Beginning in 1880, investigators carried out experiments in telepathy with an English family of five sisters aged between ten and seventeen, reporting a high degree of success. The findings were discredited eight years later after two of the girls were discovered to have cheated by secret signaling, although their method could not have produced the many positive results in those experiments for which they could not see or hear each other.
Background
The Creery family lived in Buxton, Derbyshire. In 1880, the girls Mary, Alice, Emily, Maud and Kathleen ranged in age from ten to seventeen.1 A maidservant, Jane, had lived with them for two years. According to their father, Rev AM Creery, all were mentally and physically sound.
A popular pastime during this period was the ‘willing game’, in which one member of a group tries to guess an object that has been selected by the others. In October 1880, Creery decided to try it with four of his daughters and the servant Jane:
Each went out of the room in turn, while I and the others fixed on some object which the absent one was to name on returning to the room. After a few trials the successes preponderated so much over the failures that we were all convinced there was something very wonderful coming under our notice. Night after night, for several months, - we spent an hour or two each evening in varying the conditions of the experiments, and choosing new subjects for thought-transference. We began by selecting the simplest objects in the room; then chose names of towns, names of people, dates, cards out of a pack, lines from different poems, etc., in fact any things or series of ideas that those present could keep steadily before their minds; and when the children were in good humour, and excited by the wonderful nature of their successful guessing, they very seldom made a mistake. I have seen seventeen cards, chosen by myself, named right in succession, without any mistake.2
Creery observed that other children invited to join the sessions also succeeded with a little practice. The best results followed when the person doing the guessing (percipient) was about a yard apart from the person or persons who knew the identity of the object (agent). However, success could still be attained when the percipient was in a separate room. It seemed to depend on the agent(s) keeping the object steadily in mind. Typically, the children said, two or three objects of the relevant kind would come into their minds, and they would choose the one that appeared most vividly.
William Barrett
A local doctor whom Creery invited to a session made a record of some impressive results,3 and Creery presented these to a philosophical society. Reports subsequently appeared in local newspapers, one of which Creery forwarded to William Barrett, a physicist who had been investigating cases of thought-transference. Barrett visited the family in March 1881 and spent three evenings experimenting with the children. On 3 July he published a short paper in the journal Nature, in which he argued that successful results could not be accounted for in terms of unconscious cues given by the agent and unconscious perception on the part of the percipient, using the Creery sisters as an example. He cited a session in which the group in the living room mentally selected objects somewhere in the house, and the girl who, in another room, was attempting to guess its identity, found the object and brought it to the living room (instead of simply returning there and stating the answer) – with the following results:
…hair-brush, correctly brought; orange, correctly brought; wine-glass, correctly brought; apple, correctly brought; toasting-fork, wrong on the first attempt, right on the second; knife, correctly brought; cup, correctly brought; saucer, failure. On being told this object the child said, ‘Saucer came into my head, but I thought you would never ask for that after asking for a cup, so I wasn’t sure what it was’.4
In fifteen trials using names of objects or English towns the girls only erred completely three times. Next, Barrett held short sentences in his mind – ‘What time is it?’ ‘Will you have some supper?’ ‘Will you go to bed?’ and ‘Were you at the sale today?’ – all of which the girls correctly identified.
Researchers Balfour Stewart and Alfred Hopkinson visited the Creery family in November 1881. They tested the five sisters and maidservant with objects, playing cards, two-figure numbers, names of towns and names of characters (Peter Piper, Bluebeard, Tom Thumb and Cinderella). The most striking result was the girls’ complete success with the characters.5
Society for Psychical Research
Barrett was involved in the founding of the Society for Psychical Research in 1882.6 Its first project was an investigation on thought-transference, carried out by a committee consisting of Barrett, Frederic WH Myers and Edmund Gurney, who performed experiments with the Creerys in April of that year.7
Before leaving the room, the percipient child was told the nature of the object she was to guess, for instance a playing card, number or the name of a town. On returning she was to stand close to one of the investigators, either with her face turned to the wall or eyes downcast for a period of silence ranging from a few seconds to a minute, until she called out her answer. If it was incorrect she would be given a second and occasionally a third chance.
Some trials were carried out in the Creery home, others in lodgings or a private hotel room. The girls were asked variously to identify the place in the house in which an object had been hidden, a playing card randomly drawn from a full deck, an object held in the investigator’s hand, and a name that had been made up.8 The experiments given the most weight by the investigators were those in which the child was isolated from those who knew the answer and those in which only investigators and no members of the family knew the correct answers. The authors gave consideration to non-paranormal methods as follows:
Involuntary actions,such as movement of the lips, &c., could not reach the child when she was out of sight and hearing, as was the case in the first series of experiments. Conscious or unconscious deception on the part of the subject does not apply, as the thing wished for was selected and written down by one of us. Collusion by a third party is avoided by the fact that none were allowed to enter or leave the room after we had selected the thing to be guessed, and in the second series of experiments by the exclusion of all members of the family, either from the room, or from participation in the requisite knowledge. In subsequent experiments we obtained successful results by individual trials with each of the children, that is to say, the number, word or card was known to some of us only.9
A typical set of results is as follows. (The name in the left column is the target; on the right is the statement given by the percipient trying to guess it. Mary was the percipient for the first five and Maud for the remainder.)
William Stubbs William Stubbs
Eliza Holmes Eliza H------
Isaac Harding Isaac Harding
Sophia Shaw Sophia Shaw
Hester Willis Cassandra, then Hester Wilson
John Jones John Jones
Timothy Taylor Tom, then Timothy Taylor
Esther Ogle Esther Ogle
Arthur Higgins Arthur Higgins
Alfred Henderson Alfred Henderson
Amy Frogmore Amy Freemore. Amy Frogmore
Albert Snelgrove Albert Singrore. Albert Grover10
Over six days 382 trials were made, mostly with playing cards. The investigators calculated that 71 correct answers might be made by chance. The girls were correct 127 times on the first attempt, 56 on the second and 19 on the third. With cards, first-attempt guesses that were incorrect in both suit and number were a small minority.11
The most striking result was five cards named entirely correctly on the first attempt consecutively when no family member knew the test-card. The authors calculated that this could happen by chance only once in more than a billion times. Other runs of positive results with similarly remote odds were attained.12
Pondering whether the girls received the correct answers through mental eye or mental ear, the co-authors note that a visual method is suggested by the errors showing half-correct answers for two-figure numbers and playing cards, and especially by the frequent confusions of king with knave; but on the other hand, an auditory method is suggested by similar-sounding answers given for names.13
Success rates fluctuated strongly. Conditions that seemed to favour success were the presence of the girls’ father, the percipients feeling ‘freedom from constraint and … a spice of pleasurable excitement’, and vivid simultaneous picturing of the correct answer by the sitters.14
Barrett, Myers and Gurney conducted a second set of experiments in July 1882 at Cambridge with Mary, now age seventeen, Alice, age fifteen and Maud, age thirteen. They noted that fluctuations in success rates were wide, and that more successes were gained in a casual atmosphere than a tense or anxious one. A comparison of results when one or more sisters knew the information versus when none knew showed little difference.15
In addition to tests with playing cards and two-figure numbers, the investigators tried a new test: giving the mixed up letters of an obscure Latin botanical term to one of the girls and asking her to choose the letters successively so as to form the word. One example result, with ‘1’ representing the correct letter on the first try, ‘2’ on the second and so on, is given:16
Hedypnois Physaloides
111121111 12411114111
Overall, again, the hits were far beyond chance level.17
Decline Effect
Striking success was also achieved in a second set of experiments carried out with two of the girls in Dublin in November 1882. However, the authors note: ‘The fact seems to be (and the children themselves are regretfully conscious of it) that the capacity is gradually leaving them’.18
In the committee’s third report, the co-authors gave the results for all trials in which only an investigator knew the card or number selected (except the Dublin experiments, since only one investigator was present) as follows: of 260 trials with playing cards, one in nine first responses was correct, and of 79 trials with two-figure numbers, one in nine first responses was correct, both results well exceeding chance.
Here too the investigators observed the ‘decline effect’, as it was named by a later generation of parapsychologists:
It may be noted that the power of these children, collectively or separately, gradually diminished during these months, so that at the end of 1882 they could not do, under the easiest conditions, what they could do under the most stringent in 1881. This gradual decline of power seemed quite independent of the tests applied.19
Cheating
By 1888, the Creery girls had been tested over a period of seven years. That year, Gurney published a note to the effect that two of the sisters had confessed to using a signaling code to help them identify cards in recent experiments. A third confessed to signaling in earlier experiments where one sister was the agent, although, she maintained, only rarely, when they feared they were failing and wished to avoid causing disappointment. The code was as follows:
When the two sisters were in sight of one another, the signals used were a slight upward look for hearts, downwards for diamonds, to the right for spades, and to the left for clubs. Further, the right hand put up to the face meant king, the left hand to the face meant queen, and knave was indicated by crossing the arms. It is doubtful whether there were any signs for other cards. We failed to make any out clearly. A table showing the degree of success in guessing each card suggests that there were signs for 10 and ace, but that they were either only used occasionally or used with poor success. In experiments in which a screen was placed between the two sisters, so that they could not see each other, auditory signs were used to indicate suits. A scraping with the feet on the carpet meant hearts, and sighing, coughing, sneezing or yawning meant diamonds.20
Gurney notes that he had made it clear to the girls that the most important experiments were those in which only the investigators acted as the agent, and these were also given primacy in published results. However some experiments using one or more of the girls as agents had been included in records, and Gurney acknowledges that these all were now discredited.
Detractors frequently cite the incident when making the case against parapsychology, without discussing the experiments or appearing to grasp the limitations of the girls’ signalling method.21 Other commentators argue that cheating in paranormal investigations must be understood in the social context, particularly in the case of children who feel pressured to provide successful results over a long period.22
KM Wehrstein
Literature
Barrett, W.F. (July 3, 1881). Mind-Reading versus Muscle-Reading. Nature 24, 212.
Barrett, W.F. (1911). Psychical Research. New York: Henry Holt & Co.
Barrett, W.F., Gurney, E., & Myers, F.W.H. (1882). First report on thought-reading, Proceedings of the Society for Psychical Research 1, 13-64.
Collins, H., & Pinch, T. (1982). Frames of Meaning: The Social Construction of Extraordinary Science. London: Routledge and Kegan Paul.
Creery, A.M. (1882). Note on thought-reading. Proceedings of the Society for Psychical Research 1, 43-6.
Gurney, E. (1888). Note relating to some of the published experiments in thought-transference. Proceedings of the Society for Psychical Research 5, 269-70.
Gurney, E., Myers, F.W.H., & Barrett, W.F. (1882). Second report on thought-transference. Proceedings of the Society for Psychical Research 1, 70-97.
Gurney, E., Myers, F.W.H., Podmore, F., & Barrett, W.F. (1883). Third report on thought-transference. Proceedings of the Society for Psychical Research 1, 161-216.
Inglis, B. (1988). Sir William Barrett (1844-1925). Journal of the Society for Psychical Research 55, 16-24.
McLuhan, R. (2010). Randi’s Prize: What Sceptics Say About the Paranormal, Why They Are Wrong and Why It Matters. Leicester, UK: Matador.
Stewart, B. (1882). Note on thought-reading. Proceedings of the Society for Psychical Research 1, 35-42.
Endnotes
- 1. Barrett et al. (1882).
- 2. Creery (1882), 43.
- 3. Creery (1882), 44-5.
- 4. Barrett (3 July 1881), 212, right column.
- 5. Stewart (1882), 40.
- 6. Barrett (1911), 54.
- 7. Barrett et al. (1882).
- 8. Barrett et al. (1882), 20-24.
- 9. Barrett et al. (1882), 24.
- 10. Barrett et al. (1882), 26.
- 11. Barrett et al. (1882), 26-7.
- 12. Barrett et al. (1882), 27.
- 13. Barrett et al. (1882), 28.
- 14. Barrett et al. (1882), 29.
- 15. Gurney et al. (1882), 72-3.
- 16. Gurney et al. (1882), 75.
- 17. Gurney et al. (1882), 75.
- 18. Gurney et al. (1882), 78.
- 19. Gurney et al. (1883) 171.
- 20. Gurney (1888), 269.
- 21. Examples: Wikipedia entry on Barrett’s psychical research, Skeptic’s Dictionary.
- 22. McLuhan (2010), 362-3; Collins & Pinch (1982).