The Burakumin of Japan, the Paekchong of Korea, and the Cagots of France … What do they have in common? All three were despised castes—closed groups of people who married among themselves. A despised caste is not just a low class. Otherwise, it would always be gaining and losing members, with some moving up and out and others down and in. As Gregory Clark has shown, the English lower class is descended largely from people who were middle or even upper class a few centuries before. This may seem strange if you equate the middle class with voluntary childlessness, but until the late 19th century they were the ones who had the most children—even more so if we look only at children who lived to adulthood. The resulting demographic overflow continually spilled over into the lower class.
In contrast, not much new blood flows into a despised caste, at least not on an ongoing basis. Social stigma discourages people from marrying out or marrying in. Nor does one enter simply by virtue of being poor, since the fear of losing caste keeps out most of the downwardly mobile. Despite this lack of new blood, a despised caste can perpetuate itself indefinitely because its members usually have enough resources—through their monopoly over equally despised occupations—to get married, form families, and have enough children to replace themselves. This was not the case with urban lower classes of pre-industrial times, which typically had large numbers of childless single men.
Because a caste is closed and self-perpetuating, it may preserve genetic traits that disappear everywhere else. It thus becomes more and more different not because it is changing but because its host population is changing.
But how can a population change over a few centuries? Didn’t human nature assume its present form back in the Pleistocene when cultural evolution took over from genetic evolution? In reality, these two evolutionary processes have reinforced each other. Human genetic evolution actually accelerated 40,000 years ago and even more so 10,000 years ago, apparently in response to a growing diversity of cultural environments.
What about Richard Lewontin’s finding that human genes vary much more within populations than between populations? Isn’t that proof that genetic evolution stagnated while humans were spreading over the earth and forming the many populations we see today? Lewontin’s finding is correct but does not mean what it seems to mean. Indeed, the same genetic overlap has been found between many species that are nonetheless distinct anatomically, morphologically, and behaviorally. Genetic variation between populations differs qualitatively from genetic variation within populations. In the first case, genes vary across a boundary that separates different environments and, thus, different selection pressures. This kind of genetic variation is shaped by selection and gives rise to real phenotypic differences. The situation is something else entirely when genes vary among individuals who belong to the same population and face similar selection pressures. That kind of variation matters much less, the actual phenotypic differences often being trivial or nonexistent.
Human evolution is a logarithmic curve where most of the interesting changes have happened since the advent of farming and complex societies. Homo sapiens was not a culmination but rather a beginning … of gene-culture co-evolution. There are many ways to study this co-evolution, but one way is to look at the different evolutionary trajectories followed by castes and their host populations.