On the Feynman proof of the exact relation between a class of path sums and the general nonlinear Fokker-Planck equation