Abstract
A mathematically rigorous discussion of the diffusion equation and of its connection with the Einstein relation linking the diffusion coefficient and the velocity autocorrelation function is presented. Diffusion is then propounded as a typical case for which the logical consistency of a purely mechanistic theory of dissipative phenomena can be established.