many-worlds interpretation
The many-worlds interpretation, proposed by Hugh Everett in 1957, takes the boldest possible step: it deletes collapse altogether. There is just one law, the smooth evolution of the wavefunction by the Schrodinger equation, and it never stops applying — not even during a measurement. Whatever superposition the system was in, the apparatus and the observer simply join it, becoming entangled with each possible outcome.
On this view a measurement does not pick one result and discard the others; it splits the world. Each possible outcome is realised in its own branch, complete with a copy of the observer who sees that result and believes it was the only thing that happened. The branches stop interfering once decoherence has scrambled the phase relations between them, which is why each branch looks like an ordinary, single-outcome classical world from the inside.
The appeal is economy of assumptions: one equation, no mysterious extra collapse rule, no special role for observers. The price is an extravagant ontology of countless unobservable parallel worlds, and a genuine puzzle about probability — if every outcome happens, what does it mean to say one was ninety percent likely? Defenders argue the Born rule can be recovered from rational decision-making within the theory, but whether that argument fully succeeds is still honestly debated.
No collapse: the observer entangles with every outcome, and the branches decohere into separate worlds.
The branches are not extra universes bolted on; they are parts of one big wavefunction. The main open problem is justifying why the Born rule probabilities should apply when, on the face of it, every branch simply happens.