{"id":"da16115a-eebe-4e77-bac1-9a8fb1b58504","arxiv_id":"2501.16640","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every finite normal square root crystal has a character equal to a nonnegative integer combination of symmetric Grothendieck polynomials, indexed by its highest weight elements.","lead":"The paper proves that the character of any finite normal square root crystal is a sum of symmetric Grothendieck polynomials, the K-theoretic analogue of Schur polynomials. This gives combinatorists a new crystal-based tool for proving Grothendieck positivity, with a new proof of Buch's Littlewood-Richardson rule.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.5(c) is false: Hecke-inserting the top-left entry x=1 into the two-row increasing tableau [[1,3],[2]] gives [[1,2,3],[2]], not the tableau itself, and this lemma is used in Lemma 3.12 to prove Theorem 3.11.","rationale":"The reader's weakest_assumption identified Lemma 3.5 and Lemma 3.6 as load-bearing, and noted that Lemma 3.5's proof is omitted. Our examination goes further: Lemma 3.5(c) is not merely missing a proof; it is false under the paper's own insertion rule. The counterexample [[1,3],[2]] with x=1 is a minimal two-row increasing tableau satisfying all stated hypotheses, and the paper's algorithm changes it. This false lemma is used in the combined-form subcase of Lemma 3.12, which is on the critical path to Theorem 3.11 and hence to the main theorem. We therefore cannot accept the proof as written. The central claim might still be true—the false lemma may be overstrong and the specific instances in Lemma 3.12 might satisfy extra hypotheses—but the current manuscript does not establish it. We recommend REJECT rather than CONDITIONAL because the flaw is a demonstrated incorrect statement in a proof ingredient, not merely an omitted justification. If the authors repair Lemma 3.5 or show that its false case never occurs in the needed context, a conditional acceptance could be reconsidered. Agreement with the reader is partial: they correctly located Lemma 3.5 as weak, but did not identify that the lemma is false.","tokens_in":36631,"tokens_out":38468,"duration_ms":325099,"concrete_test":"Implement the paper's Hecke insertion from Definition 3.1 and compute (1 -> T) for T = [[1,3],[2]]. The algorithm yields [[1,2,3],[2]], directly contradicting Lemma 3.5(c). To determine whether the main proof can be repaired, additionally test whether the specific tableau tab(E1(S1)) arising in Lemma 3.12 Case 3 (combined form) can equal [[1,3],[2]] for some valid input S; if such an S exists, the proof step fails as written. If no such S can occur, the false overgeneralization of Lemma 3.5(c) may be avoidable, and the verdict could be revised to CONDITIONAL after the lemma is corrected.","verdict_should_be":"REJECT","load_bearing_attack":"The paper's proof of Theorem 1.1 depends on Theorem 3.11 (P_Hecke(S)=tab(rect(S))), which is proved via Lemma 3.13, whose proof uses Lemma 3.12. Lemma 3.12's proof invokes Lemma 3.5 in two places, including the combined-form subcase of Case 3. Lemma 3.5 is stated as a 'basic exercise' with no proof, but it is not merely unproved—it is incorrect under the paper's own Definition 3.1. Concretely, let T be the two-row increasing tableau with rows [1,3] and [2]. Then T is not a rectangle, and for its unique second-row box (2,1) we have T_{2,1}=2 >= T_{1,1}+1=2, so Lemma 3.5(c) applies and predicts (1 -> T)=T. Running the paper's column-insertion algorithm: inserting 1 into the first column leaves the column unchanged (replacing 2 by 1 would violate strict increase) and bumps 2; inserting 2 into the second column replaces 3 (valid, since [[1,2],[2]] is increasing) and bumps 3; inserting 3 into the empty third column appends it, giving [[1,2,3],[2]] != T. Thus Lemma 3.5(c) is false. Since Lemma 3.5(c) is exactly what the proof of Lemma 3.12 uses to show that (m+1-j -> tab(E1(S1))) = tab(E1(S1)) in the combined-form subcase, the proof of Lemma 3.12, and therefore Theorem 3.11 and Theorem 1.1, has a concrete gap. The main theorem may still be true, but the written proof is invalid at a load-bearing technical step.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies finite normal square root crystals for gl_n and proves that the character of any such crystal is a sum of symmetric Grothendieck polynomials indexed by its highest weight elements (Theorem 1.1). The proof develops a rectification operator for square root crystals, connects it to the Hecke insertion algorithm of Buch, Kresch, Shimozono, Tamvakis, and Yong, and derives a bijection between full subcrystals and pairs of tableaux (P,Q). The introduction presents several applications: a new proof of Buch's combinatorial rule for skew symmetric Grothendieck polynomials, a K-theoretic Littlewood-Richardson rule, G-positivity for permutation Grothendieck polynomials, and a new positivity statement for set-valued decomposition tableaux.","tokens_in":36975,"tokens_out":6759,"duration_ms":67671,"significance":"If Theorem 1.1 is correct, it is a substantial contribution: it resolves a conjecture of Marberg and Tong and provides a uniform combinatorial mechanism for proving Grothendieck positivity, analogous to the role of Stembridge crystals for Schur positivity. The connection between square root crystals and Hecke insertion is elegant, and the applications in Section 1 are natural and illustrate the power of the framework. The paper is generally well structured and the reliance on prior work is transparent. However, the written proof has a load-bearing gap in a technical lemma, so the main theorem is not established by the current text.","major_comments":[{"comment":"Lemma 3.5(c) is false under Definition 3.1. Take T = [[1,3],[2]] (English notation: first row 1,3; second row 2) and x = 1. Then T is an increasing tableau, it is not a rectangle, and for the only second-row box we have T_{2,1} = 2 >= T_{1,1}+1 = 2, so the hypotheses of (c) hold. Following Definition 3.1, inserting 1 into the first column finds y = 2 and bumps it, since replacing 2 by 1 would violate strict increase. Inserting the bumped value 2 into the second column replaces 3, which is valid because the resulting tableau is still increasing. The bumped 3 is then appended to a new third column. The final tableau is [[1,2,3],[2]], not T. Thus the lemma is false as stated.","section":"§3.1, Lemma 3.5(c)"},{"comment":"The false clause (c) is load-bearing. In the combined-form subcase of Case 3, the proof reduces the desired equality to the assertion that Hecke-inserting m+1-j into tab(E1(S1)) leaves the tableau unchanged, and it states that this follows from Lemmas 2.20(b) and 3.5(c). The counterexample above satisfies exactly the same hypotheses used there, so this step is not justified. Since Lemma 3.12 is used to prove Theorem 3.11, and Theorem 3.11 is the bridge between rectification and Hecke insertion used in Theorem 3.14 and then Theorem 1.1, the main proof has a concrete gap. The main theorem may still be true, but the written derivation does not establish it.","section":"§3.2, Lemma 3.12, Case 3"},{"comment":"Lemma 3.5 is stated without proof and described only as following from the definitions as a basic exercise. It is not a basic exercise, and in its current form it is incorrect. Because the lemma is invoked in the proof of Lemma 3.12, it needs either a correct proof or a repaired statement that still supports the combined-form argument.","section":"§3.1, Lemma 3.5"}],"minor_comments":[{"comment":"Example 1.14 defines k = max{1, m-n+1}, while the caption of Figure 3 gives k = max{1, m-n-1}; these should be reconciled.","section":"Example 1.14 / Figure 3"},{"comment":"The proof of Lemma 3.6 is a long case analysis and is hard to check; a precise statement of the induction invariant, or at least a short overview of the induction, would substantially improve readability.","section":"§3.1, Lemma 3.6"},{"comment":"The wording 'bumps y' in the case where y is not actually replaced is potentially confusing; a sentence clarifying that the bumped value is passed to the next column regardless of whether the replacement was valid would help.","section":"Definition 3.1"},{"comment":"The bijectivity of the map S -> (A(S), I(S)) is labeled a straightforward exercise; a brief proof sketch would make the paper more self-contained.","section":"Proposition 3.10"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Main thing to know: this is a substantial paper with a real result in reach, but there is a load-bearing error in the proof. Theorem 1.1—that finite normal sqrt(gl_n)-crystals have Grothendieck-positive characters—is exactly the kind of structural bridge people have wanted since [MT23], and the applications (Buch's rule, G-positivity of Gw, Corollary 1.13) make it worth reading. The writing is clear and the overall architecture is sensible: rectify to highest weight elements, then identify those with Hecke insertion tableaux.\n\nThe soft spot is not a stylistic issue. Lemma 3.5(c) is false. Let T = [[1,3],[2]] and x = 1. T is a two-row increasing tableau, not a rectangle, and T_{2,1} = 2 >= T_{1,1}+1 = 2. The lemma predicts (1 H-> T) = T. Running Definition 3.1: inserting 1 into column 1 leaves the column unchanged and bumps 2; inserting 2 into column 2 replaces 3 and bumps 3; inserting 3 into the empty column 3 appends it. The result is [[1,2,3],[2]] != T. This false statement is not decorative: Lemma 3.12, Case 3, uses Lemma 3.5(c) to conclude that Hecke inserting the top-left entry fixes tab(E1(S1)) in the combined-form subcase. Lemma 3.12 feeds Lemma 3.13, Theorem 3.11, and ultimately Theorem 1.1. So the written proof has a concrete gap at a load-bearing step. The theorem may still be true—the applications are consistent and external results support the framework—but the derivation as written is not valid there.\n\nOther soft spots are minor by comparison: Lemma 3.5 is stated as a 'basic exercise' without proof, which turned out to be a bad bet; Lemma 3.6 is a long case analysis and not machine-checked; Example 1.14's bijection is shown by example rather than formally constructed. None of these matter as much as the false lemma.\n\nWho this is for: anyone working in K-theoretic Schubert calculus or crystal theory. The framework is novel and the intended applications are the right ones. But I would not trust the main theorem until Lemma 3.5/3.12 is repaired. I would send the paper to a serious referee—there is enough substance that the error may be fixable and the paper deserves a full report—but my own verdict on the current version is skeptical.","headline":"Strong and interesting framework, but the proof of the central positivity theorem relies on a Hecke-insertion lemma that is false as stated.","tokens_in":37601,"tokens_out":8205,"would_cite":false,"duration_ms":80695,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05E05","05E10","17B37"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every finite normal square root crystal has a character that is a sum of symmetric Grothendieck polynomials.","keywords":["normal square root crystals","Grothendieck polynomials","Hecke insertion","Stembridge crystals","rectification","set-valued tableaux","crystal character","K-theoretic combinatorics"],"falsifier":"Test Lemma 3.6 by hand on a three-row increasing tableau with a decreasing set of inserted numbers; for example, take P = 1 2 5 / 3 4 6 / 7 and B = {6,4,2}, and compare direct Hecke insertion with the lemma's reduction to last-row insertion.","tokens_in":36357,"feed_emoji":"💎","tokens_out":9821,"duration_ms":80748,"temperature":0.7,"pith_summary":"Normal square root crystals are a K-theoretic analogue of Stembridge crystals: their characters are meant to expand positively in the basis of symmetric Grothendieck polynomials instead of Schur polynomials. The paper proves that this is always true: for every finite normal square root crystal, the character equals the sum of symmetric Grothendieck polynomials attached to its highest weight elements. This settles a conjecture from the paper that introduced these crystals and gives a uniform mechanism for proving Grothendieck positivity of symmetric functions, just as normal crystals give Schur positivity. The proof works by showing that the rectification operator on such crystals is realized by Hecke insertion.","feed_headline":"Characters of normal square root crystals are Grothendieck sums","feed_subtitle":"A new proof shows every finite normal square root crystal's character decomposes into symmetric Grothendieck polynomials.","key_machinery":"The central object is the rectification operator rect, defined as the composition of iterated raising operators applied in a specific order, together with its realization by Hecke insertion. For each set-valued word S, Theorem 3.11 proves that the Hecke-insertion recording tableau P_Hecke(S) equals tab(rect(S)), where tab rewrites a set-valued word as a tableau. The load-bearing technical bridge is Lemma 3.6, which asserts that Hecke insertion of a decreasing set of numbers into a multirow increasing tableau reduces to inserting them into the last row and then reinserting the bumped entries into the remaining rows; this reduction is what lets the proof compare the crystal-raising process with the insertion algorithm dimension by dimension.","core_discovery":"The paper's central claim is Theorem 1.1: if B is a finite normal sqrt(gln)-crystal — normal meaning each connected component is isomorphic to a full subcrystal of a tensor power of the standard sqrt(gln)-crystal on nonempty subsets of {1,...,n} — then ch(B) = sum_{b in HW(B)} G_{wt(b)}(x_1,...,x_n). In words, the character of any finite normal square root crystal is a sum of symmetric Grothendieck polynomials, one per highest weight element. This is the exact K-theoretic analogue of the classical fact that characters of normal gln-crystals are Schur-positive. The proof establishes a bijection between elements of a normal square root crystal and pairs (P,Q) consisting of an increasing tableau P indexed by a highest weight element and a semistandard set-valued tableau Q of the same shape, mediated by Hecke insertion.","pith_inferences":["The rectification-to-Hecke-insertion correspondence suggests a broader template: any crystal family whose raising operators can be simulated by a tableau insertion algorithm should have characters that expand positively in the associated Grothendieck-type basis; testing this on other K-theoretic crystals would show how general the phenomenon is.","The paper leaves open the search for local Stembridge-style axioms that characterize normal square root crystals; the main theorem makes such axioms more valuable, since they would give an effective way to recognize when a generating function is Grothendieck-positive.","The authors verified a Lascoux-positivity conjecture for square root Demazure crystals by computer for all m,n <= 5 except (5,5); a natural next test is to check that remaining case and to see whether the same Hecke-insertion technology can prove it uniformly."],"forward_implications":["Theorem 1.1 yields a new proof of the Littlewood–Richardson rule for multiplying symmetric Grothendieck polynomials, by applying the theorem to the normal square root crystal on set-valued tableaux of skew shape.","The Grothendieck polynomial of a permutation is Grothendieck-positive, with coefficients counted by increasing tableaux whose reverse row reading word is a Hecke word for the permutation (Corollary 1.11).","The generating function for set-valued decomposition tableaux of a strict partition is Grothendieck-positive (Corollary 1.13).","Every homogeneous piece of the character of a normal square root crystal is Schur positive (Corollary 3.15), so square root crystals refine ordinary Schur positivity.","The rectification operator sends every element of a normal square root crystal to a highest weight element (Theorem 2.21), the fact that makes the highest-weight summation formula possible."],"supporting_citations":[{"why":"Introduced normal square root crystals and stated the conjecture that the main theorem resolves.","marker":"[MT23]"},{"why":"Supplied the set-valued word model and explicit crystal operators that the proof works with.","marker":"[Yu23]"},{"why":"Defined the Hecke insertion algorithm, which the paper proves is realized by crystal raising operators.","marker":"[BKS`08]"},{"why":"Standard reference for normal gln-crystals, their characters, and Stembridge axioms, providing the classical Schur-positivity formula being mimicked.","marker":"[BS17]"},{"why":"Defined symmetric Grothendieck polynomials and the Littlewood–Richardson rule whose recovery is an application of the main theorem.","marker":"[Buc02]"}],"fun_headline_variants":["Square root crystals have Grothendieck characters","Normal square root crystals yield Grothendieck sums","Grothendieck positivity for square root crystals proven","Characters of square root crystals are Grothendieck sums","Hecke insertion proves Grothendieck character formula"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is Lemma 3.6, which asserts that Hecke insertion of a decreasing set into a multirow increasing tableau reduces to insertion into the last row followed by reinsertion into the remaining rows; if this reduction fails, the paper's proof of the main theorem collapses.","fun_headline_variants_meta":{"raw":{"variants":["Square root crystals have Grothendieck characters","Normal square root crystals yield Grothendieck sums","Grothendieck positivity for square root crystals proven","Characters of square root crystals are Grothendieck sums","Hecke insertion proves Grothendieck character formula"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000219,"raw_usage":{"total_tokens":1418,"prompt_tokens":895,"completion_tokens":523,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":447}},"tokens_in":511,"tokens_out":523,"duration_ms":5106,"temperature":1.0,"reasoning_tokens":447,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T11:43:41.534103+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Test Lemma 3.6 by hand on a three-row increasing tableau with a decreasing set of inserted numbers; for example, take P = 1 2 5 / 3 4 6 / 7 and B = {6,4,2}, and compare direct Hecke insertion with the lemma's reduction to last-row insertion.","supporting_citations":[],"review_version":1}