262 The Quantum Structure of Space and Time
universe continues to expand at an accelerating rate, these black holes represent an
insignificant fraction of the volume. So, for any patch of space, no matter the value
of Xother, 4 relaxes down to and spends most of cosmic time at the minimum with
the smallest positive Rtotal.
Abbott's notion was to apply this mechanism in a standard big bang universe
where it fails utterly. In the long time it takes to tunnel, the universe expands
so much that it is completely vacuous by the time 4 begins to tunnels down the
potential. However, the concept dovetails perfectly with the cyclic model.
If the axion and the gauge fields to which it couples live on the hidden brane,
their evolution is independent of the cycling motion of the branes along the extra
dimension. Also, because the temperature generated at each big crunchlbig bang
transition is much less than f [4], the axion is not excited by the periodic reheating
of the universe. Therefore, the evolution of the axion and the cycling are completely
decoupled. The result is a cosmology with two inherent and disparate time scales:
the time for a cycle (about a trillion years) and the tunneling time for the axion
(about 1010'20 years in the final stages). The universe spends exponentially manycycles at each step down the washboard potential, with increasingly many cycles
as V, approaches zero. Once a region tunnels to a minimum with negative energy
density, the cycling becomes unstable [3] and, with ten billion years, the region
collapses into a black hole. Meanwhile, most of the universe continues to cycle.
Although different patches of space work their way downhill at different times,
the patches are uniform on scales large compared to the Hubble horizon due to the
smoothing caused during each cycle. Furthermore, every patch is locally equiva-
lent: if one measures the Hubble constant to be near GeV and the dark energy
density to be near (1 meV)4 anywhere an the universe, the physical conditions and
astronomical scene should be similar to what we see today. The situation is the
opposite of the scene suggested by anthropic/landscape scenario in which most of
the universe never looks like what we see within our horizon and is forever inhos-
pitable to the formation of galaxies, stars, planets and life. All other things being
equal, a theory that predicts that life can exist almost everywhere is overwhelmingly
preferred by Bayesian analysis (or common sense) over a theory that predicts it can
exist almost nowhere.
Remarkably, our modest proposal is subject to experimental refutation. Because
it incorporates the cyclic picture for generating perturbations, it shares the cyclic
predictions of a purely Gaussian spectrum of energy density perturbations anda gravitational wave spectrum with amplitude too tiny to produce a measurable
B-mode polarization [4]. There may be other cosmological conundra that can beresolved by having a relaxation time much longer than a Hubble time, and they may
lead to further cosmological tests. At this point, these ideas are new and formative,
perhaps to be developed and debated at a future Solvay meeting.