The Fabric of Reality David Deutch

particular implications of those laws are relevant. First, every subatomic
particle has counterparts in other universes, and is interfered with only by
those counterparts. It is not directly affected by any other particles in those
universes. Therefore interference is observed only in special situations
where the paths of a particle and its shadow counterparts separate and then
reconverge (as when a photon and shadow photon are heading towards the
same point on the screen). Even the timing must be right: if one of the two
paths involves a delay, the interference is reduced or prevented. Second,
the detection of interference between any two universes requires an
interaction to take place between 
all the particles whose positions and other
attributes are not identical in the two universes. In practice this means that
interference is strong enough to be detected only between universes that are
very alike. For example, in all the experiments I have described, the
interfering universes differ only in the position of one photon. If a photon
affects other particles in its travels, and in particular if it is observed, then
those particles or the observer will also become differentiated in different
universes. If so, subsequent interference involving that photon will be
undetectable in practice because the requisite interaction between 
all the
affected particles is too complicated to arrange. I must mention here that the
standard phrase for describing this fact, namely ‘observation destroys
interference’, is very misleading in three ways. First, it suggests some sort of
psychokinetic effect of the conscious ‘observer’ on basic physical
phenomena, though there is no such effect. Second, the interference is not
‘destroyed’: it is just (much!) harder to observe because doing so involves
controlling the precise behaviour of many more particles. And third, it is not
just ‘observation’, but 
any effect of the photon on its surroundings that
depends on which path the photon has taken, that does this. For the benefit
of readers who may have seen other accounts of quantum physics, I must
briefly make contact between the argument I have given in this chapter and
the way the subject is usually presented. Perhaps because the debate
began among theoretical physicists, the traditional starting-point has been
quantum theory itself. One states the theory as carefully as possible, and
then one tries to understand what it tells us about reality. That is the only
possible approach if one wants to understand the finer details of quantum
phenomena. But as regards the issue of whether reality consists of one
universe or many, it is an unnecessarily complicated approach. That is why I
have not followed it in this chapter. I have not even stated any of the
postulates of quantum theory — I have merely described some physical
phenomena and drawn inescapable conclusions. But if one does start from
theory, there are two things that everyone agrees on. The first is that
quantum theory is unrivalled in its ability to predict the outcomes of
experiments, even if one blindly uses its equations without worrying much
about what they mean. The second is that quantum theory tells us
something new and bizarre about the nature of reality. The dispute is only
about what exactly this is. The physicist Hugh Everett was the first to
understand clearly (in 1957, some thirty years after the theory became the
basis of subatomic physics) that quantum theory describes a multiverse.
Ever since, the argument has raged about whether the theory admits of any

---

other interpretation (or re-interpretation, or reformulation, or modification,
etc.) in which it describes a single universe, but continues correctly to predict
the outcomes of experiments. In other words, does accepting the predictions
of quantum theory force us to accept the existence of parallel universes?
It seems to me that this question, and therefore the whole prevailing tone of
the debate on this issue, is wrong-headed. Admittedly, it is right and proper
for theoretical physicists such as myself to devote a great deal of effort to
trying to understand the formal structure of quantum theory, but not at the
expense of losing sight of our primary objective, which is to understand
reality. Even if the predictions of quantum theory could, somehow, be made
without referring to more than one universe, individual photons would still
cast shadows in the way I have described. Without knowing anything of
quantum theory, one can see that those shadows could not be the result of
any single history of the photon as it travels from the torch to the observer’s
eye. They are incompatible with any explanation in terms of only the photons
that we see. Or in terms of only the barrier that we see. Or in terms of only
the universe that we see. Therefore, if the best theory available to physics
did not refer to parallel universes, it would merely mean that we needed a
better theory, one that did refer to parallel universes, in order to explain what
we see.
So, does accepting the predictions of quantum theory force us to accept the
existence of parallel universes? Not in itself. We can always reinterpret any
theory along instrumentalist lines so that it does not force us to accept
anything about reality. But that is beside the point. As I have just said, we do
not need deep theories to tell us that parallel universes exist — single-