64 The Quantum Structure of Space and Time
3.2 Discussion
A. Linde Eternal inflation indeed requires initial conditions and is not future geo-
desically complete. However, one should recall the difference between being
past eternal and past complete. For any point, any geodesic starting in it has
finite distance, however, for any geodesic, there is a longer one. So for a par-
ticular observer, the universe begins and ends, but not the universe as a whole.
This rectifies a common misconception.
G. Gibbons I agree, there is no big bang in that model.
A. Strominger Cosmic censorship proposed a solution to the singularity prob-
lems, namely that they always lie behind horizons. However, Hawking radiation
shows that black holes become smaller over time, and that presumably Planck-
ian effects take over the dynamics at some point in time. So, cosmic censorship
would not anymore seem to make general relativity complete in any way. Why
is it then so close to being true?
G. Gibbons Recall that the cosmic censorship idea came from a practical question
in X-ray cosmology. Did general relativity need to be changed at the scale these
experimenters were interested in? The conclusion they came to was that there
was no obvious reason why general relativity should break down at macroscopic
scales. Still, cosmic censorship is an important property for classical general
relativity, and it is still important to know whether it is satisfied.A. Strominger If it was not, we could see naked singularities in the sky.
G. Gibbons Which was always Penrose’s viewpoint. The counter was put forward
by more conservative opponents.S. Weinberg If only on aesthetic grounds, it seems obvious that at short distance
scales additional terms in the gravitational action will become important. Gen-
eral relativity is an effective theory as is the theory of soft pions. Obviously, it is
not the whole answer. Original singularity theorems are important in showingthat higher order terms will become relevant at some point.
G. Gibbons Yes, however, we distinguish two possibilities: they can become im-
portant already classically, or at the quantum level. The latter is more plausible.
S. Weinberg There is no problem with ghosts and the higher derivative terms.
They arise only when misusing perturbation theory.D. Gross Indeed, there is no problem. But only if you have a sensible theory to
do perturbation theory in, which is not a priori given.