914. Thus far, then, our field is clear, and with this clearance, I think, should vanish the somewhat grim associations which have gathered around the word possession. In what is now to be described there may often be cause for perplexity, but I have never seen cause for fear. Nay, how far remote from fear is the resultant feeling, the sequel will show.

Assuming then, as I think we at present may assume, that we have to deal only with spirits who have been men like ourselves, and who are still animated by much the same motives as those which influence us, we may briefly consider, on similar analogical grounds, what range of spirits are likely to be able to affect us, and what difficulties they are likely to find in doing so. Of course, actual experience alone can decide this; but nevertheless our expectations may be usefully modified if we reflect beforehand how far such changes of personality as we already know can suggest to us the limits of these profounder substitutions.

What, to begin with, do we find to be the case as to addition of faculty in alternating states? How far do such changes bring with them unfamiliar powers?

Reference to the recorded cases will show us that existing faculty may be greatly quickened and exalted. There may be an increase both in actual perception and in power of remembering or reproducing what has once been perceived. There may be increased control over muscular action, - as shown, for instance, in improved billiard-playing, - in the secondary state. But there is little evidence of the acquisition - telepathy apart - of any actual mass of fresh knowledge, - such as a new language, or a stage of mathematical knowledge unreached before. We shall not therefore be justified by analogy in expecting that an external spirit controlling an organism will be able easily to modify it in such a way as to produce speech in a language previously unknown. The brain is used as something between a typewriter and a calculating machine. German words, for instance, are not mere combinations of letters, but specific formulae; they can only seldom and with great difficulty be got out of a machine which has not been previously fashioned for their production.

915. Consider, again, the analogies as to memory. In the case of alternations of personality, memory fails and changes in what seems a quite capricious way. The gaps which then occur recall (as I have said) the ecmnesiae or blank unrecollected spaces which follow upon accidents to the head, or upon crises of fever, when all memories that belong to a particular person or to a particular period of life are clean wiped out, other memories remaining intact. Compare, again, the memory of waking life which we retain in dream. This too is absolutely capricious; - I may forget my own name in a dream, and yet remember perfectly the kind of chairs in my dining-room. Or I may remember the chairs, but locate them in some one else's house. No one can predict the kind of confusion which may occur.

916. We have also the parallel of somnambulic utterance. In talking with a somnambulist, be the somnambulism natural or induced, we find it hard to get into continuous colloquy on our own subjects. To begin with, he probably will not speak continuously for long together. He drops back into a state in which he cannot express himself at all. And when he does talk, he is apt to talk only on his own subjects; - to follow out his own train of ideas, - interrupted rather than influenced by what we say to him. The difference of state between waking and sleep is in many ways hard to bridge over.

We have thus three parallelisms which may guide and limit our expectations. From the parallelism of possession with split personalities we may infer that a possessing spirit is not likely to be able to inspire into the recipient brain ideas or words of very unfamiliar type. From the parallelism of possession with dream we may infer that the memory of the possessing spirit may be subject to strange omissions and confusions. From the parallelism with somnambulism we may infer that colloquy between a human observer and the possessing spirit is not likely to be full or free, but rather to be hampered by difference of state, and abbreviated by the difficulty of maintaining psychical contact for long together.

917. And here observe how different is the form our expectations will gradually assume from the commonplace - or even from the poetic - notion of what communication with the dead is likely to be, if it can take place at all. We now expect to have to do, not with a voice "monotonous and hollow like a ghost's, denouncing judgment"; - but rather with a voice incoherent and fugitive, like the voice of a sleeper; - with memories broken and arbitrary, like the memories of a dream.

And similarly as to what the voice is to tell us. We have no reason for anticipating either "judgment" or high revelation. We feel pretty sure, indeed, that there will be no ideas expressed which much transcend the automatist's habitual range. And, moreover, on the principle of continuity which has guided us throughout this work, we cannot assume that the departed spirit has already gained any vast increment of knowledge.

Whatever his new opportunities, we feel that his own capacity for learning may not have undergone any sudden change. We can hardly at first expect from him much more than some such account of his new state as may be intelligible to our material conceptions.

This, I say, is what we who are prepared by these previous studies are likely to expect. And I shall presently show that this is very much what we actually find. The expectations of the ordinary public, however, as seen both in fiction, and in the disappointed comments with which our actual results are greeted, are of very different scope.