Re: What _is_ rhoticity? (wa laterals (was: Pharingials etc))

>Just a possible way out of this morass:  Let us suppose that all rhotics
>originally start out as tapped or trilled [r], with all the acoustics
>thereof-- this seems likely on the basis of Romance (trill in Ital, Span,
>dialectal trill vs. uvular in Fr., trill in Portugal vs. vl.velar fric in
>Brazilian), German (dialectal trill vs. velar?/uvular), and Slavic (from
>what I'm told); even in Indonesian langs. (Ml/Indo trill, dial/related
>langs. vd or vl. velar fric.) etc. etc.----- so however a language's /r/
is
>realized, there is a sense that "this is underlying phonemic /r/" and the
>speech organs act accordingly, producing lowered 3d formant (or whatever
it
>is that marks a "rhotic"...).  Thus I suspect a language could have an
>uvular _rhotic_ fricative in addition to a non-rhotic one, and there would
>be no confusion because the non-rhotic one "comes from" a different place
in
>phonological/psycological space.
>
>A vd. velar fricative can be rhotic; but both Span. and Germ. have them
(as
>allophones of /g/) and I'm sure they're not "felt" as rhotics.
No, the thing doesn't work exactly that way. I find
that all other rhotic sounds sound to me as if they were
"alterations" or "corruptions" of an "ideal" alveolar
trill (_the_ rhotic), e.g. the tap looks to me as if
it were a "clipped-down" rhotic, the uvular trill as a
"swallowed-up" rhotic, English r sounds as a "darkened"
and "halfly-vowelized" rhotic (or "fully-vowelized"
when in rhotic vowels), Czech r^ as a "sibilant" and
"palatalized" rhotic, the bilabial trill as a "labialized"
rhotic, etc. But that's not because I feel there's an
underlying "rhotic" phoneme in those sounds, it's a
purely aural thing: I can assure you Spanish /X/ is not
thought of as any kind of "rhotic" phoneme at all, but
as I pointed out, sometimes I _do_ hear it pronounced
clearly as a rhotic [X`] (I can record a sample of this
pronunciation for you to see what I'm talking about).

Also, many pronunciations of French /R/ I hear clearly
non-rhotic, even though I'm perfectly aware that
phonemically they are considered r's by French speakers.
I hear that sometimes they come out as actual rhotic [R`]
but not infrequently the actual pronunciation comes out
merely as a voiced version of a non-rhotic [X] (or even
as [X] itself in the devoiced allophone). In fact, it
seems that rhoticizing too much your r's is not "trendy"
in current French and the more phonetically de-rhoticized,
the better. The pronunciations as uvular plosive trills
[R\] -let alone as alveolar ones-, where rhoticity is
most prominent, are already considered substandard or
'hickish' by at least some people; I've heard of a
woman who felt embarrassed of her pronunciation because
her native French dialect used [R\]. It wouldn't
surprise me that not too long from now the only trace
of rhoticity left in standard French r lies in its
etymological spelling just like in English coda r's
in non-rhotic dialects.

But, the clearest proof that it is not an underlying
phonemicity what makes me think of a sound as "rhotic"
is the example of the Haida recording: I was _expecting_
a totally non-rhotic sound, because the spelling "awáa"
was clearly telling me that the phoneme there was a /w/
and not any kind of /r/, but what I heard was clearly a
rhotic sound, I have no doubt at all about that, I heard
the "pulsiness" of the sound with disturbing crystalline
clarity: it was a voiced labiouvular trilled approximant
in all its glory. So disturbingly that I thought I had
clicked on the wrong place or the sound files had been
swapped. It was not until I heard the other files and
other recordings of Haida that I realized the weird
rhotic [w`] was simply a usual allophone of Haida /w/.

Cheers,
Javier