{"id":"bd20719c-b443-49d5-a1a1-0076bef03958","arxiv_id":"2411.10933","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The dual flagged Weyl character is zero-one if and only if the defining diagram is multiplicity-free, namely it avoids twelve listed subconfigurations up to column swap.","lead":"This paper proves a conjecture of Mészáros, St. Dizier and Tanjaya: the dual character of a flagged Weyl module has all coefficients 0 or 1 exactly when its diagram avoids twelve specified subconfigurations. Since Schubert and key polynomials are special cases, this gives one criterion and one proof for the known zero-one results for both families.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The WLOG reductions (C1)/(C2) are likely fine; the real gap is the unproved R2 Case 2 merging step and omitted Lemma 4.6, which together underwrite the iterative bijection in Theorem 4.1.","rationale":"I read the proof in good faith and agree with the reader that the theorem is plausible and the overall strategy is coherent. However, I do not think the WLOG reductions (C1) and (C2) are the most load-bearing issue: the justifications given for them, although brief, can be supplied naturally, and the reductions are not where the case analysis is likely to break. The true risk is in the iterative bijection itself, specifically the R2 Case 2 algorithm and its R3 analogue. Lemma 4.6 is explicitly unproved, and the merging operation that feeds the next iteration is not shown to preserve the normalized multiplicity-free hypotheses. These are exactly the steps that would have to be airtight for Theorem 4.1 to cover all multiplicity-free diagrams. The reader noted Lemma 4.6 and the sketchiness of Section 4.3, but chose C1/C2 as the weakest assumption, so my agreement is only partial. Because the theorem is well-motivated and the identified gaps are fillable rather than obviously fatal, the correct verdict remains CONDITIONAL, which is what the reader already gave; hence I recommend no change to the verdict.","tokens_in":20103,"tokens_out":25314,"duration_ms":273086,"concrete_test":"Implement the construction of Theorem 4.1 computationally for all normalized multiplicity-free diagrams D in [n]×[n] with n ≤ 5 and all pairs C,C' ≤ D with x^C = x^{C'}: enumerate flagged fillings, execute the R1/R2/R3 algorithms exactly as written (including the Case 2 shuffle and the merging step), and check that every intermediate object is a valid flagged filling, the merged diagram satisfies the hypotheses of Theorem 4.1, and the final map is a sign- and weight-preserving bijection from F_D(C) to F_D(C'). If any instance fails, the proof has a concrete gap; if all pass, the concern reduces to a presentational one.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader flagged the reductions (C1) and (C2) below Theorem 4.1 as the weakest assumption, but those are probably sound: padding with empty columns and an extra absent row preserves the dual character because Y is upper-triangular, and a standard interval column [m] contributes the monomial factor x_1...x_m. The more serious gap is inside the proof of Theorem 4.1 itself. In Section 4.2, Case 2 (ik = q-1), the algorithm merges columns containing q-1 or q into the k-th region by replacing D_j with [q-2] ∪ {q} and correspondingly modifying C_j and C'_j. The paper never proves that this modified diagram is still a normalized multiplicity-free diagram satisfying (C1) and (C2), nor that the region taxonomy of Section 3.2 applies to it. The subsequent iteration of Φ depends on exactly those properties, so without this verification the induction over regions is incomplete. Moreover, Lemma 4.6, which justifies the crucial shuffling of the row-q entries, is stated with 'the proof is analogous ... and so is omitted'; Section 4.3 says the Type (R3) case is only sketched. These are not merely cosmetic omissions: if the multiset equality in Lemma 4.6 fails, or if the merged columns leave the class of diagrams on which Φ is defined, the claimed sign- and weight-preserving bijection may not exist for some multiplicity-free diagram. The theorem may still be true, but the current proof does not fully establish it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves the Mészáros–St. Dizier–Tanjaya conjecture that the dual character χ_D(x) of the flagged Weyl module of a diagram D ⊆ [n]×[n] is zero-one if and only if D is multiplicity-free, that is, D avoids the twelve multiplicitous configurations of Figure 1.1. The forward direction was known; the paper proves the reverse (Theorem 1.1), yielding Corollary 1.2 and special cases recovering the known zero-one criteria for Schubert polynomials (Fink–Mészáros–St. Dizier) and key polynomials (Hodges–Yong). The proof is combinatorial: via the flagged-filling expansion of Proposition 2.2, the claim reduces (Theorem 4.1) to constructing a sign- and weight-preserving bijection between F_D(C) and F_D(C′) whenever C, C′ ≤ D and x^C = x^{C′}. The construction normalizes D, assumes two reductions ((C1): columns have at least two crossings via grid embedding; (C2): no standard interval columns), classifies multiplicity-free diagrams into Type (R1)/(R2)/(R3) regions (Lemmas 3.1–3.4), and defines an iterative operation Φ. The Type (R1) case is worked out in detail (Section 4.1); Type (R2) contains the main technical work (Section 4.2); Type (R3) is sketched (Section 4.3); and Section 4.4 argues that the iteration gives the desired bijection.","tokens_in":20373,"tokens_out":22657,"duration_ms":219571,"significance":"Assuming the proof can be completed, the paper settles a conjecture and supplies a single combinatorial mechanism for three previously separate zero-one criteria. The criterion itself is a clean, parameter-free diagrammatic condition, and the structural analysis of multiplicity-free diagrams in Section 3 (normalization, regions, Lemmas 3.1–3.4) is a useful contribution independent of the main theorem. The authors are careful to ground the bijection in the flagged-filling expansion of det(Y^C_D) (Proposition 2.2, cited from [27]), with no fitted parameters and no circularity. The main shortcomings are not in the architecture but in missing details at load-bearing points: the unproved Lemma 4.6, the unverified merging step in Section 4.2 Case 2, the sketched Type (R3) case, and the one-sentence justification of reduction (C1). Of these, (C2) is already justified in the text, and (C1) is likely repairable with a short argument; the Section 4 gaps require more substantial additional detail.","major_comments":[{"comment":"Lemma 4.6 is load-bearing but its proof is omitted with the words 'the proof is analogous ... and so is omitted.' The statement is not literally analogous to Lemma 4.4: the paper explicitly notes that the multiset of elements equal to q−1 or q need not match between (a_1,...,a_d) and (b_1,...,b_d), precisely because of the k-th region. The subsequent shuffle producing (a′) is well-defined only if Lemma 4.6 holds, and the bijection in Theorem 4.1 depends on that shuffle. A complete proof of this lemma must be supplied before the argument goes through.","section":"§4.2, Lemma 4.6"},{"comment":"The merging step in Section 4.2, Case 2 (Figures 4.15–4.17) is under-specified and inconsistent as displayed. For columns whose F′_j contains q−1 or q, the paper replaces D_j by D_j ∪ [q−2] = [q−2] ∪ {q} and 'correspondingly' replaces C^(1)_j and C′_j by C^(1)_j ∪ [p−2] and C′_j ∪ [p−2]. Three things are missing. First, the equality asserts D_j ⊆ [q−2] ∪ {q}, but the merged columns are selected because their content contains q−1 or q, and [q−2] ∪ {q} contains no q−1; the formula therefore cannot hold as written (it also uses [p−2] where [q−2] appears intended). Second, no proof is given that the modified diagram is normalized and multiplicity-free, satisfies (C1)–(C2), that the modified C-columns still lie below the modified D-columns in Gale order (the sizes of C^(1)_j ∪ [q−2] and [q−2] ∪ {q} do not obviously match), or that the region taxonomy of §3.2 applies to the next iteration of Φ. Third, the iteration in §4.4 and its termination depend on this closure, which is not stated as a lemma. The merging step thus needs to be rewritten as a lemma with explicit hypotheses, construction, and verification.","section":"§4.2, Case 2 (merging step)"},{"comment":"The Type (R3) case, one of the three exhaustive region types, is only sketched: the text says the construction is 'nearly the same' as in §4.2 and defines Φ and Φ̂ by reference to the two cases of §4.2. This case differs materially from (R2): p is defined as the second-lowest box of the Type III column D_m rather than as n_m, and the column being adjusted after the first algorithm is the Type III column itself. The analogues of Lemmas 4.5 and 4.6 in this setting are not stated, and the figures do not replace the missing verification that the resulting filling is flagged and has the claimed weight. A full treatment of this case is required for Theorem 4.1 to cover all multiplicity-free diagrams.","section":"§4.3, Type (R3)"},{"comment":"The proof of Theorem 4.1 explicitly assumes (C1) (every column has at least two crossings, achieved by embedding into a larger grid) and (C2) (no standard interval column [m]). Of these, (C2) is adequately justified below Lemma 3.3. Reduction (C1), by contrast, is asserted in a single sentence at the beginning of §3.2. Since the entire region taxonomy and the three cases of Φ rest on the presence of a second crossing, the embedding argument should be stated as a lemma: it must show that multiplicity-freeness is preserved by the embedding (checking the twelve configurations), that the embedded diagram has at least two crossings per column, and that the zero-one property of the embedded dual character descends to χ_D. The descent is immediate because the relevant monomials involve only x_1,...,x_n, but the multiplicity-free check is not written down.","section":"§3.2 and §4, reductions (C1)–(C2)"}],"minor_comments":[{"comment":"The word 'subest' appears twice in the paragraph following Lemma 4.6; it should be 'subset'.","section":"§4.2, after Lemma 4.6"},{"comment":"In the exchange argument of case (1), the sentence 'Suppose that Fj1 has column reading word u ... and Fj1 has column reading word v' should refer to Fj2 for the second word; as written the notation is inconsistent.","section":"§4.4, sign-preservation argument"},{"comment":"The claim that reordering row-q entries does not change inversion numbers states that the moved entries are 'bigger' than any entry above; the argument only requires that they are at least as large as every entry above row q in the corresponding column, so that the bottom position of each column reading word never participates in an inversion. The wording should be adjusted, and the fact that the relevant columns have no entries below row q should be stated explicitly.","section":"§4.4, operation (2)"},{"comment":"There are several typos: 'flagged Weyl models' should be 'flagged Weyl modules' (Introduction), 'northewest diagrams' should be 'northwest diagrams' (Introduction), 'eigensapce' should be 'eigenspace' (Proposition 2.1), and 'frist region' should be 'first region' (§4.2).","section":"§1, §2, §4.2"},{"comment":"Theorem 4.1 is stated for normalized multiplicity-free diagrams, but its proof begins by imposing the extra assumptions (C1) and (C2) inside the proof. Stating the reductions as a lemma before the theorem and referencing it from the theorem would make the logical structure of the proof easier to check.","section":"Theorem 4.1"}],"recommendation":"major_revision","confidential_remarks":"The central conjecture is plausible and the architecture of the proof is sound, but three load-bearing points (omitted proof of Lemma 4.6, closure of the merging operation in §4.2 Case 2, and the sketched Type (R3) case) need to be completed before the proof of Theorem 4.1 is verifiable. These are gaps that appear fillable within the manuscript's scope, so I recommend major revision rather than rejection. I also note that Proposition 2.2 is cited from the authors' own paper [27]; the identity is independent of the argument being proved, so I see no circularity issue, and the citation appears appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a plausible and genuinely new proof of the Mészáros--St. Dizier--Tanjaya conjecture, but the write-up skips over the hardest parts of the bijection. It deserves a serious referee, even though the proof as printed is not yet complete.\n\nWhat's new: Theorem 1.1 settles the open 'if' direction: multiplicity-free diagrams have zero-one dual characters. That closes the criterion from [23] and specializes to the known Schubert and key polynomial results. The strategy is coherent: normalize diagrams, classify regions by first crossings, and build an iterative sign- and weight-preserving bijection between flagged fillings. The Type (R1) case is worked out in detail with examples, and the sign/weight preservation argument is clean.\n\nWhere it gets soft. The reductions (C1) and (C2) are not the real problem; the reader worried about them, but padding with empty columns and removing standard interval columns are standard and fine, since the dual character is essentially unchanged.\n\nThe real gaps are inside Section 4.2 and 4.3. Lemma 4.6 is stated with 'the proof is analogous ... and so omitted', but that lemma is what justifies the shuffling of row-q entries in Case 2. Without a proof, the multiset equality is simply asserted. Worse, the merging step in Case 2 replaces columns Dj by [q−2] ∪ {q} and then iterates Φ. The paper never checks that the modified diagram is still normalized, multiplicity-free, and satisfies (C1)/(C2), so the induction over regions is under-justified. If the merged diagram leaves the class on which Φ is defined, the iterative bijection may not cover all cases. Section 4.3 says the Type (R3) case is 'nearly the same' and sketches it, which is a lot to hand-wave after two omitted proofs.\n\nNothing here looks circular; Proposition 2.2 from [27] is an independent expansion. The citation pattern is fine.\n\nBottom line: if the missing details are filled, this is a notable theorem. As written, it is a solid conditional result. The right move is to send it to a referee who can check the R2/R3 machinery, not desk-reject it.","headline":"Proves the missing direction of the 2021 zero-one conjecture with a clear bijective strategy, but the proof has real gaps in the Type (R2)/(R3) cases; worth refereeing with revisions.","tokens_in":20950,"tokens_out":3251,"would_cite":true,"duration_ms":29854,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05E10","05E14","05A19","14N15"],"pacs":[],"model":"deepseek-v4-flash","headline":"Multiplicity-free diagrams have zero-one dual characters, completing an if-and-only-if criterion that unifies Schubert and key polynomial cases.","keywords":["flagged Weyl module","dual character","multiplicity-free diagram","zero-one polynomial","Schubert polynomial","key polynomial","flagged filling","sign- and weight-preserving bijection"],"falsifier":"A concrete counterexample would be a multiplicity-free diagram $D$ with two diagrams $C,C'\\le D$ such that $x^C=x^{C'}$ but $\\det(Y^C_D)\\ne \\det(Y^{C'}_D)$; expanding both determinants over flagged fillings and finding different signed sums would falsify Theorem 4.1 and hence Theorem 1.1. Short of that, a direct computation of $\\chi_D(x)$ for any small multiplicity-free diagram (say, all such diagrams in a $5\\times5$ grid) that produced a monomial coefficient larger than 1 would disprove the criterion.","tokens_in":19860,"feed_emoji":"🧮","tokens_out":7780,"duration_ms":76214,"temperature":0.7,"pith_summary":"This paper proves that the dual character $\\chi_D(x)$ of the flagged Weyl module attached to a diagram $D$ in the $[n]\\times[n]$ grid has all coefficients equal to 0 or 1 exactly when $D$ is multiplicity-free, meaning it avoids twelve explicitly listed four-by-two subdiagram patterns. The equivalence settles a conjecture proposed in [23] and, because Schubert polynomials and key polynomials are special cases of dual flagged Weyl characters, it gives one uniform proof of the previously known zero-one criteria for both families. The proof is combinatorial: it shows that every eigenspace of the flagged Weyl module is one-dimensional by constructing, for any two diagrams $C,C'\\le D$ with the same monomial $x^C=x^{C'}$, a sign- and weight-preserving bijection between the flagged fillings of $D$ with column-entry sets $C$ and $C'$. If correct, the theorem reduces a character-theoretic question to a purely local inspection of the diagram.","feed_headline":"Twelve forbidden patterns decide zero-one dual characters","feed_subtitle":"The new theorem unifies Schubert and key polynomial criteria through a sign- and weight-preserving bijection.","key_machinery":"The machinery is the iterative operation $\\Phi$ on flagged fillings. A flagged filling assigns to each box $(i,j)\\in D$ an integer at most $i$, with distinct entries within each column; the determinant identity (Proposition 2.2) expands $\\det(Y^C_D)$ as the signed sum of the monomial weights of all flagged fillings whose column-entry sets are the sets $C_j$. The bijection $\\Omega$ iterates $\\Phi$, which slides entries along rows and swaps entries between columns according to labels attached to the columns from the difference sets $[n_j]\\setminus C_j$; the three cases of $\\Phi$ correspond to the three possible shapes of the first region of the normalized diagram, classified by how many boxes lie below the second crossing. Because all moved entries slide within a fixed row, the weight $y^F$ is unchanged, and the column-reading inversion counts are unchanged, so sign and weight are both preserved.","core_discovery":"The central claim is Theorem 1.1: for every multiplicity-free diagram $D$, the dual character $\\chi_D(x)$ is zero-one. Combined with the converse direction already in [23, Proposition 3.11], this yields the if-and-only-if criterion of Corollary 1.2: $\\chi_D(x)$ is zero-one precisely when $D$ is multiplicity-free. The proof works by proving a stronger statement (Theorem 4.1): if $C$ and $C'$ are diagrams below $D$ with $x^C=x^{C'}$, then $\\det(Y^C_D)=\\det(Y^{C'}_D)$, which forces the coefficient of the monomial $x^a$ to be the dimension of the corresponding eigenspace. The equality is shown by a bijection $\\Omega$ from the flagged fillings of $D$ with column sets $C$ to those with column sets $C'$ that preserves both the inversion sign and the monomial weight, so the signed expansions of the two determinants agree term-by-term. Since the coefficient of $x^a$ counts the dimension of the eigenspace, the theorem follows.","pith_inferences":["Because the bijection is constructive, it could be turned into an algorithm that computes the coefficient of any monomial $x^a$ in $\\chi_D(x)$ by iterating $\\Phi$ rather than expanding the whole character; the paper describes the iteration but does not present it as an algorithm.","The criterion suggests a matroid-theoretic reading: for a multiplicity-free diagram the support of $\\chi_D(x)$ is the set of monomials $x^C$ with $C\\le D$, and zero-one-ness means this support is exactly the indicator function of the Schubert matroid bases, so Newton polytope information determines the entire polynomial.","A natural stress test is to enumerate all diagrams on small grids and check computationally that the zero-one property coincides with avoiding the twelve configurations; agreement for $n\\le 5$ would also exercise the two reduction steps that the proof uses.","The same $\\Phi$-based bijection may adapt to other settings where a Weyl-type module has a flagged-filling expansion, potentially yielding zero-one criteria for Kohnert polynomials for northwest diagrams beyond the Schubert and key cases."],"forward_implications":["Checking whether $\\chi_D(x)$ is zero-one reduces to inspecting twelve local four-by-two patterns in $D$; no expansion of the character is needed.","The known zero-one criteria for Schubert polynomials and for key polynomials both follow from one theorem, since Rothe diagrams and skyline diagrams are special cases of diagrams.","When $\\chi_D(x)$ is zero-one, the support of the polynomial is exactly the set of monomials $x^C$ with $C\\le D$, and the Newton polytope of $D$ completely determines the character.","For northwest diagrams, whose dual characters coincide with Kohnert polynomials, the same criterion applies; this covers cases the earlier key-polynomial methods did not reach.","The converse direction already available in the literature turns Theorem 1.1 into an iff: $\\chi_D(x)$ is zero-one if and only if $D$ is multiplicity-free."],"supporting_citations":[{"why":"poses the zero-one conjecture for dual characters and supplies the converse proposition (zero-one implies multiplicity-free) that completes the criterion.","marker":"[23]"},{"why":"provides the zero-one Schubert polynomial criterion and the observation behind the necessity direction for general diagrams.","marker":"[10]"},{"why":"provides the zero-one key polynomial criterion that the main theorem recovers as a special case.","marker":"[14]"},{"why":"identifies supports of dual characters with integer points of Newton polytopes and sets up the flagged Weyl module notation used throughout.","marker":"[9]"},{"why":"supplies the determinant expansion of the generator as a signed sum over flagged fillings, which the bijection must preserve.","marker":"[27]"},{"why":"introduces standard interval columns, whose removal as monomial factors is one of the two reductions in the proof.","marker":"[25]"}],"fun_headline_variants":["Zero-one dual characters: the missing criterion","Flagged Weyl modules: when dual characters are 0-1","New theorem settles zero-one dual character conjecture","Multiplicity-free diagrams give zero-one dual characters","Unified proof for Schubert and key polynomial criteria"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof rests on two reductions—embedding diagrams with too few crossings in a larger grid, and deleting full interval columns as mere monomial factors—and if either reduction fails for some valid diagram, the case analysis that builds the bijection would not cover all inputs.","fun_headline_variants_meta":{"raw":{"variants":["Zero-one dual characters: the missing criterion","Flagged Weyl modules: when dual characters are 0-1","New theorem settles zero-one dual character conjecture","Multiplicity-free diagrams give zero-one dual characters","Unified proof for Schubert and key polynomial criteria"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001497,"raw_usage":{"total_tokens":5988,"prompt_tokens":909,"completion_tokens":5079,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":525,"completion_tokens_details":{"reasoning_tokens":5004}},"tokens_in":525,"tokens_out":5079,"duration_ms":36024,"temperature":1.0,"reasoning_tokens":5004,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T19:08:55.900546+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete counterexample would be a multiplicity-free diagram $D$ with two diagrams $C,C'\\le D$ such that $x^C=x^{C'}$ but $\\det(Y^C_D)\\ne \\det(Y^{C'}_D)$; expanding both determinants over flagged fillings and finding different signed sums would falsify Theorem 4.1 and hence Theorem 1.1. Short of that, a direct computation of $\\chi_D(x)$ for any small multiplicity-free diagram (say, all such diagrams in a $5\\times5$ grid) that produced a monomial coefficient larger than 1 would disprove the criterion.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the zero-one Schubert polynomial criterion and the observation behind the necessity direction for general diagrams."},{"cited_title":"Hodges and A","cited_arxiv_id":null,"evidence_quote":"provides the zero-one key polynomial criterion that the main theorem recovers as a special case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the determinant expansion of the generator as a signed sum over flagged fillings, which the bijection must preserve."},{"cited_title":"M´ esz´ aros, L","cited_arxiv_id":null,"evidence_quote":"introduces standard interval columns, whose removal as monomial factors is one of the two reductions in the proof."}],"review_version":1}