On 2006-07-12 12:28, Pavel A. da Mek wrote:
> What PIE ending is cognate with Gmc. -o: ?
*-ah2, as in feminines
>> And what form of the compound can be expected:
>> <k^ol-bHuh2-tlo-s>, <k^ol-o-bHuh2-tlo-s>, or maybe <k^ol-i-bHuh2-tlo-s>?
>
> So rather
> <k^l-bHuh2-tlo-s>, <k^l-o-bHuh2-tlo-s>, or maybe <k^l-i-bHuh2-tlo-s>?
The compound was invented by J.R.R. Tolkien and I'm not sure what _he_
would have expected. In productive composition, mostly of post-PIE date,
the first member, if thematic, has *-o-. In Tolkienese <holbytla> the
second member is an agent noun with a weak (nasal) stem, and the -a
reflects what's usually thought to have been a superlong (contracted)
*-õ:n. The WS verb <bytlian> is denominative, the ending -ian looks like
a secondary regularisation of older *bytlan (cf. Anglian byldan) <
*Buþl(i)jan-. In PIE verbs derived from morphologically complex nouns
were less easily formed than in Germanic, so I would hesitate to project
the reconstruction back onto PIE.
Piotr