Since brownian motion can't be defined on a differentiable manifold without assigning on it a riemannian metric by which it depends, Guerra's and coworkers' approach to the definition of the stochastic action in the framework of Stochastic Mechanics is completelly inconsistent