Personal Pronouns

[Jens Elmegaard Rasmussen]

On Sun, 16 May 1999, Patrick C. Ryan wrote:

[...]

> I believe, with Beekes, that IE was once an ergative language; and that the
> absolutive form in IE was -0. I assume that the attested accusative *te
> maintains the form of the absolute before the introduction of animate
> accusatives in -m. In the datives *toi and *tebh(e)i, we see it retained
> also. If *tu/u: were the underlying form, we should expect **twi and
> *tubh(e)i generally (although this probably explains the Greek forms, e.g.
> in s-.

I have yet to see a really cogent argument for an IE (or pre-IE) ergative.
One is constantly served mere descriptions of what the system would be
like if it is accepted, but nothing to convince one that it _must_ be
accepted. I'm not saying that pre-PIE was _not_ ergative, just that we
cannot really know. Practically all the literature on the subject simply
boasts that the author knows what an ergative is and so is without
scholarly interest. It is as if I would claim that PIE had a definite
article, just because I know what that is.

[On the refl. pronoun in the nom.:]

> Do not confuse dependence with function. "I myself am going". Is "myself"
> not a nominative depending on another nominative?

I'm sure it is, as far as function goes, but I'm also rather sure that the
IE pronoun did not have a nominative form to fulfil this function. What
was the nominative form in your opinion, and what are its reflexes in the
daughter languages on which the reconstruction is based? - Are you
_denying_ that Latin se, Greek he, German sich and Russian sebja have no
nominative? And is this fact - if so it is - to be ignored in a
reconstruction of the protolanguage?

[...(On Gk. 'us two':)]

> My dictionary shows <no:{^}e> *only* as a poetic variant of <no:{^}i>. Do
> you have different information?

Some, but that's not important; even that was IE. Admittedly, the form
no:^i is difficult. It looks like nothing we really know, except an old
dative or locative. One could imagine no:^i being on a par with the
gen.-dat. enclitics moi, toi, hoi; or it could have come from *nH3we-i, a
loc. formed like Skt. asme' 'in, among us' from *nsme-i. The immediate
development of *nH3wei into Greek would be *no:wei, whence, with loss of
the digamma, *no:ei, and with contraction finally no:^i. In the anaphoric
pl. we have acc. sphe and dat. sphi, as parallels of which one might
accept no:^e and no:^i. It would then be a sign of decay of the dual
number category that the i-form is no longer consistently restricted to
the dative. There is _no_ IE comparative support for no:^i, whereas for
no:' and no:^e there is. Thus, it is on the weakest possible grounds that
you see a pre-IE dual in *-wy (of the Egyptian kind) reflected in no:^i. I
admit that my reconstruction *nH3we' for no:^e and Avest. a:va is not
_much_ better, but still it combines forms of IE languages and so works
within the confines of a genetic frame known to be valid and for which the
rules are known. This also makes me immune to your following statement:

> I reject unequivocally your H{3} as a part of the reconstruction.

I feel we are about to reach the borders of how far this subject can
be taken.

Jens