A study published by Scientific American by University of Auckland (New Zealand) Senior Lecturer, Quentin Atkinson and his team arrived at this map of how The 100+ Indo-European languages originated in modern-day Turkey with and spread east, west, north and south.
In an interview on NPR, Atkinson explains how he arrived at this mapping of linguistic drift:
“…there are two competing theories for where these languages came from. The first theory was put forward by Marija Gimbutas, an archeologist. And then she argued that the languages spread out from the Russian steppes region about five to 6,000 years ago, and that’s linked to archeological evidence for an extension of that time. And then the alternative theory is that the languages spread much earlier, from Anatolia, what is now Turkey, with the expansion of farming. And that would be in eight to nine and a half thousand years.
“So what we were able to do was apply these methods and tests between the competing theories, and we found overwhelming support for the Anatolian farming theory. So that the languages spread from Anatolia with farming. [Ironically, the Turkish language, which domintaes the Anatolian Peninsula, today has its origins in mondern-day Mongolia].
“…we use…what are called cognates. And so we’re looking at the words in the different languages. And within the words, we’re looking for a cognate.
We’ve looked at about 200 different meanings, basic vocabulary terms like hand, foot, mother, father, fire, water, mountain, that kind of thing, across about, well, over 100 of the Indo-European languages. And through all these meanings, we’re looking for these cognates. These are words that show similar form that linguists deduced indicate that they’re related by a common ancestry.”
“..we were able to build up a large set of these cognates, over 6,000 across the 200 meanings in the 100 languages. And from that, we can build this – or it’s a data matrix, we call it. And so if you imagine down the left-hand side of the data matrix, you’ve got all your languages. And then along the top of the data matrix, you’ve got all the possible cognates that we’ve identified. And you fill in that matrix by putting a zero if that language has a – doesn’t have the cognate, and one if the language does have that cognate. And if you fill that in for all the cognates, you get this sequence of ones and zeroes for each language. And it’s a bit like telling the sequence of DNA for each language. And that’s the data that we then input into our analysis.”