On Thu, 27 Jan 2000, yl-ruil wrote (concerning /s/ -> /h/):

> I think this change occurs generally because s is voiced to z, which then
> weakens to h-. This is also really quite common in the indo-european family:
> *sawelios --> Greek. helios and Welsh haul (sun, obviously).

There is no evidence for an intermediate voiced stage in the
alternation between /s/ and /h/. In fact, such a step would make
the change less likely. What /s/ and /h/ have in common is the
spreading of the vocal folds during articulation--in fact,
that's all an /h/ is; spread vocal folds. This spreading
inhibits voicing (actually, makes it impossible). The lenition
of /s/ to /z/ puts it on a different path whose endpoint is not
likely to be /h/.


Dirk Elzinga
[log in to unmask]