gcochran commented the LW discussion on the blog:

I noticed the exchange about the genetics of schiz: lots of confusion. My take is that current data suggests that most schiz iz the result of mutational pressure: possible because it is really many different syndromes lumped together – a grab-bag. The number of loci that can mutate in a way that causes something in that grab-bag is very large. When people do find causal variants, they tend to be rare, generally specific to a particular (and shrinking) family. Average lifetime of an allele of strong effect is something like three generations.

That said, environmental factors can surely contribute. There is a real but small seasonal pattern, and a real but strong influence of prenatal starvation. The Dutch famine of 1944 and the Chinese famine associated with the Great Leap Forward each roughly doubled schizophrenia in those in-utero at the time. Then again, it’s higher in big cities than in rural areas. 2.4 times more common in Copenhagen than on a Danish farm.

Selection pressure against schiz is fairly strong today and was likely stronger in the past: people too crazy to farm or hunt generally starved, along with their children. Creativity may well be more stronger in unaffected relatives, but I doubt that creativity boosted fitness much in most past environments. While having long conversations with little silver men living behind the wallpaper was in general quite bad for fitness.