----- Original Message -----
From: Vincent Goetry
To: phoNet@egroups.com
Sent: Wednesday, August 02, 2000 1:53 PM
Subject: [phoNet] Hi
Dear Vincent,
I think the problem is that we the term "schwa" is used in two related but not identical senses.
(1) Schwa = 'mid central vowel' = [@] or [3].
(2) Schwa = 'phonetically neutral, "reduced" or "featureless" vowel'.
In many languages the neutral vowel sound tends to be mid central simply because a central vowel is not [front], [round], [high] or [low], and these are the most commonly used distinctive features. There is however quite a lot of phonetic latitude in the pronunciation of a "reduced" vowel, since there is no articulatory target to aim at. The actual pronunciation of such a sound may vary according to the segmental environment and move away from the mid central position. English [@] in "sof[@]" is very different from the identically transcribed vowel in "stom[@]ch".
As different languages favour different articulatory settings, Russian schwas sound rather differently from English ones, and even different accents of English may not have the same phonetic range of reduced vowels. Many languages have two or three reduced vowels rather than a single "schwa", thus retaining a limited range of contrasts in weak positions (e.g. between a high and a non-high central vowel).
On the other hand, a central articulation may itself be a phonetic target, as in British English "b[3:]d" (note that [3] is the preferred transcription in such cases) or "l[^]ve" as pronounced in some accents. The quality of such a "strong" sound ("Schwa in Sense 1) is then relatively stable and represents a fully contrastive vowel phoneme.
"French schwa" is defined mainly in terms of lexical instability (p'tit ~ p[@]tit, j' t[@] l' dis ~ j[@] t[@] l[@] dis, etc.). Tradition is in fact the main reason why we transcribe this vowel as [@]; in phonetic terms it doesn't strike one as a reduced vowel at all. On the contrary, it is clearly front and rounded. One could say that French has "empty slots" for vowels in certain words, which, when filled, host a *full* segment, virtually identical with the vowel of "neuf".
I'd be interested myself in other members' comments. For example, do native speakers of French detect any significant difference between the vowel of "neuf" and the surface realisation of the first vowel of "pr[@]mier"?
P.S. As for your P.S., Vincent, you needn't worry -- it's a very phoNetic problem :)

Vincent wrote:

I'm a Ph.D. student and I'm working on bilingualism and literacy
acquisition in children. Actually, I'm working on comparative phonology: it is claimed that the schwa belongs to both French and Dutch phonologies. Of course, this phoneme has very different acoustical properties in these two languages (at least because it is a reduced vowel in Dutch but not in French). Does anyone of you know any reference regarding the comparison between the schwa in these
two languages? Thanks in advance

P.S. I really don't know if my profile fits the interests of the PhoNet
group. If not, please let me know.