Re: Schwa and [V]: Learning the IPA

On 15/06/06, Larry Sulky <larrysulky@...> wrote:
> On 6/15/06, R A Brown <ray@...> wrote:
> > It is such statements made by north Americans in past threads that have
> > certainly given the impression that the two *phonemes* /@/ and /V/ have
> > fallen together in many parts of North America; indeed, your statement
> > above only confirms this.
>
> Ooooookay. I really misunderstood. I thought we were talking about [@]
> and [V] rather than /@/ and /V/.
>
> But I still think I disagree. If someone tells me that they'll [kVt]
> something, I figure they'll use a  knife to do it. But if they say
> they'll [k@t] something, I have no idea what they're talking about.
> For me the difference between [@] and [V] is usually phonemic. (I
> would probably recognise /b@t/ in unstressed position as "but", but
> not if stressed. And if someone pronounced the unstressed vowel in
> "hobbit" as [V], I would recognise it, but consider the pronunciation
> odd.)
>
> Now, I can't think of any minimal pairs for these two vowels. Would
> that make the distinction non-phonemic, even though I recognise [kVt]
> and [nVb] as words but not [k@t] and [n@b]?
Neither of those are sufficient conditions for either. For instance,
/h/ and /N/ do not contrast in English, but they are different
phonemes. On the other hand, I recognise [fu\:d] "food" and [fu:l]
"fool" as words, but not [fu:d] or [fu\:l] (other than as
funny/accented pronunciations ... actually, it occurs to me know that
when I try for [fu\:l], I hypercorrect and actually get [fy:l]...).
Worse, my normal pronunciation of /@u\/ is something like [Vu\],
whereas before /l/ it's retracted to [Ou]: The American pronunciation
[ou] sounds very much like this, and so when I unexpectedly hear an
Amercian say /ou/, I hear it as /@ul/ and misunderstanding sometimes
ensue!

So ... Even though use two different allophones for /V/, such as [@]
and [V], and you can tell them apart: This doesn't mean they're
separate phonemes.

On the other hand, even though they are in a distribution that could
be described as complementary: This doesn't mean they're a single
phoneme.

In a language like English, where /@/ basically occurs where most
other (short) vowels do not, I don't really consider it relevant to
consider whether it contrasts with a particular (short) vowel. But I
understand the intuition of some Americans is that it *doesn't*
contrast, and they feel that they are the same vowel (I assume this is
because /V/ has merged into /@/, and /@/ as something very much like
[@] now occurs in both stressed & unstressed syllables). It fact, an
American friend of mine has informed me that
introduction-to-linguistics stuff aimed at Americans tries really hard
to get Americans to spell /@/ and /V/ with separate vowels, when spelt
phonemically.

Soo..... My view on the matter is that if you feel [@] and [V] are
different in any way, be conservative and use separate vowels in
phonemic transcriptions. If, however, you feel they are the same, then
write them so. IOW, follow your intuition in this case.

By contrast, in Australian English there is an "allophone" of "/@/"
which is exactly identical to the normal pronunciation of "/a/", being
[a_"]~[6]. It's used before a (major or minor) intonation break, at
least; generally, it tends to be closer to this in word-final
syllables.* So the recital pronunciation of "filler" is [fIla_"],
versus "fillers", which is [fIl@z]. Now, it's my intuition, and the
intuition of most people I've been able to get an intuition out of in
various non-scientific ways, that this "phoneme" is in fact /a/. (I
even have to check up in a dictionary the phonemic form of words
ending in -ar (/@/ vs /a:/), and was *very* surprised to learn that
"thorough" is considered to end in /@/, not /a/; the two vowels
definitely feel basically the same.) But calling it thus completely
screws up most analyses of English vowel distributions ... Hence,
given I consider the notion of phonemes not terribly useful anyway, I
just follow the standard transcription/phonemicisation. But then my
recommendation that you follow your intuition should be taken with a
grain of salt, or some other similar caution!

[*]: Some affixes stop this, so "waters" generally has [@]; but if the
preceeding vowel is /a/ or /a:/, the affect of the affix is lessened.

(As a brief aside, I've read that the motivation for the quite
different IPA for Australian English compared with American and
British English (most notably, using something like <a> or <6> for
AmE/BrE <V>) is because an "Australian Phonetic Alphabet" was
developing in which linguistics students, when transcribing foreign
languages, would typically use <V> for a low central vowel, in spite
of being taught otherwise! I've also noticed <I> being used when <i>
is expected also, but this is not changed: Possibly because the higher
vowel in /I/ is a much more recent change. Personally I think it would
be at least as productive, given that /i:/ is nothing like [i:]. It
would probably aid teaching the actual pronunciation of [i] no end!)

(Read the stuff between angle brackets as the IPA vowel the CXS
transcribes, not the literal value.)

...
> > The question was whether or not /@/ and /V/ are separate phonemes. It
> > has become apparent to me following similar threads over many years on
> > this list (look in the archives) that in some varieties of north
> > American English, at least, the two _phonemes_ have fallen together,
> > Indeed, your mail surely confirms this.
>
> I just don't see how it does. Sure, there are probably some dialects
> where it has, but they are surely in a small minority, and my
> impression at the outset of the thread was that it was being claimed
> that these phonemes had coalesced for _most_ North American English
> dialects.
It is my impression as a non-speaker of North American English who has
never been to North American that the claim is indeed that it has
coalesced for *most* NAE dialects.

...
--
Tristan.