How/why are pocket universes created during eternal inflation?

[Chronos]

If we found observational evidence of another 'universe', would it prove the multiverse hypothesis, or alter our perception of the universe? Current evidence suggests the universe has a dark sector. Does that qualify as a 'parallel' universe, or is it just another aspect of our universe? I would argue the goal of science is to define and explain the universe in terms of all possible observations. That which has no observational consequences is not science. As in much of science, semantics are a distraction and frequent source of confusion.

 

[Chalnoth]

If we found observational evidence of another 'universe', would it prove the multiverse hypothesis, or alter our perception of the universe? Current evidence suggests the universe has a dark sector. Does that qualify as a 'parallel' universe, or is it just another aspect of our universe? I would argue the goal of science is to define and explain the universe in terms of all possible observations. That which has no observational consequences is not science. As in much of science, semantics are a distraction and frequent source of confusion.

Claiming that there are no observational consequences is just a lack of imagination.

 

[Chronos]

I agree, imagination is not required to do science.

 

[Chalnoth]

I agree, imagination is not required to do science.

??

Imagination is central to science. It is incredibly important for coming up with new theories, and coming up with clever ways of testing them.

 

[Torbjorn_L]

That it is possible to propose a test of the multiverse hypothesis is unrelated to the possibility may, in fact, be untestable. Furthermore, no such 'test' has yielded observational support. The above referenced paper, http://arxiv.org/abs/1012.1995, is no exception. In the abstract the authors note "... We conclude that the WMAP 7-year data do not warrant augmenting LCDM with bubble collisions... ". You need not concede a negative result constitutes falsification, but, you must at least be willing to concede the hypothesis may not be testable. For further discussion, see see Peter Woit's article

Woit is a known crackpot on this subject.

I said it elsewhere on Physics Forums, I don't do philosophy such as "falsification", I'm solely interested in science. What we want according to measurement theory is hypothesis testing (for observations and theories both).

I propose a hypothesis test, and the constraints go with the data into the test. It may be that some variants of the test isn't productive (we can't see bubble collisions because inflationary expansion happens to be too fast), but that is a problem of _that_ theory, not the theory with the parameters that are testable. If hypothesis testing can distinguish them based on used parameters (that go into the test constraints) they are different theories under testing.

That was the start of this thread, the question if multiverses are untestable under all conceivable circumstances using physics constraints. It is not (and neither is Woit's nemesis of string theory). And yes, I think I know what "conceivable" means, more than philosophers (and crackpots fancying philosopher's toys over scientist's tools) know what "theory" and "testing" means.