Abstract

Predictiveness (or testability) can be ascribed to empirical hypotheses in different degrees. Seven reasons why predictiveness is gradable are outlined. They are reflected in seven ways in which hypotheses can be compared according to the degree of their predictiveness. Hence additional specifications are required to accept predictiveness as a criterion of whether given hypotheses are scientific. Generally speaking a new hypothesis has little, if any, predictive power. Its predictive power will increase as science progresses. Specific problems arise when the predictive power is expected from social sciences