THERE HAVE been other changes in attitude since the early days of psychical research besides the one already discussed, the shift of emphasis from anecdote to experiment. There has also been a change from exclusive preoccupation with the task of obtaining evidence as to the reality of ESP to interest in the problem of what can be found out about ESP.
Must we not, however, be satisfied that something exists before we can be interested in finding out about its nature? Certainly no one is going to be interested in the research problems of something that he is quite certain does not exist; no zoologist is going to devote himself to study of the sex life of the unicorn. But something falling considerably short of certainty as to its reality is all that is necessary to give sufficient reason for pursuing the study of something as a research interest. Zoologists are more ready to investigate the abominable snowman than the unicorn although there is not complete certainty that the abominable snowman exists. Its existence may be judged sufficiently likely to make it an interesting topic of enquiry; so, at least, must also the existence of ESP and other
In planning an investigation, the experimental psychical researcher must choose whether he is going to try to convince sceptics by producing new and more rigid proof of the reality of the paranormal, or whether his experiments are to have some other aim, such as that of finding out about the characteristics
of ESP. This is an important choice since generally an experimental design for one of these purposes is not suitable for the other. It is true that some experiments intended to produce evidence for the reality
of ESP have, as a by-product, contributed important indications of new characteristics of ESP; the discovery of
psi-missing and of displacement are examples. Generally, however, the condition of prolonged repetition of identical experiments which is needed for new evidence as to the reality of ESP is a condition unfavourable for studying the characteristics of
ESP. For the latter purpose, the experimenter will need to try his experiments under varying conditions to discover how his subjects' scores change under these changes. Working in this way, the experimenter is not likely to get the impressive total scores that are needed in experiments to prove the reality of ESP.
One of the dangers to which experimental psychical researchers are exposed is that hostile criticism may divert them too much to the relatively unrewarding task of multiplying proofs of the reality of the paranormal. Some types of criticism seem to assume that proof of reality is always the aim of a parapsychological experiment and that it is reasonable to judge all experiments, whatever their ostensible object, by the extent to which they contribute to this aim. Such critics may suppose that they have made a case against the reality of ESP when they point out the failure of a particular experiment to prove the reality of ESP, although the experimenter himself may have made no such claim. It must, however, be admitted that experimenters do sometimes lay themselves open to this sort of criticism by not making it clear in their research reports exactly what was the object of their experiments. Sometimes even they discuss their results as if they were contributions to the total evidence for the reality of ESP, although the experiments were not carried out in a way appropriate for this purpose.
It is, of course, a fact about psychical research that there is a real question as to whether there is any reality in what it is investigating. Such a question is rarely a live one even in other fields of the biological sciences and only very exceptionally in the physical sciences. The 'abominable snowman' and the 'Loch Ness monster' are somewhat parallel examples from the biological sciences. The situation is even rarer in the physical sciences because generally the specifying of the conditions of an observation is sufficient to enable other scientists to repeat the observation.
There is, however, one parallel case in the physical sciences, that of 'ball lightning'
(Silberg, 1965). It has been frequently reported that glowing balls have been seen during a thunderstorm. These are said to persist for some seconds or minutes and then either to explode with a loud bang or to disappear quietly. The people reporting these appearances have rarely been trained meteorologists and may generally be considered to be of doubtful reliability as reporters. The phenomena of ball lightning resemble paranormal phenomena in the fact that they cannot be produced at will. They have not been produced experimentally, and so far they have not been photographed.
It is plainly not unreasonable to doubt whether ball lightning exists at all, and many meteorologists remain doubtful. Yet the general tendency of scientific research workers has been neither to ignore the reports nor to try to test their validity by a rigid process of verifying the reliability of those who claim to have seen it. The tendency has been rather to consider that the evidence is sufficient to justify a research interest in the phenomenon, and to see whether a coherent picture of the properties of ball lightning can be built up from the reports, and how far a theoretical explanation of the reported phenomena can be constructed.
In the realm of the paranormal also, it may be suggested that the primary task of those who are engaged on research in the matter is not to produce such overwhelming evidence for its reality that hostile criticism will be silenced. Presumably the psychical researcher is already convinced that the evidence for the paranormal is good enough to justify it as a research interest; otherwise he would not be engaged in psychical research. He may, in fact, think that the evidence is very much stronger than this, but he need not think that the demonstration of this evidence or the strengthening of it are matters of primary importance. He may think that it does not matter very much that there is hostile criticism; there may be some gain to psychical research that its findings are subjected to such criticism. Whether this is so or not, he may recognise that the criticism will not be silenced merely by accumulating evidence, and that it will be dangerous only if it diverts the research worker from his proper task of finding out about the nature of the paranormal to the less fruitful task of trying to defend its reality.
If it is admitted that the defence of the reality of the paranormal is not the primary task of the experimental psychical researcher, he must nevertheless be prepared to answer the question as to how strong the evidence is. The present chapter is an attempt to answer this question with respect to the reality of ESP.
An experiment which is to contribute any weight of evidence with respect to the reality of
ESP must use other methods and other precautions than those used in the more ordinary type of experiment which is investigating the nature of
ESP. Usually such experiments have involved many repetitions of experiments under similar conditions which would be an uneconomical and uninformative method for normal
ESP research. One reason for large numbers of results under similar conditions is that experimenters aiming at producing evidence for the reality of
ESP have felt that their evidence would be more impressive if the index of significance had an extremely low value. They may often have overestimated the importance of this; there is no real advantage in making the possibility of being misled by chance incomparably lower than the possibility of other sources of error.
They must also consider the possibility that other people may think the results are due to dishonesty on the part of the experimenter and have independent witnesses and procedures of independent checking to reduce this possibility. Such precautions would be unnecessary and hampering in the ordinary
ESP experiment but they are appropriate here. The certainty of every experimental result is, as was pointed out long ago by Professor
Sidgwick, limited by the honesty of the experimenter. It is unlikely that an apparently honest research worker will lie about his experimental results but the likelihood is not zero. The likelihood can, however, be reduced to a very small amount by adequate provision of independent witnesses and checking, and only those experiments which have been so safeguarded, and which also have been successful at a high level of significance, can be admitted as evidence for the reality of ESP.
There are a number of experiments which fulfil these conditions, and the choice as to which ones to cite is somewhat arbitrary. My own choice would not necessarily be that of other psychical researchers. In making the choice of experiments below, I have been guided by various considerations; partly by the extent to which particular researches have been accepted as of high evidential value, partly by the consideration that the researchers quoted should not be too remote from us in time, and partly by the aim of showing that successful experiments have been performed under different conditions and in different countries.
I have chosen the following five researches as examples of the strength of evidence for the reality of
1. The Pearce-Pratt series of experiments carried out at Duke University in 1933-4 (Rhine and Pratt, 1954).
2. The long series of experiments carried out by Dr Soal and Mrs Goldney with the high-scoring percipient Mr Shackleton in London in 1941
(Soal and Bateman, 1954).
3. An experiment carried out by Professor Lucien Warner in the thirties (Warner, 1937).
4. Card-guessing experiments carried out by Dr Ryzl in Czechoslovakia which have been witnessed or controlled by visiting investigators from other parts of the world Ryzl and
5. A fully automated series of experiments carried out in 1967 by Dr Helmut Schmidt (Schmidt, 1969a, 1969b), then a research physicist at the Boeing Aircraft Research Laboratories (now Director of the former Duke University Parapsychology Laboratory at the Foundation for Research into the Nature of Man, Durham, North Carolina).
The first of the above series of researches on the reality of ESP was an early one in the Parapsychology Laboratory of Duke University, the design of which included precautions against the possibility of the experimenter cheating. It was carried out during 1933 and the following year with Dr Pratt as agent and a divinity student called Pearce as percipient. The percipient was sent by himself to the University Library and the agent was in another building, 100 yards away in some experiments and 250 in others. The agent displaced the cards one by one from an ESP pack at an agreed time without turning them over. At the end of going through each pack the cards were turned over and recorded by Pratt. The guesses were recorded by the percipient. In order to eliminate the possibility of cheating by agent or percipient, the precaution was taken that both placed their records in a scaled package which was handed to Professor Rhine before the two lists were compared. Under these conditions, success was obtained at a good rate. The total number of guesses was 1,850 of which one would expect a fifth (i.e. 370) to be right by chance. The actual number of hits was 558 which exceeds mean chance expectation by about fifty per cent. This is a highly significant result which leaves no doubt that some factor other than chance determined the hits. If this cause was not ESP, it must have been some kind of fraud not eliminated by the design of the experiment, since this design made impossible any accidental sensory leakage or accidental effects of recording errors.
A more elaborately safeguarded experiment was that of Dr Soal and Mrs Goldney which they carried out in London in 1941. These experimenters used 'open' packs, prearranged from lists of random numbers. There were five target cards with pictures of different animals; one of these was exposed to the agent in accordance with a code connecting each of the digits 1-5 with one of the animal pictures. The actual guess made by the percipient was the name of one of the pictured animals (lion, elephants, etc.). An earlier pilot experiment had shown that Shackleton seemed to be guessing, not the card at which the agent was then looking, but the card ahead of that; the experiment was planned to discover whether this guessing one ahead would continue.
The results of this confirmatory series of experiments are summarised on p. 149 of Soal's book (Soal and Bateman, 1954). A minor mistake in the totalling was afterwards discovered by Soal; the figures given here are the corrected ones (Soal, 1956). Over a long series of guesses (nearly four thousand in all) Shackleton showed continued success in this guessing of the card ahead, with the unusual scoring rate of seven per cent above mean chance expectation. Successes were made with a number of different agents; each experiment was observed by outside witnesses to ensure that the precautions were maintained and that a true record was made of targets and guesses.
One result of the combination of a high scoring rate with a large number of guesses in this experimental series is that the odds against a chance deviation of this size are very heavy. Soal's estimate of P = 10-35 needs some correction since this was based on his original score totals but, even so corrected, the odds against chance occurrence would be of the order of many billion billions to one. Since the possibility of explanation of these results by chance is effectively eliminated by a much less extreme P value than this; one might wonder whether this series of experiments was not longer than it need have been for the purpose of demonstrating Shackleton's ability to guess the following target card correctly. One must not suppose that a low value of P gives an overall guarantee of the certainty of the result indicated, since it indicates only that the results were not produced by chance while it does not show that they were not produced by some other cause than
ESP (e.g. by experimental error or by fraud). What gives this assurance is the adequacy of the primary precautions used, and this assurance is as complete in a short experimental series as in a long one.
If the sub-atomic value of P reported in the Soal-Goldney experiments does not make these results more certain than they would be with a P of, let us say, 10-6, it does allow one to make an answer to one possible objection to their interpretation as indications of
ESP. This is the objection that this series of results might be the one accidentally positive result in an indefinitely large collection of ESP experiments performed elsewhere. Probability theory predicts that if one tried often enough to carry out an ESP experiment, any possible excess might turn up accidentally in an indefinitely long series of attempts. If, however, we ask how many experiments would have to be carried out before one could expect a deviation the size of the Soal-Goldney result to turn up by chance, the answer is that this number would have to be fantastically large, larger indeed than the whole history of the world would allow for. If one supposed that every inhabitant of the globe had done an unsuccessful
ESP experiment every month during the last sixty million years, the necessary correction that would have to be made for the Soal-Goldney series being the selected best of all these experiments, would still leave it overwhelmingly significant with odds against its chance occurrence of many million millions to one. The Soal-Goldney result could not reasonably be attributed to chance even if this was the only evidence we had that pointed to
If, however, this or any other single experimental series were indeed the only evidence we had for the reality of
ESP, the low value of P obtained in it would not, in itself, be ground for regarding this result as of any scientific importance. At best it would show that something very odd had happened, but if this oddity were never repeated it would not be a useful fact for scientific purposes; it would suggest no further step that we could take in order to understand it. The importance of this experimental series depends on the fact that it does not stand alone; it is one of a number of experiments with safeguards against the experimenters cheating which indicate the influence of ESP on the results obtained.
Another experiment of considerable importance as evidence is a relatively short one carried out by Professor Lucien Warner and reported by him in 1937. The experimenter was a psychologist who wished to make a decisive test on the matter. By himself acting as agent he could know that he was not cheating, and by using a percipient who had already shown success in
ESP experiments he hoped to ensure positive results. This hope was fulfilled. Warner placed the percipient in a locked room on a different floor from the experimenter. In order to ensure that there was no involuntary indication from himself as to the nature of the target, he arranged that there was communication from the percipient to the experimenter and not the other way. In 250 guesses the percipient had 93 hits which is an excess of 43 over the
50 right guesses that would be expected if there were only chance correspondence between the percipient's guesses and the targets. Although the series was short in comparison with the others considered here, this excess was adequately significant with odds against chance occurrence of the order of a thousand million to one. There is, therefore, ample evidence for the presence of some cause favouring right guessing and the simple but adequate precautions seem to rule out any cause other than
ESP. This experimental series therefore deserves an honourable place amongst those researches contributing evidence as to the reality of ESP.
The next set of experimental researches that will here be mentioned is that of Dr Ryzl of Czechoslovakia. Apart from its contribution to the problem of training in
ESP, Dr Ryzl research is of importance as an example of sustained high scoring which puts Dr Ryzl's subject, Pavel Stepariek, in the same class as Dr Soal's high-scoring subjects, Basil Shackleton and Mrs Stewart. It would be a disconcerting fact if all the high scorers in
ESP experiments were figures of the past.
Ryzl's researches are scattered over the pages of the technical journals but the article cited (Ryzl and Otani, 1967) gives a table summarising the results up to that date. The method of experimenting has been that the percipient made a two-target choice as to whether a card in a sealed envelope had the white or the green face uppermost. The cards had been inserted in envelopes before the experiment, then shuffled and some of them turned before placing the envelopes in an outer cardboard cover. This ensured that none of the experimenting team knew which face was uppermost in any of the envelopes. The task was, therefore, an ESP operation not involving telepathy and therefore fully safeguarded against the possibility of information being unwittingly conveyed from the experimenter by such means as unconscious whispering.
The table given in the Ryzl-Otani article shows ten series of experiments with a total of 17,648 guesses and 10,117 hits. With a fifty per cent chance of accidental success, this means that the excess of hits was 1,293 over mean chance expectation. This corresponds to an 'index of success'(1) of about 14; this may be compared with the figure of about 9 which I have calculated from Soal's + 1 results with Shackleton. The deviation from mean chance expectation in Ryzl's results is of such a high level of significance that they obviously cannot be explained as a chance effect.
(1) This index is explained in Appendix A. In the case of a choice between two targets, it is a value double the percentage deviation from expectation.
Of the results reported in these tables, some were obtained in the presence of experimenters from outside, including Pratt, Blont, Beloff, and Freeman. The only series which did not show an excess of hits was that in which Beloff was present when 1,200 guesses were made with 65 hits below mean chance expectation (a significant deficiency).
While all the sets of experiments with outside experimenters present and in control are sufficiently safeguarded against the theoretical possibility of the results having been fraudulently produced by the main experimenter, the most completely safeguarded experiment is, no doubt, one carried out in November 1963 by Dr Pratt and Dr Blom in which Ryzl himself was not present and the organisation of the experiment was entirely in the hands of the visiting experimenters (Pratt and Blom, 1964). In this series, there were 1,600 guesses with 917 hits, 117 over mean chance expectation. The level of significance is very high, a P value of the order of 10-8. The result remains overwhelmingly significant if we take into account the fact that a series undertaken by Pratt alone and one by Pratt and Stevenson showed less striking results (Pratt, 1964). The odds against the
occurrence of this rate of scoring by chance as the best of three experiments remains much more than a million to one.
At the time of writing, the most recent experimental confirmation of the reality of
ESP is a research carried out by Dr Helmut Schmidt (Schmidt, 1969a, 1969b). This is of special importance since some critics of ESP have said that they reserved judgment as to the reality of ESP until its occurrence was shown by a fully automated experiment in which selection of targets was made by an automatic randomising device and all responses by the subject were scored automatically. These conditions were fulfilled by Dr Schmidt's machine. It might be argued that, even with such a machine, the reality of the results depends on the truthfulness of the experimenter and any witnesses of his records, but the advantage of automated selection and recording is that spurious results cannot be produced by any method of cheating by the subject or by errors made by the experimenter in target selection or the recording of hits. In spite of the rigorous conditions of Schmidt's experiment, his results were overwhelmingly significant.
The apparatus used in this experiment had four lamps of different colours. The task of the subject was to choose one of four keys to indicate his opinion as to which of the four lamps would light up when the key was pressed. Which light actually lit up was determined electronically. In the first experiment, this was not determined at the moment when the subject made his decision, so the task was a precognitive one. In another experiment a tape was pre-punched with a sequence of random numbers which determined the targets, so the task could be one of clairvoyance.
Schmidt used as his subjects three individuals who had given previous grounds for supposing that they could succeed in an
ESP task; two of these were professional mediums. In the first series of experiments, a total of 63,066 guesses gave a preponderance of hits with odds against chance occurrence of about 2,000,000,000 to 1 (P>10-9). This, of course, effectively eliminates the possibility of it being a chance result, and the conditions of the experiment eliminate the possibility of any other explanation of the result than that of precognitive Esp. A later experiment with 20,000 trials gave a result of slightly greater significance (P = 10-10).
Although this series of researches has provided additional proof of the reality of
ESP, Dr Schmidt has not discovered a fool-proof method of ensuring that any experimenter can get positive
ESP results by using it. Other parapsychological experimenters have got chance results with Schmidt's machine or with machines closely resembling it. The task of devising an experimental set-up by which any experimenter can be sure of getting positive
ESP results remains a problem for the future. We cannot even be sure that it will ever be solved.
This is not, of course, a full list of researches that may be considered to supply evidence of the reality of
ESP. There were many experiments well safeguarded and carefully carried out in the early days of the Society for Psychical Research, but these took place rather a long time ago. An excellent experiment was carried out at Groningen in 1926 by Professors of Physiology and of Psychology on a young man who obtained high scores in a task which involved pointing to a square on a board as determined by the experimenters in a room above using a random method of selection (Brugmans, 1921). There was also at Duke Parapsychology Laboratory another experimental series with safeguards against deception by the experimenter in the Pratt-Woodruff experiment which also showed a high rate of success (Pratt and Woodruff, 1939). One might also include the experiments reported by the Russian physiologist Vasiliev on the telepathic induction of the hypnotic state. These and many other carefully controlled experiments show, at any rate, that the appearance of experimental
ESP is not confined to a few laboratories or to a short period of time. If the assertion of the reality of
ESP is a conspiracy of dishonest experimenters, it is a conspiracy which extends widely in space and time and which includes a variety of people who seem unlikely to co-operate.
In considering the implications of this evidence, there is more than one question that we may ask ourselves. It has already been suggested that the important question for the experimentalist is whether it establishes a sufficient case for the reality of
ESP to make it a reasonable matter for research investigation. I think there is little doubt that the answer to this question must be 'Yes'; there are few controversial issues on which the evidence in favour is stronger.
If one asks further whether the evidence is so strong as to compel belief, the answer is not so easy. It is clear that the evidence for
ESP is strong, so strong that conviction of its reality can only be avoided by supposing that a number of apparently honest and reliable investigators are deliberately deceiving the world as to their results. This is certainly extremely unlikely, but not, in principle, impossible. If it were certain that ESP is impossible, one would have to prefer this improbability of a widespread conspiracy of deception to the impossibility of
ESP, and to accept deception as the explanation of those successful results that cannot be explained by chance or by experimental error. That this is the case is the conviction of a number of critics of psychical research to whom it appears self-evident that extrasensory perception is as mythical as the unicorn. The point of view of these critics will be discussed in the next
BRUGMANS H. J. F. W. (1921), 'A Report on Telepathic Experiments Done in the Psychology Laboratory at Groningen', Report of Premier Congres international des recherches psychiques, Copenhagen.
PRATT, J. G. (1964), 'Preliminary Experiment with a "Borrowed" ESP Subject', Journal of the Society for Psychical Research, xlii, 333-45.
PRATT, J. G., and BLOM, J. G. (1964), 'A Confirmatory Experiment with a "Borrowed" Outstanding ESP Subject' J. S.P.R., xlii, 381-9.
PRATT, J. G., and WOODRUFF, J. L. (1939),'Size of Stimulus Symbols in Extra-sensory Perception', Journal of Parapsychology, iii, 121-58.
RHINE, J. B., and PRATT, J. G. (1954), 'A Review of the Pearce-Pratt Distance Series of ESP Tests', J. Parapsych., xviii, 156-77.
RYZL M., and OTANI, SOJI (1967MAn Experiment in Duplicate Calling with Stepariek' J. Parapsych., xxxi, 19-28.
SCHM1DT, H. (1969), 'Precognition of a Quantum Process', J. Parapsych., xxxiii, 99-109.
SCHMIDT, H. (1969), 'Clairvoyance Test With a Machine', J. Parapsych., xxxiii, 300-7.
S1LBERG, P. A. (196 5), 'A Review of Ball Lightning', Problems of Atmospheric and Space Electricity, (ed. S. C. Coronoti), Amsterdam and London, 436-54.
SOAL, S. G. (1956),'The Shackleton Report: an Error Discovered', J. S.P.R., xxxviii, 2 16-17.
SOAL, S. G., and BATEMAN, F. G 954), Modern Experiments in Telepathy, London
VASILIEV, L. L. (1963), Experiments in Mental Suggestion (English translation), Church Crookham.
WARNER, L. (1937), A Test Case', J. Parapsych., i, 234-8.