{"id":"eb34a913-a573-4281-9006-f859027b6bf7","arxiv_id":"1908.05446","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"An exact category satisfies the Jordan-Hölder property if and only if its Grothendieck monoid is free, with a rank-counting criterion and a permutation-combinatorics application to type A quivers.","lead":"This paper proves that a mathematical structure called an exact category has the Jordan-Hölder property exactly when its Grothendieck monoid is free, and applies this to classify when torsion-free classes of type A quivers have unique composition series. The result gives algebraists a counting test: compare simple objects against projectives, or compare supports against Bruhat inversions.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the central equivalence (Theorem 4.12) is internally coherent; the type A classification is the only external dependency and is not load-bearing for the main theorem.","rationale":"The reader's weakest_assumption points to the external Ingalls-Thomas classification used in the type A application. I agree that this is the least secure part of the paper's application section, and a concrete computational check of the support-count equality would be worthwhile. However, the paper's central claim, Theorem 4.12, does not depend on that classification: its proof is a direct consequence of the internal machinery in Sections 2-4 and Appendix A. I scrutinized the key steps: the bijection between simples and atoms (Proposition 3.6), the equivalence of length exactness with weak length-like functions (Theorem 4.4), and the monoid-theoretic criterion (Theorem A.20). No circularity or hidden assumption emerged. The freeness of M(E) gives both a length-like function and unique factorization, which yields (JHP) through the uniqueness of decompositions into atoms and the fact that atoms are exactly classes of simple objects. The converse direction constructs the monoid isomorphism from composition-series data using (JHP). These arguments are internally consistent. Thus no change to the verdict is needed.","tokens_in":50134,"tokens_out":16109,"duration_ms":162413,"concrete_test":"Enumerate all orientations of type A_n for n<=6; for each orientation, list all c-sortable elements w using the definition in Section 6.2, compute F(w)=add{M[i,j) : (i,j) in inv(w)}, compute indP(F(w)) from the definition of projectives in the exact category, and verify #supp(w)=#indP(F(w)) and the consequent equivalence #supp(w)=#Binv(w) iff (JHP) by direct inspection of composition series. A single counterexample for any n<=6 would invalidate Corollary 6.16; if all cases match, the external dependency is safe.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I find no load-bearing flaw in the central argument. Theorem 4.12 reduces (JHP) to freeness of M(E) via Proposition 3.6 (atoms = simples), Theorem 4.4 (length exactness = existence of a weak length-like function), and the monoid criterion Theorem A.20; each step is explicitly justified. The proof of (1)=> (2) in Theorem 4.9 constructs the inverse map using composition series and shows it respects conflations; the proof of (2)=> (1) uses freeness to make lengths and factor multisets unique. I checked that the linear-independence condition in Theorem 4.12(4) is sufficient because length exactness provides atomicity and reducedness of M(E). The only genuinely fragile point is in the type A application: Corollary 6.16 depends on the equality #supp(w)=#indP(F(w)) quoted as Theorem 6.13(3) from [IT, AIR, Tho]. A hidden exceptional case or off-by-one error there would affect Corollary 6.16, but not Theorem 4.12.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the Jordan-Hölder property (JHP) for exact categories. The main result (Theorem 4.12) establishes an equivalence among: (1) E is a length exact category satisfying (JHP); (2) the Grothendieck monoid M(E) is a free monoid; (3) K0(E) is free and the images of the non-isomorphic simples form a basis; and (4) the corresponding linear independence condition. The proof goes through Proposition 3.6 (simples are exactly the atoms of M(E)), Theorem 4.4 (length exactness is equivalent to the existence of a weakly length-like function), and the monoid-theoretic criterion Theorem A.20. A finite-rank counting version (Theorem 4.13) drives the applications: for extension-closed subcategories satisfying Assumption 5.6, JHP is equivalent to #sim = #indP (Theorem 5.10); Nakayama torsion-free and torsion classes satisfy JHP (Corollary 5.19). The paper then treats torsion-free classes of type A quivers: Theorem 6.15 identifies simples in F(w) with Bruhat inversions of the c-sortable element w, giving the combinatorial criterion #supp(w) = #Binv(w) for JHP (Corollary 6.16). Sections 7 and 8 contain explicit computations of Grothendieck monoids and counterexamples, including non-cancellative examples.","tokens_in":50296,"tokens_out":9507,"duration_ms":95839,"significance":"If the main theorem holds, and the proof appears sound, it reduces the Jordan-Hölder property to monoid-theoretic freeness and gives a very usable criterion in finite-rank cases: compare the number of simples with the rank of the Grothendieck group. This is a clean structural result with a genuinely useful application. The type A application is concrete and falsifiable, and the external dependence on the Ingalls-Thomas classification is clearly cited rather than hidden. Strengths include the detailed proof of the monoid criterion, the self-contained appendix on monoids, and the explicit counterexamples showing that freeness of K0 alone is not sufficient for JHP. The Bruhat-inversion result is novel and likely to be of independent interest.","major_comments":[],"minor_comments":[{"comment":"The proof bullet says that E has a length-like function if and only if M(E) has a length-like function, but this conflates length-like and weakly length-like functions: length exactness gives a weakly length-like function (Theorem 4.4), while freeness of M(E) supplies an additive length-like function on the monoid. The proof does not need the stated equivalence as written, but the wording should be corrected to avoid confusion.","section":"§4.3, proof of Theorem 4.12"},{"comment":"The construction of the inverse map K0(mod Λ) → K0(E) is summarized, but the verification that it respects short exact sequences is delegated to the reader. Since Proposition 5.8 and hence Theorem 5.10 depend on this isomorphism, please include the Horseshoe lemma and Schanuel lemma details or give a precise reference where the verification is written out.","section":"Lemma 5.7(2)"},{"comment":"The assertion that all four indecomposable objects in the example are simple objects in E is supported only by the phrase 'by checking subobjects.' Since this is the key point of the non-cancellativity example, please either display the subobject check or state it as a separate verification with enough detail for the reader to reproduce it.","section":"§8.3.2"},{"comment":"The equality #supp(F(w)) = #indI(F(w)) is quoted from support τ-tilting theory, and Corollary 6.16 hinges on this equality together with #supp(w) = #supp(F(w)). A precise statement of the quoted supporting result, for example [AIR, Theorem 2.7], would make the external dependency easier for the reader to check.","section":"Theorem 6.13(3)"},{"comment":"In the proof of Lemma 6.17(2), the assertion that the two orientation conditions follow from the closedness of M[l,l') in M[i,j) is stated without further explanation. A one-line justification of these two conditions would improve the readability of the proof of Theorem 6.15.","section":"Lemma 6.17(2)"}],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main result is real and the paper deserves a serious referee. Enomoto proves that a skeletally small exact category is a length exact category satisfying (JHP) exactly when its Grothendieck monoid is free, and gives a practical counting criterion using only the rank of K0 and the number of simples. That reduction is genuinely new and the proof is done from the definitions, using standard monoid theory in the appendix. The finite-rank counting version (Theorem 4.13) is the workhorse for the applications, and the applications are substantial: a large class of extension-closed categories, including functorially finite torsion-free classes, satisfy (JHP) precisely when #sim equals #indec. projectives. The Nakayama corollary and the type A Bruhat-inversion bijection are nice payoffs, and the combinatorial criterion in Corollary 6.16 is a genuinely new result, not just a repackaging of known classifications.\n\nWhat the paper does well beyond the main theorem: the relation-to-other-works section is unusually honest and precise about overlap with [BeGr], [BHLR], [HR], and [BHT]. The counterexamples in Section 8 are instructive, especially the idempotent-complete ones. The proofs are detailed and internally consistent; I checked the central steps of Theorem 4.9 and Theorem 4.12 and found no circularity.\n\nSoft spots are minor. Lemma 5.7(2) leaves the inverse-map details to the reader; that is a standard Schanuel/Horseshoe argument, but it is a real skipped verification. Section 8.3.2 asserts that all four indecomposables are simple after 'checking subobjects' without showing the check; again minor, but in a counterexample section you want the bad behavior explicit. The type A application depends on the equality #supp(w) = #indP(F(w)) quoted from [Tho, IT, AIR]; a hidden off-by-one there would affect Corollary 6.16 but not Theorem 4.12, which stands on its own. The footnote in Section 8.3.2 is odd but honest — it flags that extension-closedness is not trivial, and the cited argument is reasonable.\n\nWho this is for: representation theorists working with torsion-free classes, and anyone interested in exact categories as a general framework. The monoid-freeness criterion will be cited. The paper deserves peer review and, after minor revisions addressing the delegated details, acceptance. I would bring it to the reading group and cite it in my own work.","headline":"A solid, carefully proved characterization of the Jordan–Hölder property via freeness of the Grothendieck monoid, with a genuine new combinatorial application to type A torsion-free classes; minor delegated verifications do not undercut the main results.","tokens_in":50894,"tokens_out":1161,"would_cite":true,"duration_ms":14279,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18E10","16G10","16G20"],"pacs":[],"model":"deepseek-v4-flash","headline":"For any exact category, the Jordan-Hölder property holds exactly when the Grothendieck monoid is free; in type A quivers this becomes a count of supports versus Bruhat inversions.","keywords":["exact category","Jordan-Hölder property","Grothendieck monoid","torsion-free classes","c-sortable elements","Bruhat inversions","Nakayama algebras","quiver of type A"],"falsifier":"Try to construct an extension-closed subcategory of modules over an artin algebra, equivalent to an Ext-perpendicular category of a module of finite injective dimension, whose number of simple objects equals its number of indecomposable projectives but which has two non-isomorphic composition series of one object; the paper's Theorem 5.10 says this cannot exist. A lighter check is to take any c-sortable element of a type A quiver with equal support and Bruhat-inversion counts and compute the Grothendieck monoid of the corresponding torsion-free class—it must be free.","tokens_in":49878,"feed_emoji":"","tokens_out":11270,"duration_ms":98787,"temperature":0.7,"pith_summary":"The paper's central claim is that the Jordan-Hölder property for exact categories—the assertion that any two composition series of an object have the same simple factors with the same multiplicities—is equivalent to a purely monoid-theoretic statement: the Grothendieck monoid of the category is free. Since non-isomorphic simple objects survive as distinct atoms in this monoid, freeness exactly records that simple factors are uniquely determined. When the Grothendieck group is finitely generated, the criterion becomes a count: Jordan-Hölder holds precisely when the Grothendieck group is free and its rank equals the number of simple objects. For functorially finite torsion-free and torsion classes of artin algebras, this reduces the property to comparing the number of simples with the number of indecomposable projectives, and for Nakayama algebras every such class satisfies it. In type A quivers the simples of a torsion-free class are shown to be in bijection with Bruhat inversions of the corresponding c-sortable permutation, giving a purely combinatorial test for Jordan-Hölder.","feed_headline":"Jordan-Hölder property equals a free Grothendieck monoid","feed_subtitle":"In type A quivers the same check becomes a comparison of supports and Bruhat inversions.","key_machinery":"The load-bearing object is the Grothendieck monoid $M(\\mathcal{E})$, built from isomorphism classes with direct sum as addition and relations forced by conflations. Its key feature is that non-isomorphic simple objects remain distinct atoms, so a free monoid structure on $M(\\mathcal{E})$ is exactly the statement that composition factors are unique. A second machine is the combinatorial dictionary for type A quivers: torsion-free classes are indexed by c-sortable permutations, indecomposables are interval modules $M[i,j)$ indexed by inversions, and the paper shows that the simple objects among them are precisely the Bruhat inversions, that is, inversions that give covering relations in the Bruhat order. The support size of $w$ then counts the indecomposable projectives, turning Jordan-Hölder into the equality of two easily computed permutation statistics.","core_discovery":"On the paper's own terms, the discovery is a characterization theorem. For any skeletally small exact category $\\mathcal{E}$, the following are equivalent: $\\mathcal{E}$ is a length exact category satisfying the Jordan-Hölder property; the Grothendieck monoid $M(\\mathcal{E})$ is a free monoid with basis the classes of simple objects; and $\\mathcal{E}$ is length exact with $K_0(\\mathcal{E})$ free abelian on the classes of simple objects. When $K_0(\\mathcal{E})$ is finitely generated, Jordan-Hölder holds exactly when $K_0(\\mathcal{E})$ is free of rank equal to the number of simple objects. Applied to extension-closed subcategories of module categories over artin algebras, notably functorially finite torsion-free classes, this says that Jordan-Hölder is decided by comparing the number of indecomposable projectives with the number of simples. For a type A quiver with Coxeter element $c$, the torsion-free class $F(w)$ attached to a c-sortable element $w$ has simple objects in bijection with the Bruhat inversions of $w$, so $F(w)$ satisfies Jordan-Hölder exactly when $\\#\\operatorname{supp}(w)=\\#\\operatorname{Binv}(w)$.","pith_inferences":["This suggests a general heuristic: when the Grothendieck group is free of finite rank, Jordan-Hölder failures are driven by relations among simple classes that disappear after group completion, so one can search for failures by looking for non-cancellative Grothendieck monoids.","The type-A dictionary points to a testable algorithm for other Dynkin types: translate simples in a torsion-free class into the appropriate Weyl-group inversion statistic and compare it with the support count; the paper's criterion would then become a uniform Weyl-group identity.","The half-factoriality result for the unique length property suggests a finer classification of exact categories by factorization properties of their Grothendieck monoids—free, half-factorial, or merely atomic—each corresponding to a different degree of uniqueness of composition series."],"forward_implications":["In any skeletally small exact category, Jordan-Hölder, freeness of the Grothendieck monoid, and the basis condition on $K_0$ are equivalent, so JHP can be checked by group or monoid calculations instead of constructing composition series.","For extension-closed subcategories of module categories over artin algebras that are Ext-perpendicular to a finite-injective-dimension module, JHP holds exactly when the number of simples equals the number of indecomposable projectives; this covers functorially finite torsion-free and torsion classes.","Every torsion-free class and every torsion class over a Nakayama algebra satisfies Jordan-Hölder.","For a type A quiver, the torsion-free class $F(w)$ satisfies Jordan-Hölder exactly when $\\#\\operatorname{supp}(w)=\\#\\operatorname{Binv}(w)$, a condition readable directly from the one-line notation of the c-sortable element $w$.","In any category satisfying JHP, the Grothendieck monoid is free with basis the simple classes, so the Grothendieck group is free with the same basis; conversely, non-cancellative or non-freely generated Grothendieck monoids obstruct JHP."],"supporting_citations":[{"why":"Defines the Grothendieck monoid of an exact category and establishes that simple objects correspond to its atoms.","marker":"[BeGr]"},{"why":"Introduces simple objects, composition series, and the Jordan-Hölder property for exact categories, the notions the paper characterizes.","marker":"[BHLR]"},{"why":"Supplies the axioms and standard facts about exact categories used throughout.","marker":"[Büh]"},{"why":"Provides the Jordan-Hölder theorem for modular lattices, used to prove the unique length property in modular subobject posets.","marker":"[St]"},{"why":"Classifies torsion-free classes of type A quivers by c-sortable elements, the dictionary at the heart of the Bruhat-inversion application.","marker":"[Tho]"},{"why":"Gives the support-counting equality relating supports of a c-sortable element to indecomposable projectives in the corresponding torsion-free class.","marker":"[IT]"},{"why":"Provides the tau-tilting correspondence identifying functorially finite torsion-free classes with support tau-tilting modules, used in the projectives-versus-simples count.","marker":"[AIR]"},{"why":"Supplies the description of inversions and Bruhat order covering relations used to define and characterize Bruhat inversions.","marker":"[BB]"}],"fun_headline_variants":["Free Grothendieck monoids characterize Jordan-Hölder","Type A: JHP from Bruhat inversions","JHP iff free Grothendieck monoid","Torsion-free classes: JHP when projectives equal simples","Bruhat inversion count decides JHP in type A"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The combinatorial criterion for type A rests on the external classification that torsion-free classes of a type A quiver are exactly the subcategories $F(w)$ attached to c-sortable permutations, together with the equality that counts indecomposable projectives by the support size of $w$; any hidden exception or off-by-one error in that classification would break the Bruhat-inversion test, though the general monoid-freeness theorem would survive.","fun_headline_variants_meta":{"raw":{"variants":["Free Grothendieck monoids characterize Jordan-Hölder","Type A: JHP from Bruhat inversions","JHP iff free Grothendieck monoid","Torsion-free classes: JHP when projectives equal simples","Bruhat inversion count decides JHP in type A"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001704,"raw_usage":{"total_tokens":6766,"prompt_tokens":983,"completion_tokens":5783,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":599,"completion_tokens_details":{"reasoning_tokens":5701}},"tokens_in":599,"tokens_out":5783,"duration_ms":40933,"temperature":1.0,"reasoning_tokens":5701,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:13:20.048946+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Try to construct an extension-closed subcategory of modules over an artin algebra, equivalent to an Ext-perpendicular category of a module of finite injective dimension, whose number of simple objects equals its number of indecomposable projectives but which has two non-isomorphic composition series of one object; the paper's Theorem 5.10 says this cannot exist. A lighter check is to take any c-sortable element of a type A quiver with equal support and Bruhat-inversion counts and compute the Grothendieck monoid of the corresponding torsion-free class—it must be free.","supporting_citations":[],"review_version":1}