>However, *-i does occur with *-su in the thematic declension
>(*wlkWo->i-su). This suggests that *wlkWoi is the locative proper, while
>*-su >is a postposition.
Miguel in response:
>No, -oi- is here a plural (> dual) morpheme, straight from the
>pronominal declension (and hence found in the o-stems, which are to be
>interpreted as adjectives/nouns with postpositioned pronoun, much like
>the later Slavic definite adjectives).
Now why would o-stems have a postpositioned pronoun? While this might be
phonetically possible, I don't see the grammatical sense of it. I've put
forth already that thematic stems derive from stuff like *-os
adjectives/nouns where their genitive ending has been re/mis-interpreted as
nominative, leaving a new thematic stem in *-o-. No need for postpositioned
pronouns. Misanalysis answers things quite cleanly. Your pronominal
solution, which is again overly mechanical, leaves us open to even more
questions like "What would this 'pronominal' ending have conveyed
originally?", "How did it fit into IE grammar?", "How would it have
arisen?", "Why postfixed and not prefixed?", etc, etc. which would
necessitate further theories and speculation.
Plus, I was under the impression that the accusative conveyed "definiteness"
(at least this is so in Uralic where ablative conveys indefinite nouns), so
in that view, one would wonder how IE would require "definite adjectives"
for anything. Thoughts?
>All my books are packed (I'm moving tomorrow), but from memory we find >the
>-oi- in the Skt. dat./abl.pl. (-oi-bhi-o-s) [one suspects from >earlier
>*-oi-os] and in general in the ins.pl. *-o:is < *-oi-h1-s. It >also occurs
>in dual forms like Grk. -oiin/-oiun < *-oi-Hw-i-m.
... or is this simply a product of assimilation across the declensional
paradigm? Maybe *-oi- (Sanskit /e/) wasn't there to begin with in some of
these cases but was eventually used as a kind of generalized intervening
Piotr to Miguel:
>As the locative of *-i/*-u stems (and often of consonantal stems) is
> >endingless, the Loc.pl. in *-su could be analysed -- rather
> >conjecturally, I admit -- as a zero-ending locative plus *-su.
Is there not attestation of *-su itself without a preceding *-i- (like
say... Sanskrit /s'vasu/ < *k^won-su)? If so, and if the locative was
unmarked by case ending in the singular at one time, as it appears with
*k^won-su, we should expect the locative plural to have been *-(e)s like the
nominative if anything. Since *-u is also an attested locative, we have a
clear solution: *-su is composed of the plural plus a secondarily attached
locative ending *-u. Even if it is found as *-i-su from time to time, the
fact that we find *-su without *-i in defiance of the singular shows that
the latter, less ordered form must be the original form. Since the pattern
of *-i [sg] versus *-su [pl] is not immediately clear logically, a seemingly
regular plural in *-is (locative sg + plural) would have to be an ending
created to _replace_ the original ending *-su.
At what point is the above logic flawed?
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com
Share information about yourself, create your own public profile at