{"id":"df7889f2-a188-40a7-bc4a-a42275b23668","arxiv_id":"2505.15230","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A hereditary order on a curve is shown to admit a strong C-linear semiorthogonal decomposition obtained from the deformation absorption of singularities in its fiber over a ramified point.","lead":"This paper applies Kuznetsov and Shinder's deformation absorption framework to hereditary orders on curves, producing new semiorthogonal decompositions of their bounded derived categories. It gives the first example of this categorical absorption phenomenon in a noncommutative setting, with direct links to the derived categories of root stacks.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4.8(ii) is the load-bearing step: admissibility of ⟨S_i⟩ is outsourced to [KS25, Prop. 6.9] after only asserting Λ_r is Gorenstein because every injective module is projective; if the proposition's hypotheses are not verified, Theorems A and B collapse.","rationale":"The reader's weakest_assumption correctly locates the gate through which Theorem A passes to Theorem B: Lemma 4.8(ii). I agree that a citation gap at this point would invalidate the absorption theorem and hence the SOD in Theorem 4.11. I would only sharpen the concern: Λ_r is almost certainly self-injective, hence Gorenstein in the ordinary sense, so the disputed point is not the algebra's Gorensteinness itself but whether [KS25, Prop. 6.9]'s precise hypotheses are met and stated. The rest of the argument—P∞-object identification, exceptional pushforwards, and the maximal-overorder identification of D—is coherent and no internal contradiction is apparent. A short direct Ext computation with M_j would settle the issue, and it is likely to confirm the claim, upgrading the paper from conditional to acceptable. The verdict should therefore remain CONDITIONAL: the proof as written is conditional on an unverified citation, but the claim is plausible and testable.","tokens_in":23839,"tokens_out":35133,"duration_ms":319805,"concrete_test":"Check [KS25, Prop. 6.9] against Λ_r=kQ_r/(rad)^r. (1) Read the exact statement of [Jin20, Assumption 0.1] and verify that self-injectivity of Λ_r satisfies it. (2) Independently compute Ext^i_{Λ_r}(M_j,N) for M_j=(P_{j+1}→P_j) and N ranging over the indecomposable projective and injective Λ_r-modules; the spaces should be nonzero only for i=0,1, giving homological left and right finite-dimensionality directly. (3) If the cited proposition's hypotheses fail, supply a direct admissibility argument for ⟨S_j⟩ using self-injectivity; if they hold, record the missing verification in Lemma 4.8.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem B's semiorthogonal decomposition (4.11) rests on Theorem A, whose core is Lemma 4.8(ii): the subcategory ⟨S_i⟩ is admissible because the two-term complex M_i=(P_{i+1}→P_i) is homologically left and right finite-dimensional. The proof's only justification is the conjunction of [KS25, Proposition 6.9] with the claim that Λ_r is Gorenstein 'in the sense of [Jin20, Assumption 0.1]' because every injective Λ_r-module is projective. This is not verified in the text: the paper does not state what [Jin20, Assumption 0.1] requires, does not explain how self-injectivity implies that condition, and does not check that M_i satisfies all hypotheses of the cited proposition. The same admissibility conclusion is reused to assert that the complements ⊥S_i and S_i⊥ are smooth and proper. If the citation does not apply, the absorption theorem for the fiber—and with it the deformation absorption used in Theorem 4.11—has no proof. This is a correctness risk, not merely a missing reference, because the Gorenstein condition in [Jin20] may involve dualizing-complex or smoothness assumptions beyond the property that injective modules are projective.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a noncommutative analogue of Kuznetsov–Shinder's categorical absorption of singularities in the setting of hereditary orders on curves. For a hereditary O_C-order A ramified at a single closed point o of ramification index r, the author proves that the sequence of simple modules S_1,...,S_{r-1} of the fiber algebra A(o) is semiorthogonal and consists of P_{∞,2}-objects, and that the triangulated subcategory they generate absorbs singularities (Theorem A, Theorem 4.7). Using this, he constructs a strong C-linear semiorthogonal decomposition of D^b(C,A) with an exceptional collection pushed forward from the fiber and a complement D equivalent to D^b(C) (Theorem B, Theorem 4.11). An appendix proves a base-change formula for semiorthogonal decompositions on coherent ringed schemes (Theorem C, Theorem A.16), and a final section translates the results to smooth root stacks via the Chan–Ingalls dictionary.","tokens_in":24105,"tokens_out":13945,"duration_ms":113179,"significance":"If the proofs can be completed, the paper provides the first instance of deformation absorption of singularities in a noncommutative setting. It offers a new, conceptual proof of semiorthogonal decompositions for hereditary orders, links the Kuznetsov–Shinder machinery to the stacks–orders dictionary, and the appendix's base-change formula is a useful contribution to the theory of coherent ringed schemes. The exposition is careful about the framework of coherent ringed schemes, and the explicit P_{∞,2}-objects and the translation to root stacks are valuable. However, the central admissibility step in the proof of Theorem A is not fully justified, and several auxiliary results are only sketched.","major_comments":[{"comment":"The admissibility of the subcategory ⟨S_i⟩ is the load-bearing step for Theorem 4.7(iii) and for the deformation absorption argument leading to Theorem 4.11. The proof is incomplete: it asserts that Λ_r is Gorenstein in the sense of [Jin20, Assumption 0.1] because every injective Λ_r-module is projective, and then invokes [KS25, Proposition 6.9] to conclude that the two-term complex M_i=(P_{i+1}→P_i) is homologically left and right finite-dimensional. The paper does not state what [Jin20, Assumption 0.1] requires, does not explain how self-injectivity implies that condition, and does not verify that M_i satisfies all hypotheses of [KS25, Proposition 6.9]. If that proposition does not apply, the conclusion that ⟨S_i⟩ is admissible is unsupported, and the absorption theorem for the fiber collapses. Please either prove the homological finite-dimensionality of M_i directly or spell out the verification of the cited results.","section":"Section 4.2, Lemma 4.8(ii)"},{"comment":"The exceptionality of the pushforwards i_{o,*}S_k and the vanishing between them are derived from the distinguished triangle S_k[1] → Li_o^* i_{o,*}S_k → S_k → S_k[2], which is transferred from [KS23, Section 4.2] to the noncommutative coherent ringed scheme (C,A) without proof. This is not a formal consequence of the commutative statement, because pullback and pushforward for coherent ringed schemes involve the algebra structure; the comparison with the canonical self-extension in equation (39) is valid only if the noncommutative base-change triangle has the stated form. Please provide a proof of the triangle in this setting or a precise reference that covers it.","section":"Section 4.3, Lemma 4.12"},{"comment":"The identification of the component D with j_{B_i,*}D^b(C,B_i) is only sketched. In particular, the step asserting that a complex Q• belongs to the subcategory generated by L_o^{(r)} follows from the vanishing of Hom(Q•, i_{o,*}S_k) for k=1,...,r-1 is not justified; this is essentially the statement that the left orthogonal of the collection is generated by a single projective module, which should be proved by invoking the semiorthogonal decomposition of D^b(A(o)) from Theorem 4.7 and Lemma 4.10. The claim that M_o can be expressed as an iterated cone of direct sums of L_o^{(r)} also needs justification. Since Lemma 4.14 is used to prove that the decomposition in Theorem 4.11 is strong and that D≃D^b(C), this gap affects Theorem 4.11(ii).","section":"Section 4.3, Lemma 4.14"}],"minor_comments":[{"comment":"There is a typo: 'it is necessary for us to work need in the more general framework' should read 'it is necessary for us to work in the more general framework'.","section":"Section 2, after Definition 2.1"},{"comment":"The displayed fiber equation (X×_C Spec k(o), B_{Spec k(o)}) = (X,A) is inconsistent with the application in Section 4.3, where the base and the total space are the same curve C and the fiber is (Spec k(o), A(o)). Please clarify the notation.","section":"Definition 4.6"},{"comment":"The proof states without justification that S_A(P_i)=P_{i+1}[1] and S_A(i_{o,*}S_i)=i_{o,*}S_{i+1}; please add a proof or a reference. Also, in equation (47), the last component should be j_{B_i,*}D^b(C,B_i), not j_{B_i,*}D^b(C,A).","section":"Theorem 4.16"},{"comment":"The notation L(fo◦io)^*F is unclear; it should presumably be L(f∘i_o)^*F, where f is the structure morphism (C,A)→C.","section":"Lemma 4.13"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is well written and the proposed noncommutative deformation absorption is interesting. My main concern is that the proof of Theorem A rests on an uncritical application of [KS25, Proposition 6.9] and an unverified Gorenstein condition; this should be fixable by a direct argument using the structure of Λ_r. The noncommutative base-change triangle in Lemma 4.12 also needs a proof. The paper's scope is appropriate, and the single self-citation is not load-bearing. I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid paper that does what it says. It gives the first deformation absorption of singularities in a noncommutative setting, applied to hereditary orders on curves, and the main argument holds up on reading. The most delicate step is admissibility of the subcategory generated by one simple module in D^b(Λ_r), and the proof there is more compressed than it should be, but I don't think it's broken.\n\nThe genuinely new content is the construction of the simple modules S_i as P∞,2-objects in D^b(Λ_r), the proof that their span absorbs singularities, and the resulting C-linear semiorthogonal decomposition of D^b(C,A) with the leftover component equivalent to D^b(C). The paper is honest that the final decomposition can also be obtained via the Chan–Ingalls dictionary and BLS16, but the categorical-absorption route is new and gives a structural explanation. The appendix on noncommutative base change is a useful addition, and the exposition is clear throughout.\n\nSoft spots, in proportion:\n\n- Lemma 4.8(ii) outsources the key admissibility statement to [KS25, Prop. 6.9], after asserting Λ_r is Gorenstein in the sense of [Jin20, Assumption 0.1] because every injective module is projective. The assertion is true — Λ_r is self-injective, so it is Gorenstein of dimension zero — but the author should state the assumption from Jin20 explicitly and verify it, rather than leave it to the reader. This is a clarity issue, not a correctness gap in my reading. A referee should ask for the details.\n\n- Equation (15) has a typo: the ideal is generated by all cycles, i.e. rad^r of the path algebra, not (rad)^{n-1} as written. Easy fix.\n\n- Theorem 4.16, the 2r-periodicity proof, is sketched rather than fully argued. The Serre functor computation is plausible, but the mutation argument would benefit from more detail. Minor.\n\nThe citation pattern is fine; the single self-citation [BBG24] is not load-bearing. The paper does not engage with the stress-test concern that the Gorenstein step is a load-bearing flaw — I think the concern is answerable, but the author should close it in print.\n\nBottom line: this deserves serious refereeing. I would send it to a good journal. The main theorems are new, the framework is meaningful, and the technical gaps are fillable. Recommendation: engage with it, with a referee request for expansion of Lemma 4.8(ii) and the periodicity argument.","headline":"First noncommutative deformation absorption, plausibly correct; the compressed Gorenstein citation in Lemma 4.8 is a clarity issue, not a fatal flaw.","tokens_in":24644,"tokens_out":4964,"would_cite":true,"duration_ms":40870,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F08","16H10","14A22","16G20"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a hereditary order on a curve ramified to index r at one point, the paper proves the derived category splits as an exceptional block of r−1 pushed-forward simple modules plus a copy of D^b(C), giving the first noncommutative…","keywords":["semiorthogonal decomposition","hereditary orders","P∞,2-objects","deformation absorption of singularities","root stacks","cyclic quiver algebra","noncommutative derived categories","base change"],"falsifier":"Compute the Ext groups of the two-term complex M_i = (P_{i+1} → P_i) against every object of D^b(Λ_r): if some Ext^i_{Λ_r}(M_i,N) or Ext^i_{Λ_r}(N,M_i) is infinite-dimensional, the cited finiteness criterion cannot apply and the admissibility of ⟨S_i⟩—hence Theorem 4.7 and Theorem 4.11—fails. A simpler direct check is to exhibit one injective Λ_r-module that is not projective, since Lemma 4.8(ii) explicitly asserts every injective module is projective as justification for the Gorenstein condition.","tokens_in":23605,"feed_emoji":"🧩","tokens_out":8825,"duration_ms":76305,"temperature":0.7,"pith_summary":"This paper claims that the derived category of a hereditary order on a curve—a sheaf of noncommutative algebras that is Azumaya away from finitely many points—can be decomposed using the categorical mechanism that absorbs singularities of a singular fiber in a flat family. The finite-dimensional algebra obtained by restricting the order to a ramified point is Morita equivalent to the cyclic quiver algebra Λ_r, and its r−1 simple modules form a semiorthogonal collection of P∞,2-objects that absorb singularities. Pushing these modules into the order gives an exceptional collection, and the remaining piece is equivalent to the derived category of the curve. A reader should care because this supplies the first deformation absorption of singularities in a noncommutative setting and yields explicit, base-linear decompositions for hereditary orders and, via the stacks–orders dictionary, for smooth root stacks.","feed_headline":"Singularity absorption splits derived categories of hereditary orders","feed_subtitle":"Simple modules at the ramified point form the exceptional block; the rest is the curve's derived category.","key_machinery":"The load-bearing objects are the simple modules S_1,...,S_{r−1} of the cyclic quiver algebra Λ_r = kQ_r/I, where Q_r is the r-cycle quiver and I is the ideal killing all cycles; each S_i admits a 2-periodic projective resolution and has self-extension ring k[θ] with deg θ=2, making it a P∞,2-object. The canonical self-extension triangle of such an object contains the two-term complex M_i = (P_{i+1} → P_i), and homological finiteness of M_i is what makes the subcategory ⟨S_i⟩ admissible. The complement D is realized through the maximal overorder B_i of A that is purely of type i at o, whose derived category is equivalent to D^b(C), and a noncommutative base change theorem is used to ensure the decomposition is strong and compatible with fibers.","core_discovery":"The paper establishes that the bounded derived category of a hereditary OC-order A over a smooth curve C, with a single ramified point o of index r, admits a strong C-linear semiorthogonal decomposition ⟨i_{o,*}S_{i+1},...,i_{o,*}S_{i−1}, D⟩ where the first r−1 terms are pushforwards of simple modules of the fiber algebra A(o), which is Morita equivalent to the cyclic quiver algebra Λ_r, and D is equivalent to D^b(C). This is proved by showing that the simple modules S_1,...,S_{r−1} of Λ_r form a semiorthogonal collection of P∞,2-objects whose generated subcategory absorbs singularities in the deformation-absorption sense, and by identifying the complement with the maximal overorder B_i of type i at o. The paper also provides a noncommutative base change formula for semiorthogonal decompositions along flat morphisms, which controls the fibers of the decomposition over every point of the curve.","pith_inferences":["Beyond the paper: iterating the single-point argument over all ramified points should produce a decomposition with one exceptional block attached to each ramified point; the paper states only the one-point case.","Beyond the paper: the exceptional collection and 2r-periodicity suggest the existence of an explicit tilting bundle for D^b(C,A) when C is projective, extending the weighted-projective-line tilting phenomenon; the paper does not construct one.","Beyond the paper: if the homological-finiteness criterion survives unchanged, the same P∞,2-object mechanism should apply to tame orders of global dimension two on surfaces, where maximal overorders are no longer Azumaya; the paper only flags this analogue.","Beyond the paper: a concrete test of the machinery is whether the noncommutative base change formula continues to hold under base change along non-flat or non-faithful morphisms, which would let the decomposition deform over families of curves."],"forward_implications":["Theorem B yields a strong C-linear semiorthogonal decomposition of D^b(C,A) with an exceptional block of r−1 objects and a complement smooth and proper over D^b(C).","The complement D is equivalent to D^b(C), so the decomposition is a direct noncommutative analogue of semiorthogonal decompositions for root stacks.","Fibers of D over every closed point p∈C are equivalent to D^b(mod k(p)); over non-ramified points this recovers the Azumaya point.","The decomposition is 2r-periodic under right mutations, matching the known periodicity of root-stack decompositions.","For a single ramified point, the fiber subcategory S = ⟨S_1,...,S_{r−1}⟩ provides a deformation absorption of singularities of the fiber algebra Λ_r, the first such example in a noncommutative setting."],"supporting_citations":[{"why":"Supplies the definition of P∞,2-objects and the absorption theorem converting such collections into semiorthogonal decompositions.","marker":"[KS23]"},{"why":"Proposition 6.9 is the homological finite-dimensionality criterion that makes ⟨S_i⟩ admissible in Lemma 4.8(ii).","marker":"[KS25]"},{"why":"Assumption 0.1 gives the Gorenstein condition used to justify applying the finiteness criterion to Λ_r.","marker":"[Jin20]"},{"why":"Corollary 7.8 provides the stacks–orders dictionary that identifies ramification points with nontrivial stabilizers and translates the results to root stacks.","marker":"[CI04]"},{"why":"Classification of hereditary orders and their indecomposable projectives supplies the normal form of A_p and the structure of the fiber.","marker":"[Rei75]"},{"why":"Theorem 3.1 identifies the fiber of the normal form order with the cyclic quiver algebra Λ_r.","marker":"[KSS03]"},{"why":"Base change theorem for semiorthogonal decompositions whose noncommutative version Theorem A.16 controls fibers of D.","marker":"[Kuz11]"},{"why":"Provides the result that exceptional collections generate admissible subcategories, used to form D in Lemma 4.13.","marker":"[Bon89]"},{"why":"Theorem 4.7 gives the root-stack semiorthogonal decomposition that the stacky translation of Theorem 4.11 is compared with.","marker":"[BLS16]"}],"fun_headline_variants":["Split derived category of hereditary order via ramified point","Ramified point splits derived category into curve and simple pieces","Simple modules at ramified point block off curve derived category","Deformation absorption yields curve's derived category decomposition"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole result hinges on the assertion that the subcategory generated by one simple module of the cyclic quiver algebra is admissible in the derived category, which is established by a cited finiteness theorem plus the claim that every injective module over that algebra is projective; if either ingredient fails, the absorption and the decomposition collapse.","fun_headline_variants_meta":{"raw":{"variants":["Split derived category of hereditary order via ramified point","Ramified point splits derived category into curve and simple pieces","Simple modules at ramified point block off curve derived category","Deformation absorption yields curve's derived category decomposition"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001591,"raw_usage":{"total_tokens":6290,"prompt_tokens":839,"completion_tokens":5451,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":455,"completion_tokens_details":{"reasoning_tokens":5387}},"tokens_in":455,"tokens_out":5451,"duration_ms":33426,"temperature":1.0,"reasoning_tokens":5387,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T15:20:56.307976+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the Ext groups of the two-term complex M_i = (P_{i+1} → P_i) against every object of D^b(Λ_r): if some Ext^i_{Λ_r}(M_i,N) or Ext^i_{Λ_r}(N,M_i) is infinite-dimensional, the cited finiteness criterion cannot apply and the admissibility of ⟨S_i⟩—hence Theorem 4.7 and Theorem 4.11—fails. A simpler direct check is to exhibit one injective Λ_r-module that is not projective, since Lemma 4.8(ii) explicitly asserts every injective module is projective as justification for the Gorenstein condition.","supporting_citations":[],"review_version":1}