13.5 kya. For a long time the Clovis culture, associated with these spear points found across North America, looked like the earliest evidence of human occupation of the Americas. Clovis people are often thought to have entered the continent through an ice-free corridor that opened up between glaciers in Western Canada (although there are other possibilities).
But we now know that there were people in the Americas before Clovis. The Monte Verde site in Chile, dating more than a thousand years earlier (tweeted yesterday) is the best evidence. And just in the past few years, we’ve learned something about the genetics involved. Modern Amerindians are overwhelmingly a mixture of East Asian and Ancestral North Eurasian ancestry. But a small fraction of their ancestry is something else: it connects them with some relict populations in Southeast Asia and Melanesia. (In current Southeast Asia these populations have been pushed aside by later arrivals.) The best current explanation for this pattern is that an early population traveled along the Pacific Coast all the way from Southeast Asia to Chile by 15 thousand years ago, when the inland route was still blocked by ice. When Clovis and related peoples made their way to the New World, these early migrants were largely replaced, but left behind a trace of their ancestry.
There’s actually some linguistic work that’s consistent with the genetics. Johanna Nichols is a linguist who has identified various linguistic strata in the New World. These are not the same as language families. Think of it this way: the Khoi-San languages of Africa are famous for their click sounds. These languages, spoken by Bushmen and other isolated hunting and gathering groups, may be extinct before the century is out. But click sounds have been borrowed by some of the Bantu neighbors of these groups. While these non-Khoi-San click speakers do not constitute a language family, they do tell us something about pre-Bantu language history.
Nichols finds a far-flung linguistic stratum distinguished by a large number of otherwise rare features, in both Melanesia and southern South America. We might compare Nichols work on linguistic strata with Greenberg’s work on Eurasiatic. The two use different methods. Both are intensely controversial within linguistics. But in both cases, it looks like distant geographical affinities proposed by linguists get support from the latest genetic research. (We’ll be getting to more generally accepted language families later this month on Logarithmic History.)