{"id":"192871ad-211c-477e-a765-f785ccbc2819","arxiv_id":"2506.22067","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every compact Kähler threefold satisfying condition (C) admits a finite étale cover bimeromorphic to a fiber bundle over a torus fiber bundle, yielding a nowhere-vanishing holomorphic 1-form.","lead":"This paper proves that for compact Kähler threefolds, having a holomorphic 1-form without zeros is equivalent to being a fiber bundle over a circle. It settles Kotschick's conjecture in dimension 3 by classifying all such threefolds.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.2's proof depends on the unproved 'restricts non-trivially on torus fibers' property (Theorem 1.4(iv), Theorems 5.1/6.1/7.1); Remark 1.6 shows this property is not automatic, and the cited induction is omitted.","rationale":"The reader's weakest_assumption concerned the extension of Hao-Schreieder projective results to the Kaehler case. That is a legitimate concern, but the text itself points to a more direct missing proof that is load-bearing for the central claim: the 'restricts non-trivially on the fibers' property, which is exactly what the proof of Theorem 1.2 uses. The paper's own Remark 1.6 demonstrates that the property is not a formal consequence of having a torus fiber bundle and a form satisfying condition (C): the product of three elliptic curves with projection to one factor gives a form pulled back from the base, satisfying condition (C) but trivial on the torus fibers. The remark says the missing arrangement is supplied by an induction, but the induction is not given. Without the restriction property, the proof of Theorem 1.2 only shows that the induced form on the minimal model has no zeros under an additional hypothesis that is not established. This is an internal gap, distinct from the external theorem-extension issue the reader flagged, so my agreement is partial. The gap is a missing verification rather than a demonstrated falsehood, so the appropriate verdict remains conditional; the reader's CONDITIONAL verdict already requires filling such gaps before acceptance.","tokens_in":46576,"tokens_out":26699,"duration_ms":302223,"concrete_test":"Prove the descent lemma for the torus fiber bundles constructed in Theorems 5.1 and 6.1, which have trivial monodromy after replacing the base by a finite etale cover: if omega_Z = pi^* omega_B and (Z, omega_Z) satisfies condition (C), show that (B, omega_B) satisfies condition (C) for every finite etale cover B' -> B. This can be attempted by applying the Leray spectral sequence to the pulled-back bundle and using that the fiber cohomology is a nonzero vector space with zero differential under wedge with pi^* omega_B. Then write out the induction claimed in Remark 1.6 for the cases kappa = 0 and kappa = -infinity(b), exhibiting a base B with B not satisfying condition (C); if no such base exists for some example such as X = E^3 or a P^1-bundle over a ruled surface, Theorem 1.4 needs revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central implication (C) => (A) in Theorem 1.2 is reduced in the Introduction to showing that the induced form omega_Z on the torus fiber bundle Z -> B has no zeros. The proof states that this follows from Theorem 1.4(iv), because omega_Z restricts non-trivially on the torus fibers of Z -> B. The same restriction property appears as the 'Moreover' clauses in Theorem 5.1, Theorem 6.1, and Theorem 7.1, specifically the final sentence of Theorem 7.1. None of the proofs of these theorems verifies this restriction property; they construct the finite etale covers and the fibrations, but do not trace the holomorphic 1-form through the constructions. The property is not automatic: Remark 1.6 gives X = E^3 with Z -> B the projection onto the last factor, where a form pulled back from B satisfies condition (C) yet restricts trivially on the fibers. The remark explicitly says that without arranging B not to satisfy condition (C), item (iv) fails, and it refers to an induction argument that is not written out. Thus the main theorem rests on an unproved descent statement: if omega_Z is pulled back from B and (Z, omega_Z) satisfies condition (C), then (B, omega_B) must satisfy condition (C), contradicting the choice of B. If this descent statement is false, or if the promised induction cannot always produce a base not satisfying condition (C), then Theorem 1.2 is not established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies compact Kähler threefolds satisfying condition (C) from Kotschick's conjecture, i.e. carrying a holomorphic 1-form whose cup product is exact on every finite étale cover. The main Theorem 1.2 asserts that (A), (B), and (C) are equivalent in dimension 3, which would prove Kotschick's conjecture there. The structural Theorem 1.4/7.1 claims that after a finite étale cover and blow-downs, such a threefold admits a smooth morphism to a locally trivial torus fiber bundle over a base that does not satisfy condition (C), with fibers P1, P2, or Hirzebruch surfaces. The proof combines the MMP for Kähler threefolds, abundance, the Beauville–Bogomolov decomposition, and the twisted equivariant Weierstraß/tautological model machinery, extending earlier work of Hao and Schreieder from the projective case.","tokens_in":46914,"tokens_out":5558,"duration_ms":64406,"significance":"If the main theorem is correct, it resolves Kotschick's conjecture in dimension 3 and gives a concrete classification of the threefolds admitting nowhere-vanishing holomorphic 1-forms. The paper engages seriously with the genuinely new Kähler phenomena: unlike the projective case, a smooth morphism to a positive-dimensional torus need not exist, and the author identifies this obstruction explicitly. The manuscript is transparent about several delicate points, notably Remark 1.6 and Remark 6.23, and it builds on substantial established machinery rather than introducing ad hoc assumptions. No circularity or fitted parameters are apparent. The main concerns are not about the overall strategy but about unproved load-bearing assertions concerning the behavior of the holomorphic 1-form under the constructed fibrations and about the claimed Kähler versions of projective results.","major_comments":[{"comment":"The final assertion of Theorem 7.1, that the induced form ω_n restricts non-trivially on the fibers of (Y × A)^η → Y, is not proved. The proof of Theorem 1.2 reduces the implication (C) ⇒ (A) to exactly this property, and the proofs of Theorems 5.1 and 6.1 likewise state 'Moreover' clauses asserting non-trivial restriction of the induced form, but the proofs construct the finite étale covers and fibrations without tracing the holomorphic 1-form through the Weierstraß/tautological model constructions. Since Remark 1.6 shows the property is not automatic, this is a load-bearing gap, not a cosmetic one.","section":"§7, Theorem 7.1 (and proof of Theorem 1.2)"},{"comment":"The assertion that the base B can be arranged not to satisfy condition (C) 'by an induction argument' is not carried out anywhere in the paper. This arrangement is essential: it is what makes item (iv)—and hence the proof of Theorem 1.2—work. The underlying descent statement, that a pulled-back form with condition (C) on the total space forces condition (C) on the base, is neither stated nor proved. The example X = E^3 in Remark 1.6 shows that without this step the argument fails. A complete induction with base case and descent step should be written out.","section":"Introduction, Remark 1.6 and Theorem 1.4(ii)–(iii)"},{"comment":"Several load-bearing results are invoked with the explanation that their proofs 'do not require projectivity of X', but no details are given. This applies to [HS21b, Lem. 2.5, Prop. 5.3, Prop. 5.8] in the proof of Theorem 5.1, to Lemma 6.2, and to parts of Section 6.2 and Section 6.4. These results control the singular fibers, the j-invariants, and the monodromy of the Iitaka fibration, so they are central to the classification. The paper should either reproduce the arguments in the Kähler setting or identify explicitly which steps in [HS21b] use projectivity and why they can be omitted here.","section":"Sections 5–6, Kähler extension of [HS21b]"}],"minor_comments":[{"comment":"The typesetting of arrows and maps is corrupted in the supplied text (for example '/∫hortrightarrow' appears in place of arrows); this should be cleaned up in the final version.","section":"Throughout"},{"comment":"The notation for the torus fiber bundle is inconsistent: Theorem 1.4 uses Z := X_min → B, while Theorem 7.1 expresses the same structure as X'_n → (Y × A)^η → Y. The correspondence should be stated explicitly.","section":"Theorem 1.4 and Theorem 7.1"},{"comment":"The dependence on the external correction [HS25] should be integrated into the main text more clearly, since the remark indicates that a step in the projective predecessor paper was incomplete.","section":"Remark 6.23"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely correct in its overall strategy, and the missing restriction property seems fixable by writing out the induction and the descent argument that Remark 1.6 already gestures at. I would not reject on the current evidence, but the manuscript as submitted does not yet establish Theorem 1.2, because the proof relies on assertions that are explicitly stated but not verified. The Kähler-extension claims of the Hao–Schreieder results also need either detailed justification or a precise statement of which projective assumptions are used."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a substantial preprint. Pietig extends the Hao-Schreieder classification from projective threefolds to all compact Kähler threefolds, and the genuinely new part is the machinery: torus fiber bundles replace the Albanese morphism, and equivariant Weierstraß and tautological models do the heavy lifting. If the main theorem is right, it proves Kotschick's conjecture in dimension 3 and gives a clean structural classification. The paper is also honest about external issues, including Remark 6.23 about a gap in Hao-Schreieder's step 3 and an alternative fix. Credit is due for that.\n\nThe soft spot is exactly the one the author names in Remark 1.6. The proof of Theorem 1.2 reduces (C) => (A) to the statement that the induced form on the torus fiber bundle Z restricts non-trivially on the fibers, Theorem 1.4(iv). That restriction property appears as a \"Moreover\" clause in Theorems 5.1, 6.1, and 7.1, but the proofs never actually trace the form through the constructions. The author says it can be arranged by an induction that ensures the base B does not satisfy condition (C), and his own example E^3 shows the property is not automatic. The induction is not written out. This is load-bearing: without it, the proof of Theorem 1.2 does not go through. The gap may well be fixable, and the lower-dimensional classification makes me suspect the intended induction works, but \"easily arranged\" is not a proof.\n\nWho is this for: anyone working on holomorphic one-forms, Kähler threefolds, or the Schreieder-Hao circle. It deserves a serious referee, not a desk rejection. The referee should ask for the missing induction and for the restriction property to be verified inside Theorems 5.1 and 6.1. As it stands, I would not yet cite it as a theorem, but I would definitely engage with it and I expect it to become a standard reference once the gap is closed.","headline":"Pietig's Kähler threefold classification is a real advance and likely correct, but the main theorem's proof currently depends on an unproved induction that the author himself flags as crucial.","tokens_in":47436,"tokens_out":3167,"would_cite":false,"duration_ms":37104,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32Q15","14J30","14E30","32J27"],"pacs":[],"model":"deepseek-v4-flash","headline":"For compact Kähler threefolds, a holomorphic one-form without zeros exists exactly when a real closed one-form without zeros does; the resulting classification proves Kotschick's conjecture in dimension three.","keywords":["holomorphic one-forms without zeros","Kähler threefolds","Kotschick conjecture","torus fiber bundles","Iitaka fibration","Weierstraß models","fiber bundles over the circle","Kodaira dimension"],"falsifier":"A direct test is to take a non-algebraic compact Kähler threefold of Kodaira dimension 2 satisfying condition (C) — for instance a twisted product built from the examples of Remark 2.10 — and compute the local monodromies of its Iitaka elliptic fibration along the discriminant divisor. If some local monodromy fails to be finite, or if no finite étale cover of a log-desingularized base trivializes the monodromy of $R^1f_*\\mathbb{Z}$, then the asserted Kähler version of [HS21b, Prop. 5.8] fails and the classification, together with the implication (C)$\\Rightarrow$(A), collapses; checking finiteness and the étale trivialization on such an example would confirm the load-bearing step.","tokens_in":46375,"feed_emoji":"🔁","tokens_out":19117,"duration_ms":174093,"temperature":0.7,"pith_summary":"This paper proves Kotschick's conjecture in dimension three: a compact Kähler threefold admits a holomorphic one-form without zeros exactly when it admits a real closed one-form without zeros (equivalently, by Tischler's theorem, when it is a $C^\\infty$-fiber bundle over the circle), and both are equivalent to the cohomological exactness condition (C) introduced in [Sch20]. The proof proceeds by classification. Extending the projective threefold case of [HS21b], the paper shows that any Kähler threefold carrying such a form has a finite étale cover which, after blowing down elliptic curves, becomes a locally trivial fiber bundle whose fibers are positive-dimensional tori, with base a Kähler manifold that itself carries no such form; when the Kodaira dimension is negative, the fibers are instead $\\mathbb{P}^1$, $\\mathbb{P}^2$, or Hirzebruch surfaces. The decisive difference from the projective case is that a smooth morphism to a positive-dimensional torus need not exist: the product splitting used in [HS21b] has to be replaced by the more flexible notion of a twisted torus fiber bundle $(Y\\times A)^\\eta\\to Y$. The theorem settles the conjecture in dimension three and gives a complete structural description of every Kähler threefold that fibers over the circle.","feed_headline":"Proven: Kotschick's conjecture holds for Kähler threefolds","feed_subtitle":"A holomorphic one-form without zeros exists exactly when the threefold is a smooth fiber bundle over the circle.","key_machinery":"The central object is the torus fiber bundle with its twisted Jacobian fibration. A torus fiber bundle $f:X\\to Y$ of fiber dimension $g$ is encoded by a variation of Hodge structures $H$ (the local system $R^{2g-1}f_*\\mathbb{Z}$ together with its Hodge filtration) and a cohomology class $\\eta\\in H^1(Y,J_H)$; the total space is the torsor $J^\\eta_H$ (Proposition 2.4), and it is Kähler precisely when the Chern class $c(\\eta)$ is torsion (Proposition 2.7). The classification feeds the Iitaka fibration of $X$ into the theory of Weierstraß models [Nak87], twisted equivariant Weierstraß models [CHL19], and tautological models [Lin20]: once a finite étale cover has trivialized the monodromy, an elliptic fibration with local meromorphic sections is bimeromorphic to a twisted product $(Y\\times E)^\\eta$ with diagonal group action by translations (Proposition 2.22). Kollár's flop theorem [Kol89] then upgrades bimeromorphic equivalence to isomorphism of minimal models, so the constructed bundle is literally the minimal model of $X$ rather than merely birational to it. The minimal model program and abundance theorem for Kähler threefolds [HP16], [CHP16] supply the Iitaka fibration, and the Beauville–Bogomolov decomposition [Bea83] handles the $\\kappa=0$ case.","core_discovery":"The central claim is Theorem 1.2: for a compact Kähler manifold of dimension 3, the three conditions (A) a holomorphic one-form without zeros, (B) a real closed one-form without zeros, and (C) the existence of a holomorphic one-form $\\omega$ such that the cup-product maps $\\wedge\\tau^*\\omega$ are exact on the cohomology of every finite étale cover, are equivalent. The implications (A)$\\Rightarrow$(B) and (B)$\\Rightarrow$(C) were known, so the new content is (C)$\\Rightarrow$(A), obtained through the structure theorem Theorem 1.4: a threefold satisfying (C) has a finite étale cover which, after a sequence of blow-downs along elliptic curves not contracted by the Albanese map, is either a locally trivial fiber bundle over a compact Kähler base with all fibers isomorphic to a positive-dimensional torus (when $\\kappa(X)\\ge 0$), or a smooth bundle over an elliptic curve with fibers $\\mathbb{P}^2$ or a Hirzebruch surface, or a $\\mathbb{P}^1$-bundle over a torus-fiber-bundle surface (when $\\kappa(X)=-\\infty$); in the torus-bundle cases the base does not itself satisfy (C). The induced one-form restricts non-trivially on the torus fibers, which is why it has no zeros. In contrast to the projective case treated in [HS21b], the morphism to a positive-dimensional torus need not exist: the product splitting $X'\\simeq S'\\times A'$ is replaced by a twisted torus fiber bundle $(Y\\times A)^\\eta\\to Y$, whose Kählerity is governed by the torsion of the Chern class of $\\eta$.","pith_inferences":["The paper's insistence that the base of the torus bundle fails condition (C) — which Remark 1.6 shows is essential to the no-zeros conclusion — suggests a possible inductive strategy for Kotschick's conjecture in higher dimensions: turn the relevant fibration into a torus fiber bundle over a smaller base that fails (C), then the no-zeros property would follow dimension by dimension.","The examples of Remark 2.10 (a general-type Kähler surface with nontrivial $H^1(Y,\\mathcal{O}_Y)$, twisted by a non-torsion class) are the minimal demonstrations that the projective splitting must fail; one could test the classification directly on them, checking that they fall into the torus-fiber-bundle case with base failing (C).","The load-bearing external input — that the key lemmas of [HS21b] extend verbatim to the Kähler setting — invites a proof audit: tracing each cited lemma (for instance the monodromy trivialization of $R^1f_*\\mathbb{Z}$ over a log-desingularized base) would determine whether the classification rests on a genuinely Kähler argument or on an unstated projectivity assumption.","Because the classification enumerates all $C^\\infty$ circle-bundles among Kähler threefolds, it also answers the purely topological question of which Kähler threefolds admit a real closed 1-form without zeros; pairing the list with characteristic-number obstructions of the kind used in [Kot22] could show which of these bundles are realized by holomorphic forms and which by smooth forms only."],"forward_implications":["Kotschick's conjecture holds for all compact Kähler threefolds: conditions (A), (B), and (C) are equivalent, so a holomorphic one-form without zeros exists exactly when the manifold is a $C^\\infty$-fiber bundle over the circle.","Every Kähler threefold with a non-vanishing holomorphic one-form is, after a finite étale cover and blow-downs along elliptic curves, a $\\mathbb{P}^1$-, $\\mathbb{P}^2$-, or Hirzebruch-surface bundle over a locally trivial torus fiber bundle over a Kähler base that itself carries no such form (Theorem 1.4).","The structure is refined by Kodaira dimension (Corollary 1.5): for $\\kappa=2$ one gets an elliptic fiber bundle over a general-type Kähler surface, for $\\kappa=1$ a 2-torus bundle over a curve of genus at least 2 or an elliptic bundle over a Kähler surface of Kodaira dimension 1, and for $\\kappa=0$ a product of a torus with a Kähler manifold after a finite étale cover.","The non-vanishing form is produced structurally: the form restricts non-trivially on each torus fiber of the constructed bundle and on the elliptic centers of the blow-downs, which is exactly why it has no zeros (Theorem 1.4(iv) and the proof of Theorem 1.2).","The obstruction to the projective-style splitting is explicit: twisted products $(Y\\times E)^\\eta$ with non-torsion class $\\eta$ are Kähler and carry a zero-free one-form but admit no smooth morphism to a positive-dimensional torus (Remark 2.10)."],"supporting_citations":[{"why":"formulated the conjecture the paper proves and supplied the implication (A)$\\Rightarrow$(B).","marker":"[Kot22]"},{"why":"introduced condition (C), proved (B)$\\Rightarrow$(C) (Thm. 1.2 there), and established the equivalence in dimension 2.","marker":"[Sch20]"},{"why":"the projective threefold classification (Thm. 1.3 there) that this paper extends, whose Lem. 2.5, Prop. 5.3, Prop. 5.8, and Lem. 6.1 are asserted to hold in the Kähler case.","marker":"[HS21b]"},{"why":"Tischler's theorem identifying real closed 1-forms without zeros with $C^\\infty$-fiber bundles over $S^1$, giving condition (B) its geometric content.","marker":"[Tis70]"},{"why":"abundance for Kähler threefolds (with the erratum [CHP23]), used to construct the Iitaka fibration in the classification.","marker":"[CHP16]"},{"why":"the minimal model program and Mori fiber spaces for Kähler threefolds, used for the reduction to minimal models.","marker":"[HP16]"},{"why":"Weierstraß models and the uniqueness of the minimal Weierstraß model for an elliptic fibration with a meromorphic section (Thm. 2.16 here).","marker":"[Nak87]"},{"why":"twisted G-equivariant Weierstraß models and the Kählerity criterion (Prop. 2.20 here) that a boundary class must be torsion.","marker":"[CHL19]"},{"why":"tautological models (Prop. 2.21 here), representing elliptic fibrations bimeromorphic to Kähler manifolds.","marker":"[Lin20]"},{"why":"the flop theorem (Thm. 4.9) used to show the constructed torus fiber bundle is isomorphic, not merely bimeromorphic, to the minimal model.","marker":"[Kol89]"}],"fun_headline_variants":["Kähler threefolds with zero-free holomorphic 1-forms classified","Kotschick conjecture resolved for compact Kähler threefolds","Structure theorem for Kähler threefolds with no zero 1-forms","Classification of Kähler threefolds fibered over S^1","Dimension 3: Kotschick conjecture verified"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification rests on the assertion, repeated in Sections 5 and 6, that several technical lemmas proved for projective threefolds in [HS21b] (notably Lem. 2.5, Prop. 5.3, Prop. 5.8, and Lem. 6.1 there) also hold for Kähler threefolds because their proofs allegedly do not use projectivity; if any of them secretly depends on projectivity, the reduction of the Iitaka fibration to a locally trivial torus fiber bundle would break.","fun_headline_variants_meta":{"raw":{"variants":["Kähler threefolds with zero-free holomorphic 1-forms classified","Kotschick conjecture resolved for compact Kähler threefolds","Structure theorem for Kähler threefolds with no zero 1-forms","Classification of Kähler threefolds fibered over S^1","Dimension 3: Kotschick conjecture verified"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000631,"raw_usage":{"total_tokens":2964,"prompt_tokens":1046,"completion_tokens":1918,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":662,"completion_tokens_details":{"reasoning_tokens":1836}},"tokens_in":662,"tokens_out":1918,"duration_ms":15455,"temperature":1.0,"reasoning_tokens":1836,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:12:27.805537+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct test is to take a non-algebraic compact Kähler threefold of Kodaira dimension 2 satisfying condition (C) — for instance a twisted product built from the examples of Remark 2.10 — and compute the local monodromies of its Iitaka elliptic fibration along the discriminant divisor. If some local monodromy fails to be finite, or if no finite étale cover of a log-desingularized base trivializes the monodromy of $R^1f_*\\mathbb{Z}$, then the asserted Kähler version of [HS21b, Prop. 5.8] fails and the classification, together with the implication (C)$\\Rightarrow$(A), collapses; checking finiteness and the étale trivialization on such an example would confirm the load-bearing step.","supporting_citations":[],"review_version":1}