There's an interesting new research article out that is getting some buzz in the popular press. It is an open access article if you register (for free) with the Royal Society of Chemistry. I'm withholding judgement at present, but it does seem to merit further investigation.
The article looks at the autooxidation mechanism that is commonly used to describe the (photo)oxidation of polymers and why they breakdown when weathered. It's a free-radical propagation scheme not too unlike a polymerization reaction.
I certainly had no idea that the mechanism was first proposed 65 years, and on that basis alone, it probably would be a good idea to re-examine it. I also was unaware that the mechanism had been used to describing degradation of materials that a chemically distant from olefins (the materials that I am most familiar with in a degradation setting), to even include acrylates. 
Using computer modeling , the researchers found that the existing mechanism is thermodynamically unfavorable in most cases, and "in most cases" here covers an impressively wide range of chemistries. They also found that temperature doesn't change influence at all. Instead, the authors look at defects that occur in polymerization and found that the defective sites are sites where the thermodynamics are favorable for hydrogen abstraction. They then suggest that examining materials made with non-conventional and more constrained polymerization mechanisms may avoid these problems.
As I started with, I find this a very intriguing paper and one that needs more research. There is a tremendous amount of existing research in degradation that was not mentioned here, and it would be invaluable to look at it anew to see if the new mechanisms are supported by it.
I am greatly concerned that the influence of temperature was minimal. Whole books, journals and companies exist because of accelerated weathering, and temperature always plays a key role in it. (If reaction (1) above is the result of a photoinitiation, then that step is pretty much temperature independent, but all the remaining steps certainly have a temperature dependency.) Nonetheless, I still found the whole article worthwhile to read and hope that it leads to new insights quickly. I'd be curious if it leads to the development of new antioxidant chemistries to fight these "new" reactions.
 I'm glad that the researchers had the same thought I did in reading that: the proposed scheme won't work because it is well known that oxygen inhibits acrylate polymerization. The peroxy radical in step 4 can't abstract a hydrogen. So either the mechanism above is correct, or our understanding of oxygen inhibition is correct, but they both can't be correct.
 Gas phase, and some liquid phase. Ever seen a gaseous polymer? Think it might behave just a little differently in a solid? Unfortunately, even with supercomputers, modeling is still challenging and gas phase analysis is just easier to perform.