4 ms·

Language Log had a post on this https://languagelog.ldc.upenn.edu/nll/?p=59908 https://languagelog.ldc.upenn.edu/nll/?p=59908 and their commentariat seems unimp

Language Log had a post on this https://languagelog.ldc.upenn.edu/nll/?p=59908 https://languagelog.ldc.upenn.edu/nll/?p=59908 and their commentariat seems unimpressed. Linguist Andrew Garrett had some tweets about that: https://twitter.com/ndyjroo/status/1684636445854875648 https://twitter.com/ndyjroo/status/1684636445854875648 and, for traditionalists, was quoted in an article: https://www.theglobeandmail.com/canada/article-indo-european-language-origins/ https://www.theglobeandmail.com/canada/article-indo-european... The paper hasn't made it to sci-hub yet (at least not the one I checked) but it seems like they're going to have a tough row to hoe.

cmrdporcupine 3y ago
Haven't read the articles or papers yet, but I am really skeptical of the kind of statistical inferences talked about in the article -- "we were able to prove Romance from Latin using our model, so using the same model we got #..."; it seems to me this kind of thing assumes they can project a model of language mutation built from data out of the iron age back to neolithic, which is... a really giant assumption. I'd say dynamics of human history are really full of all sorts of variance and instances of punctuated equilibrium. Still... it wouldn't surprise me to find surprises around the Anatolian languages and their age and origins. They do seem to fall outside the 'norm' of the mainstream of Indo-European languages and the age of their split from the origin of whatever variant of early Indo-European that Yamnaya spoke I suspect could be a lot further back then expected. And it wouldn't also surprise me to find back and forth flow for a period of time from Anatolia to/from the Pontic-Caspian Steppe.
- tgv 3y ago
  That's been my problem with this approach too. It's an arbitrary model, fitted onto very limited data that doesn't generalize, under the very assumption that they all derive from a common root. I'm totally willing to accept that the language and DNA models can be thought of as complementing each other, but any conclusion stronger than "perhaps (almost) all European languages have a common origin" remains speculative. And that's not a very interesting conclusion.
canjobear 3y ago
The only substantive objection I see there is archeological-linguistic. Words related to chariot technology like “wheel” and “yoke” are cognate across IE languages, so the languages probably split after this technology was invented. But the invention seems to have happened later than 8000 BP.
- Archelaos 3y ago
  Can we say something about how often a particular technological adaptation is accompanied by an adoption of the corresponding verbal expression? In other words: If the wheel has been adopted by a neighbouring community, is it not very likely that the word for it will be adopted in a linguistically similar version? As an aside: I looked for the origine of the English "wheel" and the German "Rad", and they seem to have been different IE roots: kʷe-kʷlh₁-ó-/rót-h₂-o-[1], but the word is nevertheless often given as a example for a common IE origine. So what does scholars make so sure that the original Indo-Europeans already had the wheel? Could these words not have been derived somewhat later independently from IE roots of a somewhat different meaning, or when the area occupied by Indo-Europeans was still small via techno-lingustic transfer? [1] https://de.wikipedia.org/wiki/Indogermanische_Ursprache#cite_note-2 https://de.wikipedia.org/wiki/Indogermanische_Ursprache#cite...
  - bigbillheck 3y ago
    You might be interested in 'The Horse, the Wheel, and Language': https://press.princeton.edu/books/paperback/9780691148182/the-horse-the-wheel-and-language https://press.princeton.edu/books/paperback/9780691148182/th...
    - biorach 3y ago
      Excellent book
  - OfSanguineFire 3y ago
    > Could these words not have been derived somewhat later independently That would require that the independent languages maintained the exact same derivational process of reduplication + zero-grade ablaut in the root. We know from the attested history of IE language that derivational processes last only for a while before they die out. That independent languages maintained that particular derivational process (which is inseparable from historical-phonological developments, too) for thousands and thousands of years, is extremely unlikely, which is just one of the many pieces of evidence against the Anatolian hypothesis.

cmrdporcupine 3y ago

Haven't read the articles or papers yet, but I am really skeptical of the kind of statistical inferences talked about in the article -- "we were able to prove Romance from Latin using our model, so using the same model we got #..."; it seems to me this kind of thing assumes they can project a model of language mutation built from data out of the iron age back to neolithic, which is... a really giant assumption. I'd say dynamics of human history are really full of all sorts of variance and instances of punctuated equilibrium. Still... it wouldn't surprise me to find surprises around the Anatolian languages and their age and origins. They do seem to fall outside the 'norm' of the mainstream of Indo-European languages and the age of their split from the origin of whatever variant of early Indo-European that Yamnaya spoke I suspect could be a lot further back then expected. And it wouldn't also surprise me to find back and forth flow for a period of time from Anatolia to/from the Pontic-Caspian Steppe.

tgv 3y ago
That's been my problem with this approach too. It's an arbitrary model, fitted onto very limited data that doesn't generalize, under the very assumption that they all derive from a common root. I'm totally willing to accept that the language and DNA models can be thought of as complementing each other, but any conclusion stronger than "perhaps (almost) all European languages have a common origin" remains speculative. And that's not a very interesting conclusion.

canjobear 3y ago

The only substantive objection I see there is archeological-linguistic. Words related to chariot technology like “wheel” and “yoke” are cognate across IE languages, so the languages probably split after this technology was invented. But the invention seems to have happened later than 8000 BP.

Archelaos 3y ago
Can we say something about how often a particular technological adaptation is accompanied by an adoption of the corresponding verbal expression? In other words: If the wheel has been adopted by a neighbouring community, is it not very likely that the word for it will be adopted in a linguistically similar version? As an aside: I looked for the origine of the English "wheel" and the German "Rad", and they seem to have been different IE roots: kʷe-kʷlh₁-ó-/rót-h₂-o-[1], but the word is nevertheless often given as a example for a common IE origine. So what does scholars make so sure that the original Indo-Europeans already had the wheel? Could these words not have been derived somewhat later independently from IE roots of a somewhat different meaning, or when the area occupied by Indo-Europeans was still small via techno-lingustic transfer? [1] https://de.wikipedia.org/wiki/Indogermanische_Ursprache#cite_note-2 https://de.wikipedia.org/wiki/Indogermanische_Ursprache#cite...
- bigbillheck 3y ago
  You might be interested in 'The Horse, the Wheel, and Language': https://press.princeton.edu/books/paperback/9780691148182/the-horse-the-wheel-and-language https://press.princeton.edu/books/paperback/9780691148182/th...
  - biorach 3y ago
    Excellent book
- OfSanguineFire 3y ago
  > Could these words not have been derived somewhat later independently That would require that the independent languages maintained the exact same derivational process of reduplication + zero-grade ablaut in the root. We know from the attested history of IE language that derivational processes last only for a while before they die out. That independent languages maintained that particular derivational process (which is inseparable from historical-phonological developments, too) for thousands and thousands of years, is extremely unlikely, which is just one of the many pieces of evidence against the Anatolian hypothesis.