Reconstruction talk:Proto-Celtic/dubus

"*dʰewbʰ-" or "*dʰewb-"?
, we have which seems semantically appropriate. I know the *b is weird, but it's the only way to account for Germanic and. — JohnC5 21:53, 13 October 2016 (UTC)
 * I'm not sure if they are the same thing, though. —CodeCat 21:57, 13 October 2016 (UTC)
 * Well, we may have some disentanglement before us. All the Balto-Slavic and Celtic descendants that *dʰewb- has are from Derksen and Matasovic, which they reconstruct as *dʰewbʰ-. The only reason they do this, so far as I can tell, is that *b is unlikely. Indeed, among the proposed Balto-Slavic, Celtic, Tocharian, Greek(?), and Germanic descendants, only Germanic distinguishes *bʰ and *b. I think that "dark" and "deep" could make a coherent set of meanings for one unified root, but if you think they should be split, do we just leave Germanic out in the cold? — JohnC5 22:18, 13 October 2016 (UTC)
 * There's a bit of circular reasoning going on here. We say *b is rare because there are so few roots containing it, but we also refrain from reconstructing roots with *b, so that makes it rare. Balto-Slavic and Greek would also distinguish them; Greek retains the aspiration, and Balto-Slavic perhaps with an acute register. I'm not sure if unaspirated voiced consonants leave an acute register on preceding diphthongs though. —CodeCat 22:22, 13 October 2016 (UTC)
 * Oh yeah, of course AG would have the aspiration. I forgot about that for no apparent reason. I was unaware of the Balto-Slavic acute behavior. Do you have more information on that topic? — JohnC5 22:29, 13 October 2016 (UTC)
 * The main thing is . But that only creates acutes when it lengthens vowels. So I'm not sure what happens with a preceding diphthong. Probably nothing? —CodeCat 22:38, 13 October 2016 (UTC)
 * Isn't there a theory that Winter's Law applies only in closed syllables? Balto-Slavic accentology is very mysterious to me. Regardless, no change does seem like the most likely result. So, should we stick with *b or should we separate them semantically? — JohnC5 22:47, 13 October 2016 (UTC)