{"id":"b0269615-f66a-4360-918e-048b1856df0f","arxiv_id":"2606.16275","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Nakano-positive singular Hermitian vector-bundle metrics on weakly pseudoconvex complex spaces imply vanishing of higher cohomology of the associated Grauert-Riemenschneider L2-canonical sheaf.","lead":"On singular complex spaces, this paper defines Griffiths and Nakano positivity for singular Hermitian metrics on holomorphic vector bundles and proves that the associated L2-Dolbeault complex resolves the Grauert-Riemenschneider L2-canonical sheaf. This yields Nakano-Nadel-type cohomology vanishing on weakly pseudoconvex complex spaces, including compact cases where no Kähler metric is assumed.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.12 rests on imported Stein-exhaustion Theorem 5.1; if [Wat26b, Thm 4.4] does not provide an exhausting sequence with c_j→∞, the global L2-existence proof collapses.","rationale":"The reader identified Theorem 5.1 as the weakest assumption, and I agree. The central proof chain is: Theorem 7.6 uses the L2-Dolbeault isomorphism from Theorem 1.8 and the global L2-existence Theorem 5.3. Theorem 5.3 is proved entirely on the basis of Theorem 5.1, which is imported from another preprint and not proved here. No internal contradiction or circularity appears in the steps from Theorem 5.3 to the vanishing: the application to S_j is legitimate provided S_j is a Stein coordinate with a trivialization, and the L2 estimate and weak limit are standard. The condition that X admits a singular positive line bundle is used only to invoke Theorem 5.1; if this theorem is not available, the proof of Theorem 5.3 has no foundation. I found no stronger internal flaw in the vanishing proof itself. I also noted the assertion 'Since Sreg is Stein' in the converse direction of Theorem 4.12 is questionable (e.g., for a singular Stein cone, the regular locus minus the vertex is not Stein), but this does not affect the proof of Theorem 1.12, which uses Theorem 4.12 only in the forward direction (if at all). Therefore the reader's CONDITIONAL verdict is appropriate, and no verdict change is needed.","tokens_in":39576,"tokens_out":31792,"duration_ms":330704,"concrete_test":"Extract the proof of [Wat26b, Theorem 4.4] and verify two specific assertions: (1) the sequence c_j can be chosen with c_j → +∞, i.e., the X_{c_j} exhaust X; (2) for each j, A_j is a closed analytic subset of X_{c_j} and (X_{c_j})_reg \\ A_j is Stein. If either assertion is missing or relies on an extra hypothesis (e.g., normality or local boundedness), then re-run the weak-limit argument in Theorem 5.3: without an exhausting sequence of Stein subsets, the subsequence u_j cannot converge to a global u on X, so Theorem 1.11 and hence Theorem 1.12 do not follow.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central Nakano-Nadel vanishing theorem (Theorem 7.6) is proved by combining the L2-Dolbeault resolution Theorem 1.8 with the global L2-existence Theorem 5.3. The proof of Theorem 5.3 applies the definition of θ-Nakano positivity on the open subsets S_j = ((X_{c_j})_reg \\ A_j) \\ H_j, relying on the assertion that these are Stein manifolds. That assertion is exactly Theorem 5.1, imported without proof from the same-author preprint [Wat26b, Theorem 4.4]. The paper's own text does not establish this structural premise, and it is not a consequence of the vanishing claim. Moreover, the statement as written in §5 only says the c_j form an increasing sequence, not that c_j → +∞; without an exhausting sequence, the weak-limit step in Theorem 5.3 cannot produce a solution on all of X. The analytic subsets A_j are also not specified to be closed in X nor to have any particular codimension, and Stein-ness of the complement is a strong global condition that is not derived here. If Theorem 5.1 fails or is weaker than stated, Theorem 5.3 has no basis, and therefore the global ∂-solution used to prove H^q(X, ω_GR(E,h)) = 0 does not exist. This is a load-bearing external dependency, not an internal contradiction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a theory of singular Griffiths and Nakano positivity for Hermitian metrics on holomorphic vector bundles over reduced complex spaces of pure dimension. It introduces definitions via local plurisubharmonicity and L2-estimates, proves that these notions behave well under resolution of singularities, and shows that Griffiths positivity implies Nakano positivity after twisting by the determinant (Theorem 4.13). It then constructs L2-Dolbeault fine resolutions of the Grauert-Riemenschneider canonical L2-sheaf (Theorem 1.8) and proves a global L2-existence theorem on weakly pseudoconvex Kähler complex spaces (Theorem 1.11). The main application is a Nakano-Nadel vanishing theorem (Theorem 1.12/7.6): under a weak pseudoconvexity and positivity assumption, H^q(X, ω^GR_X(E,h)) = 0 for q > 0, with H^1 vanishing in the non-Kähler case. The approach reduces to known results on manifolds via desingularization, but key structural inputs are imported from several same-author preprints without proofs.","tokens_in":39871,"tokens_out":6419,"duration_ms":67331,"significance":"If the main theorems are correct, the paper would establish that the Hörmander–Andreotti–Vesentini L2 machinery works for singular vector-bundle metrics on singular spaces, substantially generalizing previous line-bundle and smooth-metric results (Ruppenthal, Shentu–Zhao, Inayama, Watanabe). The paper is honest about its limitations (Remarks 3.2, 7.7) and contains several genuinely new technical results, such as the pullback equivalence of Nakano positivity (Theorem 4.12) and the L2-resolution theorem. The strengths include careful treatment of currents and plurisubharmonic extension on non-normal spaces, and correct identification of where positivity degenerates along the exceptional divisor. However, the central claims are not independently checkable from the manuscript because of the heavy reliance on unpublished same-author preprints, and one imported theorem is stated insufficiently for its use. The significance would be high if the gaps are closed.","major_comments":[{"comment":"Theorem 5.1 is the load-bearing structural input for the global L2-existence theorem. As stated, it only asserts an increasing sequence {c_j} with c_1 > inf_X Ψ, not that c_j → +∞. The proof of Theorem 5.3 applies the local L2 estimates on S_j = ((X_{c_j})_reg \\ A_j) \\ H_j and then uses [Wat25a, Lemma 3.18] to obtain a weak limit u on all of X. This requires the domains S_j to exhaust X, i.e. c_j → +∞ (or at least ∪_j X_{c_j} = X). Without exhaustiveness the weak-limit step cannot produce a global solution. In addition, the analytic subsets A_j are not specified to be closed in X nor to have any codimension, and 'increasing' alone does not imply exhaustion. The author must either state and prove the full version of [Wat26b, Theorem 4.4] with c_j → +∞, or provide an alternative exhaustion argument. This is not a cosmetic issue: Theorem 1.12 depends on it.","section":"§5, Theorem 5.1 and proof of Theorem 5.3"},{"comment":"Several key steps are imported from same-author preprints that are not proved here and are not otherwise publicly verified: [Wat25c, Theorem 1.1] (Nakano-Nadel vanishing on weakly pseudoconvex manifolds) is used in the proof of Theorem 6.5; [Wat26b, Theorem 4.4] is used for Theorem 5.1; [Wat25b, Lemma 2.2] is used for the Negativity Lemma; [Wat24b, Theorem 3.2] is used in Theorem 7.3. While referencing preprints is acceptable, for a central theorem of this scope the referee cannot verify correctness without seeing these proofs. The burden is on the author to either include the necessary statements with complete proofs in an appendix, or to ensure these results are accepted and publicly available with full details. At minimum, the precise statements (including the exhaustive sequence issue) must be quoted correctly.","section":"§6, Theorem 6.5; §2, Lemma 2.2; §7, Theorem 7.3"},{"comment":"The argument 'the integral ... become convergent if χ grows fast enough' is too terse. For a given f ∈ Γ(X, L^{n,q}_{E,h}), one must show there exists a smooth convex increasing χ such that ∫_X ⟨B^{-1}_{θ,χ∘Ψ,ω} f,f⟩_{h,ω} e^{-χ∘Ψ} dV_ω < +∞. Since f is only locally L2 with respect to h, its growth as Ψ → +∞ is uncontrolled; the existence of such χ is plausible (as in the standard trick of choosing χ depending on f) but requires a proof. This step is essential for applying Theorem 1.11. Also, in the same proof the notation 'L^{n,q}_{L,h}' should be 'L^{n,q}_{E,h}'.","section":"§7.2, proof of Theorem 7.6"}],"minor_comments":[{"comment":"There are numerous typographical errors, most notably 'resolusion' for 'resolution' (e.g., pages 2, 4, 9, 27, 28), 'resulution' on page 4, and inconsistent spelling of 'Kähler' in a few places. The paper would benefit from a careful proofreading pass.","section":"Throughout"},{"comment":"The definition of L1_loc(U) uses integrals over K_reg with respect to a volume measure that is not explicitly defined on a singular space. It would be helpful to state that the measure is induced by a local embedding into C^N, as is done later in §2.4.","section":"§2.4, Definition 2.4 and L1_loc"},{"comment":"In the display after 'there exists u ∈ L^2_{n,q−1}(S_reg \\ H, E; ...)' the equality '= L^2_{n,q}(S,E⊗L; ...)' appears to have a typo: the left-hand side should be L^2_{n,q}(S_reg \\ H, E; ...) rather than L^2_{n,q−1}. The dimension q−1 vs q is inconsistent.","section":"§4, Theorem 4.14(b)"},{"comment":"The definition of exhaustion is incomplete: a function Ψ is an exhaustion if all sublevel sets X_c = {x ∈ X | Ψ(x) < c} are relatively compact, but one usually also requires that the sublevel sets exhaust X as c → +∞, i.e., that Ψ is proper and bounded below. The current wording does not imply that the sequence c_j in Theorem 5.1 is unbounded.","section":"§5, Definition of weak pseudoconvexity"},{"comment":"[Wat25c] is listed as 'preprint' without an arXiv number or publication status; please update if it has been accepted. Similarly, [Wat26b] is an arXiv preprint from 2026; its status should be made clear so that readers can assess the availability of the cited results.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's central theorem depends on a chain of same-author preprints, some of which are explicitly about the same class of problems. While self-citation is common, the editor should consider whether the journal's policy allows a decisive theorem to rest on unpublished, not-yet-refereed work. The specific gap in Theorem 5.1 (missing c_j → +∞) is easily fixable by restating or proving the result, but it currently blocks the proof of Theorem 5.3 and hence Theorem 1.12. I recommend asking the author to make the cited preprints accessible, include the precise statements, and possibly add an appendix with proofs of the most load-bearing imported results."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new content here is the definition of singular Griffiths/Nakano positivity on complex spaces and the pullback equivalences (Theorems 1.1–1.3), then the L2-Dolbeault resolution for general singular Hermitian metrics on vector bundles over singular spaces (Theorem 1.8). Previous work covered smooth metrics or line-bundle-type singularities, so this is a real step forward. The paper also shows that the Nakano–Nadel vanishing framework extends to this generality. The proofs are long and careful, and the author is honest about unclear points (Remarks 3.2 and 7.7). That honesty earns credit.\n\nThe soft spots are mostly about verification. The main vanishing theorem (1.12) depends on Theorem 5.1, imported without proof from the author's own preprint [Wat26b, Thm 4.4]. The statement as printed says the c_j are increasing and c_1 > inf Ψ, but it does not say c_j → +∞. The proof of Theorem 5.3 needs an exhaustion to take the weak limit over all of X. If [Wat26b] actually proves the stronger statement, fine; if not, the global L2-existence theorem has no basis. That is not an internal contradiction, but it is a load-bearing external dependency, and it should be stated precisely. The other dependencies on [Wat25c] and [Wat25b] are similar, though less alarming because they are quoted as known theorems in the field. There is no formal verification, and the analytic details are too intricate for a referee to fully check from the text alone.\n\nOverall: this deserves a serious referee. The framework is promising and the theorems fill a real gap in the literature. The referee should be asked to verify Theorem 5.1 in the cited preprint and check whether the exhaustion condition holds. If it does, this is a substantial advance. If not, the main vanishing result needs reworking. I would not desk-reject; I would send it out.\n\nThe paper is for anyone working on L2 methods on singular spaces, multiplier ideal sheaves, or vanishing theorems. I would likely cite it once the external dependencies are checked.\n\nRecommendation: engage. Send to an expert in L2 theory on singular complex spaces, with explicit instruction to check the imported Theorem 5.1 and the c_j → ∞ point.","headline":"Serious L2 theory on singular spaces; main theorems plausible, but they lean on same-author preprints and one stated theorem misses the exhaustion condition it needs.","tokens_in":40373,"tokens_out":2043,"would_cite":true,"duration_ms":23420,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32S20","14F18","32L10","32L20","32C15","32J25"],"pacs":[],"model":"deepseek-v4-flash","headline":"A Nakano-positive singular Hermitian metric on a weakly pseudoconvex complex space forces all positive-degree cohomology of its Grauert–Riemenschneider L2 canonical sheaf to vanish; without Kähler, H^1 vanishes.","keywords":["L2-estimates","singular Hermitian metrics","Nakano positivity","Griffiths positivity","complex spaces","vanishing theorems","L2-Dolbeault resolution","weakly pseudoconvex"],"falsifier":"Construct a weakly pseudoconvex complex space $X$ with a singular positive line bundle for which some $(X_c)_{\\mathrm{reg}} \\setminus A$ has non-Stein components, or otherwise prove the exhaustion premise fails; this would undercut Theorem 5.3. Alternatively, search for a weakly pseudoconvex Kähler complex space with a singular positive line bundle and a $\\theta$-Nakano positive singular Hermitian metric, plus a $\\partial$-closed $(n,q)$-form $f$ with finite $\\int \\langle B^{-1}f, f\\rangle$ norm, that has no $\\partial$-potential with the stated $L^2$ bound — that would directly falsify the global $L^2$-existence theorem and hence the vanishing theorem.","tokens_in":39409,"feed_emoji":"📐","tokens_out":10845,"duration_ms":89079,"temperature":0.7,"texified_at":"2026-08-05T21:09:19.641642+00:00","pith_summary":"On weakly pseudoconvex complex spaces, the paper establishes a vanishing theorem: a singular Hermitian metric on a holomorphic vector bundle that is Nakano positive in the sense of $L^2$-estimates forces all positive-degree cohomology of its Grauert–Riemenschneider canonical $L^2$-subsheaf to vanish; without a Kähler metric, first cohomology still vanishes. To get there, the paper introduces definitions of singular Griffiths and Nakano positivity appropriate to complex spaces — Nakano positivity is defined through solvability of the $\\partial$-equation with the standard Bochner–Kodaira–Nakano estimate on the regular locus — and proves they behave well under resolution of singularities. Two further tools are built: a global $L^2$-existence theorem on the regular locus of weakly pseudoconvex spaces, and an $L^2$-Dolbeault fine resolution of the canonical $L^2$-subsheaf, which identifies its sheaf cohomology with $L^2$-Dolbeault cohomology. A sympathetic reader would care because this is the first time the standard $L^2$ vanishing machine is made to work for arbitrary singular vector-bundle metrics on singular spaces, not just line-bundle-type singularities or tame metrics.","texify_model":"deepseek-v4-flash","texify_usage":{"total_tokens":5574,"prompt_tokens":965,"completion_tokens":4609,"prompt_tokens_details":{"cached_tokens":0},"prompt_cache_hit_tokens":0,"prompt_cache_miss_tokens":965,"completion_tokens_details":{"reasoning_tokens":3619}},"feed_headline":"All higher cohomology vanishes for Nakano-positive singular metrics","feed_subtitle":"On weakly pseudoconvex complex spaces, positive-degree L2 canonical cohomology vanishes for any singular Nakano-positive metric.","key_machinery":"The machinery has four components. (1) Singular Griffiths/Nakano positivity on complex spaces: Griffiths positivity is defined via log-norm plurisubharmonicity; Nakano positivity ($\\theta$-Nakano positivity 'in the sense of $L^2$-estimates') is defined as a property of the $\\partial$-problem — for every Stein coordinate patch, every $\\partial$-closed form with finite $B^{-1}$-norm admits a $\\partial$-potential with the standard estimate, with $B = [(\\theta + i\\partial\\partial\\psi)\\otimes id_E, \\Lambda_\\omega]$. (2) The Grauert–Riemenschneider canonical $L^2$-subsheaf $\\omega^{GR}_X(E,h)$, whose sections are $L^2$-integrable $(n,0)$-forms on the regular locus; it equals $\\pi_*(K_{\\tilde{X}} \\otimes E(\\pi^*h))$ for any resolution. (3) A global $L^2$-existence theorem (Theorem 5.3) on the regular locus of weakly ps","core_discovery":"On a weakly pseudoconvex complex space of pure dimension that admits a singular positive line bundle, a singular Hermitian metric that is Nakano positive in the $L^2$-estimate sense makes $H^q(X, \\omega^{GR}_X(E,h))$ vanish for all $q>0$; this is the content of Theorem 7.6. The paper's central insight is to formulate Nakano positivity on complex spaces not through curvature, but through the solvability of the $\\partial$-equation with the standard estimate on the regular locus. With that formulation, positivity pulls back to any resolution of singularities, the Grauert–Riemenschneider canonical $L^2$-subsheaf $\\omega^{GR}_X(E,h)$ — defined by $L^2$-integrability of sections on the regular locus — is coherent, and the global $L^2$ $\\partial$-","pith_inferences":["The L2-based definition of Nakano positivity could serve as a template for other analytic positivity notions on singular spaces, since it is insensitive to codimension-one behavior and bypasses the failure of naive curvature pull-backs.","The reliance on the exhaustion premise (Theorem 5.1) suggests a close tie between the vanishing theorem and the existence of 'enough' Stein neighborhoods on weakly pseudoconvex spaces; if that exhaustion fails for some class of spaces, the vanishing might still hold but would require a different route.","One could test the strength of the theorem by checking whether it recovers, on smooth spaces, the known Nakano–Nadel and relative vanishing results, and by seeking examples of singular metrics of line-bundle type where the L2-subsheaf coincides with the classical multiplier ideal sheaf — the theorem then reduces to a familiar statement.","The compact/Moishezon conclusion suggests a converse direction worth exploring: if all higher cohomology of ω^GR_X(E,h) vanishes for some positive singular metric, does X necessarily admit a singular positive line bundle or have a projectivizable desingularization?"],"forward_implications":["For any holomorphic vector bundle with a Nakano-positive singular Hermitian metric on a weakly pseudoconvex Kähler complex space, the higher cohomology groups H^q(X, ω^GR_X(E,h)) vanish for all q>0, giving singular-space analogues of the classical Nakano–Nadel vanishing theorem.","Without a Kähler metric, first cohomology H^1 still vanishes; in the compact case, the Kähler assumption can be dropped entirely and vanishing forces X to be Moishezon.","A Griffiths-positive (and a.e. Griffiths semi-positive) metric yields the same vanishing after twisting by determinant: H^q(X, ω^GR_X(E⊗detE, h⊗deth)) = 0 for q>0.","The L2-Dolbeault complex resolves ω^GR_X(E,h) by fine sheaves, so the structure sheaf's cohomology is computed by the L2 ∂-complex — a new cohomological isomorphism for singular spaces with general singular metrics (beyond line-bundle-type singularities and tame metrics).","Higher direct images R^qπ_*(K_ X̃ ⊗ E(π^*h)) vanish for q>0 under the local Griffiths–Nakano boundedness condition, a relative vanishing theorem for higher direct image sheaves on complex spaces."],"fun_headline_variants":["Vanishing theorem: Nakano-positive singular metrics kill L2 cohomology","Higher L2 cohomology vanishes for Nakano-positive singular metrics","Nakano-Nadel vanishing extended to singular metrics on complex spaces","Positivity via ∂-equation yields vanishing on weakly pseudoconvex spaces","Singular Nakano positivity forces L2 cohomology to vanish in positive degree"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper assumes, as an imported structural premise, that a weakly pseudoconvex complex space admitting a singular positive line bundle can be exhausted by open subsets $(X_{c_j})_{\\mathrm{reg}} \\setminus A_j$ that are Stein manifolds; this is the fact on which the global $L^2$-existence theorem and hence the vanishing theorem rest.","fun_headline_variants_meta":{"raw":{"variants":["Vanishing theorem: Nakano-positive singular metrics kill L2 cohomology","Higher L2 cohomology vanishes for Nakano-positive singular metrics","Nakano-Nadel vanishing extended to singular metrics on complex spaces","Positivity via ∂-equation yields vanishing on weakly pseudoconvex spaces","Singular Nakano positivity forces L2 cohomology to vanish in positive degree"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000772,"raw_usage":{"total_tokens":3206,"prompt_tokens":644,"completion_tokens":2562,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":388,"completion_tokens_details":{"reasoning_tokens":2479}},"tokens_in":388,"tokens_out":2562,"duration_ms":19165,"temperature":1.0,"reasoning_tokens":2479,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T11:11:10.280808+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a weakly pseudoconvex complex space $X$ with a singular positive line bundle for which some $(X_c)_{\\mathrm{reg}} \\setminus A$ has non-Stein components, or otherwise prove the exhaustion premise fails; this would undercut Theorem 5.3. Alternatively, search for a weakly pseudoconvex Kähler complex space with a singular positive line bundle and a $\\theta$-Nakano positive singular Hermitian metric, plus a $\\partial$-closed $(n,q)$-form $f$ with finite $\\int \\langle B^{-1}f, f\\rangle$ norm, that has no $\\partial$-potential with the stated $L^2$ bound — that would directly falsify the global $L^2$-existence theorem and hence the vanishing theorem.","supporting_citations":[],"review_version":1}