ABSTRACT
A “probable” inference (to repeat what has already been said) is one in which, when the premisses are true and the reasoning correct, the conclusion is nevertheless not certain but only probable in a greater or less degree. In the practice of science there are two kinds of inferences: those that are purely mathematical, and those that may be called “substantial”. The inference from Kepler’s laws to the law of gravitation as applied to the planets is mathematical, but the inference from the recorded apparent motions of the planets to Kepler’s laws is substantial, for Kepler’s laws are not the only hypotheses logically compatible with observed facts. Mathematical inference has been sufficiently investigated during the last half-century. What I wish to discuss is non-mathematical inference, which is always only probable.
