[chch]

Here's a previous thread from when this was originally posted: https://news.ycombinator.com/item?id=5669179

I remember, because I made two of my most researched comments there [1]. :)

I'm no historical linguist, but I'd take this finding with a bunch of grains of salt. Eurasiatic language families that seek to combine, say, Proto-Indo-European [2] and Altaic [3] languages are pretty controversial, and in general this paper reeks a bit of glottochronology [4], which is a pretty controversial topic in historical linguistics itself.

Wikipedia's Eurasiatic language page actually even has a large section about this very article, including some refutations[5].

[1] https://news.ycombinator.com/item?id=5670947

[2] https://en.wikipedia.org/wiki/Proto-Indo-European_language

[3] https://en.wikipedia.org/wiki/Altaic_languages

[4] https://en.wikipedia.org/wiki/Glottochronology

[5] https://en.wikipedia.org/wiki/Eurasiatic_languages#Pagel_et_....

[jcranmer]

> I'm no historical linguist, but I'd take this finding with a bunch of grains of salt. Eurasiatic language families that seek to combine, say, Proto-Indo-European [2] and Altaic [3] languages are pretty controversial, and in general this paper reeks a bit of glottochronology [4], which is a pretty controversial topic in historical linguistics itself.

It's worth pointing out that everybody who actually studies the relevant Altaic languages now agrees that the Altaic family doesn't exist, not even in 'micro-Altaic' (just Mongolic/Tunguskic/Turkic languages) form. Basically, the consensus is that the similarity between those three languages arise from deep contact rather than genetic relationships (think what happened to Old English after the Norse and Norman invasions on its way to Middle English).

[mannykannot]

As someone who knows nothing about linguistics, I was surprised that the article expressed surprise that "[these root words] can be predicted from information independent of their sounds. We showed in a sample of Indo-European languages that the frequency with which a word is used in everyday speech, along with its part of speech, can predict how rapidly words evolve, with frequently used words on average retained for longer periods of time."

I would have guessed that, on the basis of utility, the conservation of words would fundamentally be a function of their meaning, and, for the most part, that function would be fairly constant across all human cultures, regardless of the specific languages used by those cultures (again on grounds of their utility.) From that perspective, would it be all that surprising if approximately the same set of meanings were conserved even in languages with no common history?