'One gloomy February afternoon in 1936 S. G. S[oal] was sitting in the office at 13d, Roland Gardens, London,
S.W.7,' when 'suddenly the door opened and a tallish, well-groomed man of about thirty-six entered ... "I have come," he declared, 'not to be tested, but to demonstrate telepathy".'
THIS MAN was Basil Shackleton, and the present chapter concerns the experiments Dr. Soal conducted with him between that afternoon in 1936 and 15th April 1943(1).
(1) A complete account of the Shackleton experiments appears in
Modern Experiments in Telepathy, by S. G. Soal and F. Bateman, Faber, London, 1954.
Basil Shackleton was a professional photographer, and most of the experiments took place in his studio in Shaftesbury Avenue, London. The Agent sat at a card-table in the main room of the premises while Shackleton - supervised by one of the experimenters - sat in a smaller ante-room connected with the main room by a door. Both the Agent and Shackleton sat at various positions in their respective rooms in different experiments, but at no time could Shackleton see the Agent through the connecting door. In fact for the first nine sittings the door was kept completely shut while the guessing was in progress. Even had Shackleton been able to see through the door to the card-table, it would have been impossible for him to see the Agent, since he or she was always seated on the far side of a table behind a plywood screen. Furthermore no one but the Agent was able to see the cards, since these were sheltered from the view of anyone but a hypothetical person standing immediately behind the Agent by a cardboard box standing on its side.
This brings us to what is in many ways the crucial feature of Soal's experiments. Instead of a large number of cards being used on each occasion, the order of which had been randomized before the start of the experiment concerned, only five were ever required. Each of these five cards bore a picture of one of the following animals: giraffe, elephant, pelican, lion, zebra; and the nature of the card looked at by the Agent on any particular trial was determined as follows. Soal, before arriving at Shaftesbury Avenue, prepared, or caused to be prepared, a random sequence of the digits 1 to 5. At the start of the session the Agent shuffled the five picture cards and placed them face downwards inside the cardboard box. One of the experimenters (from January 1941 till June 1941, Soal himself) then showed the random numbers one by one through a hole 3 in. square in the plywood screen, and the Agent picked up and looked at the card in the position corresponding to that particular number. (The card on the left-hand end of the row in the box was called 'one', the one immediately to the right of it was called 'two', and so on up to five.) Now the Agent
re-shuffled these five cards at the end of each run of fifty guesses. Therefore no one but the Agent (not even Soal himself presenting the random numbers) knew which was the target picture for any particular guess.
The only person, therefore, who could have had normal knowledge of the order of the cards before the end of the experiment was the Agent; so if one is to suggest that Shackleton got the information by some normal means, one has to show how he could have got it (consciously or unconsciously) from the Agent.
After the first nine sittings the door between the two rooms was kept a few inches open so that the experimenter presenting the random numbers to the Agent could hear more easily when Shackleton had made his guess for one card and was ready to proceed to the next. The first possibility that suggests itself, therefore, is that the Agent might have communicated the nature of the card to Shackleton by means of some auditory signal. This signal need not have been consciously produced by the Agent, nor would it be necessary to postulate that Shackleton was conscious of obtaining information in this way.
This explanation is, however, ruled out by the nature of the experimental situation. Any noise made by the Agent that was sufficiently loud to reach Shackleton in the neighbouring room would have been detected either by Dr. Soal himself, seated just across the card-table from the Agent, or one of the twenty-one outside observers, of which there was at least one present on most occasions.
Secondly, Shackleton himself very effectively ruled out this possibility when he settled down to scoring above chance on the card
next to be turned up rather than the card the Agent was actually looking at. Between 24th January 1941 and 21st December 1941, for instance, at a total of nineteen sittings, Shackleton made 1,101 correct guesses out of a grand total of 3,789 possible correct guesses. This is an average of nearly seven correct guesses per run of twenty-five cards as compared with an expected or 'chance' average of less than five.
The difference between these two averages may not seem, on the face of it, a very startling one. In fact it is sometimes said that to the layman (i.e. anyone unacquainted with statistical theory) the results of even outstanding card-guessing experiments are unconvincing, since the individual scores differ so little from chance.
It is difficult to answer this objection since any criterion of 'littleness' in this context is bound to be purely subjective and idiosyncratic. However, anyone who has ever tried experimenting with cards, however informally, will agree with Soal that 'there is as much difference between [the performance of, say, Basil Shackleton] and that of the ordinary guesser as between chalk and cheese'(2). For instance, Soal quotes a series of six runs of twenty-five guesses each for which Shackleton's scores were 8, 10, 11, 10, 10 and 8, respectively. Perhaps there are people to whom even these scores (of which two-thirds are more than double the amount to be expected by chance) are not on the face of it surprising. But even if they are not as 'convincing' as would be a single run of twenty-five correct guesses, they are certainly large enough to be the envy of anyone who has ever tried experimenting with Zener cards himself. In fact the frequency with which such a sequence of scores may be expected to occur by chance is less than once in every two million times.
(2) Soal, S. G., 'Some Statistical Aspects of ESP', in
Wolstenholme, G. E. W., and Millar, E. C. P. (editors), Ciba Foundation Symposium on Extrasensory
Perception, J. and A. Churchill Ltd., London, 1956, p. 84.
In the case of the Shackleton experiments, however, there are certain features which must strike even those without a knowledge of statistics as requiring explanation. The first is the fact that Shackleton only achieved significant scores when certain people were acting as Agent. Altogether there were three successful Agents, while eight other people were tried with no success. (Among the latter were Soal himself and Shackleton's wife.)
Further, with one of the successful Agents, Miss Rita Elliott, Shackleton made his above-chance scores on the card one ahead of the one she was looking at, but with another successful Agent, Mr. J. Aldred, he produced significant scores in both
the '(+1)' and '(-1)' positions; in other words, he was successful in guessing both the card Mr. Aldred was about to turn up next and the card Mr. Aldred had just finished looking at.
Another fact which must appear significant even on a non-statistical criterion is the result obtained with 'two Agents in opposition'. In this situation, two of the successful Agents, Miss Elliott and Mrs. Albert, were
both looking at random series of animal pictures simultaneously, though the series were quite different from each other.
'Shackleton, who remained in the ante-room, and knew nothing whatever of the proposed experiment and the arrangements being made in the studio, was merely told that for the next three sheets of guesses Mrs. Albert was to be the agent. He was given no hint that while Mrs. Albert sat behind the screen in the usual position, Miss Elliott would be seated a few feet farther back with a second set of five animal picture cards in front of her.' The result was that with Mrs. Albert, Shackleton obtained a score with 'odds against chance' amounting to more than two million to one, while the score with Miss Elliott was a completely chance score. In other words, 'things happened just as if Miss Elliott had not been present'. The actual scores obtained with Mrs. Albert were the six scores quoted above, while with Miss Elliott the corresponding numbers of hits were 4, 3, 4, 3, 5 and 6. Thus, while there was a remarkable correlation between Shackleton's guesses and Mrs. Albert's target series, there was no correlation at all between Shackleton's guesses and Miss Elliott's series.
Another remarkable fact was that if Shackleton was speeded up - i.e. if the guessing rate was artificially stepped up from the average of one guess in every 2.6 seconds (which was the rate Shackleton found most congenial) - his above-chance scores now came not on the card
one ahead of the card he was guessing at, but the card two ahead. In other words, Shackleton's hits always seemed to be on the card that was to be turned up in about two seconds' time, regardless of the rate of guessing.
It might seem that on this basis one could predict what would happen if Shackleton had guessed at a slower rate than one card every two seconds. But what happened when Shackleton was slowed up so that he was only making one guess every four or five seconds was not that he started to score his hits directly on the card the Agent was looking at, but that he 'grew very irritable saying it was useless for him to continue, and complaining that it was enough to drive him mad. Invariably he ... failed to achieve a significant score of any kind at the "five seconds" rate'.
Finally, at one stage of the experiment Soal started to introduce runs of cards of which the Agent did not know the order. The procedure was for the Agent to shuffle the five cards and put them face downwards inside the box without looking at their faces, and when Soal showed a random number through the hole in the screen the Agent had simply to point to the appropriate card in the box without turning it up to see what it was. This was done about once a week with a run of fifty cards each time and without Shackleton's knowledge. It was found that on these surreptitious 'clairvoyance' trials Shackleton's score at once fell to chance.
Curiously enough, Shackleton himself appeared as convinced that he could do clairvoyance as he was that he could do telepathy. Soal says that 'At no time did he so much as hint that he anticipated or feared failure in experiments when the Agent did not look at the cards.... Nor did he seem aware of any difference in feeling when carrying out the two tests.' It may be argued that
Shackleton's confidence about clairvoyance was just a front; but the situation is complicated by the fact that when the first clairvoyance tests were introduced Shackleton was not aware that they had been. In other words, Shackleton would have had to exercise telepathy to have known that the Agent was not looking at the cards and to lose his confidence as a result!
These are some of the facts about the Shackleton experiments. The next chapter will consider whether they can be explained without recourse to the hypothesis of ESP.
The article above was taken from Charles McCreery's "Science, Philosophy
and ESP" (1967, Faber & Faber Ltd).