Slavic neutra ending -o

From: Вадим Понарядов
Message: 31826
Date: 2004-04-10

PIE *-om usually develops into Slavic -U. Unexpectedly, Slavic neutra of o-declension end in -o. As long as  know, there were two attempts to explain this feature.
 
1. Vyacheslav Ivanov supposes that in IE o-declension neutra originally had zero ending, just as consonant stems do. Slavic -o is the trace of this archaic feature (because phonetically it can regularly reflect PIE *-o).
 
2. Another explanaition belongs to Frederik Kortland, who writes that in Slavic the IE original *-om in neutra was replaced with -od (from pronominal paradigms). Just the same we find in late Anatolian languages, e.g. in Lydian.
 
Nevertheless, neither explanation seems to me satisfactory. Perhaps, it remains still better to assume that PIE *-om could give Slav. *-o sometimes. As a reason for such a conclusion, we find several examples where IE neutra seem to show the regular development *-om > *-U, and therefore are tracted as masculines, e.g. Slav. *darU (m.) vs. Gr. do:ron (n.) "gift". But why could *-om give both *-U and *-o, I fail to understand.
 
Does anybody has any idea on this subject?
 
==========
Vadim Ponaryadov