Explaining homogeneity and structure 149When the full range of inhomogeneities and anisotropies is taken into account
(e.g. [128]), it appears thatboth approaches are partially successful: with or
without inflation one can explain a considerable degree of isotropization and
homogenization of the physical universe (see e.g. [127]), but this will not work in
all circumstances [105,106]. It can only be guaranteed to work if initial conditions
are somewhat restricted—so in order for the programme to succeed, we have to go
back to the former issue of somehow explaining why it is probable for a restricted
set of initial data to occur.
3.9.1.2 Initial conditions are irrelevant because they never happened
Some attempts involve avoiding a true beginning by going back to some form
of eternal or cyclic state, so that the universe existed forever. Initial conditions
are pushed back into the infinite past, and thus were never set. Examples are as
follows.
- The originalsteady state universeproposal of Bondi [6], and its updated
form as thequasi-steady state universeof Hoyle, Burbidge and Narlikar
[71, 72]. - Linde’s eternal chaotic inflation, where ever-forming new bubbles of
expansion arising within old ones exist forever; this can prevent the universe
from ever entering the quantum gravity regime [90]. - The Hartle–Hawking‘no-boundary’ proposal(cf [67]) avoids the initial
singularity through a change of spacetime signature at very early times,
thereby entering a positive-definite (‘space–space’) regime where the
singularity theorems do not apply (the physical singularity of the big bang
gets replaced by the coordinate singularity at the south pole of a sphere).
There is no singularity and no boundary, and so there are no boundary
conditions. This gets round the issue of a creation event in an ingenious
way: there is no unique start to the universe, but there is a beginning of time. - The Hawking–Turokinitial instanton proposalis a variant of this scenario,
where there is a weak singularity to start with, and one is then able to enter a
low-density inflationary phase. - Gott and Liu’scausality violation in the early universedoes the same kind of
thing in a different way: causality violation takes place in the early universe,
enabling the universe to ‘create itself’ [63]. Like the chaotic inflation picture,
new expanding universe bubbles are forming all the time; but one of them is
the universe region where the bubble was formed, this being possible because
closed timelike lines are allowed, so ‘the universe is its own mother’. This
region of closed timelike lines is separated from the later causally regular
regions by a Cauchy horizon.
There are thus a variety of ingenious and intriguing options which, in a sense,
allow avoidance of setting initial conditions. But this is really a technicality: the
issue still arises as to why in each case one particular initial state ‘existed’ or