This paper criticises a line of argument adopted by peter winch, Karl popper, And others, To the effect that the course of human history cannot be predicted. On this view it is impossible to predict in a particularly detailed way certain events ('original acts') on which important social developments depend. We analyze the argument, Showing that one version fails: original acts are in principle predictable in the relevant way. A cogent version is presented; this requires a special definition for 'original act'. But, We claim, Social developments do not depend on original acts so defined. We argue separately for the possibility of a person, Or a scientific community, Predicting his or its own original acts.