My last-but-one post reminded me of Eugene Koonin’s invocation of the multiverse to explain unlikely events such as the development of DNA replication. In my first post on this I hinted at some absurdities inherent in this idea, in that one ought to expect far more instances of unlikely events than we see if all things are possible in an infinite many-worlds multiverse. Blow the irreducibly complex biology – where are the unicorns and spontaneous transmutation of lead to gold?
Nevertheless, one might conclude that a more mainstream view of the multiverse (if “mainstream” has any meaning in gauging pure speculation) could still help explain highly contingent events. After all, if the weak anthropic principle can be invoked to explain fine tuning, why not the low probabilities emerging as problems in evolutionary events?
This, however, is actually a completely different case. Koonin’s multiverse is, in a loose sense, causally linked to our own. Each quantum event, it seems, generates all the possible outcomes as new universes. So given initial conditions of a pre-DNA replicating situation, infinite numbers of possibilities can be tried and the solution found (in our Universe and a small infinity of others). In other words, every individual universe is part of the same lottery. And since there are infinite numbers of universes (with the caveat I pointed out in my earlier article that there wouldn’t be, actually), someone is bound to win.
Now it’s true that in the more usual iterations of Multiverse theory, our Cosmos is indeed part of the same lottery to the extent that it is budded, or spawned or vomited, with a random set of physical constants. Given a large enough number, one like ours might well win in the fine-tuning stakes and be predisposed to develop life. The anthropic principle explains why we observe it.
But once the Universe is formed, it is essentially cut off from the rest of the system (barring the odd proposed escape of small amounts of energy). Its constants will explain why the laws exist as they do, but not the actualisation of contingent events. If one cannot explain why DNA replication is an inevitable outworking of the Big Bang (and it’s only because we can’t that Koonin needs to conjure up the multiverse at all), then the existence of infinite numbers of similar universes doesn’t solve the problem one bit.
For an illustration consider an island in a large ocean, blissfully unaware that a host of equally isolated islands also exist. An enterprising businessman initiates a lottery in which one has to correctly guess a 3,000 digit random number. At odds of 10^3,000, with only 200 inhabitants to play, he knows he is on completely safe ground. There are not enough probabilistic resources on the island ever to produce a winner.
Does the fact that, unknown to him, 20,000 other similar islands exist, where similar entrepreneurs happen to set up similar lotteries, make any difference? Not at all. Each island is entirely dependent on its own probabilistic resources, as there is no contact between them. Any competitor who correctly guessed the number would fail to avoid the charge of fraud if he suggested that the hypothetical existence of many other islands made success inevitable somewhere.
Once “many-worlds” multiverses are discountenanced, then as far as probabilistically impossible events are concerned, we’ve got to sort them out on our own.