{"id":"5589be34-6900-426f-9c3d-130d317ee635","arxiv_id":"2501.00291","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The combinatorial shadow of any sl3-generated transitive module category is one of eight infinite graphs.","lead":"This paper maps the possible shapes that can appear when the building blocks of a symmetry called sl3 are used to generate new algebraic structures, and finds that only eight infinite patterns occur. The result is a milestone for a research program that connects Lie theory with graph combinatorics and categorification.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 20's reduction of arbitrary L to the model cases in §7.3 rests on the unpublished [Ma23] positive solution of Kostant's problem for sl3; if that result is absent or does not cover the non-integral cases actually included in the dichotomy, the classification is incomplete.","rationale":"The reader's weakest assumption identifies the same load-bearing point: Theorem 20's reduction to the model cases depends on a positive solution of Kostant's problem for sl3 that is cited only to the first author's preprint [Ma23]. My stress-test confirms this is the critical vulnerability of the paper. I add a small sharpening: the dichotomy in §7.3 is not just 'integral versus exceptional'; it also includes non-integral cosets such as (2/3,2/3)+Λ, which Proposition 18(e) shows behaves exceptionally, and the quoted support about integral central characters does not literally cover those cases. This makes the gap more precise but does not change the verdict. The case analyses and eigenvector computations in Sections 4–6 appear substantial and are plausibly correct; the issue is the missing external validation of the reduction step. Since the problem is a specific, addressable gap rather than a demonstrated false statement, the original CONDITIONAL verdict remains appropriate. No change to the reader's verdict is needed.","tokens_in":28615,"tokens_out":12899,"duration_ms":133675,"concrete_test":"Check the exact statements in [Ma23] and [MMM23]: does either prove that for every simple sl3-module L with Ann(L)=Ann(L(λ)) and λ∉(1/3,1/3)+Λ, one has L(L,L)≅U(g)/Ann(L) as algebras? In particular, verify the case λ∈(2/3,2/3)+Λ explicitly. A direct computational fallback: for the three simple Verma modules in the (2/3,2/3)+Λ block, compute dim Hom_g(V,L(L,L)_ad) and dim Hom_g(V,L(L(λ),L(λ))_ad) for all finite-dimensional V; equality for those V would confirm the missing step for that coset, while a nonzero difference would disprove the reduction as stated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is in §7.3: to classify transitive subquotients of add(C·L) for arbitrary simple L, the paper must prove that the Harish-Chandra bimodule L(L,L) is isomorphic either to U(g)/Ann(L) (when λ∉(1/3,1/3)+Λ) or to the corresponding Whittaker bimodule (in the exceptional subcase). The first isomorphism is asserted to follow from the positive solution of Kostant's problem for all simple sl2- and sl3-modules with integral central characters, cited only to the first author's unpublished preprint [Ma23]. This is the exact bridge that reduces arbitrary L to the model categories analyzed in Sections 4–6; if it fails, Theorem 20 is not established even if every case analysis in those sections is correct. Two textual gaps make the support insecure. First, [Ma23] is a preprint by the first author, not a peer-reviewed reference. Second, the dichotomy in §7.3 actually includes non-integral λ: the case λ∉(1/3,1/3)+Λ contains λ∈(2/3,2/3)+Λ, which by Proposition 18(e) is also exceptional, and the cited statement about 'integral central characters' does not by its own wording cover such λ. The sentence claiming that the corresponding part of Oχ is indecomposable is also in tension with Proposition 17 for generic λ. Thus the on-paper justification for the reduction is incomplete unless [Ma23] or [MMM23] contains a broader theorem than the text states.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the module categories add(C·L) attached to a simple sl3-module L, where C is the monoidal category of finite-dimensional sl3-modules. The authors compute, for the action of the natural module F, the oriented graph Γ_F of a transitive subquotient N of add(C·L). Theorem 20 asserts that Γ_F is always one of eight infinite graphs, collected in Figure 16. For each graph the paper provides an explicit positive eigenvector for the Perron-Frobenius eigenvalue 3, with entries coming from dimensions, Gelfand-Kirillov dimensions, Bernstein numbers, or Verma flag lengths depending on the case. The proof is organized through model cases: the regular module category, BGG category O (upper middle, lower middle, bottom, partially integral, and generic weights), Whittaker modules, and a reduction of arbitrary simple L via Harish-Chandra bimodules. Several auxiliary results, such as Theorem 1, establish rigidity properties of the regular module category.","tokens_in":28878,"tokens_out":12968,"duration_ms":119236,"significance":"If correct, the classification is a substantial extension of the sl2 results in [MZ24] and gives a finite list of infinite combinatorial shadows for finite-dimensional sl3 actions. A genuine strength is that the Perron-Frobenius eigenvectors are derived from intrinsic invariants rather than fitted to the graphs; the paper also connects the combinatorics to dimensions, GK dimension, Bernstein numbers, and Verma flag lengths. The rigidity statement for the regular module category (Theorem 1) is an additional contribution. However, the reduction in §7.3 that underpins Theorem 20 currently rests on an unpublished preprint and appears to skip the non-integral exceptional case λ∈(2/3,2/3)+Λ, so the significance is conditional on completing that argument.","major_comments":[{"comment":"The dichotomy in §7.3 is incomplete. By Proposition 18(e), both (1/3,1/3)+Λ and (2/3,2/3)+Λ have the exceptional property (W·λ)∩(λ+Λ) = {λ, sr·λ, rs·λ}, and Proposition 19(a) treats both as exceptional in the Whittaker setting. But §7.3 only separates (1/3,1/3)+Λ, and then claims that for all remaining λ the relevant part of Oχ is indecomposable. For λ∈(2/3,2/3)+Λ this claim is false: such a weight is generic in the sense of §5.7, and Proposition 17 shows that the corresponding block of category O is semi-simple. Thus the reduction of an arbitrary simple module L to the model categories is not proved for λ∈(2/3,2/3)+Λ; a separate argument (presumably parallel to the two subcases used for (1/3,1/3)+Λ) or an explicit invocation of Proposition 19(a) is needed.","section":"§7.3, case λ ∉ (1/3,1/3)+Λ"},{"comment":"The isomorphism L(L(λ),L(λ)) ≅ L(L,L) ≅ U(g)/Ann(L) for λ∉(1/3,1/3)+Λ is asserted to follow from the positive solution of Kostant's problem for simple sl2- and sl3-modules with integral central characters, with [Ma23] cited for details. This step is load-bearing for Theorem 20, but [Ma23] is an unpublished preprint by the first author, and the stated hypothesis 'integral central characters' does not by its own wording cover the non-integral weights in (2/3,2/3)+Λ that occur in the same case. Please either give a complete proof in the paper or replace the reference by a peer-reviewed source covering all central characters actually used in the argument.","section":"§7.3, first displayed isomorphism"},{"comment":"These propositions are needed for the completeness of the eight-graph classification in Theorem 20, yet each is dispatched with 'Mutatis mutandis the proof of ...'. In at least one instance the paper itself emphasizes that the two situations are not symmetric: after Proposition 5 it notes that N1 and N2 are not equivalent and have different combinatorics. For a journal-level proof, the reader needs at least a precise statement of which case analyses are replaced and why the resulting local pictures are the ones claimed; otherwise the completeness of the list is not independently verifiable from the text.","section":"Propositions 5, 6, 9, 13, 14, 15, 16"},{"comment":"Theorem 20 is stated for an arbitrary transitive subquotient N of add(C·L), but the proof establishes an equivalence for the whole category add(C·L) via L(L,L)-modules and then stops. The propositions in Sections 4–6 compute graphs for specific simple transitive module categories (N1, N2, N3, M_i^a, K, etc.) and for some quotients; they do not compute or bound the graphs of arbitrary transitive subquotients of non-simple model categories such as add(C·L(λ)) for a regular upper middle weight λ (cf. Proposition 4). If 'transitive subquotient' is intended to mean 'simple transitive subquotient' (as suggested by §1.2), the theorem should say so; otherwise a further argument showing that every transitive subquotient has the graph of one of the listed simple transitive categories is missing.","section":"§7.2–§7.3, proof of Theorem 20"}],"minor_comments":[{"comment":"Proposition 16(a) says 'For λ ∈ X4 such that λ1+λ2 = −1', but the statement concerns the category M_6^a and should read λ∈X6; X4 is defined earlier for the (a,0)+Λ case in Proposition 14.","section":"§5.6, Proposition 16(a)"},{"comment":"The term 'generic' is used in two incompatible ways: §5.7 calls all weights that are neither integral nor partially integral generic, which includes (1/3,1/3)+Λ and (2/3,2/3)+Λ, while §6.3 and Proposition 18(e) treat these two cosets as exceptional. Introducing separate terminology would avoid confusion.","section":"§5.7 and §6.3"},{"comment":"The sentence claiming that the combinatorics of tensoring with L((1,0)) between regular simples within one connected shaded component is the same as the combinatorics in C_C is not fully justified in the text beyond the assertion that a Weyl group element transforms the component; a sentence explaining why multiplicities are preserved would help the reader.","section":"§5.3, proof of Proposition 4(b)"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is potentially important, but the current proof has a concrete gap for the exceptional coset (2/3,2/3)+Λ and relies on an unpublished preprint for a load-bearing reduction. These points should be resolved before publication; the missing case appears fixable by adapting the (1/3,1/3)+Λ argument, but it must actually be written out or covered by a published reference."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper does a lot of honest work: it determines, for the monoidal category of finite-dimensional sl3-modules, the graphs of transitive subquotients of add(C·L) for L in the main families—integral, partially integral, generic, and Whittaker. The eight graphs in Figure 16 are new, and the explicit positive eigenvectors with eigenvalue 3, interpreted as dimensions, GK-dimensions, Bernstein numbers, or Verma lengths, are a nice concrete payoff. Propositions 3 and 7 are worked out in real detail, and Theorem 1 (the regular category is determined by its graph and dual graph) is clean. For the model cases, the combinatorics looks credible.\n\nThe soft spot is the bridge in Theorem 20. The reduction to the model cases in §7.3 splits on λ∉(1/3,1/3)+Λ vs λ∈(1/3,1/3)+Λ. The first case includes λ∈(2/3,2/3)+Λ, which by Proposition 18(e) has the same 3-element stabilizer behavior that the paper treats as exceptional in the Whittaker setup (Proposition 19(a)). The proof cites [Ma23] for the positive solution of Kostant's problem 'for all simple sl2- and sl3-modules with integral central characters.' Non-integral λ is not covered by that wording, and the claim that the corresponding part of Oχ is indecomposable sits awkwardly next to Proposition 17, where generic blocks are semi-simple. So the proof as written does not establish that L(L,L) reduces to the model cases for that family. This is not a fatal flaw in the whole project, but it is a real gap in the universality claim.\n\nThe other soft spot is the number of 'Mutatis mutandis' proofs (Propositions 5, 6, 9, 13–16). Some are probably fine, but together with the gap above they make the paper less self-contained than its main theorem needs to be.\n\nBottom line: the model-case classifications and the eigenvector data are worth having, and the paper deserves a careful referee. The referee should ask for a complete treatment of λ∈(2/3,2/3)+Λ and for either a proof or a peer-reviewed reference for the Kostant reduction. I would not desk-reject this; I'd send it out with the expectation of major revision.","headline":"Solid sl3 model-case combinatorics with a real gap in the universal reduction: Theorem 20 needs the (2/3,2/3)+Λ case and a peer-reviewed Kostant reference.","tokens_in":29477,"tokens_out":5438,"would_cite":false,"duration_ms":50615,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["17B10","18M05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every simple sl3-module yields one of eight infinite graphs for the category generated by tensoring with finite-dimensional modules.","keywords":["sl3-modules","transitive module categories","monoidal categories","category O","Whittaker modules","infinite Dynkin diagrams","Perron-Frobenius eigenvector","Harish-Chandra bimodules"],"falsifier":"Take any simple $\\mathfrak{sl}_3$-module with an integral central character, list the indecomposable projective functors for that character, and compute the multiplicities of each simple summand in the adjoint action on its locally finite endomorphism ring; any deviation from the multiplicities of $U(\\mathfrak{g})/\\operatorname{Ann}(L)$ or, in the exceptional coset, from the Whittaker bimodule would disprove the reduction in Section 7.3 and could produce a graph outside the eight-graph list.","tokens_in":28347,"feed_emoji":"♾️","tokens_out":12067,"duration_ms":113742,"temperature":0.7,"pith_summary":"The paper sets out to classify the combinatorial shape of every category of $\\mathfrak{sl}_3$-modules obtained by tensoring finite-dimensional modules with an arbitrary simple module. It claims that for any simple $\\mathfrak{sl}_3$-module $L$, every transitive subquotient of the category $\\operatorname{add}(\\mathcal C\\cdot L)$ — the additive closure of all summands of $V\\otimes L$ with $V$ finite dimensional — has a graph $\\Gamma_F$ taken from a fixed list of eight infinite graphs. These eight graphs are the $\\mathfrak{sl}_3$-analogues of the classical infinite Dynkin diagrams that describe the $\\mathfrak{sl}_2$ case. If the claim is right, the representation theory of $\\mathfrak{sl}_3$ has a finite combinatorial menu at this level of granularity, with the Perron-Frobenius eigenvector for eigenvalue $3$ carrying dimensions, Gelfand-Kirillov dimensions, Bernstein numbers, or Verma flag lengths depending on the case.","feed_headline":"Every simple sl3-module yields one of eight infinite graphs","feed_subtitle":"Tensoring any simple sl3-module with finite-dimensional modules always lands in one of eight combinatorial patterns.","key_machinery":"The central object is the locally finitary category $\\operatorname{add}(\\mathcal C\\cdot L)$, whose indecomposable objects carry a nonnegative integer action matrix $[F]$ for the natural module $F=L((1,0))$; the graph $\\Gamma_F$ is the directed multigraph encoding this matrix. The proof machinery is a two-part transfer. First, the algebra $\\mathcal L(L,L)$ of endomorphisms of $L$ on which the adjoint action is locally finite, a Harish-Chandra bimodule, is shown to be the algebra-object proxy for $\\operatorname{add}(\\mathcal C\\cdot L)$. Second, $\\mathcal L(L,L)$ is identified with $\\mathcal L(L(\\lambda),L(\\lambda))$ for a suitable highest weight module $L(\\lambda)$, or with the corresponding Whittaker bimodule in the exceptional coset, so the classification reduces to the model cases computed in Sections 4--6. Perron-Frobenius theory enters through the fact that the positive eigenvector for the eigenvalue $3$ is unique and its coefficients are representation-theoretic quantities such as dimension, growth data, or Verma flag length.","core_discovery":"The main theorem, Theorem 20, states: let $L$ be any simple $\\mathfrak{sl}_3$-module and $N$ a transitive subquotient of $\\operatorname{add}(\\mathcal C\\cdot L)$. Then the graph $\\Gamma_F$ recording the action of the natural three-dimensional module $F = L((1,0))$ is isomorphic to the graph in one of Figures 3, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or 15, up to the isomorphisms that identify some of them; the distinct list contains exactly eight graphs, collected in Figure 16. Because the monoidal category of finite-dimensional $\\mathfrak{sl}_3$-modules is generated by $F$, this graph determines the full combinatorial shadow of the action. The proof obtains seven graphs from simple highest weight modules in category O, the category of finitely generated weight modules with locally finite nilpotent action, and one further graph from non-degenerate Whittaker modules. Arbitrary simple modules are reduced to these models through Harish-Chandra bimodules: the locally finite part of the endomorphism algebra of $L$ is compared with the corresponding bimodule for a highest weight or Whittaker module, using equality of annihilators and a spanning result for locally finite endomorphism rings.","pith_inferences":["If the theorem is right, the eight graphs form a complete obstruction: any simple transitive module category over the finite-dimensional $\\mathfrak{sl}_3$ monoidal category not on the list cannot arise from tensoring with a simple module; whether such categories exist as abstract constructions is a separate question left open by the paper.","The eigenvalue $3$ is the dimension of the natural module, just as the eigenvalue was $2$ for $\\mathfrak{sl}_2$; by analogy, an $\\mathfrak{sl}_n$ version should have eigenvalue $n$ and probably a finite list, though the exceptional cosets and Whittaker branch suggest the list will grow with $n$ in a non-obvious way.","One can try to read the eight graphs purely combinatorially: classify strongly connected locally finite multigraphs with a positive eigenvector of eigenvalue $3$ and the local branching rules visible in the figures. A matching classification would make the Lie-algebraic reduction unnecessary.","Proposition 21 settles semi-simplicity for the four graphs without double edges; applying the same argument to the four remaining graphs would decide whether the simple transitive category itself, not only its graph, is uniquely determined."],"forward_implications":["Every transitive subquotient of $\\operatorname{add}(\\mathcal C\\cdot L)$, for any simple $\\mathfrak{sl}_3$-module $L$, has one of eight graphs; the list is finite even though each graph is infinite.","The Perron-Frobenius eigenvector with eigenvalue $3$ encodes dimensions in the regular case, growth-related quantities in the middle cases, and Verma flag lengths in the bottom case, so the combinatorics determines numerical invariants of the modules.","The eight graphs are the $\\mathfrak{sl}_3$-analogues of the infinite Dynkin diagrams that classify the $\\mathfrak{sl}_2$ case, placing the result in the same framework as the earlier $\\mathfrak{sl}_2$ classification.","For the four graphs without multiple oriented edges, the underlying simple transitive module category is semi-simple, so the classification reaches beyond combinatorics to a structural uniqueness statement.","Categories generated by regular upper or lower middle weights decompose as an extension of the corresponding singular-piece graph by the regular module-category graph, giving a compositional description of the full category."],"supporting_citations":[{"why":"Supplies the classification of indecomposable projective functors and the Harish-Chandra bimodule formalism used to transfer computations between module categories.","marker":"[BG80]"},{"why":"Gives the composition-series description of Whittaker-induced modules and the equivalence result for non-exceptional central characters that Theorem 20 extends.","marker":"[MiSo97]"},{"why":"Establishes the sl2 version of the classification via infinite Dynkin diagrams and the method of identifying simple module categories by their graphs.","marker":"[MZ24]"},{"why":"Provides the finiteness and Gelfand-Kirillov dimension facts for add(C·L) that make the action matrices and Perron-Frobenius eigenvector well defined.","marker":"[MMM23]"},{"why":"Supplies the unpublished proof that the locally finite endomorphism ring of every simple sl3-module with integral central character is the enveloping-algebra quotient.","marker":"[Ma23]"},{"why":"Shows every simple module shares its annihilator with some simple highest weight module, the first reduction step in the proof of Theorem 20.","marker":"[Du77]"},{"why":"Gives the standard facts on Harish-Chandra bimodules, enveloping-algebra quotients, and growth invariants used throughout Sections 5 through 7.","marker":"[Ja83]"},{"why":"Defines non-degenerate Whittaker modules and their basic structure, the second source of model graphs.","marker":"[Ko78]"}],"fun_headline_variants":["All simple sl3-modules land in one of eight infinite graphs","Eight infinite graphs for every simple sl3-module","sl3 simple modules: only eight infinite graph patterns exist","The eight graph shapes of all simple sl3-module actions","Simple sl3 modules: exactly eight infinite combinatorial patterns"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is an unpublished claim that every simple $\\mathfrak{sl}_3$-module of integral type has its locally finite endomorphism ring equal to the enveloping-algebra quotient, and if that premise fails the reduction to the model cases breaks.","fun_headline_variants_meta":{"raw":{"variants":["All simple sl3-modules land in one of eight infinite graphs","Eight infinite graphs for every simple sl3-module","sl3 simple modules: only eight infinite graph patterns exist","The eight graph shapes of all simple sl3-module actions","Simple sl3 modules: exactly eight infinite combinatorial patterns"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000182,"raw_usage":{"total_tokens":1279,"prompt_tokens":884,"completion_tokens":395,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":500,"completion_tokens_details":{"reasoning_tokens":313}},"tokens_in":500,"tokens_out":395,"duration_ms":4587,"temperature":1.0,"reasoning_tokens":313,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:54:31.045557+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take any simple $\\mathfrak{sl}_3$-module with an integral central character, list the indecomposable projective functors for that character, and compute the multiplicities of each simple summand in the adjoint action on its locally finite endomorphism ring; any deviation from the multiplicities of $U(\\mathfrak{g})/\\operatorname{Ann}(L)$ or, in the exceptional coset, from the Whittaker bimodule would disprove the reduction in Section 7.3 and could produce a graph outside the eight-graph list.","supporting_citations":[],"review_version":1}