One should distinguish synchronically
"neutral" vowels from their historical reflexes. For example, the schwa that
developed in Old Polish from the Slavic yers was subsequently phonologised as
either the full vowel /e/ or zero. We have the words <pies> /pjes/,
<piesek> /pjesek/ 'little dog' and <psa> /psa/ 'dog (Gen.sg.)' <
*pj&s&, *pj&s&k&, *pj&sa < *pIs-U, *pIs-Uk-U, *pIs-a.
The modern reflexes are 100% stable, i.e. <piesek> cannot be realised as
/pjesk/ or /psek/, and <psa> is never /pjesa/.
I believe French "schwa" is undergoing a
similar process of phonologisation. If retained, it becomes a full vowel --
usually, though not without some variation, merging with historical /Ø/ (let me
thus transcribe the mid-low rounded front vowel of <jeune>). Retention is
obligatory in words like <brebis>, and these have simply been reanalysed
as /brØbi/ etc. We have "unstable" /Ø/ in words like <semaine>, which
may be realised as either /smEn/ or /sØmEn/ (the informal and formal variant,
respectively). There are also pairs like <genêt> : <jeunet>, which
may be homophonous in formal styles but cannot have the same lexical
representation since the /Ø/ in <genêt> is deletable. Probably the best
solution is to use some kind of diacritic marking in the lexicon for stable
/Ø/ in this position to block its deletion (unstable /Ø/ is far more frequent).
Finally, in words like <samedi> or <médecin> (as well as in final
positions) the historical schwa has already been phonologised as
zero.
To sum up, the so-called "schwa" in French
is not a separate vowel phoneme but a set of reflexes resulting from the
phonologisation of a historical neutral vowel. The variable occurrence of /Ø/ in
some environments is untypical of French full vowels. I have no crystal ball to
consult, but the most natural course of evolution for French pronunciation would
be to eliminate such instability as well as undesirable lexical
marking, so <secours, devant, genêt> etc. will likely tend to become
invariably monosyllabic.
Piotr
----- Original Message -----
Sent: Thursday, August 23, 2001 6:31 PM
Subject: Re: [tied] Glen's Strange Rule
>As a matter of
fact, any unrounded central vowel could play the role of
schwa.
Doesn't modern French use a rounded mid
vowel as its neutral vowel? As in "le" (when the e is
sounded)? Or am I mistaken?
Peter