We shall find that these superstitions, though often held by whole tribes with the utmost assurance, differ in some subtle way from the perception-beliefs of their common sense, as that “fire burns” and that “water quenches fire.” They are unstable: (1) they become active on occasions, and otherwise are apt to be forgotten—as ghosts are only thought of at night. (2) They are modifiable merely for the sake of economy or other convenience. (3) They lose their hold on a tribe, fall off and die in course of time without any change in the evidence for them. (4) They depend a good deal upon the assent of a crowd. (5) They often vary in neighbouring countries or families, or amongst the members of a family. This is not like common sense. Superstitions or imagination-beliefs are unstable, in spite of being often held with great obstinacy (so that people die for them), and of their enduring, in the simpler forms and at a certain level of social life, for thousands of years. There is something wanting in the holdfast or anchorage of imagination-beliefs. It is necessary to explain what I mean by “imagination.” § 2. IMAGINATION Is it enough to define “imagination” as merely the having of mental “images,” pictures before the mind’s eye? This would confine imagination to visual representations, to the exclusion of auditory, olfactory, etc., which are all a man born blind can have, and which sometimes occur to those who can see, though the visual are commonest. The word “images,” therefore, is sometimes used to cover all these modes of representation; though “phantasmata” would be better. Again, a mental image or phantasm, visual or auditory, is improperly called an imagination, if there is nothing more than the reproduction of a single sense-quality. Imaginations represent not abstract sensations, but perceptions. To see an armed knight is not merely to have a visual impression of him, but to perceive a living, solid, heavy object definitely in space; and imagination reproduces the whole of this, and otherwise would be quite uninteresting. What would the tournament in _Ivanhoe_ amount to if the knights were only phantoms? Further, imagination, merely as a reproduction of perception, is not distinguished from memory; but, in use, the two are always contrasted. Memories are recognised (in their complete form) as returning to us from earlier experience, both their component pictures and the order of them, and they are relatively stable; imaginations are felt to be more or less novel, and can easily be modified. Probably all the elements of an imagination might have occurred in a memory; but the arrangement of these elements is often so different from any actual experience as to baffle every attempt to redistribute them amongst their sources. Hence, in normal cases, our attitudes toward a memory and toward an imagination are entirely different. If a seeming memory prove false, we say it was only an imagination.