We present a general formalism to characterize the probability density func
tion of a set of dynamic variables in a stationary process using conditiona
l expectations of kinematic observables on those variables. The formalism i
s exemplified with stochastic processes such as general Gaussian random pro
cesses and Brownian systems. We show that this formalism gives the Boltzman
n distribution for equilibrium processes as it should and is applicable als
o for out of equilibrium processes. (C) 1999 Elsevier Science B.V. All righ
ts reserved.