On 2006-04-08 12:04, tgpedersen wrote:
> Exactly. A root with Jens' *O- as prefix is a compound with *O as
> first element. The question is whether any first element will cause
> the same effects as an *O-prefix.
No. Other kinds of agentive compounds (especially the dHarma-bHr.t-
type) seems to be older, and no such effect is visible in them. Nor do
we have any o-grades in pure root nouns occurring as second members (the
nr.-han- type, which also includes Latin compounds with -spex, -ceps,
etc.). There is no O-infixation in typical bahuvrihis -- probably the
most widespread type of compound in PIE.
Besides, O-infixation occurs regularly in at least one morphological
category in which it can't be decompositional -- the causative/iterative
verbs like *monéje/o-.
Piotr