The Fabric of Reality and The Beginning of Infinity
Reality: more than we see
Whatever our best explanations say exists, exists. David Deutsch follows that rule into quantum physics and finds that the world we see is one thin layer of something far larger.
This page sets out his view whole, in seven steps, drawn from The Fabric of Reality (1997) and The Beginning of Infinity (2011). Four of the figures respond to you: you can fire photons through two slits, block the paths of a beam splitter, topple a domino computer and trace the four strands that hold the series together. The comparisons with other philosophers and physicists are on the thinker pages. Hugh Everett appears here as the physicist whose theory Deutsch defends; his own page tells his story.
1. The puzzle: what is real?
Everything you see is an appearance. The table in front of you is, on the physicist's account, almost entirely empty space, crossed by fields you cannot sense. Colour is not in the light, sound is not in the air, and the solid floor is a balance of electrical forces. The question of what lies behind the appearances is the oldest in philosophy, and there have always been two broad answers.
The first says: stop asking. Science predicts what we will observe, and that is all it can do. Talk of what really exists beyond observation is metaphysics, and metaphysics is empty words. This is instrumentalism, and in the twentieth century it became the official philosophy of quantum mechanics.
The second says that science explains, and an explanation is a claim about what is there. Atoms were once dismissed as a convenient fiction for doing chemistry. They turned out to be real, and microscopes now image them. Realism is the bet that our best theories are, roughly and fallibly, true descriptions of a world that exists whether or not anyone looks.
Deutsch is a realist without apology. He thinks instrumentalism is not modesty but a refusal to think, and that it has done real damage to physics. A theory that only predicts is like an oracle that answers every question and explains nothing. You could consult it for ever without understanding the world any better.
2. If it kicks back, it is real
In August 1763 James Boswell told Samuel Johnson that Bishop Berkeley's denial of matter, though obviously untrue, could not be refuted. Johnson struck his foot against a large stone until he rebounded from it, and answered, "I refute it thus." Philosophers have laughed at him ever since: the stone's resistance is just another sensation, and Berkeley never denied sensations.
Deutsch takes Johnson's side and makes the argument precise. The stone did not merely appear. It responded, in ways Johnson did not choose and could not have predicted without a theory of stones. Something that kicks back, that is complex and autonomous and does things our explanations have to account for, is real. His criterion is this: if our best explanation of what we observe refers to something, we should regard that thing as existing.
The criterion cuts in several directions at once.
- It rules out solipsism. A dream world that obeys its own laws, surprises us and resists our will is not a dream in any useful sense. Explaining it requires a world outside us, and the simplest such explanation is the world.
- It makes emergent things real. Deutsch's example is a copper atom at the tip of the nose of a statue of Winston Churchill in London. Why is it there? The true explanation runs through Churchill, the Second World War and a nation's gratitude, not through particle physics. High-level explanations are not shorthand for low-level ones. Often they are the only explanations there are.
- It makes abstract things real, a step that section 6 takes up.
- It ranks explanation above observation. We see neither the interior of the Sun nor the dinosaurs, yet our best explanations of what we do see require both. Observation is where theories are tested, not where reality stops.
3. Shadows
Light comes in indivisible lumps, called photons. Shine a single photon at a barrier with two narrow slits and it arrives at the screen behind as a single dot. Send photons through one at a time, with long gaps between them, and the dots slowly build up into stripes: bands of light separated by bands of darkness.
The dark bands are the puzzle. With only one slit open, photons do land there. Opening a second slit, which gives each photon a second way through, stops them arriving at those points. Something must be coming through the other slit and interfering with each photon. It cannot be another photon we can see, because only one is in the apparatus at a time. Deutsch calls it a shadow photon.
One photon at a time
No photons fired yet.
The shadow photons behave exactly like ordinary photons in every way except that we cannot see them. Barriers block them, lenses bend them and mirrors reflect them, just as they do the photons we see. On the criterion of section 2 they kick back: they change where the visible photon lands. So they are real.
And there are a great many of them. Deutsch argues from experiments with very fine interference patterns that there must be at least a trillion for every photon we see. They in turn interact with shadow atoms and shadow detectors that behave like ours. The shadows form a world as rich as ours. In fact they form many such worlds, which touch ours only through faint interference effects of this kind.
4. The multiverse
Hugh Everett proposed the theory in his doctoral thesis of 1957. Quantum mechanics describes the world with a wave function that evolves smoothly and deterministically. Textbooks add a second rule: when someone makes a measurement, the wave function collapses at random to a single outcome. Everett asked what happens if the second rule is simply dropped and the observer is described by the same equation as everything else. The answer: the observer comes to exist in many versions, each of which sees one outcome. Nothing collapses. Every possibility the equation allows is realised.
Deutsch speaks of the multiverse rather than of worlds that split like branches of a tree. Most of the time, histories that started the same stay the same. Instances of you that are identical in every respect are not separate things in any meaningful sense; Deutsch calls them fungible, like pounds in a bank account. When the outcome of a quantum event is recorded and amplified, as happens constantly in nature and not only in laboratories, the instances become different from one another. The differences spread outwards no faster than light, and the physics of how they spread is what physicists call decoherence.
On this account the strangeness of quantum mechanics is not that reality is fuzzy, or depends on observers, or cannot be known. It is that there is more of it than we see. Every apparent paradox, from the double slit to entanglement, is a local, deterministic process in a larger reality.
The beam splitter
Both paths open. Fire photons to see where they arrive.
Deutsch puts the case most sharply in terms of a computer. A large enough quantum computer running Peter Shor's algorithm could find the factors of a number hundreds of digits long, a task that would take any ordinary computer longer than the age of the universe. The calculation would need more steps than there are atoms in the visible universe. Where, Deutsch asks, was the number factorised? Anyone who denies the multiverse owes an explanation of how the work was done.
Where the view is contested
Most physicists use quantum mechanics without committing to any interpretation, and among those who do commit, Everett's has strong rivals: pilot waves, objective collapse, relational and belief-based readings among them. Two objections are serious. First, probability: if every outcome happens, what can it mean to say that one is more likely than another, and why should the likelihoods follow the rule that experiments confirm? Everettians have offered derivations from the theory of rational decisions, a programme Deutsch began in 1999; critics doubt that they avoid assuming what they set out to prove. Second, if every interpretation predicts the same observations, the choice seems to rest on taste. Deutsch's reply is that interpretations are explanations, and that the same predictions do not make the same explanation. This page presents the matter as live, unsettled physics, which is how it stands.
5. Reality is computable
A flight simulator is a virtual-reality generator: a machine that gives its user the experience of an environment that is not there. Deutsch asks how far the idea goes, and his answer is: all the way. In The Fabric of Reality he argues for what he calls the Turing principle, that it is possible to build a single virtual-reality generator whose repertoire includes every physically possible environment.
This is a strong claim about physics. It says that the laws of nature allow their own simulation, and that a computer is not a special kind of object but an instance of something the universe can already do. It ties two strands together. The theory of computation and quantum physics are not separate subjects, because computation is a physical process, and what can be computed depends on the laws of physics. Deutsch's own contribution, in 1985, was to describe a universal quantum computer, a machine that could simulate any physical system.
It also gives the oldest sceptical worry a new form. If a machine can render any environment, how do you know you are not in one? Deutsch's answer comes from section 2. A rendered world that kicks back is evidence of the machine that renders it, and that machine still obeys the laws of physics. Being inside a simulation would not make reality unknowable. It would mean that reality includes the simulator, and our theories would still be theories about the world.
And it explains why we can understand the world at all. Our brains are themselves rendering machines. We never perceive the world directly; we perceive a model of it, built by the brain from sense data and corrected by them. Those models can be accurate because the same laws hold for the modeller and for what is modelled. The mind site takes that thought further.
6. Abstractions are real
Numbers, proofs, the laws of physics, moral truths and the knowledge in a gene are not made of matter. Are they real?
Deutsch says yes, on the same criterion as everything else. In The Beginning of Infinity he borrows a thought experiment from Douglas Hofstadter. Build a chain of dominoes wired up as a computer and set it to test whether 641 is a prime number. A particular domino will fall only if the answer is no. Someone watches it stay standing and asks why. One answer traces the forces through every domino in the chain. The true explanation is shorter and lies at another level: it stayed up because 641 is prime. Primality is not made of dominoes, but you cannot explain what the dominoes did without it. So it is real.
The domino computer
Enter a number and run the dominoes. The last domino falls only if the number has a divisor.
The same reasoning grounds the reality of knowledge. A theory in a textbook, a gene for an eye and a program in a computer are all physical, but what makes each of them work is abstract: information that would do the same job in any medium. Constructor theory, which Deutsch develops with Chiara Marletto, makes this formal. It recasts physics as statements about which transformations are possible and which are impossible, and so gives information, knowledge and life a place in the fundamental laws instead of leaving them as afterthoughts.
Abstractions do not live in a separate realm, as Plato thought, and we do not glimpse them by a special faculty. They are real because they make a difference to what physical things do. Nor can they be reduced to physical things. Mathematical truth is independent of us, but our knowledge of it comes, like all knowledge, from conjecture and criticism, through proofs that are themselves physical processes, carried out by brains and computers that obey the laws of physics.
7. The fabric of reality
The five sites in this series are five views of a single structure. In The Fabric of Reality Deutsch names four strands of explanation, each needed to make sense of the others: the theory of knowledge, quantum physics, the theory of computation and the theory of evolution.
Epistemology. The knowledge site: we know nothing for certain, and knowledge grows by conjecture and criticism. On this site it supplies the rule that what exists is what our best explanations say exists.
Quantum physics. This site: reality is a multiverse, and its strangeness lies in its size, not its fuzziness. It also gives the time site its snapshots. Each moment of the multiverse is a vast collection of them, ordered by the records each one holds of the others.
Computation. The mind site: a mind is a program running on a brain. The Turing principle explains why such a program can model the world at all, and the quantum theory of computation turns the multiverse into a working resource.
Evolution. Knowledge grows in genes by variation and selection, and in ideas by conjecture and criticism. The free will site's freedom is what that second process looks like from inside: choices that depend on explanations not yet created, which even a complete physics could not predict in advance.
No strand stands alone. Epistemology needs physics to say what knowledge is. Physics needs epistemology to say what a theory is for. Computation needs physics to say what can be computed, and evolution needs computation to say what a replicator does. Deutsch's claim is that together they make the first theory of everything worth the name: not a final equation, but a way of understanding why the world contains understanding.
In short
- What exists is what our best explanations say exists. If it kicks back, it is real.
- Instrumentalism, the view that science only predicts, gives up on the main thing science does.
- Single-photon interference shows that something we cannot see affects what we can: shadow photons.
- The shadows make up a multiverse, of which our universe is one layer. Nothing collapses; every outcome happens.
- The laws of physics allow one machine to render every physically possible environment, which is why the world can be understood.
- Emergent and abstract things, from statues to prime numbers, are as real as atoms, because they explain what atoms do.
- Knowledge, quantum physics, computation and evolution are four strands of one explanation of the world.
Sources
- David Deutsch, The Fabric of Reality (Allen Lane, 1997), chapters 1 to 6.
- David Deutsch, The Beginning of Infinity (Allen Lane, 2011), chapters 5, 11 and 12.
- David Deutsch, "Quantum theory, the Church–Turing principle and the universal quantum computer", Proceedings of the Royal Society A (1985).
- David Deutsch, "Quantum theory of probability and decisions", Proceedings of the Royal Society A (1999).
- Hugh Everett III, "'Relative state' formulation of quantum mechanics", Reviews of Modern Physics (1957).
- David Wallace, The Emergent Multiverse (Oxford, 2012).
- Chiara Marletto, The Science of Can and Can't (Allen Lane, 2021).
- Douglas Hofstadter, I Am a Strange Loop (2007), for the domino computer.
- James Boswell, The Life of Samuel Johnson (1791), for the stone.