{"id":"79404779-c8ed-42d5-b882-49cd2a687b4e","arxiv_id":"2506.10515","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper proves effective positivity of Hodge bundles for stable families and derives uniform lower bounds on volumes and automorphism groups, in terms only of dimension and allowed boundary coefficients.","lead":"This mathematics paper proves that certain natural bundles attached to families of stable varieties become positive for an effective, not merely sufficiently large, multiplier q. That effective positivity yields uniform lower bounds on volumes of such families and on their automorphism groups, settling several open boundedness problems in moduli theory.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of Theorem 7.1 as printed is self-referential: inequality (7.1) involves only λ_q and cannot imply bigness, so the dependence of Theorem 8.1 on the effective slope bound needs disambiguation.","rationale":"I read the paper in good faith. The overall architecture is plausible: Theorem 3.18 gives nefness, Theorem 6.1 gives a lower slope bound, and Theorem 8.1 combines it with slope inequalities and known ampleness. The most load-bearing step for the central claim is the passage from the slope bound to bigness of λ_q in Theorem 7.1, and the printed text has an apparent notational or logical problem there. I am not claiming the theorem is false; the ambiguity of 'q, q' in Theorem 6.1 and the self-referential form of (7.1) must be resolved. The reader's concern about reduced fibers is valid for the non-stable statements, but it is not the most limiting issue for Theorem 8.1 because stable families have reduced fibers by definition. My recommendation is CONDITIONAL: the authors should disambiguate the indices in Theorem 6.1 and repair or rewrite the proof of Theorem 7.1, ideally using the direct movable-curve argument via Corollary 6.3. If that repair succeeds, the central claim appears credible; without it, the effective chain is not established as written.","tokens_in":40891,"tokens_out":33471,"duration_ms":403613,"concrete_test":"Re-derive the proof of Theorem 7.1 with two explicit indices q (the target) and Q (the big multiple). Check whether inequality (7.1) can be obtained with deg(g^*λ_Q) on the right when the left bundle is g^*E_q; this would require an embedding λ_q ↪ E_Q^{⊗r}, which is not provided. If no such embedding exists, test the alternative route: for every movable curve class [C], choose a general base-change curve T on which h has maximal variation, apply Corollary 6.3 to conclude λ_q|_T is ample, and verify that this gives λ_q·C > 0 and hence λ_q big. If neither derivation works, Theorem 7.1 and consequently Theorem 8.1 are unproved as written.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 8.1, the central claim, relies on Theorem 7.1 to know that λ_q is big and that λ_q·C ≥ 1 on movable curves. In the proof of Theorem 7.1, a large multiple Q is chosen with λ_Q big, then Theorem 6.1 is applied to a base-changed curve h and inequality (7.1) is recorded as μ_-(g^*E_q) ≥ ε·deg(g^*λ_q). If the two q's here are the same, this inequality is compatible with the already-known nefness of λ_q but proves no bigness: from μ_-(E) ≤ deg det(E)/rk(E) one only obtains deg λ_q ≥ rε deg λ_q, i.e. (1-s)deg λ_q ≥ 0, which is automatic for a nef divisor. The sentence 'In particular, (λ_q − ελ_q)·[C] ≥ 0' is then a no-op for bigness. If, instead, the intended inequality compares E_q with an external λ_Q, that is a different theorem: the Viehweg product trick in Theorem 6.1 constructs the determinant of the same bundle E_q, not an arbitrary big λ_Q, so the cross-inequality needs a separate proof. Because the bigness of λ_q is used in Theorem 8.1 to conclude λ_q^d ≥ 1 and to control λ_CM against λ_q, this gap is load-bearing. The reducedness issue noted by the reader is real for the non-stable Corollary 6.3, but it is not the main bottleneck for Theorem 8.1, since stable families have reduced fibers and the curve base changes in Theorem 7.1 are stable.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops effective positivity statements for Hodge bundles of families of stable varieties, and derives several boundedness applications. The central technical results are: (i) Theorem 3.18, semipositivity of f_*O_X(q(K_{X/T}+\\Delta)) over a smooth curve when the general fiber is slc; (ii) Theorem 6.1, a lower bound for the smallest Harder-Narasimhan slope of such Hodge bundles in terms of the BGV invariant; (iii) Theorem 7.1, bigness/ampleness of the determinant lambda class \\lambda_q under maximal variation and a klt fiber; and (iv) Theorem 8.1, a uniform lower bound \\lambda_{CM}^d \\ge \\delta^d for stable families of maximal variation with coefficients in a DCC set. From these, the paper derives bounds on relative volumes, automorphism groups, and Chow-Mumford volumes on moduli spaces.","tokens_in":41184,"tokens_out":14323,"duration_ms":165958,"significance":"If the main theorems are correct, the paper would replace several non-effective positivity and boundedness results by effective ones, with consequences for the moduli of stable pairs, algebraically integrable foliations, and automorphism groups. The authors also provide a number of new technical tools: effective base-change statements for Hodge bundles, a careful treatment of stable reduction, and a generalized semipositivity theorem over curves. The dependence on external results ([Fuj18], [HMX14], [BZ16], [CTV23b]) is explicit, and no circularity is apparent. However, the proof of the pivotal Theorem 7.1, on which Theorem 8.1 rests, contains a genuine logical gap, as detailed below.","major_comments":[{"comment":"The proof of Theorem 7.1 selects an integer (denoted q in the text, but evidently intended to be a different auxiliary integer Q) so that \\lambda_Q is big. It then applies Theorem 6.1 to a curve base change and records inequality (7.1): \\mu_-(g^*E_q) \\ge \\varepsilon \\deg(g^*\\lambda_q). Since for any vector bundle one always has \\mu_-(E) \\le \\deg\\det(E)/\\operatorname{rk}(E), this inequality is automatically satisfied for a nef divisor \\lambda_q and cannot imply bigness. The subsequent sentence \"In particular, (\\lambda_q - \\varepsilon\\lambda_q)\\cdot[C] \\ge 0\" is a tautology. To deduce bigness of \\lambda_q one must compare it with an external big class, either the chosen \\lambda_Q or \\lambda_{CM}; the printed argument does not supply such a comparison. The proof of Theorem 8.1 uses Theorem 7.1 to obtain \\deg\\lambda_{q,h}\\ge 1 for the uniform q coming from [HMX14], so this gap is load-bearing for the paper's central claim.","section":"§8, Theorem 8.1"},{"comment":"The uniformity of the constant \\delta depends on Theorem 7.1 being available for the specific integer q chosen via [HMX14, Theorem 1.3] and the finite coefficient set I. Because the current proof of Theorem 7.1 requires first choosing a non-effective large Q with \\lambda_Q big, it cannot establish bigness of \\lambda_q for the uniform q unless the gap described in the previous comment is resolved. In particular, the inequality \\lambda_q^d \\ge 1 used in the last paragraph of the proof of Theorem 8.1 is not justified by the printed arguments.","section":"§6, Theorem 6.1"},{"comment":"The statement of Theorem 6.1 contains a notational ambiguity that directly contributes to the conflation of two different integers in Theorem 7.1: it reads \"Let q, q \\ge 2 be positive integers such that q(K_{X/T}+\\Delta) is a Weil Z-divisor and q(K_{X/T}+\\Delta) is Cartier,\" and then formula (6.1) mixes the two q's without distinction. The theorem should use distinct symbols (e.g., q and \\bar{q}) and explicitly state which of them appears in the factor \\gamma_q and in the subscript of E_q and \\lambda_q. This ambiguity makes it difficult to check the derivation of (7.1) and should be corrected even if the main mathematical argument is repaired.","section":"§7, Theorem 7.1"}],"minor_comments":[{"comment":"The abstract contains the duplicated phrase \"several several\".","section":"Abstract"},{"comment":"There are several typos: \"a-pirori\" should be \"a priori\", \"semmi-log canonical\" should be \"semi-log canonical\".","section":"§1.1, §1.2"},{"comment":"The spelling \"Bermann-Gibbs\" appears; the standard spelling is \"Berman-Gibbs\".","section":"§5"},{"comment":"The sentence \"Observe that both \\lambda_q and \\lambda_q are Cartier by construction\" presumably refers to \\lambda_q and \\lambda_Q; the two symbols should be distinguished.","section":"§7, proof of Theorem 7.1"},{"comment":"In the proof of Corollary 8.2, the sentence \"\\lambda_q is nef and big by Corollary 3.26 and Theorem 8.1\" should refer to Theorem 7.1 for bigness, since Theorem 8.1 concerns \\lambda_{CM}.","section":"§8, Corollary 8.2"},{"comment":"In Theorem 11.1 the phrase \"over a a smooth\" should read \"over a smooth\".","section":"§11"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a substantial amount of correct and useful material, in particular Theorem 3.18 and the curve-base parts of Theorem 6.1, but the proof of Theorem 7.1 as printed is not valid, and Theorem 8.1 depends on it. The gap is identifiable and likely repairable, but the repair is not merely cosmetic: it requires either a different argument to compare \\lambda_q with \\lambda_{CM} (needing a uniform positive lower bound for the BGV invariant along B_{\\mathrm{klt}}) or a weakening of the claimed uniform statement. The authors should be asked to provide a corrected proof or to restate the theorems at the level that the proof actually supports."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline: this is a serious contribution, but don't send it to the printer yet. The main results—effective nefness of q-th Hodge bundles over curves, the Harder–Narasimhan slope bound, and the uniform Chow–Mumford volume lower bound—are genuinely new and important. The authors tie together several open boundedness questions in moduli theory, and the technical engine (Theorem 3.18, extending Fujino's semipositivity to all q with slc general fiber) looks solid. The proofs are detailed and mostly standard reductions to Hodge-theoretic semipositivity and vanishing; I see no circularity or fitted parameters. The acknowledgment of a previous mistake, fixed after X. Lü's input, is a good sign.\n\nThe soft spot is Theorem 7.1. As printed, the proof picks a large q (call it Q) with λ_Q big, then applies Theorem 6.1 to a base-changed curve and writes μ_-(g^*E_q) ≥ ε deg(g^*λ_q). If the two q's are the same (as the notation says), this gives (1-ε)λ_q·C ≥ 0, which is automatic for a nef divisor and proves no bigness. If the intended inequality was μ_-(E_q) ≥ ε deg λ_Q, that's a different theorem not established by Theorem 6.1, whose right-hand side is always the determinant of the same bundle. So the printed argument for λ_q big does not work. This matters because Theorem 8.1 uses the bigness of λ_q to conclude λ_CM^d ≥ 1. I think the gap is patchable—Corollary 6.3 plus the movable-cone duality should give λ_q big for any movable curve, since the moduli map is generically finite and pulls back an ample class to a big divisor, so any movable curve has positive degree—but the patch isn't in the paper, and the reference to Theorem 6.1 is misleading. The reducedness issue flagged in the reader's report is real for Corollary 6.3, but for stable families it's not the bottleneck.\n\nWho's this for: moduli theorists and birational geometers working on KSB-stable families, Hodge bundles, and boundedness. It deserves a serious referee—yes—but the referee should insist on a repair of Theorem 7.1's proof and a cleanup of notation (there are also typos like 'q, q' in Theorem 6.1 and 'several several' in the abstract). If the gap is fixed, it's a very good paper.","headline":"A serious, likely-correct paper whose printed proof of Theorem 7.1 has a genuine gap; worth peer review, but the referee should demand a repaired argument.","tokens_in":41800,"tokens_out":8946,"would_cite":true,"duration_ms":99972,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14J10","14E30","14D22","32M25"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that stable families of maximal variation have a uniform positive lower bound on their Chow-Mumford volume, with the bound depending only on the relative dimension and the allowed boundary coefficients.","keywords":["Hodge bundles","positivity","stable varieties","Chow-Mumford volume","Harder-Narasimhan slope","moduli spaces","log canonical pairs","effective boundedness"],"falsifier":"Construct a stable family over a curve with maximal variation, relative dimension n, one klt fiber, and coefficient set Λ, and compute (K_{X/B}+Δ)^{n+1}; if it is smaller than the paper's δ(n,Λ)=$q^{{-n-1}}$ for the q chosen from the effective birationality bound, the claimed uniformity fails. Equivalently, compute the smallest Harder-Narasimhan slope μ_-(E_q) for such a family: Theorem 6.1 predicts it is at least min{(q−1)/q, qγ_q/(1+γ_q q)}·deg λ_q/rk E_q, so a family with deg λ_q >0 and μ_-(E_q)=0 would falsify the slope bound.","tokens_in":40644,"feed_emoji":"📐","tokens_out":9043,"duration_ms":93173,"temperature":0.7,"pith_summary":"The paper's central claim is that positivity of Hodge bundles can be made effective in a uniform way for stable families of varieties, and that this single input yields new boundedness results in several areas. For each relative dimension $n$ and each DCC set of boundary coefficients $\\Lambda$, the authors prove there is a constant $\\delta(n,\\Lambda)>0$ such that every stable family $f:(X,\\Delta)\\to B$ of maximal variation with at least one klt fiber satisfies $(\\lambda_{CM})^{\\dim B}\\ge \\delta^{\\dim B}$, where $\\lambda_{CM}$ is the Chow-Mumford line bundle. When $B$ is a curve this gives a uniform lower bound $(K_{X/B}+\\Delta)^{n+1}\\ge \\delta$, and it also bounds the order of the relative automorphism group linearly in that volume. The engine is the $q$-th Hodge bundle $f_*\\mathcal{O}_X(q(K_{X/B}+\\Delta))$: the paper proves it is nef over a curve for every positive integer $q$, and gives an effective lower bound on its smallest Harder-Narasimhan slope, which then feeds a slope inequality to control volumes. A reader should care because these were previously known only non-effectively, so the result converts qualitative boundedness into dimension-by-dimension constants.","feed_headline":"Every stable family has Chow-Mumford volume at least δ^d","feed_subtitle":"One constant, fixed by dimension and boundary coefficients, controls Hodge-bundle positivity, volumes, and automorphism groups.","key_machinery":"The central object is the $q$-th Hodge bundle $E^{\\Delta}_{q,f}=f_*\\mathcal{O}_X(q(K_{X/B}+\\Delta))$, whose determinant is the $q$-th $\\lambda$ class $\\lambda_q$; the Chow-Mumford line bundle is recovered from these as $\\lambda_{CM}=(n+1)!\\lim_{q\\to\\infty}\\lambda_q/q^{n+1}$. The two pillars are Theorem 3.18, which proves $E^{\\Delta}_{q,f}$ is nef on a smooth projective curve for every $q$ when the general fiber is slc, and Theorem 6.1, which bounds from below the smallest Harder-Narasimhan slope $\\mu_-(E_q)$ by $\\min\\{(q-1)/q,\\ q\\gamma_q/(1+\\gamma_q q)\\}\\cdot \\deg\\lambda_q/\\operatorname{rk}E_q$, with $\\gamma_q$ the Bermann-Gibbs-Viehweg log canonical threshold invariant of the general fiber. The lower-bound proof runs the Viehweg product trick: it embeds $\\lambda_q$ into the $r$-fold tensor power of $E_q$, uses the BGV divisor to keep the fiber product klt, and applies a semipositivity criterion for $f_*\\mathcal{O}(L)$ that requires all fibers to be reduced so that the self-products are normal. Combining the effective slope bound with the slope inequality $L^{n+1}\\ge \\deg f_*\\mathcal{O}(L)$ gives the uniform volume bounds.","core_discovery":"The paper establishes, on its own terms, the following theorem: fix a positive integer $n$ and a DCC set $\\Lambda\\subset \\mathbb{Q}\\cap[0,1]$; then there is $\\delta=\\delta(n,\\Lambda)>0$ such that $\\lambda_{CM}^{d}\\ge \\delta^{d}$ for every stable family $f:(X,\\Delta)\\to B$ of relative dimension $n$ and maximal variation whose boundary coefficients lie in $\\Lambda$ and which has at least one klt fiber. The proof's load-bearing new input is an effective positivity statement for Hodge bundles: if the base is a smooth projective curve and the general fiber is slc, then $f_*\\mathcal{O}_X(q(K_{X/T}+\\Delta))$ is nef for every $q\\ge 1$, and under maximal variation with klt general fiber and $q\\ge 2$ it is ample whenever nonzero. From the curve case the paper derives the general base by pulling back along movable curves and using a slope inequality to convert the slope bound into an inequality of nef divisors. The same circle of ideas yields a uniform lower bound on $(K_{X/B}+\\Delta)^{n+1}$ for curve bases, an effective positivity of $\\lambda_q$ on normalizations of moduli spaces of stable pairs, and an upper bound $|\\operatorname{Aut}(f)|\\le C\\cdot \\operatorname{vol}(K_{X/B}+\\Delta)$ for fibrations over curves.","pith_inferences":["If the reduced-fiber assumption in Theorem 6.1 could be removed, the uniform volume bound would hold without passing through stable reduction, which currently replaces equalities of Hodge bundles by generically isomorphic inclusions; the paper's Corollary 10.2 suggests the slope bound itself survives non-reduced fibers.","The method predicts that an affirmative answer to the paper's Question 1.10 on effective ampleness for corank-one foliations would extend the volume lower bounds to a much larger class of foliations, since the only missing ingredient is the analogue of Theorem 3.18 for the foliated canonical bundle.","The effective positivity of $\\lambda_q$ on moduli normalizations should make the boundary of the ample cone described in Theorem 11.2 computable in examples, potentially yielding explicit birational models of moduli spaces of stable pairs.","A testable consequence is that the uniform $\\delta$ in Theorem 8.1 should be computable from the effective birationality constant $q(n,\\Lambda)$ together with the BGV invariant of the general fiber, so explicit constants could be extracted for surfaces and threefolds."],"forward_implications":["If Theorem 8.1 is correct, then the Chow-Mumford volume of every maximally varying stable family of fixed relative dimension and coefficient set is bounded below by a constant depending only on $n$ and $\\Lambda$, making effective the previously non-effective boundedness results for such families.","Over a curve base, the relative volume $(K_{X/B}+\\Delta)^{n+1}$ is uniformly bounded below, which gives a uniform lower bound for volumes of algebraically integrable foliations induced by fibrations with reduced fibers.","The $q$-th Hodge bundle is nef over a curve for every $q$ and ample for $q\\ge 2$ under maximal variation and a klt general fiber, so the Hodge and lambda classes are effective before any divisibility that was previously required.","On normalizations of moduli spaces of stable pairs containing at least one klt pair, $\\lambda_q$ is nef whenever $qI\\subset \\mathbb{Z}$ and big for $q\\ge 2$, and the Chow-Mumford divisor has volume at least $(vC)^{-\\dim M}$ in terms of the volume $v$ of the parametrized pairs.","For fibrations over a curve, $|\\operatorname{Aut}(f)|\\le \\delta\\cdot\\operatorname{vol}(K_{X/B}+\\Delta)$ with $\\delta$ depending only on $n$ and $\\Lambda$, the relative analogue of the Hurwitz-type bounds for varieties of general type."],"supporting_citations":[{"why":"Supplies the base semipositivity theorem for double semi-snc pairs that Theorem 3.18 bootstraps to singular total spaces and all q.","marker":"[Fuj18]"},{"why":"Provides the non-effective positivity and ampleness of Hodge bundles and the projectivity of moduli that the paper makes effective.","marker":"[KP17]"},{"why":"Shows λ_CM is ample on the coarse moduli space, used to characterize maximal variation and to define the CM volume.","marker":"[PX16]"},{"why":"Gives the slope inequality L^{n+1} ≥ deg f_*O(L) that converts the effective slope bound into volume bounds.","marker":"[CTV23b]"},{"why":"Yields the DCC volume lower bound and effective birationality used to choose the integer q depending only on n and Λ.","marker":"[HMX14]"},{"why":"Proves normality of r-fold fiber self-products when all fibers are reduced, the key hypothesis for the Viehweg product trick.","marker":"[CP21]"},{"why":"Supplies the framework of stable families, stable reduction, and the extension results used throughout.","marker":"[Ko23]"},{"why":"Gives a uniform ε so K_F+εΔ_F is big, allowing reduction from a DCC coefficient set to a finite set.","marker":"[BZ16]"},{"why":"Bounds birational automorphism groups by volume, used for the relative automorphism bound and the moduli volume corollary.","marker":"[HMX13]"}],"fun_headline_variants":["One δ bounds volume, Hodge positivity, and automorphism groups","Effective Hodge positivity yields uniform bounds for stable families","Uniform δ: volume, top self-intersection, and automorphisms bounded","Stable families: volume bound, Hodge nef, and automorphism bound from one δ"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole lower-bound chain presupposes that the r-fold self-product of the total space is normal, which is guaranteed only when all fibers of the family are reduced; when fibers are non-reduced, stable reduction replaces equalities of Hodge bundles by inclusions, and any loss there would weaken the uniform volume constant.","fun_headline_variants_meta":{"raw":{"variants":["One δ bounds volume, Hodge positivity, and automorphism groups","Effective Hodge positivity yields uniform bounds for stable families","Uniform δ: volume, top self-intersection, and automorphisms bounded","Stable families: volume bound, Hodge nef, and automorphism bound from one δ"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000878,"raw_usage":{"total_tokens":3881,"prompt_tokens":1114,"completion_tokens":2767,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":730,"completion_tokens_details":{"reasoning_tokens":2686}},"tokens_in":730,"tokens_out":2767,"duration_ms":22589,"temperature":1.0,"reasoning_tokens":2686,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:24:19.234157+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a stable family over a curve with maximal variation, relative dimension n, one klt fiber, and coefficient set Λ, and compute (K_{X/B}+Δ)^{n+1}; if it is smaller than the paper's δ(n,Λ)=$q^{{-n-1}}$ for the q chosen from the effective birationality bound, the claimed uniformity fails. Equivalently, compute the smallest Harder-Narasimhan slope μ_-(E_q) for such a family: Theorem 6.1 predicts it is at least min{(q−1)/q, qγ_q/(1+γ_q q)}·deg λ_q/rk E_q, so a family with deg λ_q >0 and μ_-(E_q)=0 would falsify the slope bound.","supporting_citations":[],"review_version":1}