{"id":"2c6d23f8-5dd2-47e9-b3da-9e69e115692d","arxiv_id":"2506.00760","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 finiteness of relative log pluricanonical representations for complex analytic spaces and reduces the abundance conjecture for analytic families to the classical abundance conjecture for projective varieties.","lead":"Osamu Fujino proves that relative log pluricanonical representations are finite in the complex analytic setting, and uses this to establish abundance for semi-log canonical pairs and the existence of log canonical flips. The interest for a generalist is that the hardest open problem in the minimal model program, abundance, is reduced to the well-known algebraic case.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central theorem relies on the unpublished complex analytic MMP of [EH2]; Proposition 4.6 Step 1 uses [EH2, Theorem 1.2] to run and terminate an MMP with scaling, and the internal Lemma 2.29 is only sketched. If either fails, Theorem 1.1 is not established.","rationale":"The reader's weakest_assumption identifies the dependence on [EH2] for the complex analytic MMP, and I agree that this is the single most load-bearing concern for the central claim. The paper's own internal Lemma 2.29 is also only sketched and feeds directly into the finiteness theorem (Theorem 3.2) and the gluing lemma (Lemma 4.7), but the external MMP is the more fundamental dependency: without [EH2, Theorem 1.2], Proposition 4.6 cannot be proved, and the whole structure leading to Theorem 1.1 collapses. I also note that the proof of Theorem 1.4 contains a terse reduction to the normal case ('By Theorem 1.1, we may assume that X is normal') that is not explained, but since the central claim is Theorem 1.1, I have kept the main attack on the MMP dependency. The reader's recommendation of conditional acceptance with requests for complete proofs of the relevant lemmas and verification of the external MMP is sound. I do not see an internal inconsistency in the paper's own argument; rather, the risk is that the analytical MMP prerequisites are not yet fully vetted. Thus the verdict should remain UNCHANGED: conditional acceptance, with the conditions being explicit verification of [EH2] and completion of the sketched proofs.","tokens_in":35796,"tokens_out":22475,"duration_ms":215208,"concrete_test":"Check [EH2, Theorem 1.2] in the precise situation of Proposition 4.6: a projective morphism X→Z of complex analytic spaces, a dlt pair (X,Δ), and an MMP with scaling for K_X+Δ−ε⌊Δ⌋ over a Stein compact subset, including termination of the sequence of flips/divisorial contractions. If the proof of [EH2, Theorem 1.2] contains a step that requires algebraicity or a different notion of termination, the reduction in Proposition 4.6 fails. Independently, write out a complete proof of Lemma 2.29 following Claims (A_n) and (B_n) of [Fuj1, Lemma 4.9], checking that the induction and the blowing-up argument work for arbitrary projective bimeromorphic morphisms of complex analytic spaces; if any step needs an algebraic input not available analytically, the proof of Theorem 3.2 and Lemma 4.7 is incomplete.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 1.1 is proved via Lemma 4.10, which depends on Proposition 4.9, which in turn depends on Proposition 4.6. In Proposition 4.6, Step 1 invokes [EH2, Theorem 1.2] to run a (K_X+Δ−ε⌊Δ⌋)-minimal model program with ample scaling over Z around π_Z^{-1}(W_z) and to obtain a termination after finitely many flips and divisorial contractions. This is a load-bearing external input: without termination, the reduction to Lemma 4.5 (Cases (II) and (III)) is unjustified, and the subsequent gluing argument in Proposition 4.9 Step 4 cannot go through. Since [EH2] is a recent preprint (arXiv:2404.05126) that has not yet appeared in a refereed venue, the correctness of Theorem 1.1 is contingent on the validity of that MMP in the full needed generality of projective morphisms between complex analytic spaces. Separately, the paper's own Lemma 2.29—used in Theorem 3.2 and Lemma 4.7—is given only as a sketch ('We can verify (i) by induction... the statement then follows by a direct check'), and Lemma 4.7 itself omits its proof. These are the two points where the argument is least secure: the external MMP and the internal sketch. The reader's conditional verdict is appropriate as long as these are treated as genuine prerequisites.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves finiteness of relative log pluricanonical representations in the complex analytic setting (Theorem 1.2 and Corollary 1.3), and uses these results to prove the abundance theorem for semi-log canonical pairs under a semiampleness assumption on the normalization (Theorem 1.1), freeness for nef and log abundant log canonical bundles (Theorem 1.4), and the existence of log canonical flips for complex analytic spaces (Theorem 1.7). The proof follows the strategy of [Fuj1] and [FG], replacing algebraic MMP inputs by the complex analytic MMP developed in [Fuj12], [EH1], and [EH2], and using an admissible-sections gluing argument to descend from a dlt blow-up to the semi-log canonical pair.","tokens_in":36101,"tokens_out":3576,"duration_ms":32936,"significance":"If the foundational MMP inputs are valid, the main results are substantial: Theorem 1.1 extends [HX, Theorem 2] to projective morphisms of complex analytic spaces without Kollár's gluing theory, and Theorem 1.7 provides log canonical flips in that setting. The paper also gives a clean reduction of the analytic abundance problem to the classical abundance conjecture for projective varieties (Theorem 1.10 and Corollary 1.11). The author is careful about circularity: Remark 1.13 explicitly refrains from citing works that depend on the present paper. The main weakness is that several load-bearing technical steps are only sketched or omitted, and the central argument relies on unpublished preprints.","major_comments":[{"comment":"Lemma 2.29 is stated with only a 'Sketch of Proof', yet it is used in the proof of Theorem 3.2 and again, through Lemma 4.7, in the gluing argument for Theorem 1.1. In particular, part (ii) is reduced to 'a direct check' after blowing up Z along the center S, but that check is not written out. Since Theorem 3.2 is the engine behind Corollary 1.3 and hence behind the finiteness of relative pluricanonical representations, the proof of Theorem 1.2 is not complete as written.","section":"§2, Lemma 2.29"},{"comment":"Lemma 4.7 is a key step in producing admissible sections: it asserts that averaging over the finite group G preserves preadmissibility and that the product over G restricts to the |G|-th power on the boundary. The proof is omitted entirely ('we omit the details here'), and the assertion is not immediate because the behavior of arbitrary B-bimeromorphic maps on strata requires Lemma 2.29. Since Lemma 4.7 feeds directly into Lemmas 4.8 and 4.10 and hence into Theorem 1.1, this is a load-bearing gap in the written argument.","section":"§4, Lemma 4.7"},{"comment":"Step 1 of Proposition 4.6 invokes [EH2, Theorem 1.2] to run and terminate a (K_X+Delta-epsilon floor Delta)-MMP with ample scaling over Z around W_z, and Step 4 of Proposition 4.9 repeats this dependence through Proposition 4.6; Theorem 1.7 likewise relies on [EH2, Theorem 1.2]. Since [EH2] is an unpublished preprint, the correctness of Theorem 1.1 is contingent on the full validity of that MMP in the required generality of projective morphisms between complex analytic spaces. The paper would be substantially strengthened by a precise statement of the hypotheses and termination/finiteness properties assumed from [EH2], together with a check that every invocation in this paper falls within those hypotheses.","section":"§4, Proposition 4.6 and Proposition 4.9"}],"minor_comments":[{"comment":"The proof refers to 'Theorem 1.3', but no Theorem 1.3 exists in the paper; the intended reference is presumably Theorem 1.2.","section":"§4, proof of Corollary 1.3"},{"comment":"The proof heading says 'Proof of Theorem 4.5' and should say 'Proof of Lemma 4.5'.","section":"§4, proof of Lemma 4.5"},{"comment":"There is a typo in 'B-bimermorphic involution'; it should be 'B-bimeromorphic involution'.","section":"§4, Proposition 4.6(i)"}],"recommendation":"major_revision","confidential_remarks":"The main risk to publication is the reliance on [EH2, Theorems 1.2 and 1.3], which are not yet refereed. It would be prudent for the editor to confirm that these results are accepted or otherwise independently verifiable before the paper appears."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline: Fujino has pushed the finiteness of relative log pluricanonical representations and the abundance theorem for semi-log canonical pairs into the complex analytic setting. If the supporting results hold up, this is a real step: it reduces analytic abundance to the classical projective abundance conjecture, and it gives log canonical flips analytically. The writing is careful and honest.\n\nWhat's new: Theorem 1.1 is the abundance statement for slc pairs, Theorem 1.2 the finiteness of relative B-pluricanonical representations, Theorem 1.7 log canonical flips. None of these appear in the literature for analytic spaces. The paper is also self-aware: it spells out which preprints depend on it and avoids citing them, and Section 7 corrects issues in [Fuj1] and [FG]. That is the right kind of citizenship.\n\nThe soft spots are real but proportionate. Lemma 2.29 is load-bearing—it is used in Theorem 3.2 and Lemma 4.7—yet the proof is only a sketch, and Lemma 4.7 omits details entirely. In a paper whose main theorems rest on delicate gluing of sections, those are not cosmetic gaps. A referee should ask for full arguments or exact references. There is also a minor heading typo (\"Proof of Theorem 4.5\" for Lemma 4.5).\n\nThe bigger external issue is that Proposition 4.6 uses [EH2, Theorem 1.2] to run and terminate an MMP with scaling around a compact subset; Theorem 1.7 and Theorem 1.8 depend on [EH2, Theorem 1.3]. [EH2] is a recent, not-yet-refereed preprint. If it has a gap, the analytic results here weaken substantially. That is not circularity—Fujino is transparent that [EH2] is independent—but it does mean this paper is part of a preprint web. A careful referee should cross-check the cited statements in [EH2] rather than taking them at face value.\n\nOverall, I do not see a fatal flaw. The main theorems are plausible and the internal arguments mostly detailed. The central reasoning is coherent, and the reduction of analytic abundance to the algebraic conjecture is a genuine contribution. The right response is a serious peer review with a request to complete the sketched lemmas. It is not a paper to desk reject.","headline":"Fujino delivers a mostly solid analytic MMP extension with two sketched lemmas and heavy reliance on an unpublished preprint; deserves refereeing.","tokens_in":36632,"tokens_out":2756,"would_cite":true,"duration_ms":25179,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14E30","14E07","32C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that relative log pluricanonical representations are finite for projective morphisms of complex analytic spaces, and uses this finiteness to establish the abundance theorem for semi-log canonical pairs, log canonical…","keywords":["log pluricanonical representations","complex analytic spaces","semi-log canonical pairs","abundance conjecture","minimal model program","log canonical flips","dlt blow-ups","relative canonical bundle"],"falsifier":"Take a flat family of nodal curves over a disc with smooth total space, $\\Delta$ equal to zero, and K_X relatively semiample after normalization, for example a family of elliptic curves degenerating to a nodal rational curve. Compute the base locus of O_X(mK_X) for all positive integers m over a neighborhood of the central fiber. The theorem predicts that some single multiple is pi-generated; finding a basepoint common to all m would refute Theorem 1.1, while verifying generation for one explicit m would corroborate it.","tokens_in":35550,"feed_emoji":"📐","tokens_out":10186,"duration_ms":95255,"temperature":0.7,"pith_summary":"The paper proves that, for a projective morphism between complex analytic spaces, the group of B-bimeromorphic self-maps of a log canonical pair acts on the relative pluricanonical bundle through a finite group, provided the canonical divisor is relatively semiample. From this finiteness it derives the relative abundance theorem for semi-log canonical pairs—the possibly non-normal analogue of log canonical pairs glued from normal pieces along divisors: if the pullback of the canonical divisor to the normalization is relatively semiample over a neighborhood of a compact set, then some fixed multiple becomes relatively generated over a possibly smaller neighborhood. The paper also establishes freeness for nef and log abundant log canonical bundles, existence of log canonical flips in the analytic category, and a reduction of the abundance conjecture for projective morphisms of complex analytic spaces to the classical abundance conjecture for projective log canonical pairs. This matters because abundance is the step that turns minimal models into canonically polarized families, and the paper extends its known scope from algebraic varieties to complex analytic base spaces.","feed_headline":"Relative abundance proven for analytic semi-log canonical pairs","feed_subtitle":"One multiple of the canonical divisor is generated after shrinking the base, reducing abundance to the projective case.","key_machinery":"The central object is the relative log pluricanonical representation rho_m: Bim(X/Y,$\\Delta$) -> Aut_{O_Y}(pi_*O_X(m(K_X+$\\Delta$))), which records how B-bimeromorphic self-maps of the pair act on sections of multiples of the canonical divisor. The proof that its image is finite runs through Theorem 3.2: over a polydisc, after a dlt blow-up and resolution, the evaluation of rho_m at a fiber is a representation of a projective divisorial log terminal pair, where the algebraic finiteness theorem of [FG] bounds element orders; uniformity across the base is supplied by base change to general hyperplane slices, and the bounded-order criterion for subgroups of GL(n,C) then forces the whole image to be finite. The other load-bearing mechanism is the admissible-section formalism: sections that are invariant, in a precise sense, under the finite group of B-bimeromorphic maps and restrict correctly to log canonical strata, and whose existence over the boundary strata is propagated up to the whole space by Lemma 4.10, giving pi-generation of some multiple of the canonical divisor.","core_discovery":"The central theorem (Theorem 1.1) is an analytic analogue of the algebraic abundance theorem for semi-log canonical pairs, proved without the gluing theory in [K]. The engine behind it is Theorem 1.2: for a log canonical pair (X,$\\Delta$) over a complex variety Y, with K_X+$\\Delta$ pi-semiample, the natural representation of the B-bimeromorphic automorphism group on the finite rank module pi_*O_X(m(K_X+$\\Delta$)) has finite image. The finiteness is proven fiberwise: after base change to a point, the representation factors through a projective log canonical pair on the fiber, where a prior algebraic finiteness theorem bounds the order of every element; a classical group-theoretic criterion then forces the whole image to be finite. The paper then uses admissible sections, obtained by averaging over this finite group, to piece together sections of multiples of the canonical divisor on the normalization and descend them to the semi-log canonical space. Along the way it proves the existence of log canonical flips and good dlt blow-ups in the complex analytic setting, and shows that the relative abundance conjecture for analytic morphisms is no harder than the classical abundance conjecture for projective varieties.","pith_inferences":["The finite-group averaging that replaces the gluing theory for semi-log canonical spaces suggests the same mechanism may produce canonical bundle formulas or adjunction statements for other quotients or pinched analytic spaces where a normalization splits the pair into several birationally identified components.","Because Theorem 1.10 reduces relative abundance to the classical projective conjecture, any proof of that projective conjecture—or any counterexample to it—would transfer automatically to complex analytic families, collapsing a separate analytic version of the conjecture into the classical problem.","The paper's reliance on the strict support condition in vanishing theorems indicates that the genuinely analytic content is concentrated in Lemma 4.2; if a simpler proof of that lemma were found, the whole abundance theorem would follow from the algebraic finiteness theorem plus formal minimal model program arguments.","Theorem 1.5's hypothesis that W is a Stein compact set with noetherian ring of global functions suggests that the cleanest analytic formulation of abundance goes through Stein neighborhoods; testing whether the nefness-spreading statement of Conjecture 5.2 holds for three-dimensional analytic families would make Theorem 1.5 unconditional over arbitrary compact sets."],"forward_implications":["If Theorem 1.1 is correct, the algebraic theorem of [HX] on semi-log canonical abundance follows without the gluing theory in [K], since Lemma 4.11 passes from analytic generation to Zariski generation.","Theorem 1.4 makes every pi-nef and pi-log abundant log canonical bundle relatively free over a neighborhood, so minimal models of analytic families become morphisms to a base with canonically polarized fibers.","Theorem 1.10 and Corollary 1.11 reduce the abundance conjecture for projective morphisms between complex analytic spaces to the classical conjecture for projective log canonical pairs; proving the projective case in dimension n would settle relative abundance in dimension n for analytic families.","Theorem 1.7 gives log canonical flips in the analytic category, so the birational part of the minimal model program for log canonical pairs over analytic bases is now available.","Theorem 1.8 and Theorem 6.2 give good dlt blow-ups and the ACC for log canonical thresholds for complex analytic singularities, extending singularity-theoretic tools from algebraic to analytic settings."],"supporting_citations":[{"why":"Proves the algebraic finiteness of log pluricanonical representations for projective log canonical pairs (Theorem 2.23), which the analytic proof reduces to fiberwise.","marker":"[FG]"},{"why":"Establishes the minimal model program for projective morphisms between complex analytic spaces, including dlt blow-ups and cone and contraction theorems, used throughout to reduce to dlt and Q-factorial settings.","marker":"[Fuj12]"},{"why":"Supplies the minimal model program with scaling and termination for log canonical pairs on complex analytic spaces (Theorems 1.2 and 1.3), needed in Proposition 4.6, Theorem 1.7, and Theorem 1.8.","marker":"[EH2]"},{"why":"Provides the vanishing theorems with strict support condition that make Lemma 4.2 and the connectedness arguments work in the analytic category.","marker":"[Fuj13]"},{"why":"Introduces admissible and preadmissible sections and the inductive gluing strategy for the abundance theorem that the paper adapts to analytic spaces.","marker":"[Fuj1]"},{"why":"Gives the algebraic finiteness of B-representations and semi-log canonical abundance theorem that Theorem 1.1 recovers, supplying the comparison target for the analytic generalization.","marker":"[HX]"},{"why":"Contains the canonical bundle formula and basepoint-free theorem used in Theorem 5.1 to prove freeness for nef and log abundant bundles.","marker":"[Fuj7]"}],"fun_headline_variants":["Finiteness of relative log pluricanonical representations","Analytic abundance for semi-log canonical pairs proven","Reducing analytic abundance to projective varieties","Complex analytic spaces admit log canonical flips","Finite automorphism groups bound pluricanonical maps"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the minimal model program for projective morphisms between complex analytic spaces as developed in the cited preprints—especially dlt blow-ups, termination of flips with scaling, and existence of log canonical models; if any of those foundational results has a gap, the main theorems would weaken or collapse.","fun_headline_variants_meta":{"raw":{"variants":["Finiteness of relative log pluricanonical representations","Analytic abundance for semi-log canonical pairs proven","Reducing analytic abundance to projective varieties","Complex analytic spaces admit log canonical flips","Finite automorphism groups bound pluricanonical maps"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000214,"raw_usage":{"total_tokens":1382,"prompt_tokens":858,"completion_tokens":524,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":474,"completion_tokens_details":{"reasoning_tokens":452}},"tokens_in":474,"tokens_out":524,"duration_ms":4861,"temperature":1.0,"reasoning_tokens":452,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:57:59.429647+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a flat family of nodal curves over a disc with smooth total space, $\\Delta$ equal to zero, and K_X relatively semiample after normalization, for example a family of elliptic curves degenerating to a nodal rational curve. Compute the base locus of O_X(mK_X) for all positive integers m over a neighborhood of the central fiber. The theorem predicts that some single multiple is pi-generated; finding a basepoint common to all m would refute Theorem 1.1, while verifying generation for one explicit m would corroborate it.","supporting_citations":[],"review_version":1}