Dunkl operators are parameterized differential-difference operators on ޒ N that are related to finite reflection groups. They can be regarded as a generalization of partial derivatives and play a major role in the study of Calogero᎐Moser᎐Sutherland-type quantum many-body systems. Dunkl operators lead to generalizations of various analytic structures, like the Laplace operator, the Fourier transform, Hermite polynomials, and the heat semigroup. In this paper we investigate some probabilistic aspects of this theory in a systematic way. For this, we introduce a concept of homogeneity of Markov processes on ޒ N that generalizes the classical notion of processes with independent, stationary increments to the Dunkl setting. This includes analogues of Brownian motion and Cauchy processes. The generalizations of Brownian motion have the cadlag property and form, after symmetriza-`tion with respect to the underlying reflection groups, diffusions on the Weyl chambers. A major part of the paper is devoted to the concept of modified moments of probability measures on ޒ N in the Dunkl setting. This leads to several Ž . results for homogeneous Markov processes in our extended setting , including martingale characterizations and limit theorems. Furthermore, relations to generalized Hermite polynomials, Appell systems, and Ornstein᎐Uhlenbeck processes are discussed.