Question Author
I'll try to address your comments in greater detail later, OG, but on the particular point:
//"Instead, it pays to be co-operative,"
No, as an individual it pays to look after you own interests...//
This seems to depend on your definition of what it means for something to "pay". Ultimately what this paper, and the whole topic, is about is trying to express the choice between selfish behaviour and not-selfish in terms of game theory. So the definition of "pay" they are using is from that viewpoint -- so that cooperation "pays" if it turns out to be a better game-theory choice than non-cooperation. While you can argue about using this definition, I don't think you can argue about the result, unless you are able to counter their result on these specific terms.
As usual, the problem is one of language. It's not the easiest paper to digest, and I don't know if I can understand and interpret it any better than the BBC's Science editor can. Probably not. But I'm not sure that the results in the paper can be dismissed so easily.
The other thing to say is that (as I understand it) while evolution changes species and not individuals the features of sets of individuals and differences between them are what drive a species forward. So if one set of individuals in a species can be characterised by being called "selfish", and another set is not, then that difference can have consequences for the evolution of the species. So, if altruism is more evolutionarily preferable than selfishness, then you can loosely say that it does "pay" -- in the sense of being more likely for that individual to reproduce -- to be kind.
I don't think I've expressed myself clearly, but hopefully you get the drift. It's about trying to understand how particular behaviours can emerge in a species by considering their benefits to the individual. After all, isn't it the same for all other traits? Polar bears are white because individual bears with white fur were more able to hunt in the polar regions without being detected, so that it "paid" to be a white polar bear. At least, that's the shorthand version, no? While it's certainly more complicated than that, you can loosely describe what was happening in that way.