(In which I consider more implications of the model discussed in previous posts. Also, the epistemological status of this post is extremely speculative.)
James Flynn points out quite a few reasons to doubt that generational increases in IQ test scores really mean people are getting that much smarter. If IQ truly measures intelligence across generations and cultures, that would make our average parents notably dull and our average grandparents bordeline retarded – or, if you want to look at it the other way, most college graduates from my generation could have easily joined Mensa when it started in 1946. And – to put it bluntly – someone would have noticed if old people were that dumb.