Re: (Brazilian Portuguese and Rhodrese (was French)

Hi!

Well, let me try to rephrase everything more explicitly. My books are all in
boxes still for two weeks... so let's see if I remember it right...

/r/

1. In latin there were just /r/ and /r:/... I'm not sure /r:/ is just
allophone of /r/.
2. In medieval portuguese they wrote initial /r/ as <rr>, so <rrey> for
[r:ej] 'king'. Initial /r/ and medial /rr/ merged, still as a long alveolar
trill.
3. I've heard our [R] came from the upper classes trying to imitate french
(like educated romans saying [hOma] for /rOma/ to imitate greek), but I have
to read more to be sure. Anyhow, [r:] became [R].
4. A very recent development is that this [R] is now pronounced [h], at in
Brazil's south.
5. Also, /r/ before consonant and in auslaut (VrC, Vr#) is heard like a
retroflex approximant. It is more or less a shibboleth, as it's more usual
to hear from people caming from the countryside. It's curious to hear on the
television journalists trying to fake their accent... nervously, sometimes.
6. Just to be complete, /r/ from infinitives is being syncopated, starting
with those before vowel:  'cantar uma música' [k@~ta uma muzik@]
versus [k@~taruma
muzik@]; then after consonant... rarer and sounds like careless speech:
'falar tudo' [fala tudu].

/l/

1. Once it was [l] everywhere.
2. Then those before vowel, in auslaut, came to be [w] in pausa or before
consonant, but preserved followed by a word starting in vowel: 'leal' [leaw]
and 'alma' [awma], but 'leal amigo' [lealamigu]. I still hear it from people
with 70 or more, but now it's just [leawamigu].
3. Like the /r/ item 5, those [l] before consonant and in auslaut, in
countryside speach, came to be a retroflex approximant. So 'alma' and 'arma'
sounds alike in this variety. This is so stigmatized that I rarely hear it
naturally in the wild.

Answering you more concisely...

1. an /r/ which is [R] or [h]?


Yep, but [h] is just a development of [R]. where there is [h], no [R] is
heard anymore.

2. an /l/ which is [w] or [r]?


[w] or the retroflex approximant, don't know how to write it in ASCII, but
not in the same variety of brazilian portuguese... where there is [w] for
/l/, ideally there should not be... the other, but television is mixing
everything.


> 3. free allophones which are possibly splitting
>   into phonemes thru semantic differentiation?

I would really like it, if this was cristalized as a morphological rule,
like an augmentative just by changing one phoneme... for now it is just...
novelty, and I'm afraid it will not last...


> also, is /l/ ever [l]?

In anlaut, intervocally:

lado [ladu]
alado [aladu]


Let me just take a breath before reading about Rhodrese with the care due...
now my head aches ; ).

Edgard.

Replies