{"id":"a75d7b59-c231-48ce-9ffa-6d2b11590cbc","arxiv_id":"2510.18031","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Zamolodchikov periodic B-matrices are in bijection with pairs of commuting finite-type Cartan matrices, with 29 new infinite families and 14 new exceptional types beyond the ADE case.","lead":"This paper classifies the exchange matrices in cluster algebras that stay periodic under a fixed alternating mutation pattern, extending the known classification to all finite-type Lie diagrams. It gives a complete list of new families and exceptions, and shows every case arises from the simply-laced cases by folding and transposing.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 3.22's inverse unfolding is not verified for every family; if it fails, admissible ⇒ periodic collapses.","rationale":"The reader's weakest assumption identifies exactly the reliance on Proposition 3.22 and the black-box [GP19] ADE tropical periodicity. My stress test concurs: the single most load-bearing step is Proposition 3.22, because it is the only mechanism that transfers tropical periodicity from the known ADE case to all admissible Dynkin biagrams. The proof is a brief constructive sketch without a systematic verification for the numerous families, especially those where the base nonparallel binding is obtained by multiple folds and the path contains parallel copies. If this step fails, the forward direction (1)=>(4) in Theorem 4.1 fails, and the claimed bijection is incomplete. I do not find a more basic internal inconsistency: the Vinberg-based equivalence between strict subadditivity and Dynkin components (Proposition 4.16) is sound, and the tropical-to-strict subadditivity argument (Proposition 4.18) is a natural matrix generalization of GP19. The title/abstract overclaim regarding 'cluster algebras' vs. 'B-matrices' is real but already acknowledged in Remark 1.2 and does not affect correctness. The appropriate verdict remains CONDITIONAL: the central claim is plausible and likely correct, but the proof needs a concrete verification or a detailed case-by-case proof of Proposition 3.22. Since this is exactly the reader's condition, no verdict change is needed.","tokens_in":38796,"tokens_out":15967,"duration_ms":122416,"concrete_test":"Write a SageMath script that, for each family in Theorem 3.6 (taking small parameters, e.g., n,m ≤ 5), explicitly constructs the unfolded ADE bigraph B_0 by reversing the fold/transpose sequences in Proposition 3.22 and Lemma 3.11, then checks: (i) every component of Γ_0 and Δ_0 is an ADE Dynkin diagram; (ii) Γ_0Δ_0 = Δ_0Γ_0; (iii) Γ_0 and Δ_0 share no nonzero entries; (iv) the specified bicolored automorphisms satisfy Definition 3.2 and fold B_0 back to the target biagram. If all checks pass, the gap is closed; if any fails, Proposition 3.22 is false and the classification must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The 'if' direction of Theorem 1.1 (admissible Dynkin biagram ⇒ Zamolodchikov periodic) is established via Corollary 4.13, which depends entirely on Proposition 3.22: every admissible Dynkin biagram is obtained from an admissible ADE bigraph by global flips and bicolored folds. The proof of Proposition 3.22 is an outline: for a path Γ_1^n * Γ_2^m, it says to reverse the fold/transpose sequence componentwise, setting B_{k-1} = (Λ'_1)^n * (Λ'_2)^m. It does not verify that the constructed B_0 is an admissible ADE bigraph for each of the 29 infinite families and 14 exceptional types, nor that the componentwise inverse lifts bicolored automorphisms of the base binding to bicolored automorphisms of the entire path. For instance, the fold E_6*E_6 → F_4*_1 E_6 in the (F_4 E_6^{m-1}) families identifies vertices across a binding; lifting this to a path with parallel E_6 copies requires the same orbit structure to be compatible with the parallel edges, which is not checked. If any lifted bigraph is non-admissible or has a non-ADE component, that family is not covered and the bijection is incomplete. This is the only bridge from the [GP19] ADE tropical-periodicity theorem to all non-ADE types, so the central claim hinges on it. The paper's own limitation statements (e.g., 'this proof structure was adapted from [Ste10]' and 'the proof is identical to [GP19]') reinforce that these case checks are not fully supplied.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a complete classification of Zamolodchikov periodic cluster algebras, formulated as bipartite recurrent B-matrices whose associated T-systems are periodic. The main theorem (Theorem 1.1) asserts that such B-matrices are in natural bijection with pairs of commuting finite-type Cartan matrices (Γ,Δ), and that the period divides h_Γ+h_Δ. The proof is structured as a five-way equivalence (Theorem 4.1): admissible Dynkin biagram ⇔ strictly subadditive labeling ⇔ fixed point labeling ⇔ tropical T-system periodic for all initial labels ⇔ T-system periodic. The non-ADE cases are obtained from ADE bigraphs by folding and transpose operations, extending Stembridge's classification of commuting simply-laced Cartan matrices and Galashin–Pylyavskyy's classification of Zamolodchikov periodic quivers. The paper also gives connections to W-graphs and Kazhdan–Lusztig theory, and states two conjectures supported by SageMath computations.","tokens_in":39184,"tokens_out":9972,"duration_ms":81990,"significance":"If the main theorem is correct, it resolves a natural and substantial open problem: it gives the first complete classification of Zamolodchikov periodic cluster algebras beyond the simply-laced case, and it provides a full list of pairs of commuting finite-type Cartan matrices, which is of independent interest in Lie theory and Coxeter combinatorics. The strategy of deriving all non-ADE examples from ADE bigraphs via folding and transpose is elegant and conceptually important, especially the observation that transpose preserves Zamolodchikov periodicity despite not commuting with cluster mutation. The connection to admissible W-graphs and the explicit classification into 29 infinite families and 14 exceptional types (beyond Stembridge's 6+11) gives the paper substantial scope. The proofs rely on external deep results (Keller, Galashin–Pylyavskyy, Stembridge, Vinberg, Perron–Frobenius) rather than fitting parameters, and Conjectures 6.1 and 6.3 are concrete and falsifiable. The paper is likely to become a reference for this area if the gaps identified below are resolved.","major_comments":[{"comment":"This proposition is the only bridge from the ADE tropical-periodicity theorem of [GP19] to all non-ADE admissible Dynkin biagrams, and its proof is not complete. The inverse construction in step (2)(ii) simply sets B_{k-1}=(Λ'_1)^n * (Λ'_2)^m componentwise, but does not verify that the resulting biagram is an admissible ADE bigraph, that the intermediate objects remain valid Dynkin biagrams (bipartite, no shared edges, all components ADE), or that the fold/transpose sequence reverses correctly when the original fold identifies vertices across a binding. The example of E_6*E_6 → F_4*_1 E_6 (used for the (F_4 E_6^{m-1}) families) is particularly delicate because the fold identifies vertices in the base binding, and lifting it to a path of parallel copies requires checking compatibility with the parallel edges. Similar issues occur for D_5⊠A_7 → B_4⊠C_4 and for the B_n*C_n families. Since C","section":"§3.5, Proposition 3.22"},{"comment":"The contradiction argument has a gap. It assumes a vertex k adjacent in Δ to j, and then compares the number of red-blue paths from i to k with the number of blue-red paths from i to k. The proof states that 'the number of blue-red paths i→k will be the same as the number of red-blue paths k→i, as these paths are all made up of simple edges' and then asserts 'Admissibility gives (Γ∆)_{ik}=(∆Γ)_{ik}=(Γ∆)_{ki}=(∆Γ)_{ki}'. The second equality is not justified for non-symmetric Coxeter adjacency matrices; it depends on symmetry of the simple-edge submatrices, which must be stated explicitly. Also, the existence of the Δ-neighbor k and the possibility that the nonsimple edge is incident to a vertex with no Δ-neighbor are not discussed. Since Lemma 3.18 is used to eliminate several candidate double bindings in the proof of Theorem 3.10, the argument needs to be made rigorous or replaced by a d","section":"§3.3, Lemma 3.18 (proof)"},{"comment":"Proposition 3.5 is cited from [Ste10], but [Ste10] proves the statement for ADE bigraphs, where the adjacency matrices are symmetric. Here it is applied to arbitrary Dynkin biagrams with non-symmetric Coxeter adjacency matrices, and it is used to conclude that every family listed in Theorem 3.6 is admissible. The generalization is not entirely formal: if a biagram is glued from admissible bindings, one must verify that the block matrices Γ and Δ commute when the component graph is a tree or path. This can likely be proved by the same argument as in [Ste10], but the paper should either provide the proof or explicitly state that the proof carries over verbatim. As written, the admissibility of all listed families depends on an unproved generalization.","section":"§3.4, Proposition 3.5"}],"minor_comments":[{"comment":"In the paragraph after Table 1, 'for all k∈[6]' should be 'for all k∈[5]', since the matrix is 5×5.","section":"§2.3, Example 2.6"},{"comment":"The table heading says 'for B_3 ▷ ◁1 G2' but the dynamic is of a 5-vertex biagram; it would be helpful to clarify the correspondence between the columns and the vertices, especially since some entries appear only every other row.","section":"§2.3, Example 2.6"},{"comment":"The formula for the dominant eigenvector of E_n is hard to parse; as written it appears to list fewer than n entries for E_6. Please rewrite with explicit indexing, since this vector is used in Table 3 and in the double-binding case analysis.","section":"§3.3, Proposition 3.13(v)"},{"comment":"The last line of the proof contains 'deg max(η_j)' which seems to be a typo for 'deg max(i,T_j(η_j))'.","section":"§4.3, Lemma 4.3"},{"comment":"These proofs are said to be identical to [GP19] with matrix notation. Given their importance in the 5-way equivalence, it would be helpful to include at least a detailed statement of the modifications needed in the non-symmetric, skew-symmetrizable setting, so that the reader does not have to reconstruct the argument from [GP19].","section":"§4.4, Proposition 4.18 and Proposition 4.17"},{"comment":"The statement 'Let n≥2. Let Λ∈{A_{2n}, B_{2n}, C_{2n}, E_n, F_4, G_2}' is inconsistent: E_n is not defined for arbitrary n≥2. Presumably E_6,E_7,E_8 are intended, and the parameter n should be renamed.","section":"§6, Conjecture 6.3"}],"recommendation":"major_revision","confidential_remarks":"This is a strong paper with an important main theorem, but the proof of Proposition 3.22 is the linchpin connecting the ADE case to all non-ADE families, and it is currently an outline rather than a complete proof. The editor should ask the author to supply a fully detailed proof, or to provide a certified computational verification for all families, before publication. The other main concern, Lemma 3.18, is more localized but still needs a rigorous rewrite. I see no indication of circularity or fitted parameters; the reliance on [GP19] and [Ste10] is heavy but standard for this type of classification. The conjectures in Section 6 are clearly labeled and are not used in the main proof."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline: this paper is likely correct and completes the Galashin-Pylyavskyy classification of Zamolodchikov periodic B-matrices to non-simply-laced finite type. The new results are the classification of admissible Dynkin biagrams (29 infinite families and 14 exceptions beyond Stembridge) and the reduction of every non-ADE example to an ADE bigraph by folding and transpose. That reduction is the cleanest part of the paper and it explains why the non-simply-laced cases were missing: they are shadows of the ADE picture. The equivalence theorem is adapted carefully to skew-symmetrizable matrices, and the W-graph section is a genuine extension rather than a throwaway. The conjectures are clearly labeled and supported by small checks. No circularity, no fitted parameters. The heavy reliance on GP19 is appropriate: the ADE case is the base and this paper mostly reduces the rest to it. The soft spots are in the proof of Proposition 3.22. The claim that every admissible Dynkin biagram can be unfolded to an ADE bigraph by reversing a fold/transpose sequence componentwise is the bridge that lets the ADE tropical periodicity theorem cover all non-ADE types. The proof gives the construction but does not verify, family by family, that the intermediate objects are admissible ADE bigraphs or that the lifted maps are bicolored automorphisms. I think the claim is true, because the folds are local and the parallel structures cooperate, but a referee needs those checks spelled out. There are smaller gaps: Lemma 3.18 silently assumes a vertex in a non-ADE component has a neighbour in Delta, which follows from connectedness but is not stated; and a few proofs are dismissed as identical to GP19 without enough detail to make the adaptation auditable. The title and abstract overclaim slightly: the main theorem is about B-matrices, not cluster algebras, and Remark 1.2 honestly acknowledges it. Overall, the central argument holds; the gaps are fillable rather than fatal. This paper deserves a serious referee. I would send it out and ask for an expanded proof of Proposition 3.22 and a sharper statement of which parts are new versus adapted.","headline":"A likely-correct completion of the GP19 classification to all finite-type Cartan matrices; the central bijection is sound, but the proof of Proposition 3.22 needs more detail before the paper is fully trustworthy.","tokens_in":739,"tokens_out":1562,"would_cite":true,"duration_ms":78071,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["13F60","05E99","22E46"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that Zamolodchikov periodic cluster algebras are in natural bijection with pairs of commuting finite-type Cartan matrices, with period dividing the sum of the Coxeter numbers, and completes the classification of all such pa","keywords":["cluster algebras","Zamolodchikov periodicity","T-systems","commuting Cartan matrices","Dynkin biagrams","folding","transpose","W-graphs"],"falsifier":"Take one of the newly classified non-simply-laced admissible biagrams, such as the exceptional B4 ⊠ C4, and run its tropical T-system with a generic initial labeling; if it does not repeat with period dividing 16, the period part of Theorem 1.1 is false. Alternatively, enumerate all pairs of commuting finite-type Cartan matrices up to moderate rank and look for one absent from Theorem 3.6.","tokens_in":38681,"feed_emoji":"🔄","tokens_out":5952,"duration_ms":47627,"temperature":0.7,"pith_summary":"Zamolodchikov periodicity is the property that an alternating sequence of cluster mutations eventually cycles rather than producing infinitely many new cluster variables. This paper's central claim is that the cluster algebras with this property are exactly those associated to pairs of commuting Cartan matrices of finite type, and that the period divides the sum of the two Coxeter numbers. On top of the previously classified simply-laced cases, the paper adds 29 infinite families and 14 exceptional types, covering all non-simply-laced examples. The backbone of the proof is a five-way equivalence showing that periodicity, admissibility of the associated Dynkin biagram, and existence of strictly subadditive or fixed-point labelings all say the same thing. The upshot is a complete classification: every periodic system arises from the classical ADE cases by folding and transposing.","feed_headline":"Zamolodchikov-periodic cluster algebras fully classified","feed_subtitle":"Every periodic example matches a pair of commuting finite-type Cartan matrices; period divides sum of Coxeter numbers.","key_machinery":"The central object is the admissible Dynkin biagram: a pair (Γ,Δ) of Coxeter adjacency matrices of Dynkin diagrams sharing a bipartite vertex set, with no common edges, whose Cartan matrices commute. A biagram is admissible exactly when the number of red-blue paths equals the number of blue-red paths between every pair of vertices. The proof's engine is Theorem 4.1's five-way equivalence, which chains admissibility through strictly subadditive labelings, fixed-point labelings, tropical T-system periodicity for all initial labels, and actual T-system periodicity. Two operations carry the classification: folding along a bicolored automorphism (which commutes with the relevant mutations) and gl","core_discovery":"Theorem 1.1 states that the Zamolodchikov periodic B-matrices are in natural bijection with pairs (Γ,Δ) of commuting Cartan matrices of finite type, with period dividing hΓ+hΔ. Theorem 4.1 makes this concrete: for a bipartite recurrent B-matrix, the T-system is periodic if and only if the unsigned parts Γ and Δ form an admissible Dynkin biagram — that is, each is a disjoint union of finite-type Dynkin diagram adjacency matrices, and the two matrices commute. The classification lists all such pairs: 29 infinite families and 14 exceptional types beyond the 6 infinite families and 11 exceptional types already known in the simply-laced case. The non-simply-laced members are not new primitive phe","pith_inferences":["Because transposition preserves periodicity without commuting with mutation, the periodic class likely carries a Langlands-dual symmetry; one could test whether periods and exchange graphs are transpose-invariant family by family.","The tropical chamber structure hinted at for tensor products with A1 suggests the space of initial labelings may be a fan whose chambers correspond to clusters; proving this would give a new geometric model for the cluster complex.","The folding argument suggests a practical way to compute exact periods: fold a periodic ADE T-system along orbit symmetries and track how the period divides, yielding explicit formulas instead of just the divisor bound.","The W-graph reformulation points beyond dihedral pairs: the same 29+14 list may classify nonnegative cells for other products of finite Coxeter groups."],"forward_implications":["Periods are controlled: for every periodic B-matrix, the period divides hΓ+hΔ, reducing period computations to Coxeter-number arithmetic.","The classification is complete: any future Zamolodchikov periodic example must appear among the enumerated families or their duals.","Non-simply-laced periodic systems are all folds and transposes of ADE ones, so periodicity for the entire class follows from the simply-laced base case.","The same admissible Dynkin biagrams classify nonnegative W-cells for I2(p)×I2(q), linking cluster periodicity to Kazhdan–Lusztig theory.","The fixed-point and strictly-subadditive-labeling characterizations give a concrete certificate that a given B-matrix is Zamolodchikov periodic."],"fun_headline_variants":["All Zamolodchikov periodic cluster algebras now classified","Periodic cluster algebras biject with commuting Cartan pairs","29 infinite families, 14 exceptional: periodic cluster algebras classified","Zamolodchikov periodic cluster algebras: full classification","Periodic cluster algebras: complete classification"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The proof relies on the earlier result that every admissible ADE bigraph has a periodic tropical T-system for every initial labeling, and on the claim that every non-simply-laced admissible biagram is reachable from an ADE bigraph by folds and transposes; if either assumption fails for a family, the bijection would not cover that family.","fun_headline_variants_meta":{"raw":{"variants":["All Zamolodchikov periodic cluster algebras now classified","Periodic cluster algebras biject with commuting Cartan pairs","29 infinite families, 14 exceptional: periodic cluster algebras classified","Zamolodchikov periodic cluster algebras: full classification","Periodic cluster algebras: complete classification"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001113,"raw_usage":{"total_tokens":4492,"prompt_tokens":785,"completion_tokens":3707,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":529,"completion_tokens_details":{"reasoning_tokens":3643}},"tokens_in":529,"tokens_out":3707,"duration_ms":23802,"temperature":1.0,"reasoning_tokens":3643,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T08:54:12.469607+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take one of the newly classified non-simply-laced admissible biagrams, such as the exceptional B4 ⊠ C4, and run its tropical T-system with a generic initial labeling; if it does not repeat with period dividing 16, the period part of Theorem 1.1 is false. Alternatively, enumerate all pairs of commuting finite-type Cartan matrices up to moderate rank and look for one absent from Theorem 3.6.","supporting_citations":[],"review_version":1}