{"id":"3c822b48-75d4-42cf-bdf1-6317b610061e","arxiv_id":"2602.03332","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Line bundles with singular positive Hermitian metrics have vanishing higher cohomology on weakly pseudoconvex Kähler complex spaces, with analogous vanishing on compact complex spaces carrying big line bundles.","lead":"This paper extends the L2-Dolbeault toolkit and Nadel's vanishing theorem from smooth complex manifolds to singular complex spaces equipped with singular Hermitian metrics. The result gives a route to proving cohomology vanishing on weakly pseudoconvex and compact non-Kähler spaces, including a Kawamata-Viehweg-type statement on Moishezon spaces.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Main vanishing theorem rests on an unverified same-author approximation theorem [Wat24, Thm 3.2]; if it fails, the Stein-complement construction and global L2 estimates collapse.","rationale":"The paper is internally coherent and makes a serious contribution to L2-Dolbeault theory on singular spaces. Theorems 1.1–1.5 are connected by a clear chain: Theorem 1.1 (resolution) uses Theorem 1.4 (global L2 estimates) near singular points; Theorem 1.4 uses Theorem 4.4 (Stein complements); Theorem 4.4 uses a refined Demailly-type approximation and a positivity construction that are imported from [Wat24]. The reader's weakest assumption identifies exactly this dependence, and my reading agrees: no later theorem supplies an alternative route, and the proof of Theorem 4.4 does not include the needed approximation theorem. The concern is not that the author is unreliable, but that the central claim is conditional on an unrefereed result that is not reproduced in the paper. A complete independent verification of [Wat24, Thm 3.2 and 3.5] would settle the concern; if it succeeds, the vanishing theorem should be accepted, and if it fails, the main theorem would need a different proof or a corrected statement. Therefore the reader's CONDITIONAL verdict is appropriate, and no adjustment is needed.","tokens_in":25378,"tokens_out":30892,"duration_ms":284035,"concrete_test":"Independently re-derive [Wat24, Thm 3.2 and 3.5] in the exact setting used by Thm 4.4: X weakly pseudoconvex, \\tilde h_j a singular positive metric on π*L over \\tilde X_{c_j+τ_j}. Specifically, prove that there is a sequence of approximations H_ν to \\tilde h_j with I(H_ν)=I(\\tilde h_j) on \\tilde X_{c_j+τ_j} and that after blowing up the singular locus of one H_ν0, the constructed line bundle \\mathcal L on X̂_c is positive and satisfies the hypotheses of [Tak98, Thm 1.2] (weakly 1-complete Kähler). If this cannot be proved from published results, or if a counterexample arises (e.g., a singular positive metric on a non-compact weakly pseudoconvex manifold for which every Demailly-type approximation changes the multiplier ideal sheaf), then the Stein-complement construction in Thm 4.4 and the global L2 estimates in Thm 1.4 lack support, and Thm 1.5 remains conditional.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central vanishing theorem (Thm 1.5) is reached through Thm 1.4 (global L2 estimates on Xreg), whose proof depends on Thm 4.4: one must find analytic subsets A_j so that (X_{c_j})_reg \\ A_j is Stein. The proof of Thm 4.4 is the hinge: it takes a Demailly-type approximation H of the pulled-back singular positive metric \\tilde h_j on the desingularized sublevel set \\tilde X_{c_j+τ_j}, preserving multiplier ideal sheaves, blows up the analytic singular locus of H, and then applies Takayama's embedding theorem. The existence of such an approximation and of the positive line bundle on the blow-up is imported verbatim from [Wat24, Thm 3.2 and 3.5], an unpublished, same-author arXiv preprint. No proof or statement of those results is included in this paper. If [Wat24, Thm 3.2] fails for general singular Hermitian metrics on weakly pseudoconvex manifolds—or if the positive line bundle it produces is not sufficiently positive for Takayama's theorem—then the Stein-complement sets S_j are not obtained, the L2 estimates of Thm 1.4 have no proof, and Thm 1.5 (as well as the exactness of the L2-Dolbeault resolution at singular points in Thm 1.1) collapses. This is a load-bearing dependence on an unrefereed preprint, not a peripheral citation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops an $L^2$-theory for the $ar\\partial$-operator on reduced complex spaces with singular Hermitian metrics. It proves that the $L^2$-Dolbeault complex is a fine resolution of $\\omega_X^{\\rm GR}(h)\\otimes\\mathcal O_X(L)$ (Theorem 1.1), gives a desingularized version and higher direct image vanishing (Theorems 1.2 and 1.3), establishes global $L^2$-estimates on the regular locus of weakly pseudoconvex Kähler spaces (Theorem 1.4), and derives Nadel-type vanishing theorems (Theorem 1.5) as well as compact and Moishezon variants (Theorems 1.6, 7.3 and Corollary 1.7). The proofs combine resolution of singularities, the strong openness property, Demailly-type approximation, and Stein-complement constructions. The main theorems are not circularly assumed, but the proof chain imports essential results from the author's earlier preprints, especially [Wat24, Thm 3.2/3.5] in Theorem 4.4.","tokens_in":25736,"tokens_out":16146,"duration_ms":169320,"significance":"If the results hold, they constitute a substantial generalization of the classical Hörmander–Nadel theory to singular complex spaces with singular Hermitian metrics, going beyond Ruppenthal's smooth-metric framework. The L2-Dolbeault fine resolution theorem and the global L2 estimates would be valuable tools in complex analytic and algebraic geometry. The manuscript is readable, explicitly states limitations, and avoids fitted parameters or hidden assumptions of the main theorems. However, the central chain is not self-contained: several load-bearing stones are imported from unpublished or very recent same-author preprints. In particular, the Stein-complement construction of Theorem 4.4 depends on [Wat24, Thm 3.2 and 3.5], and the proof of Theorem 6.2 depends on an unproved lemma cited to [Wat25b]. These dependencies do not by themselves prove the claims false, but they make this referee unable to certify the central result without further author-supplied verification.","major_comments":[{"comment":"The proof of Theorem 4.4 is the hinge for the global L2-estimates (Theorem 1.4) and hence for Theorem 1.5. The existence of the Demailly-type approximation $\\{H^{(j)}_\\nu\\}$ preserving multiplier ideal sheaves and the associated positive line bundle on the blow-up are quoted verbatim from [Wat24, Thm 3.2 and 3.5], an unpublished same-author preprint. No statement of these results or verification of their hypotheses in the present singular/non-normal setting is given. If these results fail for singular Hermitian metrics on weakly pseudoconvex spaces, the Stein-complement sets $S_j$ are not constructed and the global $L^2$-estimates, and with them the main vanishing theorem, collapse. The manuscript's own Remark 7.5 acknowledges this dependence. The referee requests that the full statements be included and either proved in an appendix or replaced by a refereed reference.","section":"§4, Theorem 4.4; §5, Theorem 1.4"},{"comment":"Lemma 4.3, which supplies the function $\\psi$ that is strictly plurisubharmonic near the exceptional divisor, is used in the proofs of Theorem 4.2 and Theorem 6.2, but it is cited to [Wat25b, Lemma 3.2] (another unpublished preprint) and not proved here. Although the statement appears plausible from the Negativity Lemma 2.2, it is a genuine ingredient in the higher direct image vanishing and in the exhaustion singular-positivity construction. Please include a proof of Lemma 4.3 or a detailed derivation from Lemma 2.2.","section":"§4, Lemma 4.3; §6, Theorem 6.2"},{"comment":"For a non-normal compact complex space, the proof of Theorem 1.6 constructs a metric $h$ with $h=0$ on $X_{\\rm sing}$ and with positivity $i\\Theta_{L,h}\\ge\\gamma$ only on $X_{\\rm reg}$. The vanishing is then deduced from Theorem 1.2 (or Theorem 7.1), whose hypothesis is that the local weight functions of $h$ are quasi-plurisubharmonic on $X$. The proof does not establish this extension across $X_{\\rm sing}$ when $X$ is not normal; the statement itself separates the normal case, suggesting that the extension is not automatic. Consequently, the claim of vanishing \"without assuming conditions on singularities\" is not fully justified as written. Please either prove that the pushforward weight extends qpsh across the singular locus in the non-normal case, or weaken the statement/conditions accordingly.","section":"§7, Theorem 1.6 and Corollary 1.7"}],"minor_comments":[{"comment":"There are numerous typos and grammatical slips: \"resolusion\" for \"resolution\", \"met rich\" in Lemma 5.2, \"choosen\", \"fiexd\", \"subet\", among others. A careful proofreading pass is needed.","section":"Throughout"},{"comment":"The notation is inconsistent: the text defines $H_X$ but then uses \"$HZ$\" in the last display; also $ZX := \\pi(Z)$ should be clarified as $Z_X$. Please correct.","section":"§4, Theorem 4.4 proof"},{"comment":"The proof uses the same symbol $\\varphi$ for both the test form and the weight function of $h$. This makes the integration formula hard to follow. Please rename the test form.","section":"§3, Proposition 3.1 proof"},{"comment":"The term \"$\\pi$-ample\" is used without definition. Since the paper addresses non-compact spaces, a brief definition or reference would help the reader.","section":"§2.1, Lemma 2.2"},{"comment":"Several bibliography entries (e.g., [HPS18], [Kod54]) are not cited in the body of the text; conversely, the unpublished preprints [Wat24] and [Wat25b] are cited for load-bearing results. It would be helpful to mark which references are the author's own preprints and to provide the arXiv numbers in the reference list.","section":"References"},{"comment":"In applying Theorem 1.4 to the metric $h e^{-\\chi\\circ\\Psi}$, the notation $\\varepsilon$ is kept unchanged although the positive function appearing in the curvature bound for the new metric may be different. This is harmless for the qualitative vanishing but should be stated explicitly.","section":"§7.2, Theorem 7.4 proof"}],"recommendation":"major_revision","confidential_remarks":"The paper is coherent and the main line of reasoning is transparent, but the referee is not able to certify the central theorems because of the heavy reliance on the author's unpublished preprints, in particular [Wat24, Thm 3.2/3.5] in Theorem 4.4. This is not a question of novelty but of verifiability: the refereed status of the quoted results is essential. I would also flag the non-normal case of Theorem 1.6/Corollary 1.7 as a proof gap that needs to be addressed in revision. In my view the paper merits major revision rather than rejection, provided the author either supplies the missing statements/proofs or restricts the claims to a setting where all dependencies are refereed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing: this is not a crank paper. It is a coherent, well-written attack on a real open problem, and it identifies a genuine gap in Demailly's Remark 5.17. The main new results are Theorems 1.1 and 1.2: L2-Dolbeault fine resolutions and isomorphisms for singular Hermitian metrics on reduced complex spaces, extending Ruppenthal's smooth-metric results. The Nadel-type vanishing (Theorem 1.5) and the compact/Moishezon corollaries follow naturally and are attractive.\n\nThe proof strategy is sensible: reduce to the smooth case via a canonical desingularization, get local L2-existence, then patch globally using Stein complements. The paper is honest about its limitations—Remarks 5.4, 5.5, 7.5, and 7.6 flag the places where existing tools do not directly apply, and the discussion of why Demailly's remark is nontrivial is fair.\n\nThe soft spot is exactly where the reader's report places it. Theorem 1.5 rests on Theorem 1.4, and Theorem 1.4 rests on Theorem 4.4: for each sublevel set, you need analytic subsets A_j such that (X_cj)_reg \\ A_j is Stein. The proof of Theorem 4.4 imports, verbatim, a Demailly-type approximation result and the existence of a positive line bundle on a blow-up from [Wat24, Theorems 3.2 and 3.5]. This is the same author's unpublished arXiv preprint. Nothing in this paper states or proves those results. If [Wat24] has a gap—say it does not apply to general weakly pseudoconvex manifolds with singular Hermitian metrics, or the line bundle's positivity is too weak for Takayama's embedding—then the Stein complements are not obtained, the global L2 estimates fail, and Theorem 1.5 falls. That is a load-bearing dependence, not a peripheral citation.\n\nI don't see a direct internal contradiction, and I don't see circularity: the main theorems are not assumed in the proofs that import [Wat24]. The other imported result, [Wat25a, Thm 5.3], is in a refereed journal, so lower risk. The citation pattern is heavy self-citation, but that is not itself a flaw; the flaw is that the key external fact has not been independently verified.\n\nBottom line: this paper deserves a serious referee. The right review request is to check [Wat24, Thm 3.2 and 3.5] carefully and whether they apply to the singular positive metrics used here. If they hold, the paper is a real contribution and should be accepted after the usual detail-checking. If not, it needs substantial repair. I'd bring it to reading group and would cite it with a caveat, but I wouldn't build a proof on it until [Wat24] is sorted.","headline":"This paper fills a real gap in the L2-Dolbeault theory of singular Hermitian metrics on singular spaces, but the main vanishing theorem rests on an unverified same-author preprint [Wat24, Thm 3.2/3.5] that must be checked before the central claims are taken as established.","tokens_in":26225,"tokens_out":3632,"would_cite":true,"duration_ms":51839,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32S20","14F18","32L10","32L20","32J25","32C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"A singular positive metric on a weakly pseudoconvex Kähler complex space kills all higher cohomology of the canonical twist, via a new L2-Dolbeault fine resolution.","keywords":["L2-Dolbeault complex","singular Hermitian metrics","complex spaces","L2 estimates","cohomology vanishing","weakly pseudoconvex","multiplier ideal sheaves","bigness"],"falsifier":"Take a weakly pseudoconvex Kähler complex space with a non-normal isolated singularity and a singular positive metric, resolve the singularity, and compute H^1 of the twisted canonical sheaf using the desingularization formula; a nonzero class would refute the main theorem. Alternatively, inspect the companion approximation theorem on that example: if multiplier ideal sheaves are not preserved after refinement, no Stein-complement set can be built and the global L2 estimates lack a proof.","tokens_in":25253,"feed_emoji":"🧮","tokens_out":9406,"duration_ms":97408,"temperature":0.7,"pith_summary":"The paper's goal is to build an L2 ∂-theory for holomorphic line bundles on complex spaces with singularities, where the metric is allowed to be singular. Its central claim is that the sheaves of locally square-integrable (n,q)-forms form a fine resolution of the twisted canonical sheaf, so sheaf cohomology can be computed by L2 Dolbeault cohomology on the regular locus. On a weakly pseudoconvex Kähler space, a singular positive metric then kills all higher cohomology of that twist, and even the first cohomology vanishes without the Kähler assumption. A sympathetic reader would care because this supplies a proof of the long-expected multiplier-ideal vanishing theorem on singular spaces and yields the usual vanishing consequences of positivity, for compact and non-compact singular spaces alike, without imposing smoothness.","feed_headline":"Higher cohomology vanishes on singular Kähler spaces","feed_subtitle":"New L2-Dolbeault resolutions make a singular positive metric enough to kill all higher cohomology.","key_machinery":"The load-bearing object is the L2-Dolbeault complex of (n,q)-forms whose coefficients are locally square-integrable with respect to the singular metric and whose ∂ is also locally square-integrable; exactness of this complex is what makes it a fine resolution. The decisive construction is a Stein-complement argument: after a canonical desingularization, the metric is approximated on each sublevel set while preserving the multiplier ideal sheaf, the approximation's singular locus is blown up to produce a positive line bundle and a projective embedding, and removing a general hyperplane plus exceptional loci leaves a Stein manifold. On that Stein region, standard L2 estimates for ∂ are availab","core_discovery":"On a reduced complex space X of pure dimension n, the paper establishes the following: for a holomorphic line bundle L with a singular Hermitian metric h whose local weights are quasi-plurisubharmonic, the L2-Dolbeault complex of (n,q)-forms with locally L2 coefficients and L2 ∂ is exact and forms a fine resolution of the canonical sheaf of locally square-integrable holomorphic n-forms twisted by L. The same complex also resolves the pullback of that sheaf to a canonical desingularization, so cohomology on X is isomorphic to cohomology on the desingularization with a multiplier ideal sheaf. The main application is that if X is weakly pseudoconvex Kähler and h is singular positive, then all h","pith_inferences":["Editorial inference: the Stein-complement technique should extend to L2-extension and injectivity theorems for singular Hermitian metrics on singular spaces, because the same metric-preserving approximation is the only input needed.","Editorial inference: the non-Kähler case stopping at degree one suggests the Kähler assumption is close to optimal; testing a weakly pseudoconvex non-Kähler example with a singular positive metric could reveal whether H^2 vanishing genuinely fails.","Editorial inference: the explicit 1/q constant and the ε-weighted norm suggest quantitative versions that track how fast the metric's positivity degenerates; such bounds could connect to analytic singularity invariants in concrete examples.","Editorial inference: since weak pseudoconvexity covers complex Lie groups, the vanishing theorem may apply to homogeneous complex spaces and to degenerating families where no global complete Kähler metric exists."],"forward_implications":["On a weakly pseudoconvex Kähler complex space, singular positivity of the metric gives H^q(X, ω^GR_X(h) ⊗ O_X(L)) = 0 for every q > 0; without a Kähler metric, the first cohomology group still vanishes.","The L2-Dolbeault complex is a fine resolution, so the cohomology of the twisted canonical sheaf is computed by L2-Dolbeault cohomology on the regular locus; on compact spaces it is also computed by cohomology on a desingularization with a multiplier ideal sheaf.","On compact complex spaces, a big line bundle forces the space to be bimeromorphically projective and gives the same higher cohomology vanishing without any Kähler assumption; nef and big line bundles then satisfy the standard birational vanishing statement.","Positive line bundles on weakly pseudoconvex complex spaces satisfy the classical vanishing statement for the canonical sheaf, as a direct corollary.","Higher direct images of the desingularized canonical twist vanish, giving a relative version that transfers vanishings between X and its modifications."],"fun_headline_variants":["L2-Dolbeault resolutions extend Nadel vanishing to singular spaces","Singular Hermitian metrics still kill higher cohomology","Weakly pseudoconvex and singular? Cohomology still vanishes","L2 resolutions push Nadel vanishing to weakened settings","Singular Kähler spaces: higher cohomology gone"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The global L2 estimates depend on a refined approximation theorem for singular Hermitian metrics that is stated in a companion preprint, not proved here; if that theorem fails for non-normal singularities or for this class of metrics, the Stein-complement construction and hence the main vanishing theorem collapse.","fun_headline_variants_meta":{"raw":{"variants":["L2-Dolbeault resolutions extend Nadel vanishing to singular spaces","Singular Hermitian metrics still kill higher cohomology","Weakly pseudoconvex and singular? Cohomology still vanishes","L2 resolutions push Nadel vanishing to weakened settings","Singular Kähler spaces: higher cohomology gone"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000279,"raw_usage":{"total_tokens":1430,"prompt_tokens":618,"completion_tokens":812,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":362,"completion_tokens_details":{"reasoning_tokens":726}},"tokens_in":362,"tokens_out":812,"duration_ms":8193,"temperature":1.0,"reasoning_tokens":726,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T05:00:39.665091+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a weakly pseudoconvex Kähler complex space with a non-normal isolated singularity and a singular positive metric, resolve the singularity, and compute H^1 of the twisted canonical sheaf using the desingularization formula; a nonzero class would refute the main theorem. Alternatively, inspect the companion approximation theorem on that example: if multiplier ideal sheaves are not preserved after refinement, no Stein-complement set can be built and the global L2 estimates lack a proof.","supporting_citations":[],"review_version":1}