Re: THEORY: Re : THEORY: Semivowels

From Http://Members.Aol.Com/Lassailly/Tunuframe.Html wrote:
>
> Dans un courrier dati du 08/09/99 17:44:37  , Gustavo a icrit :
>
> >     What really are the semivowels? I mean, are [j] and [w],
>
> there is another beautiful one : [|u"]
> (can't write the IPA symbol) :
> french : "suave, nuit, icuelle".
>
        I love that sound. It's a pity that it be so rare among languages of
the world.

> >  respectively, just short [i] and [u], or are they really different
> >  sounds?
> >          Can semivowels be nasalized?
>
> the surest way to get S nazalised is to nazalise
> the V after :
> french w in "loin, jouant" and |u" in "juin"
> are sometime pronounced nazalized,
> depending on locutor.
> but i don't think | in "puant" and j in "chien"
> and "lion" ever are.
> maybe there are phonetic rules for that ?
>
        Nasalisation is a strange feature in French. I think it is the only
language where nasalisation doesn't spread from nasalised sounds to
phonemically non-nasalised ones. What I mean is that in many language,
the occurence of a /m/ or /n/ will nasalise the vowel before, or the
vowel after, but it's not the case in French ('bonne' is pronounced
/bOn/, not /bO~n/, to make it different from 'bon' /bO~/). Even
languages that have phonemic nasal vowels don't behave like that (in
Portuguese, the occurrence of /n/ or /m/ always nasalise the vowel
before, for example).

        About your examples, I pronounced the w glide nasalised in 'loin', but
not in 'jouant', and I pronounce u" nasalised in 'juin' but not in
'puant'. I think the difference is that 'loin' and 'juin' are not
analysable, whereas 'jouant' and puant' are (jou-ant, present participle
of jouer, and pu-ant, present participle of puer). To keep the
connection with the verbal stem which is generally not nasalised, I
don't nasalise it. I don't know if it is a dialectical feature or not.
For 'chien' and 'lion', I think the i becomes more a mark of
palatalisation of the previous consonnant than a real glide (at least
when I don't speak carefully). As nasalisation never spreads to
consonnants in French, I think that's why it is not here.

> mathias
--
        Christophe Grandsire

        Philips Research Laboratories --  Building WB 145
        Prof. Holstlaan 4
        5656 AA Eindhoven
        The Netherlands

        Phone:  +31-40-27-45006
        E-mail: grandsir@natlab.research.philips.com