--- In
cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@...> wrote:
>
> Well-known facts:
> PIE was an accusative language.
> pre-PIE was an ergative language.
> (pre-pre-PIE was a stative language.)
>
>
> Kenneth Shields proposes that PIE acussative (*-m) was a pre-PIE
> absolutive. That makes it marked, which is odd for an absolutive.
>
> I think the PIE accusative was the pre-PIE allative. That's why
> the 'local case' system of PIE is asymmetric: ablative but no
> allative (as I was puzzled by some years back). Also that means
that
> the pre-PIE absolutive was unmarked.
>
> Which means it's easier to see all the oblique cases as developed
> from 'local cases', genitive from locative, dative and instrumental
> from ablative, and accusative from 'allative' etc.
>
I discussed it off-list with Jens. He tells me that accusative *-m is
from a genitive *-m that later became the gen.pl. (I think it was)
and that this matches well with the idea that 3rd sg *-ti, pl. *-nti
are suffixes forming verbal nouns.
That would not be good for my idea that the accusative was once an
allative (to-case), so I'll turn it on its head, like this:
pre-PIE was an OV-language. Therefore a subordinate clause came
before the main clause and had this structure (eg. as this one
serving as NP in the following main clause):
<noun>erg.-suffix <noun>abs.-suffix <finite verb><nominalising
relative suffix> <main clause> (= O V)
ie.
<noun>-s <noun>-Ø <finite verb><nominalising relative suffix> <main
clause, O V>
When it became an accusative language, the structure would be:
<noun>nom.-suffix <noun>acc.-suffix <finite verb><nominalising
relative suffix> <main clause, O V>
ie.
<noun>-s <noun>-Ø <finite verb><nominalising relative suffix> <main
clause, O V>
except that some people didn't like an unmarked accusative and
started using the allative instead, just to make themselves clear (or
Finno-Ugric substrate?)
<noun>-s <noun>-m <finite verb><nominalising relative suffix> <main
clause, O V>
When it became a VO-language the only structure acceptable before the
main clause was a nominal construction and the structure became:
<noun>-s <noun>-m <gerund> <main clause, O V>
from which one might reinterpret -s as subjective genitive marker
and -m as objective genitive marker, later to mix them up as general
genitives.
And btw, perhaps <finite verb><nominalising relative suffix> in the
above should read as <verbal noun>, given the common interpretation
of *-ti, *-nti I mentioned.
Torsten