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

From: Miguel Carrasquer
Message: 31979
Date: 2004-04-17

On Sat, 17 Apr 2004 14:51:23 +0000, elmeras2000
<jer@...> wrote:

>--- In cybalist@yahoogroups.com, "elmeras2000" <jer@...> wrote:
>
>> I read this to make us expect <*wlkWos-
>> yos + nominative> made up of genitive + inflected form of the
>> relative pronoun. Is this correct? or does it in some way support
>> Glen's interpretation of *-yo as a locative or dative (of abnormal
>> formation)?
>
>I rush to correct an error here. The relative pronoun should not be
>expected to agree with the possessed with regard to case, but should
>be constantly in the nominative: "the eye which [is] the wolf's"
>will have the word for "which" in the nominative of the number and
>gender corresponding to the word for "eye". That is, one may
>postulate earlier existence of forms like *wlkWos-yaH2, *wlkWos-yod,
>*wlkWos-yo:(w), *wlkWos-yaH2iH1, *wlkWos-yoyH1, *wlkWos-yoy, *wlkWos-
>yaH2as, *wlkWos-yaH2 also. As a representative of them all one may
>assume that first the nom.sg.masc. *wlkWos-yos was used, and
>subsequently the shortened form *wlkWos-yo appeared. If "which" is
>inflected to agree in case with "eye" we have "relative attraction"
>which is in fact non uncommom, but can hardly be the original
>structure.

But it is what is in fact attested in Anatolian.

Anatolian has a relational suffix *-assa-, *-assi-, which is
used as a genitive (but inflected as an adjective!) in C/H
Luwian (-a(:)s(s)i(:)-) and Lycian/Milyan
(-(a/e)(h/s)(e/i)-). It also occurs as an adjectival suffix
in Hittite (-assa-) and Palaic (-asa-), although these
languages have a regular genitive in -as. Luwian and Lycian
also have a rare uninflected (true) genitive in -is(s)a,
-(i)he/-(i)se (< *-(e)s(s)a). Pisidic, Sidetic and Carian
have a genitive in -s (Carian -s') which must also derive
from either inflected *-assa- or uninflected *-(e)so (final
*-s would have been lost).

Lydian, like Luwian/Lycian, also does not use a true
genitive but an adjectival construction, except that the
sufix is -al(i) (also found in Carian genitive -l).

I haven't seen Françoise Bader "Problématique du génitif
thématique sigmatique" (1991, BSL 86/1, 89-157) yet, but my
preliminary hypothesis is that these forms (adjectival
*-assa-, genitive *-es(s)a) are ultimately related (and
probably also further connected with gen. *-os).

Melchert derives *-assa-/*-assi- from *-eh2so-, but I find
the reasoning shaky. He rejects a link with *-os(-) because
Lycian -ahe would point to /a/ (> Lyc. a), not /o/ (> Lyc.
e). However, Lycian Umlaut (a > e, e > a) makes that
argument dubious (Umlaut is usually regressive [so only the
final vowel is original], and we would always expect
-ehe/-ehi. In the case of this suffix, however, Umlaut
appears to be progressive, back-vowel roots having
-ahe/-ahi, front-vowel roots -ehe/-ehi). The
Proto-Anatolian assimilation *-Vh2sV- > -VssV- is also shaky
(e.g. Hitt. pahs- etc., explained as away as
non-intervocalic), and conflicts with Palaic -asa- with
single -s-, for which Melchert does not exclude Georgiev's
proposal that *-sy- gives Palaic -s-. Perhaps then *-sy-
gives -ss- in Hittite and Luwian/Lycian. This all might
then also explain the irregular lengthening (because closed
syllable) in HLuw. -a:ssi-, for which the only example would
be precisely this suffix. But if the vowel was /o/,
lengthening is regular. I would therefore reconstruct PA
*-asya- < *-osyo- (also *-esyo-) (the alternation -o/-i
["i-motion"] in Luwian/Lycian is secondary). A similar
suffix would have been *-olo-/*-elo- (Lydian -al(i)-, Carian
-l, Hittite pronominal genitive -el).

That such a suffix develops into an inflected genitive
(Luwian, Lycian) first, then to a normal genitive (Pisidian,
Sidetic, Carian) is attested. The problem is explaining the
non-inflected forms with fixed final *-o, as seen in
Luwian/Lycian -issa/-ihe and in PIE *-esyo, *-osyo. Jens'
suggestion of dissimilation s...s (nom.masc. *-o-syo-s >
*-osyo) sounds more attractive than anything I've ever come
up with.

I still have some questions (what's the relationship between
*-osyo- and the athematic gen. *-os? Why do we find *-osyo
outside Anatolian only in the pronominal/thematic
declensions? etc.) but I'll read Bader first.


=======================
Miguel Carrasquer Vidal
mcv@...