arXiv subjects
Publications and source records attributed to Terry Gannon.
This paper is primarily intended as an introduction for the mathematically inclined to some of the rich algebraic combinatorics arising in for instance CFT. It is essentially self-contained, apart from some of the background motivation and examples which are included to give the reader a sense of the context. The theory is still a work-in-progress, and emphasis is given here to several open questions and problems.
In these notes, based on lectures given in Istanbul, we give an introduction both to Monstrous Moonshine and to the classification of rational conformal field theories, using this as an excuse to explore several related structures and go on a little tour of modern math. We will discuss Lie algebras, modular functions, the finite simple group classification, vertex operator algebras, Fermat's Last Theorem, category theory, (generalised) Kac-Moody algebras, denominator identities, the A-D-E meta-pattern, representations of affine algebras, Galois theory, etc. This work is informal and pedagogical, and aimed mostly at grad students in math or math phys, but I hope that many interested nonexperts will find something of value here -- like any good Walt Disney movie I try not to completely ignore the `grown-ups'. My emphasis is on ideas and motivations, so these notes are intended to complement other papers and books where this material is presented with more technical detail. The level of difficulty varies significantly from topic to topic. The two parts -- in fact any of the sections -- can be read independently.
This paper finds for all orthogonal algebras (i.e. the B and D series) all modular invariant 1-loop partition functions at levels 1,2,3. Previously, only those at level 1 were classified. An extraordinary number of exceptionals appear at level 2 -- indeed this is the primary motivation for this paper -- and we find infinitely many new ones there. The only level 3 exceptionals occur for $B_2$ and $D_7$, and the latter appear to be new. The $B_{2,3}$ and $D_{7,3}$ exceptionals are cousins of the "$E_6$" and "$E_8$" exceptionals in the A-D-E classification for affine $A_1$, while the level 2 exceptionals are related to the lattice invariants of affine u(1).
The problem of finding all possible effective field theories for the quantum Hall effect is closely related to the problem of classifying all possible modular invariant partition functions for the algebra $u(1)^m$, as was argued recently by Cappelli and Zemba. This latter problem is also a natural one from the perspective of conformal field theory. In this paper we completely solve this problem, expressing the answer in terms of self-dual lattices, or equivalently, rational points on the dual Grassmannian $G_{m,m}(R)^*$. We also find all modular invariant partition functions for $su(2)\oplus u(1)^m$, from which we obtain the classification of all N=2 superconformal minimal models. The `A-D-E classification' of these, though often quoted in the literature, turns out to be a very coarse-grained one: e.g. associated with the names $E_6,E_7,E_8$, respectively, are precisely 20,30,24 different partition functions. As a by-product of our analysis, we find that the list of modular invariants for su(2) lengthens surprisingly little when commutation with T -- i.e. invariance under $τ\mapsto τ+1$ -- is ignored: the other conditions are far more essential.
In this paper we explicitly classify all modular invariant partition functions for su(n) at level 2 and 3. Previously, these were known only for level 1. The level 2 exceptionals exist at n=10, 16, and 28; the level 3 exceptionals exist at n=5, 9, and 21. One of these is new, but the others were all anticipated by the "rank-level duality" relating su(n) level k and su(k) level n. The main recent result which this paper rests on is the classification of "ADE_7-type invariants".
The classification of CFTs has an important subproblem, namely classifiying the partition functions for WZW theories. This subproblem is intimately connected to the modular behaviour of the characters of affine algebras. This paper addresses the classification of the largest class of these -- the natural generalization to any affine algebra of the A-series, D-series, and $E_7$-exceptional appearing in the A-D-E classification of Cappelli-Itzykson-Zuber for $A_1^{(1)}$. We give a programme for classifying these for any algebra, and explicitly do this for the most interesting case: $A_r^{(1)}$ at any level k. The exceptionals appear at (r,k)=(1,16), (2,9), (3,8), (4,5), (7,4), (8,3) and (15,2).
The characters $χ_μ$ of nontwisted affine algebras at fixed level define in a natural way a representation $R$ of the modular group $SL_2(Z)$. The matrices in the image $R(SL_2(Z))$ are called the Kac-Peterson modular matrices, and describe the modular behaviour of the characters. In this paper we consider all levels of $(A_{r_1}\oplus\cdots\oplus A_{r_s})^{(1)}$, and for each of these find all permutations of the highest weights which commute with the corresponding Kac-Peterson matrices. This problem is equivalent to the classification of automorphism invariants of conformal field theories, and its solution, especially considering its simplicity, is a major step toward the classification of all Wess-Zumino-Witten conformal field theories.
In this paper we find all permutations of the level k weights of the affine algebra A_r^{(1)} which commute with both its S and T modular matrices. We find that all of these are simple current automorphisms and their conjugations. Previously, the A_{r,k}^{(1)} automorphism invariants were known only for r=1,2 \forall k, and k=1 \forall r. This is a major step toward the full classification of all A_{r,k}^{(1)} modular invariants; the simplicity of this proof strongly suggests that the full classification should be accomplishable. In an appendix we collect some new results concerning the A_{r,k}^{(1)} fusion ring.
We relate in a novel way the modular matrices of GKO diagonal cosets without fixed points to those of WZNW tensor products. Using this we classify all modular invariant partition functions of $su(3)_k\oplus su(3)_1/su(3)_{k+1}$ for all positive integer level $k$, and $su(2)_k\oplus su(2)_\ell/su(2)_{k+\ell}$ for all $k$ and infinitely many $\ell$ (in fact, for each $k$ a positive density of $\ell$). Of all these classifications, only that for $su(2)_k\oplus su(2)_1/su(2)_{k+1}$ had been known. Our lists include many new invariants.
The SU(3) modular invariant partition functions were first completely classified in Ref.\ \SU. The purpose of these notes is four-fold: \item{(i)} Here we accomplish the SU(3) classification using only the most basic facts: modular invariance; $M_{\laμ}\in{\bf Z}_{\ge}$; and $M_{00}=1$. In \SU{} we made use of less elementary results from Moore-Seiberg, in addition to these 3 basic facts. \item{(ii)} Ref.\ \SU{} was completed well over a year ago. Since then I have found a number of significant simplifications to the general argument. They are all included here. \item{(iii)} A number of people have complained that some of the arguments in \SU{} were hard to follow. I have tried here to be as explicit and as clear as possible. \item{(iv)} Hidden in \SU{} were a number of smaller results which should be of independent value. These are explicitly mentioned here.
The complete classification of WZNW modular invariant partition functions is known for very few affine algebras and levels, the most significant being all levels of $A_1$ and $A_2$ and level 1 of all simple algebras. Here, we address the classification problem for the nicest high rank semi-simple affine algebras: $(A_1^{(1)})^{\oplus_r}$. Among other things, we explicitly find all automorphism invariants, for all levels $k=(k_1,\ldots,k_r)$, and complete the classification for $A_1^{(1)}\oplus A_1^{(1)}$, for all levels $k_1,k_2$. We also solve the classification problem for $(A_1^{(1)})^{\oplus_r}$, for any levels $k_i$ with the property that for $i\ne j$ each $gcd(k_i+2,k_j+2)\leq 3$. In addition, we find some physical invariants which seem to be new. Together with some recent work by Stanev, the classification for all $(A^{(1)}_1)^{\oplus_r}_k$ could now be within sight.
Using the self-dual lattice method, we make a systematic search for modular invariant partition functions of the affine algebras $g\*{(1)}$ of $g=A_2$, $A_1+A_1$, $G_2$, and $C_2$. Unlike previous computer searches, this method is necessarily complete. We succeed in finding all physical invariants for $A_2$ at levels $\le 32$, for $G_2$ at levels $\le 31$, for $C_2$ at levels $\le 26$, and for $A_1+A_1$ at levels $k_1=k_2\le 21$. This work thus completes a recent $A_2$ classification proof, where the levels $k=3,5,6,9,12,15,21$ had been left out. We also compute the dimension of the (Weyl-folded) commutant for these algebras and levels.
A complete classification of the WZNW modular invariant partition functions is known for very few affine algebras and levels, the most significant being all levels of SU(2), and level 1 of all simple algebras. In this paper we solve the classification problem for SU(3) modular invariant partition functions. Our approach will also be applicable to other affine Lie algebras, and we include some preliminary work in that direction, including a sketch of a new proof for SU(2).
Thus far in the search for, and classification of, `physical' modular invariant partition functions $\sum N_{LR}\,\c_L\,\C_R$ the attention has been focused on the {\it symmetric} case where the holomorphic and anti-holomorphic sectors, and hence the characters $\c_L$ and $\c_R$, are associated with the same Kac-Moody algebras $\g_L=\g_R$ and levels $k_L=k_R$. In this paper we consider the more general possibility where $(\g_L,k_L)$ may not equal $(\g_R,k_R)$. We discuss which choices of algebras and levels may correspond to well-defined conformal field theories, we find the `smallest' such {\it heterotic} (\ie asymmetric) partition functions, and we give a method, generalizing the Roberts-Terao-Warner lattice method, for explicitly constructing many other modular invariants. We conclude the paper by proving that this new lattice method will succeed in generating all the heterotic partition functions, for all choices of algebras and levels.
A natural first step in the classification of all `physical' modular invariant partition functions $\sum N_{LR}\,\c_L\,\C_R$ lies in understanding the commutant of the modular matrices $S$ and $T$. We begin this paper extending the work of Bauer and Itzykson on the commutant from the $SU(N)$ case they consider to the case where the underlying algebra is any semi-simple Lie algebra (and the levels are arbitrary). We then use this analysis to show that the partition functions associated with even self-dual lattices span the commutant. This proves that the lattice method due to Roberts and Terao, and Warner, will succeed in generating all partition functions. We then make some general remarks concerning certain properties of the coefficient matrices $N_{LR}$, and use those to explicitly find all level 1 partition functions corresponding to the algebras $B_n$, $C_n$, $D_n$, and the 5 exceptionals. Previously, only those associated to $A_n$ seemed to be generally known.
We prove that a class of one-loop partition functions found by Dienes, giving rise to a vanishing cosmological constant to one-loop, cannot be realized by a consistent lattice string. The construction of non-supersymmetric string with a vanishing cosmological constant therefore remains as elusive as ever. We also discuss a new test that any one-loop partition function for a lattice string must satisfy.