More on Nature vs. Nurture,...

...this time from George Mason economist Bryan Caplan (my emphasis):

The central idea of twin and adoption research is that disentangling nature from nurture is hard. Our intuition isn’t very helpful. Yes, kids of college-educated parents know more words. But why? Maybe their upbringing is the reason, as you suggest. But babies from college-educated families might excel even if raised by high school dropouts, by learning a higher fraction of the words they hear, or spending more time reading.

So what does the twin and adoption data say? Language fits a standard pattern. Consistent with your skepticism, upbringing has a noticeable effect on the vocabulary of young children. But as children mature, this effect largely fades away. The Colorado Adoption Project found, for example, that 2-year-olds adopted by high-vocabulary parents had noticeably larger vocabularies. But as the kids grew up, their vocabulary scores looked more and more like their biological parents’. By age 12, the effect of enriched upbringing on vocabulary was barely visible.