tgpedersen <
tgpedersen@...> wrote:
> 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)
Well, my interpretation is completely different...
I think accusative *-m comes from, as you write above, a <relative suffix>, in fact from a relative pronoun in the oblique case.
The available constructions would have initially been something like the following:
A. <unmarked> [man I-see] "I see (a) man"
B. <marked> [man whom I-see] "The man that I see"
(B) was then reinterpreted as [man-whom I-see], i.e. [man-ACC I-see], denoting a definite object: "I see the man". This is still the function of the *m-accusative in Uralic and Altaic. In PIE, the construction with *-m completely replaced the unmarked construction (A), at least for animates.
Note that this implies a phase with zero marking of acc. (and nom.) singular in a pre-stage of IE, Uralic and Altaic.
From the formal point of view, the suffix appears as *-m in Uralic and *-b in Altaic, i.e. the same variation that we find in the 1st. person singular pronoun. In Indo-European, besides nominal *-m, the suffix appears as *-mé ~ *-wé after the personal pronoun roots (sg. *m-é ~ **mé-me [> *méne], *t-wé ~ *té-we, *s-wé ~ *sé-we, du. *n.h3-wé, *uh3-wé, *sphé [< *s(w)h3-wé], pl. *n.s-mé, *us-mé, *s-mé). In view of this, we can reconstruct the suffix as IE/Alt. *mu-á, Ural. *m-á (cf. 1st. person pronoun IE/Alt. *mu-, Ural *mi), an oblique (*-a) form of the relative/interrogative pronoun *mV "what, which" (in PIE, this pronoun has been replaced by the pronoun *ku- > *kW-is ~ *kW-os).
In Georgian, this same pronoun *mV is the basis of the ergative case (OGeo. -man > -ma), i.e *ma- + ergative suffix *-n.
The Eskimo-Aleut genitive/ergative in *-m may also be based on *ma.