The Many-Worlds Interpretation

Hugh Everett's proposal that the wavefunction never collapses — and that every quantum measurement splits the universe into branches that no longer speak to each other.

Quantum mechanics is, by some measures, the most successful theory in the history of physics. Its predictions match experiment to twelve decimal places. It explains the periodic table, the colour of the sky, the stability of matter, and the inside of a laser pointer. It also has, sitting quietly in its foundations, a problem that has not gone away in a hundred years: what happens when you measure something?

The orthodox answer — taught in textbooks and called the Copenhagen interpretation — is that the smooth quantum wave somehow snaps, instantaneously, into one definite outcome. A version of reality is selected; the others vanish. This works as a recipe. As a description of the world it is strange: nothing else in physics behaves this way. In 1957 a Princeton graduate student named Hugh Everett III proposed, in his doctoral thesis, that the recipe was wrong. There is no collapse. The other outcomes do not vanish. They simply happen, somewhere else, in another branch of the wavefunction.

The measurement problem

Quantum theory has two rules. The first, the Schrödinger equation, describes how the wavefunction of a system evolves in time. It is smooth, deterministic, and reversible — the kind of equation a 19th-century physicist would have been comfortable with. The second rule is the one that breaks the pattern. When a measurement is made, the wavefunction does not evolve smoothly; it “collapses” instantly into one of its possible outcomes, with a probability given by the Born rule.

This is awkward. What counts as a measurement? Is a photographic plate a measurement? A Geiger counter? A cat? A conscious observer? The theory does not say. It draws a fuzzy line between “system” and “apparatus” and asks you not to look too closely at where the line lives. Schrödinger, who did not believe his own equation needed to be patched this way, dramatized the absurdity with a cat in a sealed box, simultaneously alive and dead until someone peeked.

If the Schrödinger equation is universal, it must apply to the observer too. And then there is no collapse — only entanglement.

Everett's move: take the equation seriously

Everett's idea was almost embarrassingly simple. The Schrödinger equation is, on its face, a universal law. There is no special clause that switches it off when a person walks into the room. So suppose — just suppose — that it really is the only rule. There is no second rule. There is no collapse. What follows?

What follows is that the observer, made of atoms, is also a quantum system. When she measures an electron whose spin can be up or down, she becomes entangled with it. The combined state of (electron + observer) is no longer “spin up or spin down.” It is “spin up and observer-who-saw-up” superposed with “spin down and observer-who-saw-down.” Both terms persist in the wavefunction. Both are equally real. Each observer, inside her own branch, sees a single definite outcome and has no access to the other.

There is, on this picture, no moment at which the universe decides. The universe simply branches.

Decoherence: why we don't notice the other branches

An obvious objection: if all the branches are real, why don't we see them? Why does the world look classical and singular, not a smeared-out cloud of possibilities?

The modern answer, worked out by Zeh, Zurek, Joos and others from the 1970s onwards, is decoherence. A microscopic quantum system — a single electron, an isolated atom — can stay in a clean superposition because nothing else in the universe is keeping track of which branch it's in. As soon as the system interacts with anything macroscopic — a detector, the air, a photon scattering off a measuring device — the information about its state leaks irreversibly into a vast number of environmental degrees of freedom.

Once that happens, the different branches of the wavefunction stop being able to interfere with one another. They are still all present in the global state, but they have become, for all practical purposes, sealed off. An observer in the “up” branch cannot, even in principle, gather the entangled environmental record together cleanly enough to recombine it with the “down” branch. The branches diverge and never meet again.

Strange, but maybe the cheapest theory we have

Many-worlds sounds, at first hearing, extravagant. Trillions of unseen universes, for every coin flip an electron makes? But what is being multiplied is universes, not laws. The theory contains exactly one equation — the Schrödinger equation — and applies it without exception. Copenhagen, by contrast, needs the Schrödinger equation and a separate collapse postulate and a fuzzy notion of what counts as measurement. By the standards of Ockham's razor, many-worlds is the lean theory; it is reality that is being told to play along.

The view has its costs. The hardest is the probability problem: if all outcomes happen, what does it mean to say one is “more probable” than another? Why does the Born rule — probability equals the square of the amplitude — come out right? Decision-theoretic arguments by David Deutsch and David Wallace try to recover the Born rule from the rational behaviour of an observer who knows she will branch, but the question is still live.

Whatever its eventual fate, many-worlds is the cleanest existing answer to the question Schrödinger's cat made vivid. It takes the equation seriously and follows it wherever it leads. The price, if you accept it, is a particular kind of vertigo: the version of you reading this sentence is one branch among uncountably many. Each of them, equally real, equally surprised to find itself here.