{"id":"ddaea0e6-15be-46c7-b623-b48ace4c96a1","arxiv_id":"2501.15898","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Fibrant model structures on weakly idempotent complete additive categories are equivalent to fibrantly weak factorization systems, and the homotopy category is the additive quotient of bifibrant objects by trivial bifibrant objects.","lead":"This paper proves that the homotopy category of a model structure on an additive category can be computed as a quotient of the bifibrant objects by the trivial bifibrant objects, and it gives a complete recipe for building fibrant model structures from simpler factorization data. The recipe unifies two known families of model structures used in algebra and representation theory, the omega-model structures and the W-model structures.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the flagged 'epic' ambiguity is not load-bearing, and the two-out-of-three proof in §4.3 is complete as written.","rationale":"The reader's conditional verdict rests on two points: an alleged ambiguity in 'Hom_A(U,f) is epic' and a supposed missing derivation in the two-out-of-three proof of Section 4.3. The first point does not survive scrutiny: Hom_A(U,0) is always a singleton, so the map X->0 induces a map to a singleton, which is surjective in both the Set and Ab readings. Moreover, epimorphisms in Ab coincide with surjective group homomorphisms, hence with surjective set maps, so no truth-value changes between readings. The second point is also not a real gap: the two-out-of-three proof is detailed, and the step 'gp ∈ Weq, as proven before' correctly invokes the earlier proof that Weq is closed under composition, which is exactly the first case of the two-out-of-three axiom. I checked the uses of Lemma 4.6, Lemma 4.7, and Lemma 4.8 in the three cases; all applicable hypotheses are satisfied and the cancellations are valid. The retract axiom proof for Fib is valid because Hom(U,ψ1) is split epic and the equality Hom(U,g)∘Hom(U,ψ1)=Hom(U,ψ2)∘Hom(U,f) together with Hom(U,ψ2) and Hom(U,f) epic forces Hom(U,g) to be epic. No circularity appears in Theorem 1.4 or Theorem 1.5: Lemma 4.5 establishes TFib=Fib∩Weq without assuming the bijection, and the two directions of the correspondence are independent. I therefore identify no significant objection that would change the reader's verdict. The paper might benefit from an explicit note that 'epic' here means 'surjective' in the relevant concrete category, but that is a readability suggestion, not a correctness issue.","tokens_in":25546,"tokens_out":41599,"duration_ms":359427,"concrete_test":"Check that for every object U in a pointed category the set Hom_A(U,0) has exactly one element, and verify that the zero homomorphism G->0 is surjective for every abelian group G. If both hold, the 'epic' convention cannot change the validity of Theorem 1.2(3), Theorem 1.4, or the W-model examples.","verdict_should_be":"UNCHANGED","load_bearing_attack":"After a full pass over the central arguments, I find no load-bearing flaw in the classification theorem (Theorem 1.5) or in the homotopy-category description (Theorem 1.1). The reader's weakest assumption concerns whether 'Hom_A(U,f) is epic' means surjective as a set map or as a homomorphism of abelian groups. This is not a genuine ambiguity: in any pointed category Hom_A(U,0) is a singleton, so the induced map Hom_A(U,X)->Hom_A(U,0) is a map to a one-element set. Such a map is surjective in both the Set and Ab readings. More generally, epimorphisms in Ab are exactly surjective group homomorphisms, which are exactly the surjective set maps in this context, so the two readings never disagree. The W-model and ω-model examples are therefore unaffected. The two-out-of-three proof in §4.3 is present and internally consistent: the step 'gp ∈ Weq, as proven before' refers to the previously established closure of Weq under composition, and the applications of Lemmas 4.6, 4.7, and 4.8 are valid. The only presentational point is that the paper could explicitly state 'surjective' next to 'epic' for readability, but this does not affect correctness.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies model structures on weakly idempotent complete additive categories. Theorem 1.1 identifies the homotopy category Ho(A) with the additive quotient (C∩F)/(C∩F∩W) of cofibrant-fibrant objects by cofibrant-fibrant-trivial objects. The authors introduce fibrantly weak factorization systems and prove (Theorems 1.2, 1.4, 1.5) that fibrant model structures are in bijection with these systems, with fibrations given by Hom_A(TC,−)-epimorphisms, trivial cofibrations by split monomorphisms with cokernel in TC, and weak equivalences by TFib∘TCoFib. Section 5 applies the framework to rediscover ω-model structures and W-model structures and discusses their relation to exact model structures.","tokens_in":25794,"tokens_out":34174,"duration_ms":332471,"significance":"Assuming the main theorems, this is a valuable contribution. It provides an additive description of homotopy categories of model structures on weakly idempotent complete additive categories, extending known results for exact model structures and ω-model structures. The classification of fibrant model structures by fibrantly weak factorization systems is clean and genuinely useful, and the applications to ω- and W-model structures illustrate the framework well. The proofs in Section 4 are detailed and internally coherent; in particular, the two-out-of-three argument in §4.3 is complete as written, and the verification of the retract, lifting, and factorization axioms from the fibrantly weak factorization system data is sound. The main limitation of the manuscript is presentational: several conventions and diagrams need clarification, but I found no load-bearing mathematical error in the central arguments.","major_comments":[],"minor_comments":[{"comment":"The paper uses the phrase \"Hom_A(U,f) is epic\" before explicitly defining the convention. Since Hom_A(U,−) takes values in abelian groups, epimorphisms in Ab coincide with surjective homomorphisms and with surjective set maps, so the ambiguity is not mathematically dangerous; nevertheless, the authors should state explicitly that they mean surjectivity of the induced set map, for readability.","section":"Section 1.2, around Eq. (1.1)"},{"comment":"In the first lifting step of Lemma 4.8, the text states that the lifting s satisfies \"su = IdX and vus = v\". The condition \"vus = v\" appears to be a typesetting error, since the argument uses only a section of u; it should likely read \"vs = vu\" or be removed. Please correct the displayed diagram and the surrounding sentence.","section":"Lemma 4.8"},{"comment":"In Definition 1.3(3), the symbol L is used both for a class of morphisms and, immediately afterward, for the object class {X | 0→X ∈ L}; the same is true for R. This double use is confusing; please introduce separate notation such as L^ob and R^ob.","section":"Definition 1.3(3)"},{"comment":"In the factorization axiom part of the converse direction, the proof explicitly constructs only the factorization f = p∘i with i ∈ TCoFib and p ∈ Fib. The companion factorization f = q∘j with j ∈ CoFib and q ∈ TFib holds by the weak factorization system (CoFib,TFib) together with Lemma 4.5, but this should be stated explicitly so that the verification of the model-category factorization axiom is visibly complete.","section":"Proof of Theorem 1.4, §4.3"},{"comment":"The abstract and several sentences in the introduction contain grammatical errors, e.g., \"Fibrant model structures is explicitly described by trivial cofibrations\". A careful language edit is recommended before publication.","section":"Abstract and Introduction"}],"recommendation":"minor_revision","confidential_remarks":"The central theorems appear sound and the proofs are largely complete. The main work for a revision is expository: clarify the epic convention, fix diagram typos in Lemma 4.8, and disambiguate notation in Definition 1.3. I would also ask the authors to clearly distinguish, in Section 5, which results are original to this paper and which are quoted from the overlapping preprint [CLZ] and from [P] and [Bel]; this is relevant for the editor's assessment of novelty and attribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid paper with a real new structural result. Theorem 1.1 gives the homotopy category as C∩F / C∩F∩W for any model structure on a weakly idempotent complete additive category, extending what was known for exact and ω-model structures. The genuinely new piece is Theorem 1.5: fibrant model structures are classified by fibrantly weak factorization systems. That is a clean, useful correspondence, and the construction of Fib and Weq from (CoFib, TFib) is explicit enough to use.\n\nThe main proofs are careful and largely check out. Section 3's proof of Theorem 3.1 is long because it avoids the homotopy relation, but the replacement argument with C∩F objects is coherent. Section 4's construction of the model structure is structurally sound. I checked the point the reader flagged: \"Hom_A(U,f) is epic\" cannot meaningfully be read as a group-homomorphism condition that disagrees with set-level surjectivity. In Ab an epimorphism is exactly a surjective group hom, and for maps to Hom_A(U,0) both readings give surjectivity. The two-out-of-three proof in §4.3 is complete; the \"as proven before\" step refers to the composition case just handled. So the stress-test note is right.\n\nSoft spots are minor. The paper leans on the overlapping-author preprint [CLZ] for Theorem 5.1 and Theorem 5.2, and those results are stated without self-contained proofs. That is acceptable because the central theorems do not depend on [CLZ], but a referee should ask the authors to either prove Theorem 5.1 or clearly mark it as external. The paper is also honest: Theorems 5.3 and Corollary 5.4 are credited to Pirashvili and Beligiannis, and the abstract says \"rediscover.\" That is the right call.\n\nWho is this for? People working in model structures on additive and exact categories, cotorsion pairs, and representation theory. The examples in Section 5 are useful. It deserves a serious referee; I would engage with it. Recommendation: send it to peer review.","headline":"Solid new classification of fibrant model structures on weakly idempotent complete additive categories, with the reviewer's two concerns resolving on close reading; worth sending to a serious referee.","tokens_in":26303,"tokens_out":2552,"would_cite":true,"duration_ms":23509,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18N40","18N55","18E35"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that every fibrant model structure on a weakly idempotent complete additive category is classified by a fibrantly weak factorization system, and that homotopy categories are additive quotients of cofibrant-fibrant objects.","keywords":["model structures","homotopy categories","weakly idempotent complete additive categories","fibrant model structures","fibrantly weak factorization systems","additive quotient categories","exact categories","cotorsion pairs"],"falsifier":"In an artin algebra module category, take $W=\\mathrm{add}(M)$ for a nonzero module $M$ and use the W-model structure of Section 5.2; if 'epic' is read as a surjective homomorphism of abelian groups, then $X\\to 0$ is not a fibration for $X=M$, since $\\mathrm{Hom}_A(M,M)\\neq 0$, directly contradicting the Theorem 1.2(3) assertion that every object is fibrant.","tokens_in":25310,"feed_emoji":"🧩","tokens_out":19464,"duration_ms":152625,"temperature":0.7,"pith_summary":"Working in weakly idempotent complete additive categories (additive categories in which every splitting monomorphism has a cokernel, or equivalently every splitting epimorphism has a kernel), this paper establishes two structural facts. Theorem 1.1 says that for any model structure the homotopy category $\\mathrm{Ho}(\\mathcal{A})$ is equivalent as an additive category to the additive quotient $(C\\cap F)/(C\\cap F\\cap W)$ of cofibrant-fibrant objects by the cofibrant-fibrant-trivial objects. Theorems 1.4 and 1.5 go further for fibrant model structures, where every object is fibrant: they are classified by fibrantly weak factorization systems, with fibrations exactly the $\\mathrm{Hom}_{\\mathcal{A}}(TC,-)$-epic morphisms and weak equivalences the composites $\\mathrm{TFib}\\circ\\mathrm{TCoFib}$. A sympathetic reader would care because this turns the construction of model structures into a check on factorization data and recovers the known $\\omega$-model and $W$-model structures as special cases.","feed_headline":"Every fibrant model structure comes from one factorization system","feed_subtitle":"A bijection classifies fibrant model structures; homotopy categories become additive quotients of cofibrant-fibrant objects.","key_machinery":"The load-bearing mechanism is the fibrantly weak factorization system $(\\mathrm{CoFib},\\mathrm{TFib})$: a weak factorization system whose right class is right-cancellative (if $f$ and $gf$ lie in $\\mathrm{TFib}$, then so does $g$) and whose class $TC = \\{X \\mid 0\\to X\\in \\mathrm{CoFib},\\ X\\to 0\\in \\mathrm{TFib}\\}$ is contravariantly finite, meaning every object has a right approximation by an object of $TC$. From this datum the paper defines $\\mathrm{Fib}$ as the class of morphisms $f$ for which $\\mathrm{Hom}_{\\mathcal{A}}(U,f)$ is surjective for every $U\\in TC$, defines $\\mathrm{TCoFib}$ as the splitting monomorphisms with cokernel in $TC$, and sets $\\mathrm{Weq} = \\mathrm{TFib}\\circ\\mathrm{TCoFib}$. The proofs of Theorems 1.4 and 1.5 show that these three classes satisfy the retract, lifting, factorization, and two-out-of-three axioms. A complementary mechanism is the localization fact that in additive categories, localization along the morphisms inverted by an additive quotient is isomorphic to that quotient, which converts $\\mathrm{Ho}(\\mathcal{A})$ into $(C\\cap F)/(C\\cap F\\cap W)$.","core_discovery":"The central claim is that fibrant model structures on a weakly idempotent complete additive category are exactly the fibrantly weak factorization systems. Theorem 1.5 states this as a bijection: from $(\\mathrm{CoFib},\\mathrm{TFib})$ one forms the class $\\mathrm{Fib}$ of morphisms $f$ such that $\\mathrm{Hom}_{\\mathcal{A}}(TC,f)$ is epic, and sets $\\mathrm{Weq} = \\mathrm{TFib}\\circ\\mathrm{TCoFib}$, where $TC$ is the class of objects both in the left class on $0\\to X$ and in the right class on $X\\to 0$, and $\\mathrm{TCoFib}$ is the class of splitting monomorphisms with cokernel in $TC$; the inverse sends a fibrant model structure to $(\\mathrm{CoFib}, \\mathrm{Fib}\\cap \\mathrm{Weq})$. The paper uses 'epic' in the sense that the induced map of Hom-sets is surjective, the convention fixed just before display (1.1). The paper also proves in Theorem 1.1 that for any model structure the homotopy category is the additive quotient $(C\\cap F)/(C\\cap F\\cap W)$, a description that needs neither pushouts, pullbacks, nor homotopy relations. These two results together say that in this setting the homotopy category is determined by simple object-level data, and fibrant model structures are completely determined by their trivial cofibrations.","pith_inferences":["Editorial extension: The bijection gives a construction recipe: choose a contravariantly finite class $TC$ and a right-cancellative weak factorization system $(\\mathrm{CoFib},\\mathrm{TFib})$; the displayed formulas then produce a fibrant model structure, so the remaining work is checking the weak-factorization and finiteness hypotheses.","Editorial extension: Because the proof does not use pushouts, pullbacks, or homotopy relations, the same quotient description and classification are likely to survive in additive contexts with fewer limits and colimits, such as relative settings where only split exact sequences are guaranteed.","Editorial extension: The set-epic versus group-epic ambiguity is a live convention issue for the statement of Theorems 1.2 and 1.4, and the dual cofibrant classification should be scrutinized for the analogous ambiguity with $\\mathrm{Hom}_{\\mathcal{A}}(-,TF)$.","Editorial extension: When the additive quotient $(C\\cap F)/(C\\cap F\\cap W)$ is the stable category of a Frobenius category, Theorem 1.1 identifies the homotopy category with a triangulated category, giving a direct route to triangulated structure that bypasses explicit loop and suspension functors."],"forward_implications":["For any model structure on a weakly idempotent complete additive category, $\\mathrm{Ho}(\\mathcal{A})\\simeq (C\\cap F)/(C\\cap F\\cap W)$ as additive categories, so homotopy-theoretic information is readable from the object classes alone.","Fibrant model structures are classified by $\\mathrm{CoFib}$ together with $\\mathrm{TFib}$, so two fibrant model structures coincide exactly when their cofibrations and trivial fibrations coincide.","In a fibrant model structure, trivial cofibrations are precisely the splitting monomorphisms with cokernel in $TC$, and fibrations are precisely the $\\mathrm{Hom}_{\\mathcal{A}}(TC,-)$-epic morphisms; both classes are constructible from $TC$.","The homotopy category of a fibrant model structure is $C/TC$ as an additive category.","The $\\omega$-model structures on weakly idempotent complete exact categories and the $W$-model structures on weakly idempotent complete additive categories are recovered by the construction, with $W$-model structures exactly the bifibrant case."],"supporting_citations":[{"why":"It supplies the model-structure axioms, the homotopy category as a localization, and the lifting and factorization facts reused throughout the paper.","marker":"[Q1]"},{"why":"It gives the Hom-based epimorphism lemmas about fibrations and trivial cofibrations that the fibrant-model arguments depend on.","marker":"[BR, VIII, Lemma 1.1]"},{"why":"It proves that certain localizations of additive categories are additive quotient categories, which is the mechanism behind Theorem 1.1.","marker":"[K, Lemmas 2.2.1, 2.2.2]"},{"why":"It records the previous homotopy-category-as-quotient result for exact model structures that Theorem 1.1 extends.","marker":"[G2, Proposition 4.4]"},{"why":"It gave the quotient description and model structure from one hereditary complete cotorsion pair for the omega-model case, which the present classification rediscovers with a shorter proof.","marker":"[CLZ, Theorem 1.1]"},{"why":"It contains the prior W-model structure classification that the new bijection recovers as the bifibrant case.","marker":"[Bel, Theorem 4.5]"}],"fun_headline_variants":["Fibrant model structures are exactly one factorization system","Bijection between fibrant model structures and factorization systems","Homotopy category as additive quotient of cofibrant-fibrant","Fibrant structures classified by trivial cofibrations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that 'epic' means a surjective map of Hom-sets; if it instead means a surjective homomorphism of abelian groups, the fibrant-object claims in Theorem 1.2(3) fail, so the main classification depends on this convention.","fun_headline_variants_meta":{"raw":{"variants":["Fibrant model structures are exactly one factorization system","Bijection between fibrant model structures and factorization systems","Homotopy category as additive quotient of cofibrant-fibrant","Fibrant structures classified by trivial cofibrations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00067,"raw_usage":{"total_tokens":3078,"prompt_tokens":996,"completion_tokens":2082,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":612,"completion_tokens_details":{"reasoning_tokens":2013}},"tokens_in":612,"tokens_out":2082,"duration_ms":12891,"temperature":1.0,"reasoning_tokens":2013,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T13:52:45.188226+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In an artin algebra module category, take $W=\\mathrm{add}(M)$ for a nonzero module $M$ and use the W-model structure of Section 5.2; if 'epic' is read as a surjective homomorphism of abelian groups, then $X\\to 0$ is not a fibration for $X=M$, since $\\mathrm{Hom}_A(M,M)\\neq 0$, directly contradicting the Theorem 1.2(3) assertion that every object is fibrant.","supporting_citations":[],"review_version":1}