In a pretty informative piece in Gizmodo, Scientists Say New Research Tracing the Origin of Modern Humans to Botswana Is Deeply Flawed, there is an interesting quote that I would like to follow-up on:

That said, Curtis Marean, a professor of archaeology at Arizona State University who wasn’t involved with the new research, is not a fan of the idea that multiple origin locations exist for modern humans. In an email to Gizmodo, Marean said this idea doesn’t “make sense at all,” and that evolution “happens the quickest in small isolated populations—we know this—so that will always remain the favored hypothesis, until proven otherwise.”

The intuition that evolution happens fast in a small population is a common one. And, on some level it has justification. As per the neutral theory, the time until the fixation of a new allele is proportional to the effective population size. Basically, in small populations frequencies can change faster due to drift.

The idea is that you can “random-walk” your way into innovation. A lot of this thinking draws consciously or unconsciously from the shifting balance theories of Sewall Wright, where small populations subject to higher drift can stochastically explore more of the evolutionary parameter space than a large well-mixed population. The problem as outlined in Will Provine’s Sewall Wright and Evolutionary Biology is that the shifting balance theory operated as more of a metaphor than a formal model that could be tested. In The Structure of Evolutionary Theory Stephen Jay Gould argued that Wright deemphasized drift and played up gene-gene interactions in his later years because of the expectations of the Modern Synthesis. But this was possible probably because as Provine documents Wright was not always clear himself how the shifting balance worked. When drift became less fashionable he tweaked the shifting balance accordingly.

Gould was probably the most influential in promoting the idea that small populations are where the “action” in evolution is to the general public. Many evolutionary biologists demurred on this. The reasoning is that small populations may evolve fast due to drift, but drift can lead to fixation of deleterious alleles, and selection is more effective in large well-mixed populations. Additionally, in structured meta-populations, you may have scenarios where drift is periodically powerful, but a great deal of adaptation occurs during periods of mixing, so that drift may not be consequential.

Probably the dominant position in regards to speciation and differentiation of populations in an evolutionary genetic framework is that allopatry matters. Separation. Isolated populations are necessarily allopatric in relation to parent groups, but you can have two large populations that are separated as well. Rather than a small population leading to some sort of biological innovation, I suspect it will more often lead to extinction.

Which brings us back to humans, and the intuition that modern humans derive from a small distinct population. We know from ancient DNA that many human populations were often quite isolated and inbred. This holds for many Neanderthals, Denisovans, and Mesolithic European hunter-gatherers. These groups were on the edge of the range of humans, so it’s not surprising. And, it looks like there was a lot of local extinction of these populations. The archeology and genetics imply that it was Africa, with larger numbers of hominins, which was the source for new populations several times (the ancestors of Neanderthals and Denisovans replaced earlier hominin groups, and arrived from Africa ~750,000 years ago).

There are human groups that went through bottlenecks or were isolated. The ancestors of all non-Africans, for example. The ancestors of Amerindians. The ancestors of Oceanians. You can see this is aspects of the allele frequency; these populations have undergone more drift due to small populations. There is no evidence that I know that isolation was evolutionarily beneficial in a biological sense (in The Dawn of Human Culture Richard Klein leveraged Gould’s ideas to propose that a small ancestral human population was subject to a macromutation which allowed for recursive language).

Rather, let me suggest that cultural changes or geographical contingencies lead to evidence of bottlenecks being associated with successful populations. ~15,000 years ago a group of humans managed to make it to the Americas. This was a small ancestral group which by chance was afforded the opportunity to expand their range very fast. Such a demographic expansion from a small founding population has left its impact on the genetics of Native Americans. The genetics of the small isolated group did not lead to demographic expansion. Similarly, ~4,000 years ago Austronesians began to expand their range through long-distance voyages. The bottlenecks left an imprint genetically. But again, the source of innovation was not genetic isolation, genetic isolation emerged due to cultural innovation allowing for rapid fission.

Humans, at least before the Holocene, were not numerous. Meta-population dynamics and periods of isolation between demes seems to have been normal. Rather than isolation resulting in laboratories for experimentation, I suspect most of the time it just led to extinction for many groups. But not all groups. And because small isolated populations tended to be on the margins of the species’ range, you have scenarios such as where the ancestors of non-Africans expanded very fast and were quite successful (the expansion of Amerindians and Oceanians was similar). Additionally, while small populations may not be beneficial for biological evolution, the conditions for cultural revolution are different. Peter Turchin has observed, along with many others, that some of the most innovative and energetic populations tend to be on the fringe or boundary of a broader cultural zone (e.g., the rise of Rome and Qin from the fringe of the Classical and Chinese worlds).