty class of elliptic surfaces and describe the Iitaka fibration in moduli-theoretic terms. The paper also contains a useful ring-theoretic criterion (Theorem 4.1) and a careful reduction of the singular-point question to the estimate of H^1(ad(f)). However, two written proof gaps, explicitly acknowledged in the text, sit at load-bearing points: Proposition 5.11 leaves two necessary cases to the reader, and Theorem 7.1(i) asserts a key divisor equality by analogy to Lemma 5.8 without proof. These gaps prevent the canonical-singularity and Kodaira-dimension conclusions from being considered established as written.","major_comments":[{"comment":"The proof of Proposition 5.11 is incomplete. After the treatment of the branches where the first maps in (5.19) and (5.22) are nonzero, the text states: 'It is left to the reader to certify this proposition in remaining cases; the case where the second map at (5.19) is not zero, and the case where the second map at (5.22) is not zero.' The proposition is not a harmless technicality: its inequality (5.16) is used in (5.15) and Proposition 5.16, which in turn produces the final bound (6.13) in Theorem 6.1. Every nonzero f must satisfy the inequality, and the two omitted branches are exactly the remaining alternatives for the L∨-twisted cohomology. If either omitted branch can contribute more than 2, the margin in (6.13) disappears and the rank bound of Theorem 6.1, and consequently the canonical-singularity conclusion, is not justified. The remaining cases must be proved explicitly, or Proposition 5.11 must be stated as an assumption.","section":"Section 5, Proposition 5.11"},{"comment":"The proof asserts 'Then one can show that 2D + K_X - B = 0 in a similar way to Lemma 5.8' without giving the derivation. This equality is load-bearing: it is used to obtain (7.3) and (7.4), to analyze the multiplicities via Fact 5.14, and to derive the contradiction. Lemma 5.8 is nontrivial; it relies on the diagram (5.8), the Z/2-equivariant isomorphism of Lemma 5.6, and descent theory. The analogous statement in Theorem 7.1, where E is not assumed to be Case I, cannot be taken as automatic. A separate proof or a complete reduction to Lemma 5.8 is required. In addition, the final assertion after (7.7) that 'this never occur' is too terse; since it supplies the contradiction that forces E into Case I, it must be written out. Theorem 7.1(i) is essential for Theorem 7.1(ii), because the Kodaira-dimension formula uses the fact that all singular points are canonical.","section":"Section 7, proof of Theorem 7.1(i), after eq. (7.2)"}],"minor_comments":[{"comment":"The reduction to h^1(R(K_X)/G) - h^1(R(K_X)/G ⊗ L∨) = h^0(gr^{HN}_0) - h^0(gr^{HN}_0 ⊗ L∨) is justified by 'standard arguments'; this should be expanded, since the Harder-Narasimhan and Jordan-Hölder filtrations on the non-reduced scheme B_0 are delicate.","section":"Section 5, eq. (5.18)"},{"comment":"The quantities Λ(B), Λ_2(B), Λ_3(B), Λ_4(B) are introduced in the proof of Theorem 6.1 but are defined in different places; it would help the reader to collect all four definitions before (6.13).","section":"Section 6, eq. (6.13)"},{"comment":"The statement 'If M(c2) is compact (for example, c2 is odd)' would benefit from a precise reference or argument; compactness of the moduli scheme is a nontrivial condition, and the example given is only a parenthetical.","section":"Section 7, proof of Theorem 7.1(ii)"},{"comment":"The heading '9.4 (16/11/3, 12/3)' appears to contain a date or note that is not appropriate for a published lemma; please remove it or move it to an acknowledgments note.","section":"Section 9, Lemma 9.4"},{"comment":"In the line defining α', the expression 'sd-4+Λ' appears to be a typo for '2d-4+Λ'; please correct and re-read the displayed formula for α'.","section":"Section 8, proof of Proposition 8.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the overall strategy is well-motivated. My main concern is the explicit 'left to the reader' branch in Proposition 5.11, which is central to the rank estimate, and the unproved analogy in Theorem 7.1(i). These are not cosmetic; they control the canonical-singularity and Kodaira-dimension conclusions. I would be willing to look at a revised version that supplies complete proofs of those two points."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Your instinct is right: this is a genuinely new result on an open case. The Kodaira dimension formula for c1=0, kappa(X)=1 has only an upper bound from Friedman; Yamada proves the exact formula in the two-multiple-fiber case with m1=2, d=1, and supplies a clean route—rank of H^1(ad(f)) >= 2 ext^2+1 implies canonical singularities via Theorem 4.1, and then kappa(M) equals the K-dimension via Corollary 3.5. The ring-theoretic Theorem 4.1 is self-contained and looks correct; the deformation-theory translation in Theorem 4.7 is standard. Section 3's comparison of Friedman's map with the pluricanonical map is careful and is a genuine conceptual contribution.\n\nWhere I part from the reader's softer tone: Proposition 5.11 is load-bearing, not merely incomplete in a cosmetic sense. The text explicitly strands two cases ('It is left to the reader...')—exactly the alternatives for the L^vee-twisted cohomology. Inequality (5.16) enters (5.12), then (6.13), and any missing contribution could break the rank bound. The stress-test is right that these are the two branches where the first map is nonzero. I'd also flag Theorem 7.1(i): 'one can show that 2D+K_X-B=0 in a similar way to Lemma 5.8' is asserted, not derived. That equality is what lets the divisorial arithmetic in (7.4) get started. If either gap hides an error, the Kodaira dimension formula falls.\n\nOn the other hand, I do not see circularity or invented entities. The singularity criterion is proven from local algebra; Friedman's map is used only after canonical singularities are established. The dependence on the author's earlier Enriques work is for standard comparisons, and the new assertions are either stated precisely or flagged as analogous. Citations look appropriate.\n\nProportionate verdict: the paper deserves a serious referee, but the referee should send it back for a completed proof of Proposition 5.11 and a written derivation of the 2D+K_X-B=0 step. If those are supplied, this is a solid advance in the moduli-of-sheaves program. If they fail, the main theorem is unproven.\n\nWho this is for: people working on moduli of sheaves on surfaces, especially elliptic surfaces; birational geometers interested in Kodaira dimension of moduli; singularity theorists interested in the rank criterion. I'd bring it to reading group if someone completes the gaps; as is, maybe. I'd cite it once the gaps are filled.\n\nRecommendation: accept for peer review with major revisions, not desk reject. The claims are concrete, the strategy promising, and the paper is honest about what is proved vs. left.","headline":"A real new result on an open case, but the proof has two load-bearing gaps: Proposition 5.11 explicitly strands two cases, and Theorem 7.1(i) asserts an unproved equality.","tokens_in":46263,"tokens_out":3000,"would_cite":false,"duration_ms":27239,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14J60","14D20","32G13","14B05","14Exx"],"pacs":[],"model":"deepseek-v4-flash","headline":"For elliptic surfaces with a few multiple fibers, the moduli scheme M(c2) has only canonical singularities, and its Kodaira dimension is (dim M+1)/2.","keywords":["moduli of stable sheaves","elliptic surfaces","canonical singularities","Kodaira dimension","obstructed sheaves","multiple fibers","rank-two vector bundles"],"falsifier":"Compute the rank of $H^{1}$(ad(f)) for an explicit obstructed Case-I sheaf produced by Proposition 8.1 on a surface satisfying d+2 at least (7/4)Lambda(X); if some nonzero f gives rank below 2 $ext^{2}$(E,E)^0+1, then Theorem 6.1 is false. Concretely, for d=1 with two multiple fibers of multiplicities 2 and 3, take the sheaf E from Claim 8.6, write its single defining equation F=$t1^{2}$+...+$tR^{2}$+O(3), and check whether R is at least 3; R=2 would give a non-canonical hypersurface singularity and refute the theorem.","tokens_in":45239,"feed_emoji":"📐","tokens_out":6027,"duration_ms":62131,"temperature":0.7,"pith_summary":"The paper studies the moduli scheme M(c2) of rank-two stable sheaves with c1=0 on a minimal elliptic surface X of Kodaira dimension one, and asks when its singularities are canonical and what this implies for the birational classification. Its central result is that for obstructed sheaves whose generic-fiber restriction is stable (Case I), every nonzero linear combination of the defining equations has quadratic part of rank at least 2b+1 whenever X has few multiple fibers, which by a purely ring-theoretic criterion makes the point a canonical singularity. For surfaces with just two multiple fibers, multiplicities 2 and m at least 3, and chi(O_X)=1, the paper concludes every singular point is canonical, and when M(c2) is compact with c2 at least 3 its Kodaira dimension is exactly (dim M +1)/2, with the Iitaka fibration described by Friedman's moduli-theoretic map. It also shows that for Case II obstructions the quadratic part alone cannot decide canonicity, since rank-one quadratic parts occur.","feed_headline":"Canonical singularities pin down Kodaira dimension of sheaf moduli","feed_subtitle":"For elliptic surfaces with two multiple fibers, every singular point is canonical, giving kappa(M)=(dim M+1)/2.","key_machinery":"The central object is the map $H^{1}$(ad(f)): $Ext^{1}$(E,E) -> $Ext^{1}$(E,E(K_X)), defined by $\\alpha$ maps to f∘$\\alpha$ - $\\alpha$∘f for a nonzero traceless homomorphism f: E -> E(K_X). The quadratic part of the i-th defining equation F_i is the bilinear form F_fi($\\alpha$⊗$\\beta$) = tr(f_i∘$\\alpha$∘$\\beta$ + f_i∘$\\beta$∘$\\alpha$), so bounding the rank of every nonzero linear combination of these forms is equivalent to bounding the rank of $H^{1}$(ad(f)). The paper estimates this rank by factoring det(f) through a double cover Y0 -> X, introducing the divisor B recording where f factors through $E^{{∨∨}}$(K_X-B), and expressing the kernel and image of ad(f) as extensions of line bundles twisted by zero-dimensional subschemes. The decisive estimate, Proposition 5.11, bounds $h^{1}$(R(K_X)/G) - $h^{1}$(R(K_X)/G ⊗ L^∨) by 2Λ(B) using the local structure of sheaves on multiple fibers and the torsion order of O(F)|F, and this feeds into the rank bound that triggers the ring-theoretic canonical-singularity criterion.","core_discovery":"The load-bearing discovery is a rank bound on the quadratic part of the equations defining M(c2) at an obstructed Case-I sheaf E. Writing the completion as C[[t1,...,t_{D+b}]]/(F1,...,F_b), the paper proves that any nonzero linear combination G of the F_i has quadratic part of rank R at least 2b+1, provided 7(d+2)/4 is at least Lambda(X) or 2 is at least Lambda(X). This rank bound is exactly the hypothesis of the purely ring-theoretic Theorem 4.1, so E is a canonical singularity of M(c2). Applied to surfaces with two multiple fibers of multiplicities 2 and m at least 3 and with d=1, the paper proves that every singular point of M(c2) is of Case I and is canonical; for compact M(c2) with c2 at least 3, the Kodaira dimension equals (dim M(c2)+1)/2.","pith_inferences":["One consequence not drawn in the paper is that the method suggests canonicity of moduli depends mainly on the count of multiple fibers rather than on the detailed Weierstrass coefficients, so families of elliptic surfaces with fixed d and Lambda should have uniform Kodaira dimension of M(c2).","The Case-II examples with R=1 indicate that the analytic type of M at an obstructed sheaf is not determined by the infinitesimal cup product alone; higher-degree terms of the defining equations matter, so one could test this by comparing the singularity type with the prediction from the quadratic part.","The theorem that all singular points are Case I when the two multiplicities are 2 and m at least 3 suggests a deformation-theoretic obstruction: Case II sheaves require the multiplicities to satisfy stronger inequalities, and one could look for a direct cohomological reason for this exclusion.","If the rank bound extends beyond rank two, similar canonical-singularity statements might hold for moduli of higher-rank stable sheaves on elliptic surfaces whenever the generic-fiber restriction is stable, though the paper only treats rank two."],"forward_implications":["For an elliptic surface with two multiple fibers of multiplicities 2 and m at least 3 and with chi(O_X)=1, every singular point of a compact moduli scheme M(c2) is canonical; hence the Kodaira dimension of M(c2) is (dim M(c2)+1)/2 for c2 at least 3.","The Iitaka fibration of M(c2) is the Stein factorization of Friedman's morphism psi, and its general fibers are Jacobians of hyperelliptic curves, so the birational structure of the moduli scheme is described purely in moduli-theoretic terms.","For Case-I obstructed sheaves on elliptic surfaces with few multiple fibers, the moduli scheme is locally of complete intersection, normal at the obstructed point, and the singularity is canonical, so K-dimension and Kodaira dimension coincide.","Obstructed Case-I sheaves actually exist for c2 sufficiently large whenever the invariants satisfy 2d at least max(Lambda-2, 4-Lambda, 5-2Lambda), so the canonical-singularity conclusion is not vacuous.","For Case-II sheaves, the rank R of the quadratic part of the defining equation is at least 1, and examples with R=1 occur, so the degree-two part of the defining equations is insufficient to decide canonicity."],"supporting_citations":[{"why":"Supplies the presentation of the completion ring of moduli at E as C[[t1,...,t_{D+b}]]/(F1,...,F_b), the starting point for the rank analysis.","marker":"[27]"},{"why":"Constructs the coarse moduli scheme M(c2) as a projective scheme parameterizing H-stable sheaves.","marker":"[13]"},{"why":"Provides the c2-suitable polarization, the three-case classification of generic-fiber restrictions, and Friedman's morphism psi used for the Kodaira-dimension computation.","marker":"[9]"},{"why":"Supplies the torsion-order fact for O(F)|F on a multiple fiber and the count of singular fibers, used in the local estimates at multiple fibers.","marker":"[11]"},{"why":"Provides the canonical bundle formula and the logarithmic-transformation description of elliptic surfaces with multiple fibers.","marker":"[4]"},{"why":"Supplies the line-bundle formalism for moduli of sheaves, the Serre correspondence, and stability criteria used in the constructions.","marker":"[21]"},{"why":"Gives the deformation-invariance of rational singularities used in the proof of the ring-theoretic canonical-singularity criterion Theorem 4.1.","marker":"[8]"},{"why":"Bounds the size of the obstructed locus, used in proving normality and irreducibility of M(c2) for large c2.","marker":"[28]"},{"why":"Describes the relative Picard scheme and the moduli-theoretic maps on elliptic surfaces, underpinning the description of the Iitaka fibration.","marker":"[10]"}],"fun_headline_variants":["Rank bound proves canonical singularities on elliptic moduli","Kodaira dimension computed for stable sheaf moduli on elliptic surfaces","Canonical singularities: key to sheaf moduli dimension","Two multiple fibers give canonical singularities and kappa formula"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof rests on fine-grained local estimates at multiple fibers (Claims 5.12 to 5.15, Fact 5.14 on the torsion order of O(F)|F) and on the asserted equality 2D+K_X-B=0 in Theorem 7.1(i); if any of these local computations fails, the rank bound and hence the canonical-singularity and Kodaira-dimension conclusions collapse.","fun_headline_variants_meta":{"raw":{"variants":["Rank bound proves canonical singularities on elliptic moduli","Kodaira dimension computed for stable sheaf moduli on elliptic surfaces","Canonical singularities: key to sheaf moduli dimension","Two multiple fibers give canonical singularities and kappa formula"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000219,"raw_usage":{"total_tokens":1492,"prompt_tokens":1040,"completion_tokens":452,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":656,"completion_tokens_details":{"reasoning_tokens":381}},"tokens_in":656,"tokens_out":452,"duration_ms":5021,"temperature":1.0,"reasoning_tokens":381,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:25:50.280384+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the rank of $H^{1}$(ad(f)) for an explicit obstructed Case-I sheaf produced by Proposition 8.1 on a surface satisfying d+2 at least (7/4)Lambda(X); if some nonzero f gives rank below 2 $ext^{2}$(E,E)^0+1, then Theorem 6.1 is false. Concretely, for d=1 with two multiple fibers of multiplicities 2 and 3, take the sheaf E from Claim 8.6, write its single defining equation F=$t1^{2}$+...+$tR^{2}$+O(3), and check whether R is at least 3; R=2 would give a non-canonical hypersurface singularity and refute the theorem.","supporting_citations":[{"cited_title":"Laudal, Matric massey products and formal moduli i","cited_arxiv_id":null,"evidence_quote":"Supplies the presentation of the completion ring of moduli at E as C[[t1,...,t_{D+b}]]/(F1,...,F_b), the starting point for the rank analysis."},{"cited_title":"Gieseker, On the moduli of vector bundles on an algebraic surface , Ann","cited_arxiv_id":null,"evidence_quote":"Constructs the coarse moduli scheme M(c2) as a projective scheme parameterizing H-stable sheaves."},{"cited_title":"Friedman, Rank two vector bundles over regular elliptic surfaces , Invent","cited_arxiv_id":null,"evidence_quote":"Provides the c2-suitable polarization, the three-case classification of generic-fiber restrictions, and Friedman's morphism psi used for the Kodaira-dimension computation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the torsion-order fact for O(F)|F on a multiple fiber and the count of singular fibers, used in the local estimates at multiple fibers."},{"cited_title":"Barth, K","cited_arxiv_id":null,"evidence_quote":"Provides the canonical bundle formula and the logarithmic-transformation description of elliptic surfaces with multiple fibers."},{"cited_title":"Huybrechts and M","cited_arxiv_id":null,"evidence_quote":"Supplies the line-bundle formalism for moduli of sheaves, the Serre correspondence, and stability criteria used in the constructions."},{"cited_title":"Elkik, Singularit´ es rationnelles et d´ eformations, Invent","cited_arxiv_id":null,"evidence_quote":"Gives the deformation-invariance of rational singularities used in the proof of the ring-theoretic canonical-singularity criterion Theorem 4.1."},{"cited_title":"Li, Kodaira dimension of moduli space of vector bundles on surfa ces, Invent","cited_arxiv_id":null,"evidence_quote":"Bounds the size of the obstructed locus, used in proving normality and irreducibility of M(c2) for large c2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the relative Picard scheme and the moduli-theoretic maps on elliptic surfaces, underpinning the description of the Iitaka fibration."}],"review_version":1}