{"id":"2c232290-468d-4e10-a248-da845f8ee40d","arxiv_id":"2505.01228","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper proves that infinite rank cluster algebras are precisely the ind-objects of finite rank cluster algebras, and that the coordinate ring of the Sato-Segal-Wilson Grassmannian is such an ind-cluster algebra.","lead":"Cluster algebras of infinite rank are shown to be exactly the ind-objects of a category of finite rank cluster algebras, and the coordinate ring of the Sato-Segal-Wilson Grassmannian is shown to carry this infinite rank cluster structure. The result gives a categorical home for infinite-rank cluster combinatorics and ties Plücker mutations to equations of the KP hierarchy.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'precisely the ind-objects' claim is false as stated: finite-rank cluster algebras are representable ind-objects; the proof only supports Ind(mCl_f) ≅ mCl with infinite-rank objects as non-representable ind-objects.","rationale":"The reader's verdict of CONDITIONAL is appropriate, but I would shift the weight from the imported forward direction to the misstated ind-object claim. The forward direction from [Gra15, Theorem 4.6] is not reproved in the paper, yet it is likely safe: for a countable rank seed, one can form an increasing union of finite subseeds closed under neighbours of exchangeable variables, because ex-local finiteness makes each exchangeable row finite, and the inclusion maps are inducible melting cluster morphisms; the colimit is then the original cluster algebra. The uncountable case reduces to countable exchangeably connected components by the paper's own Remark 2.15. Thus the imported theorem is a dependency, but not a probable failure point. The more concrete and definite issue is that the abstract and Corollaries 1.4 and 2.35 claim that infinite rank rooted cluster algebras are precisely the ind-objects of mCl_f. In the standard sense of ind-objects, representable ind-objects are included, so finite rank rooted cluster algebras are also ind-objects. The proof itself demonstrates the stronger and correct statement Ind(mCl_f) ≅ mCl, which implies that infinite rank objects are exactly the non-representable ind-objects. This is not merely stylistic: as written, the headline claim is false, and the abstract and corollaries must be corrected. The main theorems, including the construction of filtered colimits, compactness of finite rank objects, and the identification of C[Gr] as the colimit-induced cluster algebra, are not affected; hence conditional acceptance with a required clarification remains the right verdict.","tokens_in":37668,"tokens_out":40645,"duration_ms":394440,"concrete_test":"Check the definition of ind-object in [KS06, Definition 6.1.1] for whether constant filtered diagrams are admitted. If they are, take any finite rank rooted cluster algebra, for instance the rectangle seed for Gr_{1,1}; its constant filtered diagram is an ind-object of mCl_f whose colimit is that same finite rank algebra, giving a direct counterexample to Corollary 2.35's 'precisely infinite rank' as stated. The correction is to replace 'precisely the ind-objects' with 'precisely the non-representable ind-objects' and to state the main categorical result as the equivalence Ind(mCl_f) ≅ mCl.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Corollaries 1.4 and 2.35, together with the abstract, assert that rooted cluster algebras of infinite rank are precisely the ind-objects of mCl_f. This is internally inconsistent with the proof. Under the standard definition (KS06, Def. 6.1.1), ind-objects include representable ones: for any finite-rank rooted cluster algebra A in mCl_f, the constant filtered diagram on A is an ind-object whose colimit is A itself. Thus every finite-rank rooted cluster algebra is also an ind-object of mCl_f, and the claimed 'precisely' fails as written. What the arguments actually prove is the category equivalence Ind(mCl_f) ≅ mCl, via Theorems 2.12, 2.34, 2.11 and [KS06, Cor. 6.3.5]; under this equivalence, the infinite-rank objects correspond exactly to the non-representable ind-objects, that is, objects not isomorphic to an object of mCl_f. The abstract and Corollaries 1.4 and 2.35 need to be restated in this form. This is a genuine correctness issue in the headline claim, although it does not invalidate the underlying categorical equivalence or the application to C[Gr]. The forward direction imported from [Gra15, Theorem 4.6] is a separate dependency, but I do not regard it as the weakest point: a countable-rank seed can be exhausted by finite subseeds closed under exchangeable neighbours because every exchangeable row is finite, making that direction elementary and verifiable. The misstated ind-object claim, by contrast, is false under the standard definition and appears in the central advertised result.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies infinite-rank cluster algebras through the category mCl of rooted cluster algebras with inducible melting cluster morphisms, allowing seeds of arbitrary cardinality. The main categorical results are that mCl is closed under filtered colimits (Theorems 2.12 and 2.31), that the finite-rank rooted cluster algebras are precisely the compact objects of mCl (Theorem 2.34), and, using the first author's earlier colimit decomposition of infinite-rank cluster algebras (Theorem 2.11, imported from [Gra15]), that every object of mCl is a filtered colimit of finite-rank objects. The paper then applies this framework to the Fioresi–Hacon coordinate ring C[Gr] of the Sato–Segal–Wilson Grassmannian, showing that C[Gr] is an infinite-rank cluster algebra with initial seed the infinite rectangle quiver Q∞ (Theorem 3.3), yielding Laurent/positivity statements for Plücker coordinates and a total-positivity criterion for the infinite Grassmannian, and connecting cluster mutations to three-term relations of the KP hierarchy.","tokens_in":37827,"tokens_out":12885,"duration_ms":137458,"significance":"The categorical construction in Sections 2.5–2.7 is detailed and, as far as I checked, internally consistent: the ind-seed construction, the cocone verification (Proposition 2.28), and the compactness argument (Theorem 2.34) are carefully written. If correct, the equivalence mCl ≃ Ind(mCl_f) gives a clean conceptual framework for infinite-rank cluster algebras. The application to C[Gr] is concrete and makes a strong, checkable prediction (Corollary 3.4). The main defect is that the headline statement 'infinite rank cluster algebras are precisely the ind-objects of mCl_f' is false under the standard definition of ind-object; the correct statement is the category equivalence, with infinite-rank objects corresponding to the non-representable ind-objects. The technical core of the paper is not invalidated by this misstatement, but the abstract and corollaries must be corrected.","major_comments":[{"comment":"Under the standard definition (KS06, Def. 6.1.1), the constant filtered diagram on any finite-rank rooted cluster algebra A is an ind-object of mCl_f whose colimit is A; hence finite-rank objects are also ind-objects. The claims that infinite-rank cluster algebras are 'precisely' the ind-objects of mCl_f are therefore false as written. What the proofs show is the category equivalence mCl ≃ Ind(mCl_f), via Theorems 2.11, 2.12, 2.34 and [KS06, Cor. 6.3.5]; under this equivalence the objects outside mCl_f correspond exactly to the non-representable ind-objects. Please restate the abstract, Corollary 1.4, and Corollary 2.35 accordingly, and adjust the informal sentence after Corollary 1.4.","section":"Abstract; Corollary 1.4; Corollary 2.35; paragraph after Corollary 1.4"},{"comment":"The assertion in Corollary 3.4(i) that every Plücker variable is a Laurent polynomial with non-negative integer coefficients in the rectangle variables does not follow from the general Laurent phenomenon alone, and positivity of coefficients is not a theorem for arbitrary cluster algebras. Since Theorem 1.7 and Corollary 1.10 depend on this positivity statement, the manuscript should cite the specific positivity result for finite-rank Grassmannian cluster algebras with the rectangle seed (e.g., Scott's theorem in [Sco06]) and explain that it passes through the colimit via Proposition 2.33. As written, this is a gap in the support of a central application.","section":"Corollary 3.4; Theorem 1.7; Corollary 1.10"}],"minor_comments":[{"comment":"The abstract contains a grammatical error: 'A prototypical examples' should be 'A prototypical example'.","section":"Abstract"},{"comment":"The proof of Theorem 2.31 contains a typo: 'roooted' should be 'rooted'.","section":"Theorem 2.31 proof"},{"comment":"The reduction from filtered colimits to directed colimits via [AN82, Theorem 1] is invoked without explanation; please add a sentence stating the content of that theorem so readers can follow the logical dependence.","section":"Theorem 2.12 proof"},{"comment":"The proof of Proposition 3.13 asserts the converse direction (every Γ∞-Postnikov diagram gives a Plücker cluster) rather than proving it; please spell out why a finite geometric exchange at a quadrilateral cell always produces a Plücker variable rather than a general Laurent polynomial, or make the statement conditional on that fact.","section":"Proposition 3.13"},{"comment":"Remark 4.9 contains unresolved citation placeholders '[?]' for the Kodama–Williams references; these should be filled before publication.","section":"Remark 4.9"}],"recommendation":"major_revision","confidential_remarks":"The misstatement about ind-objects in the abstract and main corollaries is the principal correctness issue; it is fixable by restatement, and the underlying categorical equivalence and Grassmannian application appear sound. The reliance on [Gra15] for the forward direction is legitimate, but the paper should be explicit that Theorem 2.11 is imported. The positivity gap in Corollary 3.4 needs a reference or proof. The paper fits the scope of the journal and makes a solid contribution to the infinite-rank cluster algebra literature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Before you read it: the advertised headline is wrong, but the paper itself is right. The abstract and Corollaries 1.4 and 2.35 say that infinite rank cluster algebras are precisely the ind-objects of the category of finite rank rooted cluster algebras, mCl_f. Under the standard definition of ind-object, every finite rank object is also an ind-object (through a constant diagram), so 'precisely' is false. What the paper actually proves is the equivalence Ind(mCl_f) ≅ mCl, with infinite rank objects corresponding to the non-representable ind-objects. That is the correct and strong statement, and the authors should restate their main claim accordingly.\n\nThe new mathematics is real and well executed. The paper constructs filtered colimits in the category mCl of rooted cluster algebras with inducible melting morphisms (Theorem 2.31), proves the compact objects are exactly the finite rank ones (Theorem 2.34), and then uses a standard theorem to get the ind-completion equivalence. I read the construction of the ind-seed and the cocone verification closely; they are detailed and coherent. The compactness proof is involved but convincing. The application to the coordinate ring of the Sato-Segal-Wilson Grassmannian is a nice payoff: C[Gr] is exhibited as an infinite rank cluster algebra with the rectangle quiver as initial seed, and the positivity statement for rectangular Plücker coordinates follows. This is the first cluster structure on that ring, as far as I know.\n\nSoft spots, in proportion. The misstated 'precisely' is the main one, and it is in the abstract and two corollaries—so a reader could be misled. It is fixable without touching the proofs. The forward direction, that every infinite rank cluster algebra is a filtered colimit of finite rank ones, is imported from [Gra15] (Theorem 2.11). That is a published result and the sketch here is plausible; I do not see it as a serious vulnerability. The paper uses the Laurent phenomenon for infinite rank seeds without much comment; it should at least cite [GG14] for that. The positivity corollary's citation is a bit vague.\n\nBottom line: this is a solid paper with a genuine flaw in its main advertised statement. The core mathematics deserves a serious referee and the paper should be sent out, with the authors asked to correct the ind-object claim. I would cite it after the correction; even now, the equivalence is the useful statement.","headline":"Strong paper on infinite rank cluster algebras, but the advertised 'precisely' ind-object claim is false as stated; the actual theorem is the ind-completion equivalence, which is still a big deal.","tokens_in":38545,"tokens_out":3710,"would_cite":true,"duration_ms":35733,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["13F60","14M15","37K10","05E05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that cluster algebras of infinite rank are precisely the ind-objects of finite-rank rooted cluster algebras, and that the coordinate ring of the Sato–Segal–Wilson Grassmannian is an infinite-rank cluster algebra.","keywords":["cluster algebras","infinite rank","ind-objects","Sato–Segal–Wilson Grassmannian","KP hierarchy","Plücker coordinates","filtered colimits","total positivity"],"falsifier":"Exhibit an infinite-rank rooted cluster algebra whose initial seed is not a filtered colimit of finite-rank seeds in mCl, for instance one with an exchangeably connected component of uncountable rank. Such a component would violate the countable-rank decomposition on which the proof of Theorem 2.11 depends, and would reduce the claimed equivalence to a one-sided inclusion.","tokens_in":37274,"feed_emoji":"🌊","tokens_out":6889,"duration_ms":63538,"temperature":0.7,"pith_summary":"The paper establishes that cluster algebras of infinite rank are exactly the ind-objects of a natural category of finite-rank rooted cluster algebras: the category of inducible melting cluster morphisms has filtered colimits, its compact objects are precisely finite-rank rooted cluster algebras, and every infinite-rank example is a filtered colimit of finite-rank ones. This gives a precise sense in which infinite-rank cluster algebras are built from finite-rank data. The paper then applies this to the coordinate ring of the Sato–Segal–Wilson Grassmannian, showing it is an infinite-rank cluster algebra whose initial seed is the infinite rectangle quiver. Consequently every Plücker coordinate is a Laurent polynomial with non-negative integer coefficients in rectangular Plücker coordinates, which yields a rectangular-coordinate test for total positivity and connects cluster mutation to the KP hierarchy.","feed_headline":"The Sato–Segal–Wilson Grassmannian is a cluster algebra","feed_subtitle":"Its Plücker coordinates are positive Laurent polynomials in rectangular coordinates, via colimits of finite Grassmannian cluster algebras.","key_machinery":"The central object is the category mCl of rooted cluster algebras with inducible melting cluster morphisms: maps that send exchangeable cluster variables to exchangeable variables or nonzero integers, may send frozen variables to exchangeable ones, and commute with mutation from the initial seed. The proof that filtered colimits exist works by constructing an ind-seed from a directed system: take the colimit of the initial cluster sets, then define the exchange matrix entries by a uniform-attainment condition on finite submatrices, with signs fixed on each exchangeably connected component. This ind-seed construction carries the argument that mCl is the ind-completion of its finite-rank part. For the Grassmannian application, the relevant seed is the infinite rectangle quiver Q∞, with vertices labelled by rectangular Young diagrams, formed as the colimit of Scott's rectangle seeds for finite Grassmannians.","core_discovery":"The category mCl of rooted cluster algebras with inducible melting cluster morphisms is closed under filtered colimits, and its compact objects are exactly the finite-rank rooted cluster algebras. Combining this with the earlier result that every infinite-rank rooted cluster algebra is a filtered colimit of finite-rank ones, the paper concludes that cluster algebras of infinite rank are precisely the ind-objects of the category of finite-rank rooted cluster algebras. Applied to the coordinate ring C[Gr], defined as the colimit of the coordinate rings of finite Grassmannians, this shows that C[Gr] is a cluster algebra of infinite rank with initial seed the infinite rectangle quiver Q∞. Every Plücker variable is therefore a Laurent polynomial with non-negative integer coefficients in the rectangular Plücker coordinates, and a point of the Sato–Segal–Wilson Grassmannian is totally positive exactly when all its rectangular Plücker coordinates are positive.","pith_inferences":["The ind-completion viewpoint suggests a universal-property approach to infinite-rank cluster algebras: any colimit-preserving invariant of mCl is determined by its values on finite-rank rooted cluster algebras, which could simplify future classification questions.","The same machinery of ind-seeds could plausibly adapt to other settings with a rooted mutation category, such as quantum cluster algebras or categories with freezing instead of melting, once filtered colimits and compact objects are identified.","The rectangular-positivity criterion may give a practical test for total positivity of KP τ-functions: check only rectangular Schur coefficients, which are in principle computable from the admissible basis of a point on the Grassmannian.","The finite four-valent subquivers constructed in the Postnikov-diagram section might provide finite-dimensional truncations whose mutations stabilize to the full infinite cluster structure, a stability property the paper does not explicitly address."],"forward_implications":["Every infinite-rank cluster algebra is a filtered colimit of finite-rank rooted cluster algebras, so structural questions about infinite-rank cluster algebras can be reduced to compatible families of finite-rank data.","The coordinate ring of the Sato–Segal–Wilson Grassmannian is an infinite-rank cluster algebra with an explicit initial seed, and its clusters give maximal algebraically independent sets of Plücker variables.","Every Plücker coordinate is a Laurent polynomial with non-negative integer coefficients in the rectangular Plücker coordinates, with support only inside the minimal bounding box of the partition.","A point of the Sato–Segal–Wilson Grassmannian is totally positive if and only if all its rectangular Plücker coordinates are positive, giving a finite-looking positivity test for KP τ-functions.","The KP equation appears as a quiver mutation in the infinite rectangle quiver, and the three-term exchange relations yield an infinite family of 3-term Plücker relations and associated PDEs for τ-functions."],"supporting_citations":[{"why":"Supplies the forward direction: every rooted cluster algebra of infinite rank is a filtered colimit of finite-rank ones, used in Theorem 2.11 and Corollary 2.35.","marker":"[Gra15]"},{"why":"Defines the category of rooted cluster algebras with melting cluster morphisms, the starting point for the category mCl.","marker":"[ADS14]"},{"why":"Provides the foundational definitions of cluster algebras, seeds, mutation, and the Laurent phenomenon used throughout.","marker":"[FZ02]"},{"why":"Establishes the cluster algebra structure on coordinate rings of finite Grassmannians, which induces the colimit cluster structure on C[Gr].","marker":"[Sco06]"},{"why":"Defines C[Gr] as a colimit of finite Grassmannian coordinate rings and identifies it with the coordinate ring of the Sato Grassmannian.","marker":"[FH04]"},{"why":"Supplies the infinite Grassmannian Gr, its Plücker embedding, and the connection between its points and τ-functions of the KP hierarchy.","marker":"[SW85]"},{"why":"Introduces the universal Grassmann manifold and the correspondence between its points and solutions of the KP hierarchy.","marker":"[Sat83]"},{"why":"Provides the finite-rank result that maximal weakly separated sets of Plücker variables are exactly algebraically independent clusters, used to compare Plücker clusters in C[Gr].","marker":"[OPS15]"},{"why":"Introduces inducible melting cluster morphisms, the morphisms that make up mCl and guarantee uniqueness by values on the initial cluster.","marker":"[CZ16]"},{"why":"Provides the ind-object formalism used to identify the ind-completion of mCl_f with all of mCl.","marker":"[KS06]"}],"fun_headline_variants":["Infinite-rank cluster algebras are ind-objects of finite ones","Sato–Segal–Wilson Grassmannian yields infinite cluster algebra","Infinite Grassmannian's coordinate ring is a cluster algebra","Colimits yield infinite Grassmannian cluster algebra","Ind-objects characterize infinite cluster algebras"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The equivalence between infinite-rank cluster algebras and ind-objects rests on the imported result that every infinite-rank rooted cluster algebra is a filtered colimit of finite-rank ones; the paper proves the converse direction but does not re-prove that decomposition.","fun_headline_variants_meta":{"raw":{"variants":["Infinite-rank cluster algebras are ind-objects of finite ones","Sato–Segal–Wilson Grassmannian yields infinite cluster algebra","Infinite Grassmannian's coordinate ring is a cluster algebra","Colimits yield infinite Grassmannian cluster algebra","Ind-objects characterize infinite cluster algebras"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001856,"raw_usage":{"total_tokens":7250,"prompt_tokens":869,"completion_tokens":6381,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":485,"completion_tokens_details":{"reasoning_tokens":6300}},"tokens_in":485,"tokens_out":6381,"duration_ms":48438,"temperature":1.0,"reasoning_tokens":6300,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:24:25.430063+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit an infinite-rank rooted cluster algebra whose initial seed is not a filtered colimit of finite-rank seeds in mCl, for instance one with an exchangeably connected component of uncountable rank. Such a component would violate the countable-rank decomposition on which the proof of Theorem 2.11 depends, and would reduce the claimed equivalence to a one-sided inclusion.","supporting_citations":[],"review_version":1}