{"id":"c2db05fd-2904-43f7-b10b-a3db0528ffdf","arxiv_id":"2608.02303","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For every n≥2, the Kazhdan–Lusztig and Z-polynomials of thagomizer matroids and graphic matroids of K_{2,n} have distinct negative real roots.","lead":"Mathematicians proved that the Kazhdan–Lusztig polynomials of thagomizer matroids and of the complete bipartite graph K_{2,n} have only real negative roots, along with their Z-polynomials. The result gives a one-parameter family of real-rooted polynomials and settles a strengthening of prior log-concavity results.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; the only soft spot is the cited K_{2,n} transfer identities, which are standard and worth a spot-check.","rationale":"The reader's weakest-assumption identification is accurate: the only place where the paper's conclusions about K_{2,n} are not self-contained is the two imported identities. However, I do not regard this as a significant objection. The identities are cited to published, peer-reviewed sources, and the internal proofs of the thagomizer cases are complete and convincing. I checked the main technical steps: the root-of-unity lower bound B_n(q) ≥ 2^n, the alternating-sign evaluations (11), the Chebyshev interlacing and zero-transfer argument, and the unit-circle criterion with the coefficient-sign lemma. No internal inconsistency or unstated assumption emerged. The concrete test proposed would settle the dependency question empirically for small n, but I would not make acceptance conditional on it without some independent reason to doubt the cited identities. Hence the reader's ACCEPT verdict should stand unchanged.","tokens_in":11613,"tokens_out":29100,"duration_ms":198339,"concrete_test":"Independently compute P_{K_{2,n}} for n = 2, 3, 4, 5, 6 using a direct matroid-KL implementation (for graphic matroids, compute the Kazhdan–Lusztig polynomial via the defining hyperplane arrangement or deletion–contraction on flats) and compare with P_n(x) + x, where P_n is obtained from the Catalan generating function (1). Similarly, compute Z_{K_{2,n}} from the flat-sum definition Z_M(x) = Σ_F x^{rk F} P_{M/F}(x) for the same small n and compare with Z_{T_n}(x) from formula (22). If all entries match, the external identities are confirmed; any mismatch would localize the failure precisely to those two imported results.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Reading the proof in good faith, the internal arguments for Theorems 1.1 and 3.5 are sound. The Chebyshev interlacing step (Step 3 of Theorem 1.1) correctly converts alternating signs at n+1 cosine nodes into n simple real zeros of the degree-n normalization, and the transfer back to the negative real axis preserves simplicity. The Lakatos–Losonczi application to D_n is legitimate: D_n is self-inversive, its constant and leading coefficients equal ρ_n, and the strict inequality ρ_n > (1/2)Σ δ_{n,r} follows from evaluating at q=1. The coefficient estimates in Lemmas 2.2 and 3.3 are carefully proven with Lagrange inversion, the von Szily identity, and an induction with nonnegative coefficients; I found no algebraic gap. The one part of the central claim not re-derived in this paper is the bridge from T_n to K_{2,n}: Theorem 1.2 for K_{2,n} uses [8, Thm 5.8] (P_{K_{2,n}} = P_n + x), and Theorem 1.3 uses [5, Prop 5.20] (Z_{K_{2,n}} = Z_{T_n}). These are published, standard identities rather than internal mistakes, but they are load-bearing for the K_{2,n} half: if either were false, that half of the main theorems would be unsupported even though the thagomizer results would stand. This is a dependency risk, not a demonstrated flaw.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies real-rootedness of Kazhdan–Lusztig polynomials and Z-polynomials of thagomizer matroids T_n=K_{1,1,n} and of the graphic matroids of K_{2,n}. The main result (Theorem 1.1) states that for every n≥2 and 0≤λ≤n/2, the pencil P_n(x)+λx has exactly ⌊n/2⌋ zeros, all negative and simple. This yields Theorem 1.2: P_n and P_{K_{2,n}}=P_n+x are real-rooted. Theorem 1.3 states that Z_{T_n}=Z_{K_{2,n}} has n+1 distinct negative zeros. The proofs introduce a rational substitution w=((q-1)/(q+1))^2, derive a mixed-coefficient estimate for the Catalan generating function (Lemma 2.2), use alternating signs at Chebyshev nodes, and apply the Lakatos–Losonczi criterion to a self-inversive transformation of the γ-polynomial.","tokens_in":11988,"tokens_out":30435,"duration_ms":190448,"significance":"The paper gives a clean, explicit proof of real-rootedness—a stronger property than the previously known log-concavity—for two natural families of matroids. The one-parameter pencil formulation is elegant, and the methods (Catalan coefficient estimates, Chebyshev interlacing, and self-inversive polynomial criteria) are elementary and potentially applicable to other families. The results are non-obvious and constitute a solid advance. In my assessment the main theorems are correct, but the written proof of Lemma 3.3 contains an algebraic error in its base case, and Remark 2.3 contains a false integral formula; both are local and fixable.","major_comments":[{"comment":"The displayed derivation of D_1(z) is incorrect. From (27) with r=1 and D_0(z)=C(z)/(1-z), the correct expression is D_1(z)=(zC(z)+√(1-4z))/(1-z)=(1-zC(z))/(1-z). The text instead writes D_1(z)=zC(z)+(1-z)√(1-4z)/(1-z)^2, which equals zC(z)+√(1-4z)/(1-z), not the correct value. Consequently the stated E_1(z)=z^3C(z)^4/(1-z)^2 is false; the correct E_1(z) is zC(z)^2/(1-z). Since this is the base case for the induction proving (28), and (28) underlies the coefficient sign pattern in (26) and hence the strict Lakatos–Losonczi inequality in Theorem 3.5, the proof must be corrected. The correction is straightforward and the lemma's conclusion remains true.","section":"Section 3.2, proof of Lemma 3.3, base case r=1"}],"minor_comments":[{"comment":"The semicircle-moment representation is false as stated. For n=2, (21) gives \\tilde P_2(y)=y^2/2-1, whereas (19) gives 2y^2-1. The integration in (21) yields coefficients C_{n-m}/2^{n-2m} rather than C_{n-m}. Since this remark is not used in the main theorems, either correct the formula or remove or rephrase the remark.","section":"Remark 2.3, Eq. (21)"},{"comment":"The K_{2,n} assertions are not re-derived in this paper: Theorem 1.2 for K_{2,n} relies on [8, Thm 5.8] and Theorem 1.3 relies on [5, Prop 5.20]. These are published identities, but the paper should explicitly flag them as imported in Section 1 so the reader is aware of the dependency.","section":"Introduction / Theorem 1.2"},{"comment":"The phrase 'alternating sign evaluations at the zeros of a Chebyshev polynomial' is slightly imprecise: the evaluations are at the nodes c_j=cos(jπ/n), 0≤j≤n, which include the endpoints ±1 in addition to the zeros of U_{n-1}. Consider rephrasing to 'Chebyshev nodes' for accuracy.","section":"Abstract and Section 2.3"}],"recommendation":"major_revision","confidential_remarks":"The main theorems appear correct and the overall strategy is sound, but the proof of Lemma 3.3 currently contains a false algebraic identity in its base case, which is load-bearing for the coefficient estimates. The fix is local (E_1(z)=zC(z)^2/(1-z)), and the false formula in Remark 2.3 is not used elsewhere. Once corrected, I expect the paper to be acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. This paper proves the Kazhdan–Lusztig polynomials of thagomizer matroids T_n (and the graphic matroid of K_{2,n}) are real-rooted, and with them the full pencil P_n(x)+λx for 0≤λ≤n/2. It also proves the shared Z-polynomial has n+1 distinct negative zeros. That is a real strengthening of the Wu–Zhang log-concavity result; I don't see these statements in the earlier literature.\n\nThe technical core is sound. The rational substitution w=((q-1)/(q+1))^2 converts the generating function into a palindromic form. For the KL pencil, the coefficient estimates reduce to one mixed-coefficient lemma for the Catalan series, proved by Lagrange inversion and the von Szily identity; I checked the telescoping and it works. The step from alternating signs at Chebyshev nodes to exact zero counting is legitimate, and the transfer back to the negative real axis preserves simplicity. The Z-polynomial side applies Lakatos–Losonczi correctly: D_n is self-inversive, the strict inequality follows from evaluating at q=1, and the simplicity on the unit circle transfers to the γ-polynomial and then to the Z-polynomial. The degree computation for Γ_n is careful.\n\nThe only genuine soft spot is that the K_{2,n} half leans on two imported identities: P_{K_{2,n}}=P_n+x from Gedeon–Proudfoot–Young and Z_{K_{2,n}}=Z_{T_n} from Ferroni–Nasr–Vecchi. The paper does not re-derive them. They are standard and published, but they are load-bearing for Theorems 1.2 and 1.3 as stated for K_{2,n}. If either had a hidden error, that half would collapse; the thagomizer results would survive. I'd want a referee to spot-check those references, but I see no sign of trouble.\n\nThis is not a paper that redefines the area, but it settles a natural question for two families and gives a reusable technique for this kind of real-rootedness problem. The writing is clear, the proofs are self-contained except for the cited transfers, and the computations are explicit.\n\nWho should read it: anyone working on real-rootedness of matroid KL polynomials, and people who use Chebyshev interlacing or self-inversive criteria in enumerative combinatorics. I'd send it to a serious referee; it deserves a careful read, and my own verdict is accept.","headline":"Real-rootedness for two infinite matroid families, proven by a clean Chebyshev/self-inversive argument; the only soft spot is a dependency on two published transfer identities for K_{2,n}.","tokens_in":12474,"tokens_out":1842,"would_cite":true,"duration_ms":14223,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05B35","05A15","05A20","26C10","30C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"A one-parameter polynomial family containing the Kazhdan–Lusztig polynomials of thagomizer and K_{2,n} matroids is shown to have only negative, simple real zeros.","keywords":["Kazhdan–Lusztig polynomial","Z-polynomial","real-rootedness","thagomizer matroid","graphic matroid","Chebyshev polynomial","self-inversive polynomial","Catalan numbers"],"falsifier":"Compute P_n(x) explicitly from the closed formula (12) for a small even n, say n=6, form P_6(x)+3x, and locate its zeros numerically. The theorem predicts exactly three negative simple zeros. If any zero is complex, repeated, or nonnegative, the central claim fails.","tokens_in":11527,"feed_emoji":"","tokens_out":3473,"duration_ms":28705,"temperature":0.7,"pith_summary":"For each n≥2, the paper studies the pencil P_n(x)+λx, where P_n is the Kazhdan–Lusztig polynomial of the thagomizer matroid T_n. It proves that for every λ in the interval [0, n/2], this polynomial has exactly ⌊n/2⌋ zeros, all negative and simple. Since the Kazhdan–Lusztig polynomial of the graphic matroid of K_{2,n} equals P_n(x)+x, the theorem immediately makes both matroid families real-rooted. The same rational substitution also shows the common Z-polynomial Z_{T_n}=Z_{K_{2,n}} has n+1 distinct negative zeros. A sympathetic reader would care because real-rootedness is a much stronger property than the known nonnegativity and log-concavity results, and it settles the question for these two graphic families.","feed_headline":"Thagomizer and K2,n matroids: all zeros negative and simple","feed_subtitle":"The one-parameter pencil P_n(x)+λx is real-rooted for 0≤λ≤n/2, so Kazhdan–Lusztig and Z-polynomials follow.","key_machinery":"The argument rests on a rational change of variables w=((q−1)/(q+1))^2, which converts each polynomial in the pencil into a palindromic polynomial. For the Kazhdan–Lusztig case, the transformed polynomial is evaluated at roots of unity; a mixed-coefficient estimate for the Catalan generating function yields alternating signs at Chebyshev nodes, giving strict interlacing with the Chebyshev polynomial U_{n−1} and hence the exact zero count. For the Z-polynomial, the same transformation produces a self-inversive polynomial whose coefficients satisfy the Lakatos–Losonczi unit-circle criterion, forcing all zeros onto the unit circle and then, after pulling back, onto the negative real axis.","core_discovery":"The central claim is that for every integer n≥2 and every real λ with 0≤λ≤n/2, the pencil P_n(x)+λx has exactly ⌊n/2⌋ zeros, all negative and simple. Here P_n is the Kazhdan–Lusztig polynomial of the thagomizer matroid T_n=K_{1,1,n}, given explicitly by the Catalan generating function. Setting λ=0 gives real-rootedness of P_n itself; setting λ=1, together with the identity P_{K_{2,n}}(x)=P_n(x)+x, gives real-rootedness for the graphic matroid of K_{2,n}. The paper also proves that the Z-polynomial Z_{T_n}=Z_{K_{2,n}} has n+1 distinct negative zeros for n≥2. All zeros are shown to be not merely real but strictly negative and simple.","pith_inferences":["The interval 0≤λ≤n/2 may well be maximal; one could test whether λ slightly above n/2 causes a pair of conjugate complex roots to form, which would pinpoint the boundary of real-rootedness in this pencil.","The Chebyshev interlacing technique, combined with the Catalan generating function, might extend to other matroid families whose Kazhdan–Lusztig polynomials satisfy similar algebraic generating functions, such as certain uniform or lattice-path matroids.","The semicircle-moment representation (21) suggests a probabilistic reading of the zeros as supporting measures, which could lead to alternative moment-sequence proofs or to generalizations where the weight Cn−m is replaced by other moment sequences."],"forward_implications":["The Kazhdan–Lusztig polynomials of the thagomizer matroids T_n and of the graphic matroids K_{2,n} are real-rooted for all n≥2, hence their coefficients are log-concave by Newton's inequalities.","The descent polynomial of 321-avoiding permutations, which equals P_n(x), is real-rooted.","The Z-polynomials Z_{T_n}(x)=Z_{K_{2,n}}(x) have n+1 distinct negative zeros for every n≥2, so their γ-polynomials have only simple negative zeros.","The whole pencil P_n(x)+λx is real-rooted throughout the interval 0≤λ≤n/2, not just at the two endpoints that correspond to the matroid families.","All zeros of these polynomials are simple, which rules out repeated-root behavior and makes associated interlacing properties particularly clean."],"fun_headline_variants":["Real-rooted Kazhdan-Lusztig for thagomizer and K2,n","All zeros negative and simple in P_n(x)+λx pencil","Z-polynomials of thagomizer and K2,n: distinct negative zeros","P_n(x)+λx real-rooted for 0≤λ≤n/2"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The conclusions about K_{2,n} rest on two identities imported from earlier papers — P_{K_{2,n}}(x)=P_n(x)+x and Z_{T_n}(x)=Z_{K_{2,n}}(x) — which the paper does not re-derive; if either were false, the K_{2,n} theorems would be unsupported, though the thagomizer results would stand.","fun_headline_variants_meta":{"raw":{"variants":["Real-rooted Kazhdan-Lusztig for thagomizer and K2,n","All zeros negative and simple in P_n(x)+λx pencil","Z-polynomials of thagomizer and K2,n: distinct negative zeros","P_n(x)+λx real-rooted for 0≤λ≤n/2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001016,"raw_usage":{"total_tokens":4158,"prompt_tokens":808,"completion_tokens":3350,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":552,"completion_tokens_details":{"reasoning_tokens":3261}},"tokens_in":552,"tokens_out":3350,"duration_ms":20639,"temperature":1.0,"reasoning_tokens":3261,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T09:54:14.955288+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute P_n(x) explicitly from the closed formula (12) for a small even n, say n=6, form P_6(x)+3x, and locate its zeros numerically. The theorem predicts exactly three negative simple zeros. If any zero is complex, repeated, or nonnegative, the central claim fails.","supporting_citations":[],"review_version":1}