Reconstruction talk:Proto-Celtic/alten

RFDO discussion: September 2016
No descendants. —CodeCat 22:05, 20 September 2016 (UTC)
 * I moved it to . — JohnC5 01:26, 21 September 2016 (UTC)
 * Striking this, but the Welsh descendant is puzzling. Why is there no t? —CodeCat 01:28, 21 September 2016 (UTC)
 * Also, it looks like this is a PIE ē (> Celtic ī) rather than a Celtic ē, judging from the Brythonic descendants. So is probably the proper lemma form. —CodeCat 01:31, 21 September 2016 (UTC)
 * Word-internal nt became nh/n in Welsh (e.g. >  > ), so maybe lt did something similar. —Aɴɢʀ (talk) 12:45, 21 September 2016 (UTC)
 * What would the appropriate declension of *altīn be? Also, what PIE forms could lead to and should it be changed to ? — JohnC5 19:34, 21 September 2016 (UTC)
 * There are a few PIE nouns with nom.sg. in -ēn, acc.sg. in -enm̥, gen.sg. in -nos, but they're pretty rare. I think the canonical example is . I have trouble believing such a rare type made it into Celtic, especially in a word whose etymology is "most likely from a non-IE substrate language". There's no way Proto-Insular Celtic borrowed a word from a neighboring language and assigned it to an extremely rare declension class that's probably otherwise completely unattested in Celtic. —Aɴɢʀ (talk) 20:38, 21 September 2016 (UTC)
 * Sorry I was unclear: I'm very familiar with the PIE nouns in *-ḗn (,, , etc.). My question pertains to the fact that the normal reflex of PIE *ē is *ī. So why was the original template (whether or not the declension existed in PC) created as *-ēn and not the expected *-īn? — JohnC5 20:53, 21 September 2016 (UTC)
 * That I don't know; we'll have to ask CodeCat since she's the one who made the template. As for this word, Matasović says the Brythonic forms are from *altinā, and although (which could come from *altīn or *alten but not *altinā) is attested in Middle Irish, the older form is actually  (which could come from *altVnā with any back vowel for V, or maybe even from *altinā if we can cobble together a reason for the lack of palatalization, since ordinarily you'd expect *altinā to give *ailten). I suspect  is a late analogical form built to look like an n-stem. —Aɴɢʀ (talk) 20:58, 21 September 2016 (UTC)
 * Probably an early stupidity. Don't mind me, I'll be over here. —CodeCat 21:35, 21 September 2016 (UTC)
 * Soooo... should we either alter or wack this table? — JohnC5 02:45, 22 September 2016 (UTC)