Language desk
< April 3	<< Mar \| April \| May >>	April 5 >

Welcome to the Wikipedia Language Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

April 4

Reconstruction of Proto-Indo-European

How confident are linguists in the current reconstructions of Proto-Indo-European? I've seen a good bit of it on Wikipedia and in other places and what I saw looks suspiciously too much like Latin to me. Are linguists sure they're not being influenced by the survivor bias? 93.136.17.76 (talk) 19:23, 4 April 2020 (UTC)[reply]

Nah, they aren't confident at all about it. That's why they use all those weird symbols to fill in the gaps. As for survivor bias? We got... a lot of surviving languages from PIE. Temerarius (talk) 20:21, 4 April 2020 (UTC)[reply]

93.136.17.76 -- They're actually fairly confident that many of the details that they've reconstructed existed at some stage of early Indo-European, but they're not too confident that they all existed together at the same chronological stage of Indo-European -- some might actually belong to earlier stages than others. Also, during much of the 19th century, most reconstructions of PIE were heavily biased towards Sanskrit; the more recent reconstructions are the result of a process of figuring out how much weight to give Sanskrit. And I don't really see the alleged specific similarity of PIE reconstructions and Latin -- Latin doesn't have original aspirated consonants, and many of the reconstructed PIE noun and verb endings are rather different from Latin. For changing conceptions of the nature of PIE, see Schleicher's Fable... AnonMoos (talk) 20:58, 4 April 2020 (UTC)[reply]

If an accomplished scholar of PIE went into a time machine and recited Schleicher's Fable to a native, there is zero chance they would be understood. But said scholar would have a much easier time establishing communication based on their knowledge than almost everyone else would have. Some degree of survivor bias is unavoidable; the reconstructions do not take extinct and unrecorded descendants into account. But by now all recorded descendants have been used in the reconstruction process. If a feature is present in several distant descendants and hardly seen outside the IE family, the likelihood that it developed independently in each is vanishingly small. So then it must have been inherited from a common ancestor with very high confidence. If it is seen in only a few descendants and fairly common elsewhere, a proposal to interpret it as a PIE feature will at best be tentative. --Lambiam 07:22, 5 April 2020 (UTC)[reply]

An example of what I meant: The king and the god - H₃rḗḱs dei̯u̯ós-kwe - Rex deusque? For example why is PIE -k^we so close to Latin -que, how much does it have to do with the fact that Latin is one of the few known written languages that have preserved this conjunction? I admit that after reading Schleicher's Fable I see common forms from many other languages too. Maybe it's just with the words used in The king and the god. 93.136.17.76 (talk) 09:53, 5 April 2020 (UTC)[reply]

For the "and" enclitic form, it's just kind of an accident that Latin preserved the original sounds better than Greek ("-te") or or Sanksrit ("-ca", with "c" pronounced as affricate). With many other language features (such as aspirated consonants) Greek and Sanskrit better preserved early Indo-European. And of course, PIE laryngeals are not preserved as such in any IE daughter branch (except partially in Hittite)... AnonMoos (talk) 10:50, 5 April 2020 (UTC)[reply]