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

by David Deutsch

By ‘known’, I meant understood.

For example, no one could possibly memorize all known observational data on even so narrow a subject as the motions of the planets, but many astronomers understand those motions to the full extent that they are understood. This is possible because understanding does not depend on knowing a lot of facts as such, but on having the right concepts, explanations and theories. One comparatively simple and comprehensible theory can cover an infinity of indigestible facts.

Being able to predict things or to describe them, however accurately, is not at all the same thing as understanding them.

Facts cannot be understood just by being summarized in a formula, any more than by being listed on paper or committed to memory. They can be understood only by being explained. Fortunately, our best theories embody deep explanations as well as accurate predictions. For example, the general theory of relativity explains gravity in terms of a new, four-dimensional geometry of curved space and time. It explains precisely how this geometry affects and is affected by matter. That explanation is the entire content of the theory; predictions about planetary motions are merely some of the consequences that we can deduce from the explanation.

Scientific theories explain the objects and phenomena of our experience in terms of an underlying reality which we do not experience directly. But the ability of a theory to explain what we experience is not its most valuable attribute. Its most valuable attribute is that it explains the fabric of reality itself.

the explanatory power of a theory is paramount and its predictive power only supplementary.

Prediction – even perfect, universal prediction – is simply no substitute for explanation.

Knowing the outcome of any given idea or experiment does not help improve the idea or experiment. Knowing that an experiment will fail does not explain why it will fail such that it can be corrected—such that a/the correct answer could be found.

Not until we already had a theory, and had thought of an experiment that would test it, could we possibly ask the oracle what would happen if the theory were subjected to that test. Thus, the oracle would not be replacing theories at all: it would be replacing experiments. It would spare us the expense of running laboratories and particle accelerators.

The oracle would be very useful in many situations, but its usefulness would always depend on people’s ability to solve scientific problems in just the way they have to now, namely by devising explanatory theories. It would not even replace all experimentation, because its ability to predict the outcome of a particular experiment would in practice depend on how easy it was to describe the experiment accurately enough for the oracle to give a useful answer, compared with doing the experiment in reality.

To put that another way: there already is one such oracle out there, namely the physical world. It tells us the result of any possible experiment if we ask it in the right language (i.e. if we do the experiment), though in some cases it is impractical for us to ‘enter a description of the experiment’ in the required form (i.e. to build and operate the apparatus). But it provides no explanations.

Is LLM/AI the natural language, theoretical complement to our physical world predictor—the former taking prompts as inputs where the latter takes experiments as inputs?

The overwhelming majority of theories are rejected because they contain bad explanations, not because they fail experimental tests. We reject them without ever bothering to test them.

What we test are new theories that seem to show promise of explaining things better than the prevailing ones do.

Passing experimental tests is only one of many things a theory has to do to achieve the real purpose of science, which is to explain the world.

explanations are inevitably framed partly in terms of things we do not observe directly: atoms and forces; the interiors of stars and the rotation of galaxies; the past and the future; the laws of nature. The deeper an explanation is, the more remote from immediate experience are the entities to which it must refer. But these entities are not fictional: on the contrary, they are part of the very fabric of reality.

Along with Einstein, Mach, and Lorentz in his belief in the things we cannot see being real parts of our reality—attempts to describe reality.

if something is, in principle, predictable, then a sufficiently complete explanation must, in principle, make complete predictions (among other things) about it. But many intrinsically unpredictable things can also be explained and understood.

The idea that evolution is a theory that explains how things get to be yet cannot predict the existence of elephants—it explains them but doesn’t predict them. Explanations then are more important than predictions because they lead to understanding of the unpredictable parts of reality.

Because understanding comes through explanatory theories, and because of the generality that such theories may have, the proliferation of recorded facts does not necessarily make it more difficult to understand everything that is understood.

explanation is a strange sort of food – a larger portion is not necessarily harder to swallow. A theory may be superseded by a new theory which explains more, and is more accurate, but is also easier to understand, in which case the old theory becomes redundant, and we gain more understanding while needing to learn less than before.

Better explanations simplify. They require fewer functional parts, which makes them harder to vary, to explain reality in a falsifiable way. They’re better fit to reality. Explanations don’t grow exponentially as we learn more. Rather, there become fewer and simpler explanations that lead to greater understanding.

But it is hard to give a precise definition of ‘explanation’ or ‘understanding’. Roughly speaking, they are about ‘why’ rather than ‘what’; about the inner workings of things; about how things really are, not just how they appear to be; about what must be so, rather than what merely happens to be so; about laws of nature rather than rules of thumb.

But at present we know of nothing that is capable of understanding an explanation – or of wanting one in the first place – other than a human mind. Every discovery of a new explanation, and every act of grasping an existing explanation, depends on the uniquely human faculty of creative thought.

Explanation and understanding at the center of creativity; essential to it. Synthesis—the collection of facts into a system that leads to an action or extrapolation—is not then a creative act. It offer no new explanation or better understanding on its own. It becomes creative when an explanation is offered—that’s the creative act.

We understand the fabric of reality only by understanding theories that explain it. And since they explain more than we are immediately aware of, we can understand more than we are immediately aware that we understand.

This is what Linus Pauling and Charlie Munger mean when they say they have a fundamental model of reality that they use to put new facts into context. They couldn’t when conceived of the facts—predicted them—before knowing about them, but they’re not surprised by them when they see the facts because they can be explained by the best explanations already known.

With a very deep theory, the recognition that it explains a given phenomenon may itself be a significant discovery requiring independent explanation.

Furthermore, it’s not obvious that new facts fit existing explanations. There is still work required to determine if a new fact can be understood with our best explanations. And Charlie Munger taught that one ought to do just that until you cannot do otherwise. Assume that new facts fit our best understanding, until it is clear it doesn’t. Then, a better, new explanation is required. This is the reason explanations need to be hard to vary. In fact, if an old explanation can be varied to fit the new fact, it might signal that the old explanation is not that good after all.

while our specific theories are becoming more numerous and more detailed, they are continually being ‘demoted’ as the understanding they contain is taken over by deep, general theories. And those theories are becoming fewer, deeper and more general. By ‘more general’ I mean that each of them says more, about a wider range of situations, than several distinct theories did previously. By ‘deeper’ I mean that each of them explains more – embodies more understanding – than its predecessors did, combined.

Also see Hardy’s mathematicians apology—his grasping at what it means for a problem to be important it has some depth. Here’s depth.

It is nevertheless quite evident that the deepening, unifying tendency I have been describing is not the only one at work: a continual broadening is going on at the same time. That is, new ideas often do more than just supersede, simplify or unify existing ones. They also extend human understanding into areas that were previously not understood at all – or whose very existence was not guessed at. They may open up new opportunities, new problems, new specializations and even new subjects. And when that happens it may give us, at least temporarily, more to learn in order to understand it all.

Thus the issue of whether it is becoming harder or easier to understand everything that is understood depends on the overall balance between these two opposing effects of the growth of knowledge: the increasing breadth of our theories, and their increasing depth. Breadth makes it harder; depth makes it easier. One thesis of this book is that, slowly but surely, depth is winning.

We are not heading away from a state in which one person could understand everything that is understood, but towards it.

prediction is not explanation.

High-level phenomena about which there are comprehensible facts that are not simply deducible from lower-level theories are called emergent phenomena.

The purpose of high-level sciences is to enable us to understand emergent phenomena, of which the most important are, as we shall see, life, thought and computation.

when two theories are logically related, logic does not dictate which of them we ought to regard as determining, wholly or partly, the other. That depends on the explanatory relationships between the theories. The truly privileged theories are not the ones referring to any particular scale of size or complexity, nor the ones situated at any particular level of the predictive hierarchy – but the ones that contain the deepest explanations.

What makes a theory more fundamental, and less derivative, is not its closeness to the supposed predictive base of physics, but its closeness to our deepest explanatory theories.

Quantum theory is, as I have said, one such theory. But the other three main strands of explanation through which we seek to understand the fabric of reality are all ‘high level’ from the point of view of quantum physics. They are the theory of evolution (primarily the evolution of living organisms), epistemology (the theory of knowledge) and the theory of computation (about computers and what they can and cannot, in principle, compute).

explanation (roughly) A statement about the nature of things and the reasons for things.

emergence An emergent phenomenon is one (such as life, thought or computation) about which there are comprehensible facts or explanations that are not simply deducible from lower-level theories, but which may be explicable or predictable by higher-level theories referring directly to that phenomenon.

Scientific knowledge, like all human knowledge, consists primarily of explanations. Mere facts can be looked up, and predictions are important only for conducting crucial experimental tests to discriminate between competing scientific theories that have already passed the test of being good explanations.

The definition of a good explanation is independent of its predictive powers. It is possible to separate most explanations from one another without experiments—separating the bad from the good.

Single-particle interference experiments such as I have been describing show us that the multiverse exists and that it contains many counterparts of each particle in the tangible universe.

In other words, particles are grouped into parallel universes. They are ‘parallel’ in the sense that within each universe particles interact with each other just as they do in the tangible universe, but each universe affects the others only weakly, through interference phenomena.

They are mostly orthogonal to one another, except for interference effects.

Each step takes the form of noting that the behaviour of objects that we observe can be explained only if there are unobserved objects present, and if those unobserved objects have certain properties. The heart of the argument is that single-particle interference phenomena unequivocally rule out the possibility that the tangible universe around us is all that exists. There is no disputing the fact that such interference phenomena occur.

The key fact is that a real, tangible photon behaves differently according to what paths are open, elsewhere in the apparatus, for something to travel along and eventually intercept the tangible photon. Something does travel along those paths, and to refuse to call it ‘real’ is merely to play with words. ‘The possible’ cannot interact with the real: non-existent entities cannot deflect real ones from their paths. If a photon is deflected, it must have been deflected by something, and I have called that thing a ‘shadow photon’. Giving it a name does not make it real, but it cannot be true that an actual event, such as the arrival and detection of a tangible photon, is caused by an imaginary event such as what that photon ‘could have done’ but did not do. It is only what really happens that can cause other things really to happen. If the complex motions of the shadow photons in an interference experiment were mere possibilities that did not in fact take place, then the interference phenomena we see would not, in fact, take place.

This is where the explanation becomes “hard to vary.” Not playing with words. The effects are real. If they aren’t, then the outcome of the experiment is not observed. Change the definition of the core explanation—tangible photons interacting with shadow photons—and the explanation falls apart.

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.

interference The effect of a particle in one universe on its counterpart in another. Photon interference can cause shadows to be much more complicated than mere silhouettes of the obstacles causing them.

In the inductivist theory of scientific knowledge, observations play two roles: first, in the discovery of scientific theories, and second, in their justification. A theory is supposed to be discovered by ‘extrapolating’ or ‘generalizing’ the results of observations. Then, if large numbers of observations conform to the theory, and none deviates from it, the theory is supposed to be justified – made more believable, probable or reliable.

It is hard to know where to begin in criticizing the inductivist conception of science – it is so profoundly false in so many different ways. Perhaps the worst flaw, from my point of view, is the sheer non sequitur that a generalized prediction is tantamount to a new theory.

It is also not true that repeating our observations is the way in which we become convinced of scientific theories. As I have said, theories are explanations, not merely predictions. If one does not accept a proposed explanation of a set of observations, making the observations over and over again is seldom the remedy. Still less can it help us to create a satisfactory explanation when we cannot think of one at all.

repeated observations cannot justify theories,

a more basic misconception: namely, that the inductive extrapolation of observations to form new theories is even possible. In fact, it is impossible to extrapolate observations unless one has already placed them within an explanatory framework.

the same observational evidence can be ‘extrapolated’ to give two diametrically opposite predictions according to which explanation one adopts, and cannot justify either of them,

What, then, is the pattern of scientific reasoning and discovery? We have seen that inductivism and all other prediction-centred theories of knowledge are based on a misconception. What we need is an explanation-centred theory of knowledge: a theory of how explanations come into being and how they are justified; a theory of how, why and when we should allow our perceptions to change our world-view. Once we have such a theory, we need no separate theory of predictions. For, given an explanation of some observable phenomenon, it is no mystery how one obtains predictions. And if one has justified an explanation, then any predictions derived from that explanation are automatically justified too.

Problem-solving does begin with an inadequate theory – but not with the notional ‘theory’ consisting of past observations. It begins with our best existing theories. When some of those theories seem inadequate to us, and we want new ones, that is what constitutes a problem.

scientific discovery need not begin with observational evidence. But it does always begin with a problem. By a ‘problem’ I do not necessarily mean a practical emergency, or a source of anxiety. I just mean a set of ideas that seems inadequate and worth trying to improve.

an unexpected observation never initiates a scientific discovery unless the pre-existing theories already contain the seeds of the problem.