Just finished reading a fascinating paper by Lee Smolin posted last night on the arXiv:
The case for background independence: hep-th/0507235.
While the paper is occassionally too jargon-laden for us non-quantum gravity researchers to follow, most of it is a very clearly written review of how our view of the “background” impacts the way we do physics. It’s really quite a journey, starting from the classic arguments between Newton and Leibniz about absolute vs. relational notions of space and time. Space and time were just early forms of “the background,” which is anything static and unchanging in your theory. We’ve moved from space and time separately, to spacetime, to fixing the dimensionality and topology of space, and so on. Field theories not only live in a fixed spacetime background, but also have conserved charges related to symmetries on a background. Fixed masses and coupling constants are all backgrounds in the Standard Model.
Smolin’s major point is that moving from background dependent to background independent theories is going to be necessary to address the problems of quantum gravity AND will lead to more predictive theories. You need only wander into one of the heated discussions on string theory that physicists frequently have to see that what we could really use is some falsifiable predictions paired with some data.
The philosophy of science aspect of this paper is a nice change from relentless pummelling of your subject with complex math you usually find on the arXiv. Not that I am complaining about math, of course. It serves as a nice anchor to prevent the hot air in your philosophies from carrying you off to crackpot-land.
One unexpected gem in the paper was the discussion of how the notion of background-independence relates to natural selection. Smolin even lays out a (very sketchy) description of what natural selection might look like when applied to cosmology. In his words (from page 35 of the paper):
- The space of parameters is the space of parameters of the standard models of physics
and cosmology. This is the analogue of phenotype. At a deeper level, this is to be
explained by a space analogous to genotypes such as the space of possible string
theories. This leads to the term the string theory landscape.
- The mechanism of reproduction is the formation of black holes. It is long conjec-
tured that black hole singularities bounce, leading to the formation of new universes
through new big bangs. There is increasing evidence that this is true in loop quan-
tum gravity.
- We may conjecture that the low energy parameters do change in such a bounce.
There are a few calculations that support this.
- The mechanism of differentiation is that universes with different parameters will
have different numbers of black holes.
He finishes this discussion by pointing out that this isn’t just idle speculation, but actually can lead to testable predictions. One prediction which he mentions is that no neutron star can have a mass bigger than 1.6 times the mass of our sun, and the largest neutron star observed so far is 1.45 solar masses. How exactly this prediction is made is a mystery to me, but he does cite another paper of his, hep-th/0407213, for more details.
(Disclaimer: This is far from an accepted theory, but is just an interesting idea that should be investigated.)
Overall, there were too many big ideas in this paper for me to absorb in one reading. It will take a couple passes to get it, I think. 
