Professional Documents
Culture Documents
Dreams1971failure PDF
Dreams1971failure PDF
Dreams1971failure PDF
'This smdy was supported through the personal generosity of Professor Gardner Murphy,
to whom we are deeply grateful. Additional support came from N I M H Grant 18315-01.
W e were given invaluable assistance by Dr. Stanley Krippner and Charles Honorton of
the Maimonides Dream Laboratory. W e also wish to acknowledge the assistance of the
following persons who conscientiously filled various roles in this study: Cherie Barkee,
James D. Larson, John McDonald, Dr. Terry Pivik, Peggy Rowland, Jean Shepherd, Dr.
Gerald Vogel, and Robert Weisz. T o our anonymous subject, who came from afar to
spend 2 wk. in winter-time Laramie in the hands of a group of roral strangers, we offer
our special thanks.
784 E. BELVEDERE & D. FOULKES
METHOD
S was the same S who had served, over 2 yr. earlier, in the Krippner and
Ullman study. The present sn~dywas conducted at the Sleep Laboratory of the
University of Wyoming. Methods were comparable to those in the Maimonides
study. Whatever differences in procedure seemed desirable or proved necessary
are elaborated below.
Target selection was done by El, who thereafter played no other role in the
study. H e was instructed to assemble a grand pool of 80 potential targets in
duplicate. Targets were magazine illustrations of a representational sort. (The
shift from*art prints to magazine illustrations was done at S's request.) El was
told to form the stimuli into 10 groups of 8 each on the basis of a lack of over-
lap in figures, settings, and emotional tone. Each stimulus then was put into a
small, sealed opaque envelope on which El signed his name and sealed the sig-
nature with transparent tape. Each group of 8 stimuli, now randomly numbered
1 to 8, was put into a larger opaque envelope, randomly assigned a letter from
A to J. That envelope, which constituted the target pool for one night, was
signed and sealed in like manner as the envelopes containing the individual stimu-
li. El then transmitted the 10 lettered envelopes to El.
E2 served as "security officer" for the experiment. H e stored the large en-
velopes off the Sleep Laboratory premises. O n the evening of each laboratory
night, he randomly selected the letter-series to be used on that night and delivered
one of the two envelopes containing the corresponding target pool to the labora-
tory. Eg verified the signan~reand seal and held the envelope for delivery to A
at the time of A's seclusion for the night.
S, meanwhile, had arrived at the laboratory in the presence of the A he had
selected for that night. The A pool consisted of three females, Al, A-, and Ag.
S h a d been given an opportunity to become acquainted with each potential A pri-
or to the onset of the study. None of the As had served in the Maimonides study.
Al claimed prior psychical phenomena, A2 was without such experience but sym-
pathetic to the idea of telepathy, and Ag had served as S in a 6-night pilot study
with inconclusive results. S selected Al for Nights 1 and 8, and A2 for Nights
2 to 7. To promote A-S rapport, S spent a variable period of time on the eve-
ning prior to each laboratory session with the A he had designated for that ses-
sion. N o practice nights were run with Al. One practice session with A2 was
conducted between Nights 1 and 2.
As the A selected for any particular night watched, Ex attached S's electrodes
for standard EEG-REM monitoring of sleep. E4 then conducted S to his bed-
room. With S in his bedroom, Eg received the target pool from E2,who did not
arrive at the laboratory until given a signal from Ex that S was abed. E3 then
TELEPATHY AND DREAMS: REPLICATION 785
conducted A to her room, where he gave her the large opaque envelope contain-
ing that night's target pool. She was instructed not to open the envelope until
after E5 had delivered a slip of paper containing a randomly selected number in-
dicating which of the 8 potential stimuli was to be the target. E5 did not enter
A's room; he merely deposited the slip under the door and immediately left the
building. E5 was otherwise out of contact with Es, As, or S during the conduct of
the study. He retained a duplicate of each night's number to permit later verifi-
cation that A did in fact open only the small envelope corresponding to the num-
ber he had selected. E5 selected each night's number just prior to delivering it
to A. His numbers were selected with replacement (i.e., the same number was
available for repeated use), while E2's letters were selected without replacement
(i.e., a target pool could only be used once).
A's suite of rooms was on the first floor northeast of the same building in
which S slept on the third floor southwest. Tape seals were affixed to all of A'S
door and window openings by E3 SO that she could not, without detection, leave
her suite and return unaided. The building itself was locked. S was locked in
his room. An intercom unit continuously monitored his room, and there was
continuoi~spolygraphic monitoring of S himself. In addition, E3 and/or E 4 al-
ways were in a position to observe the hallway leading to the doors of S'Sroom.
E3 monitored the polygraphic recordings from a control area adjacent to S's
bedroom. At the start of each REM period he signaled A via a one-way buzzer
to commence concentration on the target. The same target was "transmitted"
during the several REM periods of a given night. A indicated her receipt of the
signal by turning on a switch connected to a light in the control area. Between
"transmissions," A generally slept. She always awakened when signaled, how-
ever, as indicated by her activation of the light.
S was awakened after variable durations of REM sleep: awakening 1, M =
7.5 min. ( n = 8 ) ; awakening 2, M = 11.9 min. ( n = 8 ) ; awakening 3, M =
12.1 min. (n = 8 ) ; awakening 4, M z 16.1 rnin. (a = 7 ) ; awakening 5, 20
min. ( n = 1). At a REM burst, E3 signaled Eq, in a room adjacent to the main
control area, to initiate the awakening. He also signaled A that the REM period
was terminated, and A always indicated her receipt of the signal by promptly
activating the switch closing the signal light in the control area.
Dream interviews required little probing by E4, since S's spontaneous report
invariably was quite complete. Following each report, S was free to indicate any
tentative hypotheses he had concerning the nature of the target. The entire in-
terview was recorded for later transcription. S s final awakening was routinely
scheduled for 7 a.m. If he was in REM sleep, a dream interview was initiated.
If he was in non-REM sleep, no dream recall was solicited. In either event, S
made a free guess as to the target and any other comments or associations to his
dreams that he wished placed in the record for that night.
E4 then entered S's room to i~nplughim from the polygraph terminal box.
786 E. BELVEDERE & D. FOULKES
Meanwhile, E2 had delivered the duplicate set of target materials to EB,who veri-
fied the seal and signature. E4 took these materials into S's room and also placed
the tape recorder there, so that S had access both to the target p o l and his dream
reports of the previous night. S then ranked the target stimuli for their corres-
pondence to his dreams. Only after this judging was completed, and the judg-
ment forms were in E4's hands, was A allowed to leave her quarters. Es checked
the seals on A's suite, verified that only one of the eight envelopes in the target
pool had been opened and that that envelope corresponded to S's copy of the ran-
dom number slip delivered by E5, collected all these materials from A, and then
conducted A to S's room. Here, in the presence of E2, E3, and Eq, the match of
S's ranking to the target was ascertained, and A and S were free to discuss the
results wich one another. E3 and Eq later verified that A's random number sheets
conformed with those of E5.
S's judgments were made on a standard form. Each stimulus in the target
pool (1 to 8) was given a rank corresponding to S s estimation of the degree to
which it might have influenced his dreams. Rankings were made separately for
each REM period associated with dream recall and then for the entire dream
sample of the night. It was the latter ranking which was the dependent variable,
with the former employed simply to ensure that S's attention was drawn to each
dream report which he had given d ~ ~ r i nthe
g night. N o tied ranks were allowed.
A rank of 1 was defined as maximal dream-stimulus correspondence and a rank of
8 as minimal dream-stimulus correspondence. Hits and misses were scored as in
the Maimonides study.
Two outside judges were recruited from other laboratories. Both had had
considerable prior experience in the analysis of dream reports. These judges
were asked to rank stimuli in the target pool for their correspondence to S s total
dream output for each of the 8 experimental nights. There were two conditions
for their judgments on each night: using only S's dream reports, and using aU of
S's recorded comments during the night including associations and free guesses
as to the target.
RESULTSAND DISCUSSION
The ranks S assigned to actual targets were, in order of experimental nights,
7,7, 1, 1,6, 3,7, 5. H e thus had 3 hits ( 2 direct) and 5 misses. With P = .50
as the chance value for a hit on any single night, 4 hits would be expected by
chance in a series of 8 nights. S thus performed below chance in his accuracy of
matching dreams with targets.
Judge 1 performed identically without and with access to S's associations.
The ranks he assigned to true targets were, in order, 5 , 8, 1, 1, 7, 3, 8, 7. He
also had 3 hits ( 2 direct)-the same in each case as S s own hits-and 5 misses.
Judge 1 performed below chance in his accuracy of matching dreams with targets.
Judge 2 gave different rankings without and wich associations. Without
them, the ranks he assigned actual targets were, in order, 7, 7, 1 , 6 , 3, 7, 6, 5. He
TELEPATHY AND DREAMS: REPLICATION 787
had 2 hits (1 direct) and 6 misses, a below-chance level of accuracy. With as-
sociations, Judge 2's ranks were, in order, 5, 8, 1, 6 , 4 , 3, 7, 2. Here, Judge 2 had
4 hits ( 1 direct) and 4 misses, a chance performance.'
It might seem that the results of S's and Judge 1's rankings for Nights 3
and 4 are indicative of some transmission effect. Considering this as a self-con-
tained two-night series, the probability of two direct hits would be 1/64. But
these two nights were not, of course, self-contained, but part of a larger series,
and there is no justification for isolating these two nights merely on the basis of
the results they yielded. The probability of 2 or more direct hits in a series of 8
nights, where the chance probability of a hit on each night is l/s, is .26, and,
thus, the direct-hit data for the total series are well within the boundaries of
chance performance.
The major finding is that neither S nor the two judges were able to exceed
chance in matching dreams with targets. W e thus were unable to replicate the
results S had achieved earlier ac Maimonides.
In accounting for the discrepant findings of the two studies, at least two
broad lines of approach are possible. First, one may reject the hypothesis of
paranormal influence on dreams and then ask what procedural variations berween
the two studies might have resulted in a spurious finding of telepathic dream
influence in the Maimonides s t ~ ~ d ySecond,
. one may reject the findings of the
present study and then ask what methodologically defensible features of the first
study were modified in this study so as to interfere with the demonstration of a
telepathic influence on S's dreams.
So far as we can see, there were no methodological "flaws" in the design of
the Maimonides study, nor were there any procedural variations from that s t ~ ~ d y
to the present one which might fairly be designated methodological "improve-
ments." Thus, if one chooses to reject the Maimonides results, it still remains to
be demonstrated in just what manner they may have been spuriously generated.
Turning to the second line of approach, there were several incidental pro-
cedural changes from the original sntdy to the present study that possibly could
have influenced the failure to replicate. Some procedural variations were con-
ducted at S's own request: the use of magazine illustrations rather than art prints,
for example, and the technique of cond~lctingawakenings during, rather than at
the termination of, REM periods. Other variations were unavoidable, but S indi-
cated at the conclusion of the sntdy that he felt they may have adversely affected
his performance. Among these was the blocking of experimental nights (neces-
sitated by economic considerations). In the prior study, S had run one night at
With respect to the two approaches outlined above for interpreting the
present results, it surely is premature to be driven to either extreme: rejecting
the general uend of the Maimonides findings out of hand on the basis of a single
failure to replicate or falling back on the hypothesis that telepathic dream influ-
ence is a highly fragile phenomenon subject to the subtle interplay of many facili-
tating and inhibiting factors. While not all of the Maimonides investigations
themselves have produced statistically significant results, the trend of current evi-
dence is still positive. Any failure to replicate does raise questions that must be
answered by further research, but it cannot by itself invalidate a generally success-
ful series of apparently well-designed experiments. Still, one would have greater
confidence in the Maimonides results if the intra-subject replication within their
laboratory (Krippner & Ullman, 1969) had been paralleled here by an intra-sub-
ject replication across laboratories. There clearly are further experiments to be
done before any firm conclusion can be reached about the validity of the Maimo-
nides results. Meanwhile, one can understand the unsympathetic attitude some
observers will feel toward these results until they prove better able to "travel"
geographically and temporally than they did from the time of S's nights at
Maimonides to that of his nights in Wyoming.
REFERENCES
KRIPPNER,S. The paranormal dream and man's pliable future. Prychoanalytic Review,
1969, 56, 28-43.
KRIPPNER, S., & ULLMAN,M. Telepathic perception in the dream state: confirmatory
study using EEG-EOG monitoring techniques. Perceptual and Motor Skillr, 1969,
29, 915-918.
KRIPPNER,S., & ULLMAN, M. Telepathy and dreams: a conuolled experiment with elec-
troencephalogram-electro-oculogram monitoring. Journal o f Nervous and Mental
Di~ease,1970, 1 5 1 , 394-403.