Observe, then, that the particular difficulty which has been occupying our attention arises in the main from the assumption that our common-sense beliefs in the reality and character of material things have no other foundation than the fact that we so perceive them. From such premises it was impossible, it seemed, to infer that they exist otherwise than as they are perceived; and still more impossible to regard the immediate intuition by which we apprehend the object, and the long-drawn sequence of causes by which the object is revealed, as being the same process looked at from different ends. But this difficulty is greatly mitigated if we hold that our belief in an independent world of material objects, however it may be caused, is neither a conclusion drawn from this or that particular experience nor from all our experiences put together, but an irresistible assumption. Grant the existence of external things, and it becomes possible and legitimate to attempt explanations of their appearance, to regard our perceptions of them as a psychical and physiological product of material realities which do not themselves appear and cannot be perceived. Refuse, on the other hand, to grant this assumption, and no inductive legerdemain will enable us to erect our scientific theories about an enduring world of material things upon the frail foundation of successive personal perceptions. If this does not seem clear at first sight it is, I think, because we do not consider our experiences as a whole. A limited group of experiences—say Faraday’s experiments with electro-magnets—may guide us into new knowledge about the external world, including aspects of that world which are not open to sense perception. But then these experiences assume that this external world exists, they assume it to be independent of perception, they assume it to be a cause of perception. These assumptions once granted, experiment may be, and is, the source of fresh discoveries. But experiment based on these assumptions never can establish their truth; and if our theory of knowledge requires us to hold that “no proposition should be entertained with greater assurance than the proofs it is built on will warrant,” our fate is sealed, and we need never hope to extricate ourselves from the entanglements in which a too credulous empiricism has involved us. This means that one at least of the inevitable beliefs enumerated in the first lecture—the belief in an external world—is a postulate which science is compelled to use but is unable to demonstrate. How, then, are we to class it? It is not a law of thought in the accepted meaning of that expression. We are not rationally required to accept it by the very structure of our thinking faculties. Many people, indeed, theoretically reject it; none, so far as I know, regard it as self-evident. On the other hand, it is not an inference from experience; neither is it an analytic judgment in which the predicate is involved in the subject.