--- In cybalist@..., "Piotr Gasiorowski" <gpiotr@...> wrote:
>
> ----- Original Message -----
> From: niffabs
> To: cybalist@...
> Sent: Friday, April 19, 2002 7:20 PM
> Subject: [tied] Re: Agriculture and IE
>
>
> > The term "satemised" would seem to imply that the satem group
split from the centum group. I would be interested to know what
theories there are presently as to
> how the centum/satem split might have occured, as there seems to be
no obvious connection?
>
> In my view, the "centum" systems are closer to the primitive state
affairs, and the shift of *k^ > *c (palatal stop/affricate) and *kW >
k is a common innovation of the Satem languages. This view accounts
for the geographical distribution of the Satem branches (almost
continuous even now).
>
> > I have seen the suggestion elsewhere that the PIE word might have
started with ksh as in kshatriya - the centum group then dropping the
sh, and the satem group dropping the k. There is a case of the sh
being dropped where many people in present day Punjab would say
Khatti rather than Kshatriya. Any thoughts on this would be
interesting.
>
> This is unlikely. First, PIE had clusters of the type *k^s or *kWs,
and their development is on the whole different from that of *k^
alone. Secondly, the early Satem reflexes of the three rows of stops
*k^, *g^ and *g^H were not fricatives but affricates, preserved as
such in Nuristani, and partly in Armenian and Indo-Aryan:
>
> PIE Nur. Arm. Skt.
> *k^ ts s s'
> *g^ dz ts <c> dz' <j>
> *g^H dz dz <j> h (from *dz'H)
>
> This (and more, e.g. the phonetic shape of loans from early Satem
dialects in Finno-Ugric) is consistent with the usual theory that the
change proceded like this: [k] > [c] (palatal stop) > [ts']
(parallelly for the corresponding voiced and breathy voiced series),
from which the attested reflexes ([s'], [ts], [s], [S], [þ]) can be
derived easily via commonplace trajectories of phonetic development.
If you want a typological parallel, there is a rather similar range
of reflexes of palatalised /k/ in the Romance languages (cf. It. &
Rom. tS-, Sp. þ-, Fr., Port. & dial. Sp. s- in the word "centum",
while Sardinian retains the original k-), the difference being that
the Romance palatalisations were conditioned by a following front
vowel while the satem shift was not restricted to a specific context.
>
> Piotr
I still think it phonetically odd that the satem languages, in
contrast to the Romance ones, palatalize also before back vowels and
even consonants. Is there an explanation for that? Is there even a
parallel in the world?
I proposed earlier that the origin of the division centum/satem
languages might originate in inflectional paradigms involving a velar
stop and the Ablaut vowel, ie:
*-ke-/*-ko-/*-k-
(or in Cuny's formulation: *-kä-/*-ka-/*-k-)
developing into
*-c^e-/*-ko-/*-k- (Cuny: *-c^ä-/*-ka-/*-k-)
in other words, irregular paradigms, which were regularized thus
in satem languages
*-c^e-/*-c^o-/*-c^- (Cuny: *-c^ä-/*-c^a-/*-c^-)
from where it developped as outlined by Piotr
and
in kentum languages
*-ke-/*-ko-/-*k- (Cuny: *-kä-/*-ka-/*-k-)
As to why -k- would be replaced by -c^- everywhere in satem
languages: I believe the distinction developped into a shibboleth
between neighboring languages, which would force the speakers to
adopt the "new pronunciation" or face the consequences.
Has this taken place elsewhere? Except for between Danish and Swedish
(see
http://www.angelfire.com/rant/tgpedersen/Shibbolethisation.html
)
I can think of one partial example: Lat. ka- -> (c^a ->) French s^a-
(spelled cha-). There is to my knowledge no phonological explanation
for this. One might however imagine a state of affairs where the
French (and this in this case means Northern French) elite insists on
following Piotr's sequence (see above) with their underlings
following them, and catching up with them, phonologically, in the
sequence. This would push the elite speech more towards -s-, even
involving ka- -> c^a etc, if that's what it takes to keep the social
distance. (Similarly, as far as my ears tell me, Stockholm Swedish
has taken original /s^/ furthest towards /hW/, away from the
dialects).
Torsten