Abstract. The maximal rank hyperbolic Kac-Moody algebra e10 has been conjectured to play a prominent role in the unification of duality symmetries in string and M theory. We review some recent developments supporting this conjecture.
IntroductionHidden symmetries of exceptional type in the reduction of supergravity theories were first discovered in maximal N = 8 supergravity in d = 4 [1,2]. The unexpected emergence of the coset E 7 /SU (8) describing the scalar sector of the theory was soon generalised to other dimensions and other theories [3]. The most prominent example remains the chain of hidden symmetries occurring in the dimensional reduction of d = 11 maximal supergravity on a torus T n . For 1 ≤ n ≤ 8 the scalars always in appear in group cosets which have become known as E n /K(E n ) where K(E n ) designates the maximal compact subgroup of E n . For n > 8 it was soon conjectured that the resulting symmetry groups become infinite-dimensional [4] and formulations using the centrally extended loop group E 9 [5, 6] and partial results on E 10 [7] have since been obtained.In an initially unrelated development, the study of the asymptotic behaviour of d = 11 supergravity (and IIA and IIB supergravity) near a space-like singularity also revealed evidence for infinite dimensional symmetries, and the hyperbolic Kac-Moody algebra E 10 , in particular. Namely, in this limit, the dynamics can be asymptotically described by a cosmological billiard taking place in the fundamental Weyl chamber of E 10 [8, 9]. The dynamical variables in this case are the spatial scale factors. This led the authors of [10] to propose a one-dimensional nonlinear σ-model based on a coset E 10 /K(E 10 ) and [10,11] uncovered a remarkable dynamical equivalence between a truncation of the bosonic d = 11 supergravity equations and a truncated version of this infinite-dimensional coset model. It is the aim of the present contribution to review this correspondence and similar correspondences to the d = 10 maximal supergravities which were derived in [12,13]. We will not present the relation of higher derivative corrections to the E 10 model which can be found in [14]. Related work in [15,16] discusses the role of imaginary roots of E 10 from a brane point of view and orbifolds.We note that already in [17,18] it was proposed that d = 11 supergravity is a non-linear realisation of the bigger group E 11 (and the conformal group via a Borisov-Ogievetsky-type construction [19]). The non-linear E 11 model is thus supposed to operate directly in eleven (or even more) dimensions, and is therefore very different from the one-dimensional E 10 model which we will present below. In addition, as will be discussed in section 5.3, space-time is thought to