Possible phonological changes (was: Rate of change)

[Ante Aikio]

On Thu, 14 Jun 2001, R.S.Georg wrote:

> In some New Indian languages, under the so-called ruki-rule, *s went
> as far as (at least) /x/. The ruki-rule decrees that /s/, after /r/,
> /u/, /k/, /i/ goes to /sh/ (in Iranian, and Indic /s-retrofl./.
> Ah, yes, (I'm not fully awake yet), it holds in Slavic, too, where
> the outcome, at least after /u/, is /x/, too  (ru. snoxa, lat. nurus,
> skt. snuSa:, Gk. nuos).
> The phonetic ratio, if any, has to do with retroflexion. If you
> retroflect your tongue *a lot* when pronouncing, say, Sanskrit /S/
> (s-subscript dot), the acoustic impression is already very close to
> (velar, not uvular) /x/. IOW, /S/ and /x/ have quite adjacent places
> of articulation. The whole process from /s/ to /x/ is, thus,
> describable as a gradual movement of place of articulation. OK, in
> the end /x/ has to lose its friction to turn into the stop /k/, but
> then you have it.

I don't think the Mator *s > *k shift has to do with retroflexion; the
context (/_e,i) rather suggests palatalization. So one would rather think
of something like *s > *sj > *cj > *kj > *k, but this feels quite
awkward.

Regards,
Ante Aikio