certainty is), one part of them cannot lie within the bounds of possible experience and another outside all possible experience.

Consequently, since merely empirical grounds of proof lead to nothing supersensible, and since what is lacking in the series of them cannot in any way be completed, we do not approach in the least nearer in our attempt to attain by their means to the supersensible and to a cognition thereof. Thus in any judgement about the latter by means of arguments derived from experience, probability has no place.

(4.) If an hypothesis is to serve for the explanation of the possibility of a given phenomenon, at least its possibility must be completely certain. It is sufficient that in an hypothesis I disclaim any cognition of actuality (which is