A stochastic interpretation of the so called membership functions of fuzzy logic is presented. The membership functions are used to define a probability space, and the values of them may be interpreted as probabilities. It is believed that the stochastic interpretation will facilitate the analysis and understanding of fuzzy control.