_After-images_ belong to sensation rather than to imagination; so that the most immediate phenomena of imagination would seem to be those tardier images (due to what the Germans call _Sinnesgedächtniss_) which were spoken of in Vol. I, p. 617,--coercive hauntings of the mind by echoes of unusual experiences for hours after the latter have taken place. The phenomena ordinarily ascribed to imagination, however, are those mental pictures of possible sensible experiences, to which the ordinary processes of associative thought give rise. When represented with surroundings concrete enough to constitute a _date_, these pictures, when they revive, form _recollections_. We have already studied the machinery of recollection in Chapter XVI. When the mental pictures are of data freely combined, and reproducing no past combination exactly, we have acts of imagination properly so called. OUR IMAGES ARE USUALLY VAGUE. For the ordinary 'analytic' psychology, each sensibly discernible element of the object imagined is represented by its own separate idea, and the total object is imagined by a 'cluster' or 'gang' of ideas. We have seen abundant reason to reject this view (see Vol. I, p. 276 ff.). An imagined object, however complex, is at any one moment thought in one idea, which is aware of all its qualities together. If I slip into the ordinary way of talking, and speak of various ideas 'combining,' the reader will understand that this is only for popularity and convenience, and he will not construe it into a concession to the atomistic theory in psychology. Hume was the hero of the atomistic theory. Not only were ideas copies of original impressions made on the sense-organs, but they were, according to him, completely adequate copies, and were all so separate from each other as to possess no manner of connection. Hume proves ideas in the imagination to be completely adequate copies, not by appeal to observation, but by _a priori_ reasoning, as follows: