En réponse à Mark J. Reed :

>Incidentally, to my non-native-French ear it seems that the closest French
>vowel to [U] is the one in "oeuf".

Really? "oeuf" is [9f], i.e. the vowel is the rounded version of [E]. I 
would have thought you would compare [U] rather with [u] and [o] than with 
[9]. The distance with that one looks rather far... And to my French ear, 
[U] sounds nothing like [9]. As Philippe showed very well, we confuse it 
rather with [u].

To Philippe: you see, a native speaker think [U] is closer to the vowel in 
"oeuf" than the vowel in "coup". So for the native speaker, the difference 
between [u] and [U] is just as big as the difference between [u] and [9] is 
for us.
En réponse à Philippe Caquant :

>Can't understand that point about "cure". I always
>pronounced it something like "kju:r", and my Harraps's
>says someting similar. What has it to do with a nurse

If you had truly read what Trebor wrote, you would have seen how he 
pronounces "cure": [kjr\=]. No [u] in there! And probably he pronounces 
"nurse" as [nr\=s] or something similar, so in his speech "cure" and 
"nurse" are assonant (i.e. they have the same vowel. Note that [r\=] is 
vocalic). That's what it has to do with "nurse". And stop hiding behind 
this dictionary. The Harrap's only presents word pronunciation in the 
"British Received Pronunciation", a sort of neutral British flavour which 
has an existence only in the mouth of teachers teaching English to 
non-English speakers. It doesn't say anything about the English spoken by 
most native speakers. Remember one thing: English *doesn't* have an 
Academy, it *doesn't* have a hierarchy of dialects, it *doesn't* have a 
standard form. It doesn't even have a majority dialect! English is a jigsaw 
puzzle, with each piece being a dialect, and no piece is much bigger than 
the other. So your dictionary doesn't describe a "better" version of 
English, nor a "standard". It just describes a variant of British English 
that is commonly taught to foreigners. This variant is not majoritary, and 
not standard.
En réponse à Roger Mills :

>But another large segment of the population has unrounded and centralized
>the vowel to a schwa (actually [3^] or [3:] depending on rhoticity), and
>rhymes this word with _her, fur, burr, nurse_ etc.

Philippe has difficulties realising that English is not structured like 
French, with a standard dialect enforced by school and under control of the 
Academy (and forced on people with violence until not that long ago). He 
cannot really understand the variety of English dialects and the big 
differences they can have. And he doesn't get that dialects are not just 
accents (i.e. identical, except that some sounds are changed a bit, like 
Southern French has an alveolar trilled r, while Northern French has an 
uvular fricative r) but have differences up to the phonemic level, the 
morphological level, and the syntactic level. Centuries of centralised 
French education do that. But Philippe has to understand that the 
linguistic monolith that French is is an exception rather than the rule. 
Most languages, including all written ones, are no more than collections of 
reciprocally intelligible dialects, with  *none* being truly standard.

>"Surely you've studied 'Hamlet'?"
>"Don't call me Shirley!"
>Ha ha groan

LOL. I think Philippe should get it now ;))) . The example was very good :)) .

Christophe Grandsire.

You need a straight mind to invent a twisted conlang.