Quote:
Originally Posted by Cougar
The argument goes something like this: All of the "pocket universes" have different constants, physical laws, and number of dimensions. Only certain configurations will conceivably allow for "life" or "observers". But there is a range of values within each configuration where life could emerge. If by observing our own values we find that our pocket universe is much, much more bio-friendly than it "needs to be" for life to emerge, then... something's fishy. The multiverse idea predicts that our values for the vital parameters that affect life should be fairly close to the "edge," beyond which the pocket universe would not be bio-friendly. So if we find our pocket universe is a "million miles" from that edge on the side of bio-friendliness, then the multiverse idea would be falsified. Therefore, it's (sort of) scientific.
Well, let's see how that goes over....  |
Like a lead balloon! Thanks for summarizing that argument, but if that's all they've got, it's not gonna fly. It has as many holes as a sieve, including the fact that we know so little about (1) what ranges in parameter space can lead to intelligent life, (2) what the imaginary distribution over those ranges are (Vilenkin apparently assumes the distribution is a steep power law of some kind, so we should be bumped into one edge, but how can we know that-- if it turns out we're not, they just pick a new distribution and are back in business-- no falsifiability that way), and (3) the validity of making the usual "Carter catastrophe hypothesis" that we are somehow "generic" intelligences, even though we have just now come up with the idea of multiverses (which may make us special-- what if there are parameter regimes where civilizations who come up with the multiverse idea are near the end of their viability, whereas other parameters have civilizations that live way past this concept, would our parameters not be more likely in the former group even if the latter group is otherwise more numerous?). Those uncertainties make it a can of worms.
But here's the real kicker. The argument presented in no way distinguishes a "landscape" approach to the parameters (where all the parameters are actualized somewhere in the multiverse) from any other algorithm you could imagine for how the universe got its parameters. Say, for the sake of argument, that the universal "parameter choice algorithm" is that it makes one single universe by randomly sampling over the parameter range, but if the result is nonviable it just throws it out and tries again (remember, the parameter distribution process is outside of any physics we know, so we have no way to rule out the possibility that it can interrogate the entire future of the universe and ask if intelligent life appears).
Note this approach is very similar to the core idea of the multiverse, that there is a purely statistical algorithm of some kind for choosing the fundamental constants, but it simply has no need to assume that more than one version is ever actualized. The statistics are identical, so all the Vilenkin-like arguments apply equally well, but there's no "multi" in the "verse", it's still a "uni". How would one ever distinguish these models scientifically? It would purely come down to how you imagine the algorithm, does it need to create a universe for every parameter to get life in some of them, or can it know about the future as part of the algorithm? That's the trouble with leaving the testable realm to create your fantasies-- you cannot constrain them enough to make them anything but pure guesswork and philosophical prejudice. You end up looking in the mirror and saying, "ah yes, I see".