[tied] Re: Nominative Loss. A strengthened theory?

--- In cybalist@yahoogroups.com, enlil@... wrote:

>[...] The accusative of */wlkW&-s ya hWa:kW-s/
> would be */wlkW&-s ya hWa:kW-m/ because *ya is a LOCATIVE
> and _seperate_ from hWakW-. The clause is showing _where_ the
> eye is and hence cannot be expected to agree in case with
> *hWa:kW-s. It wouldn't make sense in fact if it did.
>
> 1. *wlkW&-s here must be analysed as BOTH a nominative
> AND a genitive (a case merger)
> 2. *ya is added for clarity, just as we similarly say "own"
> as in "John's book" / "John's own book"

2. contradicts 1. If *yo (your horrible "*ya", presumably written
this way to patch over the lethal flaw of the final vowel) is
superfluous to the construction, the first noun must be a genitive.
Otherwise the genitival relationship is not expressed, and, quite
honestly, this is a language with case marking.

> 3. As per example in 2, "own" carries the genitive sense.
> Likewise, since *wlkW&-s cannot elucidate the true case
> of the noun, whether gen. or nom., *ya is the only thing
> that can disambiguate the cases.

If speakers stuck to the genitive in its old form *-e-s, there would
be no such problem with *-esyo; and, when *-esyo has been
established as the normal genitive of thematics, there is no problem
in changing it to *-osyo so as to comply with other cases where the
thematic vowel is *-o-.

> 4. If *ya were a nominative, there's nothing "genitive"
> about the morpheme to convey the genitive. In English,
> "own" DOES have a possessive meaning inheirantly.

This is not about English. In the Iranian Izafet, the element
corresponding to *yo (your horrible "*ya"), viz. ya-, the relative
pronoun, is not in the genitive, but either in the nominative
(underlying construction) or in agreement with the possessed (by
attraction). Therefore assuming the same for IE is not introducing
something unheard-of. True, this is not about Iranian either, but it
*is* about the relative pronoun.

> 5. Ergo, even if we can get around inventing an endingless
> animate nominative for a thematic stem that doesn't
> exist, *ya cannot be nominative for functional reasons
> because, simply, a nominative can't convey a genitive.

Well, that's exactly what the construction does in Iranian, and with
the article also in Albanian and Greek. In these languages the
connecting particles resume a genitive in the cases that are
parallel to this. That works just fine.

> 6. The only case with an endingless form to explain all of
> the above is a locative which CAN convey a genitive.

Not if some other case lost its ending for whatever reason.

> So *ya is clearly locative and DOES NOT need to agree with the
> possessed noun. Rather, _Jens_ predicts these forms of **-syo-z,
> **-syo-m, **-syo-i, etc because he insists in a nominative
> ending that is conclusively not there.

I insist the form has *-o which is at variance with both analyses:
An endingless locative should have *-e, and an inflected nominative
should have *-os. Since both are a bit off, we are free to pick what
we like - but we are NOT free to base great portions of the rest of
the morphological analysis upon it, for whatever we choose will be
very weak here (still, mine is supported by the high probability of
*-so < *-so-s now, while yours has no basis).

> I accept what I see in Reconstructed IE, don't predict forms
> that don't exist

You postulate a locative *yo which certainly does not exist, and
does not even fit our expectations. That is worse.

> and I end up explaining the origin of *-syo
> efficiently.

"The eye at which there is a wolf"?? Or even, "the eye at which
there is the wolfe's [soul? shadow? smell? aura?]" ??

> I can't see the problem here.

I'm afraid you are quite right saying this.

> Shouldn't we be
> questioning Jens' solution then for its inefficiency?

You have been doing that all along. I would have given it up if you
had produced anything to indicate I should.

Jens