11/26/2007

STENGER INVERTS OCKHAM

There's a nice little kerfuffle going on in the Science Blogs about a piece in the New York Times in which the cosmologist Paul Davies stubs his toe in the philosophy of science. As usual, Wilkins has an excellent analysis, which in this case demonstrates that Davies is guilty of at least two logical errors.


Interestingly enough, Davies is already something of thorn in the side of skeptics for taking the Templeton Foundation's money and endorsing a limited 'fine-tuning' argument in behalf of some vague deism. During the discussion on Pharyngula, a reader brought up an argument by another physicist:

"Victor Stenger at Talk Reason* makes the argument that a single universe runs afoul of Occam's Razor by postulating an arbitrary and unnecessary limit of 1 to the number of universes. That limit itself is the unparsimonious entity. After all, we have the existance of one universe known - more than one isn't multiplying entities, it's repeating a single entity that is known to exist in at least one case..."

To which I say, ha. By extension, the value of 6.67 E -11 for the gravitational constant is also an arbitrary limit. Just because this is among the narrow range of possible values actually observed under experimental conditions shouldn't rule out wildly different values! In other words, I don't think much of Stenger's argument.

After all, we don't posit a limit on the number of entities observed just by the act of observation, and extradordinary claims require extraordinary evidence. The claim that there is one universe known to exist is not an extraordinary claim, and is certainly supported by a very large number of claims. The existence of a multiverse is an extraordinary claim for precisely the opposite reason. It is not enough to bask in the beautiful mathematics of another untested notion, string theory, and observe that some versions demand a multiverse. At some point, we need some actual evidence...right?


Maybe the various detectors on the Large Hadron Collider, in the final stages of construction as we speak, will shed some light on the present paucity of data for both string 'theory' and multiverse scenarios. Right now, their main scientific virtue with respect to origins is not that they 'explain' anything, but that they in principle supply a naturalistic redoubt for cosmogenesis. In the meantime, we do have one universe known to exist and we do not need to posit a theological redoubt in order to ask why it has the properties that it does.
Davies goes too far in equating the axioms of science with the dogma of faith, but that says nothing about the utility of the anthropic principle or the apparent 'fine tuning' of certain parameters. Scientists should be free to propose testable hypotheses drawn from different metaphysical axioms, as long as they test them and let the chips fall where they may.

* I should point out that Stenger's article is a good read, well worth a person's time, and it has a lot more in it than just the argument I reference above. For the record, the link takes one to a PDF file, and one can find Stenger's discussion of parsimony with reference to multiverse scenarios beginning on page 17.

5 comments:

Anonymous said...

HI, I was the original commenter. Glad you found it interesting, amusing enough to post on MT

The existence of a multiverse is an extraordinary claim for precisely the opposite reason. It is not enough to bask in the beautiful mathematics of another untested notion, string theory, and observe that some versions demand a multiverse. At some point, we need some actual evidence...right?

I lost a previous comment, so I'll be a bit briefer.

The evidence points to the existence of this universe. I kick a rock and prove it thus. But it's one thing to say 'this universe exists' and another to say 'ONLY this universe exists'. That's another extraordinary claim, and lacking evidence itself. MU has explanatory power in QM - what does ONLY1U have?

Stenger, I think, is cleverly pointing out this is a bit of a vase/two faces situation. You could express it another way - which has the most information:

1

or

1111...(recurring)?

The answer is (I think - my info theory-fu is rusty and never was strong) - well, that depends if you're talking finite strings or infinite sequences.

The gravitational constant bit is apples and oranges, things and relationship values The value of Pi says nothing about the cardinality of circles.

I'm also certain my little knowledge will prove to be a dangerous thing - but it's a fun brain stretch.

Scott Hatfield . . . . said...

Thanks for your comment, which provided some depth I hadn't previously considered. I agree that MU has explanatory power within QM. You wrote, however:

'what does ONLY1U have?'

Well, that's easy. Parsimony, from one side of the two faces / two vases.

Now, as this post acknowledges, you can invert parsimony as Stenger does. But the question I would ask is: to what end? Parsimony of ONLY1U leads to just one universe with one set of parameters to investigate. Parsimony of MU leads to...what? Perhaps a way to resolve some paradoxes in QM, as beings within this universe attempt to understand same. But it also resolves paradoxes for universes in which everyone is Mike Brady. Or Charles Dickens. Or turtles, all the way down, in universes which don't use gravity at all, but the 'apples and oranges' of one's choice. The multiverse scenarios, as I understand it, predict every sort of possible universe. Whatever you can dream up, there is a universe in which said is true.

Anything that predicts everything really predicts nothing. So it seems to me that the upshot of MU is any actual application is baroquely non-parsimonious. It only appears not to multiply entities if we consider it solely from the standpoint of our own universe. If that's the case, parsimonious MU really depends upon the simplifying case of 1U, which (as a practical matter) reduces to ONLY1U. Or at least the ONLYU we can test.

Anonymous said...

While I take your point I think you're confusing 'imaginable' with 'possible'. I'm willing to bet you could never have a universe of Charles Dickenses because Charles Dickens is not the product of a universe of Charles Dickenses, but of Victorian England. So that's a limitation of prediction there, albeit only removing the illogical and the impossible.

Whether the laws of physics are the same, as in a branching quantum multiverse, or different, as in a quantum foamy bubbly seeding multiverse, or some combination thereof, there will always be some sort of physics involved.

I think what you do gain is perspective. Occam's razor, in itself, is not a sufficiently useful tool to distinguish between
MU and ~MU. But it was never a finely honed weapon at the best of times, it's a rule of thumb, and not of law. To be scientific and pragmatic we can, indeed, only work with what we have - but we have to keep in mind the alternate possibilities and not just dismiss them with appeals to Mr Occam.

So on with the LHR and evidence - wherever it may be!

Scott Hatfield . . . . said...

Right on!

Blake Stacey said...

The subtlety is that we have decent reasons to expect other universes to exist, or at least other regions of our own so far away that we can't observe them (which amounts to the same thing). This is a prediction of standard models of inflationary cosmology, models which are consistent with everything we've observed so far. If you take one of those models as your working hypothesis, then you really do need extra postulates to keep the number of universes down to 1.

I don't expect this question to be resolved for a good many years to come, of course.