Fortunatov's Law and tolkAppiyar's rules

[Miguel Carrasquer Vidal]

Vidhyanath Rao <vidynath at MATH.OHIO-STATE.EDU> wrote:

>Miguel Carrasquer Vidal <mcv at WXS.NL> wrote:
>> *s         =>    => h  => h
>> *sC        =>    =>    => s
>> *{ruki}s   => s^ =>    => s^
>> *k^        => c^ => c  => s
>> *k^C       => c^ => s^ => s^
>
>I missed this point till now.
>
>The change c^ -> c cannot be later than the second palatalization (change
>of PIE *k/*k-super-w to c^ before e/i) which must have preceded the
>merger of e/o/a. So, under the above scenario, the merger of e/o/a must
>be later than the split of Indian from Iranian.
>
>I find this quite incredible.

It depends on the exact phonetics of Proto-Indo-Iranian.

Clearly, the satem and the ruki-phenomena are the oldest, as they are
shared by Indo-Iranian, Balto-Slavic and Armenian.  For "proto-satem"
(proto-East-Indo-Eropean), we can assume:

PIE    PEIE
*k^ >  *c^  (satem)
*ke >  *k^e (palatal allophone of /k/ before e, i)
*s  >  *s
*rs >  *s^  (or *is, *us, *ks: RUKI-s)


In Indo-Iranian we have:

PIE   PEIE   Avest  Skt
*k^   *c^    s      s'
*ke   *k^    c^     c^
*s    *s     h      s
*rs   *s^    s^     s.

After the (Indo-Iranian) merger of e/a/o, PEIE *k^ (i.e. PIE *k, *kw
before *e, *i) became phonemic rather than allophonic.  But was it
already [c^] or simply [k^]?  If it was [c^], then PIE *k^ (the satem
sound) must have already been something else in PII, probably *s',
from which it follows that RUKI-s was already *s. ("retroflex"
shibilant) in PII:

PIE   PII
*k^   *s'  (> s in Iranian)
*ke   *c^
*s    *s   (> h in Iranian)
*rs   *s.

[PII == Skt. in this analyis, at least for these four sounds]

On the other hand, if we reconstruct *k^ (palatalized [k]) for PIE
*ke/i, *kwe/i, we get:

PIE   PII
*k^   *c^  (> s' in Skt., > c > s in Iranian)
*ke   *k^  (> c^ in Skt, Iranian)
*s    *s
*rs   *s^  (> s. in Skt.)

Both options are possible.

The same two analyses are possible for Balto-Slavic, where we have:

PIE    Balt  Slav  PBS?
*k^    s^    s     c^ ~ s'
*ke    k     c^    k  ~ k  (allophonically: [k^])
*s     s     s     s  ~ s
*rs    s^    x     s^ ~ s.

We know the 2nd. palatalization in Slavic is recent.  Ruki-s has
shifted back to /x/ in Slavic, which must have happened under the
influence of satem-*k^ becoming /s^/, which means that it wasn't /s^/
initially.  What was it?  The two likely candidates are /c^/ and /s'/
(and consequently, RUKI-s becomes either /s^/ or /s./).


=======================
Miguel Carrasquer Vidal
mcv at wxs.nl
Amsterdam