14.6-13.8 thousand years ago.
The last glacial phase looks like it’s coming to an end, and people in the Natufian culture of the Levant look like they’re gearing up to invent agriculture. They’ve settled in villages, and are harvesting and storing grain, but not yet sowing it. Possibly they’re brewing beer for feasts. This turns out to be a false start though. In a thousand years or so the glaciers will come back for a final hurrah (the Younger Dryas event), and only after this will farming actually get going.
If you plug different frequencies of different genes from a bunch of populations into a computer and ask it to generate a tree where genetically similar populations share closer branches, you get something like this:
This is from the pioneering synthesis of genetics by Cavali-Sforza and co-workers, back in 1994. This looks like a nice diagram of humans spreading out of Africa, maybe some taking a southern route (the Southeast Asian branch), and others a northern route (North Eurasian), and I used to teach it this way in anthropology classes. But as we look at ancient DNA, we’re finding that things are more complicated. Even 14,000 years ago, the structure of populations is different from what we’re used to today. We’ve already mentioned the Ancestral North Eurasians earlier, who just maybe could have spoken a language ancestral to Greenberg’s Eurasiatic family.
In the Near East, too, things were complicated. A paper out just last year shows that there were three very different hunting and gathering populations in Anatolia, Western Iran, and the Levant. Folks in Iran and the Levant were as genetically distinct as modern Europeans and Chinese! Either the Near East during this period had just been settled by migrants from widely separated places, or there had been strong barriers to gene flow in place for some time. Since then people in the area have mixed a lot.
Each of these populations of hunters and gatherers will give rise to its own set of farmers. The Natufians will contribute a lot to the ancestry of later farmers in the Levant. And apparently each set of farmers will send migrants off in a different direction: the Anatolians to Europe, the Iranians (or some Caucasian relatives) to the Eurasian steppe, and the Levantines to East Africa. It’s possible that the Natufians were speakers of a language ancestral to the Afro-Asiatic family, one of the oldest widely accepted language families, including Arabic and Hebrew, Somali and Oromo.
And here’s Razib Khan on our new understanding of how different ancient population structure was.