Re: OT: Phonetics (IPA)

Hi,

Andreas Johansson a écrit :
> So what would be the problem with analyzing an acoustical affricate as two
> phonemes? I've seen _unitary_ phones analyzed as phonemic clusters in a
> variety of languages (eg [S] as /sj/ in Japanese).
I see no problem at all, and one should not pay too much attention to
acoustics (or at least one should consider phonology first). I strongly
believe that phonology is grounded into acoustics, but there is (or must
be) in the mind a core phonological level where there are only formal
objects (whatever they are : articulatory manoeuvres, acoustic
effects...). For instance, french vowels are considered by several
authors (including my director) as being /VN/ underlyingly. Even when
they appear as true nasal vowels as in Standard French (e.g. _cousin_
[kuzE~] "cousin", while Midi French has [kuzE(~)_N]), there are
arguments for considering them as VN sequences : the feminine of
"cousin" is _cousine_ [kuzin(@)], vowels are systematically decomposed
as VN when they are borrowed, even for oral borrowings (there is an
interesting study of the phenomenon within the framework of the Theory
of Constraints and Repair Strategies)... So, acoustics are by no means a
perfect mirror of the phonological pattern of a language.

Julien.

--
"Well be a lot longer discovering the future if we dont recover the
past" John Anderson