Theory of multiverse - it is really scientific theory? In book Hidden Reality: Parallel Universes and the Deep Laws of the Cosmos Brian Green shows that many current physical theories (e.g. quantum mechanics - many world interpretation of QM, superstring theory or theory of inflation) allow a possibility that our Universe is a part of bigger structure - a multiverse. 
He also added some other theories like a holographical principle or simulated universe (meaning universes simulated on computer) to the list of theories above.
However, the issue is that many of these theories presented in the book predict that it is impossible to verify that the multiverse exist because universes cannot communicate each other.
My question is: Are theories of multiverse really scientific theories if we take into account that their predictions cannot be verified? These theories seems to me to be rather philosophical concept than serious science. Or do I miss something?
 A: It depends on whether a theory will ultimately lead to testable predictions. For many of them that is not obvious but more a matter of opinion, and advocates of different theories tend to disagree about which ones might some day be testable.
As for the multiverse, many different kinds of multiverse have been suggested. Some are probably more testable than others. For example if there are many "brane" universes floating in the multiverse plenum and ours is one such, perhaps even created through a collision between two others, then this might leave signatures in the makeup of our universe that we can detect. In this case the theory is valid science because it is in principle testable. But perhaps no such signatures can be left, in which case the exact same theory turns out to be just metaphysics.
So, in reality-checking the wacky concepts drawn from advanced mathematical constructions, are we doing physics or metaphysics or some other branch of philosophy?
I'd say that we are doing a bit of all three.
