Reconstruction talk:Proto-Indo-European/krewh₂-

s-stem
The Italic, Hellenic, and Indic all have sigmatic stems. For Balto-Slavic, Derksen reconstructs *krúh₂-s ~ *krewh₂-, and LIN list both the zero-grad root noun and the full grade sigmatic. I'd say that Celtic, Balto-Slavic, and Germanic need to be put under the zero-grade root noun and Italic, Hellenic, and Indic under the sigmatic stem. — JohnC5 14:57, 16 May 2016 (UTC)
 * What evidence is there for an -s- in Greek? Also, an ablaut *krúh₂-s ~ *krewh₂- would be very strange. The stem never has a stressed zero grade in the nominative of athematic nouns. —CodeCat 15:37, 16 May 2016 (UTC)
 * : Greek has ~  /  and a later reformulated genitive . Yeah, I was curious about the zero-grade too. I suppose the athematic root noun could easily be thematicized from the oblique in all the daughter languages. — JohnC5 19:00, 16 May 2016 (UTC)
 * There's no strict need for an -s- there though. Morphological analogy is also feasible, i.e. when krea- was analysed as the stem (based on the nominative), it could have triggered the creation of other consonant stem forms based on that stem. —CodeCat 19:03, 16 May 2016 (UTC)
 * But the genitives κρέαος / κρέως are older and do represent PH *-as- stems, which are exactly what you expect. The later *-t- stem is also very rare. Beekes says that, besides the accent, the Sanskrit and Greek are formally identical and sigmatic stems. Curiously, Derksen and Beekes both reconstruct a zero-grade root; though I agree that it is weird. I didn't notice before but Derksen says that the Baltic are from a diminutive *krowh₂-yos whereas the Slavic are the root (zero-grade?) noun. — JohnC5 19:34, 16 May 2016 (UTC)
 * Slavic reflects a zero grade, yes. Slavic -y- comes from an older long -ū-. However, Slavic simply reflects earlier -ūC- ~ -uwV- throughout, so this could easily be the result of levelling. So Derksen's reconstruction is quite strange as it is exactly the opposite of the ablaut you'd expect, and a zero-grade nominative isn't needed to explain the Slavic outcome. In light of the Greek full grade, it has to be levelled in some way, and it makes more sense that the full-grade nominative was levelled away in Slavic, rather than a full-grade oblique stem. —CodeCat 20:35, 16 May 2016 (UTC)
 * Let me clarify slightly. Beekes specifies a zero-grade for Slavic and not for Greek. Shall we go for an athematic *kréwh₂ ~ *kruh₂- and sigmatic *kréwh₂-s ~ kruh₂-s-? That seems the most parsimonious to me. As for the Latin, De Vaan wants *kruh₂-ṓs (PI *kruwōs > L cruor), which looks like Klimp and Kloekhorst's hysterokinetic collective, but *kréwh₂ōs might still develop something like PI *krouwōs > L cruor. — JohnC5 20:56, 16 May 2016 (UTC)
 * Keep in mind that an amphikinetic noun would have zero grade of the root and suffix in oblique cases regardless: genitive *kruh₂-s-és. So it may not be necessary to posit *krouwōs > cruor, it may simply be Italic *krūōs with the long u introduced from the oblique cases. Of course it's always nice when a form phonetically results directly from a PIE form, and I do think *kréwh₂ōs can directly become cruor (or more precisely, the earlier unlevelled form *(?)cruos), but it's certainly not the only possible explanation. —CodeCat 21:34, 16 May 2016 (UTC)