This mode of regarding them (and the reservations with which the word "subjective " must be used) will be fully explained in the sequel. It is enough for the present to note that the witness who would be an unsafe authority if be said, "Sea-serpents exist," may be a safe authority if he says, "I saw what appeared to be a sea-serpent"; and this amount of assertion is all that the telepathic evidence involves. All the accuracy of observation required of the witness has to do with what he seemed to himself to see, or to hear, or to feel.

1I may say here, once for all, that our gratitude to an informant is none the less because his or her experience may not have appeared relevant to the direct argument of this book. Such cases have often been very useful and instructive in other ways.

Nor in our cases is the danger of errors of inference so serious as might-be imagined. A man may, no doubt, see something odd or indefinite, at the time that his mother dies at a distance, and may infer that it bodes calamity; and if, after he hears of the death, he infers and reports that he saw his mother's form, the error will be a very grave one. But it will be more convenient to treat retrospective mistakes of this sort under the head of errors of memory. And with a percipient's interpretation of his impression at the moment we have really very little concern. He may see the apparition of a relative in his room, and infer first that it is the relative's real figure in flesh and blood, and next that it is the relative's spirit. Neither inference has any relation to our argument. The only fact that concerns us is the fact that he had the subjective impression of seeing his relative. I may refer once more, by way of contrast, to the case of witchcraft, where the very basis of the superstition was error of inference - error shown (and by the more intelligent class exclusively shown) not in the giving but in the interpreting of testimony.

§ 5. The tendencies to error which more vitally concern us fall broadly into two classes - tendencies to error in narration, and tendencies to error of memory. Let us ask, then, what are the various conscious or unconscious motives which may cause persons who belong to the educated class, and who have a general character for truthfulness, to narrate experiences of telepathic impressions in a manner which is not strictly accurate?

One motive which has undoubtedly to be allowed for in some cases is the desire to make the account edifying. This danger naturally attaches to the evidence for any class of facts which can be regarded, however erroneously, as transcending natural law. Enthusiastic persons will value an unusual occurrence, not for its intrinsic interest, but for its tendency, if accepted, to convert others to their own way of belief; and they will be apt to shape and colour their account of it with a view to the desired effect. Intent on pointing the moral, they will unconsciously adorn the tale. This source of error is one which it is specially necessary to bear in mind where some particular type of story is connected with a particular religious sect. The literature of the Society of Friends, for instance, is remarkably rich in accounts of providential monitions and premonitions; and it supplies also a considerable number of telepathic cases. But we have already seen that telepathy does not specially lend itself to the support of definite articles of faith. Nor is any one who takes the trouble to study our evidence likely to maintain that errors of narration have largely entered into it under the influence of a propagandist zeal.

It is rather for the sake of completeness than on account of its practical importance that such a possibility has been mentioned.

A far more frequent and effective source of error in narration is the tendency to make the account graphic and picturesque. Among human beings, the motives which prompt narration of matters unconnected with business or the mere machinery of life are mainly two - a desire to interest one's auditor; and a desire to put oneself en evidence, to feed one's own self-esteem by attracting and retaining the attention of others. The influence of each of these motives is towards making the story as good a one as possible. And though, as I have already said, a good deal of our evidence comes from persons who profess to have had no bias in favour of the reality of such events as they describe, and wish rather that they had not occurred, still the instinct to make what one says seem worth saying is too general for it ever to be safe to assume its absence. In such a subject as ours, this instinct will find its chief opportunity in making things appear marvellous. The reader must decide for himself how far the evidence to be here presented bears the stamp of the wonder-monger or raconteur. The desire to make people open their eyes is no doubt perfectly compatible with a habit of truthfulness in the ordinary affairs of life.