CONTENTS
AbstractThe Kashiwara-Vergne method reduces the proof of a deep result in analysis on a Lie group (transferring convolution of invariant distributions from the group to its Lie algebra, by means of the exponential mapping) to checking two formal Lie brackets identities linked to the Campbell-Hausdor¤ formula. First, we expound this method and a proof, for quadratic or solvable Lie algebras, of the Kashiwara-Vergne conjecture allowing to apply the method to those cases. We then extend it to a general symmetric space S = G=H. This leads to introduce a function e(X; Y ) of two tangent vectors X; Y at the origin of S, allowing to make explicit, in the exponential chart, G-invariant di¤erential operators of S, the structure of the algebra of all such operators, and an expansion of mean value operators and spherical functions. For Riemannian symmetric spaces of the noncompact type, otherwise well-known from the work of Harish-Chandra and Helgason, we compare this approach with the classical one. For rank one spaces (the hyperbolic spaces), we give an explicit formula for e(X; Y ).Finally, we explain a construction of e for a general symmetric space by means of Lie series linked to the Campbell-Hausdor¤ formula, in the spirit of the original Kashiwara-Vergne method. Proved from this construction, the main properties of e thus link the fundamental tools of H-invariant analysis on a symmetric space to its in…nitesimal structure.The results extend to line bundles over symmetric spaces.
PrefaceLet G=K denote a Riemannian symmetric space of the noncompact type.« The action of the G-invariant di¤ erential operators on G=K on the radial functions on G=K is isomorphic with the action of certain di¤ erential operators with constant coe¢ cients. The isomorphism in question is used in Harish-Chandra's work on the Fourier analysis on G and is related to the Radon transform on G=K.For the case when G is complex a more direct isomorphism of this type is given, again as a consequence of results of HarishChandra.» This is a quote from Sigurður Helgason, Fundamental solutions of invariant di¤ erential operators on symmetric spaces, Amer. J. Maths. 86 (1964), p. 566, one of the …rst mathematical papers I have studied. As a consequence he showed that every (non-zero) G-invariant di¤erential operator D on G=K has a fundamental solution, whence the solvability of the di¤erential equation Du = f . Those results were fascinating to me, they still are and they fostered my interest in invariant di¤erential operators as well as Radon transforms. Yet I was dreaming simpler proofs could be given, without relying on HarishChandra's deep study of semisimple Lie groups... In the beginning of 1977 Michel Du ‡o [18] gave an analytic interpretation of the celebrated Du ‡o isomorphism he had exhibited six years before by means of algebraic constructions in an enveloping algebra. As a consequence he proved that every (non-zero) bi-invariant di¤erential operator on a Lie group has a local fundamental solution and is locally solvable. His work us...