The Fabric of Reality: The Science of Parallel Universes--and Its Implications

by David Deutsch

Feynman conjectured that it would not be necessary to use a literal copy of the environment being rendered: that it would be possible to find a much more easily constructed auxiliary device whose interference properties were nevertheless analogous to those of the target environment. Then a normal computer could do the rest of the rendering, working through the analogy between the auxiliary device and the target environment. And, Feynman expected, that would be a tractable task. Furthermore, he conjectured, correctly as it turned out, that all the quantum-mechanical properties of any target environment could be simulated by auxiliary devices of a particular type that he specified (namely an array of spinning atoms, each interacting with its neighbours). He called the whole class of such devices a universal quantum simulator.

Classical environments rendered by analog computers while the quantum mechanical bits to be rendered are calculated using auxiliary interference components to get the interference effects that are intractable for rendering/computing on analog computers.

In 1985 I proved that under quantum physics there is a universal quantum computer. The proof was fairly straightforward. All I had to do was mimic Turing’s constructions, but using quantum theory to define the underlying physics instead of the classical mechanics that Turing had implicitly assumed. A universal quantum computer could perform any computation that any other quantum computer (or any Turing-type computer) could perform, and it could render any finite physically possible environment in virtual reality. Moreover, it has since been shown that the time and other resources that it would need to do these things would not increase exponentially with the size or detail of the environment being rendered, so the relevant computations would be tractable by the standards of complexity theory. The classical theory of computation, which was the unchallenged foundation of computing for half a century, is now obsolete except, like the rest of classical physics, as an approximation scheme. The theory of computation is now the quantum theory of computation.

any solid body consists of an array of atoms, which are themselves composed of electrically charged particles (electrons, and protons in the nuclei). But because of classical chaos, no array of charged particles could be stable under classical laws of motion. The positively and negatively charged particles would simply move out of position and crash into each other, and the structure would disintegrate. It is only the strong quantum interference between the various paths taken by charged particles in parallel universes that prevents such catastrophes and makes solid matter possible.

Among the many technical difficulties of working at the level of a single atom or single electron, one of the most important is that of preventing the environment from being affected by the different interfering sub-computations. For if a group of atoms is undergoing an interference phenomenon, and they differentially affect other atoms in the environment, then the interference can no longer be detected by measurements of the original group alone, and the group is no longer performing any useful quantum computation. This is called decoherence. I must add that this problem is often presented the wrong way round: we are told that ‘quantum interference is a very delicate process, and must be shielded from all outside influences’. This is wrong. Outside influences could cause minor imperfections, but it is the effect of the quantum computation on the outside world that causes decoherence. Thus the race is on to engineer sub-microscopic systems in which information-carrying variables interact among themselves, but affect their environment as little as possible.

It turns out that, unlike classical computation, where one needs to engineer specific classical logic elements such as AND, or and NOT, the precise form of the interactions hardly matters in the quantum case. Virtually any atomic-scale system of interacting bits, so long as it does not decohere, could be made to perform useful quantum computations.

All those computations are performed in parallel, in different universes, and share their results through interference.

To those who still cling to a single-universe world-view, I issue this challenge: explain how Shor’s algorithm works. I do not merely mean predict that it will work, which is merely a matter of solving a few uncontroversial equations. I mean provide an explanation. When Shor’s algorithm has factorized a number, using 10500 or so times the computational resources that can be seen to be present, where was the number factorized? There are only about 1080 atoms in the entire visible universe, an utterly minuscule number compared with 10500. So if the visible universe were the extent of physical reality, physical reality would not even remotely contain the resources required to factorize such a large number. Who did factorize it, then? How, and where, was the computation performed?

My only grievance with an argument based on finite atoms in the universe as the limit of computation is Deutsch’s own argument that knowledge transforms our environment to things greater than the collection of atoms. That’s said, since computation is a finite task, bound in physical reality, then there is indeed a limit on what computations can be done, or as he said there are computations that are physically intractable. I suppose the knowledge required here to get around those limits is in fact interference and the explanation is to use the physical limits of the multi-verse to increase what can be computed.

a new, absolutely secure system of quantum cryptography. In 1989, at IBM Research, Yorktown Heights, New York, in the office of the theoretician Charles Bennett, the first working quantum computer was built. It was a special-purpose quantum computer consisting of a pair of quantum cryptographic devices designed by Bennett and Gilles Brassard of the University of Montreal. It became the first machine ever to perform non-trivial computations that no Turing machine could perform.

quantum computation Computation that requires quantum-mechanical processes, especially interference. In other words, computation that is performed in collaboration between parallel universes.

tractable/intractable (Rough-and-ready rule:) A computational task is deemed tractable if the resources required to perform it do not increase exponentially with the number of digits in the input.

decoherence If different branches of a quantum computation, in different universes, affect the environment differently, then interference is reduced and the computation may fail. Decoherence is the principal obstacle to the practical realization of more powerful quantum computers.

The laws of physics permit computers that can render every physically possible environment without using impractically large resources. So universal computation is not merely possible, as required by the Turing principle, it is also tractable.

Quantum computation is a qualitatively new way of harnessing nature.

To put it bluntly, the reason why the common-sense theory of time is inherently mysterious is that it is inherently nonsensical. It is not just that it is factually inaccurate. We shall see that, even in its own terms, it does not make sense.

So there is no single ‘present moment’, except subjectively. From the point of view of an observer at a particular moment, that moment is indeed singled out, and may uniquely be called ‘now’ by that observer, just as any position in space is singled out as ‘here’ from the point of view of an observer at that position. But objectively, no moment is privileged as being more ‘now’ than the others, just as no position is privileged as being more ‘here’ than other positions. The subjective ‘here’ may move through space, as the observer moves. Does the subjective ‘now’ likewise move through time?

Nothing can move from one moment to another. To exist at all at a particular moment means to exist there for ever. Our consciousness exists at all our (waking) moments.

They are all conscious, and subjectively they are all in the present. Objectively, there is no present.

What we experience are differences between our present perceptions and our present memories of past perceptions. We interpret those differences, correctly, as evidence that the universe changes with time. We also interpret them, incorrectly, as evidence that our consciousness, or the present, or something, moves through time.

What’s incorrect is our interpretation of our experience of time.

Time cannot flow. The idea of the flow of time really presupposes the existence of a second sort of time, outside the common-sense sequence-of-moments time. If ‘now’ really moved from one of the moments to another, it would have to be with respect to this exterior time. But taking that seriously leads to an infinite regress, for we should then have to imagine the exterior time itself as a succession of moments, with its own ‘present moment’ that was moving with respect to a still more exterior time — and so on.

But the sequence of moments itself, in pictures like Figures 11.1 — 11.3, is an exceptional entity. It does not exist within the framework of time — it is the framework of time.

For example, in saying that Faraday discovered electromagnetic induction ‘in 1831’ we are assigning that event to a certain range of moments. That is, we are specifying on which set of snapshots, in the long sheaf of snapshots of world history, that discovery is to be found. No flow of time is involved when we say when something happened, any more than a ‘flow of distance’ is involved if we say where it happened. But as soon as we say why something happened, we invoke the flow of time.

Common wisdom concepts of time: When—an instance—vs cause and effect—changes involving past events effects on future happenings. Static vs dynamic. Deutsch claims these two perceptions of time contradict; contradict how?

Our theories of physics are, unlike common sense, coherent, and they first achieved this by dropping the idea of the flow of time.

But if the future really is open (and it is!), then that can have nothing to do with the flow of time, for there is no flow of time. In spacetime physics (which is, effectively, all pre-quantum physics, starting with Newton) the future is not open. It is there, with definite, fixed contents, just like the past and present. If a particular moment in spacetime were ‘open’ (in any sense) it would necessarily remain open when it became the present and the past, for moments cannot change. Subjectively, the future of a given observer may be said to be ‘open from that observer’s point of view’ because one cannot measure or observe one’s own future. But openness in that subjective sense does not allow choices.

In reality, we make no choices. Even as we think we are considering a choice, its outcome is already there, on the appropriate slice of spacetime, unchangeable like everything else in spacetime, and impervious to our deliberations. It seems that those deliberations themselves are unchangeable and already in existence at their allotted moments before we ever know of them.

Let us imagine ourselves, magically and impossibly, outside spacetime (and therefore in an external time of our own, independent of that within spacetime). Let us slice spacetime into snapshots of space at each moment as perceived by a particular observer within spacetime, then shuffle the snapshots and glue them together again in a new order. Could we tell, from the outside, that this is not the real spacetime? Almost certainly. For one thing, in the shuffled spacetime physical processes would not be continuous. Objects would instantaneously cease to exist at one point and reappear at another. Second, and more important, the laws of physics would no longer hold. At least, the real laws of physics would no longer hold. There would exist a different set of laws that took the shuffling into account, explicitly or implicitly, and correctly described the shuffled spacetime.

Of course the inhabitants could not tell the difference. If they could, they would. They would, for instance, comment on the existence of discontinuities in their world, and publish scientific papers about them — that is, if they could survive in the shuffled spacetime at all. But from our magical vantage-point we can see that they do survive, and so do their scientific papers. We can read those papers, and see that they still contain only observations of the original spacetime. All records within the spacetime of physical events, including those in the memories and perceptions of conscious observers, are identical to those in the original spacetime. We have only shuffled the snapshots, not changed them internally, so the inhabitants still perceive them in the original order.

What Deutsch is saying is that while the external observer would see time out of order, the experience of the internal observer would be to see events as they happen continuously because that is how, in each snapshot of time, even out of order, they consciously observe their reality.

So the snapshots have an intrinsic order, defined by their contents and by the real laws of physics. Any one of the snapshots, together with the laws of physics, not only determines what all the others are, it determines their order, and it determines its own place in the sequence. In other words, each snapshot has a ‘time stamp’ encoded in its physical contents. That is how it must be if the concept of time is to be freed of the error of invoking an overarching framework of time that is external to physical reality. The time stamp of a snapshot is the reading on some natural clock that exists within that universe.

The determinism of physical laws about events in spacetime is like the predictability of a correctly interlocking jigsaw puzzle. The laws of physics determine what happens at one moment from what happens at another, just as the rules of the jigsaw puzzle determine the positions of some pieces from those of others. But, just as with the jigsaw puzzle, whether the events at different moments cause one another or not depends on how the moments got there. We cannot tell by looking at a jigsaw puzzle whether it got there by being laid down one piece at a time. But with spacetime we know that it does not make sense for one moment to be ‘laid down’ after another, for that would be the flow of time. Therefore we know that even though some events can be predicted from others no event in spacetime caused another. Let me stress again that this is all according to pre-quantum physics, in which everything that happens, happens in spacetime. What we are seeing is that spacetime is incompatible with the existence of cause and effect. It is not that people are mistaken when they say that certain physical events are causes and effects of one another, it is just that that intuition is incompatible with the laws of spacetime physics. But that is all right, because spacetime physics is false.

If the multiverse were literally a collection of spacetimes, the quantum concept of time would be the same as the classical one. As Figure 11.6 shows, time would still be a sequence of moments. The only difference would be that at a particular moment in the multiverse, many universes would exist instead of one. Physical reality at a particular moment would be, in effect, a ‘super-snapshot’ consisting of snapshots of many different versions of the whole of space. The whole of reality for the whole of time would be the stack of all the super-snapshots, just as classically it was a stack of snapshots of space. Because of quantum interference, each snapshot would no longer be determined entirely by previous snapshots of the same spacetime (though it would approximately, because classical physics is often a good approximation to quantum physics). But the super-snapshots beginning with a particular moment would be entirely and exactly determined by the previous super-snapshots. This complete determinism would not give rise to complete predictability, even in principle, because making a prediction would require a knowledge of what had happened in all the universes, and each copy of us can directly perceive only one universe. Nevertheless, as far as the concept of time is concerned, the picture would be just like a spacetime with a sequence of moments related by deterministic laws, only with more happening at each moment, but most of it hidden from any one copy of any observer.

This is the distinctive core of the quantum concept of time: Other times are just special cases of other universes. This understanding first emerged from early research on quantum gravity in the 1960s, in particular from the work of Bryce DeWitt, but to the best of my knowledge it was not stated in a general way until 1983, by Don Page and William Wooters. The snapshots which we call ‘other times in our universe’ are distinguished from ‘other universes’ only from our perspective, and only in that they are especially closely related to ours by the laws of physics. They are therefore the ones of whose existence our own snapshot holds the most evidence.

What we call the past and future are indistinguishable from other universes—in fact, we think of them in the same way, except that we relate to past and future by laws of physics and make exception for other universes by artificial boundaries between them and the one we experience this moment.

Quantum theory does not in general determine what will happen in a particular snapshot, as spacetime physics does. Instead, it determines what proportion of all snapshots in the multiverse will have a given property. For this reason, we inhabitants of the multiverse can sometimes make only probabilistic predictions of our own experience, even though what will happen in the multiverse is completely determined.

The determinism of quantum theory, just like that of classical physics, works both forwards and backwards in time. From the state of the combined collection of ‘heads’ and ‘tails’ snapshots at the later time in Figure 11.7, the ‘spinning’ state at an earlier time is completely determined, and vice versa. Nevertheless, from the point of view of any observer, information is lost in the coin-tossing process. For whereas the initial, ‘spinning’ state of the coin may be experienced by an observer, the final combined ‘heads’ and ‘tails’ state does not correspond to any possible experience of the observer. Therefore an observer at the earlier time may observe the coin and predict its future state, and the consequent subjective probabilities. But none of the later copies of the observer can possibly observe the information necessary to retrodict the ‘spinning’ state, for that information is by then distributed across two different types of universe, and that makes retrodiction from the final state of the coin impossible.

In this case the laws of quantum mechanics predict that no observer who remembers seeing the coin in the ‘predictably heads’ state can see it in the ‘tails’ state: that is the justification for calling that state ‘predictably heads’ in the first place. Therefore no observer in the multiverse would recognize events as they occur in the ‘spacetime’ defined by the line. All this goes to confirm that we cannot glue the snapshots together in an arbitrary fashion, but only in a way that reflects the relationships between them that are determined by the laws of physics. The snapshots along the line in Figure 11.8 are not sufficiently interrelated to justify their being grouped together in a single universe. Admittedly they appear in order of increasing clock readings which, in spacetime, would be ‘time stamps’ which would be sufficient for the spacetime to be reassembled. But in the multiverse there are far too many snapshots for clock readings alone to locate a snapshot relative to the others.

Seems to be a similar idea to the uncertainty principle of quantum mechanics—that an observer of the final state of a probabilistic event cannot determine the path of the object from its initial state.

In this jigsaw-puzzle multiverse, which neither consists of a sequence of moments nor permits a flow of time, the common-sense concept of cause and effect makes perfect sense. The problem that we found with causation in spacetime was that it is a property of variants of the causes and effects, as well as of the causes and effects themselves. Since those variants existed only in our imagination, and not in spacetime, we ran up against the physical meaning-lessness of drawing substantive conclusions from the imagined properties of non-existent (‘counter-factual’) physical processes. But in the multiverse variants do exist, in different proportions, and they obey definite, deterministic laws. Given these laws, it is an objective fact which events make a difference to the occurrence of which other events.

spacetime physics Theories, such as relativity, in which reality is considered to be a spacetime. Because reality is a multiverse, such theories can at best be approximations.

Time does not flow. Other times are just special cases of other universes.

Time travel may or may not be achieved one day, but it is not paradoxical. If one travels into the past one retains one’s normal freedom of action, but in general ends up in the past of a different universe.

Kuhn’s theory suffers from a fatal flaw. It explains the succession from one paradigm to another in sociological or psychological terms, rather than as having primarily to do with the objective merit of the rival explanations. Yet unless one understands science as a quest for explanations, the fact that it does find successive explanations, each objectively better than the last, is inexplicable. Hence Kuhn is forced flatly to deny that there has been objective improvement in successive scientific explanations, or that such improvement is possible, even in principle:

So the growth of objective scientific knowledge cannot be explained in the Kuhnian picture. It is no good trying to pretend that successive explanations are better only in terms of their own paradigm. There are objective differences.

the truth is not obvious, and that the obvious need not be true; that ideas are to be accepted or rejected according to their content and not their origin; that the greatest minds can easily make mistakes; and that the most trivial-seeming objection may be the key to a great new discovery.

Pragmatic instrumentalism has been feasible only because, in most branches of physics, quantum theory is not applied in its explanatory capacity. It is used only indirectly, in the testing of other theories, and only its predictions are needed. Thus generations of physicists have found it sufficient to regard interference processes, such as those that take place for a thousand-trillionth of a second when two elementary particles collide, as a ‘black box’: they prepare an input, and they observe an output. They use the equations of quantum theory to predict the one from the other, but they neither know nor care how the output comes about as a result of the input. However, there are two branches of physics where this attitude is impossible because the internal workings of the quantum-mechanical object constitute the entire subject-matter of that branch. Those branches are the quantum theory of computation, and quantum cosmology (the quantum theory of physical reality as a whole). After all, it would be a poor ‘theory of computation’ that never addressed issues of how the output is obtained from the input! And as for quantum cosmology, we can neither prepare an input at the beginning of the multiverse nor measure an output at the end. Its internal workings are all there is. For this reason, quantum theory is used in its full, multiverse form by the overwhelming majority of researchers in these two fields.

(An artificial intelligence is a computer program that possesses properties of the human mind including intelligence, consciousness, free will and emotions, but runs on hardware other than the human brain.)

a basic tenet of rationality — that good explanations are not to be discarded lightly.

Popper’s epistemology has, in every pragmatic sense, become the prevailing theory of the nature and growth of scientific knowledge. When it comes to the rules for experiments in any field to be accepted as ‘scientific evidence’ by theoreticians in that field, or by respectable journals for publication, or by physicians for choosing between rival medical treatments, the modern watchwords are just as Popper would have them: experimental testing, exposure to criticism, theoretical explanation and the acknowledgement of fallibility in experimental procedures. In popular accounts of science, scientific theories tend to be presented more as bold conjectures than as inferences drawn from accumulated data, and the difference between science and (say) astrology is correctly explained in terms of testability rather than degree of confirmation. In school laboratories, ‘hypothesis formation and testing’ are the order of the day. No longer are pupils expected to ‘learn by experiment’, in the sense that I and my contemporaries were — that is, we were given some equipment and told what to do with it, but we were not told the theory that the results were supposed to conform to. It was hoped that we would induce it.

There is an explanatory gap.

Now, as I have said, my guess is that the brain is a classical computer and not a quantum computer, so I do not expect the explanation of consciousness to be that it is any sort of quantum-computational phenomenon. Nevertheless, I expect the unification of computation and quantum physics, and probably the wider unification of all four strands, to be essential to the fundamental philosophical advances from which an understanding of consciousness will one day flow.

But Darwin was able to wonder how laws of nature that did not mention elephants could nevertheless produce them, just as Newton’s laws produce ellipses. Although Darwin made no use of any specific law of Newton’s, his discovery would have been inconceivable without the world-view underlying those laws. That is the sense in which I expect the solution of the ‘What is consciousness?’ problem to depend on quantum theory. It will invoke no specific quantum-mechanical processes, but it will depend crucially on the quantum-mechanical, and especially the multi-universe, world-picture. What is my evidence? I have already presented some of it in Chapter 8, where I discussed the multiverse view of knowledge. Although we do not know what consciousness is, it is clearly intimately related to the growth and representation of knowledge within the brain. It seems unlikely, then, that we shall be able to explain what consciousness is, as a physical process, before we have explained knowledge in physical terms. Such an explanation has been elusive in the classical theory of computation. But, as I explained, in quantum theory there is a good basis for one: knowledge can be understood as complexity that extends across large numbers of universes.

Freedom has nothing to do with randomness. We value our free will as the ability to express, in our actions, who we as individuals are. Who would value being random? What we think of as our free actions are not those that are random or undetermined but those that are largely determined by who we are, and what we think, and what is at issue. (Although they are largely determined, they may be highly unpredictable in practice for reasons of complexity.)

The fruitfulness of the multiverse theory in contributing to the solution of long-standing philosophical problems is so great that it would be worth adopting even if there were no physical evidence for it at all. Indeed, the philosopher David Lewis, in his book On the Plurality of Worlds, has postulated the existence of a multiverse for philosophical reasons alone.

I have pointed out one possible contributory cause, namely that individually, all four theories have explanatory gaps that can make them seem narrow, inhuman and pessimistic. But I suggest that when they are taken together as a unified explanation of the fabric of reality, this unfortunate property is reversed.