§ 6. The next predisposing condition of hallucinations that we have to consider is awe, in that special form which is connected with the near sense of death, and with which elements of grief and regret are often mingled. It is remarkable how large a proportion of phantasms of the recognised sort represent friends or relatives whose recent death is being mourned. Out of 231 cases, I find that 28 are of this type; of which 6 took place on the day or the morrow, 4 within a very few days, and the rest within a very few weeks, of the death.1Now the reader may ask how an emotional condition due to a death can affect the interpretation of any phantasms that could possibly be regarded as telepathic; for telepathy, as treated in this book, is an action between the minds of living persons. But it must be remembered that we have already assumed the possibility of a certain period of latency in telepathic impressions. That a certain period has followed a death before the occurrence of the hallucination representing the person who has died, is not, therefore, fatal to a telepathic explanation of the case; and the question how far the percipient's own emotional state is to be preferred as an explanation will depend, to some extent, on the length of this intervening period.

We are, therefore, led on to the wider question - by how large an interval must two experiences be separated before the possibility that they may be telepathically connected is excluded? This, it will be seen, is the very question involved in No. 3 of the four grounds of ambiguity above set forth.

Putting aside for the moment any special theory, it will be seen that the argument for a causal connection between an event of one class, A, and an event of another class, B, arising from the frequency with which such events approximate to one another in time, is not an argument whose force stops suddenly at any predictable point. It is an argument whose force gradually diminishes in proportion as the correspondences of time throughout a series of cases get less and less precise, and only becomes inappreciable when marked gaps begin to occur in the series. Thus the fact that certain pyschical phenomena form a cluster, comparatively thick at first and gradually becoming more and more sparse, in the few days that follow deaths, would strongly indicate some common bond of connection between the phenomena and the deaths, even if such a thing as telepathy in connection with living persons had never been observed. But as a matter of fact, we find the cluster of cases as thick just before life has ceased as just after.

Hence the presumption of a single common cause for the whole group.

Now not only is telepathy the only common cause which it seems possible to name; and not only is it the cause which is, so to speak, in possession of the field: it is a cause with which, up to a certain point after death, the theory of an emotional origin cannot even come into competition.

1 I am excluding some cases where the interval is known to have been as great as two months; but including some which are stated to have occurred "soon" or "very soon" after the death, without specification of time.

For however efficacious grief, or the general awe which death inspires, may be to provoke hallucinations representing the dead person, the grief and awe cannot operate till the fact of his death is known; and for anyone at a distance from him, this knowledge must necessarily follow the death at a very appreciable interval. Up to that point, then, we must hold that there is an appreciable chance that the hallucination is telepathic in character, and that an impression, received from the agent before or at death, has lain latent for a while in the percipient's mind. At the same time this theory of latency is one that ought not to be strained. The fact that in the majority of our cases there seems to be no latency at all, and that in another large group the period of latency is short, establishes a sort of presumption that it is not likely ever to be very long; and while admitting that any limitation that we can make is arbitrary, we prefer - for evidential purposes - to draw the line early, and (as explained in Chapter IV (General Criticism Of The Evidence For Spontaneous Telepathy)) we have drawn it at 12 hours after death.

The cases, however, where knowledge of the death, and the emotion caused by it, have preceded the hallucination, stand on very different ground. A new element is now introduced, which has seemed to many weighty authorities a full and sufficient explanation of all post-mortem phantasms. I am not aware, indeed, that any crucial instance has ever been forthcoming - that anyone, believing a friend to be dead who was really in a perfectly normal state, has seen his "ghost." Moreover, most of those who have attributed post-mortem phantasms to the percipient's emotional condition have done so under an a priori conviction that with physical death all possibility of affecting others must necessarily cease - a conviction which my colleagues and I do nor share.1 At the same time, death is so wholly unique a fact in human experience that it seems reasonable to believe the ideas and emotions connected with it capable of producing very unique effects; and personally I am disposed to regard such ideas and emotions as probably the sufficient cause of any hallucinations that occur while they are present, as long as such hallucinations present no features which the percipient's mind would have been unable to supply.

The presumption is at any rate so strong, that no experiences of any sort - even though otherwise admissible, as following within the 12 hours' limit - have been included in our telepathic evidence, if the fact of the death was already known to the percipient, and if his experience was unshared by any one else; and this principle is the easier to justify in that the day immediately succeeding a death seems the most likely of all periods for abnormal subjective impressions connected with the death to occur.

1As our telepathic theory is a psychical one, and makes no physical assumptions, it would be perfectly applicable (though the name perhaps would be inappropriate) to the conditions of disembodied existence. And it may be quite fairly asked why this possibility was not taken account of above, in connection with the phantasms that have shortly followed deaths. What need is there, it may be said, to trace these phenomena to a state of the agent preceding or exactly synchronising with his physical decease, when his psychical life may be supposed to be continuing after the great change? The answer is that the point is not one of theoretic possibilities, but of evidence; and that the evidence for post-mortem communications seems to us inconclusive. In all the evidential cases cited in this book, the very keystone of the argument is the coincidence - more or less close, but always close enough to be remarked - with death or some other external fact. That - and often that alone - is the obstacle to regarding the cases as purely subjective: failing that, very special and peculiar features must be present, to establish even a presumption of some exciting external to the percipient's own mind.

For example, the same hallucination might affect several persons independently and at different times; or the phantasm might convey information, afterwards discovered to be true, of something which the percipient had never known - this last condition being probably the only one which could prove an intelligent external cause. A certain amount of evidence of both these types exists, of a quality which makes it imperative on us to keep our minds open for more; but at present it will bear no sort of comparison with the evidence for telepathy. For a sketch and criticism of the present state of the question, see Mrs. Sidgwick's paper "On the Evidence, collected by the S.P.R., for Phantasms of the Dead," in vol. iii. of the Proceedings.