{"id":"28ca0174-4135-48f5-91b9-bb176b625569","arxiv_id":"2412.16820","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Weyl alternation sets, the group elements that actually matter in Kostant's multiplicity formula, are always order ideals, and for sl_{r+1} they are fully described and counted for negative roots.","lead":"The paper proves that the nontrivial contributors to Kostant's weight multiplicity formula always form an order ideal in the weak Bruhat order, and gives the complete description of these contributors for the highest root and every negative root in type A. This turns a factorially hard alternating sum into a manageable combinatorial skeleton, with Fibonacci numbers and a closed generating function for the counts.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Definition 3.5's independence relation is too weak: with it, Theorem 3.10(2) is false (s1 and s2s3 in A4 are independent, but s1s2s3 ∉ A(α̃,−α̃)), so the type-A BAS lists and enumerations are unsupported as stated.","rationale":"The reader's conditional verdict targeted the omitted case computations behind Corollaries 4.9–4.11. My concern lands earlier and is more specific: the printed independence relation in Definition 3.5 makes Theorem 3.10(2) false, with an explicit type-A4 counterexample. This does not refute the headline theorem (Theorem 3.1), whose proof is correct, but it removes the current support for the paper's second advertised contribution, the complete characterization of A(α̃,μ) for negative roots, and for the enumerative results built on it. The flaw is likely fixable by redefining independence via extended influence (as Example 3.11's independent-set count implicitly does), but that change is not cosmetic: the existence proof in Theorem 3.10 and the coefficient argument in part (2) would both need substantial revision, and the BAS lists in Section 4 would need to be recomputed under the corrected relation. Hence the paper should not be accepted unconditionally; a conditional revision addressing the independence definition and the derived lists is appropriate.","tokens_in":33472,"tokens_out":21231,"duration_ms":176722,"concrete_test":"Enumerate A4(α̃,−α̃) directly from Lemma 4.1/4.2 (or by brute force over the 24 elements of S4): if s1s2s3 is not in the resulting set while {s1,s2s3} is I-disjoint, Theorem 3.10(2) fails as printed. As a second check, enumerate all subsets of the seven BAS elements in Example 3.11 whose printed influences are disjoint; if the count is not 11, the independence relation must be strengthened and every BAS list in Section 4 re-derived with the corrected relation.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 3.1 is sound: the covering-relation argument correctly shows σ(λ+ρ)−τ(λ+ρ) is a nonnegative simple-root combination. The load-bearing failure is in Theorem 3.10. Definition 3.5 says σ and τ are independent when I(σ)∩I(τ)=∅. With that definition, Theorem 3.10(2) asserts every product of pairwise independent BAS elements lies in A(λ,μ). This is false. In type A4, for λ=α̃ and μ=−α̃, Theorem 4.7 places s1 and s2s3 in BAS, and I(s1)∩I(s2s3)=∅. But for σ=s1s2s3, a direct computation gives σ(α̃+ρ)+α̃−ρ = 2α̃−4α1−2α2−α3, whose α1 coefficient is negative, so σ∉A(α̃,−α̃). Correspondingly, Example 3.11 lists only 11 independent sets, not the full family of I-disjoint subsets of its seven BAS elements. The proof's coefficient argument is invalid because a factor with an index adjacent to I(b) can change the coefficient of α_i; the intended condition must separate extended influences, not merely influences. Since Proposition 3.17 and Corollaries 4.9–4.11 inherit this flaw, the completeness of the type-A characterization and the Section 5 enumerations is not established by the arguments given.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the Weyl alternation set A(λ, μ) appearing in Kostant's weight multiplicity formula. The first main result, Theorem 3.1, proves that A(λ, μ) is an order ideal in the left and right weak Bruhat orders for any dominant integral λ, by showing that for a cover τ ⋖ σ the difference σ(λ+ρ)−τ(λ+ρ) is a nonnegative integer combination of simple roots. The paper then introduces the notion of basic allowable subwords (BAS) and claims in Theorem 3.10 that every element of A(λ, μ) corresponds bijectively to a pairwise independent subset of a unique BAS. For type A it characterizes BAS(α̃, μ) for negative roots μ (Corollaries 4.9–4.11), proves Fibonacci-type recurrences for the cardinalities |A_r(α̃, μ)|, and gives a bivariate generating function in Theorem 5.14.","tokens_in":33817,"tokens_out":24502,"duration_ms":192443,"significance":"The order-ideal theorem is a clean and apparently correct structural result; if it stands, it gives a useful global constraint on the support of Kostant's formula and reduces alternation-set computations to downward-closed sets. The BAS framework is a natural strengthening, and the type-A characterizations would extend earlier results of Harris and Harry. However, the central BAS theorem is not correct as stated, because the proposed independence relation is too weak; the type-A classification and all Section 5 enumerations rest on that theorem. The Appendix A case analysis is extensive and suggests that the intended combinatorial picture is plausible, but the manuscript in its current form does not establish it. No code or machine-checked artifacts are provided; the main theorem's proof is short and verifiable by hand.","major_comments":[{"comment":"Definition 3.5 defines independence by I(σ) ∩ I(τ) = ∅, and this is too weak for Theorem 3.10(2). In type A4 take λ = α̃ and μ = −α̃. Proposition 4.3 places s1 and s2s3 in A(λ, μ), and Definition 3.5 makes them independent because I(s1) = {1} and I(s2s3) = {2, 3}. A direct computation with ρ = 2α1 + 3α2 + 3α3 + 2α4 gives s1s2s3(α̃+ρ) + α̃ − ρ = 2α̃ − 4α1 − 2α2 − α3, whose α1-coefficient is negative; hence s1s2s3 ∉ A(λ, μ). The coefficient argument in the proof of part (2) is therefore invalid: a factor whose support is adjacent to I(b) can change the coefficient of α_i even when i ∉ I(b). The independence relation and the proof of part (2) must be repaired before the BAS results can be used.","section":"Definition 3.5 and Theorem 3.10(2)"},{"comment":"Example 3.11 is inconsistent with Definition 3.5. Under Definition 3.5, {s2, s3} is an independent subset of the seven listed BAS elements, because I(s2) = {2} and I(s3) = {3}; yet it is not among the 11 listed independent sets, and its product s2s3 would duplicate the BAS element s2s3. The stated 5-clique and the 11 independent sets are exactly what one obtains when independence is interpreted as the two influence sets being at Dynkin distance at least two. The definition, the proof of Theorem 3.10, and Example 3.11 therefore cannot all be correct. This is not a cosmetic issue, because a corrected independence relation is precisely what the coefficient argument in Theorem 3.10 needs.","section":"Example 3.11"},{"comment":"Corollaries 4.9–4.11 are the type-A payoff, but each is introduced as being obtained by computing the intersection BAS(α̃, −α̃) ∩ A(α̃, μ), and no computation is supplied; Example 4.8 verifies a single excluded element. Since Proposition 3.17 inherits the current BAS theory, and since all Section 5 enumerations (Propositions 5.2, 5.4, 5.5 and Theorem 5.14) are based on these BAS lists, the reader cannot check the most load-bearing step of the type-A classification. The authors should either prove the intersection statement systematically, for each family (a)–(e) determining exactly the index ranges for which the element lies in A(α̃, μ), or provide an exhaustive table or certificate for all index ranges.","section":"Corollaries 4.9–4.11"}],"minor_comments":[{"comment":"The two influence sets I(σ) and the extended influence set are easy to confuse throughout the text; Definition 3.5 and Example 3.6 are especially hard to parse. Use visually distinct symbols and always state explicitly which influence is meant when the word 'independent' is used.","section":"Definition 3.4"},{"comment":"The proof of Proposition 5.4 explicitly states that the enumerative arguments are omitted. Since this is a formal proposition, the omission should be filled, or the result should be marked as a conjecture or moved to an appendix with a complete proof.","section":"Proposition 5.4"},{"comment":"In the proof of Lemma 5.7 the bijection φ is described only by its images; a short verification that the five images are disjoint and exhaust the cokernel would make the recurrence checkable.","section":"Lemma 5.7"}],"recommendation":"major_revision","confidential_remarks":"The main theorem, Theorem 3.1, is solid and likely publishable on its own. However, the paper's principal new machinery, the BAS decomposition, has a false formulation of independence, and the type-A claims are downstream of it. The counterexample in A4 is small, and the fix (Dynkin distance at least two between influence sets) is apparent from Example 3.11, so I believe a careful revision can repair the paper; I therefore recommend major revision rather than rejection. The enumerative sections should be re-checked after the repair because their derivations depend on the BAS lists."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, Theorem 3.1 — that A(λ,μ) is an order ideal in the weak Bruhat order for dominant integral λ — is true and cleanly proved. The covering-relation argument works, and it is a genuine structural result. Second, the paper's central apparatus around \"basic allowable subwords\" has a load-bearing flaw. The definition of independence (disjoint influence sets) is too weak. In type A4, s1 and s2s3 are in BAS(α̃,-α̃) and have disjoint influences, but their product s1s2s3 is not in A(α̃,-α̃); direct computation gives σ(α̃+ρ)+α̃-ρ = 2α̃ - 4α1 - 2α2 - α3 with a negative α1 coefficient. So Theorem 3.10(2) — the bijection between A(λ,μ) and pairwise independent subsets of BAS — is false as stated. The proof's coefficient claim fails because adjacent factors can change coefficients of simple roots outside their own influence, and the assertion that independent elements commute is also false when influences are adjacent but disjoint.\n\nThis flaw propagates. Proposition 3.15, Proposition 3.17, and the type-A characterizations in Theorem 4.7 and Corollaries 4.9–4.11 all inherit it. The enumerative results in Section 5 count independent subsets of the proposed BAS sets, so they overcount if independent subsets do not all land in A(α̃,μ). The case-by-case verification in Appendix A may be right, but it is checking the wrong condition.\n\nWhat the paper does well: the order ideal theorem is solid and worth publishing on its own, and the BAS idea is promising. The type-A computations are explicit and suggest the correct lists may be salvageable with a stronger independence relation, likely involving extended influence. The generating functions are worked out carefully, conditional on the characterization.\n\nBottom line: this needs major revision, not acceptance. The advertised type-A results are not established. A referee should be sent, because the order ideal theorem deserves scrutiny and the BAS framework may be repairable. But I would not cite the type-A characterization in its current form, and I would not put it on a reading group list until the independence condition is fixed.\n\nRecommendation: send to peer review, but expect the referee to demand a corrected definition of independence and a re-proof of Theorem 3.10 before the type-A claims can be trusted.","headline":"The order ideal theorem is true and clean, but the BAS independence framework has a false step that undermines the type-A characterization and enumerations.","tokens_in":34362,"tokens_out":4965,"would_cite":false,"duration_ms":37595,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["17B10","05E15","17B20","20F55"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every Weyl alternation set is an order ideal in the weak Bruhat order of its Weyl group.","keywords":["Weyl alternation set","Kostant weight multiplicity formula","weak Bruhat order","order ideal","basic allowable subwords","Fibonacci numbers","type A Lie algebra","vector partition function"],"falsifier":"Compute A(α̃, μ) for a concrete unlisted case in type A, say r = 6 and μ = −α_{2,4}, by direct evaluation of the Kostant partition function on every Weyl-group element; if any element not of the forms in Corollary 4.11 contributes, or any listed element fails to contribute, the characterization is wrong. More broadly, for any simple Lie algebra, find a dominant integral λ and a weight μ for which some element below an element of A(λ, μ) in the weak Bruhat order has ℘(σ(λ+ρ)−μ−ρ) = 0; that would refute Theorem 3.1.","tokens_in":33311,"feed_emoji":"","tokens_out":6662,"duration_ms":52359,"temperature":0.7,"pith_summary":"This paper proves a structural constraint on Kostant's weight multiplicity formula: for any integral dominant weight λ and any weight μ, the set of Weyl group elements that contribute nonzero terms to the multiplicity—the Weyl alternation set—is an order ideal in both the left and right weak Bruhat orders. That means if a Weyl group element contributes, every element below it in the weak order contributes as well, so the support of the formula is always downward-closed. The paper then specializes to sl_{r+1}(C), fully characterizing the alternation sets when λ is the highest root and μ is a negative root, and provides generating functions for their sizes. The motivation is computational: knowing the support and its structure lets one evaluate Kostant's alternating sum over a fraction of the Weyl group.","feed_headline":"The support of Kostant's formula is always downward-closed","feed_subtitle":"New proof that contributing Weyl-group elements form order ideals, with full type-A enumeration for negative roots.","key_machinery":"The load-bearing objects are the weak Bruhat order on the Weyl group, defined by covering relations σ ⋖ σ s_i when length increases (right order) and σ ⋖ s_i σ (left order), and the order ideal property: with any contributor τ, all prefixes and suffixes in reduced expressions contribute too. The type-A analysis is organized around 'influence' I(σ), the set of simple-reflection indices appearing in a reduced word for σ, and its connected extension I(σ), together with the decomposition of A(λ, μ) into basic allowable subwords BAS(λ, μ), elements with connected influence that cannot be split into independent factors. Proposition 3.15 gives sufficient conditions for a candidate set S to equal BAS(λ, μ); Lemma 4.6 verifies those conditions case by case for μ = −α̃, and Corollaries 4.9–4.11 obtain other negative roots by intersecting with A(α̃, −α̃).","core_discovery":"On its own terms, the paper's central result is Theorem 3.1: for a simple Lie algebra g with Weyl group W, if λ is dominant integral and μ is any weight, then A(λ, μ) is an order ideal in the left and right weak Bruhat orders. The proof shows that when τ covers σ, the difference σ(λ+ρ) − τ(λ+ρ) is a nonnegative integer combination of simple roots; hence any vector partition of τ(λ+ρ)−μ−ρ can be extended to one for σ(λ+ρ)−μ−ρ. Theorem 3.10 sharpens this: every nonempty alternation set is the set of products of pairwise independent subsets of a unique collection BAS(λ, μ) of 'basic allowable subwords' with connected influence. In type A, for λ = α̃ and μ = −α_{i,j}, the paper lists BAS(α̃, μ) explicitly in Corollaries 4.9–4.11 and shows that |A_r(α̃, μ)| satisfies the Fibonacci recurrence in the rank, with a closed generating function over all negative roots.","pith_inferences":["The order-ideal theorem is not restricted to highest-root weights; if it survives beyond the adjoint representation, it gives a general pruning rule: any algorithm that walks the weak order can terminate at the maximal contributors instead of scanning all of W.","The Fibonacci-type growth suggests a broader phenomenon: for fixed λ, μ in type A, |A(λ, μ)| grows like c^r rather than r!, so Kostant's formula could in principle be evaluated in subfactorial time by enumerating ideals.","A direct testable extension is to produce analogous forbidden-subword lists for types B, C, and D and compare the resulting cardinalities to sequences such as the Lucas numbers the paper already recovers in Corollary 5.3.","If the q-analog conjecture is proved with these methods, the same support structure would also predict which powers of q appear for other classical types."],"forward_implications":["Corollary 3.2: any reduced word that appears as a consecutive subword of a reduced expression for a contributing element is itself a contributor, so membership in A(λ, μ) can be checked by downward closure.","Theorem 3.10: every element of A(λ, μ) corresponds uniquely to a pairwise independent subset of BAS(λ, μ), making the alternation set an abstract simplicial complex.","Corollaries 4.9–4.11: for sl_{r+1}(C), the sets A(α̃, −α_{i,j}) are fully enumerated, answering Harry's 2024 question for negative roots.","Proposition 5.5 and Theorem 5.14: the cardinalities satisfy |A_r| = |A_{r−1}| + |A_{r−2}| and are captured by an explicit bivariate generating function.","Section 6: these lists are the input for a planned proof of Harry's conjecture that m_q(α̃, μ) = q^{r+j−i+1} + q^{r+j−i} − q^{j−i+1}."],"supporting_citations":[{"why":"gives Kostant's weight multiplicity formula (1), the alternating sum whose support is studied.","marker":"[17]"},{"why":"defines the Weyl alternation set and computes A(α̃, 0), the Fibonacci base case that the paper extends.","marker":"[5]"},{"why":"characterizes A(α̃, μ) for positive roots and poses the question on negative roots that Corollaries 4.9–4.11 answer.","marker":"[14]"},{"why":"supplies the Lemma 4.1 root-action computations and the Proposition 3.15 criterion for identifying basic allowable subwords.","marker":"[10]"},{"why":"provides the length-versus-root positivity fact (Equation 4.25) that carries the proof of Theorem 3.1.","marker":"[1]"},{"why":"supplies the order-ideal and generating-function background used in the definitions and enumeration.","marker":"[19]"}],"fun_headline_variants":["Weyl alternation sets proven to be order ideals","Downward-closed: Kostant's support is an order ideal","Type A: full list of alternation sets for negative roots","Fibonacci recurrence for Kostant's alternation sets"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the case-by-case verification in Lemma 4.6 covers every possible pair of nonindependent basic allowable subwords and that the intersections used to derive Corollaries 4.9–4.11 were computed correctly; if any case or intersection is wrong, the type-A characterization and its enumerations fail.","fun_headline_variants_meta":{"raw":{"variants":["Weyl alternation sets proven to be order ideals","Downward-closed: Kostant's support is an order ideal","Type A: full list of alternation sets for negative roots","Fibonacci recurrence for Kostant's alternation sets"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000908,"raw_usage":{"total_tokens":3974,"prompt_tokens":1088,"completion_tokens":2886,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":704,"completion_tokens_details":{"reasoning_tokens":2818}},"tokens_in":704,"tokens_out":2886,"duration_ms":19232,"temperature":1.0,"reasoning_tokens":2818,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T10:15:25.546640+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute A(α̃, μ) for a concrete unlisted case in type A, say r = 6 and μ = −α_{2,4}, by direct evaluation of the Kostant partition function on every Weyl-group element; if any element not of the forms in Corollary 4.11 contributes, or any listed element fails to contribute, the characterization is wrong. More broadly, for any simple Lie algebra, find a dominant integral λ and a weight μ for which some element below an element of A(λ, μ) in the weak Bruhat order has ℘(σ(λ+ρ)−μ−ρ) = 0; that would refute Theorem 3.1.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives Kostant's weight multiplicity formula (1), the alternating sum whose support is studied."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defines the Weyl alternation set and computes A(α̃, 0), the Fibonacci base case that the paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"characterizes A(α̃, μ) for positive roots and poses the question on negative roots that Corollaries 4.9–4.11 answer."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the Lemma 4.1 root-action computations and the Proposition 3.15 criterion for identifying basic allowable subwords."},{"cited_title":"Bj¨ orner and F","cited_arxiv_id":null,"evidence_quote":"provides the length-versus-root positivity fact (Equation 4.25) that carries the proof of Theorem 3.1."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the order-ideal and generating-function background used in the definitions and enumeration."}],"review_version":1}