{"id":"9ba5fade-2b8e-44ef-82b4-d40aabf56941","arxiv_id":"2411.17549","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The exceptional divisor of a divisorial Mori contraction of submaximal length is birational to a projective bundle or a quadric bundle.","lead":"An algebraic geometry paper proves a structure theorem for contractions of submaximal length: the exceptional locus is birational either to a projective bundle or to a quadric bundle. The result completes a natural relative version of classical characterizations of projective spaces and quadrics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.1 applies a smooth-target deformation bound to singular fibres of E→Z; on a quadric cone this bound is false, so the existence of f-minimal curves in every fibre and the dichotomy −E·Γ∈{1,2} are not yet secured.","rationale":"The reader's weakest assumption is exactly the delicate deformation-family argument in Lemma 3.1, and my analysis converges on the same point. The other parts of the paper appear internally coherent: the ray argument shows that in the −E·Γ=2 case Lemma 4.2's global degree condition is satisfied, the semicontinuity step in Section 5 is plausible once one notes that a proper map with a zero-dimensional central fibre is finite in a neighbourhood, and the conductor/quadric arguments in Proposition 3.3 are consistent. The unresolved issue is that Lemma 3.1 applies [Kol96, II.1.3] to a singular fibre without proving that the minimal rational curve is free or that the singularities of E_z do not obstruct its deformations. Since a quadric cone provides a concrete normal lci Fano variety where the asserted dimension bound is violated, the possibility that such a fibre occurs in the setup has not been excluded. This does not amount to a demonstrated counterexample, but it is a genuine gap in the proof of a statement on which the main dichotomy rests. Therefore the conditional verdict remains appropriate; the paper needs either a corrected deformation-theoretic argument valid for singular fibres or an explicit proof that the fibres in question are smooth along their minimal curves.","tokens_in":16279,"tokens_out":46544,"duration_ms":479252,"concrete_test":"Test the deformation-theoretic step in Lemma 3.1 on a cone fibre: let E_z be an n-dimensional quadric cone (n≥3), with −K_{E_z}=O(n), and compare the dimension of the deformation family of a ruling line (n−1) with the bound −K_{E_z}·ℓ + n − 3 = 2n−3 used in the proof. Then determine whether a quadric cone can occur as a fibre in Setup 1.4; if it can, Lemma 3.1 fails as written; if it cannot, identify and verify the extra hypothesis that excludes such a fibre, for instance that the minimal curve must be free inside the smooth locus of E_z.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central theorem and both classification branches depend on Lemmas 3.1 and 3.2, which provide f-minimal curves in every fibre and then show they cover Eeq. The proof of Lemma 3.1 uses [Kol96, Prop. II.1.3] to assert, for a generically reduced fibre E_z of dimension n over a smooth point of Z, that a minimal rational curve C meeting the smooth locus has a deformation family in E_z of dimension at least −K_{E_z}·C + n − 3. This is a smooth-target deformation estimate, but E_z is only an lci fibre of a Mori contraction and may be singular or even non-normal. The later classification explicitly permits normal quadrics, including quadric cones. On an n-dimensional quadric cone Q (n≥3), the ruling lines are minimal rational curves that meet the smooth locus, yet their deformation family has dimension n−1, whereas the displayed bound would be −K_Q·ℓ + n − 3 = n + n − 3 = 2n−3, which is strictly larger for n≥3. Thus the numerical estimate cannot hold without an additional hypothesis on C or E_z, such as freeness of C inside the smooth locus. Since the proof of Proposition 3.3 and hence the dichotomy −E·Γ=1 or 2 relies on this bound, and Lemma 3.2's covering statement transfers it to all fibres, the missing justification is load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a relative numerical characterization of divisorial elementary Mori contractions of submaximal length. Under Setup 1.4, where f:X→Y is birational, divisorial, elementary, and has length l(f)=n−1, the main theorem (Theorem 1.5) claims that for an f-minimal curve Γ in the equidimensional locus Eeq one has either −E·Γ=1 or −E·Γ=2. In the first case it asserts that Eeq→Zeq is birational to a quadric bundle with reducible fibres having two components normalized by P^n; in the second case it asserts that all fibres are normalized by P^n and, for n even, that Eeq→Zeq is a projective bundle. The proof is built on Lemmas 3.1 and 3.2, which supply f-minimal curves in every fibre and show that they cover Eeq, and on a case analysis in Section 4. Section 5 constructs a nonequidimensional example of a divisorial elementary Mori contraction of submaximal length whose exceptional divisor is generically a quadric bundle with one projective-space fibre.","tokens_in":16450,"tokens_out":22674,"duration_ms":230594,"significance":"If the main theorem is correct, it is a natural and valuable extension of the Höring–Novelli treatment of maximal length to the submaximal case, and it fits into the line of Cho–Miyaoka–Shepherd-Barron, Kebekus, and Dedieu–Höring. The paper is clearly structured and the birational constructions in Section 4 are coherent. The author also provides an explicit example in Section 5, which is useful for understanding why the equidimensional locus is needed. However, the central proof relies on deformation-theoretic estimates for singular fibres, and these estimates are not justified in the text; since the dichotomy −E·Γ∈{1,2} and the subsequent fibre descriptions depend on them, the main theorem is not yet established as written.","major_comments":[{"comment":"The proof applies [Kol96, Prop. II.1.3] to a fibre Ez that is only assumed to be a generically reduced local complete intersection in the smooth variety X. The cited deformation estimate is a smooth-target statement; indeed, Lemma 3.2 explicitly says 'As X is smooth, we may apply [Kol96, Theorem II.1.3]', while Lemma 3.1 suppresses the smoothness issue. On a singular fibre the estimate cannot be used as a black box: for an n-dimensional quadric cone Q with n≥4, the family of lines has dimension n−1, whereas the formula −KEz·C+n−3 would give at least 2n−4. The line does not satisfy the numerical hypothesis −KEz·C>n+1 of the subsequent contradiction, so this example does not by itself disprove Lemma 3.1, but it shows that the cited estimate is false for arbitrary singular lci fibres and that the proof needs an additional argument controlling the minimal curve C inside a singular fibre. This point is load-bearing because Lemma 3.1 supplies the f-minimal curves used in Lemma 3.2, Proposition 3.3, and the whole dichotomy −E·Γ∈{1,2}.","section":"Lemma 3.1"},{"comment":"The proof invokes [CMSB02, Theorem 0.1] to bound the dimension of an unsplit deformation family in a possibly singular fibre F by 2n−2 and to infer that equality of the bound implies that F is normalized by P^n. The cited theorem is a statement about smooth projective varieties, and the text does not explain how to reduce to the normalization of F, which need not be smooth. The same use of [CMSB02, Theorem 0.1] appears again in Lemma 4.2 for a component D1 of a reducible or nonreduced fibre. Without a singular analogue of this dimension bound, or a separate proof, the claimed dichotomy and the fibre descriptions in Theorem 1.5 are not established.","section":"Proposition 3.3"},{"comment":"The sentence 'we may conclude that Ez is isomorphic to a quadric' is ambiguous and appears inconsistent with the preceding equality −KEz = −nE|Ez if 'quadric' is allowed to mean a quadric cone. On an n-dimensional quadric cone Q with hyperplane class H, one has −KQ = (n−1)H and H·ℓ = 1, whereas −nE|Ez would have degree n on the ruling line ℓ. Thus the equality forces a smooth quadric, or at least rules out the vertex cone. The terminology should be clarified and the proof adjusted accordingly.","section":"Proposition 3.3, case (ii)"}],"minor_comments":[{"comment":"The step 'Fixing p ∈ Ez a general point ... we have dim Uz,p ≥ n+1' is not justified as written: if p is not in the image of the evaluation map, the fibre over p is empty. The argument should either choose p in the image of ev or explain how dominance of ev is obtained.","section":"Lemma 3.1"},{"comment":"The reference [Kol96, Prop. II.1.3] in Lemma 3.1 and [Kol96, Theorem II.1.3] in Lemma 3.2 should be checked and cited consistently, since the exact hypotheses of the cited statement matter for the main argument.","section":"Section 2"},{"comment":"The smoothness verification of X is relegated to 'a straightforward calculation' without showing the Jacobian or its rank. Since the example in Section 5 is meant to illustrate the main result, it would be helpful to include the computation or a reference.","section":"Lemma 5.1"},{"comment":"There are typographical issues such as 'eiher' for 'either' in Lemma 3.1 and 'K ebekus' in the abstract; these should be corrected.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"I recommend major revision rather than rejection because the geometric conclusions are plausible and the paper is well organized, but the core proof relies on deformation estimates for singular fibres that are not justified. The concern about Lemma 3.1 is real: the author should either supply a precise singular version of the cited deformation bound or restructure the proof to avoid it. The quadric cone example in the stress-test is not exactly a fibre of the setup, so it should be presented as evidence that the cited estimate is not a general fact, not as a direct counterexample to Lemma 3.1. I would ask the author to address the singular applications of [CMSB02, Theorem 0.1] as well."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read Dewer's paper. The intended result is worth having: a submaximal-length divisorial Mori contraction whose exceptional locus is birational to a projective bundle or a quadric bundle, with a clean projective-space versus quadric dichotomy and an explicit non-equidimensional example. The organization is good, the Section 4 constructions are plausible, and the citation pattern is appropriate—the paper really is extending [HN13] rather than repackaging it. No circularity, no fitted parameters.\n\nThe soft spot is load-bearing. The proofs of Lemma 3.1 and Proposition 3.3 use [Kol96, II.1.3] to lower-bound deformation spaces of rational curves in the fibres of E→Z. That bound is a smooth-target deformation estimate. The fibres are only lci and may be singular; the classification later explicitly allows normal quadrics, including quadric cones. On an n-dimensional quadric cone (n≥3), the ruling lines meet the smooth locus, but their deformation family has dimension n−1, not the claimed −K_Q·ℓ+n−3 = 2n−3. So the estimate is false in exactly the kind of fibre the theorem permits. Lemma 3.1 uses this bound to rule out −K_Ez·C > n+1 and to produce f-minimal curves in every fibre; Proposition 3.3 uses it again to force −E·Γ∈{1,2}. Without a freeness or smoothness condition on the curve, the dichotomy is not secured as written.\n\nThe rest of the paper is in better shape. The projectivization and quadric-bundle constructions in Section 4 are coherent given the dichotomy. The §5 example is instructive but the semicontinuity step proving that ε is elementary is too terse and should be expanded, though that is secondary.\n\nWho gets value from this? Specialists in Mori theory and Fano classification. The result is likely true and the gap may be repairable—for instance by adding a suitable hypothesis on the minimal curves or by a more careful deformation-theoretic argument for non-free curves. I would not cite it as established until that is fixed, but I would want a referee to see it. Recommend sending to peer review, with a referee who knows singular deformation theory.","headline":"A meaningful submaximal analogue of Höring–Novelli, but the central dichotomy rests on a deformation estimate that fails on the singular fibres the paper explicitly allows.","tokens_in":17067,"tokens_out":8911,"would_cite":false,"duration_ms":91654,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14E30","14J45","14D06"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a birational divisorial elementary Mori contraction of submaximal length, the exceptional locus is birational either to a projective bundle or to a quadric bundle.","keywords":["divisorial Mori contraction","submaximal length","exceptional locus","projective bundle","quadric bundle","deformation family of rational curves","Fano manifold","birational geometry"],"falsifier":"A concrete observation that would settle the claim: find a birational divisorial elementary Mori contraction satisfying Setup 1.4 with an $n$-dimensional fibre over a smooth point of $Z$ containing no $f$-minimal curve, or an $f$-minimal curve $\\Gamma$ with $-E\\cdot\\Gamma\\ge 3$; either outcome would refute the dichotomy, and a computation showing Lemma 3.1's surjectivity fails would also falsify the main theorem.","tokens_in":15952,"feed_emoji":"📐","tokens_out":7433,"duration_ms":73486,"temperature":0.7,"pith_summary":"This paper studies birational divisorial elementary Mori contractions whose length is one less than maximal, i.e., $l(f)=n-1$ where $n$ is the dimension of the general fibre of the exceptional divisor. It proves that the equidimensional part of the exceptional locus is birational either to a family of projective spaces or to a quadric bundle. In the projective-space case, for even $n$ the family is actually the projectivization of a vector bundle. In the quadric case, reducible fibres have exactly two components, each normalized by projective space. The result extends the known maximal-length statement (projective bundle) and gives a relative analogue of the absolute characterizations of projective spaces and quadrics.","feed_headline":"Submaximal length: exceptional loci are projective or quadric bundles","feed_subtitle":"Paper: under length n−1, the exceptional locus is birational to a P^n-bundle or a quadric bundle","key_machinery":"The central objects are the length $l(f)=n-1$ (the minimal degree of $-K_X$ on contracted rational curves) and the $f$-minimal curves realizing this degree. Lemmas 3.1 and 3.2 show that these curves form an unsplit, surjective deformation family covering the equidimensional locus. The dichotomy $-E\\cdot\\Gamma\\in\\{1,2\\}$ comes from comparing the bend-and-break lower bound $\\dim H\\ge 2n-4-E\\cdot\\Gamma$ with the upper bound $\\dim H\\le 2n-2$ supplied by the Cho--Miyaoka--Shepherd-Barron characterization of projective space. The conductor divisor (Lemma 2.5) then forces the fibre structure, and the surjectivity of the relative evaluation map (Theorem 2.7, due to Andreatta and Wiśniewski) builds the projective or quadric bundle model.","core_discovery":"Under Setup 1.4, for any $f$-minimal curve $\\Gamma$ the intersection number $-E\\cdot\\Gamma$ can only be $1$ or $2$. When it is $2$, every $n$-dimensional fibre of $E_{\\mathrm{eq}}\\to Z_{\\mathrm{eq}}$ is normalized by $\\mathbb{P}^n$ and the whole fibration is birational to a family of projective spaces; for $n$ even it is isomorphic to the projectivization of a vector bundle over $Z_{\\mathrm{eq}}$. When it is $1$, the fibration is birational to a quadric bundle, every reducible fibre has two irreducible components whose reductions are normalized by $\\mathbb{P}^n$, irreducible generically reduced fibres are quadrics, and nonreduced fibres have reduction $\\mathbb{P}^n$.","pith_inferences":["The generic fibre dichotomy (projective space or quadric) mirrors the absolute pseudoindex characterization of Fano manifolds, suggesting that the length deficit $n-1$ is the relative avatar of pseudoindex $n$.","The parity obstruction for $n$ suggests that whether the family is globally projectivized may depend on the vanishing of an even-degree cohomology class (e.g., a Brauer-type invariant); for odd $n$ the paper leaves open whether non-projectivized families with $-E\\cdot\\Gamma=2$ exist.","The §5 example indicates that dropping equidimensionality creates special fibres isomorphic to $\\mathbb{P}^3$ inside a quadric-bundle family, so a full global statement would need to track non-equidimensional fibres separately, presumably as additional blow-ups.","The same degeneration-of-minimal-curves technique could be applied to contractions of length $n-k$ for $k>1$, plausibly yielding a hierarchy of fibre models (complete intersections of quadrics, etc.), though the paper stops at $k=1$."],"forward_implications":["The equidimensional part of the exceptional divisor of a submaximal divisorial Mori contraction is always birational to a $\\mathbb{P}^n$-bundle or a quadric bundle.","When $-E\\cdot\\Gamma=2$ and $n$ is even, the birational modification can be removed: $E_{\\mathrm{eq}}$ is the projectivization of a vector bundle over $Z_{\\mathrm{eq}}$.","In the quadric-bundle case, reducible fibres have exactly two components (each normalized by $\\mathbb{P}^n$) and nonreduced fibres have reduction $\\mathbb{P}^n$, so the local structure of the exceptional divisor is completely pinned down.","Together with the maximal-length theorem, this gives a relative dichotomy: maximal length forces projective bundles; submaximal length forces projective or quadric bundles."],"supporting_citations":[{"why":"Supplies the numerical characterization of projective space via the dimension bound on deformation families of rational curves, the key upper bound $\\dim H\\le 2n-2$.","marker":"[CMSB02]"},{"why":"Gives an alternative proof of the projective space characterization used alongside [CMSB02].","marker":"[Ke02]"},{"why":"Characterizes smooth quadrics by pseudoindex $n$, the absolute analogue of the quadric-bundle dichotomy.","marker":"[DH17]"},{"why":"Establishes the maximal-length case (projective bundle up to birational modification) which this paper extends to submaximal length.","marker":"[HN13]"},{"why":"Provides Ionescu's inequality bounding the length of an elementary Mori contraction, used in Setup 1.4 and the terminology of maximal/submaximal length.","marker":"[Io86]"},{"why":"Wiśniewski's version of the length inequality, cited together with [Io86] for Theorem 1.3.","marker":"[Wi91]"},{"why":"Provides the theory of deformation families, the bend-and-break argument, and degree constancy on closed families used in Lemmas 3.1 and 3.2.","marker":"[Kol96]"},{"why":"Theorem 2.7 gives the surjectivity of the relative evaluation map used to construct the projective bundle and quadric bundle models.","marker":"[AW93]"},{"why":"Supplies the conductor divisor formalism used to rule out nonnormal fibres and determine the fibre structure.","marker":"[Kol13]"}],"fun_headline_variants":["Submaximal length: exceptional locus birational to projective or quadric bundle","When length is submaximal, exceptional loci are projective or quadric bundles","Submaximal Mori contractions: exceptional loci become projective or quadric bundles","Exceptional loci of submaximal contractions: projective or quadric bundles","Submaximal length yields projective or quadric bundle exceptional loci"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole dichotomy rests on the claim in Lemmas 3.1 and 3.2 that the $f$-minimal curves form an unsplit deformation family that covers every $n$-dimensional fibre of the exceptional divisor; if some fibre lacked a minimal curve, or the family degenerated into reducible curves, the dichotomy $-E\\cdot\\Gamma=1,2$ would collapse.","fun_headline_variants_meta":{"raw":{"variants":["Submaximal length: exceptional locus birational to projective or quadric bundle","When length is submaximal, exceptional loci are projective or quadric bundles","Submaximal Mori contractions: exceptional loci become projective or quadric bundles","Exceptional loci of submaximal contractions: projective or quadric bundles","Submaximal length yields projective or quadric bundle exceptional loci"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001583,"raw_usage":{"total_tokens":6258,"prompt_tokens":829,"completion_tokens":5429,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":445,"completion_tokens_details":{"reasoning_tokens":5333}},"tokens_in":445,"tokens_out":5429,"duration_ms":31185,"temperature":1.0,"reasoning_tokens":5333,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:00:16.390916+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete observation that would settle the claim: find a birational divisorial elementary Mori contraction satisfying Setup 1.4 with an $n$-dimensional fibre over a smooth point of $Z$ containing no $f$-minimal curve, or an $f$-minimal curve $\\Gamma$ with $-E\\cdot\\Gamma\\ge 3$; either outcome would refute the dichotomy, and a computation showing Lemma 3.1's surjectivity fails would also falsify the main theorem.","supporting_citations":[],"review_version":1}