{"id":"e9849635-4cd3-448a-a3c7-1e41c0c9bfe6","arxiv_id":"2607.29074","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"PGF objects in a functor category are exactly the diagrams whose pointwise values are PGF modules, yielding a unified hereditary abelian flat model structure whose homotopy category is the Q-shaped derived category.","lead":"Pure mathematics paper proving that PGF objects — a broadly useful class of diagrams — are completely controlled by their values at each point of the diagram, and using that to build flat model structures, the framework for doing homotopy theory on diagram categories. A generalist should care because it unifies previously separate constructions in homological algebra, subsuming the classical flat model for chain complexes and known Q-shaped variants.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 1.20's proof is the load-bearing seam: it reduces arbitrary injectives to a single product ∏_q G'_q(I), skipping multiplicities; products with repeated factors need not equal coproducts, so the objectwise-to-global Tor step behind Theorem B is not justified.","rationale":"The central claim is the objectwise characterization of PGF objects and its use in the flat model structure. The cleanest load-bearing step is Lemma 1.20. I checked the Hovey-triple part: Theorem 5.6 really does reduce to Lemma 5.5, and the equality ⊥(Cot∩W_G)∩W_G=Flat is a standard splitting argument; Theorem 2.5 is a clean consequence of [11, Thm 7.29] and Lemma 2.4. The PGF bridge, however, rests on Lemma 1.20, and its proof has an unaddressed multiplicity problem: cogeneration by the set {G'_q(I)} only produces a product of copies indexed by a possibly repeated family, not a single copy per q; [12, Prop 3.7] requires local finiteness, which fails for repeated factors. The one-object category shows the proof's asserted reduction is false as written, though not the lemma itself. Since Theorem 4.2's reverse inclusion and Theorems B/C depend on this lemma, this is the weakest assumption. I am not claiming the theorem is false; I am claiming the written proof is incomplete at this seam. The reader's CONDITIONAL verdict is appropriate; no change.","tokens_in":18713,"tokens_out":28425,"duration_ms":302353,"concrete_test":"Independently derive Lemma 1.20 without the single-product reduction. Concretely: take C with one object and endomorphism ring k, A=k, X arbitrary, and let E=I^{(ℕ)}; verify that the proof's assertion 'E is a summand of ∏_{q∈C}G'_q(I)' fails, while the lemma holds. Then check whether [11, Prop 3.12] (or a direct argument) implies every injective E in Mod C,A is a direct sum (not just a product summand) of G'_q(I)'s. If it is, replace the product=coproduct step by additivity of Tor; if not, Theorem 4.2 lacks a proof of the reverse inclusion, and Theorems B/C are unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 1.20 is what converts objectwise PGF into global PGF in Theorem B/4.2. In (i)=>(ii), the proof takes an injective E and says it is a summand of ∏_{q∈C} G'_q(I), then replaces product by coproduct via local boundedness. But [11, Prop 3.12] only gives that the class {G'_q(I)} cogenerates: an arbitrary injective E is a summand of a product of possibly many copies of these objects. If a given q occurs with infinite multiplicity, the family is not locally finite, and [12, Prop 3.7] cannot be invoked. In the one-object case C=pt, A=k, G'_q(I)=I and E=I^{(ℕ)} is injective but is not a summand of a single I; the proof's claimed reduction is literally false there (the lemma itself survives tautologically). Since Theorem 4.2's reverse inclusion needs exactly this lemma to pass from Tor_1^A(I,X(q))=0 to Tor_1^{C,A}(E,X)=0, and since Theorem C needs S_q(A) to be PGF, this gap is load-bearing. The statement may be true by a different argument, but the text as written does not supply it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a unified method for constructing flat model structures on functor categories C,AMod using projectively coresolved Gorenstein flat (PGF) objects. The main results are: Theorem A (existence of a hereditary abelian model structure whose trivial cofibrant objects are the flat objects, plus a homology/cokernel characterization of flat objects), Theorem B (PGF objects in C,AMod are exactly the functors that are objectwise PGF A-modules), Theorem C (a flat model structure on Q-shaped diagram categories whose homotopy category is the Q-shaped derived category), and Theorem D (an explicit description of the cofibrant objects under a finite weak global dimension assumption). The proof strategy uses Hovey triples, left Frobenius pairs, and the theory of PGF modules from Šaroch–Šťovíček.","tokens_in":18863,"tokens_out":23863,"duration_ms":212026,"significance":"If the results hold, the paper provides a clean, uniform framework that simultaneously recovers the classical flat model structure on chain complexes and the Q-shaped flat model structure, with a tractable description of cofibrant objects. The objectwise characterization of PGF objects (Theorem B) is a strong and useful result, and the construction via left Frobenius pairs is conceptually appealing. The paper also honestly records its limitations (Remark 6.12). The main technical concern is the proof of Lemma 1.20, which is load-bearing for Theorems A and B and currently has a gap.","major_comments":[{"comment":"The proof asserts that an arbitrary injective E in Mod C,A is a direct summand of the single product ∏_{q∈C} G'_q(I), citing [11, Proposition 3.12]. That proposition only provides that the family {G'_q(I)} cogenerates the category, meaning E is a summand of a product with possibly repeated copies of the same G'_q(I). The subsequent replacement of the product by the coproduct via [12, Proposition 3.7] requires local finiteness, which fails when a given q occurs infinitely often. For instance, when C=pt and A=k, E=I^{(ℕ)} is injective but is not a summand of a single I. Since Lemma 1.20 is used in Theorem 4.2 to pass from objectwise Tor vanishing to global Tor vanishing, this gap is load-bearing for Theorem B and its applications. The statement may be repairable by grouping repeated factors as ∏_q G'_q(I^{M_q}) and applying the objectwise hypothesis to the injective module I^{M_q}; the pro","section":"Lemma 1.20, proof of (i)=>(ii)"},{"comment":"The equality W_G = {X | Ext^1_{Q,A}(S_q(A),X)=0 for all q} is imported from [17, Lemma 5.1], which is an unpublished preprint. The text's 'indeed' only proves Ext^1_{Q,A}(S_q(A),X) ≅ H^1_[q](X); it does not justify why the higher syzygies Ω^n S_q(A) appearing in the definition of W_G impose no extra conditions. This equality is essential for the identification W_G = E in Theorem 6.4, so the proof is incomplete as written. Please either give a full proof of the equality or replace the reference with a published source.","section":"Lemma 6.2"}],"minor_comments":[{"comment":"The running header and PDF title contain the typo 'FLA T MODELS' instead of 'FLAT MODELS'.","section":"Title/header"},{"comment":"Lemma 6.2 applies Lemma 1.17(a), which requires A to have finite projective dimension over k. In Setup 6.1 this follows from k being hereditary, but the hypothesis is not stated explicitly; please add a remark.","section":"Lemma 1.17 / Setup 6.1"},{"comment":"The notation Ω^n G for n∈Z is used freely. A brief explanation that Ω^n for negative n is defined via the right half of the complete resolution would improve readability.","section":"Definition 5.1 / Lemma 5.4"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely correct in its main claims, but the gap in Lemma 1.20 is central and must be fixed before publication. The reliance on an unpublished preprint in Lemma 6.2 should also be addressed. I recommend major revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The first thing to know: the main theorems are probably right, but the proof of Lemma 1.20 as written does not work. The paper gives a clean objectwise description of PGF objects in functor categories (Theorem B) and uses it to build hereditary abelian flat model structures (Theorem 5.6), subsuming known constructions. Theorem 3.7, lifting left Frobenius pairs from A to functor categories, is elegant and likely to be useful independently.\n\nWhat is genuinely new: the objectwise characterization of PGF objects, the Frobenius-pair lifting, and the unified construction of flat model structures. The paper is well organized and honest: it states explicitly that the main application reobtains the authors' earlier flat model, and Remark 6.12 flags the finite weak-global-dimension restriction.\n\nThe soft spots are concentrated in two places. Lemma 1.20 is the bridge from objectwise Tor vanishing to global Tor vanishing; Theorem B and Theorem 4.2 both lean on it. In the proof of (i)=>(ii), the authors take an injective object E, invoke cogeneration by the objects G'_q(I), and then assert that E is a summand of the single product ∏_q G'_q(I), which they identify with the coproduct using local boundedness. That reduction skips multiplicities. Cogeneration gives E as a summand of a product of possibly many copies of the G'_q(I); a product with repeated factors is not generally the same as one copy per q. The one-object case C=pt, A=k shows the claimed identification is literally false: E = I^{(ℕ)} is injective but not a summand of a single I. The lemma itself is still true in that case, so the statement may be salvageable, but the proof as written does not establish it.\n\nThe other soft spot is Lemma 6.2, where the key equality W_G = E is delegated to [17, Lemma 5.1], an unrefereed preprint; the surrounding computation does some work but the initial reduction is one line.\n\nIf Lemma 1.20 can be repaired—say, by a direct argument using local finiteness or finite filtration—the central claims look sound. The proof architecture is coherent, and the Hovey-triple verification in Section 5 checks out.\n\nThis paper is for people working on Q-shaped derived categories, flat model structures, and Gorenstein objects. It deserves a serious referee, and the referee should be asked to verify Lemma 1.20 and the dependence on [17]. My own verdict: conditional acceptance after those points are resolved.","headline":"Genuinely useful objectwise PGF criterion and a unified flat model construction, but Lemma 1.20's proof has a load-bearing gap around products of injectives that should be fixed before the main theorems are cited.","tokens_in":19579,"tokens_out":4364,"would_cite":true,"duration_ms":42000,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18G25","18A25"],"pacs":[],"model":"deepseek-v4-flash","headline":"A diagram is PGF exactly when its values at each object are PGF, and this pointwise test yields flat model structures on diagram categories.","keywords":["Q-shaped derived category","abelian model structure","projectively coresolved Gorenstein flat object","PGF objects","flat model structure","functor categories","cotorsion pairs","left Frobenius pairs"],"falsifier":"Take C to be the path category of a finite cycle quiver (Hom-finite, locally bounded, with Serre functor), A a finite-dimensional algebra, and compute whether every objectwise-PGF diagram is globally PGF; a single diagram whose objectwise components are PGF but which has nonzero Tor against some injective diagram would disprove Theorem 4.2.","tokens_in":18404,"feed_emoji":"⚙️","tokens_out":8895,"duration_ms":71645,"temperature":0.7,"pith_summary":"This paper establishes that projectively coresolved Gorenstein flat (PGF) objects in a diagram category over a Hom-finite, locally bounded category with a Serre functor are completely determined by their objectwise components: a diagram is PGF if and only if each of its values is a PGF module. This characterization is used to construct hereditary abelian model structures on diagram categories whose trivial cofibrant objects are precisely the flat objects. Applied to Q-shaped diagram categories satisfying a strong retraction property, the construction recovers a flat model structure whose homotopy category is the Q-shaped derived category, and, when the coefficient ring has finite weak global dimension, identifies the cofibrant objects as exactly the objectwise flat diagrams. A sympathetic reader cares because this gives a unified, checkable route to flat model structures that subsumes the classical flat model on chain complexes.","feed_headline":"A diagram is PGF exactly when each of its values is PGF","feed_subtitle":"This objectwise test yields flat model structures on Q-shaped derived categories and describes their cofibrant objects.","key_machinery":"The central objects are projectively coresolved Gorenstein flat (PGF) objects: objects appearing as cokernels in an exact sequence of projectives that remains exact after tensoring with any injective object. The argument is carried by three pieces of machinery: the left Frobenius pair (Fun(C,PGF(A)), ^Prj(A)) in C,AMod, which lifts the PGF/projective pair from modules to diagrams; the bridge lemma (1.20), which equates objectwise Tor-vanishing against injective modules with global Tor-vanishing against injective diagrams using Hom-finiteness, local boundedness, and a Serre functor; and the cotorsion-pair/Hovey-triple formalism that converts these ingredients into a hereditary abelian model s","core_discovery":"The core discovery is Theorem 4.2: for C Hom-finite, locally bounded, with a Serre functor, an object of the diagram category is PGF precisely when its value at every object is a PGF module. The proof lifts the left Frobenius pair (PGF(A), Prj(A)) to (Fun(C,PGF(A)), ^Prj(A)) and uses the Tor-vanishing bridge lemma (1.20) to pass from objectwise to global injective-Tor vanishing. Theorem 5.6 then yields, for any set G of PGF diagrams, a hereditary abelian model structure (⊥(Cot(C,A)∩W_G), W_G, Cot(C,A)) whose trivial cofibrant objects are exactly the flat diagrams. In the Q-shaped case, G={S_q(A)} makes the trivial objects exactly the exact diagrams, so the homotopy category is the Q-shaped d","pith_inferences":["We infer that the same objectwise-lifting mechanism might characterize other Gorenstein-type properties (e.g., Gorenstein projective objects) in diagram categories, by replacing the left Frobenius pair with the corresponding pair for that property.","We speculate that the Serre functor hypothesis in Lemma 1.20 could be relaxed to a weaker duality if one only needs the Tor-vanishing equivalence for a restricted class of injectives; a testable extension is to check whether Theorem B still holds for index categories that are Hom-finite and locally bounded but lack a full Serre functor.","The theorem leaves open the behavior of cofibrant objects when the coefficient ring has infinite weak global dimension; constructing an explicit diagram that is objectwise flat but not cofibrant over such a ring would delineate the boundary of Theorem 6.9.","Since the model structure is hereditary, its homotopy category is triangulated, and one could try to identify it with a known triangulated category of Q-shaped complexes; a natural next question is whether the cofibrant-object description upgrades to a Quillen equivalence with a more classical derived category."],"forward_implications":["PGF-ness of a diagram can be checked objectwise; no global resolution is needed to test the property.","For any set G of PGF diagrams, one obtains a hereditary abelian model structure on C,AMod whose trivial cofibrant objects are exactly the flat diagrams.","In the Q-shaped case, the trivial objects are exactly the exact diagrams, so the homotopy category is the Q-shaped derived category, and the model structure is hereditary abelian.","When the coefficient ring has finite weak global dimension, the cofibrant objects of the Q-shaped flat model are exactly the objectwise flat diagrams.","For the path category of the quiver with ∂^2=0, this recovers the classical flat model structure on chain complexes, where cofibrant objects are dg-flat complexes."],"fun_headline_variants":["PGF diagrams are pointwise PGF: flat models on Q-shaped categories","Objectwise PGF test yields flat model structures for diagram categories","PGF iff every component is PGF: flat model structures on diagrams","A pointwise PGF test yields flat model structures on Q-shaped categories","Flat model structures from a simple pointwise test on PGF objects"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The whole argument depends on Lemma 1.20's bridge: vanishing of Tor against every injective module at each object is equivalent to vanishing against every injective diagram, a step that requires the index category to be Hom-finite, locally bounded, and equipped with a Serre functor; if that equivalence fails, the objectwise PGF criterion and the model structure collapse.","fun_headline_variants_meta":{"raw":{"variants":["PGF diagrams are pointwise PGF: flat models on Q-shaped categories","Objectwise PGF test yields flat model structures for diagram categories","PGF iff every component is PGF: flat model structures on diagrams","A pointwise PGF test yields flat model structures on Q-shaped categories","Flat model structures from a simple pointwise test on PGF objects"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000923,"raw_usage":{"total_tokens":3763,"prompt_tokens":679,"completion_tokens":3084,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":423,"completion_tokens_details":{"reasoning_tokens":2990}},"tokens_in":423,"tokens_out":3084,"duration_ms":21460,"temperature":1.0,"reasoning_tokens":2990,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T14:06:53.168011+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take C to be the path category of a finite cycle quiver (Hom-finite, locally bounded, with Serre functor), A a finite-dimensional algebra, and compute whether every objectwise-PGF diagram is globally PGF; a single diagram whose objectwise components are PGF but which has nonzero Tor against some injective diagram would disprove Theorem 4.2.","supporting_citations":[],"review_version":1}