Braid Group Representations and Modular Data: Non-Uniqueness, Finite Images, and the Ising Test Case

In the paper's words

A central question in the theory of topological order asks whether the braid group representation B_N carried by N non-Abelian anyons uniquely determines the modular data — the S and T matrices — of the underlying topological quantum field theory. We reconcile three lines of analysis into a single assessment. Structurally, we formalize the question as a reduction map Φ from modular tensor category data to projective braid group representations, whose kernel is the locus of potential counterexamples.

The abstract, verbatim, and its opening sentence is the claim. Every line below is quoted from the text.

Numbers it reports

1. Computed values (upper bound / projection): $N=4$: $2^4\cdot2!=16\cdot2=32$ / $2^3\cdot2=16$. $N=6$: $2^6\cdot3!=64\cdot6=384$ / $2^5\cdot6=192$. $N=8$: $2^8\cdot4!=256\cdot24=6144$ / $2^7\cdot24=3072$.

   from 4. Analysis

2. R4 (deformation family). $\mathrm{Ising}(7)$ shares all fusion rules with Ising but has $\theta_\sigma=e^{7i\pi/8}$ and distinct $S,T$ (its unitarity as a Galois conjugate is not established, and no chiral central charge is assigned to it); its $B_3$ representation is inequivalent to Ising's (trace ratio $e^{-3i\pi/4}$), though projectively intertwined by the parity operator of [13].

   from 5. Results

3. (5) Connect the hexagon-equation analysis [13] and the memory-relaxation model [14] to test whether finite braid images predict a specific error-correction threshold.

   from 6. Discussion

How it is done

1. The metaplectic anyon models of Bombin and collaborators [9], combining an Ising sector with $SO(m)_2$ Chern–Simons theory, provide a complete account of quasiparticle types, fusion rules, and braiding for a family of models; we use the $m=1$ (Ising) member as our worked example.

   from 2. Background and Related Work

2. A unitary modular tensor category $\mathcal{C}$ has simple objects $a,b,\dots$, fusion coefficients $N_{ab}^c\in\mathbb{Z}_{\ge0}$, quantum dimensions $d_a$ satisfying $d_a d_b = \sum_c N_{ab}^c d_c$, total quantum dimension $\mathcal{D}=\sqrt{\sum_a d_a^2}$, topological spins $\theta_a\in U(1)$ (diagonal of $T$), and an $S$ matrix.

   from 3. Methods

3. 4.1 The Ising anyon model.

   from 4. Analysis

4. The claim that braiding is a complete invariant of modular data (advanced in one of the reconciled drafts) holds, if at all, only under additional hypotheses (full set of $R$-symbols on all channels, multiplicity-free fusion, fixed chirality/gauge) and is not established as a general theorem here.

   from 5. Results

What would falsify it

1. Section 6 discusses limitations and falsifiability.

   from 1. Introduction

2. Define $$ \Phi: \{\text{MTC data }(\mathcal{C},S,T)\} \longrightarrow \{\text{projective representations } B_N \to PU(\mathcal{H}_N)\}. $$ The uniqueness conjecture is injectivity of $\Phi$; a counterexample is a pair of distinct MTCs with equivalent projective braid images.

   from 3. Methods

3. The non-uniqueness assessment would be falsified by a proof that $\Phi$ is injective on a class containing the finite-image families.

   from 6. Discussion

Where the claim stops

1. (ii) The finiteness results [2,4] are special to their families; a generic MTC may have infinite braid image, where the finite-image argument does not apply.

   from 6. Discussion

2. (iii) "Finite image" does not by itself prove "non-unique modular data": two finite groups can distinguish rich algebraic data if the representation is faithful enough.

   from 6. Discussion

3. (2) Determine whether $\Phi$ is injective when restricted to infinite-image representations.

   from 6. Discussion

Where to read

1. Abstract258 words · 1 min
2. 1. Introduction291 words · 1 min
3. 2. Background and Related Work615 words · 3 min
4. 3. Methods305 words · 1 min
5. 4. Analysis727 words · 3 min
6. 5. Results218 words · 1 min
7. 6. Discussion271 words · 1 min
8. 7. Conclusion176 words · 1 min
9. Appendix A. Divergence report162 words · 1 min