John Wheeler 67

have been reported to me, indicate that Bohr's expectations are

indeed fulfilled.

I believe that you suggested a hypothetical cosmologicalversion of

this experiment. Can you tell us about that?
Yes. The experiment that I've just been talking about is on a
laboratory scale. However, there's nothing in principle to pre-
vent the scale being some five billion light years if we use a light
source such as a quasar. Fortunately there is a quasar by chance
situated in space in such a way that its light comes to us by two
different routes- each on a different side of an intervening galaxy
which just happens to be in the same line of sight in the sky. The
two beams are bent by the gravitational field of the intervening
galaxy so that they converge on the eye of an observer here on
Earth. This so-called gravitational lens effect provides us in
principle with a means to do the delayed-choice experiment at
the cosmological level, even though technically it is beyond us.
The photons reaching us start out more than five billion years
ago - that is, before there was anyone on Earth. Waiting here on
Earth we can today cast a die and at the very last minute decide
whether we will observe an interference photon (that is a photon
which has come, as we jokingly describe, 'both ways') or change
our method of registration so that we will find out which way the
photon has come. And yet the photon has already accomplished
most of its travel by the time we make this decision. So this is
delayed choice with a vengeance!
But let me insert a caveat. The very words, we have to
appreciate, are wrong. It's wrong to say that the photon goes by
one route or another, or by both. Picturesque though they are,
those phrases are only suggestive. The elementary quantum
phenomenon - the Great Smoky Dragon - is only brought to a
close by an irreversible act of amplification in the counter, at our
telescope. That Great Smoky Dragon had its heart back in that
distant quasar. To the extent that it forms a part of what we call
reality, we have to say that we ourselves have an undeniable part
in shaping what we have always called the past. The past is not
68 The ghost in the atom

really the past until it has been registered. Or put another way,
the past has no meaning or existence unless it exists as a record
in the present.
Does that mean that we, as conscious observers, are responsiblefor
the concretereality of the universe?
That's probably too loosely stated. I would rather return to the
concept of meaning"as a joint product of all the information that
is exchanged between those who communicate. And that infor-
mation comes back to a set of many ele~entary quantum
phenomena. Naturally most of t~ese elementary quantum
phenomena are individually at too low a scale of energy to be felt.
But we know that many a chance yes/no adds up to the definite-
ness of how much. We know that the pressure of one hand on
a~other goes back to atomic impact from the atoms on one hand
to the atoms on the other. And each atomic impact in the last
analysis is close to a yes/no process~ So the pathway between the
quantum and knowledge of meaning is a long one. The work of
Hubel and Wiesel on the visual system of the brain has been
enough to show that.
If we understoodthe workings of human or even animal brains,would
that help to solve the problem of understanding the link between
quantum theory and meaning? -
The nature of the brain is certainly a sophisticated and challeng-
ing subject and extremely important. But I cannot believe that the
physical elements for arriving at the description of the physical
world are to be found there. Meaning, yes. That may depend on
our unravelling the detailed mechanisms of the brain.
But it's interesting to recognize that the human brain may not
be so impressive as we're accustomed to believe. The studies of
the evolution of ~he eye have shown that the eye has indepen-
dently evolved in other species over 40 times. And insofar as the
eye is the window of tl;le mind, we can believe that the evolution
of the mind too is not so special as we might think.
Can you foresee any further experiments to test the conceptual
foundations of quantum theory?
 John Wheeler 69

I am too stupid to see any immediate experiment or test. I would

rather hope that we shall find a deeper conceptual foundation
from which we can derive quantum theory; a foundation based
upon distinguishability on the one hand and complementarity
on the other. Complementarity restricts our freedom in the
asking of questions, whilst distinguishability clarifies the result
of answering these questions. But spelling out the details of any
such derivation is still beyond us. So it's on this conceptual side
- the derivation side - rather than the experimental side that I
think the greatest hope of progress lies.
 5
Rudolf Peierls

Sir Rudolf Peierls retired from th~ Wykeham Chair of

Physics at the University of Oxford in 1974. His interest in
quantum mechanics dates from his early studies under
Sommerfeld, Heisenberg and Pauli, and his frequent visits
to Bohr's institute in Copenhagen. He has therefore been
acquainted with the 'Copenhagen interpretation' for half a
century, and he still finds it satisfactory.

How did you first become interested in the conceptualfoundations of

quantum mechanics?
I was a graduate student when quantum mechanics started and
of course it was an exciting time: together with trying to under-
stand how to use quantum mechanics, which obviously we all
wanted to do, we also had to understand what it meant.
And were you influenced by the Einstein-Bohr debate,which we read
about now as a great historical episode?Is this something which had
an impact on you?
No, because I learnt about it only later, not while it was taking
place, but we, of course, were sure that on that particular debate
Bohr was right and Einstein was wrong.
Well, I'd like to ask you about that, because as far as the textbooks
would have us believe it's certainly consideredthese days that Bohr's
Copenhagen interpretation is the official view. Yet, strangely, it
seems rather difficult to find people these days who are prepared to
pin their coloursto the Bohr mast. Do you think that the Copenhagen
interpretation is still the official view?
 Rudolf Peierls '/ I

Well, first of all I object to the term Copenhagen interprett11thm.

Why is that?
Because this sounds as if there were several interpretation ■ of
quantum mechanics. There is only one. There is only one way in
which you can understand quantum mechanics. There are a
number of people who are unhappy about this, and are trying to
find something else. But nobody has found anything else which
is consistent yet, so when you refer to the Copenhagen interpre-
tation of the mechanics what you really mean is quantum
mechanics. And therefore the majority of physicists don't use the
term; it's mostly used by philosophers.
So Bohr's interpretation, in your opinion, is the only one that we can
really take seriously at present, although there are attempts to find
alternatives. Perhaps you could tell us what you understand by the
Copenhageninterpretation, if I dare call it that?
I have to say, to start with, that it is a little hard to get used to
because it seems to contradict our intuition in many respects.
Our intuition tells us, for example, that if you have an electron or
some other particle, it is always to be found at some place with
some particular velocity. Quantum mechanics, however, tells
you that you have to use these concepts - position and velocity-
with caution because they may not necessarily have a meaning in
the context of an experimental situation. Our intuition, of course,
has developed from the experience of everyday life, which is on
such a different scale from the atom, that these quantum effects -
these complications - are unimportant. Sometimes we mee~
everyday situations where concepts that seem intuitively obvi-
ous suddenly lose their meaning. The simplest is the concept of
up and down, which at first has an obvious intuitive meaning,
until you ask the question of whether Australia is below us or
above us. You then realize that there is no answer to the question.
Now, quantum mechanics meets situations like that all the
time, and therefore we have to realize that we should invoke only
concepts which have a meaning inasmuch as they can be con-
nected to some real or at least possible experiments. That gives
72 The ghost in the atom

gives many people the trouble that they want to see reality; for
example, they want to say 'Well, maybe I can't observe where the
electron is, but really it ought to be somewhere'. The word 'real'
here is a concept which is not clearly defined, and in my opinion
all the worries about quantum theory are due to the fact that one
is using terms which are not defined. Reality, of course, in
everyday life is quite clear; this table we are sitting at is real
because I can see it, I ·can touch it, and I feel a pain if I knock it;
but obviously when you talk about the reality of an electron this
is not whaf you mean.
Could I interrupt you, and ask you whether you believe that if we
don't look at the table because we're in another room perhaps,
whether it is still really there?
Oh yes. Because there are many ways in which the table's
existence makes itself felt. On the everyday scale of classical
physics, where. an observation does· not appreciably interfere
with the object that's being observed, you can talk lightly about
all these concepts without gettjng into trouble. But in quantum
mechanics, it's different because any process of observation
must involve an interference with the object you are observing.
And therefore, in talking about what the object is doi,ng, we must
be very specific about what we observe, or what we're free to
observe.
Bohr, as I understand it, expressed it this way: if we're to talk about
reality, it's always within the context of a specified experimental
arrangement; you've got to say precisely what you're going to
measure, and how you're going to do it, before you can say what is
actually going on.
That is right.
Is it correct then that we can't think of an electron as beingjust like a
scaleddown version of say a billiardball, in the sense that we can't say
that it has a position or has a motion, until we've actually measured
either its position or its motion? In the absence of a measurement we
can't say that it has either of these qualities?
 Rudolf Peierls

Yes, I entirely agree ~ith this way of putting it.

Of course this makes the external world appear rather ghostly,
because it seems to require the presence of an observer before it
exhibits definite attributes. Many people have drawn the conclusion
that therefore mind must play some fundamental role in physics
becauseit's only when we're talking about observations that we can.
really talk about reality. Do you think mind has such a role to play in
physics, or couldwe replacethe observerwith some inanimatedevice?
No we couldn't. That brings in a very interesting question. In
quantum mechanics we always talk in terms of what is called the
wave function, or a state function of a system, and this is a
mathematical object which represents our knowledge of the
system; knowledge of an electron for example. Now, when we
make an observation, ~e have to replace that wave function
description by the new one which takes into account our new
knowledge of the system. And there has been a lot of speculation
of what's involved in this 'collapse of the wave packet' as it is
sometimes called.

Which is the sudden change in the wave function that occurswhen an

observation is made?
That's right. Now, let's think about the way an experiment- an
observation - is carried out. Suppose you have an apparatus
which tells you whether or _nota radioactive atom has decayed by
the position of a pointer on a dial. You cpuld describe this
apparatus in terms of conventional physics. But before you look
at the apparatus there are two possibilities for what the result
might be, and quantum mechanics will give you the probability
of the pointer being in one position or the other. And th~n you
say, alright, you have to look at it, so you shine light on the
pointer, but again you only know the probability of the light
being reflected in one direction or the other. So it goes on until
the moment at which you can throw away one possibility and
keep only the other is when you finally become consciousof the
fact that the experiment has given one result.
74 The ghost in the atom

So you think consciousness plays a crucial role in the nature of

reality?
I don't know what reality is.
Well, let me put it this way. Supposing that rather than getting a
human being to conduct t~is experiment, we replace him by a very
advanced computing system, or maybe even something more modest
like a camera.Can the operation of a camerarecordingthe position of
the pointer on film be considered to collapse the wave function,
putting that radioactiveatom into a concrete condition?
No, it will not, because you can of course describe the operation
of your camera or of your computer by the laws of physics, and
what you will find is that when the camera has been exposed or
when the computer has taken in the information, both options
are still open, and so there's no collapse of the wave packet.
So the ghostliness of the micro-worldgets amplified up and becomesa
ghostliness of the cameraor the computer?
Well, I wouldn't call it ghostliness.
Indecision?
Yes, that relates to knowledge. You see, the quantum mechanical
description is in terms of knowledge. And knowledge requires
somebody who knows.
But can a computer know?
I would say not.
So this does seem to suggest that there is a quality of human beings,
call it the mind, that distinguishes us from the other objects in our
environment and which is absolutely crucial for making sense of
fundamental physics?
I think that's right. And, in fact, that has an interesting conse-
quence because some people say, 'Well, supposing you include
the observer in yqur quantum mechanical or wave function
description, you could write down equations describing the
motion of every electron in every cell of the brain of the observer.'
 Rudolf Peierls 75

You couldn't actually do it, but in principle such equations

should exist. Having established this wave function the question
would be: 'whose knowledge does this represent?' There's no
easy answer to that.
No, I'm sure there isn't!
I think the way out of this is simply that the premise that you can
describe in terms of physics the whole function of a human being
(or any other living being), including its knowledge, and its
consciousness, is untenable. There is still something missing.
Well, surely one difficulty with this point of view is that there was
obviously a time beforethere were any human beings and presumably
beforethere were any observersof any sort. Can we think that in some
sense the universe was unreal or undecided before there were any
human beingsaroundto exorcisethe ghost worldsof quantum theory?
No. Because we now have some information about the origins of
the world. We can see around us in the universe many traces of
what happened there before. We haven't understood all of it
clearly, but the information is there. We can therefore set up a
description of the universe in terms of the information which is
available to us.
This is a very interesting idea. You're saying that in a sense our
existence as observers here and now, 15 billion years after the big
bang, is in some sense responsible for the reality of that big bang
becausewe're looking back and seeing the traces of it.
Again I object to your saying reality. I don't know what that is.
The point is I'm not saying that our thinking about the universe
creates it as such; only that it creates a description. If physics
consists of a description of what we see or what we might see and
what we will see, and if there is nobody available to observe this
system, then there can be no description.
That seems reasonable.But, of course,Einstein would have disagreed,
I'm sure, very strongly with what you're saying because he believed
that reality was something that we just uncover by our observations.
Do you think Einstein was completely wrong on this one?
 ,
76 The ghost in the atom

I think so. Despite the enormous respect we all have for Einstein
for the many things he discovered in physics, we have to accept
that he was not willing to adjust himself to the implications of
quantum mechanics. You see, there is no clear way in which we
can define this concept of reality. Now, it is true there are many
valid concepts which we cannot clearly define, and therefore it
doesn't necessarily follow that reality can't have a meaning. But
if we try to maintain this Einstein ideal that there must be such a
thing as reality, then we get into lots of logical troubles with
quantum mechanics. People have spent 60 years or so, trying
very hard to find a way out of those troubles but they haven't
found it. It seems to me very unlikely that such a way exists.
There does seem to be a very powerful, even emotional, appeal for
believing in concrete reality, or objective reality as Einstein would
have it; that is, somehow to write ourselves out of the picture.
Personally,I've always found it curious that scientists should want to
displace the mind or the observerfrom the centre of things becauseit
seems to me appealing to have us there. Why do you think there is this
restless search by many physicists to find some sort of vestige of
Einstein's vision of objective reality which doesn't depend on the
mind?
Well, for the reasons you've just given, except that I do not think
there are so many physicists who are worried about this. I think
it is only a very small minority.

Perhaps they arejust vocal.

They are vocal. In fact, I was asked the other day why it is that so
few people are willing to stand up and defend Bohr's view,.and I
didn't have an answer on the spot. But the answer is, of course,
that if somebody published a paper arguing that two and two
makes five, there wouldn't be many mathematicians writing
papers to defend the conventional view!
Of course,one of the things that Einstein did was to invent a thought
experiment which promised to throw new light on this problem.Now,
recently, with the results of Aspects's experiments and other similar
 Rudolf Peierls 77

experiments, we've seen that these ideas can actually br lt'Hlrd. l Jo

you think that the results fit into an expected pattern or do you think
Aspect's experiments tell us anything new about quantum mechanics?
No, they don't. Of course it is always nice to have a theoretical
prediction verified by experiment, because in the past we've had
surprises. But physics is not changed by the fact that these
experiments have given results which agree with the predictions
of quantum mechanics. If they had come out otherwise, then we
would have been in real trouble, because then we would have
had to abandon at least some part of our existing scheme, and in
fact it's very difficult to imagine a theory which can reproduce all
those many results of quantum mechanics which have already
been very accurately verified, and introduce something new
with these few erudite experiments. But, fortunately, that's not
what happened, for the experiments have agreed with quantum
mechanics.
Now there was a time when people suggested the possibility of
hidden variable theories - theories in which one could think of the
indecisionor indeterminacy of a quantum particle as being due to the
particle being jiggled around by a set of complicated random forces
that we couldn't perceive, muchisas in thermodynamics one has
particlesjiggled around by a collection of complicatedforces due to
molecular bombardment. Do you think that Aspect's experiments
have put an end to such theories, or is there still a way in which they
can survive?
Well, if people are obstinate in opposing the accepted view they
can think of many new possibilities, but there is no sensible view
of hidden variables which doesn't conflict with these experimen-
tal results. That was proved by John Bell, who has great merit in
having established this. Prior to that there was a proof due to the
mathematician von Neumann, but he made an assumption
which is not really necessary. So, I think the answer is that these
experiments, at least, dispose of all existing hidden variables
theories, but perhaps somebody can still come up with one
which is compatible with these experiments.
78 The ghost in the atom

One possibility is to abandon the idea of locality, that is, to entertain

the possibility of some sort of faster-than-light signalling,so that the
events taking place at separated places can somehow conspire
together simultaneously; I think Einstein referred to this as 'spooky
action at a distance'. If one is prepared to entertain the possibility of
such instantaneous communication, then I suppose it would be possi-
ble to retain an objective view of reality and yet still be in keeping with
the results of Aspect's ex11-eriment.
It becomes a very funny view of reality if you do that. First of all,
if there were any real possibility of transmitting signals instan-
taneously or faster than light, then of course we'd again be in
very great trouble with the theory of relativity.

And it might then be possible to transmit signals backwards in time,

and perhaps influence our own past with all the attendant paradoxes
that that would produce.
Indeed. But of course if you think about the consequences of
these new types of experiments they open up no way of transmit-
ting signals faster than light.

But if we have correlation between separated events, it seems as

though they must be conspiring in some way, and letting each other·
know what's going on. Can you think of an easy way of explaining
why that isn't so?
The original thought experiment we are talking about, involyes
two particles with spin. When you measure the spin in any
particular direction you get a definite result - either plus or
minus. The amusing point is that if you measure the spin of one
particle in the vertical direction, say, then you can predict the
spin of the other side equally in the vertical direction, and if you
were to measure its horizontal component then you can also
predict the horizontal component of the other side. And this
makes people think that by choosing whether to measure either
the vertical or the horizontal component this is changing the
situation on the other particle in some way. But, in fact, it doesn't.
Of course if you know the answer for one particle, then you also
 Rudolf Peierls 79

know what's true on the other. But if you make a measurement of

either vertical or horizontal spin,· and do not disclose the answer,
then nothing has changed for the other particle, and therefore
this cannot be used to transmit a signal.
You have no control over the outcome of your particular measure-
ment, and so you can't control the outcome of the other one; all you
know is that having per/ormed your measurement then the result of
the other correspondingmeasurement is fixed.
Yes, but whether you perform tli.e measurement in one direction
or the other doesn't alter the state of affairs on the other side. So
there is no way in which this allows rapid transmission of
signals. If you think of your hidden variables then you would
have to invent some variables which you can never measure,
which you can in principle never know the answer to, and which
somehow or other communicate over long distances but without
at the same time affecting the physical situation. To me that
seems to be such an unattractive view that, even if you could
make it consistent with Aspect's experiments, I would prefer the
present interpretation.
You make such a position seem very contrived. Could I turn to the
subject of cosmology, because these days there's a great deal of
interest in applying quantum theory to the entire universe. And there
we run into a severe interpretational problem because if the entire
universe consists of everything including observers, we have a diffi-
culty with how a measurement of the quantum state of the whole
universe can ever be made. What's your reaction to that?
Well, I think it's quite clear that this would never be possible.
One difficulty is that we do most of our illustrations, most of our
exercises, in quantum mechanics by thinking of a system and
saying 'Right, now we have measured the state of the system
completely and that's where we're going to start'. Technically,
that's what we call a pure state. Now, we never meet that in any
practical situation. There are always so many variables left over
which could, in principle, be measured but which we don't have
the time or the energy to follow up. This is similar to the situation
80 The ghost in the atom

we have in classical physics where one seldom claims to have

measured everything possible. Or to statistical mechanics, where
you leave the behaviour of many of the individual molecules
undetermined, and, instead, consider only their average
behaviour. In the univer·se we have this problem, only more so.
But it is possiblewithin the context of classicalphysics to make sense
of what the entire universe is doing. We could envisage, in principle
(as indeed Laplace did), having information about all of the particles
and their trajectories and in some sense predicting all future
behaviour. If you try to do that sort of thing with quantum mech~nics
you come up against the stumbling block of including the observer.
Well, you could never predict all the future behaviour, that's in
,the nature of quantum mechanics, which is not determin~stic.
But you could in principle write down a quantum mechanical
equation, a wave equation, for the whole universe.
People do, yes. But the thing is, does it mean anything?
Well, it would mean something only if you .could ascertain the
initial conditions, and find out what the state was at some
particular time in microscopic detail.
But although one might not be able actually to gain that information
is it meaningful to even think in terms of the wave function of the
universe?
You can think about it.
But does it make any sense?
I think so, because there would still be some consequences you
could derive from its behaviour which would be independent of
the nature of the original observation that somebody might have
made. Those consequences are going to be useful, so I think it is
legitimate to speculate about such a wave function. But we can't
actually carry this out in practice because we'll never be able to
do the whole observation. And you said, when you specified that
wave function of the universe, it would include all observers, and
there again we come onto this question of whether biology is a
 Rudolf Peierls 81

part of physics in the same sense in which we now know that

chemistry is ultimately a part of physics. And that is not proven.
Many people tend to assume it, but it may not be true.
You mean that there may be qualitatively new features which emerge
when structures becomesufficiently complex?
When they become alive.
And life itself is not something that can be reducedto the properties of
atoms?
I don't think anything terribly mysterious is to be expected here;
it's rather like what happened to physics in the nineteenth
century when, at first, scientists believed that any explanation
had to involve a mechanism, and that mechanics was the whole
of physics. When they met electric and magnetic phenomena,
physicists tried to explain these in terms of some kind of
mechanism. Maxwell even tried to do that, but then he realized,
and other people realized, that this didn't make sense because
electricity and magnetism were physical concepts in their own
right, not contradicting but adding to and enriching mechanics.
And in that sense I think we won't have finished with the
fundamentals of biology until we have enriched our knowledge
of physics with some new concepts. What these will be I wouldn't
like to say.
Some of the people who have thought about the wave function for
the whole universe have felt obliged to adopt the so-called many-
universes interpretation of quantum mechanics,where one envisages
that all possiblequantum alternatives in some sense co-exist. What do
you think about this interpretation?
That's making things unnecessarily complicated. Since we have
no means of seeing or ever communicating with the other
universes, why invent them? There is one way of thinking along
those lines which I regard as sensible but unnecessarily high-
brow. Quantum mechanics can only make predictions from
given initial information. When you've made some observations,
you know something about your system, and then quantum
82 The ghost in the atom

mechanics can tell you what are the chances of future experi-
ments giving another set of results. So, in a sense, you can say
that quantum mechanics can be represented as a dictionary
listing all possible outcomes of all possible initial conditions.
Now, if instead of 'dictionary' you simply use the term 'many
universes', then we are in line with Everett and the other pro-
tagonists of the many-universes idea. In other words, there are
certainly many possibilities allowed for by quantum mechanics
and we, by our observations, find out which possibility we
actually see. In the more conventional language of Everett's
interpretation, what you say is that you have to make an observa-
tion to see which of the universesyou are in. But I prefer to use the
word 'possibilities' or 'dictionary of possibilities', rather than
'universes'.
 6
David Deutsch

David Deutsch is a Research Fellow at the Department of

Astrophysics, Oxford, _and at the University of Texas at
Austin. He has had a long-standing interest in the concep-
tual foundations of physics in general and quantum
mechanics in particular. He argues here for the many-
universes interpretation.

Can I begin by asking you to give a brief description of the many-

universes interpretation?
The idea is that there are parallel entire universes which include
all the galaxies, stars and planets, all existing at the same time,
and in a certain sense in the same space. And normally not
communicating with each other. But if there were no communi-
cation at all then there wouldn't be any point to our postulating
the other universes. The reason why we have to postulate them is
that, in experiments on a microscopic level in quantum theory,
they do in fact have some influence on each other.
Before we get into that, can I just clarify that this is correct:in some
sense, 'out there' there are other universes much the same as this one
existing alongside ours but unconnected to it through our own time
and space?
That's right.
So where are these other universes?
As I said, in a sense they are here sharing the same space and time
with us. But in another sense they are 'elsewhere' because the
same theory which predicts their existence also predicts that we
84 The ghost in the atom

can only detect them indirectly. We can never go there or com-

municate with them in any large-scale way.
Now why should we believe in such a monstrous suggestion?
I suppose the first reason is that the theory which predicts them
is the simplest interpretation of quantum theory, and we believe
quantum theory because of its enormous experimental success:
it really has been the most successful physical theory in history.
You say it's the simplest interpretation of quantum theory, but it
seems to me like a very complicated interpretation, or at least an
interpretation that involves some pretty bizarre ideas. In what sense
is it the simplest?
It is by far the simplest in that it involves the fewest additional
assumptions beyond those which correctly predict the results of
experiments. All theories in physics predict some things which
are directly amenable to experiment and some which aren't. For
example, our theories of the stars predict things one could
measure, like how brightly they will shine, and when they're
going to go supernova. But they also predict things like the
temperature at the centre of the star, which we cannot measure
directly. We believe these ideas, including their unobservable
predictions, because they are the simp,est way of explaining the
things we can observe within a consistent physical theory.
Now the other interpretations of quantum theory also involve
rather counter-intuitive assumptions about reality. In some of
them consciousness - human consciousness - has a direct bear-
ing on the nature of physical reality, so that nothing exists until
it is observed. This is, in my view, a far more spectacular and
actually unacceptable consequence of the theory than the idea of
parallel universes.
So the parallel universes are cheap on assumptions but expensive on
universes.
Exactly right. In physics we always try to make things cheap on
assumptions.
 David Deutsch 85

How many of these other universes are there?

The exact number depends on the details of physical theories
which we don't know well enough yet. I think it's safe to say that
there is a very large, probably infinite, number of these univer-
ses. Many of them are very different from ours, but some of them
differ only in some minute detail like the position of a book on a
table, and are identical in every other respect.

Can you say something about how these universes come into being.
Are they there all the time, or do they increaseor perhaps decreasein
number?
In my favourite way of looking at this, there is an infinite number
of them and this number is constant; that is; there is always the
same number of universes. Before a choice or decision is made,
in which more than one outcome is possible, all the universes are
identical, but when the choice is made, they partition themselves
into two groups, and in one group one outcome happens and in
the other group another outcome happens. Normally these two
groups don't affect each other thenceforward, but as I have said,
occasionally they do.

It is sometimes said that the many-universes interpretation is also a

branching-universes interpretation. That is, that when the world is
faced with a quantum alternative it splits into all of the different
alternativeswhich arepresentedto it. Your view is slightly different?
That's right. When Hugh Everett first proposed this interpreta-
tion in 1957, this was the sort of language that he used, he spoke
about branching universes. The reason was that, if there was a
collection of identical universes, he preferred to speak of it as
being one universe. If they were all identical and remained
identical then he thought it would be pointless to speak of them
as being· 'many' - they would merely be a different way of
describing just one u_niverse. So when I say the universes parti-
tion themselves into two groups, Everett said that one universe
splits into two universes. My way of speaking about this is to say
86 The ghost in the atom

that there's always the same number of universes, and that they
repartition themselves.
Is it true that as time goes on, these universes differentiate more and
more; and that we can envisage them as somehow existing in parallel
and not changing in number, but changing in content?
That's right, they change in complexity, and this increase in
complexity is the reflection in quantum theory of the second law
of thermodynamics, which is that entropy always increases, or
that there is a 'forward arrow' of time.
I wouldn't quarrel with that, but one of the things that puzzles me is
that the underlying structure of quantum mechanics is symmetrical
under time reversal, and I can't see why it is that the changes which
we're talking about should occur in a preferred direction in time.
Might we not find an equal number of other universes where the
complexity is decreasingin time?
It is perfectly true that the Everett formulation of quantum theory
allows universes to fuse, to use his old language, or to become
identical again in the language I prefer; and a priori in the theory
there is no reason why they shouldn't predominantly do that,
instead of predominantly differentiating towards the future. Or
indeed why they shouldn't do both in a haphazard way. Why
there should be preferred forward time direction in the differen-
tiation of universes is the same problem as occurs in all branches
of physics in explaining why there is an arro\v of time.

This problem is not solved in your theory?

No. I believe that there are promising avenues of research in
quantum theory which may lead to a solution, but it's not solved
yet. But remember, this is the same problem which exists in
every branch of physics and it's not directly solved by the Everett
interpretation, nor by anything else as yet. I should add that the
coming together of universes on a small scale does indeed occur,
and has in effect been observed, because every time there is an
interference experiment, this provides indirect evidence of the
fusion of two groups of universes into one.
 David Deutsch H7

That sounds like an amazing statement. Could you give us a precise

example of where you considerthat two universes have been observed
to be fusing together.
Yes, in the classic Young's two-slit experiment in optics. What
one does is to pass a very weak beam of light, one photon at a time
(this can also be done with other particles nowadays) through a
pair of slits in such a way that some of the properties of the
photon are destroyed by its passing through either of the slits
separately. That is if the photon passes through one slit then
some of the information stored in it is destroyed, and also if it
passes through the other slit this information is destroyed.
According to quantum theory some aspect of this particle - the
wave function - passes through both slits simultaneously and
the information is not lost. Now this reminds one of the old
argument: is light corpuscular in nature or is it wave-like in
nature? The experiment that I'm now speaking of demonstrates
the 'wave-like property of ph_otons'. However, if one then puts a
detector close to either one of the slits, one always finds that the
photon is detected wholly, 100%, coming out of either one slit or
the other. But the very, presence of this detector prevents one
from successfully operating the apparatus which would· detect
the wave-like nature of the motion. Now, the way that Everett
interprets this is to say that the result of the observation of
wave-like motion tells us that at a previous instant there were
two groups of universes - and that in one group the photon
passed through one slit, and in the other it passed through the
other slit, but that 1,ter both these photons appeared at the same
position, and from then on all the universes were the same again.
Let's just get this right. We present a particle with a choice of either
going through one slit or the other, and in the Everett interpretation
these represent two quite separate worlds. But if we allow the system
to bring these two pathways back to overlap each other, then this is
like bringing the two worlds back into fusion.
Correct. And when one observes the fused photon afterwards, it
has properties which rule out the possibility of it having specifi-
cally been through either the one slit or the other.
88 The ghost in the atom

So these worlds that we're talking about, then, although not part of
our space and time, nevertheless appear to be able to communicateat
the level of atoms. Is it possibleto envisage ever exploringthese other
universes?Can we ever gain any information about them, even at the
atomic level? Can we look at the properties of atoms and find out
anything at all about these other universes?
To a limited extent, we do. The only experiments in which we can
detect the presence of the other universes are indirect experi-
ments - rather like detecting the temperature of the inside of the
sun which is 16 million degrees - by looking at the exterior at
5000 degrees. In other words, the way we detect them is via our
theory.
As for exploring these other worlds, our present theory states
that this is impossible. We cannot travel to them any more than
we can directly travel into the past or into the future.

But nevertheless these other universes have inhabitants that ·1ook

very much like you and I?
Just as the past and future do. In fact Don Page and William
Wooters recently explored this connection between the' different
universes' of the past and future and the different universes
existing now side by side with us and described these on ~
uniform mathematical footing which shows that the past and
future are just special cases .of Everett's other universes.

But travel into the past would involve certain types of paradoxes-
causal paradoxes- that maybe are circumventedin the case of these
parallel universes. One can envisage going to one of the other
universes and meeting, as it were, another copy of oneself. But it
wouldn't strictly be oneself, because it would be slightly different.
And you could alter the future events of such a universe without
contradictingyour own future in your own universe when you went
back. Doesn't it in fact enableyou to escapefrom someof thesefamous
time travel paradoxes that sciencefiction writers love?
It would, if quantum theory were slightly different. The reasqn
why quantum theory doesn't allow this in its present form is that,
 David Deutsch 89

just as the past causes what happens in the present, in some

sense, and the present causes what happens in the future, so the
different parallel universes are linked by being part of a common
physical object. Physical reality is the set of all the universes
evolving together, like a machine in which some cogwheels are
connected to other cogwheels; you cannot move one without
moving the others. So the parallel universes are connected as
inextricably as the universes of the past and the future.
If you went to one of these other l!,niversesand stamped on a beetle,it
would have a knock-on effect in your own universe?
That's right.
So it may be more complicatedthan we thought in these time travel
paradoxes?
Yes. One could of course speculate that with a modification to
quantum theory one could travel to the·past, or to other universes
in the present, but since quantum theory is the only reason that
we have for believing in these universes at all, it seems rather
wild to change quantum theory just to make the universes
behave in a slightly different and even stranger way than they do
~~dy. .

You've explained part of your attraction to the many universes

interpretation,but what in your opinion is wrong with the standard
Copenhageninterpretationof quantum mechanics?
Well, I've mentioned that the Everett interpretation is more
natural in the formal sense. But the best physical reason for
adopting the Everett interpretation lies in quantum cosmology.
There one tries to apply quantum theory to the universe as a
whole, considering the universe as a dynamical object starting
with a big bang, evolving to form galaxies and so on. Then when
one tries, for example by looking in a textbook, to ask what the
symbols in the quantum theory mean, how does one use the
wave function of the universe and the other mathematical objects
that quantum theory employs to describe reality? One reads
there,_'The meaning of these mathematical objects is as fol!~~s:
90 The ghost in the atom

first consider an observer outside the quantum system under

consideration ... ' And immediately one has to stop short. Post-
ulating an outside observer is all very well when we're talking
about a laboratory: we can imagine an observer sitting outside
the experimental apparatus looking at it, but when the experi-
mental apparatus - the object being described by quantum
theory - is the entire universe, it's logically inconsistent to
imagine an observer sitting outside it. Therefore the standard
interpretation fails. It fails completely to describe quantum cos-
mology. Even if we knew how to write down the theory of
quantum cosmology, which is quite hard incidentally, we liter-
ally wouldn't know what the symbols meant under any interpre-
tation other than the Everett interpretation.
Let me add that, in my experience of physicists changing
their views about the interpretation of quantum theory, it is
very often when they begin to consider quantum cosmology that
they are finally convinced that there is no alternative to the many-
universes interpretation.

If we're dealing with quantum cosmology, we're in trouble with the

conventional interpretation. But with the many universes we have an
interpretation which appears to face up to the problem of quantum
cosmology squarely and come to grips with it. It gives us, at least in
principle, a consistent way of being able to talk about the quantum
behaviour of the entire universe. It therefore opens up the prospect
that w~ could look to quantum mechanics as an explanation for the
very existence of the universe. That is, to talk about the coming into
being of the entire universe as some sort of quantum phenomenon. Do
you think this could be correct?

Yes, although I must stress that unlike most of the other things
I've been saying, this is speculative. (The other things I've been
saying in my opinion are not.) I think that, just as there's a strong
possibility of understanding the second law of thermodynamics
by using this branching structure of the Everett interpretation,
there is also a possibility of understanding something about the
problem of the existence of the universe as a whole.
 David Deutsch 91

Now, in the many-universes interpretation, one seems to hang on to

some vestige of objective reality, although it's a multiplied reality.
Yes, that's one of its main advantages.
Nevertheless, it's not necessary to introduce any subjective features,
like consciousness and the mind and so on. Does this theory have
anything at all to say about what an observer actually is?
No. Another advantage of the Everett interpretation is that it is
not necessary to present within the theory a working model of an
observer. That is, it's not necessary to state in fine detail what is
the difference between an observer and any other physical
system. One thing that the interpretation does shed light on, by
the way, is what we mean by a measurement. There are topics in
the theory of measurement which are far more easily dealt with
in the Everett interpretation. But these are straightforward
matters compared to the question of what· is consciousness. I
regard it as one of the advantagesof the Everett interpretation that
it has nothing to say about this. The interpretation will work
even before we have an exact knowledge of what consciousness
is all about. The other interpretations will not work properly
until we understand consciousness.
But of course, for many people, one of the endearing aspects of
quantum mechanicsis precisely that it puts the observerback into the
centre of the stage. It involves mind in a non-trivial way in the
operation of the universe. And they like that, becauseit has a certain
mystical appeal. You're banishing mind from the universe, or at least
you're not making it indispensableto the operation of the universe.
Yes, this is an interesting controversy. I would actually put it
entirely the other way round. I believe that it is the conventional
interpretations which banish minds from the realm of physical
reality.
Why do you say that?
Because in them the mind is supposed to obey physical laws
different from the rest of reality, and, secondly, in all the versions
92 The ghost in the atom

of the conventional interpretation that I know of, the exact nature

of this new property of the mind - this new mystical property- is
not specified. It is more a hope than a real theory, that maybe one
day we might find new laws to describe the mind which are just
such that the conventional interpretation of quantum theory will
work! In the Everett interpretation, the existing laws of physics
are supposed to describe the mind properly, and until we find a
contradiction, it's perfectly reasonable for us to carry on believing
that. It's only in the Everett interpretation that the observer is
considered to be an intrinsic part of the universe that he is
measuring.
But he seems to be there just for the ride, he doesn't play a role in
determining reality.
He doesn't play a special role in determining reality, no more
than does any other physical system.
So it doesn't help us to understand what consciousnessis? We can
simply say that brains are more ~omplex systems than individual
atoms, and that for some unknown reason they endow the universe
with consciousness.
That's right. But I don't see how it is an advantage for the
interpretations competing with Everett's that they require this
knowledge if they don't present it.
I think it's probably only an advantage to the mystics! So, let me put
to you, then, the following: one could of course simply say that in
dealing with the world, at least the world of physics, all we have are
our observations;we can do experiments, we can make measurements
and we try to relate the'!' with a model;quantum mechanicsprovides
us with an excellent modelfor relating the results of observations:we
can just regardit as an algorithm, a method of connecting together all
the things that we observe, and it works very well. So why do we need
these elaborateideas of many universes? Can't we just take quantum
theory at face value?
The disadvantage of interpreting a theory purely as an instru-
ment for predicting the results of an experiment, rather than
 David Deutsch

regarding it as a true description of an objective reality, l ■ that

such a view would paralyse future progress. If I can give- 1n
analogy from an earlier era of physics: when Galileo was pressed
by the inquisition to renounce his theory that the Earth moves
around the sun and that this causes the apparent motion of the
lights in the sky, he wasn't asked to go the whole way to saying
that his theory was false, he was only asked to go half way. He was
asked to say that although his theory was a good algorithm for
predicting the positions of bright spots in the sky, he shouldn't
go further and say that these spots were caused by things which
actually had an objective existence as radiating material bodies
in space.
Well, I wonder whether there is any real difference between these two
approaches.It seems to me that in modern physics there really is no
difference at all. For example, people regularly talk about virtual
photons - are they really there, or aren't they really there? I don't
think that question has any meaning. It seems to me all we have is a
means of computing the results of different observations, and to talk
about whether virtual particles are really there or not is a fruitless
enterprise.
Yes. This is a slightly different sense of the term 'really there'.
Whether we describe a virtual photon as a particle or a wave or as
something that exists in ordinary spacetime or not is merely a
difficulty in translating our physical knowledge into ordinary
everyday language. But I think we do have to say that something
is really there. If I can go back to the example of Galileo, had other
physicists at that time truly been willing to accept the idea that
his theory was merely an algorithm for predicting the locations of
spots of light in the sky, further progress, towards Newt<?n's
theory, would have been paralysed, because although Newton's
theory is one substantial intuitive step beyond Galileo's actual
theory, it bears no comparison at all with the old theory of the
celestial spheres. And if Newton had been content to·stay with
the old ontology of a celestial sphere, he would never have been
able to formulate his own theory, even as an 'instrument' or
'algorithm'.
94 The ghost in the atom

We have a dual reason for regarding the quantum theory as

describing reality. One is because that's what we wanted the
theory for in the first place, and the other is that not to do so
inevitably stultifies progress.
I'm not completely convinced, because after all one could claim that
the electromagnetic_fieldis just an invention, just a word - it's not
really there- and yet it hasn't impededprogressin electromagnetism.
Again, I think you're using the term 'really there' in two different
senses. When we speak about electromagnetic fields, for example
radio transmissions, the language that we customarily use to
describe this is the language of these waves really being there;
we say that they emanate from the transmitter and are received
in the receiver. In fact it's quite hard to reformulate classical
electromagnetic theory so that this is not so, although it is
possible - one can speak entirely of the motion of the electrons in
the receiver and in the transmitter without ever speaking of what
transmits the influence between them. But this is a mistaken way
of speaking, because had we, at Maxwell's time, forced ourselves
into this way of looking at the world, subsequent developments
in field theory where, for example, an energy density was
ascribed to the field itself, and later the quantum theory of fields,
would not have been possible.
But the field is still an abstract construct, isn't it?
It is certainly an abstract construct, but it gains its place in
physics when a physical theory says that it corresponds to
something real. What words we then attach to this real thing that
it corresponds to, is a subsidiary matter.
But surely any model of reality in which we can have confidence,
ultimately comes ~ack to our observations?It's only at the receiving
end - the observer - that one makes contact with reality, whatever
sort of elaborate abstract machinery one may invent to talk about
disturbancespropagating and influences connecting together.I mean,
ultimately, we come back to our observations-and that's allwe've got
isn't it? Why should we want any more?
 David Deutsch 95

I don't believe so. I don't believe that we would even have the
observations, if they were really 'ultimately' all we had. The way
we really observe things is via an intimate relationship between
theory and experiment. We need both. After all, our very sense
organs are the physical embodiments of certain theories; our
eyes are the embodiments of certain optical theories, certain
theories about colour, and. three-dimensional space. One way we
can tell that these are only theories is that some of them are false
- some of the theories embodied in the eye are actually false
theories. And when we see things, we do not rely solely on our
sensory perceptions, otherwise we would never have found out
that there are two types of green light, one directly green, and the
other a mixture of blue and yellow.
Yes, but we only find out about them by extending the range of our
capabilitiesthrough technology.
Exactly. We find out about them by extending our knowledge of
the world through a combination of theory and observation:
never just by observation, and never just by theory.
Well, interesting though the many-universes theory may be, is it
perhaps simply a way of speaking about the world, or can it actually
be tested? You've said that we can't visit these other universes, but
can we devise any sort of experiment to show that they really do exist?
When Everett first put forward his interpretation, he believed
that it was a pure interpretation in the technical sense of the
word. In other words, that the physical predictions of quantum
theory under his system were precisely identical with those
under any other system. Now, I believe that this is not so, and I
have recently done some work trying to elaborate the exact
experimental difference between the Everett and the conven-
;tional 'interpretations'. I now have to say 'interpretations' in
quotes because I believe that there are actually different formal
structures for quantum theory.
So we're talking, not about two different ways of looking at the same
theory, but two completely different theories?
96 The ghost in the atom

Yes. Once I realized that at the mathematical level the two

formalisms are in fact slightly different, and that there is therefore
a hope in principle of constructing a crucial experimental test, my
greatest difficulty in trying to think of a test was that the
conventional interpretations are so woolly and imprecise that it
is hard to pin down exactly what their predictions are! However,
eventually I got to what I believe is the common core of all the
conventional interpretations, which is this: they all say that at
least by the time the result .pf a measurement has entered the
consciousness of an observer, the wave function will have col-
lapsed (or whatever this irreversible loss of information is called
in the various versions of the conventional interpretation). Also,
we know from experiment that so long as the information is still
in a sub-atomic system, one which is still capable of exhibiting
atomic interference, this collapse has not yet happened. So the
collapse must be supposed to happen at some point in between
the atomic level and the moment when an observer becomes
conscious of it. Where, we don't know. The reason why we don't
know is that the conventional interpretations are very vague on
this point. Now, in the Everett picture this collapse of the wave
function would be described as the sudden disappearance of all
the universes except one.
But that of course doesn't happen?
Well, we believe it doesn't happen. But what we want is an
experiment which will detect whether it happens or not. Here's
how it works: we first find a situation in which the conventional
interpretation predicts that all the other universes suddenly
disappear, and where the Everett interpretation predicts that
they don't disappear but they're all there in parallel. Then we
find some observable consequence of their subsequently
interacting with each other in an interference experiment. And
we then observe one result if the Everett interpretation is true,
and another result if any of the conventional interpretations is
true. Simple as that.
Unfortunately, this experiment requires the observation of
interference effects between two different states of an observer's
 David Deutsch 97

memory. The reason why it has to be an observer's memory

rather than just any old physical system, is not Everett's fault. It
is that the conventional interpretation makes a special reference
to observers as being different. The way in which these interpre-
tations differ from the Everett interpretation is that they say that
observers obey different physical laws, and Everett says they
obey the same physical laws. So the place where we would expect
a crucial experimental test is with quantum effects inside an
observer's brain.
We're talking about quantum memory?
We're talking about quantum memory, and presumably electro-
nic artificial intelligence.
Becauseour own brains really work at a.classicallevel rather than a
quantum level?
That's right. As far as we know. There are theories that they
don't, but whether they do or not, it seems unlikely that we shall
get control over the electronic functions of the human brain at
such a fine level. Whereas, when it comes to electronic compo-
nents, it's already commonplace to use some of their quantum
properties; every microchip works on those principles. But even
microchips are at present too crude for interference phenomena
to be observed in them. ·
But we can envisage buildingsome sort of artificialsuperbrainwith a
memory at the quantum level, and ask it to carry out this experiment
for us, and tell us what it feels?
That's right. And it could record the results of this experiment in
any way we like. It could perhaps write them down, or tell us the
results; the difference - rather like in Aspect's experiments -
between quantum theory and the rivals is not a matter of a small
percentage, it is an all or nothing thing. In the experiment I
describe, one would observe a certain atomic spin, and if it was
pointing one way, Everett's interpretation wouk~. be true, and if
it was pointing the other way, the conventional interpretation
would be true.
98 The ghost in the atom

Now, you've explained how one might construct this superbrain to

play the role of an observer who has a quantum memory, but can you
just tell us exactly what it is he's going to observe? Exactly what
experiment does he perform, if we can call him he?
Yes. The experiment hinges on observing an interference
phenomenon inside the mind of this artificial observer. This can
either be done by someone else looking inside him, or more
elegantly, by his trying to remember various things so that he can
conduct an experiment on his own brain while it's working.

He can observe himself?

He can observe part of himself, yes. And what he tries to observe
is an intereference phenomenon between different states of his
own brain. In other words, he tries to observe the effect of
different internal states of his brain in different universes
interacting with each other.

How would these different internal states be set up?

They are set up in the first instance by a special sense organ
which is essentially just another quantum memory unit. This
sense organ is used to observe the state of an atomic system - a
system with two possible states, such as an atomic spin for
example. Now, quantum theory predicts that, having observed
this atomic system, the observer's mind will differentiate itself
into two universe branches.
So, we have an atomic system with two possiblestates, each of which
would trigger the brain of this artificial observer to be in either one
state or the other. And accordingto the Everett interpretation you're
saying that these two brain states somehow co-exist - or at least they
exist in parallel universes. But we don't let the universes get out of
touch with each other. We bring them back to overlap, to interfere
with each other, and this poor observer is, as it were, schizophrenic
and observing both of these possibilitiesat once.
That's right. In effect he is feeling himself split into two copies.
And he feels himself merge again?
 David Deutsch 99

Yes, in effect. Of course, we don't have sense organs of this type,

so it's hard to say what this would feel like, but when this
observer exists we can ask him!
It sounds most uncomfortable!
Perhaps it will be, but then presumably he'll be a physicist so
he'll enjoy doing this experiment!
How exactly will he go about it?
At the intermediate stage he will write down an affidavit to the
effect that 'I am hereby observing one and only on_eof the two
possibilities'.
And what he writes down will be different in the two different
universes?
No. What he writes down will be the same in the two different
universes because he won't actually say which of the two pos-
sibilities he observes. Instead, he will write: 'So that this experi-
ment can be continued, I will not actually say which of the two I
am observing but I do certify that I'm observing only one of the
possibilities.' He can then continue with an interference experi-
ment between the two parallel universes containing the different
brain states and he should get a result which is compatible only
with the presence of both of these brain states in his past. So, if
interference occurs he can infer that these two possibilities must
have existed in parallel in the past - supporting the Everett
interpretation. However, if the conventional interpretation is
true, then at some time during his deliberations all the universes
but one will have disappeared. And although it'll still be true that
he will write down 'I am observing only one', by the time he gets
on to the interference phenomenon it won't work (i.e. the inter-
ference won't occur). And so he will have demonstrated that the
Everett interpretation is false.
Because by his having this certain knowledge of a particular outcome
he will have completely modified the wave properties of the system,
and therefore altered the subsequent quantum development of the
system, which could be checked by subsequent measurements?
100 The ghost in the atom

Yes, he either will or won't, and if he has altered them in that way
then the conventional interpretation is true, and if he hasn't then
the Everett interpretation is true.

Which means that, in the Everett interpretation, it's possiblefor the

observer to make up his mind, but he's got two minds.
Yes.
He's in two minds_aboutit! When the experiment is complete, and this
machine observer is asked to rememberwhat it was he observedeven
though he didn't write it down at the time, what will he remember?
Will he rememberboth?
No, he will remember neither in fact. It is a necessary conse-
quence of the other things he does, that he must wipe out the
memory of which o~e of those two possibilities he observed.

But he still has the memory that he only observedone of the two?
Yes, this is the key feature of my experiment: the memory that he
knew one and only one of these possibilities can be retained even
though he is obliged to forget which one.
You're saying he can deduce that he must have been split, becausehe
knows that the outcome involves both possibilitiesco-existing?
Exactly.

If it's true that all these other universes which exist around us can
couple to our universe at the atomic level, why don't we feel their
presence?
In principle we could. There's no fundamental reason why we
don't. It's only that our brains are sufficiently large to operate on
essentially a classical level. If we had fine enough senses, then,
rather like the mechanical observer in my thought experiment,
we could detect or feel (whatever that would mean) the presence
of the other universes.
You mean, if we couldfeel all the atoms creeping about in our brains,
then we would indeedfeel these other universes?
 David Deutsch 101

Yes. In fact, as I said, Everett often compared critics of his

interpretation to the opponents of Galileo, who said that they did
not feel the Earth move beneath them. The point being that
Galileo's very theory predicted that one doesn't feel the Earth
move, unless one uses sufficiently fine apparatus. And just as
with a Foucault pendulum or with sufficiently delicate astronom-
ical measurements one can detect - 011:ec_an,in effect, 'feel' - the
motion of the Earth, so with sufficiently fine senses we would
indeed feel the presence of the other universes.
Anyway, to do the test you've just described we would need this
supercomputer to tell us that the Everett interpretation is or is not
correct.
Unfortunately yes. And it seems rather a long way beyond
present technology to construct such a computer. Although
when I say a long way I don't mean millions of years away, I mean
decades away.
Well, it's certainly fascinating that there might be some prospect of
actually testing these ideas within the foreseeablefuture. But why is
it that Everett overlooked this possibility?
Now, I've never though~ about that! Perhaps one of the reasons
was that he had an additional idea connected with quantum
theory, which was that its interpretation ought to follow directly
from the formalism. That is, if you write down the mathematical
rules of quantum theory, he thought there ought to be only one
way of interpreting these. This is an extremely strong assumption
to make, and it would, had it been true, have been the first
physical theory in history ever to have this strong property. He
hoped it was true, and therefore I think he concentrated on the
similarities between the predictions of his theory and those of
the rivals, thereby highlighting the fact that the rival conven-
tional interpretation requires additional metaphysical parapher-
nalia whereas his doesn't. So he said, 'I take the pure formalism,
and I add nothing, and I obtain my interpretation. In contrast
they (the supporters of the conventional interpretation) have to
add all this stuff about consciousness and so on.' Now, I think
102 The ghost in the atom

Everett was slightly wrong. I think that even in his interpretation,

one requires a little bit of extra structure in order to arrive at the
interpretation. But not much - very much less than in the
conventional interpretation.

Can you summarize in a few words what this little bit of extra
structure is?
Yes. It is the little piece of mathematics which provides the
connection between the wave function or state vector, which is
the mathematical object describing the universe, and the concept
of the many parallel universes. I don't think one can do without
this extra structure. But I do agree with Everett so far as to say that
his is the simplest possible addition to the purely instrumental
quantum theory.

I'm not sure if I've understood this correctly. Are you saying that
Everett's extra assumption is necessary to tell us something about
how any individual universe in this vast stack of cosmic alternatives
fits into the stack?
That's right, yes.
You have already explained the advantages that the many-universes
theory has over,_say, the conventional Copenhagen interpretation.
What advantages do you think it has over the other rival interpreta-
tions?
There again, they are rather diverse when it comes down to
details. I suppose you're referring mostly to the hidden variables
interpretations?
Yes, or it's modern variant: the so-calledquantum potential.
Yes. One objection is that to append to the quantum formalism
an additional structure which is supposed to correspond to
physical reality (this additional structure being far more compli-
cated than the original physical theory), solely for the purpose of
interpretation, is I think a very dangerous thing to do in physics.
These structures are being introduced solely to solve the interpre-
tational problems, without any physical motivation. From the
 David Deutsch 103

point of view of a physicist, I'd say the chances of a theory which

was formulated for such a reason being right are extremely
remote.
But aren't you introducingmany universesfor precisely that reason-
to solve.the interpretationalproblem·?
Well the problem of having an interpretation in the first place is
itself an unavoidable problem, and I would gladly dispense with
the many u·niverses were there a simpler interpretational
assumption. But the many-universes assumption is in fact so
simple from th~ point of view of the underlying physical laws,
that, as I said earlier, Everett, DeWitt and others were misled into
thinking that there was no additi<?nalstructure at all. It really is
the most natural interpretation of the formalism yet thought of.
By contrast, the hidden variables theories are very complicated.
One of the reasons why they are is that from Bell's theorem and
Aspect's experiments we know that the simplest forms of hidden
variable theories simply cannot mock up the effect of quantum
theory.
Instead, we need to have some sort of non-local hidden variable
theory, which is what Bohm and Hiley are attempting to do.
A non-local hidden variable theory means, in ordinary language,
a theory in which influences propagate across space and time
without passing through the space in between.
Without passing through? Or should we simply say they propaga~e
instantaneously? Is that perhaps the same thing?
Yes. To say they propagate instantaneously in the context of
relativity means that they cannot be passing through the inter-
vening space-time because if they did their description would
then be inconsistent with relativity.
They don't deny this. They say of course their description is inconsis-
tent with relativity, but when it comes to actually making measure-
ments the results of all these measurements are consistent with
relativity. It's only the mechanism itself which seems to contradictthe
spirit of relativity.
104 The ghost in the atom

Yes. This is a defence only if you're willing to back completely

into the corner of saying that quantum theory is merely an
instrument. And if it is merely an instrument, th~n the hidden
variable theories lose their main ·advantage, which is that they
cling to the notion of objective reality, just as ~verett does.
But look, one feature which the many-universes interpretation and
these non-local,or let's say faster-than-light, interpretationshave in
common is that they are both attempting to retain some vestige of
objectivereality. In both cases,accordzngto Bell's inequality, and the
Aspect experiments, we have to make a choice. You either have
faster-than-light signalling, or you throw away objective reality.
Now, in my opinion, it doesn't seem terribly awful to have to throw
away objective reality. Why should we insisf so much that the
external universe is independent of our observations?Surely it's not
surprisingthat we ourselvesplay a part in reality, becausewe seem to
be very important to ourselves?At least it doesn't surpriseme, on the
basisof my personalexperience,that we areplaying a part in reality.
So why this desperate urge to cling on to some vestige of objective
reality, if it means introducing complicatedthings like faster-than-
light signalling,or other universes?
Well, I agree that Aspect's experiment forces us to change our
view of reality. The reason why I would want to cling on to the
philosophical notion of objective reality.in itself, regardless of
whether this looks unfamiliar or not is the same reason that I
mentioned before for not changing to an instrumentalist view .of
physical theories. Firstly, because if we can have a theory that has
objective reality in it, it is philosophically superior. Therefore we
should at least try that approach before we thr~w away the notion
of reality. And secondly, from the point of view of science,
especially physics, I believe that changing to an instrumentalist
interpretation of a theory makes it impossible to obtain the next
theory, because the next theory will be a step forward from the
ontology of our present theories. The likelihood is that it'll be
even wilder, it'll tell us that the universe is even stranger than
Everett said it was, not less strange. And if we abandon the
notion of reality, we are depriving ourselves of the mechanism
 David Deutsch 105

by which we construct conceptual models of the universe. It's

only by altering our present conceptual models that we will
discover the new theory.
I'm not saying we should abandon reality, but abandon a type of
reality which is independent of ourselves. It just means that future·
modelswill have to incorporatethe observer at a fundamental level.
Yes. I wouldn't object ~othat in principle. But I don't believe that
quantum theory drives us to this. Perhaps I can stress again that
the conventional interpretations of quantum_ theory which do try
to give the observer a special place in forming reality haven't
actually done so yet. They merely claim that they will one day.
Yes, of course.And they can't cope with quantum cosmo!ogywithout
having an observer o_utsidethe universe.
Yes. One day, perhaps, somebody will be able to write down
exactly what physical laws the mind does obey, if it doesn't obey
quantum theory. And perhaps that new physical theory (it
would not be quantum theory any more, it would be a new
physical theory) might be testable against quantum theory.
Well, perhaps, but nobody's writt·enit down yet!
No, and when speaking about the supposed advantages of the
conventional interpretation - that it gives the observer a funda-
mental view which is perhaps philosophically attractive to you (I
don't know whether it is or not) - you are _forgetting that in fact
they don't do this yet.. This is merely a claim, a promise, which
over 50 years has not been fulfilled. Whereas the Everett interpre-
tation is unproblematic. It works perfectly well without making
these promises.
 7
John Taylor

John G. Taylor is Professor of Mathematics at King's College,

Universtity of London, and the author of a number of books,
both specialist and popular. His main research interest is
quantum gravity, but he is also interested in the physics of
the brain. In this interview he adopts a hard-headed
approach to the more outlandish ideas of quantum
mechanics, and opts strongly for the statistical interpreta-
tion.

What is the ensemble(or statistical) interpretation?

Well it is a concept that lives up to its name: when we're making
a measurement of any observable in a system what we're actually
doing, according to the ensemble interpretation, is that we're
making a measurement on an aggregate or ensembleof identically
prepared systems. We thereby obtain a whole set of measure-
ments, one for each of the identical versions of our particular
experiment in the ensemble. Hence our results take the form of a
probability distribution of particular values for that measure-
ment.
So you just look at the statistics, and you don't care about any
individual event?
That's right. And indeed it's amazing that Einstein, if I can quote
him now, really did in the end settle for this ensemble interpreta-
tion. In his writings, in reply to criticisms, he said: 'One arrives
at very implausible theoretical conceptions if one attempts to
maintain the thesis that the statistical quantum theory. is in
prin~iple capable of producing a complete description of an
individual system. On the other hand these difficulties of
 John Taylor ·107

theoretical interpretation disappear if one views the quantum

mechanical description as the description of ensembles of sys-
tems.' So Einstein was, in fact, one of the forerunners of what I
think is regarded by most physicists as a natural interpretation of
quantum mechanical measurements. This is that we are actually
doing large numbers of measurements on identical systems and
we take the frequencies of particular measurement values as the
probability distribution of those values.
So you make no attempt at all to describe what is going on in an
individual system? ·
We're not allowed to. That is quite clear when we look at the
various paradoxes. If we take the Einstein-Podolsky-Rosen
experiment, which really is the basis of the Aspect experiment,
it's clear that a paradox arises there, because we're assuming that
when a measurement is made, say, of the spin of a particular
particle, we can also measure the spin of a far away particle
whose properties are correlated according to the usual quantum
mechanical ideas. For example, we might find that the particle
nearby has a spin pointing up. From that, we can conclude that
the other particle far away must (if it is correlated) therefore have
spin down. This would be paradoxical if you believe that you are
indeed measuring individual systems because it would seem
that you're actually able to influence that far away particle, and in
some ways determine its spin simply by making a measurement
on the nearby particle.
The ensemble interpretation says, however, that we're looking
at a whole ensemble of such systems. Some 50% of them may
have (when we're measuring them) nearby particles with spin
up and far away ones with spin down, while the other 50% have
the opposite spins. But we can't say in any particular case what
that spin of the far away particle is from the measurement
nearby, because we don't know about it; we only know about
ensembles of such situations.

Could I press you further on this question of whether, in the ensemble

interpretation, one is still supposed to think of an individual system as
actually possessing well-defined properties. For example, does an
108 The ghost in the atom

electronat any given moment actually possessa well-definedposition

and a well-defined momentum, even though of course we can't
measurewhat they are?
The answer is no, the electron cannot have both of those attri-
butes. All we can ever measure, through the uncertainty princi-
ple, are the lower bounds on the dispersion of measurements of
that position, and of that velocity or momentum over an ensem-
ble. We can never measure these quantities for a particular
electron. It's ruled out of court, and I think we've got to accept
that from Aspect's experiment.

But if electronsor atoms don't actually possessthese propertiesbefore

we make a measurement, it seems to suggest that the observer mu~t
therefore be involved in a fundamental way, because these particles
certainly do have well-defined properties after we make the approp-
riate measurements,and we of coursecan choosewhich measurements
to make, either position or momentum.
Yes, but we're making that by setting up an ensemble: a set of
identical copies of the particular situation we're measuring.

But we might not. We might choose to look at an electron and, say,

measure its position, and find a position, and then that's quite
satisfactory. But if we argue that it didn't have a well-definedposition
before we made that measurement, then the measurement itself is
playing a crucialrole.
One has to be very careful in distinguishing between measuring
and preparing. That's something which some physicists have
considered quite carefully.

Could you give us a brief introduction to the distinction between

them?
Well, if you're preparing a state of an ensemble, then you know
that it will have properties identical with that preparation in the
future. If you make a measurement, then you will have been able
to gather what it was like just before the measurement in the past.
There is a rigid distinction between these two. And I think you
 John Taylor 109

have to be very careful that you don't fall into the trap of always
equating the measurement process with the preparation process.
Once you've prepared a system, then you can begin to look at
what the ensemble of states you have prepared will look like. For
example, you might choose to measure the positions of a set of
electrons. On the other hand, you may wish to measure their
momenta. But always the dispersion of these measurements is
related by the uncertainty principle. If you're going to prepare
the electron in a given position, then you know that through its
dispersion in the ensemble it won't have any definite value for its
momentum. And that's the nature of the beast.
So you don't believe that if, for example, we prepare a quantum state
with an electronat a particularposition, it actually has a well-defined
momentum, even though we can't ourselves measure it?
\

No. We must accept that it's given by all possible ranges of

momenta. In other words the momentum cannot even be
defined-.
Yes, but that brings me back to the feeling that if its momentum can't
be defined and yet after a measurement by a person it will have a
well-defined momentum, then it seems that the act of measurement
itself is absolutely crucialin promoting the system from a sort of fuzzy
indeterminatestate to one of concrete reality.
Ah, but you've then reprepared the system if you want to look at
it in a state of given momentum.
But if you put it in a state of given position, and then decideto measure
the momentum, of course you get a particular value, although the
value can't be predicted.
Ah, but again you're making it sound as if you're looking at an
individual electron.
But in practice we can do that; we can decideto make a measurement
on an individual electron.
Yes, but then you will know that if you try to measure its
momentum, there will be an infinite range of possibilities. There
110 The ghost in the atom

will, of course, be a particular value for a particular case in the

ensemble.
Which seems to make it look like the observer intruding.
Of course. But you know very well that the preparation you make
when you put the electron in a particular place gives you an
ensemble in which the momentum is completely undefined. If
you wish now to look at a particular case of measuring that
momentum, you will get a particular value, but that value has no
meaning at all in quantum mechanics. In effect, you're preparing
another ensemble (if you make a number of such measurements),
and if you wish to start again and say I look at all those electrons
with a given momentum, now they are in no definite place.
So in this scheme the electron's wave function doesn't collapseon to
one of particular momentum when you measure its momentum.
No. You're now setting up a new ensemble. You can't take a
particular electron, at a given place, and say I'm now measuring
its momentum, because that doesn't mean anything. That's not
allowed.
If you abandon any attempt to describe what is going on in an
individual system, isn't that a bit of a cop-out?
Well, I think you have to ask whether indeed you're in more
trouble if you cop-in, than if you cop-out! And as far as this
paradox is concerned - the Einstein-Podolski-Rosen paradox -
clearly you're in greaf trouble. Likewise if you take the
Schrodinger cat paradox. This too depends on a thought experi-
ment.
According to any interpretation of quantum mechanics which
attempts to describe individual systems, the wave function of the
system including the cat must show that after about one lifetime
of radioactive decay, there is an equal probability of the cat being
alive or dead. That means the quantum mechanical state is
composed of the cat being alive for half the time, and the cat
being dead for the other half. In other words the cat doesn't know
whether it's dead or alive, which is absolutely absurd! Now if
 John Taylor 11I

you take the ensemble interpretation, then in 50% of thr l ANttN

1

the cat is alive and 50% it's dead. That's quite reasonable.
So if we take one individual case, and ask whether the cat is aliveor
dead, then the answer - or your answer - would be that there is no
answer?
Well, the answer would be that in fact, according to quantum
mechanics, there is no way of saying whether it is alive or dead
in any particular case. It's a meaningless question. We can only
say it has a 50-50 chance of being dead (or alive). I think we have
to accept that feature, especially now if we go to the Aspect
experiment. Because there we see that quantum mechanics
agrees with the results of that experiment and any other interpre-
tation is not satisfactory. Except possibly for a non-local type of
explanation (such as the interpretation of Bohm and Hiley). But
then you have to be very careful because there are new features
that enter.
If you're looking for alternative versions of quantum theory
which agree with the results of the Aspect experiment, these
alternatives must also stand up to the level of success we have
obtained when we have gone beyond quantum mechanics to
what is called quantum field theory. Quantum field theory is a
whole new bag of tricks; it involves levels of success in explaining
what we see in nature to one part in at least a million if not more.
There are whole regimes of successes which are almost impossi-
ble to conceive of as being explained any other way.
If you think, for example, of the problems of quantum electro-
dynamics: one of the great successes in the late 1940s and the
early 1950s was of understanding why it is that there are very
delicate shifts of energy levels in the hydrogen atom which are
not explained in conventional quantum mechanical terms. These
energy level shifts could only be explained in terms of 'virtual'
processes involving virtual photons, virtual electrons and posit-
rons - virtual meaning not actually existing in our real world
because we can't observe these virtual particles directly.
Nevertheless, quantum field theory predicts very precisely the
effect of these virtual processes and the results agree with the
112 The ghost in the atom

observed energy shifts to at least one part in a million. How

you're going to replicate that with alternatives to quantum
mechanics, my mind boggles~
To continue this train of thought, let's tum to the recent
discovery of the W and Z particles - the intermediate vector
bosons. These particles were predicted by a theory which unifies
electromagnetism and radioactivity - a theory that was a direct
product of our quantum field theories. Only by considering the
implications of these quantum field theories would we have
been led to the existence of these particles and to a prediction of
their masses, all of which has been confirmed by high energy par-
ticle experiments at CERN. To say that any alternative to quantum
mechanics could have done this is, I think, pie in the sky.
Then there are problems which I think are even more funda-
mental; not questions of one part in a million but questions of
principle. For example, classical mechanics cannot describe the
annihilation or the creation of particles. And yet we observe this
all the time in partjcle accelerators. How on Earth is anybody
going to describe this in classical terms? No amount of non-local
quantum potentials or what have you will explain how matter
can be created or destroyed.
So you're saying this -impressiverefinement of quantum mechanics
called quantum field theory, which gives a very satisfactory descrip-
tion of large areas of modernparticle physics, would simply collapse
if we didn't retain the traditionalnotion of quantum mechanics?
Yes, I would say that any attempt to replace the uncertain
quantum mechanical observables by certain but uncontrollable
or hidden ones is doomed to failure. I've known physicists who,
during their careers, have attempted to replace these amazing
successes of quantum field theory by the classical approach.
Several people spring to mind. They have all failed. And their
failure has become worse and worse as the successes of quantum
field theory have grown. An_dyet at the same time we've seen
Nobel prizes awarded to some of our colleagues for their succes-
ses with quantum field theory and.in particular for unifying the
forces of nature. It's now very hard for me to see any other way.
This avenue of research is almost unique.
 John Taylor 113

Following on from these remarks one can conclude that

Aspect's experiment need not have been performed because
confirmation of quantum mechanics was virtually guaranteed by
the enormous successes we've had with the theory so ~ar. If
you think of the understanding of locality we have gained in
terms of the so-called dispersion relations based on quantum
field theory, again the Aspect experiment needn't have been
carried out. High energy scattering experiments have verified
that locality is preserved down to billionths of a centimetre,
right inside the photon. It's just impossible to conceive of any
violation.
Aspect's results came as no surprise to you then?
In a way, no. Of course, there could have been a surprise round
the comer, but I think I could go back to Einstein, who said that
the Lord is subtle but not malicious.
Could I take you back to the Schrodinger cat paradox and ask you
whether, in this ensemble interpretation, the cat is actually alive or
dead but it's just that we can't know about it? Should one think of the
cat being alive or deadin ·aparticular case, even though we can't ever
find the answer?
Well, we can always record the answer. And the cat itself knows
whether it's alive or dead. I would have thought that the only way
of avoiding a paradox here is ~osay that we're not allowed to find
out, in any individual case. It comes, I think, to this question of
the nature of consciousness. Is consciousness important in the
measuring process in quantum mechanics? I think it's something
that a number of physicists have claimed is a crucial feature.
Yes. Do you believe the observer is involved in the measurement
process in a fundamental way?
No, because it would seem to me that we can just as well observe
by means of machines, cameras, video tape recorders, and the
recording equipment running here in this particular programme!
I don't see consciousness as relevant at all.
I think this maybe brings us on to the question of how
quantum mechanics can ever have been misused for explaining
114 The ghost in the atom

extrasensory perception, for explaining the phenomena

associated with Uri Geller, spoon bending, telepathy, precogni-
tion, and all those paranormal events which of course have great
public interest - great interest, for example, from the point of
view of our own survival after death. All of this is related to the
question of whether consciousness plays a role in fundamental
physical phenomena. If consciousness is important, then maybe
we can use our minds to control some very delicate physical
processes, and hence explain how psychokinesis, spoon bend-
ing, and other peculiar phenomena might occur. If consciousness
is not relevant then this possible connection seems to be cut.
Arthur Koestler argued in his book The Roots of Coincidence*
that, because quantum mechanics seems to have these bizarre
features associated with the Einstein-Podolsky-Rosen experi-
ment and the Schrodinger cat paradox, that therefore other
bizarre phenomena can also occur in the world. This is, I think, a
very dangerous, specious argument.
Guilt by association?
Yes. But of course, with the remarkable successes that we've had
in high energy physics that I have already described there is little
evidence for any bizarre phenomena. High energy physics is a
very precise, watertight area to work in. And, moreover I would
say that there is no hard evidence at all.for extrasensory percep-
tion.
A lot of people have been very impressedby how the spirit of modern
quantum theory appears to be rather in tune with ancient oriental
mysticism. So, quite apart from paranormal phenomena, do you
regardideas of mysticism as being of any value in modern physics?
No, I don't at all, in fact I'm rather horrified by these develop-
ments. It seems to me that there's a large amount of vague and
woolly thinking that is contained in Eastern mysticism. No
matter how modem science had developed, the mystics could
have said, 'Aha, I told you so!' It's rather like doing the same

* Hutchinson, London, 1972.

 John Taylor 115

exercise on the Bible, and picking out certain words and saying,
'Aha!, this contains all the works of James Joyce!' It's absolutely
ludicrous. The detailed precision of modem theoretical physics
surpasses anything that is continued in Eastern mysticism. On
the other hand if these mystical ideas are used as an entree into
modern physics then they may have value, but only if used as
stepping stones to the greater precision of the real thing.
Fine.Now you've said that you don't see consciousnessas relevant to
quantum theory but nevertheless there are a number of contending
interpretations of quantum mechanics in which consciousness is
involved in a fundamental way, Wigner's interpretationfor example.
There are also other types of interpretation such as the many univer-
ses interpretation. Now Aspect's experiments don't actually rule out
these alternative interpretations because they're meant to be pure
interpretations and hence be consistent with all of the known results
of quantum theory. What's more they also attempt to account for
what happens in individual cases. In other words, they seem to go
beyond what the ensemble·interpretation can do, providing more
complete information about a system and getting to grips with those
paradoxes.So what's your reaction to that?
Well, if it is right that they do satisfactorily come to grips with the
paradoxes then I would be delighted, but I don't believe that. I
am very doubtful about the consciousness interpretation mainly
because it involves one in an infinite regress. Also I can't see why
consciousness is so special because all it requires is an aggregate
of nerve cells. In fact, consciousness involves such large aggre-
gates of cells that it would be"'difficult to see how quantum effects,
which involve uncertainties in rather small objects, could be
significant.
As far as the many-universes interpretations are concerned, I
would have felt that, again, I'm not satisfied with their avoidance
of the various paradoxes, the EPR paradox and the Schrodinger
cat paradox. And as far as interpretations which .are dealing
essentially with an idea of hidden variables, or uncontrollable
variables, I would say that they could not even get as far as
quantum field theory.
116 The ghost in the atom

But in defence of the many-universes theory, it would be claimed, I

think, by their proponents, that the paradoxes such as that of
Schrodinger's cat are easily resolved because if you ask in any
individual case whether the cat is alive or dead, the answer is both.
0

And in one universe the cat is alive and in the other universe it's dead,
and that seems to be a. perfectly satisfactory explanation. In the
ensembleinterpretation the answer is ... well, we can't answer.
I'm afraid I don't see it as satisfactory. I really must confess that I
find the many-universes interpretation as bizarre. No, I'm sorry,
I'm a hard-nosed physicist. Since one has no idea of what goes on
in the other universes, they shouldn't be brought in.
It does of course have the other advantage that it may be able to make
sense of the notion of the quantum mechanicsof the entire universe -
quantum cosmology. Now, in the ensemble interpretation, doesn't
that presen"tyou with a difficulty, becausewe only have one universe,
so how can we ever talk about the quantum mechanics of the whole
universe?
Well, I think that is a problem, but it's one that we can face if we
have a universe that goes on for an infinite extent (i.e. is spatially
infinite). Because then we can only think of making local mea-
surements. We can't ever, in a universe of infinite extent, expect
to measure the whole of it. We make measurements in our
laboratories over finite ranges. I think it would really be too much
to expect that we can have a wave function to describe an
ensemble of infinite universes. It would be beyond our com-
prehension.
Quantum cosmologyreally is a non-starter then?
Well, no I'm not saying that, because we can have a wave
function to describe the whole universe, but we can only measure
bits of it, and so our ensemble interpretation can still work.
Provided we have a universe which is infinite in extent. If it is
finite in size, then we may have problems, in that we could then
think of a laboratory which would cover the whole of the univer-
se. So, indeed, if we find, by observing the deceleration of distant
 John Taylor 117

galaxies, that the universe is in fact going to collapse again (and

hence be finite in size), then we may be in trouble as far as our
quantum mechanical interpretation of an ensemble nature is
concerned. The difficulty with the many-universes interpreta-
tion, though, is that one is bringing in so many additional things
that we can never find out about. You can never work in these
other universes.
Of course,once again, the-proponentswould argue that, although the
physical conglomeration of universes might appear to be a rather
bulky and unwieldy structure, nevertheless the epistemology of the
theory is extremely elegant and slim because we don't need to make
many assumptions.
But the assumptions are so bizarre that I would say it is not at all
slim, and I would also reiterate that unless you can actually
observe anything in these other universes they should not be
introduced. You see, in the ensemble interpretation one is saying
that there's only a limited amount of information one can obtain.
But in the many-universes interpretation one is saying that there
is a plethora of information which one cannot obtain. And that's
because most of it is in other universes - in fact, an infinite
amount of it is in other universes.
So really you're saying that both interpretationsforsake information;
in the case of the ensembleinterpretation we simply say that we can't
answer questions about individual systems, in the case of the many
universes one we can't answer questions about the other universes?
That's right, yes. I would choose having less information than
having information that I can never find out about. But then I
wouldn't ~ven call it information, I would call it hallucination!
 8
David Bohm

Before his retirement David Bohm was Professor of Theoret-

ical Physics at Birkbeck College London and for 30 years has
been an acknowledged world authority on quantum
mechanics. He was responsible for recasting the EPR experi-
ment in its modem form. Throughout his career Bohm has
been a leading advocate of the hidden variables school of
thought, and has written many papers attempting to formu-
late a detailed theory. More recently, along with his co-
worker Basil Hiley, he has constructed a non-local theory of
quantum mechanics based on the idea of 'quantum poten-
tial'. Bohm is also well known for his philosophical deliber-
ations on modem physics.

Can you explain how your interpretation differs from, let's say,
Bohr's Copenhagen interpretation of quantum mechanics, which I
think we can call the official view?
Yes. Well, there really is no very clear official view. I would say
that there are several variations. But the general idea is that
quantum mechanics cannot describe 'actuality' - that is, what
actually happens as a self-referent process. You see, if we say
something 'actually happens', quantum mechanics can only
describe what can be observed in a measuring apparatus.
Isn't that all that one would needfrom a theory-what we can observe
or measure?
Well, yes; if you presuppose that's all you need. But there is a
difficulty with this view. The Copenhagen interpretation only
gives a formula describing the probability of what can be
observed in a piece of apparatus. Yet the apparatus itself is
 David Bohm 119

supposed to be made of the very same sort of things we're

studying (i.e. particles subject to quantum effects).
Atoms?
Yes, atoms. Therefore if you want to discuss the existence of the
apparatus, you should in principle use another piece of apparatus
to look at it, and so on and so on.
This is the famous infinite regress?
Yes. Now Wigner has ended that regress by saying that only
when somebody becomes conscious of a phenomenon is it really
'actual'.
How do you feel about that particular interpretation?
Well, it is one way of looking at things. My own feeling is that
there is an area where it is true, especially in human relation-
ships; people becoming conscious of each other can have a
tremendous effect on each other. But I don't think it's really true
for the experimental situations on which physicists work in the
laboratory. It does seem to me that at this level the universe is
independently actual, and that we are part of it.

Do you think that the external world exists in some sense indepen-
dently of our existence, and independently of our observations?
Every physicist really believes that. For example, he talks about
the universe having evolved before there was anybody around to
look at it, except possibly God. Now unless you want to attribute
it to God, as Bishop Berkeley did (and most physicists don't want ·
to do that), you're unable to solve the problem of how the
universe exists without physicists to look at it - or without
somebody else to look at it.

As I understand it, the dispute between Einstein and Bohr was

that Einstein insisted that our observations merely uncover the
reality which already exists, whereas Bohr says that our observa-
tions actually create that reality. So you 're closer to Einstein's
position?
120 The ghost in the atom

Well, Bohr doesn't even say that. He says that we deal with
nothing but phenomena, appearances and regularities in
phenomena. And he says, essentially, that reality is ultimately
ambiguous and unspecifiable.

But you would find yourself much more in sympathy with Einstein's
point of view, that our observationsuncover a reality which in some
sense alreadyexists?
Well, I've already put myself in between Einstein and Bohr. I say
there is an area where our observations do create the reality, as in
human relationships: when people become aware of each other
and communicate they create the reality of society. But I think
that the universe as a whole does not depend on us to do that.

It seems to me that by adopting this position, you abolish mind

altogetherfrom the universe.
No, I say mind is real, mind may be very real. I specifically said
that between people mind has a tremendous effect. It affects the
body, it affects human relationships, it affects society.
But it doesn't affect atoms?
I don't think it has a significant effect on atoms. At least the
human mind doesn't. Perhaps you could take the view, as Bishop
Berkeley did, that the mind of God was responsible for the
creation of all matter. But then we must not equate ourselves with
God!
When you talk in your book Wholeness and the Implicate Order,
this wholenessis referringto both mind and matter-the mat.terwhich
exists around us. Could you say how mind and matter fit together in
this view of wholeness?
Yes, you're referring to the implicate order. Perhaps we could
talk first about Descartes, who made a distinction between mind
and matter. He said that there was thinking substance which we
call mind, and extended substance which we call matter. Now
they are so different it's hard to s~e how they could be related.
You see, our thoughts have no extension.
 David Bohm 121

Yes, we can't find where they're locatedin space, for example.

That's right. So what Descartes proposed was that God put clear
and distinct thoughts into the mind of man. God was able to do
it because he created both mind and matter- man and everything
else - and therefore he could put these thoughts into the minds
of men so that they could understand extended substance. Now.
when the notion that God may be used as an explanation of
things was dropped, then there was nothing left. Mind and
matter were left totally unrelated. However, the implicate order-
the enfolded order - shows that mind and matter may neverthe-
less be looked at in a similar way. Quantum mechanics may see
mind and matter as enfolded.

Could I ask you to explain what you mean by implicate or enfolded

order?Can you give a simple example?
Yes. Well, the simplest example is that if you fold a piece of paper
and make a pattern on it, and then unfold it you get all sorts of
new patterns. While the paper was folded the pattern was
implicit- in fact the word implicit means enfolded in Latin - and
therefore we could say the pattern was enfolded. Now quantum
mechanics suggest that this is the _way that phenomenal reality
comes about from a deeper order in which it is enfolded. Reality
unfolds to produce the visible order and folds back in. It is
constantly unfolding and enfolding at such a rate that it apears to
be steady. And now you can say that I'm proposing that thoughts
and feelings and mind work in a similar way. The very fact that
we say a thought is implicit means it contains another enfolded
thought. Right?

Yes, but enfolded in what? In what are our thoughts enfolded?

I'm going to avoid that question for the time being, and say I
want to show a si~ilarity of form between ming and matter. Now
this is what Descartes did not have. His belief was akin to saying
that thought is enfolded, and that matter is extended. However,
I'm saying both are enfolded and both unfold, therefore they are
similar in their basic structure, though they may be very different
122 The ghost in the atom

in many other ways. Their similarity in basic structure is what

enables us to understand the possibility of their being related.
What you're saying sounds to me very much like orientalphilosophy.
Perhaps some students of Zen would find these ideasvery familiar. Do
you see your thinking in this subject areaas giving support to oriental
mysticism?
Well it might do so, yes. But I think that this idea of enfoldment
has also been present in the West. You see, Nicholas of Cusa
proposed a similar idea several centuries ago. He had three
words: implicatio(enfolded), explicatio(unfolded) and complicatio
(all folded together). And he was saying that reality has this
enfolded structure: that eternity both enfolds and unfolds time.
Now, I don't think that we should categorize things as East or
West, but look at the ideas on their own merit. And I think that
quantum mechanics specifically is suggesting this enfolded
order. If you go into it as I have done, you can begin to make sense
of some of the strange properties of quantum mechanics by
looking at it this way.
Can you say why? What is the crucialfeature of quantum mechanics
that leads you to believe in the idea of enfolded order?
Well, it's the wave-particle duality: you may say that something
can unfold either as a wave-like or a particle-like entity. The
mathematics of quantum mechanics - if you look at it carefully-
corresponds to this enfoldment. It's very similar to the mathema-
tics of the hologram, you see.
I was going to suggest that the hologram seems like a very good
example of implicate or enfolded order.
Yes; it's one of the best ones we have where we see that a pattern
is enfolded into the photographic plate, and when you shine
light on it it's unfolded into a visible image. Each part of the
photographic plate contains information about the whole. So the
whole is unfolding from each region.
So your view of the atomic world is that in some way all of the
information about a particularphysical system is encodedsomewhere,
 David Bohm

but it's encodedin an obscureway that we don't normally have acc~ss

to?
Yes, but by definition it must be obscure when we look at it in the
usual way, because I think all encoding, such as that found in
DNA for example, is very obscure when looked at on the large
scale.

If we take the famous case of the position and the momentum of a

particle: according to the Heisenberg uncertainty principle, we can
choose either one or the other to be well defined, but not both at the
same time.
That's right, we can encode these properties so as to let one or the
other develop.

But you 're saying that in reality both of these quantities have
well-defined meanings, well-defined values, but somehow we can
only get at one or the other in experiments?
No, not exactly. You see, another example of the unfolded order
is a seed. If you take a seed it contains encoded information and
what happens is that when it's put in the ground the substance
of a plant develops from the air, the water, the soil, and the
energy from the sun. These materials are just moving in their
usual way, but with this tiny seed of information they start to
make a tree instead of whatever they would have done otherwise.
And now that tree can produce a seed which makes another tree,
and so on. Now, you can't say definitely that the tree was in the
seed: the kind of tree that developed - its shape and size -
depended not only on the seed, but also on the whole environ-
ment. And now if you go to a forest you can see trees are
continually growing, dying, and being replaced by new trees
and, if you visit your forest every hundred years, you would say
that apparently trees had moved from one place to another. In
fact, they're continually unfolding and enfolding, and that is the
picture I want to give of the motion of matter at the most
fundamental level. I want to say that life, mind and inanimate
matter all have a similar structure.
124 The ghost in the atom

Now, as far as I'm aware there's no known experiment which cannot

be satisfactorily explained by quantum mechanics as -iHs. Do you
disagreewith that?
No, but that's begging the question. If the only purpose of
physics is to explain experiments, then I think physics would
have been a great deal less interesting than it has been. I mean,
why do you want to explain experiments? Do you enjoy doing it
or what?
Well, I think my position, if I can be so boldas to put it, is that physics
concerns making models, that we make models of the world about us,
to help us to relate one type of observation to another. And we either
have good modelsor lessgood models.And that there is no such thing
as a 'real world' in the sense qf something which exists 'out there' to
which our modelsare mere approximations.As all we can ever do is to
make observations,what more can we ever.want of physics?
I think that observations and experiments are guided by our way
of thinking, and the questions we ask are determined by our way
of thinking. And for thousands of years people haven't asked
themselves the right questions. In quantum theory we're now
asking a certain kind of question and we're getting a certain kind
of answer. We may be putting ourselves into a trap, you see, by
restricting ourselves to this way of thinking.
So you think that by adopting a new way of thinking, a new approach
to the subject of microscopicphysics, we could perhaps construct a
very different set of questions .to ask, and maybe end up with a very
different theory?
Oh yes. And that's happened before many times. If you go back
to the motion of the planets - the old idea of epicycles led people
to ask certain questions, and then Newton's laws led us to ask a
very different set of questions. Statistical mechanics leads to one
set of questions; and quantum mechanics makes another set of
questions and so on. The questions you ask are determined
mainly by the theory, by theoretical conceptions.
But usually a particular way of approachinga topic is followed until
some experiment comes along that doesn't fit in with it.
 David Bohm 125

I think that's presupposing that that's the only way. You may
have to be banged on the head for 200 or 300 years before you'll
change your ideas. For example, I think non-locality was obvious
50 years ago, but now only a very few physicists realize it's there.
If they'd get.banged on the head for another 50 years maybe more
will realize it's there.
Let's talk a little more about non-locality. I wanted to ask you about
your response to the Aspect experiment, which htis been performed
recently. Now, as 1 understand it, given Aspect's results we have to
relinquish either what we might call objective reality - the external
wor!d existing independently of our observations - or locality - the
idea that different regions of the universe can't signal to each other
faster than light, crudely speaking. Which of these two are you
prepared to relinquish?
I would be quite ready to relinquish locality; I think it's an
arbitary assumption. I mean in the last few hundred years it has
been given tremendous weight. If you went back 1000 or 2000
years, almost everybody was thinking non-locally.
But now don't we run into horrendousparadoxessuch as being able to
signal our own past?
No, that's only if we assume that the present theories are the last
word. That's the wh_olepoint of considering new ways of looking
at things: put your questions in a different form, and you won't
get into these paradoxes.
So you want to abandon the special theory of relativity?
I don't say abandon relativity theory. I'm saying it's going to be
an approximation to a much broader point of view, just as
Newtonian mechanics is an approximation to relativity.
But you must certainly entertain the idea of faster-than-light
signalling.
Yes. I have a view which would entertain that and yet not
contradict any experiments which have been performed.
Can you think of any new experiments which could test this non-
localfeature of your theory?
126 The ghost in the atom

That's a bit premature, because we're in a peculiar situation, as

when Democritus proposed the atomic hypothesis several
thousand years ago. If you had said then that we will not think
about this unless we can propose an expe!iment to verify the
hypothesis then that would have been the end of that idea.
There was no way at that time to propose an experiment. Even
if anybody had been ingenious enough to propose one, there
was no equipment available that would have made it possible
anyway. And yet the idea was still valuable.
So are you saying that, inf act, not only are we not capableof testing
this faster-than-light signalling, but that you can't off-hand think of
a way in which it could be done?
I think you must entertain a notion for a long time before you can
do something new. If you say, 'I will only think of something the
minute you propose an experiment - otherwise I won't think of
it', how will you ever propose anything new? It often takes many,
many years to be able to see what sort of experiments co~ld be
done. It took 2000 years to get enough content into the atomic
theory to propose an experiment. So what would you have said?
That nobody should think about it until suddenly an idea would
occur for an experiment? Experiments would never have been
done if nobody had thought about it.
But do you believe that by using quantum effects and bringing about
faster-than-light contact betw.eenseparated systems, it will ever be
possible to literally send signals into the past?
No, I think that those paradoxes will not arise in the way I have
formulated the question. Those causal paradoxes only arise if
you say that relativi_ty is the absolute truth.
How precisely could this faster-than-light signalling come about?
Well, you see this would require some historical explanation. I
proposed another interpretation of quantum mechanics, an alter-
native interpretation, in 1951. There are two stages to this: first,
as applied to particles and, then, as applied to fields. Now, in the
first stage, I said that a~ electron is essentially a particle, but, in
 David Bohm 127

addition to all the other potentials it has, such as the electro-

magnetic potential, it has a new kind of potential, which at that
time I called the quantum potential.
Which, crudely speaking, we can think of as being something that
would jiggle the electronaround?
Yes. Now the quantum potential has new properties, and the
first of these is that its effects do not depend on its magnitude but
only on its form, so that it may have big effects at long distances.
In this way you can explain, say, the two-slit experiment.
This is normally explained, of course, by proposing the interference
between waves passing through the two slits.
It's not explained, it is merely described. If you said it was a
wave, that would be an explanation. But since the electrons
arrive as particles, it is no explanation. It is merely a sort of a
metaphorical way of talking. Right? There is no explanation. We
should say that quantum mechanics does not explain anything;
it merely gives a formula for certain results. And I'm trying to
give an explanation.
How does the quantum potential explain the interference?
Well, the quantum potential, which is carried as a wave, can
affect particles even quite far away from the slits, because, as I've
said, its influence depends on the form not the magnitude. Now,
the quantum potential or wave is quite different if the second slit
is open than if it's closed. So particles which pass through can be
deviated by the quantum potential even a long distance from the
slit, in sue~ a way that we produce these interference patterns.
Now this shows a new property of wholeness that in some ways
I agree is similar to what Bohr said, but I'm proposing to give an
explanation of it.
So part of the information carriedin this wave or potential would be
the experimental arrangement?
The experimental arrangement, yes. Also the states of all the
other particles in the system, and so on. So therefore you have
128 The ghost in the atom

what I call a non-local connection. This information brings about

the new quality of wholeness, in the sen~e that each part_now
moves in a way which reflects the state of the whole. It may be
that the connection is very weak under ordinary circumstances,
but special conditions exist in which it can become quite strong,
such as superconductivity or the two-slit experiment that I just
described.
This wave that you introducedmany years ago is clearly not the same
as the wave which we're familiar with when we talk about the wave
aspects of matter?
No, it's a new kind of wave which I call 'active information'. The
notion of active information is already familiar to us from com-
puters. Also, if I tell you something and you do something, that's
obviously active information. If I shouted 'fire', everybody
would move, so we know that in living intelligent systems, and
in computers, active information is a useful concept. Now what I
am proposing is that matter in general is not so different.
We're familiar with other types of potent~alslike electric potentials,
and gravitational potentials, how does your quantum potential com-
pare with those?
Well, you see it's similar in that it obeys certain equations,
though more subtle. It's different in that it does not necessarily
fall off with the distance, and that its effect is active and does not
depend on the inten_sity of the potential but only on the form.
So there really is nothing else like this in physics?
Yes, but we've often been in that situation, where something has
been introduced that wasn't there before.
Earlieryou implied that although the quantum potential idea enter-
tains the notion of faster-than-light signalling, it would not conflict
with our present experimental results. Can you tell us how this is
possible?
Yes. Well, this involves an extension of the idea of quantum
potential to a field, the entire field of the universe, which I call the
 David Bohm 129

super quantum potential. And this would take some explanation.

But, basically it will bring about a connection of fields at different
points instantaneously. Now it does not violate the principle of
relativity in any experiment, because one can show that the
statistics of experiments as done so far in the present system of
quantum mechanics will still come out in agre~ment with the
theory of relativity.
That is, forbiddingsignallingfaster than light?
There's no way to signal because we are using only sta~istical
experiments anyway.
We have no control over the influences that propagate faster than
light?
Yes that's right. So long as the present type of experiment is
done, the theory of relativity will still be saved. But if we could
manage to get deeper than that then we might find that there was
something faster than light. You see, we would then say that
relativity and quantum mechanics have the same limit, namely
the limit of statistics.
The usual objection about faster-than-light signalling is that, if we
can encode and transmit information, then that wot!,ld lead to
paradoxes; whereas you're saying that basically we don't have that
control over the microworld and everything is fuzzed out by the
unpredictablenature of quantum phenomena?
Yes, and one can even demonstrate that therefore there is no way
to get any inconsistency, and that, if we had some hold on the
thing that was deeper, we could then get beyond these limits.
It seems a little ironical that you are, if not contradicting Einstein's
special theory, at least drastically modifying it, perhaps against the
spirit of the original theory. What do you think Einstein would make
of this?
Yes, well I don't think anybody can necessarily expect everything
to turn out the way he expects it. Quite a few things turned out
that way for Einstein, but he can't have everything right!
130 · The ghost in the atom

One argument against the use of your quantum potential is that it

sounds like a very .complicatedthing: it doesn't have a simple set of
equations in the same way as, say, an electricfield does.
The equations are just Schrodinger's equation either for the one
or the many-body problem. Nature is telling us that the simple
idea of the electric field is too simple! And the point I'm trying to
make is that nature has a complexity and subtlety which
approaches that of the mind. I'm trying to say we have had too
simple a view about nature.
You think that this is perhaps due to the Newtonian tradition of
reductionism,of chopping the world up into lots of little pieces?
That's right. I don't know whether Newton was behind it, but
those who followed him certainly did that.
Whereas you would feel more in sympathy with, let's say, a synthetic
or holistic view, where we have to take into account the total system
to understand any component of it?
That's right. Yes I'm glad you brought that up, because now we
have to ask, 'how do we explain the ability to analyse the world
into independent parts in ordinary mechanics?' The answer is
that when the wave function has a certain property which I call
factorizability~ that's a mathematical term - you.Jind that the
various parts behave independently. Now, under ordinary cir-
cumstances that's a good approximation, but quantum mechani-
cal experiments are so designed as to produce situations where
the wave function is not factorizable, so they can demonstrate
wholeness. ·
Could I come back to the Aspect experiment? Are you saying that,
when the two photons travel in opposite directions and reach rela-
tively widely separatedpoints, their cooperationcan be attributed to
a signal passing between them faster than light?
Well, I think the word signal is wrong, because it has a certain
connotation which means that you can transmit messages. This
would not be that definite, but there would be a connection, I
prefer to use the word connection. You see, a connection is esta~-
 David Bohm 1:l 1

lished such that what happens to one particle will affect what
happens to the other one. Now, conventional quantum mechanics
doesn't explain the Aspect experiment. It merely gives you a
system of calculating (the results of that experiment). You see, I
think we should distinguish between explanation and systems of
calculus, and quantum mechanics is a calculus that enables you to
predict statistical results. But it has no explanation, and Bohr
emphasized that there was no explanation of any kind.
But is there ever explanation in physics? I mean, don't we simply
make models and invent languagefor them?
But models explain the thing in the sense that they show how it
comes about; the explanation makes it intelligible. Quantum
mechanics says that nature is unintelligible except as a calculus,
that all you can do is to compute with the equations and operate
your apparatus and compare the results.
Can you think of another area of physics, say a si'mplearea, where
you think that we actually have an explanation?
Well, a lot of classical physics gives an explanation in so far as it's
correct.
In what way though? Isn't it just language and models relating
observations? Where is the real explanation? We use the word
'explanation', but it seems to me rather meaningless, and that all
you 're really doing is relating observations together successfully.
I don't think so. You see, I think observations are a secondary
affair. I can't understand the tremendous emphasis in modern
physics for putting observations first. I think it's the positivist
philosophy which has done it. You must admit it mostly began
this century. If you had gone back 200 or 300 years, everybody
would have understood what explanation is and nobody would
have understood what the positivists were trying to do.
It's true. But supposing we take a particular example - why does the
apple fall - and we say the explanation is because there's a gravita-
tional field, and the Earth is acting on the apple, then we're still left
with the problem of explaining the gravitational field.
132 The ghost in the atom

Yes, but we are at least giving an account of what actually

happens: we say there is an apple, and it follows a path, and we
understand how the apple gets from here to there by passing
through stages in between. Now if we were to take quantum
mechanics, we would say that explanation has gone, we have an
apple here, we have another apple on the ground, we have no
notion of how one would connect up with the otheri we don't
even know whether it's going to happen, but we have a calculus
which gives the statistics of the number of apples arriving in
certain places. Now this is similar to the insurance company
saying we have statistics on how many people of a certain
category will die in a certain year, and that's all we care about! But
that is not an explanation.
But if we go back to the apple and we think of this purely classically,
after all we can only make observationson the apple, and we can make
measurements of where it is at certain times and so on: at the end of
the day, if we have a successful theory, then it will relate these
observations together.
I think that's a secondary affair. It does that, but more impor-
tantly it gives a conception of what is happening.
Oh, it gives one. It gives us a simple mental image of what's going on,
namely the apple is falling in a continuous trajectory to the ground.
But isn't this image simply an illusion?
Well, what are the calculations then?
The calculationsare a modelthat enableus to relate these observations
together.
Why do you want to relate them?
Because it seems to me that physics is about making observations of
the world.
Why is it about making observations? I mean, that's an idea
which started a few hundred years ago. People hold on to it
because they've been taught it by their teachers. But why do you
say that?
 David Bohm 133

Well, because for experimental physicists, it's their profession to

make measurementsof the world.
But physics did not begin purely with experiments, it began
with people asking questions. I mean, there would have been no
experiments if nobody had asked these questions. People were
interested in the world from a much broader point of view.
This raises Popper's idea about what we can regard as scientific. He
argues that you have to be able to show the theory is potentially
falsifiable, and that dependson being able to make observationswhich
could contradict the theo.ry.
That's Popper's idea. I'm trying to say, why should we take him
as the authority? There are all sorts of ideas which people have
had, and Popper has proposed an idea which has some merit, but
it needn't be the absolute truth. If one says that Popper has given
the absolute last word as to what science is, then why should I
accept that?
To summarize, then, I think that in the absence of any experiment to
the contrary, all we are debating here are really different philosophi-
cal standpoints?
Yes, well, originally the word philosophy meant love of wisdom.
Now it becomes a sort of technique. Also I think our modern age
is falling into reducing everything to techniques, and it takes
away the significance of everything. I think that people have
gradually fallen into that, and have said that anything else which
doesn't fit that simply is of no consequence. You must notice this
has _developed historically. You can't regard it as an absolute
truth.
But although we sit here discussingwhat we might call philosophy -
and there's been a great deal of discussion about the conceptual
foundations of quantum mechanics which seems to me to be mere
philosophy - nevertheless if I'm correct, you do foresee a time in the
future (we don't know how long in the future) where real experiments
will be done which will actually expo~ethe weaknesses of the present
interpretation of quantum mechanics.
134 The ghost in the atom

Yes, but I think that any fundamental new experiments arise

from philosophical questions. If you go into history, in the Greek
times, science was largely speculative. People then corrected that
by bringing in experiments. Now we've gone the other way and
said experiments are almost the only thing there is to it. So, in
effect, we have gone to the opposite extreme. Science surely
involves several things? It ~nvolves insight into ideas, and this
insight precedes experiment. If you exclude philosophy you will
eventually exclude these things too. The only insight available
now is through mathematics: that's the only place people allow
themselves any freedom. They can play around with mathema-
tics as much as they like without experiments. I saw an article in
the New York Times a few months ago where t~ey said we have
supergravity, and they said it looks promising, but that we won't
be able to say anything definite for 20 years! So nobody minds,
as long as it's mathematics. People believe that mathematics is
truth, but anything else is not.
Well, it's true that mathematical eleganceis a criterionwhich people
have used in support of a theory where experiment is lacking.
Bu~ if you will allow mathematical elegance, will you not allow
elegance in the conception? Every physicist has at least a tacit
philosophy but the present generally accepted philosophy is
extremely inelegant. It's really crude.
But nevertheless- and I'm sorry to keep coming back to this - do you
feel that there will be a stage in the future where it will be possibleto do
experiments to discriminatebetween these different interpretations?
I think there will be, but there won't be if you don't first consider
these ideas seriously in the absence of experiments.
But you_don't have any particular experiments in mind to propose at
this stage?
No, but I'm trying to say that if everybody took that attitude,
saying we will not listen to anything anybody has to say until he
proposes an experiment, then nobody can ever propose anything
fundamentally new.
 9
Basil Hiley

Basil Hiley is Reader in Physics at Birkbeck College, Univer-

sity of London. His research interests are in solid state,
liquid state and polymer physics, as well as the conceptual
basis of quantum mechanics. A long-standing collaborator
of David Bohm, he has for many years rejected the conven-
tional interpretation of quantum mechanics and attempted
to construct a theory more in keeping with 'common sense'
realism. His recent work with Bohm on the non-local 'quan-
tum potential' is a direct challenge to the orthodox view.

Aspect's recent experiment suggests that the traditionalapproach to

quantum mechanicsis alive and well, aridthat we can continue to use
it with confidence.Now, in your quantum potential theory you seem
to take a radically different app.roach.Why is it that you distrust the
conventional interpretation of quantum mechanics?
I think distrust is the wrong word. If anybody came to me and
said I want to solve a certain physical problem, I would recom-
mend that they go ahead with the conventional interpretation
because we. know it works and gives the correct answers. But
when you look at the conventional interpretation, and you try to
undertand what is going on when electrons produce an interfer-
ence patte~ for example, you have no physical way of explaining
the formation of this pattern.
Why do you feel it is necessaryto say what the electronis doing?After
all, in physics, and not just in quantum mechanics, the only accesswe
have to the world is through our instruments and experiments,and the
only data that we have to go on are our experimental results. Why do
you 1-pantto push the model of the external world so far that we need
136 The ghost in the atom

to talk about what the electron is doing even though we can't actually
observe this? Aren't our observationsenough?
No. I think that what we have to try and do is to build up a model
with which we can reinforce our intuitive ideas about the physi-
cal world. I've been brought up as a physicist, and I feel that
intuition has always helped a great deal. When I look at quantum
mechanics I find it completely counterintuitive. We just have a
prescription - a set of rules: we have a wave function which is
supposed to describe the state of the system; we have an operator
which we then apply to the wave function; and we get certain
predicted experimental values out. But this doesn't help me in
understanding the two-slit experiment, for example. What
exactly is the electron doing when it passes through the slits?
Does it go through one slit, or does it go through both? Now,
these are questions which are extremely important in trying to
get a feel of what is actually happening.
Let's try to make this absolutely clear. In the conventional, or
Copenhagen, interpretation one can talk about either the position of
an electron, say, or the momentum of an electron, but not both
simultaneously. And that's b~causewe don't know where the electron
is or how it's moving; it's meaninglessto even talk about the electron
having a well-defined position and a well-defined motion simultane-
ously. Now you're saying that the electron really does have a
well-defined position and motion even though in practice we can't
determine both of these quantities at the same time. Is that right?
Yes. The model I have been looking at was. originated by de
Broglie and subsequently developed by David Bohm. The diffi-
_culty with the usual approach is that we can only talk about
'observations' or 'measurements' . and cannot say what happens
in between. I feel we need to explore ontologies in which we can
raise such questions and this could mean that we can attribute a
precise position and momentum to a particle even though for the
observer these are unknown.
This is the so-calledquantum potential idea? Can you summarize the
essentialf eatures of this approach?
 Basil Hiley 137

First of all we imagine that there is an actual particle that has both
a definite momentum and p definite position. We then take its
wave function, and, rather than regarding this as a means of
calculating probabilities, we treat it as a real field, something
analogous to the electromagnetic field. The field can then influ-
ence the behaviour of this or other particles. Technically this is
achieved by an equation of motion derived from Schrodinger' s
equation. It contains an additional potential which we call the
quantum potential, since it modifies the classical behaviour of
the particles to produce results consistent with quantum
mechanics.
What sort of wave or field is this?
Although I have used the analogy of the electromagnetic field, it
actually has properties which are very different from the electro-
magnetic field.
What properties are these?
Perhaps I can best illustrate them through an example. We know
that if we pass electrons through a screen with two adjacent slits,
the result on the other side looks very much as if we're getting
waves interfering with each other. And, indeed, the orthodox
theory actually uses the wave function to describe this particular
wave phenomenon. But what we actually see at th 7 other end is
the arrival of individual electrons. So the wave is re"allyan average
of how a beam of individual electrons behave, and the intensity
of the wave corresponds to the number of electrons arriving at
that particular spot in a given time interval.
Now, orthodox theory says that you cannot actually predict
how each electron will arrive at the screen. But what the quantum
potential does is to enable you to calculate the set of individual
trajectories that gives rise to the interference pattern. You can
therefore look at the form of the quantum potential from the
calculations that you use. The quantum potential will contain
things like the slit width, the distance between the slits, and the
momentum of the particle; in other words it appears to have
some information about the environment of the particle. It is for
138 The ghost in the atom

this reason that one tends to regard the quantum potential as

arising from a field that is more like a field of information than a
physical field.
Perhaps I can take this analogy a little bit further. Suppose we
imagine that we have a ship which is guided by radar waves; th~
radar waves are fed into the ship's computer, and the ship then
adjusts its direction depending upon the information that it
receives from the radar waves. Now, we're trying to suggest that
the quantum potential arises from waves that are more like radar
waves. The quantum potential carries information about the
environment which is fed into the electron so that the electron
then adjusts its movement in order to produce the bunching
effect we observe on the screen.
So the motion· of the electron isn't forced upon it by the quantum
potential. The potential just carriesthe information to tell the electron
how to move?
Yes, it's an information potential. The more traditional way in
physics is to think that the electron is pushed around by the field,
just as water waves can push a ship around. The quantum
potential doesn't work like that, because you can actually multi-
ply the field by a constant and it doesn't change the force on the
particle. So it's not an ordinary classical force pushing the
electron around.
This quantum potential seems to be quite·unlike anything we have
encountered before in physics. Indeed, it seems to be rather remarka-
ble. If we think of the electronas being like the ship, moving under the
guidance of the information carriedby this potential, it rather makes
the electron seem a bit like a super-computer. Can we really imagine
that such a simple thing as an electron, which is supposed to be
structureless and have no internal parts, can respond in such a
complicatedway?
When I first began to think about this idea, I remember that
Richard Feynman had already pre-empted us in saying that he
thought of a point in sp~cetime as being like a computer with an
input and output connecting neighbouring points. The point
 Basil Hiley 139

would have a memory for all the fields and particles that are
possible and would actually act like a computer. So he has each
point in spacetime acting like a computer! I am only suggesting
that the electronmay act like a computer!
. Of course present experiments fail to reveal any structure
within the electron down to a distance of about 10- 16 centimetres.
But remember we've still got to go down to gravitational lengths.
which are about 10- 33 centimetres, so there's still room for a lot of
structure, even though it is going to be rather small on our scale.
So, you think that a particle like an electron could actually be a
composite body with internal parts that can act rather like the
components of a computer?
I wouldn't like to push the analogy too far, but it is a possibility.
I have a rather naive question now. I think it's very nice this analogy
of the ship with the radar,but of course to enable a ship to respond to
radarsignalsit still has to have some motive power of its own. So if the
electron gets the messagefrom this quantum potential which says:
'Move to the left!', well, how doesit move? What is its motive power?
The motive power comes from ..the quantum potential itself.
But I thought the quantum potential just triggereda response in the
electron,not drove the electron?
I have not made myself clear. It is the wave field that triggers a
response in the electron. This is translated into a quantum
potential which is part of an equation of motion. In terms of this
equation the quantum potential does give rise to a driving force
with energy coming from the self-activity of the electron. But I
don't like pushing to_ofar down this line, because I have a slightly
different image of the electron. I don't think the electron can be
completely separated from its environment. You see, one of the
things about quantum theory that Bohr emphasized is that we
have to look at the whole experimental situation. And what seems
to be coming across from the quantum potential approac~ is that
we can actually take his idea a bit further. If we cannot separate
particles and treat them as independent entities, we have to
140 The ghost in the atom

regard them -asaspects of the total situation. It's the whole system
that responds, so we shouldn't think of the electron as having
something which drives it from within. That would be like going
back to a mechanistic view of cogwheels or computer parts inside
the electron.

There was a time when it was proposed that maybe the qu~ntum
uncertainty of an electronwas due to it beingjiggled about by random
forces in its environment, in the conventional way that a wave may
toss a cork around on the surface of the sea. If we think of the electron
as following a zig-zag path, then it's easy to see if it's subject to
randomforces, it can be forced along a zig-zag path. But you seem to
be saying that the quantum potential tells the electron how it's got to
zig-zag about, but we can't find any motive power to cause the
. .
zig-zagging.
We always have the zero point energy. We know the vacuum
state is actually full of energy, and the orthodox theory exploits
that energy.

Yes. That's hard to push through in detail though, isn't it? For
example, you would expect there to be a difference between neutrons
and protons, and yet their quantum mechanical behaviour is very
similar.
But I'm not thinking o~
this in terms of the electromagnetic
background, because the quantum potential arises from a field
that is not like an electromagnetic field. It seems to be very
different; it seems to be much subtler than that.

So this zero point backgroundthat you're talking about is some sort of

background of the quantum potential field rather than the zero point
energy associatedwith other types of morefamiliar fields like electro-
magneticfields?
That's right.
If we turn directly to Aspect's experiment, in which one is dealing
with a two-particle system rather than a one-particle system, the
experiment shows that we have to make a choice:we can either reject
 Basil Hiley 141

what we might call 'reality', the idea of the external world existing
independently of our measurements,or we can reject locality, the idea
of all signalsand influencestravelling no faster than the speed of light.
Now, the quantum potential idea as I understand it, attempts to retain
at least a vestige of the old idea of objective reality, but the price one
has to pay is that you end up with a measure of non-locality. Is that
right?
Are you suggesting that quantum mechanics does not have that
non-locality in it?
No, I realize that quantum mechanics has an element of non-locality
as well, but of coursein the Copenhageninterpretation one is usually
quite happy to abandon the naive version of reality. And so it's
possible to make the Aspect experiment consistent with the absenceof
faster-than-light signalling.
If you're essentially saying that we can calculate the probabilities
using quantum calculus then I agree with you wholeheartedly.
We can do that. What is not clear to me from the orthodox theory
is how to understand Aspect's distant correlations. What the
quantum potential does, is to ~how unambiguously that there is
a non-local connection between the two. I know that if you go
back to Einstein's view, namely that reality is a description in
spacetime with only local interactions, then this would rule out
the quantum potential point of view. That, incidentally, is one of
the reasons why Einstein didn't think too much of the quantum
potential point of view.
_Doesthat worry you?
No, it doesn't. We now have experimental evidence to show that
reality does have some form of non-local element in it. What we
have to do is to ask why most experiments only reveal local
connections. We have already begun to see how that could be
explained by extending the idea of the quantum potential to the
quantum theory of fields.
Supposing one could push this programme through - and it is of
course tentative at this stage - but supposing we could push it
142 The ghost in the atom

through, it does seem to lead to the possibility off aster-than-light

communication.If we accept the theory of relativity this could enable
us to communicatebackward in time. Now this seems to be a recipefor
all sorts·of causal paradoxes, and it seems a high price to pay for
hanging on to some vestige of naive reality.
The quantum potential will not have any causal paradoxes in it
because it essentially requires an absolute spacetime in the
background, a quantum aether of the type suggested by Dirac.
Let me explain. We take the field theory and construct a superpo-
tential from the fields. You can then show that the superpotential
(which is governed by a Schrodinger superwave equation) is in
instantaneous contact with all particles (i.e. non-local). But when
you work out the statistical results of typical quantum experi-
ments you find that they are still Lorentz invariant (that is, they
obey the theory of relativity). So in other words, relativity in the
quantum potential approach comes out as a statistical effect, not
as an absolute effect.
So there is no way in practice of being able to ~endsignalsfaster than
light?
That is not clear. There is no way that we can see at the moment.
But if you have an absolute spacetime, or an absolute time, in the
background theri you don't get into causal loops. So the causal
paradoxes won't arise in this theory. But you will have instan-
taneous connections, and the question is: what do those
instantaneous connections mean? It is possible that we might be
able to find other experiments that exhibit those instantaneous
connections.
But if we take the behaviour of ordinary clocksas normally understood
within the context of the theory of relativity, then instantaneous
communication would amount in practice to communication back-
wards in time, would it not?
The point is that clocks are actually macroscopic collections of
particles; they are statistical in the way they function, and they
will not be able to detect these instantaneous connections.
 Basil Hiley 143

No, a clock wouldn't, but is it not possible that one can devise a
communication system which, although in your absolute spacetime
would produce an instantaneous connection, nevertheless in the
referenceframes as normally used by clocks in special relativity this
would amount to signallingbackwards in time?
It's not clear to me that such a possibility exists. If we go back to
Aspect's experiment, although the quantum potential shows
there is an instantaneous connection, when we look at the
statistical properties of the particles at each end of the connection,
they (the particles) appear to be independent; it's only in the
correlations that we see the non-locality. It's not clear to me that
those correlations can ever be transformed into a signal which
makes things go backwards in time.
At the moment, of course, it's not possible to use these correlations
actually as a signallingdevice.
Correct.
And in the conventional interpretation of quantum mechanics that
would never be the case. But it seems that with your interpretation it
is in principle possible, although in practiceyou can't think of how to
do it.
Well, I think this is of some merit for our theory, because it'll
make us think very carefully about whether we can do this kind
of thing or not.
It does seem as though you're aimed for a head-on clash with the
theory of relativity.
I don't see it that way because, as I say, at the moment it looks as
if it's the statistical effects which give us relativity. The problem
is how are we going to design experiments which will go beyond
this level to see these instantaneous connections. That's not clear
at the moment. What is clear is that the quantum potential
faithfully reproduces the results of quantum mechanics in our
present experimental regime. It is not at this stage doing any-
thing different.
144 The ghost in the atom

So am I right that the only place where the results of quantum

mechanics would differ from your theory is in the area of these
instantaneous communications- in the very area that's going to get
you into troublewith relativity?
The trouble is that in the orthodox interpretation of quantum
theory we can't ask the question of what is happening between
two separated systems. I cannot even think about the problem in
the formalism as it is now, because I just have a wave function.
And from that wave function I know how to work out the
correlations, but I don't know what is going on underneath, so I
can't raise the question. Now, maybe you think that we shouldn't
raise the question, but if we've got a theory which produces
exactly the same results as the orthodox theory, then it seems to
me that we should explore this further and find out whether we
will get any new physics. Maybe we won't and then you could
argue that it's all been a waste of time. But at least we have a
different view on this question.
OK. We'll let that point rest. But what advantagesdo you think your
approachhas other than providing us with a tidy model of reality?
What the orthodox approach has always left us with is this
so-called measurement problem. And if you look back through
the literature you'll find almost 300 papers on trying to solve the
measurement problem. What's more, the exponents of the
orthodox theory disagree over whether there even is a measure-
ment problem or not.
This is where we bringthe observerinto quantum theory in an explicit
way.
Yes. Now, when you're talking about the measurement problem,
you have to remember that the orthodox theory says that the
wave function describes the state of the system. And you then
use your apparatus to determine how this state develops. When
you use the apparatus you find that the state develops into what
is called a linear superposition. Let me take the following situa-
tion: suppose you have an experiment which gives you two
possibilities . . .
 Basil Hiley 145

Suppose we take live cat/dead cat.from the Schrodinger cat experi-

ment?
... That would do, yes. You've got two possibilities: either the
cat is alive or the cat is dead. If you now try to cal~ulate what
h~ppens in the quantum mechanical formalism, you find that the
state function for the cat at the end of the experiment is a linear
superpos~tion of a cat alive and a cat dead.
That means these two states are somehow overlapping each other.
These two states exist together in some way, yes. Now, when you
open up the box containing the cat, you then see whether it's
alive or dead, and that is referred to as 'the collapse of the wave
function'. You cannot bring about that collapse of the wave
function within the orthodox theory. So this has tempted people
as distinguished as Wigner to suggest that perhaps (the act of)
looking is a very important feature of quantum mechanics;
namely, that somehow consciousness enters into the situation.
When consciousness enters, the cat is either alive or dead, but
before that it's in a sort of a state of suspended animation, being
neither one thing nor the other.
I take it you don't like the idea of introducing mind into physics?
I don't see why mind should be introduced into physics at this
level. Another idea people have is the many-universes interpre-
tation of quantum theory. That is to say, when you look in the box
what you discover is whether you're in one branch of the
universe or another. One branch would correspond to the cat
being alive, the other branch would correspond to it being dead.
The world is split into its two alternatives?
That's right, and we just happen to be following one of those
alternatives. I'm not very keen on that idea because we seem to
be producing many universes of which only one is observed by
us. And so we have a rather strange situation. Now in terms of
the quantum potential, we don't run into that difficulty. Because
we have an actuality, namely the particle, and if the particle is in
one of those waves, then as far as the information (quantum)
146 The ghost in the atom

potential is concerned no information feeds back from the other

wave packet that we normally use in quantum mechanics (i.e. the
part of the wave function corresponding to the other branch of
the bifurcated universe).
Don't they interfere with each other?
They could eventually have the possibility of interfering with
each other, but the point is that, when the particle is in one wave
packet, as long as it is well separated from the other wave packet,
they will not interfere. However, if those two wave packets are
now allowed to overlap- then of course there is the possibility of
interaction between the two. But now when -rNemake a measure-
ment, one of the things that happens is that· an irreversible
process takes place; it is that irreversible process which is the key
to the collapse of the wave function in the quantum potential
approach. The 'empty' wave packet now cannot ever be brought
back again to overlap the wave function of the one that had the
particle in.
Why not? Becauseit's suddenly disappearedfrom the universe?
Perhaps we shouldn't talk about it actually disappearing from
the universe. Rather the information in the 'empty' wave packet
no longer has any effect, because during the act of measurement
the irreversible process introduces a stochastic or random distur-
bance which destroys the information of quantum potential of
the wave packet.
So it's not so much that a part of the wave disappears,it's that it gets
scrambled up among other things in an irreversible way. The wave
hasn't gone, it's just got completely interwoven with other waves,
losing its originalinformation.
I would accept that, yes. It's not active any more. And we've tried
to introduce a distinction between active information and inac-
tive information. That is, when an apparatus has undergone this
irreversible change, one wave packet becomes inactive.
So instead of a portion of the wave disappearing(as a result of the act
of measurement),it just becomesimpotent?
 Basil Hiley 147

Yes, that's right.

Let's look at the microscopicquantum scale again. You've said that a
particle such as an electron does in fact have both a well-defined
position and momentum. Yet we know from Heisenberg'suncertainty
principle that both of these cannot be measuredsimultaneously. How
do you explain that?
Well, that would just be a statistical effect. You see, when you
bring the measuring apparatus into an experiment you have a
many-body system. And the many-body system is necessarily
thermodynamic in nature, so you can never hope to know where
all the particles of the apparatus are. The very process of measure-
ment or preparation of a system in a given momentum state, say,
means you will then have all this uncertainty in it, and you can
never be sure exactly where the particle is. We've always got an
ambiguity because of this thermodynamic situation.

The uncertainty is introducedby the apparatus?

By the apparatus, yes.
It's our clumsinessin probing the system?
That's right. So in principle it would be causal in this interpreta-
tion. But in practice, because we are a thermodynamic system
and our apparatus is a thermodynamic system, we cannot hope
to determine the precise effect.
I can't then see how Planck's constant comes in becauseit seems that
if quantum uncertainty is purely thermodynamicalit's just a classical
effect, and I can't see why there should be any preferred scale of
action.
To me the value of Planck's constant is not really relevant to
quantum mechanics. I know I'm committing a heresy here,
because a lot of people are under the impression that if you put
Planck's constant equal to zero you will just.recapture classical
mechanics from the quantum formalism, and nothing could be
further from the truth.
But nevertheless it is a fundamental constant of nature which has a
148 The ghost in the atom

value, and if its value was different then the world would be a rather
different place.
I agree, but the quantum potential does contain Planck's con-
stant. And therefore if Planck's constant changed its value the
quantum potential would change its value.
But the question we were dealing with a moment ago concerned
Heisenberg'suncertainty principle in that if we carry out a measure-
ment on a system and it's the clumsiness of the apparatus - in a
classical thermodynamic sense - which introduces the apparent
quantum uncertainty, why is it on a scale determined by Planck's
constant? It seems a little bit mysterious why it should be that
particular scale- if the uncertainty is purely a classicaleffect.
But at the moment we're essentially getting this quantum poten-
tial out of the Schrodinger equation, which already has Planck's
constant in it.
Yes, but after all if we just go to a sort of classicalinterpretation of
measurement, and we have a particle, and we're trying to measureits
position and its momentum and so on, and we find that we're dong this
in a rather clumsy way, then there's a degree of uncertainty in the
results. And, of course, we know from thermodynamics that this is
often the case. But if we imagine refining our apparatus more and
more, and getting ever more precise results, quantum mechanics tells
us that there is an irreducibleuncertainty, and that's where Planck's
constant enters. Fromwhat you say about the apparatus causing this
disturbance, I can't see any reason for an irreduciblelevel of uncer-
tainty. Why should there be any particularscale of action?
This is a good question. I take your point. I think I agree with you
that it can't be just the irreversibility. But remember we are using
Schrodinger's equation to derive the quantum potential and
since it contains Planck's constant our analysis will also contain
it. So you are essentially asking me to explain why we need the
Schrodinger equation. I don't know the answer to that one.
 GLOSSARY

Action at a distance. The concept of two, separated systems

exerting physical effects on each other. In modem physics direct
action at a distance is replaced by field theory in ~ich separated
systems interact only by stimulating influences \~O propagate
through a field which extends across the space between them.
For example, the moon's motion, acting through-"..the inter-
mediary of its gravitational field, raises the ocean tides.
Aether. A hypothetical medium once thought to fill all space,
thereby defining a universal frame of reference relative to which
a material body's velocity through space could be defined.
Electromagnetic waves were regarded as vibrations of the aether.
The aether concept was rendered irrelevant by the special theory
of relativity.
Aspect's experiment. An experiment performed in 1982 by Alain
Aspect and co-workers to test the conceptual foundations of
quantum mechanics by checking Bell's inequality for pairs of
photons emitted simultaneously in single atomic transitions.
(Seep. 17 for full description.)
Bell's theorem (or inequality). After John Bell, who in 1965 proved
certain very general restrictions, in the form of mathematical
inequalities, of the degree to which the results of measurements
performed simultaneously on separated physical systems can be
correlated, given certain assumptions about the nature of physi-
cal action and the nature of reality.
Causality. The relationship between cause and effect. In classical
physics, an effect is restricted merely to follow a cause. In
relativistic physics causal connections are additionally limited
by the finite speed of light. Events which cannot be connected by
150 The ghost in the atom

influences travelling at the speed of light or less are causally

independent. One cannot affect the other.
CERN. Acronym for the Centre Europeen pour la Recherche
N ucleaire near Geneva in Switzerland, where some of the world's
most powerful subatomic particle accelerators are situated.
Conservation of momentum. A fundamental law of both classical
and quantum physics, which requires the total momentum of an
isolated system to remain constant, whatever internal changes
may occur in the system. In classical Newtonian mechanics,
momentum is defined as mass x velocity.
Copenhagen interpretation. The interpretation of quantum
mechanics associated with the name of Niels Bohr and his
research school in Copenhagen during the 1930s. The Copenha-
gen interpretation is usually accepted to be the conventional
viewpoint in spite of the continuing challenge to its position.
(Seep. 31 for full description.)
Einstein-Podolsky-Rosen experiment. A thought experiment
devised by Einstein and colleagues in 1935 designed to expose
the pecularities of quantum mechanics as interpreted by Bohr.
The experiment, which consisted of measurements performed
simultaneously on two quantum systems that at one time
interacted and then moved far ~part, forms the basis for Aspect's
real experiment. (Seep. 17 for full description.)
Electrody.namics.The theory that treats electromagnetic fields
together with their sources - electric charges, currents, and
magnets. Electrodynamics takes into account the motion of the
sources, the propagation of the fields and the interaction between
sources and fields.
EPR paradox (or experiment). See Einstein-Podolsky-Rosen
experiment.
Faster-than-light signalling. Hypothetical mechanism involving
physical effects propagating faster than light, therel?y enabling
events to be causally connected that would otherwise be regarded
as physically independent according to the theory of relativity.
Heisenberg'suncertainty principle. After Werner Heisenberg, this
is a mathematical formula that describes an irreducible level of
uncertainty that is always present (for quantum reasons) in
 Glossary 151

certain pairs of dynamical quantities when they are measured

together, e.g. the position and momentum of a particle.
Infinite regress.Philosophically unpalatable outcome of an argu-
ment in which each step depends logically on a succeeding step,
continuing in an unending sequence.
Irreversibleprocess. In some physical systems, e.g. the swinging
pendulum, the processes of interest can also occur in reverse. In
others, e.g. the diffusion of two different gases into each other,
the process is irreversible.
Locality. A physical restriction on the way in which events can
causally influence each other. In a general context, locality refers
to the idea that events can only influence other events in their
immediate vicinity. It also has a more restrictive meaning. If all
physical effects are assumed to propagate no faster than light,
two spatially separated events that occur simultaneously cannot
be causally connected. Hence an event can only be instantane-
ously connected to another if it is at the same spatial location.
Lorentz invariance. After H. A. Lorentz, this is a mathematical
concept connected with the symmetry properties of theories. It
relates the values of physical quantities observed in one frame of
reference to those observed in another, in a way consistent with
the principles of the special theory of relativity. A theory must
possess Lorentz invariance if it is to comply with the special
theory.
Non-locality. The hypothetical circumstances in which locality
fails. Some quantum processes have non-local flavour in that
spatially separated events can be correlated, but usually this is
assumed not to violate the more restrictive definition of locality
concerning instantaneous causal connection between spatially
separated events.
Planck's constant. A universal constant of nature, denoted h,
which quantifies the scale at which the quantum effects are
important. It is present in all mathematical descriptions of quan-
tum systems, and may appear in a variety of contexts, e.g. as the
ratio of the energy of a photon to the frequency of the light wave.
Quantum field theory. The quantum theory applied to fields, such
as the electromagnetic field. Quantum field theory forms the
152 The ghost in the atom

basis for current understanding of high-energy particle physics

and the fundamental forces that control subatomic matter.
Quantum potential. Mode of description of quantum systems
favoured by Bohm, Hiley and co-workers in which the erratic
and unpredictable fluctuations associated with quantum
behaviour are regarded as a consequence of a 'potential' field
analogous to, for example, the gravitational potential.
Relativity, theory of. The currently accepted description of space,
time and motion, and a cornerstone of twentieth-century
physics. The 'special' theory, first published by Einstein in 1905,
introduced some unusual ideas such as time dilation and the
equivalence of mass (m) and energy (E = mc2). A key result of the
special theory is t~at rio material body, physical influence, or
signal can exceed the speed of light (c). The later (1915) 'general'
theory included the effects of gravitation on spacetime structure.
Schrodinger's cat paradox. A paradox arising from a thought
experiment in which a quantum process is used to put a cat into
an apparent superposition of live and dead states. (See p. 28 for
full description.)
Schrodinger's equation. After Erwin Schrodinger, this is an equa-
tion, similar to that for a conventional wave, which describes the
behaviour of the quantum wave function.
State function. Abstract mathematical object that encodes all the
physical information needed to give the most complete available
physical description of a quantum system. In many cases the
state function can be represented as a wave function obeying
Schrodinger's equation.
Two-slit experiment. An experiment first performed by Thomas
Young in which light falls on two nearby narrow slits in a screen
and produces an interference pattern on an image screen, thereby
demonstrating the wave nature of light. (See p. 7 for a full
description.)
Virtual particles. The Heisenberg uncertainty principle permits
particles to appear and disappear again spontaneously, having
survived for only very short durations. These fleeting entities are
called 'virtual' to distinguish them from the more familiar, long-
lived, 'real' particles.
 Glossary 153

Wave function. A mathematical obje€t that describes the state of a

quantum system. In simple cases the behaviour of the wave
function is described by Schrodinger's equation.
Wave function, collapse or reduction of. The process that occurs
when a measurement is made of a quantum system, whereby the
wave function abruptly and discontinuously alters its structure.
The significance of this 'collapse' is contentious.
Wave packet. Sometimes the wave function of a quantum system
is concentrated in a narrow region of space. This configuration,
which implies that the particle being described is relatively
localized, is called a wave packet.
Zero point energy. An irreducible quantity of energy which,
according to quantum mechanics, always resides in a system that
is confined in some way. Its existence can be regarded as a
consequence of Heisenberg's uncertainty principle.
 FURTHER READING

T. Bastin (ed.), Quantum Theory and Beyond (Cambridge University

Press, Cambridge, 1971).
D. Bohm, Wholeness and the Implicate Order (Routledge & Kegan Paul,
London 1980).
J. F. Clauser and A. Shimony, 'Bell's theorem: experimental tests and
implications' in Reports on Progressin Physics 41, 1881-1927 (1978).
B. d'Espagnat, The Conceptual Foundationsof Quantum Mechanics
(Benjamin, New York, 1971); In Search of Reality (Springer-Verlag,
New York, 1983); 'Quantum theory and reality' in Scientific American,
November 1979, 158--81.
B. S. DeWitt, 'Quantum mechanics and reality', in Physics Today,
September 1970, 30-5.
B. S. DeWitt and N. Graham, The Many-Worlds Interpretation of Quantum
Mechanics (Princeton University Press, Princeton, N.J., 1973).
W. Heisenberg, Physics and Philosophy (Harper & Row, New York, 1959).
M. Jammer, The Philosophy of Quantum Mechanics Oohn Wiley, New
York, 1974).
N. D. Mermin, 'Is the moon there when nobody looks? Reality and the
quantum theory', in Physics Today, April 1985, 38--47.
A. I. M. Rae, Quantum Physics: Illusion or Reality (Cambridge University
Press, 1986).
G. Ryle, The Concept of Mind (Barnes & Noble, London, 1949).
J. von Neumann, Mathematical Foundations of Quantum Mechanics
(Princeton University Press, Princeton, N.J., 1955).
J. A. Wheeler and W. H. Zurek, Quantum Theory and Measurement
(Princeton University Press, Princeton, N.J., 1983).
E. P. Wigner, 'Remarks on the mind-body question', in The Scientist
Speculates -An Anthology of Partly-Baked Ideas, ed. I. J. Good, 284-302
(Basic Books, New York, 1962).
 INDEX

Bold numbers indicate whole chapters

aether 48-9, 142, 149 computers 34, 74, 138-9

Alley, Caroll 9, 66 intelligent 37, 97
arrow of time 86 consciousness 31-3, 54-5, 63-4, 68,
artificial intelligence 97 74-5,84,91-2,96, 101, 113-15,
atoms 2, 3 119, 145
Aspect, Alain vii, 2, 4~, 52 conventional interpretation, of
Aspect's experiment vii, 17-19, 39, quantum mechanics 102, 105,
2,46-8,52,76-9,97, 103-4, 107-8, 135-6, 144; see also Copenhagen
111, 113, 115, 125, 130-1, 135, interpretation
140-1, 149-50 Copenhagen interpretation, of
quantum mechanics 11, 12, 15, 20,
Bell, John 15, 17, 3, 77, 149 24, 26, 39, 70-1, 89, 102, 118, 136,
Bell's inequality, see Bell's theorem 141,150
Bell's theorem, 15--17, 38, 42, 45, 48,
103-4, 149
experimental test of 17-19, 39ff. Davisson, Clinton 3
Berkeley, Bishop 119-20 de Broglie, Louis 3, 56, 136
big bang 35, 38, 75, 89 delayed-choice experiment 9-11, 64-7
biology 32-3, 64, 68, 75, 80-1, 115, 123 Democritus 126
black holes 35 Descartes, Rene 120-1
Bohm, David 39, 43, 56, 103, 111, 8, deterministic universe 47
135--6, 152 Deutsch, David 36, 6
Bohr, Niels vi, 2, 11, 12, 15, 20-1, 24, DeWitt, Bryce 36, 103
26,31,34,39,58--61,65-7,70,72, Dirac, Paul 4, 142
76, 118-20, 127,131,139 dispersion relations 113
Born, Max 4 DNA 123
brain 32-4, 68, 74, 92, 98-100 dreams 62
branching universe 85
Eastern mysticism, see mysticism
causality 149-50 Eddington, Arthur 33
breakdown of 78-9, 141-4 Einstein, Albert vi, 2, 13-17, 20, 30,
CERN 52, 55, 112, 150 42-50,59,61,65,70, 75--8, 106-7,
cat paradox, see under Schrodinger 113, 119-20, 129,141
classical machanics 3, 4, 81, 112, 124 Einstein separability, see separability
collapse of the wave function 21, 28, Einstein-Podolsky-Rosen (EPR)
31-4,37,59,73-4,96, 110, 145-6, experiment 13-17, 40-1, 45, 48-9,
153 53, 55, 64, 107, 110, 114-15, 118,
communication, in the establishment 150,152
of meaning 62-3; see also electromagnetic field 94
knowledge electromagnetic radiation 2, 3
complementarity, principle of 11, 12, electromagnetic waves 2, 3, 5, 149
34, 38, 65, 69 electromagnetism 39
156 Index

electrons, Josephson, Brian 45

atomic 2-4
as computers 139 knowledge 73-5, see also meaning
momentum.and position of 6, 7, 11, Koestler, Arthur 114
13,58, 108,136,147
photo-emitted 2 Laplace, Pierre 80
in two-slit experiment 7-9 Leggett, Anthony 53
wavelike behaviour 3-11 life 123; see also biology
energy levels in atoms 2,3 locality 17, 39, 48, 78, 113, 141, 151
ensemble interpretation of quantum logical positivism 25, 131
mechanics, see statistical Lorentz invariance 48, 142, 151
interpretation
EPR experiment, see Einstein- many-universes interpretation of
Podolsky-Rosen experiment quantum mechanics 34-8, 55,
Everett, Hugh 35--6, 59--60, 82, 85-92, 59--60,82-105, 115-17
95-101 mechanics, see classical mechanics and
extrasensory perception 114 quantum mechanics
Maxwell, James Clerk 2, 3, 81, 94
faster-than-light signalling 16, 22, Maxwell's electromagnetic theory 2-4
42-5,48-50,55,78, 104, 125-30, meaning, and quantum mechanics
141-2, 150 62-3, 68; see also knowledge
Fermi, Enrico 4 measurement, .
Feynman, Richard 138 in classical physics 20
Fisher, R. A. 62 paradoxical nature of 26-30
Follesdal, D. 62, 64 in quantum physics 21ff., 47-8, 55,
free will 32, 47 59,61,91,96-7, 106-9, 146-7,
153
Galileo Galilei 93, 101 see also observation
Geller, Uri 114 measuring apparatus 27, 35, 147-8
ghost in the atom 34 mind 31-4, 47, 54-5, 59, 68, 73--6,91-2,
ghost in the machine 32, 34 98, 120-1, 130,145
God 46, 119-21 mind-body problem 32-4
gravitational lens effect 67 mysticism 12, 114-15, 122
gravity 35, 39
Newton, Isaac 3, 4, 93, 124, 130
Heisenberg, Werner 4, 6, 58 Newtonian mechanics, see classical
Heisenberg's uncertainty principle, mechanics
see uncertainty principle Nicholas of Cusa 122
hidden variables 13, 19, 38, 43, 77, 79, non-locality 22, 39, 111-12, 125, 128,
102-3, 112, 115, 118££. 141-3, 151
Hiley, Basil 39, 103, 111, 118, 9, 152 nucleus 2-4
holism, see wholeness
hologram 122 observation,
Hume, David 51 in classical physics 20, 72
in quantum physics 25-6, 31, 54, 56,
implicate order 121-3 59--63, 72-3, 96££.,124, 131-3, 136,
indeterminism 6££.,46; see also 144
uncertainty principle and see also measurement
unpredictability observer, role of 20ff., 47, 54, 59, 62, 68,
infinite regress 28, 30, 119, 151 74,76,80,90,94-100, 105,110,113
interference 5, 8, 9, 11, 27, 64-5, 96-9, origin of the universe, see big bang
127, 135, 137, 152
irreversibility 146, 151 Page, Don 88
 Index 157

paradoxes, Schrodinger, Erwin 4, 12, 28-30, 32

causal 125--6, 129,142 Schrodinger's cat paradox 28££., 110,
of quantum measurement 26--8; see 113-16, 145,152
also Schrodinger' s cat paradox Schrodinger's wave equation 130, 137,
parallel worlds 35, 55, 83ff. 142,148, 152-3
Peierls, Rudolf 5 separability 17, 42, 43
phenomenon, quantum 24, 120 singularities, in spacetime 35
photoelectric effect 2, 5 spectra, of atoms 4
photons, spin, intrinsic 78-9
in delayed-choice experiment 9, 10 statistical interpretation of quantum
introduction of 2, 3 mechanics 38, 106-17, 132
virtual 93, 111, 152 statistical mechanics 12, 80, 124
in Young's experiment 8, 9 superconductivity 4, 53, 128
Planck, Max 2 supergravity 134
Planck's constant 147-8, 151
Podolsky, Boris 13 Taylor, John G. 7
polarization measurements 5, 6, thermodynamics 12, 60-1, 77, 147-8
16-19, 21, 41 second law of 60-1, 86, 90
polarization of light 5, 6, 22-3, 28 time,
Popper, Karl 133 arrow of 86
positivism, philosophy of, see logical and the mind 55
positivism time travel 78, 88-9
preparation, of quantum states 108 two-slit experiment, see Young's
psychokinesis 114 experiment

quantum cosmology 34-5, 79, 89-90, uncertainty principle 6, 9, 13, 14, 33,
105, 116 58,62, 108-9, 123, 147-53
quantum electrodynamics 111 unpredictability, in quantum
quantum field theory, see under mechanics 9
quantum theory
quantum mechanics 1, 4 virtual particles 93, 152
quantum memory 97-8 von Neumann, John 27-31, 77
quantum potential 39, 102, 112, 118,
127, 130, 135££., 152 wave function 73-5, 87, 102, 136-7,
quantum theory 1 144, 152-3
of fields 111-13, 141, 151 wave function of the whole universe
origins of 2-4 79-81, 89, 116
quasars 67 wave-particle duality 2-12, 34, 122
Wheeler, John 9, 10, 23, 4
radioactivity 4 wholeness 39, 127-8, 130, 139-40
reality 17££., 62, 67, 72-6, 102-5, 109, Wigner;Eugene 31, 54-5, 63, 115, 119,
142 145--6
nature of 20--6, 30-1, 43, 50, 84, 89, Wooters, William 62, 88
91,94, 119-21, 144,149
objective 17, 20, 38, 42, 48, 50, 84, 89, Young, Thomas 2, 7, 152
91,94, 119-21, 144,149 Young's experiment 2, 7, 10, 11, 22, 87,
reductionism 130; see also wholeness 127-8, 136-7, 152
relativity, theory of 16, 48-50, 78, 103,
125--6, 129, 142-4, 149-52 Zen 122
Rosen, Na than 13 zero point energy 140, 153
Ryle, Gilbert 32
ISBN978-0-521-45728-6

11111111111
9 780521 457286 >