Re: Order of Some Indo-Iranian Sound Changes

From: Piotr Gasiorowski
Message: 63411
Date: 2009-02-24

On 2009-02-24 04:29, Andrew Jarrette wrote:

> To tell the truth, I actually have a few more questions concerning
> the Russian postalveolar/medio-palatal spirants and affricate.
>
> 1. If Russian <c^> comes from palatalized *k(W) and Russian <z^>
> comes from palatalized *g(W) and *g(W)H, why do they differ in
> articulation in that <c^> is soft medio-palatal as you described it,
> while <z^> is apical postalveolar? They both arise from the
> palatalization of (labio)velars which only differed in voicing in
> Proto-Slav, did they not?

Yes. Already in Common Slavic a sligh asymmetry developed as *3^ was
deaffricated, becoming *z^. Other Slavic languages developed new
affricates later on, but Russian deaffricated all the voiced ones and
was left with /c^/ and /c/ (the latter from the second palatalisation or
progressive palatalisation of PSl. *k). I suppose this skewedness of the
obstruent system made /c^/ available for individual treatment.

In standard Polish, by the way, the sounds represented <sz>, <z.>, <cz>,
<dz.> (s^, z^, c^, 3^) area all apico-postalveolar; so are the reflexes
of palatalised *r (spelt <rz>), which have merged with /z^/ and /s^/
(once upon a time it was a trilled fricative, like Czech r^). There is
also a regional phenomenon known as "mazurzenie" (Masovian
pronunciation) -- the merger of the whole apico-postalveolar series with
the dental sibilants <s, z, c, 3>. The reflexes of Old Polish /r^/ were
not affected, since they developed more recently.

> 2. In many grammars of Slavic languages, <s^> is listed as the
> palatalization of [x].

Or rather, the RUKI sound, *s^, was retained in palatalising
environments and shifted to /x/ elsewhere (cf. Spanish and some Modern
IA languages). *Synchronically*, /s^/ has been grammaticalised as the
palatalised allomorph of /x/, but diachronically it must have been the
other way round.

> Is this the main origin of <s^> (<sz>, etc.) in Slavic languages? I
> always noticed that in dictionaries at least, <s^> seems to be
> noticeably less frequent than <z^> or <c^>, except in words of
> foreign origin (thus especially less frequent in OCS). Is this
> because it really is primarily the result of palatalization of [x] of
> whatever origin? Or are there other major sources for the <s^> sound?
>
Another major source is *sj (including PIE *k^j). In some, but not all,
Slavic languages, the palatal *s' from the second palatalisation of *x
before new front vowels, merged with *s^ (in the rest of Slavic, with
*s). *z^ is more frequent because of its affricate origins (<
palatalised *g, *gH, *gW and *gWH) with *zj (with *z < *g^, *g^H) as an
extra source.

> And is there any reason why it aligns with <z^> rather than <c^>
> apart from the fact that it is a fricative rather than an affricate
> (i.e. because of the IE or Proto-Slav consonants or consonant-vowel
> combinations it derives from)?

I don't think so. It was just attracted into the same natural class.

> 3. Apart from *s after RUKI, what _is_ the origin of Slavic [x]? It
> sometimes seems to correspond to IE initial *ks- or *kWs-,

This is a RUKI environment (plus cluster simplification, probably *ks^ >
*s^ > x). Note, however, that *k^s, *k^รพ (the "thorny" cluster) > Slavic *s.

> sometimes to IE *sk- (at least before consonants), yet sometimes it
> seems IE *sk- is preserved before consonants. What could be the
> conditioning factors? I fully realize that these questions have been
> addressed by many scholars, and that the question is not resolved. I
> just seek your opinion or the opinions of the other members of the
> list, or maybe you could cite the journalistic articles or books
> which deal with this issue.

Still debated and pretty far from being quite resolved, just as you say.
In a few cases we may be dealing with a generalisation of RUKI from
prefix-root sandhi (esp. in the verb root *s^Id-/*xod- 'go, walk'), in
others, with Iranian loanwords. It is also widely believed that there
was a pre-Slavic aspirated *kH (from *k(W)h2 and sometimes from PIE *kH,
if it did exist, at least as an allophone), and that it merged with *x
from *s^. I haven't got a bibliography to hand, byt I'll try to make a
short list later on.

Piotr