So much for the consistency of the extracts themselves. I now turn to a criticism on a different basis. The difficulties connected with this inconsistent doctrine naturally arise out of the standpoint occupied by the author. He accepts as final, and as justifiable in metaphysics, the convenient psychological assumption that the group of sensations gained from an object is a something distinct from the object itself; that the object may be external to the organism, but that the mind, with its sensations, is, or may be treated as if it were, somewhere within the organism; that the sensations are gained from real things, but are not themselves real things, so that a world of sensations--things being abstracted--must be an unreal and phantom world. Now, the man who has thus distinguished between things and sensations, if he regard the sensations as our only immediate representatives of the things, will find it difficult, without making an evidently gratuitous assumption somewhere, to prove his right to reach things at all. Dr. McCosh sees this difficulty, and so he assumes that consciousness reveals both sensations and things. He allows us "perceptions mingled with sensations."[71] Where are these mingled perceptions and sensations? In the mind. Where is the mind? In the body. In what body? The body perceived. Is this body perceived itself in the mind and mingled with sensations? No. It is then distinct from the mental percept--the perception of it is somewhere in it, but is not it. How do we know, then, that there is a body? We infer it from the percept; consciousness (the percept) "reveals" it. On what principle is it inferred? The question is a just one if the knowledge be not immediate. Our author does not even see that there is a question. The fact is that this doctrine seems to avoid the difficulties of a representative perception only while it is allowed to remain loose and vague. If things are not to be known representatively, they must either be themselves in consciousness--_and then they are not extra-mental_--or they must be directly known in some other way than as in consciousness, _and then consciousness does not reveal them and cannot be appealed to_. An appeal to consciousness, unless the thing itself is in consciousness, is fatal. [71] Chap. III, p. 75.