A difficulty for Everett's many‐worlds theory

Abstract

An argument originated by Brandon Carter presents humankind's imminent extinction as likelier than we should otherwise have judged. We ought to be reluctant to think ourselves among the earliest 0.01 %, for instance, of all humans who will ever have lived; yet we should be in that tiny group if the human race survived long, even at just its present size. While such reasoning attracts many criticisms, perhaps the only grave one is that indeterminism means there is not yet any firm, theoretically predictable fact of how long the human race will survive. This, though, might not save Everett's many‐worlds theory from a variant on Carter's point. Everett seemingly pictures observers as splitting into ever more versions, which explains away all apparent indeterminism; but then, absurdly, all except a vanishingly tiny proportion of one's versions would come into existence near one's death, outweighing all apparent evidence that death was not imminent. The need to avoid such a result severely constrains Everett‐type theories. We need theories in which observer‐versions diverge without increasing their numbers in any straightforward way.