Polchinski's review of Woit and Smolin

[Severian]

I thought you might be interested in Polchinski's review of Woit's and Smolin's books:

 

http://cosmicvariance.com/2006/12/07/guest-blogger-joe-polchinski-on-the-string-debates/

 

 

Rather predictably, he doesn't like them much, but he brings up some interesting points.

[Martin]

I think it is certainly constructive for (prominent) string theorists to respond publicly to Smolin's book!

 

Was glad to see that Polchinski had spoken up.

 

I think his point where he seems to dismiss the importance of explicit background independence because it is (merely ) a property of the formulation of a theory is questionable and will draw some comment. So we may see Smolin or someone like that replying----as he has in the past at that blog when the discussion concerned his writings.

 

Have to see how it goes. Good idea to start this thread!

[Martin]

Basically, I'm happy to wait and see how Smolin responds (as I expect he will) but you may have some ideas relating background independence you might want to discuss in the meantime.

 

There is a subtle point about how do you distinguish between a physical theory and its math formulation.

 

(Does a physical theory even exist apart from its formulation? )

 

I think it does if you say that two math models are equivalent if they lead to exactly the same predictions.

 

So then the problem with string (that it so far has only formulations which invoke some prior geometry----maybe several alternative geometries, but some prior choice of background) is that it is so far explicitly background dependent.

 

But theorists counter that by stating their confidence that some version may be shown equivalent to some as-yet-unformulated version which can be formulated entirely without fixed geometry. So their hunch is that the theory is essentially background independent. They also mention AdS/CFT correspondence, involving some conjecture, and a limited form of background independence where IIRC only assymptotic geometry needs to be specified.

 

Smolin (whose book is not all that anti-string) examines these claims in his book. He already has, in effect, answered Polchinski on that point, so if he replies he would just need to cite pages.

 

His point is that even if a background DEpendendent version of string were to resolve the landscape predicament and make testable predictions and successfully survive some tests etc it would STILL NOT BE A FUNDAMENTAL THEORY. It would still just be an effective theory and people would still need to develop a fundamental background independent QG---that is a fundamental theory of what spacetime is and how it interacts with matter.

 

It may sound to some like a finicky point, and it is not the only one in Smolin's book of course , but I think it is a major issue.

 

Basically Smolin is saying that Gen Rel tells us that space is dynamic, not rigid and prior-determined, so any theory that requires for its formulation a prior-determined geometry is fundamentally unrealistic!

 

Anyway---I just want to spotlight that issue. Polchinski brought it up already in his paragraph 6 or 7, early on. And AFAICS he did not say anything that Smolin had not already answered in his book. So this is a major bone of stubborn contention and we should be on the lookout for more comment about it.

 

Whoopee!

 

thanks again for starting this thread, Severian. good idea