Ask and ye shall receive. In a recent post on eugenics, I claimed that the connection between early 20th century genetics and early 21st century genetic engineering was weak. I asked if anyone thought I was wrong, and in no time I got a comment from Razib at Gene Expression.

He suggests that I'm limited by conventional preconceptions, taking issue on both my points--first about the prospects of engineering intelligence and second about the prospects of a new species of engineered humans. I think he's got a stronger argument on the first point than the second.

On the first point, Razib argues that it wouldn't be as hard as I think to engineer more intelligence. I said maybe thousands of genes would have to be tinkered with, and he pointed out that if individual genes typically accounting for around 1% of variation, then it shouldn't take thousands of genes to engineer significantly brighter people. OK--I'll give in on the thousands, although nobody can really say what the exact number is. But even with hundreds (or even dozens) of genes, you're still dealing with a level of complexity that dwarfs anything I've seen reported in this area, even in mice. And if I'm blindered by conventional preconceptions, then at least I'm in good company. Here's an essay Steven Pinker wrote last summer that lofts the same bucket of cold water.

On the second point--making a new species--Razib thinks that you could get enough barriers up around the new population of engineered humans to get speciation. He writes:

"...those barriers can be social, if some religious nutsos decided to create biphallic sons, there would be issues with these sons being able to get mates from the mono-phallic majority. Additionally, GE [genetic engineering] would by its very nature alter the ground rules for speciation as mutation in the context of genetic drift and natural selection plus physical barriers thrown up by geography, etc. might not be the only sources of reassortment & segregation of genes within a population...."

It's true that barriers can be social--songbirds develop new tunes that make them sexy only to certain females, for example. But you still need some serious isolation to get them singing a new song before you bring the new population back in touch with the old one. (Like putting them on another island for a while.) Otherwise, the differences just wash out. I suppose you can try to imagine some Dr. Moreau engineering men with twin-penises (along with bivaginal women, I guess?), but it just shows how far you have to go into X-Files territory to make an argument for speciation. Genetic engineering is certainly a form of mutation that the world has never really seen before. But that doesn't mean that it cancels all the rules about how new species form.