Many-World Interpretation

An Argument against the Many-World Interpretations#

I've always hated these sorts of interpretations (along with all the possible-world things which sound silly enough), and since I hated them a lot I didn't even bother to argue against them, now Tim Maudin in his Quantum Non-locality and Relativity argues:

The many-worlds theory is incoherent for reasons which have been often pointed out: since there are no frequencies in the theory there is nothing for the numerical predictions of quantum theory to mean. This fact is often disguised by the choice of fortuitous examples. A typical Schrödinger-cat apparatus is designed to yield a 50 percent probability for each of two results, so the “splitting” of the universe in two seems to correspond to the probabilities. But the device could equally be designed to yield a 99 percent probability of one result and 1 percent probability of the other. Again the world “splits” in two; wherein lies the difference between this case and the last?
Defenders of the theory sometimes try to alleviate this difficulty by demonstrating that in the long run (in the limit as one repeats experiments an infinite number of times) the quantum probability assigned to branches in which the observed frequencies match the quantum predictions approaches unity. But this is a manifest petitio principii. If the connection between frequency and quantum “probability” has not already been made, the fact that the assigned “probability” approaches unity cannot be interpreted as approach to certainty of an outcome. All of the branches in which the observed frequency diverges from the quantum predictions still exist, indeed they are certain to exist. It is not highly likely that I will experience one of the frequencies rather than another, it is rather certain that for each possible frequency some descendants of me (descendants through world-splitting) will see it. And in no sense will “more” of my descendants see the right frequency rather than the wrong one: just the opposite is true. So approach of some number to unity cannot help unless the number already has the right interpretation. It is also hard to see how such limiting cases help us: we never get to one since we always live in the short run. If the short-run case can be solved, the theorems about limits are unnecessary; if they can’t be then the theorems are irrelevant.

A remark should be made: the deficiency lies actually in the frequentist interpretation of probability; many-mind theories do not have the same deficiency.

Also a joke: maybe in the 1 versus 99 example, there should be 1 world corresponding to 1 percent and 99 worlds corresponding to 99 percent.