Re: Are gauge theories possibly approximations?

Thomas Larsson wrote:

> May I remind you that the successful theories in physics, namely
> general relativity and the standard model in 4D, are field theories.
> You don't expect worldsheets in a field theory, do you?
> 
> There is a simple reason why natural theories involve fields rather
> than strings. Namely, the irreducible modules of the correct symmetry
> principles (diffeomorphisms and gauge transformations) are fields.
> E.g., tensor fields transform irreducibly under diffeomorphisms, so they
> are nice and natural objects whose relevance to general relativity is
> hardly in doubt. String fields exist classically but are extremely
> reducible and hence ugly; basically, every point on the string
> transforms independently from its neighbors.

How about reducible, but not decomposable reps? Are they necessarily 
fields? Or how about nonlinear realizations (there's no reason for the 
operator algebra to be generated by a linear rep, is there?)? Are you 
absolutely sure locality is a necessary consequence of diffeomorphism 
covariance?

Why are you biased against reducible reps, like string fields?