BACKWARD CAUSATION

• Document Type: CREST
• Collection: STARGATE
• Document Number (FOIA) /ESDN (CREST): CIA-RDP96-00792R000701040002-4
• Release Decision: RIFPUB
• Original Classification: U
• Document Page Count: 5
• Document Creation Date: November 4, 2016
• Document Release Date: May 17, 2000
• Sequence Number: 2
• Content Type: RP

Approved For Release 2000/08/15: CIA-RDP96-00792R000701 -,4sation 79
Backward Causation
Perhaps no otherpsi phenomenon arouses so much philosophical disquiet
as does precognition. Although, in a sense, all our activity is directed towards
the future, the idea that it might be possible occasionally to catch a glimpse
of that future, as opposed simply to making inferences about it on the basis
of past experience, is an idea that many find outrageous, not least because it
seems to imply a backward causation. When, therefore, I was invited, again
by the Parapsychology Foundation, to participate in a conference on philoso-
phy and parapsychology, in Copenhagen in August 1976, I decided that back-
ward causation would make a suitably controversial topic for such an occasion.
Bob Brier (1974) had already set the cat among the pigeons with the publica-
tion of his book on this theme. I, for one, had found the book persuasive.
Moreover, the recent advent of observational theory was making a bid to inter-
pret all psi phenomena as a retrocausal effect produced at the point of feed-
back.
The two professional philosophers who took part in the discussion which
followed my paper (Parapsychology Foundation 1977), namely Shivesh Thakur
and Peter French, seemed amenable to the principle of backward causation,
unlike Anthony Flew who at a previous conference (Parapsychology Founda-
tion 1974) had attacked the notion as logically absurd and, with it, the very
idea of precognition in a literal sense. I had tried to challenge him at the time
but he would not concede anything to my objections; so did Brier, himself,
even more forcibly, pointing out that Flew had failed to recognize the elemen-
tary distinction between changing the past, which was a logical absurdity, and
influencing the past which was not. To this day Flew has clung to his position
and refuses to recognize this distinction.
Every paranormal phenomenon must, in the nature of the case, arouse
suspicion and skepticism, but the claim to be able to foretell events which have
not yet occurred, which is what we mean by precognition, is apt to raise doubts
78
Approved For Release 2000/08/15: CIA
as to whether this concept is even coherent. A number of eminent
philosophers, including some like the late C.D. Broad (1967) or C.W.K. Mun-
dle (1964) who were known for their interest in and involvement with para-
psychology, came reluctantly to the conclusion that precognition, taken in its
literal sense as a knowledge of the future, was not just impossible in point of
fact but was logically impossible.'
I want to start by considering some of these logical objections that have
been brought against the concept of precognition. It has been pointed out, for
example, that if an event has not yet occurred, then it cannot have any effects
and, in particular, it cannot be an object of knowledge. Something which does
not exist may be imagined, but it cannot literally be known. Moreover, even
if we disregard this obvious point, in order for a future event to become the
cause of my present precognition of it, it would have to exert a backward effect
in time and this, it has been said, is inadmissible on at least three counts. First,
because it is part of the meaning of the word cause that a cause should precede
its effect, hence a cause which followed its effect would be a contradiction in
terms. Secondly, and more seriously, if, forgetting about the semantics of the
case, we do posit a backward causation, this would be tantamount to asserting
that we could alter the past. But what has been, has been and nothing that
anyone can do after the event can change it in any way. Hence, the very idea
of backward causation is a radical absurdity. And, thirdly, even if we could
somehow overlook the two previous objections and persist with the idea of
backward causation, we should find ourselves committed to the following
paradox: If it were possible to intervene in the past then, in principle, we could
intervene in order to prevent our own birth. But, in that case, who was it
that intervened in the first place? We would thereby have negated our own
existence!
These are not by any means the only objections that have been raised with
respect to precognition and I shall have more to say on the point towards the
end of my talk but, in the meanwhile, I want to look very closely at the concept
of backward causation which I believe is the real root of the trouble. I hope
to show that, contrary to what many philosophers have maintained, there is
nothing radically wrong with this concept and hence it cannot be invoked as
a reason for rejecting precognition on a priori grounds. Furthermore, I shall
try to show that, once we have come to terms with this seemingly paradoxical
idea, we shall be ready to appreciate an exciting new development in ex-
perimental parapsychology, namely the study of backward or retroactive PK.
In what follows, I would like to acknowledge my debt to a young American
philosopher, Bob Brier, who alerted me to this problem and convinced me that
the a priori case against precognition could be met. Bob Brier propounded his
views in a brief but important monograph entitled Precognition and the
Philosophy of Science (Brier 1974), with the subtitle "An Essay on Backward
Causation." Brier, in turn, was following the lead of the English philosopher
RDP96-00792R000701040002-4
 Approved For Release 2000/08/15: CIA-RDP96-00792R000701040002-4
80 The Relentless Question ' Backward Causation 81
Michael Dummett (1964), whose "Bringing About the Past" first appeared in
the Philosophical Review. But, having acknowledged my intellectual debts, I
consider myself free to develop my theme in my own way and draw my own
conclusions. Let us return, therefore, to the objections which I have already
listed and see whether they are really as cogent as they may at first appear.
The idea that future events do not exist and can, therefore, have no con-
ceivable influence upon the present is a case of begging the question. For, in
fact, many if not most contemporary philosophers of science who have written
on the problem of time take a realist or objectivist view of the temporal order.
I mean by this that they deny that the expressions past, present, future have
anything more than a subjective or, at least, mind-dependent basis, indicating
the relationship between the observer and the event in question, like the ex-
pressions here and there. In a mindless universe, they argue, no particular in-
stant of the time series would have a preferential status over any other instant.
Each event would still stand to any other event in the relationship earlier than
or later than but none would stand out from the rest as representing a unique
now, that is, the instant which the universe had attained on the time series.
All events would simply coexist in the timeless sense in which we talk of facts
as existing or being the case. Now, I am not saying that the realist view of time
is necessarily the correct one or that the concept of now-ness or becoming, as
something extra which an event attains when it actually takes place, is a con-
cept with no clear meaning. I mention this view simply to show that one cannot
just assume, without further argument, that future events are any less real than
past or present events. And, certainly, those cosmologies which treat the
universe as a single four-dimensional continuum, in which every object is
represented by a certain determinate world-line, recognize no division of the
universe into a past, present and future.
Let us agree, then, that there are no a priori objections to our treating
future events as sufficiently real to have causal consequences. The question we
must then consider is whether the direction of such causation must always ex-
tend from earlier to later, never from later to earlier. Let us look again at the
reasons that are offered for refusing to admit any sort of backward causation.
First, I do not think that the purely verbal objection need to detain us for long.
It may well be the case that, in ordinary language, a cause is always understood
as preceding its effect, since this, after all, is what happens in ordinary ex-
perience. But, given new circumstances, linguistic usage tends to become
reasonably pliable and if we can show that there are paranormal cases where
the cause comes after its effect, I am quite sure we would not be deterred from
acknowledging them by the constraints of language. At worst we:might have
to write the word "cause" in adverted commas, to signal that this was a rather
special kind of cause that we are dealing with.
What is much more problematic is whether the admission of backward
causation would lead us intA1JpwG W1 aFoOtRe a4*200 Ole aftg
indeed commit us to asserting that the past could be altered and this, I would
agree, is nonsense. But, although we cannot alter the past in the sense of mak-
ing something to have happened which did not in fact happen or, conversely,
preventing something from having happened which did in fact happen, it does
not follow that what did happen might never have happened, or might have
happened differently, but for certain later events. We may find this supposi-
tion strange, but it is certainly not illogical. In fact, the case is no different
logically when we are considering the future. If what has been, has been, then,
equally, what will be, will be and in this sense one can no more alter the future
than one can alter the past. But, of course, we quite rightly believe that what
will be may depend on what we now do and, in this sense, we can influence
the future even if it makes no sense to talk of altering it. The fallacy which
previous philosophers have fallen into arises from the fact that they failed to
observe Brier's distinction between altering the past and influencing the past.
The point to grasp is that the past might not have been what it was, had it not
been for some action in the present.
We are now in a position to deal with the other paradox of the man who
by dint of backward causation cancelled his own birth. Given the fact that a
person was born on such and such a date, it follows that nothing thereafter can
alter this fact. But it does not follow logically that his having been born might
not have been due to some subsequent action that was taken. For example, a
pious Christian mother might well believe that her child would never have
been born, but for the fact that, at the appointed time, she had it baptized.
Now, we may mock her for her credulity as much as we please, but we cannot
fault her logic. The fact that we cannot use backward causation to cancel our
birth is no more problematic than the fact that we cannot cancel any event once
it has occurred.
Critics of backward causation point out that causation is an essentially
pragmatic concept. Thus, we say that A is the cause of B when we are satisfied
that we can bring about B by taking action involving A. However, if we know
that B has already occurred, then any action we take involving A would be
superfluous, since we have now obtained B whether or not we do A. Con-
versely, if B has failed to materialize at a given time t1. Hence, from the
pragmatic point of view, every instance we take involving A is then futile in-
sofar as it was designed to make B occur at time t1. Hence, from the prag-
I matic point of view, every instance of backward causation must be either super-
fluous or else futile.
Brier offers an amusing example to show, nevertheless, that backward
causation need not be devoid of pragmatic import. Suppose a man were to
discover, by sheer trial and error, that whenever he received a letter through
the post he had only to clap his hands and, when he then opened the envelope,
he would invariably find there a check made out in his favor, while, conversely,
CUBRDP#~'- eft 7 S 2*6 U U 7df'there would be no check. In this situation,
 Approved For Release 2000/08/15: CIA-RDP96-00792R000701040002-4
Brier asks, would any of us, whatever our philosophical prejudices, refrain from
clapping our hands? And, if so, could any of us honestly deny that clapping
our hands at the right moment was a necessary condition or cause of the check
having been inserted into the envelope at an earlier time? Now, you may say
at this point: well, supposing the envelope were transparent, what then? The
man would then know as soon as he caught sight of the letter either that there
was a check inside, in which case he would not bother to clap his hands, or that
there was no check, in which case no amount of hand-clapping would help
him! And yet, surely, a causal connection which depended for its validity on
the knowledge or ignorance of the agent would be a very queer sort of cause.
But consider, any causal connection whatsoever is circumscribed by cer-
tain boundary conditions. Indeed, as Ducasse (1969) and other philosophers
have stressed, causation actually involves a triadic relationship, A causes B
under conditions C. Striking a match will cause a flame but not if the match
is damp or the surface is too smooth. It is not, therefore, surprising if what we
might call the epistemic aspects of the situation should constitute the boun-
dary conditions where human actions are involved. Here, again, a comparison
with actions directed towards the future may help to clarify this point. Thus,
if I know that tomorrow I shall be killed, it follows logically that nothing I can
do today can prevent it, since knowing something logically implies that it is
the case, but, granted that I do not know whether I shall be killed, it is entirely
rational for me to take steps to avoid getting killed.
Similarly, if the letter I receive arrives in a transparent envelope, so that
I know that there is no check inside, no magic on my part can conjure one into
having been inserted, but, granted that the envelope is opaque and that I do
not know, it is entirely rational for me to clap my hands. What makes these
two cases appear different is that normally we do not know what lies in the
future, whereas normally we do know, or at least we can very easily find out,
what has happened in the past. But this is an empirical difference as between
past and future and., of course, there are many such empirical differences. The
logic of the two situations, however, is entirely symmetrical and it is the logical
aspect that now concerns us. From the foregoing we are, I submit, entitled to
conclude that there are no logical objections to saying that A is the cause of
B, so long as we believe that B would not have happened but for A, regardless
of Whether A is earlier than B or B is earlier than A.
Up to this point we have discussed the question of backward causation in
purely hypothetical terms, using examples drawn from the realm of fantasy or
superstition. I now want to turn to the actualities of parapsychological research
and consider its relevance in that context. It has been widely held by para-
psychologists that ESP is essentially independent of space, time and matter, so
that when we use our extrasensory powers, instead of relying on our sensory
channels and their associated brain-mechanisms, there is no inherent reason
as much as events occurring contemporaneously, just as there is no inherent
reason why we should not become aware of events occurring in remote places
or events shielded from us by intervening matter. In the case of PK, however,
it has been the practice to use target-systems located in the subject's immediate
vicinity and much less has been made of the potential time and space tran-
scendence of the phenomenon. And yet it would be very odd if PK behaved
any differently in this respect from ESP, considering that the two are so closely
linked that it is customary to subsume both phenomena under the generic
term psi. The trouble is, however, that it would, in the nature of the case, be
very difficult, if not impossible, to demonstrate unequivocally the existence of
a PK effect directed towards the future that would be the analog o:
precognitive ESP. For, whereas in a precognition experiment the subject ha
simply to record his guess at time t1, and then wait for the target event to oc-
cur at time t2, against which he can compare it, in the case of PK there is n(
objective way of registering mental effort at time t1 that might be responsible
for the target even at time t2. The point is that PK is known only through in
effects. It cannot be identified with any sort of conscious effort on the part o
the subject. Moreover, even if one went to the extreme of killing the subjec
between the time t1 when he is given the instruction to try exerting a Pk.
effect on the future and the time t2 when the effect is observed, one coulc
still not rule out the possibility that the effect was due to a delayed-action Pk
or, even, that the deceased subject was exerting a PK effect from the lift
beyond, as it were.
We may note in passing that the problem of demonstrating forward PF
is exactly matched by the problem of demonstrating backward ESP, otherwis,
known as retrocognition. In the latter case, although the subject's response ea
be recorded prior to verification, the verification would not be possible but fo
the existence of certain records in the present. Hence, there can be no un
equivocal test of retrocognition. Fortunately, there is no such logical barrier ii
the way of demonstrating unequivocally the existence of a backward or retroac
tive PK, which is what I want to talk about now. The way it can be done is a
follows: Some physical effect can be automatically recorded at time t1, bu
kept secret until later. At time t2, the subject is instructed to try producin,
this particular physical effect or outcome on this particular target-system any
then, but only then, is the result checked against the original record at tim
t1. This, basically, is the method used by Helmut Schmidt in his recent ex
periments on PK with time displacement, but others, too, have been workin
independently along similar lines elsewhere (Janin 1974). For a straightforwar
test of PK, Schmidt uses as his target system a device known as a "randor
number generator." This usually has a binary output such that, when it is opei
ating in isolation, the two possible outcomes will occur with approximate]
why we should not becooOJYffePbP jVa~ 2DDlf'S'Ue:IA-RDP96-0 0,OMM M ooothe one digit or the other, depending on th
equal frequency. The PK task is to upset the randomicity of the machine b
 arbitrary instruction that is laid down. A prominent feature of the Schmidt set-
up is the feedback system. This may take a variety of forms, visual or auditory,
for example, a recording pen that shifts from side to side or a fluctuating tone
etc. The principle, in every case, is that the feedback is controlled by the random
number generator in such a way that an excess of hits or successful trials will
produce a shift in one direction, while an excess of misses or unsuccessful trials
will produce a shift in the opposite direction. This feedback is not just an
amenity designed to engage or sustain the interest of the subject; it is, accord-
ing to Schmidt's (1975) mathematical model of psi, a crucial part of the whole
process. It is, indeed, the feedback which is said to activate the psi source.
However that may be, the subject is encouraged to concentrate his attention
on the feedback and to try getting it to move in the desired direction, and not
to bother about the random number generator which actually governs it.
This, as I say, is what happens in the straightforward or contemporaneous
type of PK test. To convert this into an experiment on retroactive PK, only one
modification is necessary. The output of the randomizer is now determined by
a solid state memory on which a sequence of digits has been recorded, based
on a previous output from the machine. Hence, instead of generating a fresh
sequence of random digits by being triggered from a source of quantum in-
determinacy or of electronic noise, it now merely repeats the series recorded
at this earlier time. So far as the subject is concerned, however, nothing has
changed; he does not need to know that the situation is any different from
what it is in the standard case. Nevertheless, if the subject is to succeed in mak-
ing a significant score under these new conditions, it is necessary for him to in-
fluence the behavior of the machine, not as it is currently, since it is now strictly
determined, but as it was when the target sequence was originally recorded.
His PK, in other words, is now operating retroactively. Although, so far, very
few such experiments have been reported, it is beginning to look as if PK with
time displacement may be a fact (Schmidt 1976).
In our department at Edinburgh, Richard Broughton, one of my
postgraduate students, has already started work on a project which depends on
the retroactive properties of PK. His point of departure is the notorious ex-
perimenter effect, that is, the idea that the experimenter himself, rather than
the ostensible subject, is the real psi source which is responsible for the positive
results which are observed. Without some such hypothesis, it is hard to explain
why certain parapsychologists obtain positive results time and time again while
others hardly ever do so. Indeed, Broughton had begun to wonder whether his
own positive results, which he had himself obtained earlier, while working on
his thesis project designed to demonstrate a hemispheric lateralization effect
in ESP, might not have been the product of such a psi experimenter effect. In
this new project, what he does is to generate a certain expectation in his sub-
jects regarding the kind of scores which they may expect, more specifically, that
they may expect to do better under one condition in which they perform their
r Causation
rw_ qir Felease 2000/08/15: CIA-RDP96-00792R00070104d1
9
6214
ESP task than they will under another condition, on account or certain
unspecified artifacts. Actually, there is no objective difference between the two
conditions under which they perform, and it is only when they go to the
computer to receive a printout of their scores at the termination of the experi-
ment, that they are told which was the high-scoring condition and which the
low-scoring condition, the assignment being arbitrarily determined by the
computer. This means that the expectation is generated in the subject, not at
the start of the experiment, but only when it is all over. The idea, if you follow
me, is that any significant difference that is then found as between these two
arbitrary categories must be due to the subject's retroactive PK. In other words,
what he expects to find at time t2 must influence the scoring level recorded at
time t1. A pilot experiment along these lines has already been run which did
in fact produce some significant differences and Richard Broughton has been
reporting on them to the Parapsychological Association Convention at Utrecht
(Broughton 1976). There are further refinements in his experiment which I
have not bothered to mention and a much larger experiment is planned for
next year, which the Parapsychology Foundation has generously agreed to sup-
port, but I have taken it as an illustration of the fact that backward causation
is beginning to enter into the current thinking of experimental parapsy-
chologists.
Nor is it only PK that operates retroactively. If we adopt the Schmidt
axiom, that all psi interactions depend critically on feedback, then ESP, too,
involves backward causation, proceeding from the moment at which feedback
is received to the earlier moment in time at which the response was given. Even
precognition, on the Schmidt model, is elicited not, in the first instance, by
the future event that is precognized, but rather by the subsequent confirma-
tion, whether by the subject himself or the experimenter, that the event in
question has come to pass.
Armed with the concept of backward causation let us finally turn back to
the problem of precognition which I introduced at the outset of this paper. Un-
doubtedly, one reason why precognition has engendered so much resistance is
that to many people it seems to cut at the root of the belief in free-will. If our
future actions could be known, it is feared, this would make a mockery of our
claim to be able to determine our actions by a spontaneous act of will. Apply-
ing the concept of backward causation, however, this fear would seem to be
unfounded. For, when I spontaneously choose or decide upon some course of
action, then my choice or decision could just as well have caused you to have
had a precognition of it the week before, as it could be the cause of your
remembering it a week later. Whether this entirely disposes of the uneasiness
we all feel at the thought of the future being already in some sense laid down
and knowable, I do not know. Certainly, we like to feel that the future lies
open before us waiting to be created, but whether this intuition is
philosophically justifiable, I must leave you to judge. I will say, however, that
Approved For Release 2000/08/15: CIA-RDP96-00792R000701040002-4
 86 pr19nyiWe4trOFZeIease 2000/08/15 : CIA-RDP96-00792R00070104004,zAcausation 87
there is one restriction we must impose on the scope of precognition if we are
to retain the idea of free-will. We clearly cannot both precognize our own future
actions and, at the same time, freely decide upon them when the time comes.
For, deciding implies being in a state of uncertainty which is terminated only
when the decision is made. But, if we already know what we are going to do,
it would be absurd to talk of our then deciding what to do. Having perfect pre-
cognition, which would include knowing all one's future actions, would in-
deed be incompatible with living one's life in the sort of way that characterizes
human existence.
I am not claiming that the concept of backward causation can solve all the
problems connected with the idea of precognition-there is, for example, the
notorious intervention paradox which raises quite different questions-but
what I have tried to do in this paper is to show that precognition cannot be
dismissed on logical grounds alone, and that the idea of backward causation,
so far from being nonsensical, is now being taken seriously by more and more
experimentalists as a basis for testing psi in the laboratory.b
It is curious that, as a consequence of what I regard as their faulty logic, both
Eisenbud and Braude (who acknowledges his indebtedness to Eisenbud) adopt
what they themselves call an "active" interpretation of precognition. This means
that the events said to have been precognized were actually brought about,
psychokinetically, by the precognizer! This has long been a favorite fall-back posi-
tion for those parapsychologists who, for whatever reason, cannot come to terms
with literal precognition. Alternatively, when the active interpretation becomes
too far-fetched even for its proponents, recourse can be had to superinferential
powers of extrapolation from contemporaneous clairvoyant knowledge. Such
antics, it would seem, are the price one has to pay for rejecting backward causa-
tion.
Notes
a. To be fair to Broad, he does attempt to rescue the concept ofprecognition
in the literal sense by toying with the idea of an additional dimension of time
(Broad 1967, pp. 199-204).
b. Despite this confident conclusion, resistance to the idea of backward causa-
tion and, hence, literal precognition, is still far from having subsided. Both Jule
Eisenbud (1982) and Stephen Braude (1986, chap. 5) repudiate it as an incoherent
idea. Braude, a professional philosopher, unlike Flew, does not fall into the same
trap, that offailing to distinguish between altering the past and affecting the past.
His argument is much more subtle, namely that `Any causal connection we iden-
tify will always be part of a larger causal nexus spreading indefinitely into the past
and future"whereas the putative counterclockwise causes are never part ofany such
wider network. My own response to this objection is to challenge the premises of
Braude' argument. I do not agree that normal causes are necessarily part of any
wider network. The obvious model for precognition is memory, indeed one is
tempted to regard precognition as simply memory operating in reverse. Now the
relationship between the events remembered and the present memory need be no
different from the relationship between the events precognized and the present
precognitive experience. It is true that, in the case of memory, one can fill in the
time gap by invoking memory traces in the brain; indeed this is what enables us
to regard memory as a normal rather than paranormal phenomenon (see my paper
on pages 110-122). But, clearly, this is not part of that causal nexus to which
Braude alludes and, in any case, Braude has elsewhere (1979) repudiated the trace
theory of memory. The point is that when we are dealing with an epistemic rela-
tionship there is simply no need for Braude's causal nexus, rather it is a case of the
percipient's mind being directed either onto its past experiences or, in the para-
normal case, onto its future experiences.
Approved For Release 2000/08/15: CIA-RDP96-00792R000701040002-4