On 04-11-08 14:12, tgpedersen wrote:
> "sit" is a common verb and occurs in hundreds of compounds in modern
> IE languages.
Hundreds of compounds? Isn't this an exaggeration?
> In original PIE as opposed to constructions in later
> stages, you'd expect that root to ablaut, thus in compounds you'd get
> zero grade *-sd-. To my knowledge there are three examples of that.
> That's isolated.
Well, there are also very few examples of *-bd- from the nil grade of
*ped-, such as Av. frabda-, abda- < *pro-pd-o-, *n.-pd-o-. Does it mean
that they are loanwords? Quite the opposite. They are relict forms
surviving from a time when such ablaut was fully productive. In less
archaic but far more numerous compounds reduction was avoided in order
to preserve the morphological and semantic transparency of the word.
Asyllabic nil grades of roots with two obstruents are generally
characterised by extremely low survival rates. This is why Lat.
prae-se:s, -sidis always keeps its vowel, though we'd expect PIE
*-sé:d-s, *-sd-ós. Three examples of nominal compounds, plus
reduplications like *se-sd- and *si-sd-e- add up to decent attestation,
as such roots go.
>>Still, PIE seems to have had a native word for the thing.
>
> By your reckoning.
By communis opinio, for what it's worth.
Piotr