1) I am afraid that, in my view, you have rather underestimated the power and Global scope of Mass Extinction Events. - What do mass extinction events have to do with human evolution? No mass extinction events have occurred since hominids first evolved.
And yeah as to the second point, I suppose an evolutionary...type, process might apply to socially constructed behaviours. Obviously this is not a new idea, Richard Dawkins has his memes and what not. Umm...I'm not actually sure what I think of that. I suppose in one sense, there's something inescapably logical, even tautological in the idea that "those behaviours which exist, exist because something about them made them likely to spread and be repeated." At the same time however, in evolution what allows genes or traits to persist in the gene pool is their effect on an organism's ability to survive. The reasons why ideas or behaviours persist often seem much more arbitrary than that. Take for example written language. The reason we use the alphabet we use isn't necessarily because something about it gives us an advantage. Having SOME system of writing or recording information on a permanent or semi-permanent basis is an advantage, but in and of itself the latin alphabet that we use, is no more advantageous than language written with the Cyrillic alphabet, or the greek alphabet, or in Kanji, or hieroglyphs.
If I might be allowed to go on a slight diversion here, there is no evidence that any Andean civilizations ever developed any system of "written language" as in symbols carved into stone or marked on a paper like substance. However in the absence of writing, a number of technologically, and culturally complex societies were able to develop, the most impressive of which was of course Tahuantinsuyo; the Inca Empire, which was the largest empire in pre-Columbian America and was quite administratively complex. In the place of writing they used a system of record keeping which stored information as knots, tied in various places on a series of ropes.
What this digression illustrates is the complete arbitrariness of social evolution. The only reason we write like we do is because people before us did too, even if other behaviour patterns might work just as well, or better.
For another example, take for instance the QWERTY keyboard. It's inefficient. In fact, it was designed to be that way. There are a number of alternative keyboards that are in fact more efficient. A purely natural selection argument would say that a more efficient keyboard would allow people who used it to produce more, and in turn make more money, and then other people seeing the advantages would adopt that keyboard type and it would become more common. But that simply hasn't happened. So, I guess in my opinion, for what that's worth, behaviour and ideas do evolve, "social evolution" if you like, does occur, however the process that forces change is much too arbitrary, and is therefore not terribly analogous to "natural selection."