{"id":"5a8de9a9-1eef-498a-8dcd-b8b4679db561","arxiv_id":"2501.18088","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A compact Kähler manifold is uniruled exactly when its canonical line bundle is not pseudoeffective, proved by carrying Bost's algebraicity criterion to the Kähler setting.","lead":"A new proof settles a long-open classification criterion: a compact Kähler manifold is covered by rational curves exactly when its canonical line bundle is not pseudoeffective. The paper adapts Bost's algebraicity criterion to Kähler germs and proves new integrability criteria for foliations, with consequences for complex geometry and classification.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 8.1 is false as stated: it omits a semipositivity hypothesis on the Hermitian metric h; the applications in Proposition 8.8 appear to satisfy that hypothesis, so the proof is repairable but the manuscript currently contains a false key lemma.","rationale":"The reader's weakest assumption flagged the Section 8.D positivity imports, and our analysis confirms that this area contains a real defect: Lemma 8.1 is a key analytic input to Proposition 8.8, and its statement is false in the given generality. The counterexample is elementary and directly contradicts the claimed inequality. This is not merely a missing reference or an overly bold conjecture; it is an internally inconsistent statement. However, the defect is repairable: the applications in Proposition 8.8 use Hermitian metrics h_i with curvature equal to a Kähler form, so the semipositivity hypothesis that makes the cited Bergman positivity theorem valid is present in all actual uses. Thus the central theorem 1.1 is not overturned by this flaw, but the manuscript as written contains a false lemma in a load-bearing position. A referee should require the lemma to be corrected and the application to be re-verified against the corrected statement. The formal extension mechanism (Lemma 5.6) and the overall proof architecture appear coherent, and no issue of comparable concreteness was found there. Hence the appropriate verdict is CONDITIONAL acceptance, not rejection or unmodified acceptance.","tokens_in":37688,"tokens_out":46697,"duration_ms":454318,"concrete_test":"Construct the counterexample: let E = C/(Z+iZ), X = C × E, Y = C, f the projection, L = O_X, and h = e^{-|y|^2 v(w)} with v(w) = 1 + 0.9 cos(2π Re w). Verify that v has average 1 and minimum 0.1, and compute the Bergman metric H for m = 1. Check the inequality Θ_g ≥ -Θ_h at a point (0,w_0) with v(w_0) < 0.5; if it fails, Lemma 8.1 is false as stated. Then, to settle the impact on Theorem 1.1, confirm that in Proposition 8.8 each h_i satisfies Θ_{h_i} = ω|_{V_i} ≥ 0 and that the cited Bergman positivity theorem applies with this semipositive metric, so that the main proof survives after correcting Lemma 8.1.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Lemma 8.1 claims that for any projective morphism f:X→Y, any smooth Hermitian metric h on a line bundle L, and any m with H^0(X_y, ω_{X_y}^{⊗m}⊗L|_{X_y}) ≠ {0}, the Bergman-type metric g on ω_{X/Y}^{⊗m} defined by g = H·h^{-1} satisfies Θ_g ≥ -Θ_h. This is equivalent to Θ_H ≥ 0 for the Bergman kernel metric H on ω_{X/Y}^{⊗m}⊗L. That positivity is false without a semipositivity condition on h. Concretely, take X = C × E, Y = C, f the projection, E an elliptic curve, L = O_X, m = 1, and h = e^{-|y|^2 v(w)} where v:E→R is smooth with average 1 and minimum < 1/2. Then H(y) = ∫_E e^{-|y|^2 v(w)} dV(w). At (0,w) one computes Θ_g = i(v(w)-1) dy∧d\\bar{y} and -Θ_h = -i v(w) dy∧d\\bar{y}; the inequality Θ_g ≥ -Θ_h fails wherever v(w) < 1/2. The cited source [BP10, Theorem 0.1] requires the metric on L to have semipositive curvature, a hypothesis missing from Lemma 8.1. In Proposition 8.8 the metrics h_i have Θ_{h_i} = ω|_{V_i}, a Kähler form, so the missing hypothesis is satisfied there. Nevertheless, the written proof invokes Lemma 8.1 in its false full generality, creating a genuine gap in the proof of Theorem 1.1 as stated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Theorem 1.1, which gives a complete characterization of uniruled compact Kähler manifolds by the non-pseudoeffectivity of their canonical bundle: X is uniruled if and only if ω_X is not pseudoeffective. The proof develops a Kähler analogue of Bost's algebraicity criterion for germs, introduces a notion of formal extension of subvarieties along a submanifold, establishes criteria for algebraic integrability of foliations (Theorems 1.3–1.5), and then applies these to a fibration induced by a holomorphic 2-form. The final step uses relative Bergman kernel metrics to force the relative canonical bundle to be pseudoeffective and derives a contradiction by descent and induction on dimension. The argument is long and carefully structured, with many auxiliary lemmas, some of which are delegated to external references.","tokens_in":38080,"tokens_out":7313,"duration_ms":82816,"significance":"If correct, Theorem 1.1 resolves a central conjecture in the classification of compact Kähler manifolds, extending the projective theorem of Boucksom–Demailly–Păun–Peternell to the non-projective Kähler setting. The paper introduces several potentially influential tools: the formal extension criterion for subvarieties, quasi-psh functions with large restricted Lelong numbers, and a positivity concept for torsion-free sheaves on Kähler manifolds. The proof is largely self-contained and the structural organization is clear. However, as discussed below, a key lemma (Lemma 8.1) is stated in false generality, and although the applications appear to satisfy the missing hypothesis, the manuscript as written contains a genuine gap that needs to be addressed.","major_comments":[{"comment":"Lemma 8.1 is false as stated. The lemma claims that for any smooth Hermitian metric h on a line bundle L, the Bergman kernel metric g on ω_{X/Y}^{⊗m} satisfies Θ_g ≥ -Θ_h. This is equivalent to positivity of the Bergman kernel metric H on ω_{X/Y}^{⊗m}⊗L, which requires the curvature of h to be semipositive; the cited source [BP10, Theorem 0.1] indeed assumes semipositivity of the metric on L. Without that hypothesis the statement is contradicted, for example, by X = C × E, Y = C, L = O_X, and h = e^{-|y|^2 v(w)} where v is a smooth function on the elliptic curve E with average 1 and minimum < 1/2: at points where v(w) < 1/2, Θ_g = i(v(w)-1)dy∧d\\bar{y} does not dominate -Θ_h = -i v(w)dy∧d\\bar{y}. In the only application of the lemma, Proposition 8.8, the metrics h_i have Θ_{h_i} = ω|_{V_i}, a Kähler form, so the semipositivity hypothesis is satisfied there and the proof of Theorem 1.1 can be repaired by adding that hypothesis to Lemma 8.1. Nevertheless, the paper currently invokes Lemma 8.1 in its false full generality, so the written proof of Theorem 1.1 contains a genuine gap that must be corrected.","section":"8.A, Lemma 8.1"}],"minor_comments":[{"comment":"The statement of Theorem 1.3 contains a grammatical error: \"Then S0 is has the same dimension\" should read \"Then S0 has the same dimension\".","section":"1. Introduction, Theorem 1.3"},{"comment":"The phrase \"We endow ω_{X/Y}^{⊗ab}|_{V_i} the metric g_i\" is awkward; it should say \"we endow ω_{X/Y}^{⊗ab}|_{V_i} with the metric g_i\".","section":"8.D, Proposition 8.8 proof"},{"comment":"The proof of the first paragraph of Lemma 8.1 is a bare citation to [BP10, Theorem 0.1] and [Cao17, Theorem 3.5]; the statement should explicitly record the semipositivity hypothesis so that the reader can verify the cited result applies.","section":"8.A, Lemma 8.1"},{"comment":"The proof of Lemma 4.4 refers to external results (GKP16, Toma16) without giving the precise statements; a brief indication of how these results combine would improve readability, though the argument itself appears sound.","section":"4., Lemma 4.4"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a very significant problem and the overall strategy appears sound. The false statement in Lemma 8.1 is a load-bearing issue, but it is local and repairable since the actual use in Proposition 8.8 satisfies the missing semipositivity condition. I recommend major revision rather than rejection, and I would be willing to accept a corrected version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague—\n\nThis is a substantial paper. It proves the Kähler analogue of the Miyaoka-Mori/BDPP criterion: a compact Kähler manifold is uniruled iff its canonical bundle is not pseudoeffective. That has been open in general, and the proof is a real technical advance, not an incremental twist. The adaptation of Bost's algebraicity criterion through formal extensions, quasi-psh functions, and Lelong numbers is genuinely new, and the foliation criteria in Theorems 1.3 and 1.4 are likely to be reusable in the Kähler MMP. I found no post-hoc fitting or circular use of the target theorem; the induction on dimension and the import of the projective case from BDPP are clean.\n\nThe soft spot is real. Lemma 8.1 is false as stated: it claims Θ_g ≥ −Θ_h for the Bergman kernel metric for any smooth Hermitian metric h, but that positivity needs h semipositive. The stress-test counterexample with X=C×E, Y=C, L=O_X and h=e^{−|y|^2 v} is correct—the inequality fails where v(w)<1/2. The cited results in [BP10] and [Cao17] do require semipositive curvature, and the proof of Lemma 8.1 does not supply it. In the application (Proposition 8.8), the metrics h_i have curvature equal to a Kähler form, so the missing hypothesis is satisfied there. The fix is straightforward: add semipositivity to the statement, and verify it for the h_i in the application. But the manuscript as written contains a false key lemma, so the proof of Theorem 1.1 is not formally correct until that is patched.\n\nA lesser issue: a few lemmas are delegated with 'the same argument works' in the Kähler setting (e.g., 4.4 and 6.2). I think those are likely fine, but I would want the author to spell out the Kähler adaptions before signing off.\n\nRecommendation: deserve serious peer review, conditional on fixing Lemma 8.1 and a careful pass over the delegated positivity results. If the fix lands, this is a major result.","headline":"Important new Kähler uniruled criterion; the main theorem is likely correct, but Lemma 8.1 is false as stated and needs a routine repair.","tokens_in":38602,"tokens_out":3080,"would_cite":true,"duration_ms":34028,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14E30","32Q15","14M20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that a compact Kähler manifold is uniruled if and only if its canonical line bundle is not pseudoeffective.","keywords":["uniruled manifolds","compact Kähler manifolds","pseudoeffective canonical bundle","foliations","algebraic integrability","Lelong numbers","meromorphic maps","formal extensions"],"falsifier":"A compact Kähler manifold $X$ with $\\omega_X$ not pseudoeffective that is not dominated by rational curves would refute Theorem 1.1, and the paper's final contradiction argument is precisely an attempt to rule out such an object.","tokens_in":37506,"feed_emoji":"","tokens_out":21137,"duration_ms":218576,"temperature":0.7,"pith_summary":"This paper proves a complete positivity characterization of uniruled compact Kähler manifolds: $X$ is covered by rational curves if and only if its canonical line bundle $\\omega_X$ is not pseudoeffective, meaning no positive closed current can represent its class. The projective version of this statement was already known, so the new content is the converse direction in the full Kähler category, where algebraic techniques such as Chow's theorem are unavailable. The proof develops a Kähler analogue of an algebraicity criterion for germs, then uses it to show that foliations whose cotangent bundle is not pseudoeffective are induced by meromorphic maps, and finally runs a dimension-by-dimension contradiction built on holomorphic 2-forms, relative Albanese reductions, and positivity of relative canonical bundles. If correct, the theorem turns uniruledness into a cohomological positivity condition that can be checked without knowing the rational curves themselves.","feed_headline":"Kähler uniruled iff canonical bundle is not pseudoeffective","feed_subtitle":"The hard direction, proved via foliations and Lelong numbers, now works in the full Kähler category","key_machinery":"The engine is Theorem 1.2, a Kähler counterpart of the projective algebraicity criterion. It says that if a locally closed submanifold $S_0$ contains a Zariski open piece $C_0$ of a submanifold $C$ and extends formally along $C$ — meaning the pair survives an infinite sequence of blowups along the successive intersections — then for any $\\lambda>0$ there is an $\\omega$-plurisubharmonic function $\\varphi$ with analytic singularities whose restriction Lelong numbers along $C_0$ exceed $\\lambda$. The mass-concentration and regularization theorems for plurisubharmonic functions produce such functions, and Lemma 2.5 extracts their vanishing order by an explicit blowup cascade. Theorem 1.3 then turns largeness of restriction Lelong numbers into an algebraicity conclusion: $S_0$ has the same dimension as its Zariski closure whenever the extended conormal bundle is non-pseudoeffective. For a foliation, $S_0$ is the analytic graph, and Lemma 5.6 shows it extends formally along the diagonal, so Theorem 1.4 follows: a foliation with non-pseudoeffective cotangent, or with positive minimal slope with respect to a movable class, is induced by a meromorphic map. The final section imports Bergman-kernel positivity of relative canonical bundles to force pseudoeffectivity of $\\omega_{X/W}$ and close the induction.","core_discovery":"The central claim, Theorem 1.1, is that for every compact Kähler manifold $X$, $X$ is uniruled if and only if $\\omega_X$ is not pseudoeffective. One direction is standard: a free rational curve intersects $\\omega_X$ negatively, so a positive current representing $\\omega_X$ would have to be nonnegative on that curve, forcing the canonical class to be pseudoeffective only when no such curve exists. The reverse direction is the paper's achievement: assuming $\\omega_X$ is not pseudoeffective, it constructs the rational curves that dominate $X$. The bridge is a new algebraicity criterion for locally closed submanifolds of a compact Kähler manifold, proved through formal extensions and restriction Lelong numbers, and applied to the analytic graph of a foliation so that non-pseudoeffective cotangent bundles force the foliation to come from a meromorphic map. With that in hand, the proof rules out the non-uniruled alternative by an induction on dimension that uses holomorphic 2-forms, relative Albanese fibrations, and Bergman-kernel positivity.","pith_inferences":["The paper leaves implicit that the classical conjecture 'uniruled iff Kodaira dimension is negative' is not formally settled by this theorem: the theorem uses the stronger hypothesis 'canonical bundle not pseudoeffective', so any remaining gap would be a non-uniruled manifold whose canonical class is pseudoeffective yet admits no pluricanonical sections.","The formal-extension and Lelong-number machinery is a transferable tool: the same two conditions — formal extendability plus non-pseudoeffectivity of the conormal bundle — could be tested for other algebraicity questions, such as which germs or leaves are algebraic under weaker positivity hypotheses.","Because the proof separates the obstruction into formal extension, conormal non-pseudoeffectivity, and relative canonical positivity, a natural next step is to check whether the same trichotomy survives for mildly singular Kähler spaces or for foliations with only weak positivity of their cotangent bundles."],"forward_implications":["Uniruledness and canonical pseudoeffectivity now split the class of compact Kähler manifolds into two disjoint families: every non-uniruled example has a pseudoeffective canonical bundle.","Foliations with non-pseudoeffective cotangent bundle on a compact Kähler manifold are induced by meromorphic maps, so their leaves are algebraic; the same holds if the foliation has positive minimal slope with respect to some movable class.","For a foliation restricted to a subvariety, non-pseudoeffectivity of the pulled-back tangent distribution forces the local leaves through that subvariety to have Zariski closures of the same dimension.","The known threefold case is recovered, and the uniruledness criterion is now uniform in every dimension across projective and Kähler manifolds."],"supporting_citations":[{"why":"Supplies the projective theorem this paper extends and the projective case used in the final step of the proof of Theorem 1.1.","marker":"[BDPP13]"},{"why":"Provides the algebraicity characterization for germs in projective varieties that the paper adapts into Theorem 1.2.","marker":"[Bos04]"},{"why":"Gives the algebraicity criterion for leaves of foliations that is extended to the Kähler setting in Theorems 1.3 and 1.4, and supplies the analytic-graph formulation.","marker":"[Bos01]"},{"why":"Defines the analytic graph and graphic neighborhood of a foliation used in Lemma 5.6 and Theorem 1.4, and supplies the algebraic-integrability criterion being generalized.","marker":"[BM16]"},{"why":"Provides the mass-concentration construction used in Lemma 6.2 to produce $\\omega$-psh functions with arbitrarily large restriction Lelong numbers.","marker":"[Dem93]"},{"why":"Provides the regularization of closed positive currents used in Theorem 2.4 to turn mass concentration into functions with analytic singularities while controlling Lelong numbers.","marker":"[Dem92]"},{"why":"Supplies divisorial Zariski decompositions and the exceptional-family technology used to separate classes from the pseudoeffective cone in Lemmas 3.1–3.3 and 6.1.","marker":"[Bou04]"},{"why":"Supplies the positivity of direct images of adjoint relative canonical bundles imported as Lemma 8.1 to force $\\omega_{X/W}$ to be pseudoeffective.","marker":"[BP10]"},{"why":"Provides Bergman-kernel positivity and extension results for compact Kähler fibrations used in Lemmas 8.1 and 8.7.","marker":"[Cao17]"},{"why":"Provides local freeness of higher direct images for proper Kähler morphisms, used in Lemma 8.9 to identify $R^1f_*\\mathcal{O}_X$ and glue the local $f$-ample line bundles.","marker":"[Tak95]"}],"fun_headline_variants":["Kähler uniruled iff canonical bundle is not pseudoeffective","Uniruled Kähler manifolds: canonical bundle must fail to be pseudoeffective","Equivalence proven: uniruled Kähler iff canonical bundle not pseudoeffective","New criterion: Kähler uniruled iff canonical bundle not pseudoeffective","Hard direction solved: Kähler uniruled iff canonical bundle not pseudoeffective"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the germ $S_0$ extends formally along $C$, equivalently that it survives infinitely many blowups, and that for the analytic graph of a foliation this formal extension is guaranteed by Lemma 5.6; if that property failed in any needed case, the blowup cascade and the Lelong-number contradiction that prove Theorems 1.2 and 1.3 would collapse.","fun_headline_variants_meta":{"raw":{"variants":["Kähler uniruled iff canonical bundle is not pseudoeffective","Uniruled Kähler manifolds: canonical bundle must fail to be pseudoeffective","Equivalence proven: uniruled Kähler iff canonical bundle not pseudoeffective","New criterion: Kähler uniruled iff canonical bundle not pseudoeffective","Hard direction solved: Kähler uniruled iff canonical bundle not pseudoeffective"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001253,"raw_usage":{"total_tokens":5088,"prompt_tokens":847,"completion_tokens":4241,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":463,"completion_tokens_details":{"reasoning_tokens":4123}},"tokens_in":463,"tokens_out":4241,"duration_ms":34712,"temperature":1.0,"reasoning_tokens":4123,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T00:43:01.198772+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A compact Kähler manifold $X$ with $\\omega_X$ not pseudoeffective that is not dominated by rational curves would refute Theorem 1.1, and the paper's final contradiction argument is precisely an attempt to rule out such an object.","supporting_citations":[],"review_version":1}