Reconstruction talk:Proto-Indo-European/séh₂ls

Move
I disagree with this move. The root doesn't just form this noun, it also forms verbs. Germanic is also transparently a t-adjective derived from a verbal root. 20:29, 24 March 2014 (UTC)
 * A t-adjective would have had to be ×salþaz or ×saldaz; has to be from something ending in -ldos, which doesn't look verbal at all. It just looks like one of the thousands of noun stems with a consonantal extension that show up all over IE. —Aɴɢʀ (talk) 21:13, 24 March 2014 (UTC)
 * You're right, I missed that. But even then, Germanic has, which is a strong verb. If it's strong, that must mean it's primary rather than derived. 21:21, 24 March 2014 (UTC)
 * Not necessarily; it could have been derived an early date before the usual way of forming denominative verbs was settled on, and the -Hd- extension seems to be present only in Germanic and Balto-Slavic, while all the other branches just have a noun based on *seh₂l-/sh₂el-/sh₂l-. —Aɴɢʀ (talk) 15:54, 25 March 2014 (UTC)
 * You mean in Pre-PIE times? That's a hard claim to test... 18:19, 25 March 2014 (UTC)
 * Or even just pre-PGmc, but since the verb has no cognates there's no knowing how old it is. —Aɴɢʀ (talk) 18:47, 25 March 2014 (UTC)
 * But primary verbs were, by definition, always formed from roots and not from other parts of speech. And in Germanic, primary verbs became strong while secondary verbs became weak. So a strong verb derived from anything other than a root is exceptional and needs explaining. 19:01, 25 March 2014 (UTC)
 * Fair 'nuff, but it can't be from a bare root anyway because of the -Hd- suffix. The semantics just scream denominative, and class VII is full of unsolved mysteries. —Aɴɢʀ (talk) 20:12, 25 March 2014 (UTC)
 * Not to me (see this and this paper). The most plausible explanation for *saltaz in my opinion is a bahuvrīhi compound *sh₂(e)l-dh₃-ó- "giving taste/salt(iness)", also continued in Proto-Balto-Slavic *saldus "sweet".
 * That said, Beekes reconstructs the nominative of this noun differently, I think *soh₂l, which makes sense to me. --Florian Blaschke (talk) 00:42, 10 August 2014 (UTC)