ABSTRACT
In seeking the postulates of scientific inference there are two kinds of problems. On the one hand there is analysis of what is generally accepted as valid inference, with a view to discovering what principles are involved; this inquiry is purely logical. On the other hand there is the difficulty that there is, prima facie, little reason to suppose these principles true, and still less to suppose them known to be true. I think that the question in what sense, if any, these principles can be known, requires an analysis of the concept of “knowledge”. This concept is too often treated as though its meaning were obvious and unitary. My own belief is that many philosophical difficulties and controversies arise from insufficient realization of the difference between different kinds of knowledge, and of the vagueness and uncertainty that characterizes most of what we believe ourselves to know. There is another thing which it is important to remember whenever mental concepts are being discussed, and that is our evolutionary continuity with the lower animals. “Knowledge”, in particular, must not be defined in a manner which assumes an impassable gulf between ourselves and our ancestors who had not the advantage of language.
