{"id":"da4212a0-037c-4d32-8cb2-6f291d544180","arxiv_id":"2608.09619","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For s_λ(μ_t, z, z^{-1}), the evaluation is zero or a signed product of three hyperbolic sine factors read from the t-residue profile, for every t and every shape.","lead":"This paper evaluates Schur polynomials on an alphabet made of all t-th roots of unity plus one free reciprocal pair (z, z^{-1}), proving a compact product formula. The value sees only three integer distances and a sign, so arbitrarily large partitions sharing those data give identical evaluations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The factorization formula is proved, but the abstract's 'nothing else' minimality claim is only checked for t≤6, |λ|≤14; the proof does not rule out distinct triples with equal values.","rationale":"The reader's weakest assumption identifies exactly the same load-bearing concern: the abstract's strong minimality statement is only numerically checked, not proved, while the main product formula is proved for all t. My reading of the proof confirms this. I do not see a substantive gap in Theorem 3.1 itself: the Laplace expansion, Lemma 4.1, Lemmas 4.4 and 4.5, and the sign computation in Proposition 3.10 are internally coherent, and the extensive numerical checks give independent support. The zero-locus converse in Section 8 is explicitly labelled conjectural outside the proved ranges, so it is not a hidden defect. The only point at which an advertised claim outruns the proof is the 'nothing else' compression. Since the paper itself flags this as a check rather than a proof, the CONDITIONAL verdict is appropriate: the core evaluation is proved, but the headline information-theoretic claim should either be proved or narrowed. No change to the reader's verdict is therefore needed.","tokens_in":41009,"tokens_out":18022,"duration_ms":160825,"concrete_test":"For t=2, enumerate all interval triples (d1,d2,d3) arising from partitions with |λ| up to, say, 40; compute the reduced Laurent polynomial Φ_2 from equation (5) together with ε_λ, and test whether two distinct multisets {d1,d2,d3} with the same sign produce identical polynomials. If any collision appears, the minimality claim is false and the abstract must be weakened. To go beyond search, factor each u^{2d_i}−1 into cyclotomic polynomials and determine whether the map from a multiset of three d_i to the multiset of cyclotomic factors is injective on the attainable triples; an injectivity proof would replace the finite check.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 3.1, with the closed form in equation (5), is proved by a Laplace expansion and checked over 10,959 exact cases; the sign and the three-factor structure are on solid ground. The load-bearing concern is a different claim, made in the abstract: that Φ_t(λ;z) sees 'a multiset of three integers and a sign, and nothing else.' The theorem proves that Φ_t factors through the ordered invariant I_t = (d1,d2,d3,ε_λ) defined in (7). Reducing the ordered triple to the multiset is immediate from the symmetry of the numerator in (5). But the minimality direction — that two different multisets cannot give the same value — is not proved. The paper says after Proposition 3.10 that the orientation/minimality part 'remains a check rather than a proof,' and the verification table in Section 9 reports no collisions only for t≤6 and |λ|≤14. This is not a cosmetic gap: each factor u^{d_i} − u^{−d_i} in the numerator is reducible, since u^d − u^{−d} = u^{−d} ∏_{m|2d} Φ_m(u). Hence additive identities among divisor multisets could in principle make two distinct triples produce the same rational function, and Theorem 3.1 does nothing to exclude that possibility. The phrase 'and nothing else' is the paper's headline compression claim; its truth is not settled by the proof of the evaluation formula.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the Schur polynomial s_λ(1, ζ, ..., ζ^{t-1}, z, z^{-1}) for t ≥ 2 and arbitrary partitions λ with at most t+2 parts. Theorem 3.1 gives a closed form: the value is zero if some residue class modulo t is empty, and otherwise equals a signed product of three hyperbolic sine factors divided by sinh^2(tθ/2) sinh θ, with the three integers d_1, d_2, d_3 and the sign read off the beta set of λ. The proof is a Laplace expansion along the t frozen rows, with explicit lemmas for the Vandermonde minors, the column move, and the cancellation identities. The paper derives several consequences: a geometric vanishing criterion, an extension of an independence criterion of Ayyer–Kumari for two-row shapes, a signed enumeration of plane partitions at t = 2 refined by a free parameter, a discussion of four deformations that destroy the factorization, and a partly conjectural determination of the zero locus of Ψ_r = s_λ(1, -1, z_1^{±1}, ..., z_r^{±1}). The main evaluation theorem is supported by 10,959 exact numerical checks, and all computational claims are documented in an ancillary archive.","tokens_in":41262,"tokens_out":12477,"duration_ms":111027,"significance":"If the main claims hold, Theorem 3.1 is a clean and remarkably complete evaluation for this mixed alphabet, with an explicit sign and no hypothesis on the shape of λ. The proof is detailed and appears sound: it gives a genuine Laplace-expansion argument, explicit lemmas, and a derived (not fitted) sign formula. The paper also provides reproducible code and archived outputs, and it carefully labels conjectural versus proved statements, which is exemplary. The zero-locus section is more speculative but honestly framed. However, the headline compression claim that the value sees 'a multiset of three integers and a sign, and nothing else' is not proved in the printed text; it is verified only for finite ranges. Since this minimality claim is central to the abstract's 'determine exactly how much' assertion, the paper needs either a proof of that claim or a corresponding weakening of the claim.","major_comments":[{"comment":"The assertion that Φ_t(λ;z) depends on a multiset of three integers and a sign 'and nothing else' is not established by Theorem 3.1. Theorem 3.1 proves that the value factors through the ordered invariant I_t = (d_1, d_2, d_3, ε_λ) defined in (7); symmetry of the numerator in (5) then reduces the ordered triple to a multiset. But the converse direction — that two distinct multisets cannot produce the same rational function — is only checked computationally: §3.4 reports the absence of collisions only over |λ| ≤ 20 for t = 3, 4, and the verification table in §9 contains no row proving minimality for all t. This is not a cosmetic gap, because each factor u^{d_i} − u^{−d_i} in the numerator is reducible (u^d − u^{−d} = u^{−d} ∏_{m|2d} Φ_m(u)), so additive identities among divisor multisets could in principle make distinct triples coincide. The paper should either prove the minimality claim, for example by a cyclotomic-factor argument in the spirit of Lemma 5.1, or explicitly weaken the abstract and §3.4 to the factorization statement that is actually proved.","section":"Abstract; §3.4; §9"},{"comment":"The claim that the converse of Theorem 8.1 holds for every r when |λ| ≤ 2r + 2 outside Littlewood's range is asserted in a single sentence: 'there Littlewood's rule applies verbatim and the converse follows by the argument of Theorem 8.4.' That is not demonstrated. The proof of Theorem 8.4 uses ℓ(λ) ≤ N/2 in several essential places: it uses the β' = ∅ term to get m_μ ≥ 1 for every μ ⊆ λ, it uses ℓ(μ*) > ℓ(λ) when μ = λ, and it uses horizontal-strip constructions that depend on the row structure. When ℓ(λ) > N/2, none of these steps is automatic, and the text gives no witness construction for the unstable band. Since the abstract explicitly claims this converse as proved, the argument needs to be supplied, or the claim should be moved to the conjectural part of Section 8.","section":"§8.4, sentence after Lemma 8.9"}],"minor_comments":[{"comment":"In the displayed identity (21), the exponent s in (-1)^s is undefined. From the preceding expression it should be the parity of \\binom{N}{2} + r (or the sum itself), and this should be stated explicitly.","section":"§8.2, Eq. (21)"},{"comment":"The row labelled 'Theorem 8.1, both directions' is misleading: Theorem 8.1 as stated contains only the sufficient direction for all r, while the converse for all r is Conjecture 8.6. The numerical verification of the converse over r ≤ 3 should be labelled as a check of the conjecture, not of the theorem.","section":"§9, verification table"},{"comment":"The parity computation in the proof of Lemma 4.4 is hard to audit as printed, especially the step that obtains κ_{1j}/κ_{2j} = -1 from the exponent '2j_{A2} − 1' and the indicator [a_2 < b_j < a_1]. Please expand that display into explicit parity bookkeeping, since the lemma is load-bearing for the sign in Theorem 3.1.","section":"§4, proof of Lemma 4.4"}],"recommendation":"major_revision","confidential_remarks":"The core evaluation theorem and its proof are strong, and the numerical verification is unusually thorough. The main obstacle to accepting the paper's headline claim is the unproved minimality statement in the abstract and §3.4; that is fixable either by proving it or by carefully restating the claim. The Section 8.4 converse for |λ| ≤ 2r+2 outside Littlewood's range is also asserted too quickly. If these two points are addressed, I would support publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main theorem is real, and the paper deserves referee time. The evaluation of s_lambda at mu_t union {z, z^{-1}} with the three-factor closed form and explicit sign is genuinely new, and the proof is complete: Laplace expansion, two lemmas, the cancellation identity, all spelled out. The numerical checks are extensive (10,959 exact cases) and the scripts are provided. The paper is also honest about what it has not proved: the zero-locus converse is labelled a conjecture, and the failed certificate is stated plainly.\n\nThe soft spot is the abstract's compression claim. 'A multiset of three integers and a sign, and nothing else' is presented as a settled result, but the minimality direction -- that two different multisets cannot give the same value -- is not proved. The theorem proves the value factors through the ordered triple; symmetry of the numerator gives the multiset reduction; but nothing rules out collisions coming from factorizations of u^d - u^{-d} into cyclotomic factors. The paper itself says after Proposition 3.10 that the orientation/minimality part 'remains a check rather than a proof,' and the verification covers only t <= 6, |lambda| <= 14. That is a genuine gap between the abstract and the results. It does not affect the evaluation formula itself, which stands fully proved.\n\nAlso worth noting: the zero-locus converse (Conjecture 8.6) is not proved, and the proposed certificate fails; that is clearly labelled, so it is a soft spot only in the sense that the abstract's phrasing could be read as stronger. The abstract does say 'The rest is conjectural,' so that part is fair.\n\nWho gets value: anyone working on Schur evaluations, factorization theorems, or core-quotient structure. The t = 2 enumerative reading is a nice byproduct. I would send this to a serious referee; the main theorem is a real result and the proof is checkable. The revision should either prove the minimality claim or soften the abstract to say the value depends on the multiset and sign (which is what is proved), not that it depends on nothing else.","headline":"Solid new evaluation theorem with a complete proof; the abstract's 'nothing else' minimality claim outruns what is proved and should be fixed in revision.","tokens_in":41827,"tokens_out":2276,"would_cite":true,"duration_ms":23052,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05E05","05A15","05E10","20G05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Schur polynomials at roots of unity plus a reciprocal pair evaluate to a signed product of three hyperbolic factors, so every partition is compressed to three integers and a sign.","keywords":["Schur polynomials","roots of unity","reciprocal variables","factorization","cores and quotients","beta sets","signed enumeration"],"falsifier":"For $t=7$, enumerate all partitions of size at most $20$ and compare the exact bialternant with the right-hand side of (5), then search for pairs with the same multiset $\\{d_1,d_2,d_3\\}$ and same sign whose exact values differ; either search can falsify the paper's claims at the first counterexample.","tokens_in":40765,"feed_emoji":"🧮","tokens_out":11881,"duration_ms":84588,"temperature":0.7,"pith_summary":"The paper establishes that the Schur polynomial of any partition $\\lambda$, evaluated at the alphabet $\\{1,\\zeta,\\ldots,\\zeta^{t-1},z,z^{-1}\\}$ consisting of all $t$-th roots of unity plus one free reciprocal pair, is either zero or a fixed product of three hyperbolic sine factors divided by a fixed denominator. The three arguments and the overall sign are read directly from the residue profile of $\\lambda$'s $\\beta$ set, so the value depends on a multiset of three integers and a sign, and on nothing else. This means partitions of unrelated sizes can share one value, and the paper writes down the infinite fibers of this collapse explicitly. A vanishing criterion follows: the value is zero exactly when a residue class modulo $t$ is empty, or (only for even $t$) when the two distinguished $\\beta$-intervals are concentric. The result also yields an extension of the independence criterion for two-row shapes and, at $t=2$, a $(-1)$-enumeration of plane partitions refined by a free parameter.","feed_headline":"Schur values collapse to three integers and a sign","feed_subtitle":"One free variable survives the signed count; two-row shapes gain exactly one extra family.","key_machinery":"The proof runs through the bialternant formula for Schur polynomials. Expanding the numerator by Laplace along the $t$ 'frozen' rows that carry the root-of-unity variables leaves one term for each choice of two un-frozen columns; the fact that there are $t+2$ columns in total and $t$ residue classes forces a trichotomy of residue profiles, and a single cancellation lemma in the symmetric group collapses the surviving terms into the three-factor product. The geometric reading used throughout identifies $d_1,d_2$ as the lengths of two intervals on the beta line and $d_3$ as twice the distance between their centres; this 'interval triple' is what the evaluation sees, and its symmetry is the reason the value records a multiset rather than an ordered triple.","core_discovery":"The central claim is Theorem 3.1: for $t\\ge 2$ and $\\lambda$ a partition with at most $t+2$ parts, with $\\beta(\\lambda)$ the shifted $\\beta$ set and $n_i(\\lambda)$ the counts of $\\beta$ parts congruent to $i$ modulo $t$, the value $\\Phi_t(\\lambda;z)=s_\\lambda(1,\\zeta,\\ldots,\\zeta^{t-1},z,z^{-1})$ vanishes if any $n_i=0$; otherwise the residue profile is either 'two-class' (two residue classes each contribute two $\\beta$ parts) or 'size-three' (one class contributes three), and in both cases one has $$\\Phi_t(\\$\\lambda$;z)=\\varepsilon_\\$\\lambda$\\,\\frac{\\$\\sinh$(d_1\\$\\theta$/2)\\$\\sinh$(d_2\\$\\theta$/2)\\$\\sinh$(d_3\\$\\theta$/2)}{\\$\\sinh$^2(t\\$\\theta$/2)\\$\\sinh$\\$\\theta$},\\quad z=e^\\$\\theta$,$$ where $d_1,d_2$ are the gaps inside the two distinguished classes, $d_3$ is the distance between their sums, and $\\varepsilon_\\lambda=\\pm1$ is an explicit sorting sign. The paper proves this for every $\\lambda$ with no hypothesis on its shape, and shows that the formula is symmetric in $d_1,d_2,d_3$, so the actual invariant is the multiset $\\{d_1,d_2,d_3\\}$ plus the sign. Since the right-hand side is a Laurent polynomial in $z$, the theorem also gives, via $\\mathfrak{sl}_2$ characters, a uniform character-ratio form with no exponentials.","pith_inferences":["If the compression holds for all $t$, the same evaluation invariant may govern other specializations that adjoin exactly one free direction to a full orbit, suggesting a general rank-one evaluation theorem for characters of classical type.","The counting argument that three factors are forced by translation invariance of an interval pair indicates that any alphabet with excess two, for instance in flagged or skew settings, should exhibit a three-factor product independently of the Laplace-expansion proof.","The $t=2$ signed enumeration with a free parameter invites a cyclic-sieving refinement: a $q$-analogue of $\\Phi_t$ with a cyclic action whose fixed points are counted by the refined signed count; the paper leaves this as an open problem.","The determinant dichotomy of Section 8, where alphabet determinant $+1$ gives factorization on self-complementary shapes while determinant $-1$ gives vanishing, may transfer to other groups and other order-two fixed letters, providing a test for universal-character analogues."],"forward_implications":["For any $t\\ge2$, computing $s_\\lambda$ at this alphabet reduces to reading three integers and a sign off the $t$-residue profile; no expansion of $\\lambda$ is needed.","The value vanishes exactly when a residue class modulo $t$ is empty, or when the two distinguished intervals are concentric (the latter possible only for even $t$).","For two-row shapes, the value equals $\\pm s_\\lambda(z,z^{-1})$ not only on the $t$-cores but on one additional family classified by its core and quotient; this is the full correction to the independence criterion on the reciprocal locus.","At $t=2$, the theorem gives a product formula for a $(-1)$-weighted count of plane partitions in a box, refined by a free parameter $z$; at $z=1$ it recovers signed counts such as $(c/2+1)^2$ for $2\\times2\\times c$ boxes with $c$ even and $0$ for $c$ odd.","The factorization is isolated: adding a second reciprocal pair, enlarging the root-of-unity orbit, replacing the orbit by a coset, or replacing the reciprocal pair by a free pair all destroy the product; only the zero locus survives for arbitrary numbers of pairs, with two explicit conditions that are proved sufficient and, in the proved ranges, necessary."],"supporting_citations":[{"why":"Supplies the classical evaluation of Schur polynomials at all t-th roots of unity, which the new formula restricts to when the reciprocal pair is removed.","marker":"[LR34]"},{"why":"Provides the residue-count notation n_i and the root-of-unity factorization setting the present alphabet extends.","marker":"[AK22]"},{"why":"Gives the independence criterion and the sorting permutation that the paper extends and refines on the reciprocal locus.","marker":"[AK25]"},{"why":"Establishes the core–quotient correspondence used to locate the vanishing conditions in the core versus the quotient.","marker":"[GKS90]"},{"why":"Gives the skew form of Littlewood's theorem with ribbon signs, on which the signed enumeration section relies.","marker":"[Mac95]"},{"why":"Identifies the self-complementary shapes whose factorization under a determinant +1 alphabet becomes vanishing under the determinant −1 alphabet of Section 8.","marker":"[AB19]"}],"fun_headline_variants":["Schur at roots of unity: a triple product or zero","Three numbers plus a sign decide Schur values","Reciprocal twist: Schur becomes a signed triple","No shape hypothesis: Schur product of three factors","Vanishing and a triple product for twisted Schur"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the evaluation invariant (multiset of three integers plus sign) captures all the information in the value for every $t$; the formula is proved for all $t$, but the 'nothing else' part is only checked for $t\\le6$, so the compression claim for larger $t$ rests on an unproved minimality statement.","fun_headline_variants_meta":{"raw":{"variants":["Schur at roots of unity: a triple product or zero","Three numbers plus a sign decide Schur values","Reciprocal twist: Schur becomes a signed triple","No shape hypothesis: Schur product of three factors","Vanishing and a triple product for twisted Schur"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00078,"raw_usage":{"total_tokens":3624,"prompt_tokens":1303,"completion_tokens":2321,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":919,"completion_tokens_details":{"reasoning_tokens":2243}},"tokens_in":919,"tokens_out":2321,"duration_ms":15228,"temperature":1.0,"reasoning_tokens":2243,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T13:59:43.837777+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For $t=7$, enumerate all partitions of size at most $20$ and compare the exact bialternant with the right-hand side of (5), then search for pairs with the same multiset $\\{d_1,d_2,d_3\\}$ and same sign whose exact values differ; either search can falsify the paper's claims at the first counterexample.","supporting_citations":[],"review_version":1}