{"id":"5a2f36bf-5e89-48ae-af2f-851aaa5b6dab","arxiv_id":"2607.10097","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The q-Onsager algebra's PBW bases and alternating centralisers, conjectured by Terwilliger and by Baseilhac–Belliard, are proven over any field with q not a root of unity.","lead":"This paper proves five long-standing conjectures about the q-Onsager algebra, the algebra behind boundary quantum integrable models, showing that its elements have unique ordered normal forms for every admissible ordering and that certain commutative subalgebras are maximal. The results give researchers unique normal forms and maximal commuting subalgebras in a structure central to boundary integrable systems.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 10.3 contains an incorrect constant: γ_r has ⟨r⟩_t! in the numerator, but it belongs in the denominator; as printed, (10.11) fails for r=2 and the simplification to (10.22) is inconsistent. The fix is mechanical and restores the centraliser argument.","rationale":"The reader's weakest assumption is precisely Lemma 10.3, and I agree. The constant error is real and occurs in a lemma on which the negative centraliser theorem depends. Contrary to a merely cosmetic typo, the printed γ_r makes Lemma 10.3 false, and the proof's simplification to (10.22) is inconsistent with that printed value. However, the corrected γ_r is obtained by a straightforward iteration of (10.12), and it restores the equality in (10.22) without changing any other step. I found no further load-bearing flaw in the root-vector PBW theorem, the alternating PBW bases, or the imaginary centraliser argument; the reliance on Terwilliger's Theorem 3.1 and Damiani's straightening relations is explicit and standard. The reader's CONDITIONAL verdict is therefore appropriate: the paper should be accepted after the constant in Lemma 10.3 is corrected and the surrounding scalars are rechecked.","tokens_in":29046,"tokens_out":27648,"duration_ms":254802,"concrete_test":"Take r = 2, v = 1, and t = q^2 generic in Lemma 10.3. Compute S_2 1 = C_t C_{t^2}1 = (1 − t)(1 − t^2) x⋆x and D^2(S_2 1) = (1 − t^2)^2. The identity D^2 z = v forces γ_2 = (1 − t^2)^{-2}; the printed γ_2 in (10.11) is (1 + t)/((t − 1)(t^2 − 1)) = (t − 1)^{-2}. Re-run the derivation of (10.22) with the corrected γ_r = (−1)^r/(⟨r⟩_t! ∏_{j=1}^r (t^j − 1)) and verify the scalar becomes ⟨r+1⟩_t. This settles both the existence of the error and the sufficiency of the correction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The negative-centraliser proof in Section 10 depends on Lemma 10.3. Iterating the recurrence (10.12) gives the coefficient γ_r = (-1)^r / (⟨r⟩_t! ∏_{j=1}^r (t^j − 1)), not the printed γ_r = (-1)^r ⟨r⟩_t! / ∏_{j=1}^r (t^j − 1). A concrete failure: for r = 2, v = 1, one computes D^2(S_2 1) = (1 − t^2)^2, so the requirement D^2 z = v forces γ_2 = 1/(t^2 − 1)^2, whereas the printed value is 1/(t − 1)^2. Thus Lemma 10.3 as stated is false. This is load-bearing because Proposition 10.5 uses Lemma 10.3 to write z = γ_r S_r v and then claims the scalar in (10.22) equals ⟨r+1⟩_t. With the printed γ_r, that scalar is ⟨r⟩_t! ⟨r+1⟩_t; the displayed simplification is algebraically inconsistent. Since Theorem 10.11 and Corollary 10.13 (Conjecture 16.7) rely on Proposition 10.5, this spot must be repaired. The surrounding argument appears sound: with the corrected γ_r, the factors cancel and (10.22) is true, so the intended conclusion is preserved.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the q-Onsager algebra O_q over an arbitrary field F and q not a root of unity. It proves three main families of results: (1) the Baseilhac–Kolb root vectors form a PBW basis for every total order on the positive roots of ŝl_2 (Theorem 1.1), removing a previous transcendence hypothesis; (2) twelve PBW bases in the three alternating families W^-, W^+, G/eG, together with persistence under arbitrary scalar central specialisation of the alternating central extension (Theorem 1.2, Corollary 1.3); and (3) the negative alternating subalgebra and the algebra generated by eG_1 are self-centralising, so the four single-family alternating polynomial subalgebras are maximal commutative (Theorem 1.4). The overall strategy is a filtered-lifting argument: initial forms are identified with alternating words in the q-shuffle realisation of U_q^+(ŝl_2), and PBW bases are lifted using general filtered-lifting lemmas. The centraliser results are proved by a deletion-operator argument in the q-shuffle algebra and by PBW-profile arguments.","tokens_in":29405,"tokens_out":29486,"duration_ms":293315,"significance":"If correct, these results settle four conjectures of Terwilliger (Conjectures 16.1, 16.2, 16.7, 16.8) and, for char F ≠ 2, the Baseilhac–Belliard W G-basis conjecture. The arbitrary-field formulation and the arbitrary-total-order PBW statement are genuine improvements over previous approaches that required a transcendence hypothesis. The paper is carefully structured: scalar initial forms are computed explicitly, denominators are checked systematically through Lemma 2.1, and the filtered-lifting lemmas are clean and reusable. No circularity is apparent: the argument uses, but does not assume, the conjectures being proved. There is, however, one serious algebraic error in the negative-centraliser section: Lemma 10.3 states an incorrect constant. This is local and repairable, and the surrounding argument survives with the corrected constant, but the manuscript as written contains a false lemma at a load-bearing point.","major_comments":[{"comment":"Lemma 10.3 is false as stated. Iterating the recurrence (10.12) gives γ_r = (-1)^r / (⟨r⟩_t! ∏_{j=1}^r (t^j-1)), not the printed γ_r = (-1)^r ⟨r⟩_t! / ∏_{j=1}^r (t^j-1). A concrete failure is r = 2, z = x^2. Then [z,x] = 0, d(z) = 2, and v = D^2 z = 1, Dv = 0. For the asserted relation z = γ_2 S_2 v to hold, applying D^2 gives 1 = γ_2 D^2(S_2 1), and D^2(S_2 1) = (t^2-1)^2, so γ_2 = (t^2-1)^{-2}. The printed value is (t-1)^{-2}. This is not a harmless typo in isolation: the lemma is used in Proposition 10.5 and Theorem 10.11. The corrected formula restores the intended argument.","section":"§10.1, Lemma 10.3, Eq. (10.11)"},{"comment":"Equation (10.22) is algebraically inconsistent with the printed value of γ_r. With the printed γ_r, the scalar on the right is ⟨r⟩_t! ⟨r+1⟩_t, not ⟨r+1⟩_t; the displayed simplification to (t^{r+1}-1)/(t-1) omits a factor of ⟨r⟩_t!. With the corrected γ_r, the factors cancel and (10.22) becomes exactly ⟨r+1⟩_t [v,G]. Fortunately, Proposition 10.5 only needs the scalar to be non-zero, and both the printed and corrected scalars are non-zero. Thus the main conclusion of Theorem 10.11 is salvageable, but the proof as written relies on a false identity. The manuscript must be revised to correct Lemma 10.3 and to re-state or re-derive (10.22).","section":"§10.2, Proposition 10.5, Eq. (10.22)"}],"minor_comments":[{"comment":"At the start of Section 10 only W_-n and G_n are defined explicitly, but Corollary 10.7 uses eG_n. Add the definition eG_n = (xy)^n to avoid ambiguity.","section":"§10, beginning"},{"comment":"The triangularity step using (3.13) is stated very tersely. It is correct: the diagonal coefficients are (-1)^n [n]_q and [2n]_q, both non-zero. Please spell this out, since the equality F[J_1,J_2,...] = F[eG_1,eG_2,...] is load-bearing for the imaginary centraliser.","section":"§10.3, Corollary 10.7"},{"comment":"Reference [7] is cited as an arXiv preprint; if a journal version exists, it should be updated.","section":"References"},{"comment":"The notation eA^{{δ}}_{q,ρ} is introduced twice with slightly different typography; a short remark that {δ} is not a set but a sequence would improve readability.","section":"§9.1"}],"recommendation":"major_revision","confidential_remarks":"The main issue is the incorrect constant in Lemma 10.3. It is local, and the intended argument is recoverable with the corrected scalar, so this is not a reject. I also checked the other potentially delicate points — the arbitrary-order straightening in Section 4, the initial-form calculations in Sections 5–6, and the PBW-profile arguments in Sections 10–11 — and found no additional errors. The paper is well within the journal's scope and should be acceptable after the lemma and equation (10.22) are repaired and the surrounding text is updated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis paper deserves a serious look. It proves four Terwilliger conjectures and one Baseilhac–Belliard conjecture about PBW bases and centralisers for the q-Onsager algebra. The arbitrary-order Damiani PBW theorem, the exact scalar initial forms, persistence under scalar central specialisation, and the centraliser arguments are genuinely new. The filtered-lifting method is clean, and the left-deletion operator in Section 10 is a nice new tool. I checked several denominators and scalar factors in Sections 4–9; they look correct.\n\nThe soft spot is exactly where the reader flagged it. Lemma 10.3 has a wrong constant: iterating (10.12) gives γ_r = (-1)^r / (⟨r⟩_t! ∏_{j=1}^r (t^j−1)), not the printed (-1)^r ⟨r⟩_t! / ∏_{j=1}^r (t^j−1). For r=2 and v=1, D^2(S_2 1) = (1−t^2)^2, so γ_2 should be 1/(t^2−1)^2, not 1/(t−1)^2. As written, the lemma is false. This is load-bearing because Proposition 10.5, Theorem 10.11, and Corollary 10.13 depend on it. The good news is that the fix is mechanical: with the corrected γ_r, the scalar in (10.22) simplifies to ⟨r+1⟩_t as claimed, and the centraliser argument goes through. I found no other issue of this magnitude. The paper relies heavily on Terwilliger’s and Damiani’s external theorems, but it does not assume the conjectures it proves, and the citation pattern looks honest.\n\nWho is this for? Anyone working on q-Onsager algebras, boundary integrable models, or PBW bases for quantum symmetric pairs. It deserves a serious referee. My recommendation: send it to review, with a request that the author fix the constant in Lemma 10.3 and re-check everything that depends on it. The intended results are very likely correct, but the current text needs repair.\n\nBest.","headline":"A strong paper that settles several open conjectures, but the negative-centraliser section contains a fixable yet load-bearing constant error in Lemma 10.3.","tokens_in":29933,"tokens_out":5329,"would_cite":true,"duration_ms":45864,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["17B37","17B67","81R50"],"pacs":[],"model":"deepseek-v4-flash","headline":"Over any field in which the deformation parameter q is not a root of unity, the q-Onsager algebra admits a PBW basis in every ordering of its root vectors, twelve alternating PBW bases that survive central specialisation, and four maximal c","keywords":["q-Onsager algebra","PBW basis","alternating generators","root vectors","centraliser","q-shuffle algebra","central specialisation","maximal commutative subalgebra"],"falsifier":"Fix F=Q and q=2. In the q-shuffle algebra, compute the subspace of words of length at most six that commute with both x and xyx (i.e., W0 and W-1). The paper predicts this subspace is exactly spanned by the negative alternating words; finding any extra commuting element would disprove the negative centraliser theorem and the dependent PBW claims.","tokens_in":28881,"feed_emoji":"🧮","tokens_out":12249,"duration_ms":124570,"temperature":0.7,"pith_summary":"The paper proves that the q-Onsager algebra — the q-deformed symmetry algebra of boundary integrable systems — has a Poincaré–Birkhoff–Witt (PBW) basis of ordered monomials in its root vectors for every possible ordering of the positive roots, and twelve PBW bases in the alternating generators. This removes a transcendence hypothesis that previously limited such statements. The author also determines the centralisers of the negative alternating family and of the first imaginary alternating generator, showing that four polynomial subalgebras are maximal commutative. These results settle four conjectures from the literature on the algebra's alternating central extension and, outside characteristic 2, the 'WG-basis' conjecture for current-algebra quotients. The argument works by lifting PBW bases from the associated graded algebra, identified with the q-shuffle algebra, and by a new deletion operator for centraliser computations.","feed_headline":"Five conjectures settled: q-Onsager PBW in every root order","feed_subtitle":"Arbitrary-field proof also pins down centralisers and proves the WG-basis conjecture away from characteristic 2.","key_machinery":"The q-shuffle realisation of the positive part U_q^+(sl2-hat) as an algebra of words in x and y under the q-shuffle product, together with the alternating PBW theorem for that algebra. The paper computes the initial forms (leading homogeneous components) of the root vectors and alternating generators as explicit scalar multiples of the alternating words, then applies a filtered-lifting lemma: any PBW basis in the associated graded algebra lifts to O_q. For the centraliser half, a left-deletion operator D on the shuffle algebra reduces the joint centraliser of W0 and W-1 to the centraliser of the first imaginary generator; a filtration by the number of G-factors completes the negative case, a","core_discovery":"The central claim: over any field in which q is not a root of unity, the q-Onsager algebra admits PBW bases in every total order of its distinguished root vectors, and twelve PBW bases in its alternating generators; these bases survive arbitrary scalar central specialisation. The paper further proves that the four polynomial subalgebras generated by the alternating families are self-centralising, hence maximal commutative. This resolves a package of open conjectures about the algebra's PBW structure and, in characteristic not 2, establishes the 'WG-basis' conjecture for the current-algebra quotients.","pith_inferences":["The arbitrary-order PBW property hints at a universal PBW structure for quantum symmetric-pair coideal subalgebras; the same filtered-lifting strategy may transfer to other examples of that class.","The left-deletion operator is a new tool for computing centralisers in q-shuffle algebras and could extend to higher-rank shuffle realisations where such centraliser problems have resisted direct computation.","In finite-dimensional representations of O_q, the maximal commutative subalgebras identified here are natural candidates for complete sets of commuting integrals; this could be tested numerically in low dimensions.","The characteristic-2 restriction on the WG-basis arises only from the comparison of central coordinates; a characteristic-free reformulation of those coordinates may lift the restriction."],"forward_implications":["Every element of the q-Onsager algebra has a unique normal form in the root vectors for any chosen total order, and twelve normal forms in the alternating generators.","The alternating PBW bases persist under scalar central specialisation, so the WG-basis normal form for current-algebra coefficients holds outside characteristic 2.","The four single-family alternating polynomial subalgebras are maximal commutative, making them maximal sets of pairwise-commuting elements in O_q.","The associated graded algebra has the Hilbert series Hilb = ∏_{m≥1} (1-t^{2m-1})^{-2}(1-t^{2m})^{-1}, giving a numerical invariant of O_q.","The removal of the transcendence hypothesis means the PBW results are valid over arbitrary fields, not just over the complex numbers."],"fun_headline_variants":["q-Onsager PBW bases in every root order, any field","Twelve PBW bases for q-Onsager, surviving central specialisation","Centralisers pinned down: four polynomial subalgebras maximal","Five conjectures settled in q-Onsager PBW theory","Arbitrary-field proof: PBW bases for q-Onsager in all orders"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the straightening relations, the alternating PBW theorem, and the deletion-operator constant used to compute initial forms and centralisers hold verbatim over every field with q not a root of unity; any coefficient error or hidden characteristic restriction there would void the lifted PBW bases and centraliser equalities.","fun_headline_variants_meta":{"raw":{"variants":["q-Onsager PBW bases in every root order, any field","Twelve PBW bases for q-Onsager, surviving central specialisation","Centralisers pinned down: four polynomial subalgebras maximal","Five conjectures settled in q-Onsager PBW theory","Arbitrary-field proof: PBW bases for q-Onsager in all orders"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000219,"raw_usage":{"total_tokens":1232,"prompt_tokens":647,"completion_tokens":585,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":391,"completion_tokens_details":{"reasoning_tokens":490}},"tokens_in":391,"tokens_out":585,"duration_ms":5782,"temperature":1.0,"reasoning_tokens":490,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T07:25:46.986021+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fix F=Q and q=2. In the q-shuffle algebra, compute the subspace of words of length at most six that commute with both x and xyx (i.e., W0 and W-1). The paper predicts this subspace is exactly spanned by the negative alternating words; finding any extra commuting element would disprove the negative centraliser theorem and the dependent PBW claims.","supporting_citations":[],"review_version":2}