Missing Modules, the Gnome Lie Algebra, and $E_{10}$

TL;DR

This paper explores the embedding of Kac-Moody algebras into Borcherds algebras realized through string theory, identifying missing modules and roots, and providing explicit constructions for the gnome Lie algebra and $E_{10}$.

Contribution

It introduces explicit realizations of Borcherds algebras as string state Lie algebras and characterizes their imaginary simple roots, including the novel gnome Lie algebra.

Findings

Complete set of imaginary simple roots for gnome Lie algebra $g_{1,1}$.

Explicit basis for $g_{1,1}$ using string states.

Characterization and conjecture of imaginary simple roots for $g_{9,1}$ and $E_{10}$.

Abstract

We study the embedding of Kac-Moody algebras into Borcherds (or generalized Kac-Moody) algebras which can be explicitly realized as Lie algebras of physical states of some completely compactified bosonic string. The extra ``missing states'' can be decomposed into irreducible highest or lowest weight ``missing modules'' w.r.t. the relevant Kac-Moody subalgebra; the corresponding lowest weights are associated with imaginary simple roots whose multiplicities can be simply understood in terms of certain polarization states of the associated string. We analyse in detail two examples where the momentum lattice of the string is given by the unique even unimodular Lorentzian lattice $II_{1,1}$ or $II_{9,1}$, respectively. The former leads to the Borcherds algebra $g_{1,1}$, which we call ``gnome Lie algebra", with maximal Kac-Moody subalgebra $A_1$. By the use of the denominator formula a complete set of imaginary simple roots can be exhibited, whereas the DDF construction provides an explicit Lie algebra basis in terms of purely longitudinal states of the compactified string in two dimensions. The second example is the Borcherds algebra $g_{9,1}$, whose maximal Kac-Moody subalgebra is the hyperbolic algebra $E_{10}$. The imaginary simple roots at level 1, which give rise to irreducible lowest weight modules for $E_{10}$, can be completely characterized; furthermore, our explicit analysis of two non-trivial level-2 root spaces leads us to conjecture that these are in fact the only imaginary simple roots for $g_{9,1}$.