Hi, Jerzy,
There are a number of natural trajectories
of sound change. It's common for [s] to change into [s.] (a postalveolar or
"retroflex" sound) -- it's enough for the tip of the tongue to be slightly
retracted towards the palate. Both sounds are defined as "coronal", which means
that they are articulated with the front part of the tongue. Some languages
typically realise their /s/ phoneme as a remarkably retracted fricative in all
positions, while in others retraction is conditiones by certain well-defined
phonetic contexts. In the latter case we may eventually get a split of an
original /s/ into two phonemes with quite different phonetic
realisations.
The acoustic effect of retraction is
enhanced if the sound is "dark" (velarised), i.e. if the tongue-tip articulation
is accompanied by the raising of the back of the tongue towards the soft palate
(the back part of the roof of the mouth, a.k.a. the velum). If now the
tongue-tip "gesture" is weakened while the tongue-back raising is emphasised,
the fricative may shift towards a predominantly "dorsal" (tongue-back) rather
than coronal articulation, becoming velar [x].
This is what happened in the prehistory of
Slavic: first, in some environments (most notably after *r, *u, *k and *i) PIE
*s was retracted to [s.] (the same or only minimally different development,
known mnemonically as the "r-u-k-i" change, is visible in Indo-Iranian and
Baltic). Then -- and this is a uniquely Slavic development -- this [s.]
changed into velar [x]. When we say that PIE *s became Slavic *x in
this or that context, we in fact take a mental shortcut, taking the intermediate
stages for granted.
Incidentally, when followed by a front
vowel or *j, *x was realised as a palatovelar fricative [รง], which evolved
smoothly into Slavic palatoalveolar *s^ (= [S], pronounced like English "sh").
The latter sound was then dispalatalised in some languages, becoming again [s.]
(= Polish "sz", Russian "sh"). This is why we have ucho [uxo] 'ear' but uszy
[us.I] 'ears [originally dual]' < *aus(-os), *ausi:.
The articulation of [s] may also be
"lenited" or weakened by reducing the tongue-tip movement; as a result, the
fricative may change directly into the glottal glide ("aspirate") [h]
(which is regarded by many phonologists as a "basic fricative" deprived of any
place-of-articulation components). This development is familiar from the
prehistory of Greek, Iranian and Brythonic Celtic; it also takes place (in
syllable-final positions) in some Spanish dialects. Of course more complex
evolution is also possible, e.g. [s] > [s.] > [x] > [h].
Our Spanish and Portuguese members could
tell fascinating stories about the development of coronal fricatives in their
languages.
Piotr
----- Original Message -----
Sent: Sunday, November 26, 2000 12:33 AM
Subject: [tied] /s/ to /x/ sound change
Hello!
My question may seem a bit amateurish, but I am
an amateur
in this field (being an software engineer by profession):
When
reading the discussion on slavic endings I started to
wonder how easy/likely
is the /s/ to /x/ sound change?
I know it happens, but these sounds are no so
similar (well,
both voiceless fricatives), so this sound change is not
that
trivial as palatalization, for example. I would rather imagine
/s/
> /S/ > /x/. What (phonetically) help such a
change?
Jerzy
p.s. I follow this list for a quite a long time and
I like it
very much, but never dared a post before.