Classification of Rank 6 Modular Categories with Galois Group $\langle (012)(345)\rangle$

David Green

REVIEW 2 major objections 5 minor 14 references

Classification of Rank 6 Modular Categories with Galois Group $\langle (012)(345)\rangle$

T0 review · 2 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Two families exhaust rank-6 modular data with Galois group (012)(345)

desk verdict 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. read the letter →

arxiv 1908.07128 v2 pith:VGZ2P4T5 submitted 2019-08-20 math.QA

David Green This is my paper · ORCID

classification math.QA MSC 18M2018M1581R50

keywords modulartensorcategoriesrank6classificationGaloisgroupdataGrobnerbasisnonunitarizablequantumgroupsfusionrules

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

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.

What carries the argument

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.

What would settle it

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.

Watch

Extended reading notes

Core claim

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.

Load-bearing premise

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.

Editorial extensions

If this is right

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.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

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.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the 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.

The 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.

The 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".

I 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.

This 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.

Referee Report

2 major / 5 minor

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.

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 (2)

[Section 6, Sign Choice 1, Cases 3–5] 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 6, general] 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.

minor comments (5)

[Section 6, Sign Choice 7] 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.
[Proof of Theorem 3.5] The equation 'r1r2 + r2r3 + r1r2 = −3' should presumably read 'r1r2 + r2r3 + r1r3 = −3.'
[Section 6, final relations] 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.
[References] Reference [11] is missing the author name in the bibliography.
[Abstract] 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.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the modular-data classification is derived from Galois symmetry, Eholzer's external classification, and Gröbner elimination, with independent constructions used only for realization.

full rationale

The central derivation is self-contained. Lemma 3.2 establishes irreducibility by reducing nondegeneracy of t-spectra to thirty-five Gröbner basis computations printed in Section 6, and Theorem 3.5 then uses Eholzer's external classification [7] of irreducible SL(2,Z) representations together with Galois symmetry to construct the two families of (S,T). No parameter is fitted to the target modular data and no target statement is used as an input: the realized categories in [12] and [13] are cited only after the data have been derived. The reliance on Creamer's thesis [4] for the surrounding list of possible rank-6 Galois groups is background context for the broader program, not an ingredient that reappears as the theorem's conclusion. I note a separate, non-circular correctness risk: the Section 6 tables appear to contain internal inconsistencies (for example, Sign Choice 1, Cases 3-5 add factors forcing θ2=-1 or θ1=-1 while claiming reduction to Case 1, which is defined by θ0=θ1=θ2=1), and the repeated 'Sign Choice 15' headers obscure which sign choice is being eliminated. These issues undermine the printed support for Lemma 3.2, but they are proof-completeness defects, not circular reasoning.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The paper does not fit free parameters or invent entities. It imports the admissible data conditions, the Galois symmetry theorem, the rank 2/3 classifications, Eholzer's representation tables, and Creamer's Galois group list as background. The main non-transparent inputs are the unpublished Creamer thesis and the Macaulay2 computations.

assumptions (5)

standard math Admissible modular data conditions of Bruillard-Ng-Rowell-Wang (Definition 2.2) hold for realizable (S,T).
Assumed throughout as the definition of admissible modular data; used to set up Grobner basis ideals in Section 6.
standard math Galois symmetry theorem: for each sigma in Gal(C), sigma permutes columns of tilde S, sigma^2(t_i)=t_{sigma(i)}, and S transforms up to a sign function (Theorem 2.5 from [3]).
This is the core symmetry enabling the reduction to sign choices and degeneracy cases in Lemma 3.2.
domain assumption The list of possible Galois groups for rank 6 non-self-dual and self-dual non-integral cases from Creamer's thesis [4] is complete.
The paper assumes this list to restrict attention to the Galois group (012)(345); if the list is incomplete, the classification statement is incomplete.
standard math Eholzer's classification [7] of modular fusion algebras: in the 9|N case the only irreducible representation (up to character) is of type B9, and Table 12 lists the 7|N representation.
The proof branches on these two representation-theoretic facts from [7] to identify potential fusion rules.
standard math All rank 2 and rank 3 MTCs are classified as in [13] (Theorem 3.2), used in the 7|N branch.
Used to conclude the first modular data family is the only solution after factoring.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Classification of Rank 6 Modular Categories with Galois Group $\langle (012)(345)\rangle$." pith.science (2026). https://pith.science/paper/VGZ2P4T5

@misc{pith2026190807128,
  author       = {Pith},
  title        = {Pith review of: Classification of Rank 6 Modular Categories with Galois Group $\langle (012)(345)\rangle$},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VGZ2P4T5}},
  note         = {Machine review of arXiv:1908.07128}
}

abstract

Modular Tensor Categories (MTC's) arise in the study of certain condensed matter systems. There is an ongoing program to classify MTC's of low rank, up to modular data. We present an overview of the methods to classify modular tensor categories of low rank, applied to the specific case of a rank 6 category with Galois group $\langle(012)(345)\rangle$, and show that certain symmetries in this case imply nonunitarizable (hence, nonphysical) MTC's. We show that all the rank 6 MTC's with this Galois group have modular data conjugate to either the product of the semion category with $(A_1, 5)_{\frac{1}{2}}$ or a certain modular subcategory of $\mathcal{C}(\mathfrak{so}_5, 9, e^{j\pi i/9})$ with gcd$(18, j) = 1$.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[12]

E. C. Rowell. From quantum groups to unitary modular tensor categories , Contemp. Math., (2006), 413, 215–230

work page 2006
[4]

Creamer, A computational approach to classifying modular tensor cat egories, Texas A&M University, 2018

D. Creamer, A computational approach to classifying modular tensor cat egories, Texas A&M University, 2018

work page 2018
[1]

Bruillard, S

P. Bruillard, S. H. Ng, E. Rowell, J. Plavnik, and Z. Wang, On the classiﬁcation of weakly integral modular categories, J. Pure. Appl. Algebra (2014), 220-226

work page 2014
[2]

Bruillard, S

P. Bruillard, S. H. Ng, E. Rowell, and Z. Wang, Rank-ﬁniteness for modular categories , J. Amer. Math. Soc. 29 (2016), 857–881

work page 2016
[3]

Bruillard, S

P. Bruillard, S. H. Ng, E. Rowell, and Z. Wang, On classiﬁcation of modular categories by rank, Int. Math. Res. Notices (2016)

work page 2016
[5]

Deligne, Categories tensorielles , Moscow Math

P. Deligne, Categories tensorielles , Moscow Math. Journal 2 (2002) no. 2, 227-248

work page 2002
[6]

Drinfeld, S

V. Drinfeld, S. Gelaki, D. Nikshych, V. Ostrik, On Braided Fusion Categories I ,Selecta Math., New Ser. 16 (1) (2010) 1-119

work page 2010
[7]

Eholzer, On the classiﬁcation of modular fusion algebras Comm

W. Eholzer, On the classiﬁcation of modular fusion algebras Comm. Math. Phys. 172 (1995), no.3 623–659

work page 1995

Show all 14 references

[8]

Etingof, On Vafas theorem for tensor categories

P. Etingof, On Vafas theorem for tensor categories . Mathematical Research Letters 9, no 5, 651–657
[9]

M¨ uger,On the structure of modular categories

M. M¨ uger,On the structure of modular categories. Proc. London Math. Soc. (3) 87 (2003), no. 2, 291–308

2003
[10]

Ostrik, Fusion categories of rank 2 , Math

V. Ostrik, Fusion categories of rank 2 , Math. Res. Lett. 10 (2003), no. 2-3, 177-183

2003
[11]

, Pre-modular categories of rank 3 , Mosc. Math. J. 8 (2003), no. 1, 111-118

2003
[13]

E. C. Rowell, R. Stong, and Z. Wang. On Classiﬁcation of Modular Tensor Categories , Commun. Math. Phys. (2009) 292–343

2009
[14]

Schopieray, Classiﬁcation of sl3 Relations in the Witt Group of Nondegenerate Braided Fusion Categories, Comm

A. Schopieray, Classiﬁcation of sl3 Relations in the Witt Group of Nondegenerate Braided Fusion Categories, Comm. Math. Phys. 353 (2017),no. 3, 1103-1127 E-mail address : green.2116@buckeyemail.osu.edu Department of Mathematics, The Ohio State University, Colu mbus, Ohio, U.S.A

2017

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [12]

E. C. Rowell. From quantum groups to unitary modular tensor categories , Contemp. Math., (2006), 413, 215–230

work page 2006

[2] [4]

Creamer, A computational approach to classifying modular tensor cat egories, Texas A&M University, 2018

D. Creamer, A computational approach to classifying modular tensor cat egories, Texas A&M University, 2018

work page 2018

[3] [1]

Bruillard, S

P. Bruillard, S. H. Ng, E. Rowell, J. Plavnik, and Z. Wang, On the classiﬁcation of weakly integral modular categories, J. Pure. Appl. Algebra (2014), 220-226

work page 2014

[4] [2]

Bruillard, S

P. Bruillard, S. H. Ng, E. Rowell, and Z. Wang, Rank-ﬁniteness for modular categories , J. Amer. Math. Soc. 29 (2016), 857–881

work page 2016

[5] [3]

Bruillard, S

P. Bruillard, S. H. Ng, E. Rowell, and Z. Wang, On classiﬁcation of modular categories by rank, Int. Math. Res. Notices (2016)

work page 2016

[6] [5]

Deligne, Categories tensorielles , Moscow Math

P. Deligne, Categories tensorielles , Moscow Math. Journal 2 (2002) no. 2, 227-248

work page 2002

[7] [6]

Drinfeld, S

V. Drinfeld, S. Gelaki, D. Nikshych, V. Ostrik, On Braided Fusion Categories I ,Selecta Math., New Ser. 16 (1) (2010) 1-119

work page 2010

[8] [7]

Eholzer, On the classiﬁcation of modular fusion algebras Comm

W. Eholzer, On the classiﬁcation of modular fusion algebras Comm. Math. Phys. 172 (1995), no.3 623–659

work page 1995