Thread:User talk:CodeCat/Proto-Indo-European/reply (15)

o-grade in Germanic actually becomes a-grade (we still call it o-grade though), because of the change o>a in Germanic; Proto-Germanic didn't even have a short o. Short u (and in descendants also o) always goes back to a syllabic (zero grade) form of w, l, r, m or n: u stays u, and l̥, r̥, m̥, n̥ become ul, ur, um, un in Germanic. So the o-grade that you mentioned on the page for is wrong, and it seems that the Germanic forms that were originally listed (before you removed them) were actually correct. Aside from that, there is no ə in PIE, it is just a way to indicate the sound of a laryngeal between consonants (which we write with the laryngeal letter on Wiktionary). But I don't see why it would appear in some of the verb forms. Personally I wonder what source you have for the verb forms. Does that source specifically mention that the root's present aspect had athematic root inflection?

In any case, a root that is ancestral to 'rock' would have to contain one of those sounds in it somewhere. Since there is no l, r, m, n in the Germanic words, the conclusion is that it must have contained w. By Kluge's law this must go back to either rewk/ḱ-, rewg/ǵ- or rewgʰ/ǵʰ-. M. Philippa's etymology dictionary for Dutch suggests h₃ruk-néh₂- from the root h₃rewk- "to dig up, to pull out". She mentions the following cognates: Latijn, Ancient Greek , Sanskrit , Latvian.

One note: you say that the change from e to a is typical of Indo-Iranian. While this is true, it does not apply in this particular case. Already in PIE, e was pronounced as a when it was next to h₂, and e was pronounced as o next to h₃ (but we still write e in both cases). So the vowel change in this particular word is actually much older and has nothing to do with the Indo-Iranian change.