Indo-Iranian

[Jens Elmegaard Rasmussen]

On Sat, 29 Jan 2000, Rick Mc Callister wrote:

> Looking at Watkins's chart of correspondences
> IE	Sanskrit	Avestan		Old Persian	OCS	Lithuanian
> k	s'		s		th		s	s
> kw	k/c		k/c		k		k/c^/c	k
> g	j		z/g		g/d		z	z
> gw	g/j		g/j		g/j		g/z^/z	g
> gh	h		g/z		g/d		z	z
> gwh	gh/h		g/j		g/j		g/z^/z	g

> 	Why does Old Persian look as far removed [or more] from Avestan as
> any of the others? What's the time difference? Is the difference between
> Avestan and Old Persian as great as the chart would indicate?

First, the chart is not accurate, probably due to inaccurate copying (I
don't know the exact source). In the IE column, kw, gw, gwh are obviously
meant to include also plain velars (i.e. non-palatal k, g, gh), while the
"IE k, g, gh" of the chart stand for palatals only. Second, some haceks
are missing in the Lithuanian part of the chart (and of course understood
in Sanskrit, Av., OP where "c, j" are traditional notations of c^, j^). It
should be:

> IE	   Sanskrit	Avestan		Old Persian	OCS	Lithuanian
> k^	   s'		s		th		s	s^
> k/kw	   k/c		k/c		k/c		k/c^/c	k
> g^	   j		z		d		z	z^
> g/gw	   g/j		g/j		g/j		g/z^/z	g
> g^h	   h		z		d		z	z^
> gh/gwh   gh/h		g/j		g/j		g/z^/z	g

Thus, as regards palatalization (th alternative options following the
slants), the two Old Iranian languages behave the same, and just like
Sanskrit. The only difference between Avestan and Old Persian is now in
the phonetics of the old palatals. It is known that the Proto-Iranian
pronunciation of the old palatals was that of dental affricates [ts],
[dz], because (1) there are loanwords in Tocharian and Armenian that
retain this form, (2) inherited *-ts- comes out the same (Skt. matsya-
'fish' : Av. masiia- : OP unattested *mathiya- behind Mod.Pers. mahi:g),
(3) the Iranian dialect distribution of zasta-/dasta- 'hand' shows d- in
dialects that otherwise have z-, this being interpreted to represent a
dissimilated reflex of Proto-Iranian [dzasta-] with loss of the first
sibilant prior to the change of dz to z (Klingenschmitt's observation).
Thus, the Avestan (and "Medic") developments ts > s and dz > z are as in
French (Lat. palatalized c > OFr. [ts] > Mod.Fr. [s]), or Greek or Latin
(nepo:s from *nepo:t-s), while the Old Persian ts > th is as in Spanish
(and Albanian). The sibilant release of the old palatal was assimilated to
the dental more and more, giving a change from ts' to ts and further to
tth (i.e. stop t + fricative th, all within the duration of the old
palatal), the final step being loss of the occlusion whereby ts became s,
and tth became th. The voiced palatal must have a similar history: dz' >
PIran. dz > Av. (etc.) z, pre-OPers. ddh (stop d + spirant dh release) >
plain spirant dh > stop d (or ddh > d directly?).

   Thus the Avestan and Old Persians reflexes are straightforward
further developments of the same Proto-Iranian stage. In most other
respects, Old Persian is so plainly a close relative of Avestan that this
has never been questioned, and indeed cannot seriously be.

   The Indic development, then, cannot represent the Proto-Indo-Iranian
pronunciation unchanged. The palatal s' must have had some occlusion in
PII, since there is still some in PIran. [ts], so this phoneme will have
to be posited as [ts'] for PII; in parallel fashion, the reflex of IE *g^
(which has retained its occlusion in Indic j) must be PII [dz'], and its
aspirated counterpart [dz'h]. In Iranian the aspiration was conbsistently
lost in voiced consonants, so that both of these became [dz] (> z/d),
while the late palatals j^ and j^h (i.e. [dz^h]) fell together as j^
(conventional notation j). In Indic the voiced assibilated affricates lost
their buccal traits and simply gave a voiced [h].

   The Proto-Indo-Iranian phonetic values of the IE palatals thus were
[ts'], [dz'], [dz'h], or, in a simpler notation *c', *j', *j'h with an
acute accent on the affricate to indicate prepalatal position (as opposed
to the hacek on *c^, *j^, *j^h denoting a more central palatal position of
the younger palatals). It appears that the IIr. occlusion of the old
palatals is also retained in the Finno-Ugric renditions of the oldest
loanwords that have been taken from a prestage of the Indo-Iranian
protolanguage.

Jens