CIA-RDP96-00787R000100130003-6

.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
:
APPENDIX 1
PRECOGNITION - A MEMORY OF THINGS FUTURE ?
G. Feinberg
Department of Physics
Columbia University, New York, N. Y. 10027
To be presented at the Conference on Quantum Physics and Parapsychology
Geneva, Switzerland, August 26 -27, 1974
.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

2.
Approved For Release 3000/08/10 : CIA-RDP96-00787R000100130003-6
theories is quite different than the above picture would suggest. Instead of forbidding
precognition from happening, these theories typically have sufficient symmetry to sug-
gest trict phenomena akin to precognition should occur in a manner qualita .. vely,
although not necessarily quantitatively, similar to the occurrence of retrocognition.
Indeed, phenomena involving a reversed time order of cause and effect are generally
excluded from consideration on the ground that they have not been observed, rather
ihen because the theory forbids them. This exclusion itself introduces an element of
csymmetry into the physical theories, which some physicists have felt was improper,
or required further explanation.
2
Thus, if such phenomena indeed occur, no change
in the fundamental equations of physics would be needed to describe them. Only a
change in the solutions used would be necessary.
The details of these aspects of physics relevent to this possibility will be
given below. However, it is worth noting first that the occurrence of physical effects
:hat propagate backwards in time may be related to precognition very indirectly. To
see this, we note that the information about the past that is available to any person
at a given time does not mainly consist of his sense data at that instant. Indeed, we
usually do not think of sense data as giving information about the past, although
strictly speaking it is the past we are observing, because of the finite time required
for any known type of signals to propagate across space. Instead, our information about
the post comes either from inferences we make from these observations, or through the
Poor! hoursfood process w= cat. merry, through which we can bring into our pres-
ent awareness observations that we, or others, have made in the past, and which have
somehow been stored in our brains.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
3.
$
A plausible analogy between information about the past and future would
suggest that if information about the future is available to a person at all, the main
source of it might well be observations that he or others will make in the future, and
which will then be stored in his brain. It might be expected that whatever the mecha-
nism of precognition, it could work more easily upon the future state of the percipi -
ents own brain than on the world outside. In other words, I am suggesting that pre-
cognition, if it exists, is basically a remembrance of things future, an analogy to
memory, rather than a perception of future events, an analogy to sense perceptions
of the very recent past. This suggestion has at least the merit of being fairly easy to
:
test through simple experiments, or perhaps even through a careful literature search.
D
i shall spell out below some of the simple consequences of this model for precogni-
tion, and how to test it. If it is correct, it would not directly indicate the physical
mechanism for precognition, any more than the existence of memory indicates its
physical mechanism. However, if it does turn out that memory can operate into the
future as well as into the past, it would suggest that the symmetry of physical laws
mentioned above is involved, and that physicists have been premature in discarding
those solutions to their equations that describe reversed time order of cause and
effect.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
1

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
4.
1
11. Time Symmetry of the Equations of Physics
The equations that describe the evolution in time of physical phenomena
ne a rather simple form according to relativity theory. A typical example, which
Hlestratos the main points, is the wave equation in one space dimension, whose form
: :
2ªø
ax
2
2
1
2
C
at
2
2
= p(x,+)
(1)
where c is the velocity at which the waves move through space.
In this equation,
represents the amplitude of some wave phenomenon, and p a material source for
e wave, For instance, Ø might represent an electric field strength, and p the
tribution of charge or current that produces the field. In a physical application of
Til eauction, we would take p to be a prescribed function of space and time, and
une the equation to calculate @ for all values of x and t . The values obtained
* 3 will depend on the value of p , but in a rather complex way. However, gen-
valy speaking, a change in p at one point in space and time, will lead to a change
2 at many points in space and time, in a way prescribed by the equation. A human
"ning, or an instrument, sensitive to the value of $ in some region of space-time will
m'afore receive different impulses depending on the value of p in other regions of
sese-time, and hence will know something about what is happening in those other
tions. Clearly, the relation between p and O is a critical factor in determining
"# regions of space-time are accessible to a particular observer through measurement
texting of the value of A at his location.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
5.
Because equation (1) is a second order partial differential equation in the
time, it has in general two sets of solutions. The particular form of equation (1) is
such that one set can be obtained from the other set by the change of t into -t ,
in both Ø and in p . We can study the character of these solutions by considering
the simple case in which o is a transient disturbance, such as a light bulb that is
turned on and off in a short period of time, and is limited to a small region of space.
We shall call the point at which p is localized xo, to . The solutions for this case
can then be described as follows. One solution, called retarded, has Ø = 0 for all
times earlier than to , everywhere in space. For times after t , Ø is non-zero
at the points
x ± ct . This corresponds to the generation of two pulses
of radiation, each travelling outward from the source point at velocity c . An ob-
server at a distance d from the point x would detect this radiation at a time t,
later than to by d/c , the time taken for the radiation to travel the distance d .
This retarded solution is the one generally chosen to represent the physical phenomena
described by the wave equation.
The other solution, obtained from the retarded solution by letting + -+ -t ,
is known as the advanced solution. It has the property that Ø = 0 for t later than
to , everywhere in space. For t earlier than t , Ø is non-zero at the points
x = x # ct . This solution may be interpreted either as two pulses of radiation
0
travelling outwards from the source, but backwards in time, or as two pulses coming
from sparial infinity, but forwards in time, to meet at the source at t . With either
0
interpretation of the advanced solution, there is associated with the disturbance et
to , effects at times earlier than : , rather than later than t , as for the retarded
0
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
1
6.
solution. An observer at a distance d from the point x would detect the radiation
corresponding to the advanced solution at a time t , earlier than t by d/c . In
0
the case of an electromagnetic wave, travelling at the speed of light, this time is
usually very short. When d corresponds to a distance of a few meters, d/c is about
1,/100 of a microsecond, so that the advanced notice of a disturbance available in
this way would not be' very useful. If one considers waves propagating more slowly,
such as sound waves, the advanced notice would be somewhat longer, but still too
short to be directly useful for precognition. However, indirect effects of advanced
waves are more promising, and will be discussed below.
Whatever use we could make of advanced waves, we must first ask whether they
actually occur in the world, as against occurring as mathematical solutions to equa-
tions. The general solution to equation (1) is a linear combination of the retarded
and advanced solutions, with unknown coefficients. As mentioned above, physicists
have usually, although not always, - supposed that the coefficient of the advanced
solution is zero, and only the retarded one is present. The reason for this is that ad-
vanced effects do not appear to occur, at least within some range of accuracy. The
evidence for this is simple. If there were advanced effects comparable in size to re-
tarded ones, many bizarre astronomical phenomena would be observed. For example,
two images would be seen of a planet, or other astronomical objects, displaced by
the distance that the object moves in twice the time it takes light to go from the ob-
ject to earth. For the planet Mars, these images would be displaced by more than
the planets apparent diameter, and would have been easily detected. Another exam-
ple is that phenomena that occur ar a well defined time at the piace of origin, such
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
7.
es eruption of a solar prominence , would appear to occur twice in the same spot,
once corresponding to the arrival of the advanced wave, and then later corresponding
to the arrival of the retarded wave. Since these phenomena have not been reported by
astronomers, we may conclude that advanced waves are not as strongly produced as
retarded waves.
However, this does not imply that they are not produced at all. Conceivably,
the ratio of strength of advanced and retarded waves is quite small, but not zero.
This would not necessarily make the advanced waves useless for precognition, but
would rather imply that precognition would not be as effective as ordinary percep-
tion, or as memory of the past, a conclusion for which there is ample evidence. Ex-
periments to detect a relatively small amount of advanced light wave are not hard
to imagine, and I have suggested some, that may soon be carried out by Prof. Riley '
Newman of the University of California. In the simplest such experiment, a light
source is turned on at a time that is very sharply defined, say to within 10-9 s
seconds.
A detector is placed at a distance of 10 meters from the source. The detector will
ordinarily indicate the presence of the retarded wave after about 3x 10-8 seconds
have passed, corresponding to the transit time of the light over the 10 meters. If
an advanced wave is also present, the detector would react to it at a time 3 x 10-8
seconds before the light is turned on, and this time is large enough that the ad-
vanced and retarded signals are easily distinguished. The consequence of not turn-
ing on the light after the advanced signal is derected is left to the reader to con-
sider. Professor Newman believes that an advanced wave of intensity as little as
one part in 10
19
of the retarded wave could be detected in this way, so we should
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
3.
ccon know if advanced electromagnetic waves occur. Advanced solutions fo: other
. kinds of wave motion, such as sound can be treated by similar mathematics. How-
over, since these motions generally involve a real medium through which the wave
moves, such as the atmosphere, it is unclear whether the interpretation would be the
some. No experimental evidence about such advanced solutions is known to me.
.
.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
9.
.
III. A Model for Precognition
1
In the following, I shall outline a very speculative model for precognition
that relies on advanced waves. The model is qualitative rather than quantitative,
because it involves workings of the brain, where detailed physical information is
unavailable. Yet I believe that model is sufficiently precise that it can easily be
tested, providing that precognition can be demonstrated at all.
We assume that when some sensory input reaches the brain, an oscillatory
variation of some internal patterns in the brain occurs, which is specific to the input.
This oscillation persists for some period of time in at least part of the brain. When
the person involved remembers the stimulus, what has happened is that the stored
oscillatory pattern has influenced another part of the brain, bringing the memory
into consciousness, or at least into something accessible to consciousness. Those
familiar with the literature on memory will recognize that I have given a very sketchy
description of one model for short term memory. There is some indication that long
term memory involves rather different mechanisms.
.
Suppose now that the oscillarory pattern set up by an external stimulus has
not only a retarded part, which propagates forward in time, but also an advanced
part, propagating backwards in time. Although we do not know what equations this
pattern would satisfy, it is not implausible that these equations are sufficiently
similar to Eq. (1) that both types of solution exist. As in the case of light waves,
the relative amounts of the two that are Involved in an actual situation is not de-
termined by the equation, and must be decided by experiment. I shall assume that
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
10.
the advanced part is non-zero, but presumably smaller than the retarded part, since
precognition in practice is not a very effective way of getting information. Since
the retarded part of the oscillation, which in this model allows memory of the past
to occur, is known to persist for at least some time, without great attenuation, it is
possible but not certain, that the advanced oscillation would be able to propagate
for a corresponding time into the past before the stimulus occurs. So at least over
this period of time, by a process similar to memory of the past, it could be possible
for the advanced pattern to be brought into consciousness, so that the person involved
would "remember" the future stimulus connected with the advanced pattern.
This in brief outline is the model for precognition that I wish to consider.
There are several qualitative features of this model that can be simply recognized.
1). One can only "remember" things that one will eventually sense, or learn about
through someone else's report. At least this is the case if one disregards the possibil-
ity of extrasensory information to be obtained at a later time and rememberd by this
mechanism at an earlier time. While this neglect may not be entirely justified, it
would seem a useful working hypothesis, since, in any event, the amount of informa-
tion obtained by extrasensory means is small compared to the other sources I am con-
sidering.
2). If the retarded oscillatory pattern is correlated with short term memory, and if
the latter has a relatively short term of operation, then we will expect that the ad-
vanced pattern would also have a similarly short range into the past. This would imply
that precognition would be effective only for events in the not very distant future,
perhaps on the scale of hours. Within this time period, precognition would be expected
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
11.
to show a "decay" curve similar to that shown by ordinary memory. Thus precognition
would become easier as the percipient approached to events more closely in time.
However, if precognition became a well controlled ability, it would become
possible to "pass" information back indefinitely into the past. For example, suppose
someone were going to observe an earthquake at noon, and became aware of it pre-
cognitively at 11:45. He could write out the sentence "There will be an earthquake
at noon", and show it to other people. The recording of this sentence would then
itself become a new stimulus, which could be recognized precognitively sometime
before it was real, or ideally, more than the 15 minutes warning gained by the imag-
ined precognition.
This process could be repeated indefinitely, and so the .
warning time increased indefinitely. Of course, several paradoxical results can be
reached in this way, and these will be alluded to below.
3). There should be little or no correlation between the spatial location of the
primary event and the ability to precognize it. This is because precognition is oper-
ating on the future state of the percipients own brain, rather than directly on the dis-
tant event. For example, if there were a supernova explosion in a distant galaxy,
whose light will reach earth 15 minutes from now, precognition would be able to
give a warning of that explosion sometime before the light reaches earth, about as
well as it could about an event that occurs in the percipients immediate vicinity.
4). We would expect that the same types of external and internal factors that are
known to effect ordinary memory, such as drugs, fatigue, age and training might be
expected to have similar effects on precognition. The demonstration of such differ-
ential effects would of course be very exciting for parapsychological studies.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
.

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
12.
5). If we omit from consideration the precognitive "chains" discussed under 2) above,
it should not be possible for anyone to precognize about any event that will occur
after that persons death, since no sensory input about that event could ever reach his
brain. This conclusion is independent of the length of time that the advanced pattern
can propagate into the past. It is consistent with one old legend to the effect that
prophets cannot foretell their own death, but inconsistent with other legends. Of
course, even if it is true that precognition cannot be used to foresee ones own death,
other explanations are available to account for this, and it is therefore not a pre-
diction very specific to the present model.
These properties that precognition should satisfy according to this model
suggest a number of experimental tests of the validity of the model. Several of these
tests will be discussed in the next section.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
13.
IV. Tests of the Future Memory Model of Precognition
In order for a model or explanation of any phenomena to have any value, it
must be possible to confront it with experimental tests, or to make new observations
of the phenomena about which the model makes specific predictions. This is not
hard to do for the "future memory" model of precognition, provided always that we
have fairly definite evidence that precognition is occurring in a specific instance.
The simplest aspect of the model to test is probably the prediction that a
percipient can precognize only those things he will eventually know through ordinary
perception .. In order to test this prediction, one might first make a search of the
literature on precognition to see whether accurate predictions have been made under
conditions that preclude the obtaining of the information by the percipient at anytime
after the prediction was made. If this turns out to be the case, it would be evidence
against the model.
A more convincing test would involve an experiment designed for the purpose.
The simplest version of this might be a precognition test in which the results are not
:
ever revealed to the subject. A slightly more sophisticated version would involve a
randomized decision pattern for revealing the data a fixed time after the trial. A
comparison of the rate of success when the data are revealed as against those in which
they are not could indicate the validity or invalidity of the model even if the level
of precognition was low.
A possible objection to such experiments is that it is difficult to ensure that
the object will never have access to the data at any future time. However, if point
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
14.
2) above is correct, information obtained long after the trial has taken place would
be useless, because of the decay of the advanced pattern at times long before it is
established. A test of this point can also be carried out along the lines described
above, if it is found that the basic effect exists. To do this, it would be necessary to
give information about the data to the percipient at various time intervals after the
trial, and investigate how the success rate of precognition would vary with this time
dolay. If the model is correct, there should be a dependence on time delay that is
similar to the dependence of short term memory on the time lapse after the initiating
stimulus. I am assuming here that there is no precognitive equivalent for long term
memory, as the latter appears to involve a kind of static chemical storage, rather than
Q
cn oscillating pattern in the brain. If this assumption were wrong, the particular test
just described would give negative results, and precognition would be possible of any
event up to the death of the percipient. This possibility, while it should be kept in
mind, seems less likely to me.
Another testable aspect of the model is that the success rate of precognition
should not depend on the spatial location, or any other physical attributes of the
event being precognized. This could be tested by varying such attributes of the tar-
get, but keeping the information about it eventually furnished to the percipient, and
the time advance, constant. Under these conditions the success rate would not be ex-
pected to vary, even if the target is at astronomical distances, or is extremely well
shielded. These properties are in qualitative agreement with some anecdoted reports
of precognition.
Finally, if the model is correct, we would expect precognitive ability to vary
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
15.
greatly from person to person, just as short term memory does. In fact, it is possible
that the same people that have good short term memories would also be good at pre-
cognition, although that connection is not definite. Nevertheless, it would be worth
testing people with good memories for precognitive abilities. Furthermore, it should
ho possible to improve precognitive ability by using the techniques that are used to
improve short term memory. Probably, these techniques would improve the accessibil-
ity of the advanced pattern to the conscious mind, rather than affecting the absolute
. amount of advanced pattern generated by the event. The latter amount is probably
determined by the basic laws by which the brain operates, and not subject to altera-
tion by training.
I believe that if a series of experiments of the type described is carried out
with a subject who has real precognitive ability, it would definitely decide whether
the memory model of precognition is valid. Perhaps what is even more important,
such experiments would furnish much new information about precognition, which would
be useful in any case, even if the model should prove wrong.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

16.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
V. Conclusions and Discussion
Since I believe in a materialist description of natural phenomena, including
those involving human beings, I believe that if advanced effects occur in the human
brain, they must occur elsewhere in the world, since brains are made of the same kind
of matter as other objects are. It therefore appears plausible to me that if the future
memory model of precognition is valid, that it should also be possible to detect ad-
vanced effects outside of the human brain, perhaps in the type of experiment that Dr.
Newman plans to carry out. Conversely, if his experiments gave a positive result,
showing that advanced effects do occur, it would lead more credence to the idea that
.
they are what is involved in precognition. Even if the advanced effects are very
:
small compared with the retarded ones, this would not rule out their playing a role in
brain processes, provided that they are larger than the "noise" background. It would
be interesting to estimate how small the ratio of advanced to retarded effects could
be in the brain, and still have the advanced effects be useful; but I have not tried to
do this.
Physicists have sometimes raised the objection that any occurrence of ad-
vanced effects in nature would lead to unavoidable paradoxes, and causal anomalies.
Careful analysis of this question has not substantiated this claim,"
2
but the question
is not completely closed. However, it should be recognized that if such problems
exist, they would also occur just from the existence of precognition, whatever the
physical interpretation of the phenomenon. The analyses that have been carried out
of possible causal anomalies due to advanced effects could usefully be applied to the
precognition directly, rather than to its physical interpretation. I believe that the
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
17.
limited accuracy of precognition, and especially the impossibility of knowing whether
a given precognition will turn out to be accurate until after the event has occurred,
eliminate the possibility of such causal anomalies, but it would be worthwhile to carry
the analysis through.
Finally, it would be interesting to follow up on a suggestion that is sometimes
made, and investigate the extent to which all valid reports of extrasensory perception
can be explained in terms of ordinary perception, combined with precognition. My
impression is that many such reports can be so explained, but I do not know if they
all can be. If it were possible to do so, this would mark a substantial advance in our
understanding of these phenomena, and in linking them to other aspects of the physical
world that we know better.
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6

Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6
18.
References
:
. .
See for example, J. B. Rhine, The Reach of the Mind , (William Sloane
Associates, New York, 1947).
For example, R. P. Feynman and J. A. Wheeler, Review of Modern Physics,
17, 157 (1945).
Approved For Release 2000/08/10 : CIA-RDP96-00787R000100130003-6