[tied] Re: Origin of Thematic Neuter -om (was: 1sg. -o:)

> That isn't obvious at all. It it were a postposition, like
> is usually claimed, it would behave like a postposition,
> which it doesn't.

A PIE postposition, yes, not a pre- or non- one.

> A postposition in PIE is added to a case form of a noun, so
> we would expect *p&2tros-bhi, not p&2tr.-bhi-os,
> *p&2tris-bhi, not p&2tr.-bh-is. It is true that some
> postpositions are themselves inflected case forms of nouns
> (e.g. *h2ant-i, which is the locative of *h2ant-
> "forehead"), but not in this case, because a noun of the
> shape *bhi is impossible in PIE.
>

If we look at the PIE pre-/post- -verb/-positions themselves, they
seemed to have suffixes -o, -i, -er. That's not part of the PIE case
system, but might be part of a pre-PIE or non-IE case system. Even
Basque loans postpositions.

> A postposition can be added freely to the singular or the
> plural of a noun. There are no postpositions that attach
> only to the plural.
>
> A postposition can be added to nouns no matter what their
> declensional class. There are no postpositions that attach
> exclusively to athematic nouns.

The fact that *bhi attaches only to athematic nouns should tell you
something about its age.

> Obviously, *-bhi- cannot be a postposition.

of PIE. *abhi/*obhi with variants *pi etc might though. Note the
similarity with the *xap- "water", *poGW- "(make) drink" complex; I
think it that set of postpositions etc had to do with postion
relative to 'our' river, as is common in some communities.



Torsten