630 A. Note On " Number-Habits"

Some critics of the experimental evidence for thought-transference have attempted to show that a large proportion of the successes obtained may be due, not to telepathy, but to the mental idiosyncrasies of the experimenters. This is a possible source of error which should always be kept in mind, but its actual bearing on the evidence has been widely misunderstood. A brief general review of the subject of "mental habits" is therefore given here.

It has long been recognised by psychologists that most - if not all - persons have unconscious preferences for certain objects or ideas over others of the same class; so that, if one is asked to guess or to think of, say, a colour, a playing-card, or a number, - certain colours, cards, or numbers occur to the mind more frequently than others, and are therefore guessed more often. These idiosyncrasies are called "mental habits," or - if we are referring to numbers only number-habits," and they may vary in the same person at different times. But different persons may exhibit the same preferences, and when this is so in the case of two experimenters, a certain proportion of the diagrams drawn by the percipient may resemble those drawn by the agent, and thus simulate the phenomenon of thought-transference. Similarly, if the cards or numbers to be guessed are chosen by the agent, his mental habits may lead him to choose a large proportion of those that happen also to be favourites with the percipient, who will therefore have a better chance of guessing right.

This is, of course, one reason why it is always best for the agent in experiments with cards or numbers to draw them at random from a batch and not to choose them.

It is sometimes found, in fact, that in experiments with diagrams of which only a small proportion have succeeded, the successes relate to the most familiar forms, such as circles and triangles. To explain, however, by "mental habits" the large proportion of success in the series cited below, it would be necessary to prove that almost all the diagrams used were general favourites.8

In order to test empirically how far mental habits might have simulated thought-transference in the experiments with diagrams, Colonel G. L. Le M.

1 The incident of the feet occurred at 4.30 to 7.30 p.m. The crystal picture was about 10 p.m.

2 [Mr. Lang gives further footnotes referring to these cases as follows:] " Miss Angus had only within the week made the acquaintance of Mrs. Cockburn and the Bissetts. Of these relations of theirs at a distance she had no knowledge. This account I wrote from the verbal statement of Mrs. Bissett. It was then read and corroborated by herself, Mr. -Bissett, Mr. Cockburn, Mrs. Cockburn, and Miss Angus, who added dates and signatures. The letters attesting each of these experiments are in my possession. The real names are in no case given in this account, by my own desire, but (with permission of the persons concerned) can be communicated privately".

3 An attempt was made by Professor C. S. Minot (see his paper "Upon the Diagram Tests" in the Proceedings of the American S.P.R., vol. i. p. 302) to find statistically what forms occur most often when people are asked to draw or to think of diagrams.

Taylor carried out a series of dummy experiments, made in the same manner as the experiments in thought-transference, but with the element of thought-transference eliminated (see Proceedings S.P.R., vol. vi. p. 398). He got eighty persons to draw twenty-five diagrams each, and so obtained 1000 pairs of diagrams, which could be compared according to a prearranged plan. The comparison showed how many resemblances were actually produced by chance, combined with similarity in the mental habits of the persons who drew the diagrams; and the number of resemblances were found to be proportionately far less than those found in the experiments in thought-transference.

To this it might be objected that the persons who drew the diagrams being taken at random, there was no reason to expect similarity in their mental habits; whereas - since some experiments in thought-transference fail, while others succeed - it might be argued that only those succeed where the mental habits of agent and percipient happen to be similar. Similarity of mental habits could not, of course, in any case ensure success, because it would be very unlikely that the percipient would think of his favourite forms in the same order as the agent; but it might increase the chance of success. In many of the experiments in thought-transference, however, the diagrams were drawn or selected by some person other than the agent, and not always the same person, so that a general similarity in mental habits - as well as a general tendency to think of the favourites in the same order - would have to be assumed, and this seems to be negatived by Colonel Taylor's experiments.

In the case of experiments with such objects as numbers,1 the effect of "mental habits " can be more precisely tested. Supposing first that the same numbers happen to be the favourites of both agent and percipient. If, then, the agent selects numbers to think of, some successful guesses may be made which are due - not to thought-transference, but to similarity in the number-habits of the two experimenters.

This source of error, however, may be absolutely excluded if the numbers to be guessed are not selected voluntarily by the agent, but drawn at random from a batch of numbers. As early as 1886, therefore (see Phantasms of the Living, vol. i. pp. 31-35, and vol. ii. p. 653), experimenters who worked in connection with the Society for Psychical Research were accustomed to use the method of drawing numbers at random.

On the other hand, supposing again that the agent selects the numbers and that his number-habits are markedly dissimilar from those of the percipient, then the successes would probably be decidedly fewer than they would be if due to chance alone.

Now, confining ourselves to cases where the numbers to be guessed are drawn at random, it is clear that the existence of any decided number-habit does not affect in any way the probability of guessing right by chance, since the number drawn at any moment is neither more nor less likely to be one of the percipient's favourites than to be any other number. On the average, therefore, the number of accidental successes would be the same, whether a number-habit existed or not.

1 For examples of the experimental study of number-habits, see Professor and Mrs. Sidgwick's articles on their experiments in thought-transference in the Proceedings S.P.R., vol. vi. p. 170, and vol. viii. p. 548; also a further discussion by Professor Sidg-wick in vol. xii. pp. 303-4; and a brief account in a review of Dr. Dessoir's Das Doppel-Ich in vol. vi. p. 209.

A decided number-habit may, however, affect prejudicially the number of successes produced by telepathy (assuming, for the sake of the argument, that successes may sometimes be due to telepathy), because the idea of the favourite number, constantly obtruding itself into the mind, would tend to obscure or replace the impressions derived telepathically; just as, when a material object is perceived in the ordinary way through the senses, a preconceived idea as to what the object is may often make us perceive it wrongly.

Thus, in experiments of the kind under consideration, there is only one case in which the existence of number-habits can increase the successes and so make the evidence for telepathy in that case appear stronger than it really is; namely, the case in which (1) the agent selects the numbers to be guessed and at the same time (2) his number-habits are similar to those of the percipient. In all other cases, number-habits would decrease those successes which are due to any other agency than chance.