This paper defends the view that a correct analysis of knowledge must take empirical data into consideration. The data here provided is from experimental psychology, namely from phenomena involving unconscious cognition.