{"id":"9803d8fa-df42-4e99-9b06-2bb4653d18b1","arxiv_id":"2411.16309","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For every Grothendieck category, the subcategories closed under coproducts, subobjects, and essential extensions, as well as the cohomologically stable thick subcategories, are classified by clopen subsets of a new Boolean spectrum.","lead":"Mathematicians get a new way to assign a support location to every object in a large class of additive categories, using a Boolean spectrum built from localizing subcategories. The support classifies certain natural subcategories and extends known results about modules over rings and the Ziegler spectrum.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.8's axiom (S1) fails: for A=Mod Z, the generator Z has support {Q}, not all of Spc(Mod Z).","rationale":"The reader's ACCEPT overlooks a concrete counterexample to a highlighted theorem. The concern is not about the imported spectral-category equivalence; it is an internal failure of the support datum axioms. Theorem 3.3 (the classification of essentially closed subcategories) may still be correct and is logically independent of Theorem 3.8, but the paper's strongest claim, as summarised, includes the universal property of the support datum. Since (Spc A, supp) is not a support datum for A=Mod Z, Theorem 3.8 is false as stated. This is load-bearing because the abstract and introduction advertise the support datum and its universal property as a main contribution. A correction would require either modifying (S1), for example by requiring supp(X)=Spc A only for a distinguished generator or for generators of Spec A, or substantially weakening Theorem 3.8. The classification theorems and the exact support results remain of independent interest, but the paper in its current form cannot be accepted without major revision.","tokens_in":23544,"tokens_out":38424,"duration_ms":361891,"concrete_test":"Compute supp(Z) for A=Mod Z directly. Since Spc A = Spec(P(Sp A)) and supp(Z) is the clopen set of ultrafilters on Sp A containing the singleton {Q}, verify that this clopen set is not all of Spc A by exhibiting the principal ultrafilter at Z(p∞) for a prime p in its complement. Equivalently, show that in Spec(Mod Z) there is no nonzero morphism from P(Q) to P(Z(p∞)): any such morphism is represented by a span Q←M→Z(p∞) with M essential in Q; choosing an essential submodule M'⊆M∩Z on which the map to Z(p∞) vanishes makes the span equivalent to zero. This confirms ⟨P(Z)⟩ is the subcategory of Q-vector spaces, not all of Spec(Mod Z), so (S1) fails.","verdict_should_be":"REJECT","load_bearing_attack":"Theorem 3.8 asserts that (Spc A, supp) is a support datum, in particular satisfying (S1): supp(X)=Spc A for every generator X. This is false. Take A=Mod Z, which is locally noetherian, so Example 2.6 gives Spec A discrete and S(A)=L(Spec A) the Boolean lattice of subsets of Sp A. The object Z is a generator. Since Z→Q is an essential monomorphism, P(Z)≅P(Q). In Spec(Mod Z), Q is an indecomposable injective with endomorphism ring Q, so P(Q) is a simple object. Hence the localising subcategory ⟨P(Z)⟩ generated by P(Z) corresponds to the subset {Q} of Sp A, not to all of Sp A (which also contains Z(p∞) for every prime p). Therefore supp(Z) is the clopen subset of Spc A corresponding to {Q}, a proper subset of Spc A. Thus (S1) fails and (Spc A, supp) is not a support datum; the universal property in Theorem 3.8 does not hold as stated. The root cause is that P does not send generators of A to generators of Spec A: inverting essential monos kills the information that Z maps to every nonzero Z-module, so the generated localising subcategory in Spec A is only the Q-linear part, not all simples.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a Boolean spectrum Spc A = Spec(L(Spec A)) for every Grothendieck category A, where Spec A is the spectral category obtained by inverting essential monomorphisms. It defines a support supp(X) = <P(X)> in the Boolean lattice L(Spec A), identifies this lattice with the clopen subsets of Spc A via Stone duality, and proves classification theorems: essentially closed subcategories of A correspond to clopen subsets of Spc A (Theorem 3.3), cohomologically stable subcategories correspond to clopen subsets via an exact support (Theorem 5.5), and pure analogues hold for exactly definable categories (Theorems 6.6 and 6.7). The paper also studies direct-sum decompositions of injective objects through a Boolean lattice D(X) (Theorem 4.5) and states a universal property for (Spc A, supp) as a support datum (Theorem 3.8).","tokens_in":23730,"tokens_out":17062,"duration_ms":217472,"significance":"The classification theorems, if correct, are a significant contribution: they provide a uniform Boolean-lattice parameterization of subcategories closed under coproducts, subobjects, and essential extensions, as well as a derived analogue and an extension of Crawley-Boevey's correspondence to pure-essentially closed subcategories. The proofs are largely constructive, with explicit inverse maps in Theorems 3.3, 5.5, and 6.6, and the paper is transparent about its reliance on the standard equivalence (Inj A)/Rad(Inj A) ≅ Spec A. However, the manuscript currently contains concrete errors in the advertised point-set description of Spc A and in the universal support-datum property, so it requires a substantive revision before these claims can be accepted as stated.","major_comments":[{"comment":"The claim that Spec(L((Spec A)_d)) is Sp A with the discrete topology is false in general. By Lemma 2.5, L((Spec A)_d) is the full power-set Boolean lattice P(Sp A), and the Stone spectrum of P(S) is the Stone–Čech compactification βS, not the set S; when S is infinite there are nonprincipal ultrafilters. For example, when A = Mod Z, Spc A is not the countable discrete set Sp A. The lattice-level classifications survive because Clop(βS) ≅ P(S), but the point-set statements in the introduction, in (2.5), and in Proposition 6.9 need to be corrected. In particular, the canonical embedding Sp A → Spc A is not a bijection merely because Spec A is discrete; it is a bijection only when Sp A is finite.","section":"§2, eq. (2.5)"},{"comment":"The pair (Spc A, supp) is not a support datum as defined, because axiom (S1) fails. In A = Mod Z, the object Z is a generator, but the inclusion Z → Q is an essential monomorphism, so P(Z) ≅ P(Q). Since Q is an indecomposable injective with endomorphism ring Q, P(Q) is a simple object of Spec(Mod Z), and supp(Z) = <P(Z)> is the clopen subset of Spc A corresponding to {Q}, not to all of Spc A. Thus (S1) is false, and the universal property asserted in Theorem 3.8 does not follow. The sentence \"It is clear that (Spc A, supp) is a support datum\" in the proof is exactly where the failure occurs. The exact analogue in Theorem 5.9 needs a separate verification, since the spectral functor does not automatically send generators of A to generators of Spec A. A possible repair is to replace (S1) and (E1) by the condition σ(X) = T for objects X with <X> = A (respectively <X>_ex = A), which is what the proof of Lemma 3.7 actually needs.","section":"§3, Definition 3.6 and Theorem 3.8"},{"comment":"Proposition 6.9 is affected by the point-set error in (2.5). The equivalence between \"X = X_d for every pure-injective X\" and \"the embedding Ind Λ → PSpc Λ is a bijection\" cannot be correct as stated: if the pure spectral category is discrete and Ind Λ is infinite, then PSpc Λ is the Stone–Čech compactification of Ind Λ, so the embedding into the full Stone space is not surjective. The proposition should be reformulated in terms of the discrete part of PSpc Λ or in terms of the principal ultrafilters, and the same correction should be carried through Remark 6.12 and the introductory claims about Sp A ≅ Spc A for locally noetherian categories.","section":"§6, Proposition 6.9"}],"minor_comments":[{"comment":"There are several typographical/OCR artifacts, including \"Groth endieck\" and \"ex tensive\" in the abstract and \"integerdivide\" in Remark 3.18; these should be cleaned up in the final version.","section":"Throughout"},{"comment":"It would help the reader if the paper explicitly distinguished between the Boolean lattice L(Spec A), its Stone space Spc A, and the subset of principal ultrafilters corresponding to Sp A. Many statements read as if Spc A were equal to Sp A in the locally noetherian case; this is only true when Sp A is finite.","section":"§2 and §3"},{"comment":"The proof says that conditions (S0)–(S∧) imply that the induced map is a lattice homomorphism, but the treatment of the top element uses (S1). Once (S1) is modified as suggested, this lemma should be rewritten to state explicitly which condition forces the top element to map to T.","section":"§3, Lemma 3.7"}],"recommendation":"major_revision","confidential_remarks":"The core classification theorems appear to be sound and the paper is likely publishable after a careful revision, but the false universal property and the incorrect point-set identification of the Stone spectrum are serious and must be fixed before the claims as stated can be accepted. The dependence on [13] and [27] is standard, but the author should be asked to verify all assertions that identify Spc A with a space of ordinary points."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear Colleague,\n\nQuick take: the Boolean spectrum and the two main classification theorems are a real contribution, but Theorem 3.8 is false as stated. Your stress-test note lands. For A = Mod Z, Z is a generator and Z ↪ Q is essential, so P(Z) ≅ P(Q). In the spectral category, P(Q) is simple, hence supp(Z) is the clopen subset {Q} of Spc A, not Spc A. So (S1) fails and the universal property in Theorem 3.8 does not hold; the proof is only the sentence \"It is clear.\" That is a real bug, not a nitpick.\n\nWhat is genuinely good: Theorem 3.3, the classification of essentially closed subcategories via S(A) = L(Spec A), is new and elegant. It gives a single invariant refining both Gabriel and Ziegler spectra, and the inverse map U ↦ P^{-1}(U) is explicit. The exact support in Theorem 5.5 and the pure versions in Theorems 6.6 and 6.7 look plausible and correctly differentiate themselves from Dauns, Takahashi, and Wu–Ma. The paper is admirably explicit about which statements are new and which are imported.\n\nSoft spots, in addition to the S1 bug: the proof of Lemma 5.4 is compressed; the induction deserves a few more lines. The reliance on the spectral category equivalence (2.1) is heavy, but those are standard facts and I do not doubt them. The author's footnote in Remark 4.3 is unprofessional but harmless.\n\nThe failure of S1 does not destroy the classification theorems, because Theorem 3.8 is not used later except as a conceptual analogue. But it does mean the announced \"universal support theory\" is overstated. The author could fix it by either removing (S1) or finding a different assignment that sends generators to the top; the current definition of supp based on P(‐) cannot satisfy (S1) even in the simplest non-semisimple example.\n\nWho this is for: anyone working on infinite-dimensional module theory, Ziegler spectra, or categorical invariants of Grothendieck categories. A serious referee should engage: the classification results deserve publication, but Theorem 3.8 must be corrected or explicitly weakened before that.\n\nMy vote: accept for peer review, with the expectation of a revision that repairs the support-datum claim.\n\nBest,\n\n[Signature]","headline":"The classification theorems are solid, but Theorem 3.8's support-datum claim is false: for A = Mod Z, the generator Z has support {Q}, not Spc A.","tokens_in":24329,"tokens_out":5668,"would_cite":false,"duration_ms":185711,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18E10","16D70","16E50","18E40","18E45"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that in any Grothendieck category, subcategories closed under arbitrary coproducts, subobjects, and essential extensions correspond exactly to clopen subsets of a new Boolean spectrum, and each such subcategory is the…","keywords":["Boolean lattice","Boolean spectrum","Grothendieck category","spectral category","support","exact support","definable subcategory","Ziegler spectrum"],"falsifier":"Find a nonzero object X in a Grothendieck category whose support in Spc A is empty, or find two objects with equal support but different essentially closed closures. Either observation would contradict Theorem 3.3 and its corollary that every essentially closed subcategory is singly generated.","tokens_in":23260,"feed_emoji":"🧩","tokens_out":10220,"duration_ms":88727,"temperature":0.7,"pith_summary":"The paper introduces a support for every object of any Grothendieck category—a class of abelian categories with coproducts and exact filtered colimits that includes module categories and categories of sheaves. The support of an object is a clopen subset of a new space, the Boolean spectrum, built from the category's spectral category. The main result is a bijection: subcategories closed under arbitrary coproducts, subobjects, and essential extensions correspond exactly to clopen subsets of the Boolean spectrum, and every such subcategory is the support of a single object. A derived variant, exact support, classifies the cohomologically stable thick subcategories. If the paper is right, decomposition questions about objects can be read off a complete Boolean lattice, and the classical correspondence between definable subcategories and closed subsets of the Ziegler spectrum extends to all pure-essentially closed subcategories.","feed_headline":"The Boolean spectrum classifies all essentially closed subcategories","feed_subtitle":"One support function now classifies subcategories closed under coproducts, subobjects, and essential extensions.","key_machinery":"The central object is the spectral category Spec A = A[$Ess^{{-1}}$], the category obtained by formally inverting all essential monomorphisms; every short exact sequence in it splits. Its lattice of localising subcategories is a complete Boolean lattice, with complements coming from swapping the two halves of a torsion pair, and Stone duality turns this lattice into the Boolean spectrum Spc A. The identification (Inj A)/Rad(Inj A) ≅ Spec A and the consequence P(X) ≅ P(Y) if and only if E(X) ≅ E(Y) are what let decompositions in the spectral category be lifted to decompositions of injective objects; this is the mechanism that makes the support map both total and faithful, so the support of a nonzero object is never empty.","core_discovery":"On the paper's own terms, the central discovery is that the spectral category Spec A, obtained from A by inverting all essential monomorphisms, carries a complete Boolean lattice L(Spec A) of localising subcategories, and its Stone space Spc A = Spec(L(Spec A)) is a universal target for support. Each object X gets supp(X) = ⟨P(X)⟩, and the assignment C ↦ supp(C) = P(C) is a lattice isomorphism from essentially closed subcategories to S(A) = L(Spec A), with inverse U ↦ $P^{{-1}}$(U). The paper proves that (Spc A, supp) is the initial support datum: any other support map into clopen subsets of a space factors uniquely through it. It then proves an exact version using the right derived functor of P, classifying cohomologically stable subcategories, and a pure version for exactly definable categories, extending the correspondence between definable subcategories and Ziegler-closed sets.","pith_inferences":["Editorial inference: since Spc A is a Stone space attached functorially to A, it can serve as a new comparison invariant for Grothendieck categories; non-homeomorphic Boolean spectra would certify non-equivalent categories even when the classical spectra of indecomposable injectives agree.","Editorial inference: the universal property suggests a way to relate the Boolean spectrum to tensor-triangulated support theories when A carries a monoidal structure, since the tensor product of objects should correspond to intersection of supports; one could test whether the resulting map from the tensor-triangulated spectrum is a homeomorphism onto a closed subspace of Spc A.","Editorial inference: the decomposition lattice D(X) gives a practical test for the paper's open problem on strictly wild algebras—construct pure-injective modules whose summand lattices realise prescribed complete Boolean lattices, and the classification predicts exactly which lattices are possible.","Editorial inference: for categories whose classical spectrum of indecomposable injectives is empty, the paper's framework predicts a nonempty Boolean spectrum with enough points; computing Spc A for such a category would give the first concrete support theory on it."],"forward_implications":["Every essentially closed subcategory of a Grothendieck category is generated by a single object, namely any object whose support is the corresponding clopen set.","The Boolean spectrum is universal among all spaces carrying a support datum, so any other support theory on the same category factors uniquely through Spc A.","For injective objects, direct summands up to essential equivalence form a complete Boolean lattice, isomorphic to the central idempotents of End(X)/J(End(X)); coproduct decompositions are therefore governed by Boolean rings.","For module categories over any ring, pure-essentially closed and pure-cohomologically stable subcategories are classified by clopen subsets of the pure Boolean spectrum, extending the classical definable-subcategory correspondence with the Ziegler spectrum.","The frame of localising subcategories embeds into the Boolean spectrum's clopen subsets, so the Boolean spectrum has enough points even when the original localising frame has none."],"supporting_citations":[{"why":"Introduces the spectral category of a Grothendieck category and the criterion P(X)≅P(Y) iff E(X)≅E(Y), which the support construction uses.","marker":"[13, Satz 1.5]"},{"why":"Supplies the equivalence (Inj A)/Rad(Inj A) ≅ Spec A and the lifting of decompositions used in Lemma 2.1 and Theorem 4.5.","marker":"[27, Proposition 2.5.9]"},{"why":"Provides the classical spectrum of indecomposable injectives and the localising-subcategory correspondence that the Boolean spectrum extends.","marker":"[12]"},{"why":"Gives the definable-subcategory to Ziegler-closed-subset correspondence for module categories that Theorems 6.6 and 6.7 generalise.","marker":"[7]"},{"why":"Introduces the Ziegler topology and the indecomposable pure-injective spectrum used as the target of the module-theoretic extension.","marker":"[40]"},{"why":"Defines the notion of support datum and its universal property, which Theorem 3.8 adapts to Grothendieck categories.","marker":"[1]"},{"why":"Provides the classification of localising subcategories of derived categories that motivates and anchors the exact support comparison.","marker":"[30]"}],"fun_headline_variants":["Boolean spectrum yields universal support for Grothendieck categories","One space classifies all essentially closed subcategories","Support via Boolean lattice: a complete classification","Beyond Ziegler: Boolean spectrum for definable subcategories"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the spectral category of a Grothendieck category is the same thing as its injective objects up to the coarsest reasonable equivalence; if that identification were wrong, the classification's inverse map would not be defined.","fun_headline_variants_meta":{"raw":{"variants":["Boolean spectrum yields universal support for Grothendieck categories","One space classifies all essentially closed subcategories","Support via Boolean lattice: a complete classification","Beyond Ziegler: Boolean spectrum for definable subcategories"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000567,"raw_usage":{"total_tokens":2635,"prompt_tokens":847,"completion_tokens":1788,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":463,"completion_tokens_details":{"reasoning_tokens":1725}},"tokens_in":463,"tokens_out":1788,"duration_ms":13435,"temperature":1.0,"reasoning_tokens":1725,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:16:54.003950+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a nonzero object X in a Grothendieck category whose support in Spc A is empty, or find two objects with equal support but different essentially closed closures. Either observation would contradict Theorem 3.3 and its corollary that every essentially closed subcategory is singly generated.","supporting_citations":[{"cited_title":"Gabriel, Des cat´ egories ab´ eliennes, Bull","cited_arxiv_id":null,"evidence_quote":"Provides the classical spectrum of indecomposable injectives and the localising-subcategory correspondence that the Boolean spectrum extends."},{"cited_title":"Crawley-Boevey, Infinite-dimensional modules in the representation theory of finite- dimensional algebras, in Algebras and modules, I (Trondheim, 1996) , 29–54, CMS Conf","cited_arxiv_id":null,"evidence_quote":"Gives the definable-subcategory to Ziegler-closed-subset correspondence for module categories that Theorems 6.6 and 6.7 generalise."},{"cited_title":"Ziegler, Model theory of modules, Ann","cited_arxiv_id":null,"evidence_quote":"Introduces the Ziegler topology and the indecomposable pure-injective spectrum used as the target of the module-theoretic extension."},{"cited_title":"Balmer, The spectrum of prime ideals in tensor triangulated categor ies, J","cited_arxiv_id":null,"evidence_quote":"Defines the notion of support datum and its universal property, which Theorem 3.8 adapts to Grothendieck categories."},{"cited_title":"Neeman, The chromatic tower for D(R), Topology 31 (1992), no","cited_arxiv_id":null,"evidence_quote":"Provides the classification of localising subcategories of derived categories that motivates and anchors the exact support comparison."}],"review_version":1}