He wakes, and I now tell him to do anything he pleases, but that he is to act entirely of his own free-will; at the same time I give him a folded paper on which I have written what he is to do. X. does what I suggested, and is then much astonished to find written on the paper the action he has carried out and in the performance of which he firmly believed he acted of his own free-will.

Of course, we must not exaggerate the importance of these hypnotic experiments, neither should we draw general conclusions about free-will from them; for we may be nearly certain that there will be a feeling of free-will only when the action is agreeable to the individuality of the subject. If, for instance, a morally normal man were told to take another's watch or to commit some other improper act, the feeling of aversion would be strong enough to suppress the feeling of free-will. A subjective constraint would then be felt. The feeling of freewill is much more likely to be aroused by post-hypnotic suggestion when the acts suggested are indifferent and unimportant than when they are momentous and at the same time opposed to the subject's individuality. We come across something similar in the normal waking state; in that, as Freud has rightly pointed out, the feeling of free-will arises soonest when the proceedings are unimportant, while when more important decisions are concerned the feeling of being unable to act in any but one way almost invariably predominates. Finally, we must also beware of exaggerating the importance of hypnotic experiment, since we do not by any means find these deep hypnoses and subjective delusions of the judgment in all subjects.

Such subjects are invariably very few in number, and after repeated experiments most of them begin to observe themselves, which makes them conscious of the constraint put upon them. Finally, I must not omit to mention that hypnosis has taught us nothing absolutely new in this respect; for in every-day life, also, we are often able by an adroit manipulation of circumstances and conditions to give definite guidance to the will of some people without their knowing it. Goethe's saying, "Du glaubts zu schieben, und Du wirst geschoben," 1 expresses this fact. I will, moreover, take this opportunity of referring to the art of conjuring. A well-known trick of the conjurer is to allow a card to be drawn from a pack and to guess it at once. The trick is thus explained: the spectator thinks he has freely chosen the card, but in reality the conjurer has directed him to one in particular, and compelled him to take it. The conjurer attains this end by putting the card he wishes chosen where it will naturally be the first to be taken up. We have as little right to assert that the hypnotic experiments already described disprove the doctrine of free-will, as we have to consider that the above experiments in waking life do so.

Few can have made such experiments often without fresh doubts of freedom of will arising, but from these doubts to scientific proof is an immense step.

1 Roughly, "You think the force is yours, - and you are forced."

In one very particular respect hypnotism has given us a lesson of the greatest importance to psychology: it has proved that special precautionary measures must be taken in planning psychological experiments. It has shown that an essential source of error in psychological experiments was formerly overlooked or considerably underrated. The training of hypnotics has thrown light on this source of error. I must refer the reader to what I said on this question on page 156 et seq.; here I will only point out once again that a hypnotizer may, often without knowing it, by the tone of his voice or by some slight movement, cause the hypnotic to exhibit phenomena that at first could only be produced by explicit verbal suggestion, and that altogether the signs used by the hypnotizer to cause suggestions may go on increasing in delicacy. A dangerous source of error is provided by the hypnotic's endeavour to divine and obey the experimenter's intentions. This observation has also proved useful in non-hypnotic experiments. We certainly knew before the days of hypnotism that the signs by which A. betrays his thoughts to B. may gradually become more delicate.

We see this, for example, in the case of the schoolboy, who gradually learns how to detect, from the slightest movement made by his master, whether the answer he gave was right or not. We find the same sort of thing in the training of animals - the horse, for instance, in which the rough methods at first employed are gradually toned down until in the end an extremely slight movement made by the trainer produces the same effect that the rougher movements did originally. But even if this lessening in the intensity of the signals exists independently of hypnosis, it is the latter that has shown us how easily neglect of this factor may lead to erroneous conclusions being drawn. The suggestibility of the hypnotic makes these infinitesimal signals specially dangerous in his case. But when once this danger was recognized, greater attention was paid to this source of error in non-hypnotic cases than before. It is certain that many psychological experiments are vitiated by the fact that the subject knows what the experimenter wishes. Results are thus brought about that can only be looked upon as the effects of suggestion; they do not depend on the external conditions of the experiment, but on what is passing in the mind of the subject.

Just as we saw on page 343 et seq., that to certain medicaments chemical activity is only ascribed because the doctor expects a therapeutic action which he intentionally suggests to the patient, so is it with psychological experiments.

Perhaps many of the results obtained experimentally in connection with the injurious effects of alcohol belong here, since the subject is only too inclined to present the condition expected by the experimenter. Many experiments have been made - Kohlschutter's were the first - to measure the depth of sleep by the intensity of the sound necessary to wake the .sleeper. Forel has pointed out that these experiments prove nothing, since a noise to which a person is accustomed soon loses its power of waking him, whereas gentle but unwonted noises wake one at once. In addition to this, we have to consider auto-suggestion as a source of error, particularly the subject's expectation that he will be wakened by some noise or other, a circumstance that considerably enhances the probability of his being awakened. As early as 1831 Kohl-schutter himself had pointed out that the subject cannot get rid of the feeling of expecting the signal agreed upon. Naturally, these objections call for earnest consideration. Certainly Michelson, who has also studied the question, thinks that the unanimity of the results already obtained disposes of any importance that might attach to this source of error.