{"id":"c7181a2a-6360-4257-bce9-71c729dac886","arxiv_id":"2606.10817","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every H-regular d-dimensional tropical subvariety of a tropical abelian variety has (−1)^d χ(X) ≥ 0, with counterexamples when H-regularity is dropped.","lead":"This paper proves a tropical analogue of the Green–Lazarsfeld theorem: well-behaved tropical subvarieties of tropical tori have Euler characteristics whose sign is forced by dimension. It transfers a Hodge-theoretic sign law into combinatorial topology, shows the law fails without the well-behavedness hypothesis, and resolves a 2008 question of Hacking.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Prop 6.1 is under-specified: the proof omits v∈int(cone(v_i)) and relies on [AP24]; the reader's 'wide cone cannot be contained in ℓ>0' is inaccurate, but the gap is real.","rationale":"The reader identifies Proposition 6.1 as the weakest point, and I agree: it is the combinatorial base of Theorem 2.8, and its proof is the least secure part of the paper. However, the reader's specific internal-consistency objection is not accurate. The proof does not claim σ_S is a thin cone around v; σ_S is the polar of the cone generated by the v_i. If v is placed in the interior of that primal cone, the polar is wide but still satisfies u·v>0 for every nonzero u, so it can be simultaneously wide and contained in {ℓ>0}∪{0}. The genuine flaw is that the proof never states the primal-cone condition; it only says σ_S contains v, which is weaker. This is a missing hypothesis, not a contradiction. The [AP24] dependency is also real but likely replaceable: one can construct a stellar refinement of Δ making σ_S a union of cones using Alexander's theorem on stellar subdivisions of a simplex, so the common-stellar-subdivision theorem is not essential. Because both issues are repairable without altering the main strategy, and because the Morse-theoretic and weight-filtration steps I checked are internally consistent, I do not change the CONDITIONAL verdict. I mark partial agreement: same lemma flagged, but the stated impossibility is not correct.","tokens_in":20411,"tokens_out":35119,"duration_ms":374695,"concrete_test":"Implement a search over rational unit vectors v_i in R^3 for a complete simplicial fan Δ and a generic linear form ℓ, requiring v ∈ int(cone(v_i)) and every ℓ-positive ray of Δ inside int(σ_S). Verify σ_S ⊆ {ℓ>0}∪{0} and explicitly construct the deformation retractions from C(S\\σ_S) onto C(S≤0) and C(Σ≤) on the link complex. Run this for the 6-ray fans (e.g., octahedral fan and 3-cube fan). If no such v_i exists for some fan, Proposition 6.1 fails; if it succeeds, the printed gap is repairable, and the [AP24] dependency can likely be replaced by a direct stellar-subdivision construction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is Proposition 6.1: its cone σ_S and the stellar refinement Δ' with |Δ+|=σ_S underlie the homotopy equivalence L(T≤0)≃K, which is the core of Theorem 2.8 and hence of Theorems 1.1 and 1.4. As printed, the proof chooses unit vectors v_i close to v and sets σ_S = {u : (u,v_i)≥0}. The condition that σ_S contain v in its relative interior is just (v,v_i)>0, automatic for v_i close to v; it does not imply σ_S ⊆ {ℓ>0}∪{0}. If the v_i all lie on one side of v, the polar cone contains rays with ℓ<0. To force σ_S ⊆ {ℓ>0}, one needs v ∈ int(cone(v_i)), i.e., the v_i must surround v. Then σ_S is a wide, nearly half-space cone, but its extreme rays lie strictly inside ℓ>0, so the reader's claim that a wide cone 'cannot' be contained in ℓ>0 is not correct. The proof omits this surrounding condition, so the construction is incomplete as written. Separately, the 'moreover' statement invokes [AP24, Thm 1.1] (common stellar subdivisions of fans), a 2024 preprint resolving a classical conjecture; if that theorem is unverified, the existence of Δ' is unsupported. Both gaps sit at the base of the Hodge-theoretic machinery, not in a peripheral corollary.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a tropical analogue of the Green–Lazarsfeld theorem: for an H-regular d-dimensional tropical subvariety X of a tropical abelian variety, the signed Euler characteristic satisfies (−1)^d χ(X) ≥ 0. The proof is based on a local vanishing theorem (Theorem 2.8) for H-regular tropical fans, proved via mixed-Hodge-module weight vanishing (Prop. 4.1), a comparison between dual complexes and weight-graded cohomology (Prop. 5.1), and a Morse-theoretic argument (Thm. 7.1). The paper also gives a Lefschetz-type theorem for affine H-regular tropical varieties, discusses the necessity of H-regularity through a quotient construction, and constructs a 3-dimensional tropical fan whose link has nontrivial homology in an unexpected degree, addressing a question of Hacking.","tokens_in":20567,"tokens_out":14491,"duration_ms":157073,"significance":"If the main theorem is correct, it establishes a clean combinatorial sign law for a natural class of tropical subvarieties of tropical abelian varieties, exactly parallel to the classical Green–Lazarsfeld inequality. The proof strategy is well structured and uses sophisticated tools (Saito's mixed Hodge modules, decomposition theorem, toric transversality) in a way that is likely influential. The paper also gives a new proof of a version of the Adiprasito–Björner vanishing for realizable matroids and supplies a counterexample to a proposed generalization. The writing is careful and the modular architecture (Prop 4.1 → Prop 5.1 → Thm 2.8 → Thm 7.1 → Thm 1.1) is coherent. However, the main local vanishing theorem rests on Proposition 6.1, whose proof as written has a real gap, and on an external 2024 preprint ([AP24]). These issues are load-bearing and must be resolved before the central claim can be considered fully established.","major_comments":[{"comment":"The construction of the cone σ_S is incomplete. The proof chooses unit vectors v_1,…,v_n “sufficiently close to v” and sets σ_S = {u : (u,v_i) ≥ 0}. For the subsequent claims (σ_S ⊂ {ℓ>0}∪{0} and every ℓ-positive ray of Δ lying in the relative interior of σ_S) one needs v ∈ int(cone(v_i)); “sufficiently close” does not force this, because the v_i could all lie in a half-space through v. This is not a mere technicality: if σ_S contains a ray with ℓ<0, the subsequent link retraction argument can fail. Please add the surrounding condition explicitly and prove that such v_i exist. In addition, the bullet property asserting that the homotopies H^S_σ and H^Σ_σ stay outside σ_S is asserted without proof; since the retraction moves points along straight segments, this requires verification. A complete proof of Prop. 6.1 is necessary for Theorem 2.8.","section":"§6, Proposition 6.1 (page 15)"},{"comment":"The existence of the common stellar refinement Δ′ is imported from [AP24, Theorem 1.1], an arXiv preprint from 2024 that resolves a classical conjecture. This places a substantial external result at the base of the main theorem. If the result has since been accepted for publication, please cite the published version; if not, either prove the special case needed here (where one of the fans is a subdivision containing the cone σ_S) or clearly flag the dependence. As it stands, the unsupported preprint makes the proof of Theorem 2.8 conditional.","section":"§6, Proposition 6.1, “moreover” statement (page 16)"},{"comment":"Proposition 5.1 requires each D_I to be connected and rationally smooth. Corollary 3.8 establishes rational smoothness of intersections with toric subvarieties, but connectedness of D_I = \\bar{U} ∩ (intersection of boundary divisors) is not proved. The H-regularity assumption only guarantees that U∩O is connected for each torus orbit O; it does not, as written, imply connectedness of the intersection of \\bar{U} with an orbit closure, which may meet several orbits. This gap affects the identification H^0(D_I,Q)^∨ ≅ Q used in the proof of Prop. 5.1. Please supply a proof or a precise argument showing that the hypotheses of Prop. 5.1 hold in the setting of Section 6.","section":"§6, application of Proposition 5.1; §5, Proposition 5.1"}],"minor_comments":[{"comment":"The stated homology is internally inconsistent: a wedge of ten 2-spheres has H_2 ≅ Z^{10}, not H_3 ≅ Z^{10}. Moreover, the homology of the 2-skeleton of a 6-simplex has H_2 ≅ Z^{20} (Euler characteristic 21 − 1 = 20), not 10. Please correct the Betti numbers and the wedge count. The main point of the example (non-vanishing H_1) is unaffected.","section":"§8, Lemma 8.1"},{"comment":"The abstract says the link is “not homotopy equivalent to a bouquet of 2-spheres,” while the introduction says “not homotopy equivalent to a bouquet of spheres.” The stronger statement with arbitrary spheres is false for the constructed example, which is a bouquet of S^1 and 2-spheres. Please make the wording consistent and accurate.","section":"Abstract and Introduction"},{"comment":"“Sch¨ on” should be “Schön.”","section":"§2, Remark 2.5"},{"comment":"The proof refers to Figure 1, but the figure is not reproduced in the text. If this is an arXiv rendering issue, please ensure the figure appears in the final version.","section":"§2, Example 2.11"}],"recommendation":"major_revision","confidential_remarks":"The paper is promising and the architecture is sound, but the gap in Proposition 6.1 and the reliance on [AP24] are serious enough that the main theorem is not yet proven. I believe the Prop. 6.1 gap is likely repairable by imposing v ∈ int(cone(v_i)) and carefully checking the homotopies, and the [AP24] dependence may be avoidable or at least more carefully documented. The Section 8 homology numbers should also be corrected. I would encourage revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Thanks for the report. I read the paper with your notes. My take: this is a real result, not a dressed-up special case. The main novelty is Theorem 1.1 plus the local vanishing Theorem 2.8, and the proof route through MHM, toric compactifications, and Morse theory is coherent. I checked the weight arithmetic in Prop 4.1 and the spectral sequence in Prop 5.1; those hold. The counterexamples matter—Example 2.11 shows H-regularity is needed, and Section 8 addresses Hacking's question. I agree with you on the load-bearing issue: Proposition 6.1 is under-specified. The stress-test note is right that your \"wide cone cannot be contained in ℓ>0\" is not accurate—a dual cone can be wide and still lie inside ℓ>0 if its generators surround v. But the gap is real: the proof never states that v must lie in the interior of cone(v_i). Without it, the chosen σ_S can have rays with ℓ<0. Since n vectors can form a simplicial cone around v, the fix is easy: choose v_i so that v ∈ int(cone(v_i)) and all are close to v. Add that sentence and the argument goes through.\n\nTwo smaller things. First, the \"moreover\" in Prop 6.1 uses [AP24] (common stellar subdivisions). That is a 2024 preprint resolving a conjecture; the paper cites it without comment. The rest of the paper does not depend on that specific theorem, but it should be flagged. Second, Lemma 8.1 has a typo/arithmetic slip: the 2-skeleton of a 6-simplex has H2 = Z^20, not Z^10, and the displayed H3=Z^10 should be H2=Z^10 (or Z^20). The example's conclusion is unaffected because H1=Z is what violates the bouquet-of-2-spheres property, but the count should be corrected.\n\nI also noted two standard facts used without citation (purity of cohomology of rationally smooth projective varieties; the fiber-bundle claim in Theorem 7.1). These are minor and fixable.\n\nNet: the main claims are believable, the architecture is sound, and the flaws are repairable. The paper deserves a serious referee; I'd send it to someone comfortable with mixed Hodge modules. I'd cite it once the Prop 6.1 fix is in.","headline":"A serious tropical-geometry paper proving a Green–Lazarsfeld analogue for H-regular subvarieties; the main theorem is plausible and the architecture is sound, but Proposition 6.1 needs a small fix and the Section 8 sphere count has a typo.","tokens_in":21384,"tokens_out":12365,"would_cite":true,"duration_ms":127110,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14T05","14C40","14F08"],"pacs":[],"model":"deepseek-v4-flash","headline":"H-regular tropical subvarieties obey the Green–Lazarsfeld sign law: (−1)ᵈ χ(X) ≥ 0.","keywords":["tropical abelian variety","Euler characteristic","H-regular tropical fan","Green–Lazarsfeld theorem","tropical homology","Lefschetz-type theorem","Bergman fan","Morse theory"],"falsifier":"A direct combinatorial check: for any complete simplicial fan Δ and linear functional ℓ, try to exhibit a rational simplicial cone σ_S contained in {ℓ>0}∪{0} whose relative interior contains every ℓ-positive ray of Δ, while also being wide enough for the deformation retraction argument. If no such cone can exist (e.g., for a fan with two opposing ℓ-positive rays nearly spanning the complement of a thin cone), Proposition 6.1 fails.","tokens_in":20047,"feed_emoji":"","tokens_out":1308,"duration_ms":14307,"temperature":0.7,"pith_summary":"This paper proves a tropical analogue of a classical theorem of Green and Lazarsfeld: on a complex abelian variety, the signed Euler characteristic of a smooth subvariety X satisfies (−1)^{dim X} χ(X, O_X) ≥ 0. Here, for a `d`-dimensional tropical subvariety X of a tropical abelian variety, the paper establishes (−1)ᵈ χ(X) ≥ 0 under an `H-regularity` condition that locally models X on tropicalizations of well-behaved very affine varieties. The sign law is combinatorial: χ(X) is computed from tropical Hodge numbers, and the paper's local vanishing theorem also yields a Lefschetz-type statement for affine H-regular tropical varieties. The authors further show the sign inequality fails for general tropical subvarieties and give an explicit non-H-regular tropical fan whose link is not a bouquet of spheres.","feed_headline":"Tropical subvarieties obey Green–Lazarsfeld sign law","feed_subtitle":"For H-regular d-dimensional tropical cycles in a tropical abelian variety, (−1)ᵈ χ(X) ≥ 0.","key_machinery":"H-regularity: a closed irreducible subvariety U of the complex affine torus (C*)ⁿ is H-regular if its closure in some smooth projective toric variety is smooth, meets every torus orbit transversely, and each nonempty orbit intersection is connected. A tropical fan is H-regular if it is the tropicalization of such a U. This condition is the load-bearing geometric input that guarantees the existence of a toric compactification whose boundary has rationally smooth strata, allowing the authors to compare the tropical link to a dual complex and apply mixed-Hodge-module vanishing results.","core_discovery":"Theorem 1.1: If A is a tropical abelian variety (a real torus with positive definite quadratic form) and X is an H-regular, pure-dimensional tropical subvariety of dimension d, then (−1)ᵈ χ(X) ≥ 0. This is proved via a Morse-theoretic argument built on a local statement: for an H-regular tropical fan T of dimension d and a linear function ℓ that does not vanish on any ray, the reduced homology of the link of T ∩ {ℓ ≤ 0} vanishes in all degrees except possibly d−1 (Theorem 2.8). The local vanishing is established by relating the link to the dual complex of a toric compactification, converting the problem into a vanishing of top-weight cohomology for a quasi-projective variety that admits a pr","pith_inferences":["A natural testable extension is whether the signed inequality generalizes to larger classes of tropical varieties, e.g., those locally connected in codimension 1 but not H-regular; the authors suspect it does not, and their Example 2.11 gives a starting point.","The dependency of Proposition 6.1 on a 2024 preprint about common stellar subdivisions of fans means that if that result were incomplete, the local vanishing theorem would lack a fully established proof; a reader might check whether the cone construction can be simplified to avoid that citation.","The paper's explicit 3-dimensional tropical fan whose link is not a bouquet of spheres shows that H-regularity is not merely a technical condition but controls fundamental topological obstructions; it would be interesting to see whether such non-bouquet links obstruct any combinatorial Hodge-theoretic positivity.","The authors raise the question of a tropical Chern-class refinement; if a signed Chern-class positivity holds for H-regular tropical varieties, it would give a stronger structural constraint than the Euler characteristic inequality alone."],"forward_implications":["If the main theorem holds, every H-regular d-dimensional tropical subvariety of a tropical abelian variety has signed Euler characteristic of the predicted sign, determined solely by dimension.","The signed Euler characteristic of such a subvariety equals the signed alternating sum of tropical Hodge numbers, linking the topological statement to tropical Hodge theory.","The local vanishing theorem (Theorem 2.8) yields a Lefschetz-type result: for a general affine hyperplane H, the relative homology H_k(X, X∩H) vanishes for all k≠d.","The sign law fails without H-regularity: the paper constructs a 2-dimensional tropical cycle in a quotient of a product of Jacobians with Euler characteristic −4.","The local vanishing gives a new proof of a slight weakening of a result of Adiprasito–Björner for Bergman fans of matroids realizable over C, and answers a question of Mikhalkin–Ziegler in that case."],"fun_headline_variants":["H-regular tropical cycles: Euler sign ≥ 0","Green–Lazarsfeld tropical: Euler sign nonnegative","Signed Euler nonnegative for H-regular tropicals","Euler sign law for H-regular tropical subvarieties"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The existence of a sufficiently fine simplicial cone σ_S in the positive half-space of ℓ that contains all ℓ-positive rays of the ambient fan—used to build the retraction to the link of T∩{ℓ≤0}—is asserted by Proposition 6.1; if this cone cannot be constructed, the local vanishing theorem has no proof.","fun_headline_variants_meta":{"raw":{"variants":["H-regular tropical cycles: Euler sign ≥ 0","Green–Lazarsfeld tropical: Euler sign nonnegative","Signed Euler nonnegative for H-regular tropicals","Euler sign law for H-regular tropical subvarieties"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001117,"raw_usage":{"total_tokens":4458,"prompt_tokens":683,"completion_tokens":3775,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":427,"completion_tokens_details":{"reasoning_tokens":3716}},"tokens_in":427,"tokens_out":3775,"duration_ms":27024,"temperature":1.0,"reasoning_tokens":3716,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T11:58:20.857869+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct combinatorial check: for any complete simplicial fan Δ and linear functional ℓ, try to exhibit a rational simplicial cone σ_S contained in {ℓ>0}∪{0} whose relative interior contains every ℓ-positive ray of Δ, while also being wide enough for the deformation retraction argument. If no such cone can exist (e.g., for a fan with two opposing ℓ-positive rays nearly spanning the complement of a thin cone), Proposition 6.1 fails.","supporting_citations":[],"review_version":2}