{"id":"56fc00c0-fc0b-4f77-af73-d4fdb3464387","arxiv_id":"1908.07128","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"All rank 6 modular categories with Galois group (012)(345) have modular data conjugate to one of two explicit families, one of which is nonunitarizable.","lead":"This paper proves a complete list of the rank 6 modular tensor categories whose Galois group is the permutation (012)(345), showing only two families exist. The result is a step in a classification program for these objects, which model certain condensed matter systems, and it identifies the smallest rank of fusion rules with no physically realizable version.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section 6's reductions to Case 1 are internally inconsistent: in Sign Choice 1, Cases 3–5 the derived relations force θ2 = −1, not the Case 1 hypothesis θ2 = 1, so Lemma 3.2's nondegeneracy proof is incomplete as printed.","rationale":"The central claim is a complete classification, and the classification proof splits on irreducibility of the SL(2,Z)-representation. That irreducibility is Lemma 3.2, whose only evidence is Section 6's Gröbner tables. The tables are not merely unauditable; one of the reductions they print is internally inconsistent: in Sign Choice 1, Case 3, the zero factors added imply θ2=−1 while the text claims reduction to Case 1 (which requires θ2=1). This is exactly the kind of concrete gap that could invalidate the nondegeneracy conclusion. The reader's weakest_assumption points at the same computational core, so my agreement is partial: I sharpen the auditability concern into a specific algebraic discrepancy. The paper may still be correct—a fresh Macaulay2 run might produce unit ideals—but as written the proof does not go through. I therefore move the verdict from CONDITIONAL to UNVERDICTED pending the recomputation; if the recomputation reproduces the claimed contradictions and the labels are fixed, the classification would again be viable.","tokens_in":12333,"tokens_out":17569,"duration_ms":163946,"concrete_test":"Implement the algorithm described in Section 6 in Macaulay2 from scratch, using only the stated initial ideals (twist relations, orthogonality, sign-choice equations, and case equations). For Sign Choice 1, Cases 3, 4, and 5, add the printed zero factors and compute a Gröbner basis; determine whether the ideal becomes the unit ideal. In particular, for Case 3, the ideal after adding θ4+θ5=0 and θ5+1=0 contains θ1=1, θ2=−1, θ3=1; test whether this ideal is already contradictory. Also run the full set of 35 cases and compare the unit-ideal outcomes with the table claims, then release the script. If any case fails to reach the unit ideal, Lemma 3.2 and Theorem 3.5 are unsupported.","verdict_should_be":"UNVERDICTED","load_bearing_attack":"Lemma 3.2 is the load-bearing step: without it, Theorem 3.5 cannot invoke Eholzer's classification of irreducible representations, and the whole dichotomy (2×3 versus 6-dimensional irreducible) collapses. The lemma is proved only by the Gröbner basis tables in Section 6. As printed, those tables contain a concrete logical error. For Sign Choice 1, Case 3, the initial ideal includes θ3−1=0, θ4−θ1=0, θ5−θ2=0. The table adds the zero factors θ4+θ5=0 and θ5+1=0, so θ5=−1, θ4=1, hence θ1=1 and θ2=−1. The text then states 'We now have θ4 = θ1 = 1, so we reduce to case 1 and are done,' but Case 1 is defined as 1=θ0=θ1=θ2, which requires θ2=1. Sign Choice 1, Cases 4 and 5 have the same defect: the printed zero factors force, respectively, (θ2=1, θ1=−1) and (θ1=1, θ2=−1), neither of which is Case 1. Several other table blocks (e.g., Sign Choice 1, Case 1) never display the unit-ideal marker '1', and the repeated 'Sign Choice 15' headers make it impossible to tell which sign choice is being eliminated. Since no Macaulay2 input is shipped, the nondegeneracy of the t-spectra—and hence the irreducibility claim—is not established by the text. This is a correctness gap, not merely a preference for reproducible code.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper classifies rank 6 non-integral self-dual modular tensor categories with Galois group ⟨(012)(345)⟩. The main theorem (Theorem 3.5) states that, up to relabeling and Galois conjugation, the only realizable modular data are the product of the semion category with (A1,5)_{1/2} and a modular subcategory of C(so5,9,e^{jπi/9}) with gcd(18,j)=1. The proof combines Galois symmetry (reducing the number of sign choices), a Gröbner basis computation to prove irreducibility of the SL(2,Z) representation (Lemma 3.2), and Eholzer's classification of irreducible modular fusion algebras to pin down the S and T matrices. The paper also includes the Gröbner basis tables in Section 6 and a short discussion of future work.","tokens_in":12670,"tokens_out":10038,"duration_ms":82275,"significance":"The classification result is significant for the ongoing rank-by-rank program: it identifies exactly two families of modular data for this Galois group, and it exhibits a rank-6 family that is nonunitarizable, complementing the structure theory of modular tensor categories. The paper provides explicit S and T matrices, computes the relevant fusion rules, and connects the resulting modular data to independent constructions in [12], which gives evidence that the statement is correct. The main weakness is that the crucial irreducibility lemma is outsourced to computational tables that are neither internally consistent nor reproducible from the text.","major_comments":[{"comment":"The displayed 'zero factors added' contradict the claimed reduction to Case 1. In Case 3, the zero factors θ4+θ5=0 and θ5+1=0 give θ5=−1 and θ4=1; with the initial relations θ4−θ1=0 and θ5−θ2=0 this forces θ1=1 and θ2=−1. The text then says 'We now have θ4 = θ1 = 1, so we reduce to case 1 and are done,' but Case 1 is defined by 1=θ0=θ1=θ2, which requires θ2=1. Cases 4 and 5 have the same defect: the zero factors force (θ2=1, θ1=−1) and (θ1=1, θ2=−1), respectively, neither satisfying the Case 1 hypothesis. Consequently, Proposition 6.1 cannot be invoked for these reductions, and the nondegeneracy of the t-spectra for these sign choices is not established by the text. Since Lemma 3.2 is the sole basis for the irreducibility used in Theorem 3.5, this is a load-bearing correctness gap.","section":"Section 6, Sign Choice 1, Cases 3–5"},{"comment":"The proofs of Lemma 3.2 and the final T-matrix computation rely on thirty-five Gröbner basis computations that are summarized only in tables; no Macaulay2 input scripts or output logs are provided, and the algorithm is stated to be 'essentially unchanged from [4]', an unpublished thesis. The printed tables also contain repeated 'Sign Choice 15' headers after the 'Sign Choice 7, Case 3' block and after the 'Sign Choice 8, Case 5' block, so the correspondence between rows and sign choices is ambiguous. As a result, the manuscript does not provide a verifiable proof of the computational claims on which the classification rests. The authors should either include the complete code and detailed outputs, or replace the table summary with a fully explicit and internally consistent derivation.","section":"Section 6, general"}],"minor_comments":[{"comment":"The two blocks following 'Sign Choice 7, Case 3' are headed 'Sign Choice 15, Case 4' and 'Sign Choice 15, Case 5'; this appears to be a labeling error that should be corrected.","section":"Section 6, Sign Choice 7"},{"comment":"The equation 'r1r2 + r2r3 + r1r2 = −3' should presumably read 'r1r2 + r2r3 + r1r3 = −3.'","section":"Proof of Theorem 3.5"},{"comment":"The displayed relations 'θ2 2 + θ2 + 1' and 'θ2 5 + x + θ5 − d5' are typeset ambiguously; they should be θ2² + θ2 + 1 and θ5² + x + θ5 − d5.","section":"Section 6, final relations"},{"comment":"Reference [11] is missing the author name in the bibliography.","section":"References"},{"comment":"The phrase 'nonunitarizable (hence, nonphysical)' uses 'hence' in a way that is not mathematically justified; the correct statement would be that such categories do not admit a unitary structure, which rules out certain physical realizations.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The computational core of this paper is not included as executable code, and the tables contain enough inconsistencies that a careful reader cannot verify Lemma 3.2. I would encourage the editor to request the Macaulay2 scripts and outputs, or an independent derivation, before considering acceptance. The reliance on Creamer's unpublished thesis is acceptable in principle but should be made explicit and the relevant statements quoted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a genuine contribution to the rank-6 classification program: it settles the Galois group <(012)(345)> case, gives two explicit families of (S,T), and shows one family is nonunitarizable, which yields the smallest known rank of fusion rules with no unitary realization. The proof strategy—Galois symmetries plus Gröbner bases—is sensible, and the final match to the independent constructions in [12] gives me real confidence that the classification is correct.\n\nThe soft spot is the computational core, and it is more than a style complaint. Lemma 3.2 is load-bearing: it needs nondegenerate t-spectra to apply Eholzer's classification, and the only evidence is the Section 6 tables. As printed, those tables contain a concrete logical error. In Sign Choice 1, Cases 3–5, the zero factors added (θ5+1, θ4+θ5, etc.) force θ2 = −1 or θ1 = −1, but Case 1 is defined as θ1 = θ2 = 1. The text says \"we reduce to case 1 and are done\" in each of those rows, and that inference is false. The same pattern appears in several of the later Sign Choice blocks. If the underlying Macaulay2 computations actually produce those zero factors, then the reductions need to be reported differently; if the zero factors are misprints, the tables need to match the real runs. Either way, the printed proof does not establish Lemma 3.2.\n\nThe mislabeled \"Sign Choice 15\" headers only make this worse. No scripts are shipped, and the algorithm is described as copied from an unpublished thesis. That combination means the reader cannot audit the one step that the whole classification depends on. There are also minor typos: \"r1r2 + r2r3 + r1r2\" should surely be \"r1r2 + r2r3 + r1r3\".\n\nI think the result is very likely right, and the nonunitarizability consequence is interesting. But a classification theorem whose key lemma is supported by irreproducible and internally inconsistent tables should not be accepted as-is. The fix is straightforward: ship the Macaulay2 input, correct the Section 6 labels and the Case 1 reductions, and make the Creamer thesis publicly available or summarize its relevant results.\n\nThis paper deserves a serious referee—the question is not whether to engage, but how strongly to require the computational audit before publication. Send it to review, but tell the author the scripts and tables have to be repaired first.","headline":"Plausible and likely correct classification, but the printed Gröbner basis proof has a concrete gap in the Case 1 reductions that needs fixing before the result is trustworthy.","tokens_in":13191,"tokens_out":1833,"would_cite":false,"duration_ms":19086,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18M20","18M15","81R50"],"pacs":[],"model":"deepseek-v4-flash","headline":"Two families exhaust rank-6 modular data with Galois group (012)(345)","keywords":["modular tensor categories","rank 6 classification","Galois group","modular data","Grobner basis","nonunitarizable categories","quantum groups","fusion rules"],"falsifier":"Run the 35 Gr\\\"obner basis computations with the same sign choices and degeneracy cases using a computer algebra system: if any computation returns a non-unit ideal, Lemma 3.2's irreducibility claim is false. Alternatively, construct or find a rank 6 non-integral self-dual MTC with this Galois group whose modular data is not conjugate to the two displayed pairs.","tokens_in":12105,"feed_emoji":"🧮","tokens_out":8003,"duration_ms":76325,"temperature":0.7,"pith_summary":"The paper sets out to finish one branch of the low-rank classification of modular tensor categories: rank 6, non-integral, self-dual categories whose Galois group is the three-cycle group generated by $(012)(345)$. It proves that, up to relabeling and Galois conjugation, exactly two families of modular data $(S,T)$ are realizable. The first is the product of the semion category (the basic rank-2 modular category) with a rank-3 category from the $(A_1,5)_{1/2}$ series; the second is a modular subcategory of a quantum-group category built from $\\mathfrak{so}_5$ at level 9. A byproduct is that the second family, forced by the symmetries, is nonunitarizable, so its modular data cannot come from a unitary physical theory. Completing this case leaves only two Galois groups open in the full rank-6 classification.","feed_headline":"Two families exhaust rank-6 modular data with Galois group (012)(345)","feed_subtitle":"The symmetry analysis also rules out unitary realizations, leaving a semion product and an so(5) level-9 subcategory.","key_machinery":"The load-bearing object is the Galois action on modular data. The Galois group is identified with a permutation of the columns of the normalized $S$-matrix, and the symmetry $\\sigma^2(t_i) = t_{\\sigma(i)}$ constrains the twist eigenvalues. After reducing the sign choices $\\epsilon_i$ to seven equivalence classes, the proof uses 35 Gr\\\"obner basis computations over the admissible-data polynomial equations to show that any degenerate t-spectrum leads to a contradiction; that establishes irreducibility of the resulting $\\mathrm{SL}(2,\\mathbb{Z})$ representation. Irreducibility splits the problem into the $2\\otimes 3$ branch when $7\\mid N$ and a 6-dimensional irreducible of type $B_9$ when $9\\mid N$, where the fusion rules force the $S$ and $T$ matrices and the Verlinde formula checks consistency.","core_discovery":"Theorem 3.5 states that for any rank 6 non-integral self-dual modular tensor category with $\\mathrm{Gal}(\\mathcal{C}) = \\langle(012)(345)\\rangle$, the modular data must be conjugate to one of two explicit pairs. In the $7 \\mid N$ case, where $N$ is the order of the twist matrix $T$, the data is a tensor product: the $2\\times 2$ Hadamard matrix times a $3\\times 3$ matrix involving $d = 2\\cos(\\pi/7)$, with twists that are 7th roots of unity. In the $9 \\mid N$ case, the normalized $S$-matrix is built from three algebraic numbers $r_1,r_2,r_3$ in $\\mathbb{Q}(e^{i\\pi/9})$ satisfying $r_1+r_2+r_3=0$ and $r_1r_2+r_2r_3+r_1r_2=-3$, and the twists are 3rd and 9th roots of unity. Both displayed data are realizable: the first by the semion product, the second by subcategories of $\\mathcal{C}(\\mathfrak{so}_5,9,e^{j\\pi i/9})$ with $\\gcd(18,j)=1$. Consequently this Galois group yields no new unitary physics; the second family is nonunitarizable.","pith_inferences":["An independent audit of the 35 Gr\\\"obner computations, with the sign choices and degeneracy cases spelled out and with corrected table headers, would directly test Lemma 3.2; the repeated 'Sign Choice 15' headers in Section 6 make such an audit worthwhile.","The mechanism that makes sign choice 15 nonunitarizable may generalize: any rank where Galois-sign symmetries force a particular sign pattern could yield nonunitarizability before the full modular data are computed.","Testing Lemma 3.3 at higher rank would show whether factorization of a non-integral fusion ring automatically gives modular factors, which would shorten higher-rank classifications if true.","The $9\\mid N$ branch is classified only at the level of modular data; identifying the underlying categories beyond the constructed subcategories remains open."],"forward_implications":["Every rank 6 non-integral self-dual MTC with $\\mathrm{Gal}(\\mathcal{C})=\\langle(012)(345)\\rangle$ has modular data conjugate to one of the two displayed families.","The $9\\mid N$ family is nonunitarizable, so the symmetries of this Galois group rule out unitary, and hence physical, realizations.","The fusion rules of $\\mathrm{PSO}(5)_{3/2}$ are the smallest-rank fusion rules with no unitary realization.","Together with prior work, the only remaining Galois groups in the rank-6 classification are $\\langle(012345)\\rangle$ and $\\langle(01)(23)(45),(02)(13)\\rangle$.","Both families are known to exist as actual categories, not merely admissible data: the first as a semion product, the second as quantum-group subcategories."],"supporting_citations":[{"why":"Supplies the admissibility conditions for modular data and the rank-finiteness theorem that justifies classification by rank.","marker":"[2]"},{"why":"Provides the Galois symmetry theorem, including the permutation action and the relation $\\sigma^2(t_i)=t_{\\sigma(i)}$, and the divisibility result forcing 7 or 9 to divide $N$.","marker":"[3]"},{"why":"Supplies the list of possible Galois groups for rank 6 and the Gr\\\"obner basis algorithm used, essentially unchanged, in Lemma 3.2.","marker":"[4]"},{"why":"Provides the irreducibility criterion via nondegenerate t-spectra, the table of $\\mathrm{SL}(2,\\mathbb{Z}/7\\mathbb{Z})$ representations, and the type $B_9$ fusion algebra.","marker":"[7]"},{"why":"Gives the norm bound $N \\mid D^5$ used to show the character is trivial in the $9\\mid N$ case.","marker":"[8]"},{"why":"Supplies the decomposition theorem used in Lemma 3.3 to split a fusion-ring product into modular factors.","marker":"[9]"},{"why":"Classifies pre-modular categories of rank 3, used in Lemma 3.3 to eliminate candidate fusion rules.","marker":"[11]"},{"why":"Constructs the quantum-group categories realizing the $9\\mid N$ modular data as subcategories of $\\mathcal{C}(\\mathfrak{so}_5,9,e^{j\\pi i/9})$.","marker":"[12]"},{"why":"Classifies modular categories through rank 4 and gives the rank 2 and 3 classifications used to identify the $7\\mid N$ solution.","marker":"[13]"}],"fun_headline_variants":["Two families exhaust rank-6 modular data with (012)(345), both nonunitary","Rank-6 Galois (012)(345): symmetry leaves no room for unitary MTCs","Nonunitarity forced: every rank-6 MTC with Galois group (012)(345) classified","Two families, no unitary physics: rank-6 Galois (012)(345) classification"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification stands on Lemma 3.2's proof of irreducibility, which is checked by 35 computer algebra calculations whose setup is not fully reproduced in the paper and whose printed tables contain repeated 'Sign Choice 15' headers; if any of those calculations is incomplete or mislabeled, the proof of nondegeneracy, and with it the whole classification, fails.","fun_headline_variants_meta":{"raw":{"variants":["Two families exhaust rank-6 modular data with (012)(345), both nonunitary","Rank-6 Galois (012)(345): symmetry leaves no room for unitary MTCs","Nonunitarity forced: every rank-6 MTC with Galois group (012)(345) classified","Two families, no unitary physics: rank-6 Galois (012)(345) classification"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001179,"raw_usage":{"total_tokens":4902,"prompt_tokens":1005,"completion_tokens":3897,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":621,"completion_tokens_details":{"reasoning_tokens":3797}},"tokens_in":621,"tokens_out":3897,"duration_ms":27334,"temperature":1.0,"reasoning_tokens":3797,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:25:06.269148+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the 35 Gr\\\"obner basis computations with the same sign choices and degeneracy cases using a computer algebra system: if any computation returns a non-unit ideal, Lemma 3.2's irreducibility claim is false. Alternatively, construct or find a rank 6 non-integral self-dual MTC with this Galois group whose modular data is not conjugate to the two displayed pairs.","supporting_citations":[{"cited_title":"Bruillard, S","cited_arxiv_id":null,"evidence_quote":"Supplies the admissibility conditions for modular data and the rank-finiteness theorem that justifies classification by rank."},{"cited_title":"Bruillard, S","cited_arxiv_id":null,"evidence_quote":"Provides the Galois symmetry theorem, including the permutation action and the relation $\\sigma^2(t_i)=t_{\\sigma(i)}$, and the divisibility result forcing 7 or 9 to divide $N$."},{"cited_title":"Creamer, A computational approach to classifying modular tensor cat egories, Texas A&M University, 2018","cited_arxiv_id":null,"evidence_quote":"Supplies the list of possible Galois groups for rank 6 and the Gr\\\"obner basis algorithm used, essentially unchanged, in Lemma 3.2."},{"cited_title":"Eholzer, On the classification of modular fusion algebras Comm","cited_arxiv_id":null,"evidence_quote":"Provides the irreducibility criterion via nondegenerate t-spectra, the table of $\\mathrm{SL}(2,\\mathbb{Z}/7\\mathbb{Z})$ representations, and the type $B_9$ fusion algebra."},{"cited_title":"Etingof, On Vafas theorem for tensor categories","cited_arxiv_id":null,"evidence_quote":"Gives the norm bound $N \\mid D^5$ used to show the character is trivial in the $9\\mid N$ case."},{"cited_title":"M¨ uger,On the structure of modular categories","cited_arxiv_id":null,"evidence_quote":"Supplies the decomposition theorem used in Lemma 3.3 to split a fusion-ring product into modular factors."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Classifies pre-modular categories of rank 3, used in Lemma 3.3 to eliminate candidate fusion rules."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Constructs the quantum-group categories realizing the $9\\mid N$ modular data as subcategories of $\\mathcal{C}(\\mathfrak{so}_5,9,e^{j\\pi i/9})$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Classifies modular categories through rank 4 and gives the rank 2 and 3 classifications used to identify the $7\\mid N$ solution."}],"review_version":1}