The philosophy is fascinating. Symplectic G-manifolds correspond to classical mechanical systems with G symmetries. Unitary representations of G correspond to quantum mechanical systems with G symmetries. Since g* is the final object in the category of Poisson G-manifolds, we should expect its space of leaves in the symplectic foliation by coadjoint orbits to parametrize irreducible unitary representations IF a sufficiently good notion of "quantization" of a classical mechanical system is possible.