Sketch of an argument against the many worlds interpretation

— Essentially, as I understand it, the many worlds interpretation seeks to reduce probability distributions in observation to probability distributions among parallel universes.

— So, if quantum theory predicts, and experiments observe, that some quantum event occurs with 25% probability, then many worlds explains this as saying that in 25% of universes it occurs, and in the other 75% it does not.

— Now, first of all, the number of universes predicted by many worlds is unimaginably vast, if not actually infinite. Thus, it would seem, that almost every conceivable universe does by many worlds actually exist.

— Many worlds does place some constraints on the set of actual universes — it is not the same as modal realism or ultimate ensemble theory, in that there are actual physical laws which all universes must obey. But, let us distinguish two different types of potential parallel universes:

o those which a physicist or mathematician might be interested in, say those with radically different physical laws from our own

o those of maybe more quotidian interest – those which in broad concerns, such as physical laws, are largely identical to our own, and yet in numerous specific details are different; say, in which your favourite TV show is actually real, or in which that special someone you had a crush on in high school ends up being the love of your life, rather than someone you never ever see again

— Clearly, many worlds rules out at least some of the first group. It may well be the case that, even by many worlds, there is no universe supporting sentient life having at a macroscopic level a 174-dimensional spacetime. And yet, it seems unlikely that it rules out every, or even most or all, possible universes in which observed physical laws are identical to our own, but nonetheless certain events which in this universe some wish would occur and yet do not, actually do.

— More to the point, however; it seems there must be at least some universes in which, even if purely by the remotest chance, intelligent life exists despite radically different laws of physics being observed. For any finite probability, however small, if there are a sufficiently large number of universes, then the probability that at least one of them contains the desired state of affairs is as close to certain as we would like; and if there be an infinity of universes, almost surely such a universe exists.

— Now, the underlying assumption behind the many worlds theory is that the observed probability distribution within this universe can be extrapolated to the distribution of universes within the ensemble. But let us suppose, we are in an extremely unlikely yet nonetheless actual universe — we may observe within our own universe an extremely rare probability distribution, and thus by this assumption we would assume that this distribution is the norm for the universe ensemble.

— It might be objected that the aforementioned scenario is extremely unlikely; and yet, to so object is to engage in circular argumentation: for our judgement that it is unlikely is dependent on our judgement that it is false.

— And, going back to the point about many worlds, having certain pan-multiverse physical laws, thereby excluding certain mathematically / logically possible universes from those it claims to exist — it would seem, even if no such universes are actual, that there would be universes in which rational agents would be justified in believing they were in such a universe. But, by the same argument as earlier, that we can have no knowledge of the distribution of universes, thereby we have no reason to believe we are not in such a universe, thus we have no reason to believe in any pan-multiverse laws.

— What I am essentially trying to prove here, is that many worlds, as distinguished from modal realism, is self-defeating. Taking into account the existence of observers, and that if many worlds is true, then we are one of those observers (and, our belief in many worlds is itself subject to many worlds), it follows that we have no reason to believe in any version of many-worlds which is non-identical to Lewisian modal realism, or Tegmark’s ultimate ensemble theory.

— Thus it follows that many worlds is not a scientific theory, since modal realism certainly is not a scientific theory.