An Observation on Finite Groups and WZW Modular Invariants

Bo Feng, Yang‐Hui HearXiv 2000 · DOI · arXiv

Abstract

In this short note, inspired by much recent activity centered around attempts to formulate various correspondences between the classification of affine $SU(k)$ WZW modular-invariant partition functions and that of discrete finite subgroups of $SU(k)$, we present a small and perhaps interesting observation in this light. In particular, we show how the groups generated by the permutation of the terms in the exceptional Affine $SU(2)$-WZW invariants encode the corresponding exceptional $SU(2)$ subgroups.

Introduction

In the study of Rational Conformal Field Theory (RCFT), specifically Wess-Zumino-Witten (WZW) models, a striking "ADE meta-pattern" has long been observed. The classification of modular-invariant partition functions for the $SU(2)$ algebra follows the exact same ADE (Algebraic-Differential-Enumerative) classification as simply-laced Lie algebras and the discrete finite subgroups of $SU(2)$.

While this correspondence has been recognized since the late 1980s, the underlying mechanism remains a subject of investigation. Previous attempts to bridge the gap between the modular invariants and the finite subgroups often relied on complex graph-theoretic constructions, such as the McKay Correspondence, or involved cumbersome identities between theta and eta functions. This paper seeks a more direct, group-theoretic observation: identifying whether these finite groups can be seen directly in the behavior of the partition function under modular transformations.

Methodology

The research focuses on the "exceptional" cases of the $SU(2)$ WZW models, specifically those labeled $\mathcal{E}_6, \mathcal{E}_7,$ and $\mathcal{E}_8$. The authors propose a method to extract finite group structures from the modular group generators $S$ and $T$.

Basis Redefinition and Matrix Representation

Standard analysis of WZW models involves a $(k+1)$-dimensional space of characters, where $k$ is the level of the affine algebra. However, the authors restrict their focus to a specific subspace. This subspace is spanned by the unique combinations of characters that appear as terms within the modular-invariant partition function. By representing the generators of the modular group—the $S$ and $T$ matrices—within this reduced basis, the authors can observe how these transformations permute the terms of the invariant.

Group Generation and Quotient Analysis

Using the reduced matrix representation, the authors generate a finite group $G = \langle S, T \rangle$. To analyze the structure of $G$, they utilize the GAP (Groups, Algorithms, and Programming) system to identify normal subgroups $H$. The objective is to determine if a quotient group $G/H$ exists that is isomorphic to the expected finite subgroups of $SU(2)$, namely the binary tetrahedral, octahedral, or icosahedral groups.

Key Insights: Permutation as Encoding

The core insight of this work is that the requirement for modular invariance—which dictates that $S$ and $T$ must permute the terms of the partition function—effectively encodes the structure of the corresponding finite group.

While the group $G$ generated by the modular transformations is typically much larger than the target finite group, the target group is found to be a unique quotient of $G$. For example, in the case of the $\mathcal{E}_6$ invariant, the generated group $G$ has 1152 elements. By quotienting $G$ by a specific normal subgroup of order 48, the binary tetrahedral group is recovered. This relationship appears remarkably unique in the $SU(2)$ context; for each exceptional invariant, only one normal subgroup yields the correct corresponding finite group, and it does not yield any of the others.

Results

The authors successfully mapped the exceptional $SU(2)$ modular invariants to their corresponding finite subgroups through the following quotient relationships:

  • $\mathcal{E}_6$ (Level $k=10$): $G_{1152} / H_{48} \cong \hat{E}_6$ (the binary tetrahedral group).
  • $\mathcal{E}_8$ (Level $k=28$): $G_{720} / H_{6} \cong \hat{E}_8$ (the binary icosahedral group).
  • $\mathcal{E}_7$ (Level $k=16$): This case yielded $G_{324} / H_{27} \cong E_7$ (the ordinary octahedral group). The authors note that $\mathcal{E}_7$ is frequently the "difficult" case in classification schemes, and this slight deviation from the binary group reflects that inherent complexity.

Discussion and Limitations

While the results for $SU(2)$ are compelling, the authors note significant challenges in generalizing this "naive" approach to higher-rank algebras. When applied to $SU(3)$ invariants, the method fails to produce the expected exceptional subgroups. This suggests that the relationship for $SU(3)$ and beyond is significantly more complex, likely requiring a deeper understanding of the truncation of fusion graphs.

Furthermore, while the paper identifies a "pleasant surprise" in the group structure, it does not yet provide a fundamental physical or geometric explanation for why these specific quotients appear. The study concludes that while the permutation of terms in the partition function provides a direct link to the ADE classification, the deeper mechanism driving this correspondence remains an open question for future research.

Summary updated

Abstract

In this short note, inspired by much recent activity centered around attempts to formulate various correspondences between the classification of affine SU(k) WZW modular-invariant partition functions and that of discrete finite subgroups of SU(k), we present a small and perhaps interesting observation in this light. In particular we show how the groups generated by the permutation of the terms in the exceptional Affine SU(2)-WZW invariants encode the corresponding exceptional SU(2) subgroups.
BibTeX
@misc{feng2000observation,
  author = {Bo Feng and Yang‐Hui He},
  title = {{An Observation on Finite Groups and WZW Modular Invariants}},
  year = {2000},
  eprint = {hep-th/0009077},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/hep-th/0009077},
}

Loading related content...

Figure