While this may or may not be the case, Pinker is a linguist and has neither the expertise nor evidence to back up this assertion. I don't think modern American suburbs resemble the African savanna. But the chaparral terrain on much of the California coast where I often hike is a lot closer to it.
He is a psycho-linguist and his field is evolutionary psychology. He never claimed a complete resemblance between the savannah and suburbia, but just that both contain many of the key features which would make it highly suitable for the species, at least from an evolutionary perspective.
