Emergent Spacetime^179
5.2 Discussion
S. Shenker I would like to speak in defense of redundancy. It is certainly true
as you say that gauge symmetry really does reflect just a redundancy in the
description. But sometimes, as we know, this redundancy is very useful. For-
mulating quantum electrodynamics with four gauge fields is good, because you
can choose different ways of resolving the redundancy to make things like uni-
tarity, or Lorentz covariance, clear in different gauges. And so, well, some of us
have been worrying for a long time, and Gerard ’t Hooft made these ideas quite
explicit, perhaps what we should be thinking about is a description where we
vastly enlarge the number of degrees of freedom in our description of whatever
the thing we’re trying to describe is, quantum gravity. And so we would then
be able to resolve this redundancy in a way in which locality, to the extent it
exists, is manifest, or holography is manifest in another gauge, or some other
property is manifest. And then it would be too much to ask for a description in
which all the desirable properties of the theory are clear, in one presentation.
N. Seiberg I would like to respond to that. I sympathize with your point of view,
but recall that you wrote a paper about the matrix model, where the local
reparametrization on the world sheet was absent. You had a description of two
dimensional gravity, in the sense of the world sheet, without reparametrization
freedom. So this is an example where gravity and general covariance emerge.
S. Shenker What is this statement: to do as I say, not as I do.
J. Harvey There is also saying that consistency is the hobgoblin of small minds.
E. Silverstein I just have a small comment on non-locality appearing in string
theories. So another example occurs in AdS/CFT. You said we don’t have any
argument against non-locality at the string scale, but there is evidence for non-
locality at a much larger scale, if you consider multi-trace deformations of the
field theory. So if you include the internal space, the simplest versions being
spheres or other Einstein spaces, the boundary conditions are grossly non-local
on those dimensions, at a scale of order the Ads radius scale. So this is another
indication that we might need to incorporate non-locality in a serious way.
S. Weinberg I have a comment and a question. The comment is that for a long
time, I thought that general covariance was a red herring, because any particle
with mass zero and spin two would have the properties that we derive from
general covariance, as shown also by Feynman, and the great thing done by
string theory, in this area, is to show that there has to be a particle of mass
zero and spin two, while before string theory we didn’t know why there had
to be one. My question had to do with your remark that the S matrix in
string theory is more analytic than in quantum field theory. I thought that in
quantum field theory the S matrix was as analytic as it possibly could be, given
the constraints of unitarity ~ I do not see how anything...
N. Seiberg What I had in mind is the good high energy behaviour. And I think
