{"id":"eab57c2a-6ce5-4331-9332-a123f13bf59a","arxiv_id":"2506.19443","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every rectangular semistandard Young tableau induces a positroidal subdivision of the hypersimplex, and for Gr(2,n) the non-frozen prime tableaux are exactly the coarsest such subdivisions.","lead":"This paper connects two areas of combinatorics: the special tableaux used to build the dual canonical basis of the Grassmannian ring, and the geometric subdivisions of the hypersimplex polytope. It proves that in the simplest nontrivial case the matching is one-to-one, and it proposes a formula for counting these subdivisions in general.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.2's proof assumes the unproved additivity Conjecture 5.3, so the rank-two correspondence in the abstract is not established as written.","rationale":"The central Theorem 4.1 is well-supported: it follows from Speyer-Williams and the equality of the positive tropical Grassmannian with the positive Dressian, and I do not object to it. The concern is isolated to the advertised Gr(2,n) refinement. The proof of Theorem 5.2 explicitly relies on an equality that the authors themselves leave open in Conjecture 5.3, and the converse is asserted via a phylogenetic-tree indexing rather than a formal bijection. This is a genuine correctness risk in the paper's strongest advertised claim, not a stylistic issue. The construction is plausible and the paper has computational evidence, so a conditional verdict rather than rejection is appropriate. This matches the reader's weakest-assumption analysis and does not change the reader's CONDITIONAL verdict.","tokens_in":11420,"tokens_out":14144,"duration_ms":149744,"concrete_test":"Independently verify Conjecture 5.3 for k = 2: enumerate all weakly separated tuples of one-column tableaux in SSYT(2,[n]) for n <= 8, compute both sides using an independent implementation of the Speyer-Williams map (or the authors' linked Sage/polymake code), and compare. Any failure falsifies the inference in Theorem 5.2; if all pass, prove the identity from the explicit 2-row web-matrix formula and re-derive Theorem 5.2 with that lemma inserted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 5.2 contains the paper's most load-bearing gap. After invoking the split decomposition theorem to write the weight w_Sigma of a non-split subdivision as w_1 + ... + w_j, the proof says that because F_{k,n} is a bijection, \"surjectivity\" gives tableaux T_1,...,T_j with T = union_i T_i and F_{k,n}(v_{T_i}) inducing S_i. Surjectivity gives no such tableau decomposition: F_{k,n} is piecewise-linear, and the needed identity F_{k,n}(v_{union T_i}) = sum_i F_{k,n}(v_{T_i}) is exactly the open question posed in Conjecture 5.3, which is not proved. Without this identity the contradiction that a one-column prime tableau cannot decompose is not obtained, so the forward direction of the claimed rank-two correspondence fails. The converse direction is also only asserted informally via phylogenetic trees and not derived from a stated theorem. Both are likely repairable and supported by examples, but as written the abstract's precise \"correspond precisely\" claim is conditional.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper connects dual canonical basis elements of Grassmannian cluster algebras, indexed by rectangular semistandard Young tableaux, to positroidal subdivisions of the hypersimplex. The main theorem (Theorem 4.1) asserts that for every T in SSYT(k,[n]), the Speyer--Williams weight F_{k,n}(v_T) lies in the positive Dressian and hence induces a positroidal subdivision. For k=2, Theorem 5.2 claims that non-frozen prime tableaux induce split (coarsest) subdivisions, and the abstract states a precise correspondence with coarsest subdivisions. The paper also states Conjectures 4.2 and 5.8 about the k>2 case and supplies SageMath and polymake data.","tokens_in":11626,"tokens_out":6057,"duration_ms":62524,"significance":"The potential significance is substantial: if Theorem 4.1 holds, every dual canonical basis element acquires a canonical polyhedral subdivision, giving a new bridge between the representation-theoretic basis of Grassmannian cluster algebras and tropical geometry. The paper's main positive contribution is Theorem 4.1, which is a concise corollary of Speyer--Williams' parametrization together with the equality of the positive tropical Grassmannian and the positive Dressian. The computational evidence and the public code repository are concrete strengths. However, the advertised rank-two correspondence is not established as written: the proof of Theorem 5.2 relies on the unproved additivity Conjecture 5.3, and the converse direction is only asserted informally. The core idea is promising and likely repairable, but the abstract's precise 'correspond precisely' claim currently outruns the proof.","major_comments":[{"comment":"The proof invokes the split decomposition theorem to write wΣ = w1 + ... + wj and then asserts that surjectivity of F_{k,n} gives tableaux T1,...,Tj with T = ∪ Ti and with F(v_{T_i}) inducing S_i. Surjectivity alone does not yield such a tableau decomposition of T; it only gives existence of some preimage for each weight. The needed identity F(v_{∪ Ti}) = Σ F(v_{Ti}) is exactly the content of Conjecture 5.3, which is stated later in the paper and is not proved. Without this identity, the contradiction that a one-column prime tableau cannot decompose is not obtained. This is a load-bearing gap in the forward direction of the rank-two theorem.","section":"§5.1, proof of Theorem 5.2"},{"comment":"The converse of the claimed correspondence is only asserted informally. The text says that split positroidal subdivisions correspond to phylogenetic trees with exactly one internal edge and that non-frozen prime tableaux provide a canonical indexing, and it then lists the cells of the subdivision. No theorem or proof is supplied that every split positroidal subdivision of Δ(2,n) arises from the image of a non-frozen prime tableau under F_{2,n}. Since the abstract claims these tableaux 'correspond precisely' to the coarsest subdivisions, this direction must be stated as a theorem and proved.","section":"§5.1, paragraph after Example 5.5"},{"comment":"Even if Conjecture 5.3 were proved, it is not immediately applicable in the proof of Theorem 5.2 as written: the conjecture assumes a union of pairwise weakly separated one-column tableaux, but the proof's split decomposition yields weights w_i whose preimages are not shown to be pairwise weakly separated. Thus the additivity identity used in the proof is conditional on an additional structural assertion that is not stated or proved.","section":"§5.1, Conjecture 5.3"}],"minor_comments":[{"comment":"There is a notation mismatch: the split subdivisions are indexed by j, while the tableau decomposition is written as T = ∪_{i=1}^m T_i; the proof should use a single index consistently.","section":"§5.1, proof of Theorem 5.2"},{"comment":"The statement that 'F_{k,n} is a bijection' should specify the domain and codomain; F is a piecewise-linear bijection from the positive tropical Grassmannian fan to the corresponding subfan of the secondary fan, and this precision matters for the subsequent surjectivity argument.","section":"§5.1, proof of Theorem 5.2"},{"comment":"The phrase 'The empty tableau is denoted by 1' is confusing because 1 is also used as an entry of tableaux; consider denoting the identity tableau by ∅ or by a separate symbol.","section":"§3.2, definition of SSYT"},{"comment":"The enumeration formula (k-1)/2 · n(n-k-1) is stated as a conjecture for all k≥2, but for k=2 it coincides with the number of split subdivisions explicitly verified in Section 5.1; a short comment clarifying the relationship would help the reader calibrate the strength of the conjecture.","section":"§5.2, Conjecture 5.8"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: this paper gives every rectangular dual-canonical-basis tableau a positroidal subdivision of the hypersimplex via the Speyer-Williams map, and it does it honestly—Theorem 4.1 is a corollary of known results (Speyer–Williams, positive Dressian equals positive tropical Grassmannian, ray classification), not a hidden-fitting argument. The new part is the dictionary itself and the conjectural enumeration of split subdivisions, and that is worth taking seriously.\n\nWhat is good. The main construction is simple and credible: each tableau gives a weight in the positive Dressian, so each gives a positroidal subdivision. The paper then specializes to Gr(2,n). The structural lemma that non-frozen prime tableaux are two-element one-column tableaux is clean, and it does establish a genuine bijection between those tableaux and split subdivisions—if the additivity step holds. They also provide code and computational evidence for the higher-rank split conjecture for nine parameter pairs, which is more reproducible than many papers in this area. The conjectural formula (k-1)n(n-k-1)/2 is concrete and testable.\n\nSoft spots. The rank-two proof has a load-bearing gap that the stress-test note catches exactly. In Theorem 5.2, after decomposing the weight of a subdivision as a sum of split weights, the proof claims that surjectivity of the Speyer-Williams map gives tableaux T_i whose union is T and whose individual images are the splits. Surjectivity does not give that. The needed identity F(v_{union T_i}) = sum_i F(v_{T_i}) is exactly Conjecture 5.3, stated later and never proved. Without it, the contradiction for a one-column prime tableau does not go through. The converse direction, identifying split subdivisions with those one-column tableaux via phylogenetic trees, is also asserted informally rather than derived from a stated theorem. These are likely repairable—the examples all work, and the additivity conjecture might follow from known weak-separation facts—but as written the abstract's precise \"correspond\" claim is stronger than the proof supports. That should be fixed before publication.\n\nThe authors lean on several of their own prior results, including the tableau-basis correspondence. That is normal here and not a red flag; the results are established and the code is available. The paper is not circular in any fitting sense; there are no fitted parameters, and the claims are not tuned to examples.\n\nWho this is for: anyone working on tropical Grassmannians, matroid subdivisions, and cluster algebras. The dictionary is worth having. My own verdict: send it to a serious referee. It needs a revision, but the core construction is sound and the main gap is identifiable and plausibly fixable. If the authors add the missing additivity lemma or clearly demote the Gr(2,n) claim to a conjecture, the paper is publishable in a good combinatorics venue.","headline":"A sound dictionary between dual canonical tableaux and positroidal subdivisions, but the Gr(2,n) bijection in the abstract is ahead of the proof as written.","tokens_in":12145,"tokens_out":4720,"would_cite":true,"duration_ms":46638,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52B11","52B40","05E10","14M15","13F60"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that every dual-canonical-basis tableau of the Grassmannian cluster algebra induces a positroidal subdivision of the hypersimplex $\\Delta(k,n)$, and that non-frozen prime tableaux give precisely the coarsest subdivisions…","keywords":["Grassmannian cluster algebra","dual canonical basis","hypersimplex","positroidal subdivision","semistandard Young tableau","tropical Grassmannian","split subdivision","matroid polytope"],"falsifier":"Compute $F_{k,n}(v_{T_1\\cup T_2})$ for any two weakly separated one-column tableaux not covered by Example 5.4, using the same Plücker-coordinate formulas; if the weight is not the sum of the individual weights, the additivity behind Theorem 5.2 is false. A direct check of the rank-two bijection is also possible: list all non-frozen prime tableaux in $\\operatorname{SSYT}(2,n)$ and all split positroidal subdivisions of $\\Delta(2,n)$ for $n=6$; any split not indexed by a pair $\\{i,j\\}$, or any pair whose weight is not split, would falsify the correspondence.","tokens_in":11181,"feed_emoji":"📐","tokens_out":9030,"duration_ms":83743,"temperature":0.7,"pith_summary":"Every rectangular semistandard Young tableau with $k$ rows and entries in $[n]$ labels a distinguished element of the dual canonical basis of the coordinate ring of the Grassmannian $\\operatorname{Gr}(k,n)$. The paper proves that the Speyer–Williams tropical map assigns to each such tableau a weight that induces a positroidal subdivision of the hypersimplex $\\Delta(k,n)$, the $0$\\textendash$1$ polytope whose vertices are the $k$-subsets of $[n]$. For $k=2$, it proves that non-frozen prime tableaux are exactly the one-column tableaux attached to the split, or coarsest, positroidal subdivisions of $\\Delta(2,n)$. The paper also conjectures that the number of split positroidal subdivisions of $\\Delta(k,n)$ is $\\frac{k-1}{2} n(n-k-1)$ and verifies this by computation for a range of small $k,n$. If the main theorem is right, every dual canonical basis element acquires a canonical polyhedral incarnation, tying representation theory to discrete geometry.","feed_headline":"Grassmannian tableau basis cuts hypersimplex into positroidal cells","feed_subtitle":"Every basis tableau gets a polyhedral subdivision; in rank two the non-frozen primes are exactly the splits.","key_machinery":"The load-bearing object is the Speyer–Williams map $F_{k,n}\\colon \\mathbb{R}^{(k-1)(n-k)} \\to \\operatorname{Span}_{\\mathbb{R}}\\{e_J : J\\in \\binom{[n]}{k}\\}/L$, evaluated on $v_T=\\sum_{i,j} c_{i,j} e_{i,j}$, where $c_{i,j}$ counts how many times the fundamental tableau $T_{i,j}$ (a one-column tableau with entries $[j,j+k]\\setminus\\{i+j\\}$) appears in a factorization of $T$. Its tropicalization of Plücker coordinates sends tableaux to weight vectors. The argument's second machine is the positive Dressian: a weight vector induces a positroidal subdivision exactly when it lies on a cone of the positive Dressian, so the equality between the positive tropical Grassmannian and the positive Dressian converts tableau weights into subdivisions. For the rank-two theorem, the split decomposition theorem of polytopes reduces any subdivision to a refinement of compatible splits, and one-column prime tableaux encode the internal edges of the phylogenetic trees that index those splits.","core_discovery":"Formally, the central assertion is Theorem 4.1: for every tableau $T\\in \\operatorname{SSYT}(k,[n])$, the vector $F_{k,n}(v_T)$ obtained by evaluating the Speyer–Williams map on the tableau's fundamental-tableau multiplicities lies in the positive Dressian, hence its lower hull is a positroidal subdivision of $\\Delta(k,n)$. The proof routes through the equality of the positive tropical Grassmannian and the positive Dressian and through the classification of cones of the positive Dressian. In rank two, Theorem 5.2 identifies the non-frozen prime tableaux $\\{i,j\\}$ with the split positroidal subdivisions, those with exactly two maximal cells; the paper's stated correspondence is that these tableaux are precisely the coarsest subdivisions of $\\Delta(2,n)$. The paper further presents a conjectural enumeration and computational evidence for the split subdivision count in higher rank.","pith_inferences":["A natural test is whether the additivity identity of Conjecture 5.3 can be upgraded to a tableau-calculus rule: if it holds, the subdivision of a multi-column tableau is the common refinement of the splits of its weakly separated one-column factors, making the subdivision visibly computable from the tableau.","The exceptional prime tableaux in $\\operatorname{Gr}(3,8)$ that are not coarsest align with the difference between the positive tropical Grassmannian and the cluster complex; one could test whether, in general, non-coarsest prime tableaux are in bijection with cluster variables of degree greater than one.","The conjectural split count has the flavor of a simple closed form; a bijective proof might come from encoding a split by a pair $(i,j)$ together with a cyclic-gap datum, extending the rank-two phylogenetic-tree picture to higher $k$.","The framework suggests a dictionary between dual canonical basis factorizations and common refinements of subdivisions: prime tableaux should correspond to indecomposable subdivisions, so the tableau poset and the subdivision refinement poset may be compared directly."],"forward_implications":["Every element of the dual canonical basis of $\\mathbb{C}[\\operatorname{Gr}(k,n)]$ comes with a canonical positroidal subdivision of $\\Delta(k,n)$, so representation-theoretic data encoded by tableaux can be read polyhedrally.","For $\\operatorname{Gr}(2,n)$, the coarsest subdivisions of $\\Delta(2,n)$ are indexed by pairs $\\{i,j\\}$ with $i