The Trickster and the Paranormal

by George P. Hansen

  The implicit assumptions didn’t stop theoretical development, but the emphasis was on experiments. Parapsychology was a small field (for most of its history, probably fewer than 10 full-time researchers in the U.S.), and with Rhine’s commanding presence, theoretical work was not a high priority. Rhine kept the field focussed on experiments, and without his leadership it may well have disintegrated. Yet the focus had a price.

  New Approaches

  The late 1960s and the 1970s were years of growth for parapsychology, and spurred by several factors, that period proved to be a watershed for theoretical development. With Rhine’s retirement as

  director of the Institute for Parapsychology, his dominance in the field lessened, and new laboratories were established by those who found Rhine’s approach too restrictive. The influx of new researchers broadened the field, and established scientists from other areas brought perspectives from their disciplines. Physicists introduced quantum mechanics to parapsychological problems; biologists undertook more studies of psi in animals, and computer technology allowed new and more sophisticated experiments. During that fruitful but turbulent period, older experiments were reevaluated, and alternative theoretical models were developed.

  Helmut Schmidt and Evan Harris Walker, a physicist with the U.S. Army at Aberdeen Proving Grounds in Maryland, produced some of the most important theories. They were based on quantum mechanics and attracted the attention of a number of scientists. Other theories were introduced as well. Rex Stanford proposed his psi-mediated instrumental response (PMIR) and conformance behavior models. J. E. Kennedy published conceptual articles on task complexity, redundancy, and experimenter effects. William Braud introduced the ideas of lability and inertia in relation to psi processes. All were developed in the 1970s.

  These theoretical efforts occurred in tandem with experiments that explicitly probed the limits of psi and raised questions as how to best conceive it. The driving topics included psi-mediated experimenter effects, retroactive PK, the goal-oriented nature of psi, divergence, task-complexity independence, conformance behavior, and the source of psi problem. Probably only a few dozen people in the world have a working knowledge of these terms. They will be explained below. They are interrelated, and one cannot draw clear distinctions between them, but all speak to the problem of boundaries, constraints, and limitations. Their meaning and implications are difficult to grasp without extensive familiarity with the literature of experimental parapsychology. As a result, most receive little or no coverage in introductory texts, and since the early 1980s, they have attracted relatively little notice, even within the field.

  To illustrate the points, several additional experiments need to be presented. Some were quite clever and produced striking results, and the reader can profit even if he or she chooses to concentrate on the experimental procedures and ignore the conceptual ambiguities. Nevertheless, theoretical issues are central here, and as with all scientific experiments, varying interpretations are possible. Yet when all parapsychological research is viewed together, a picture emerges that is decidedly odd.

  Experimenter Effects

  Parapsychologists had long known that some of their colleagues were more successful than others in obtaining positive results. They suspected that this was due to differences in interpersonal skills, abilities to make people feel comfortable yet highly motivated in experiments. The experimenter effect was so well recognized that Rhine urged new people coming into the field to try some preliminary tests in order to determine whether or not they could elicit psi in a lab. If a potential parapsychologist was unable to do so, Rhine counseled that the person might be more effective in something other than experimental psi research. Other parapsychologists experimentally tested experimenter effects, and it turned out that they involved more than just social skills. In some studies, experimenters unconsciously exerted a psychic influence on scoring patterns.

  In the early 1950s Donald J. West and George W. Fisk had an interest in the issue. Fisk was a British physicist who spent many years in China and later served as editor of the Journal of the Society for Psychical Research. West is a psychiatrist, criminologist, and former president of the Society. Each had conducted a number of studies. Fisk frequently obtained positive results, but West did not. They decided to collaborate and explore the differences in results. They set up an experiment that resulted in one of the most frequently cited papers on psi experimenter effects; it was published in 1953.

  Twenty subjects were each sent 32 sealed packs of 12 clock cards (clock cards have the face of a clock printed on them and a hand pointing to one of the hours). West and Fisk each prepared half of the targets for each subject, and the subjects were unaware of who had prepared any of the packs or even that two different experimenters were involved. Fisk sent out all the packs, and in random order. It turned out that the cards prepared by Fisk were much more successfully guessed than those prepared by West. The tabulated results

  showed that no single subject was responsible for the overall outcome.10

  The experiment was generally well conducted, but unfortunately the report gave essentially no details about precautions against cheating. However, there were no especially extreme scores that might provoke such suspicions. Furthermore even if one of the subjects did cheat, it would be expected that all packs were vulnerable (the authors report that even if the subjects had known the packs to have been prepared by two experimenters “they would not have been able to tell which were which.” Thus while the experiment was not perfectly controlled, cheating seems unlikely.

  The question arises—were the subjects the ones responsible for the results? Some probably were, for one of them did very well on targets sealed by both Fisk and West. Had Fisk or West left some kind of residue that either stimulated or inhibited psychic ability? This might be a possibility, but later work by other experimenters found similar patterns in precognition experiments where the targets were not selected until after the guesses were made. Significant differences were found between experimenters who checked the results, and it appeared that it was the experimenters’ psychic influences that were at work. These kinds of findings provoked discussion.

  In 1976 two major reviews appeared on experimenter effects in psi research. One was by Rhea White in the Journal of the American Society for Psychical Research and the other by J. E. Kennedy and Judith Taddonio was published in the Journal of Parapsychology. These were independently written, but covered much of the same material, indicating that the matter had come to the fore. White, a librarian, had worked at Rhine’s laboratory in the 1950s. She later served as editor of the Journal of the American Society for Psychical Research and now edits her own periodicals and databases. over the years she has produced a number of important reference books for the field as well as significant conceptual articles. Kennedy was a young researcher at Rhine’s lab who published several major theoretical papers in the 1970s. He was one of a trio who caught laboratory director, Walter J. Levy, M.D., faking results in 1974.

  Part of the stimulus for their papers on experimenter effects came from parallel concerns in ordinary psychological research. In the preceding 10 years, Robert Rosenthal’s study of experimenter bias (previously mentioned in the chapter on reflexivity) attracted considerable media interest, and it sparked lively debates. Experimenter effects alarmed psychologists. They often compared groups in experiments, and an investigator might unconsciously treat one group differently than another. If a procedure allowed an experimenter’s bias to be unintentionally communicated to the subjects, results could be contaminated. Similar problems arise in medical research with the well-known placebo effect. The problems in parapsychology are even more severe than in psychology or medicine. The West and Fisk study indicated that experimenter effects can be due to psi.

  Experimenting about experiments is a reflexive activity, and experimenter effects are one of the problematic manifestations of reflexivity. They challenge the validity of experimental methods in all fields. They r
aise problems that people want to avoid. As discussed in the last chapter, experimenter effects in ordinary psychology were difficult to replicate, and the replication problem in parapsychology is even more severe. One cannot be sure that only the putative subjects use psi; experimenters and others may contribute psi influences. This, of course, exacerbates problems of replication. Experimenter effects also raise the issue of participation. Psi permits (and maybe requires) experimenters to psychically participate in experiments, even unwillingly. They cannot fully separate themselves from their task.

  Participation is a crucial issue for parapsychology. It is also found in ethnomethodology and in the sociology of scientific knowledge. It will again be addressed in the next chapter on primitive mentality, particularly with the work of Levy-Bruhl.

  Non-Intentional Psi

  Some people seem lucky and have a knack for being at the right place at the right time. Parapsychologists noticed this and suspected that those people used their psychic abilities unconsciously. The idea initially appeared straightforward, and a number of experiments tested it.

  In 1975 Rex Stanford and his colleagues published a paper on non-intentional PK. Stanford is a professor of psychology at St. John’s University in New York, and he has been an active experimentalist and theoretician in parapsychology for many years. Earlier in his career he worked at Rhine’s laboratory and then with psychiatrist Ian Stevenson at the Division of Parapsychology at the University of Virginia Medical School. Rex has an identical twin brother, Ray, an artist and psychic who has written a number of books on paranormal topics. Ray has also made substantial contributions to paleontology by discovering previously unknown fossilized species.

  In Stanford’s experiment 40 male college students were each given the long, boring task of tracking a slow-moving pursuit rotor, for up to 45 minutes. Unknown to each student, a random number generator was running during the time he was being tested. If at any point during the 45-minute session the RNG produced a sufficiently extreme result, the student was released from the tracking task, and given an interesting one—rating erotic pictures. It turned out that a significant number of students escaped the boring tracking task and had the opportunity to view the erotica. They were not aware that they were in a parapsychology experiment at the time. Apparently they used PK unconsciously to influence the RNG; that seemed to be the most reasonable interpretation, at least initially.

  Martin Johnson at Lund University in Sweden conducted another study on non-intentional psi in 1971 and 1972 (ESP in this one).15 Johnson was in charge of a psychology course, and he prepared an examination for his students. Each copy of the exam was pasted to an envelope. Unbeknownst to the students, inside each envelope was a sheet with some randomly selected answers. Each student had a different set answers, and the sheets were covered with aluminum foil to assure the answers couldn’t be seen. The students were not told the contents of the envelopes or even that they were participating in an ESP test. They did significantly better on the questions that were answered in their own envelopes.

  That outcome raises a question. If, as much evidence indicates, ESP is not limited by distance, why didn’t the students use ESP to learn the answers to all the questions equally well (e.g., by telepathy from the teacher, or from answers in other students’ envelopes)? Why did they score higher on questions that had answers in their own envelopes? of course one could suggest reasons that were somehow compatible with the findings. But this and other research suggest that it wasn’t necessarily the subjects’ psychic abilities that accounted for the results. The experimenter’s desire for a successful outcome, coupled with his own unconscious psychic ability, may explain the result. Johnson randomly selected answers for each envelope and in distributing the envelopes for the test, perhaps they were matched with students who knew those particular answers. This entire procedure can be construed as a complex random process, and his psi may have influenced it. A number of conceptually comparable studies also suggest that experimenter psi can operate in such circumstances.

  Task-complexity Independence

  Early laboratory researchers suspected that ESP and PK were aspects of a unitary process. One way to investigate that idea combined ESP and PK tests into one experiment by using hidden targets. Karlis Osis published one of the earliest of those experiments in 1953 while he was working at Rhine’s laboratory at Duke. Osis was a major figure in parapsychology, and he deserves some introduction. He was born in Latvia and obtained his Ph.D. in psychology in Germany. He immigrated to the U.S. after WWII and first did manual labor because of his poor English, but during that time he tried some ESP tests with animals. Rhine had an interest in that and invited Osis to work at Duke. After his stint there, Osis served as director of research for the Parapsychology Foundation in New York City and later in a similar capacity at the American Society for Psychical Research. He is one of very few who made a nearly life-long career in parapsychology, and he was an innovator. He pioneered research on ESP in dreams; he studied near-death experiences long before Raymond Moody and Elisabeth Kubler-Ross popularized the topic, and he investigated the effects of distance on ESP performance. Osis was one of few who devoted himself almost exclusively to research and had little to do with administration, teaching, or management.

  Osis’ testing procedure used dice, and it was simple and straightforward. Lists of target faces (numbers 1 through 6) were prepared by utilizing tables of random numbers. On each list, 16 targets were indicated. Each list was placed in an envelope, and a similar but blank sheet was clipped to the outside. Subjects threw a die for each trial, trying to make the uppermost face of the die match the associated target in the envelope. The throws’ outcomes were recorded on the outer sheet.

  If ESP and PK act like perceptual and motor processes, then in Osis’ study, one presumes that a person first uses ESP to learn the target face and then uses PK to affect the fall of the die. If this is so, PK scores with hidden targets should be much lower than those with targets known to the subject. For instance, if ESP and PK were independent and each worked accurately 10% of the time, then they would work together only 1% of the time (.10 x .10 = .01). The scores would be much reduced.

  Osis found that scores on hidden targets were about the same as for known targets, indicating that PK may not be comparable to other forces in nature. This preliminary experiment was certainly not sufficient to prove the conclusion, and further, it was not as well controlled as formal studies in parapsychology, a fact Osis made very clear in his report. (Better controls would include machine tumbling of the die, sealed envelopes with target lists covered with aluminum foil to assure opacity, and greater monitoring by the experimenter. It should be noted, however, that Osis and his wife contributed the bulk of the trials, which partly mitigates the shortcomings.) The conclusions were strengthened when a variety of later studies pointed in the same direction.

  Hidden-target PK experiments are now seen as testing the effects of what is called task complexity. Complexity in this sense refers to the amount of apparent information required to duplicate a psi task by normal human sensorimotor activities (e.g., in the Osis experiment, one first needed to learn the identity of the target and then influence the fall of the dice; two steps were required). Making a task more complex did not lower the scores, and this was confirmed with other studies with high complexity that still achieved success.

  These experiments stimulated theoretical work, and in 1974 Helmut Schmidt suggested that psi was “goal oriented.” He stated: “This suggests that PK may not be properly understood in terms of some mechanism by which the mind interferes with the machine in some cleverly calculated way but that it may be more appropriate to see PK as a goal-oriented principle, one that aims successfully at a final event, no matter how intricate the intermediate steps.”21 This idea spurred others.

  In June 1978 J. E. Kennedy published another important conceptual paper entitled: “The Role of Task Complexity in PK: A Review” in the Journal of Parapsychology. Kenned
y asked whether there was any substantial difference between a subject throwing a die and wishing for a six in a PK experiment and an experimenter conducting a study and desiring a significant result. Both attempt to influence random processes.

  Kennedy pointed out that any experiment involves a hierarchy of goals: a subject aims for success on each individual trial; each subject wants to do well on the sum of his or her trials; the experimenter desires a statistically significant experiment, and each experiment is typically part of a larger line of research that the investigator hopes to be successful. The goals are not contradictory, but they do cause problems for interpretation. If an experimenter is the primary source of psi, and the subjects do little more than generate random data, statistical significance might be detected only when a study is analyzed overall. For instance, it is possible that no individual subject achieves a significant result yet the entire experiment could be statistically significant. The pattern of the data generated randomly, via random number tables and subjects’ guesses, may conform to the experimenter’s expectations. Kennedy suggested that in order to maximize psi in an experiment (statistically speaking), all parties involved should give particular focus and effort to each individual trial rather than on the overall experiment. He elaborated upon this idea in a paper on redundancy published the next year.