On 17/03/06, Mark J. Reed <[log in to unmask]> wrote:
> One thing that I only became aware of after reading JC Wells' book is
> that there are several sounds I had thought of as foreign, non-English
> sounds, that actually occur in my own everyday variety of English!
> This came as something as a shock, but I cannot dispute it.  Careful
> attention to my speech reveals, for instance, that the word
> "cucumber", which I think of as beginning with /kj/, actually comes
> out of my mouth with an aspirated palatal affricate [c_hC], although
> the initial stop feels closer to [k] than [c] - maybe it's [k_j].  And
> while words like "huge" and "human" normally have a real [hj] cluster
> to match my phonemic /hj/,  they likewise occasionally start with [C]
> instead-  perhaps an overcorrection in my desire to avoid the
> to-me-distasteful (despite being historically correct!) pronunciation
> with a bare initial [j].
>
> Then there's the rampant labialization, palatalization, and
> nasalization!  "Queen" may be /kwin/ phonemically, but it comes out as
> [k_w_hi~J_}].  Holy crap!
>
> Anyone else had an experience like this?  Maybe if I'd had more than
> one quarter of linguistics in college I would have had this a-ha
> moment in class instead of when I read the book, but it really sort of
> blew me away.


When I first worked out (some time ago) that vowel length, and not
vowel tension, was phonemically relevant in my dialect, I found it odd
& surprising. Right now I'm pretty sure that /u\:/ (food) and /3:/
(bird) are actually more like [y:] and [2:], distinguished from (one
pronunciation of) "feared" and "bared" primarily by rounding. Also I'm
of the growing opinion that the way I at least say /O/ (hot) is
usually as an unrounded and short counterpart to /o:/ (caught), thus
[7]~[V], rather than height/"tenseness". Certainly all of /u\: 3: U
o:/ are strongly rounded, but if /O/ is rounded at all, it's very
weak.

Most of these changes actually seem pretty reasonable given the
conversion of tense/lax to long/short; the pairs either need to merge
in quality (as did "cud"/"card") or gain an extra distinction, and
+/-round seems as good as anything... (Though, I'll admit, the
exchange of +/-tense for +/-round *and* +/-long seems a little
inefficient.)

Following from the loss of tense/lax, I was also very surprised when I
first realised I couldn't hear the difference between either /I/ and
/i/ or /I/ and /e/ in languages like Dutch and Icelandic, in which
length does not (necessarily) co-occur with tension. More recently
when I've been listening to how, rather than what, is said, I've also
noticed that I have the same difficulty in the speech of some
Americans.

As far as consonants are concerned, /nj/ is more like [n_j] or [J] but
I think that's relatively normal for English. My /k/ is heigher in
"kit" or "*c*ucumber" than in "cut" or "could", but I don't think it's
as far forwards as [c]...

I can hardly[*] distinguish between voiced and unvoiced stops (or [tS]
vs [dZ] rather than the English [tS_h] vs [dZ_h]), word-initially &
intervocalically. That seems to be common amongst my Australian
classmates in my phonetics & phonology class, too. John Cowan once
mentioned on this list I think that the distinction in AusE was indeed
one of aspiration, but I didn't believe him at the time!

[*] That is, I can hear that they are different if you say [pa ba],
but which is which is hard and, I will have great difficulty telling
if you're saying [pa] or [ba] if you're just saying one or the other.
I'll have to be able to readily hear and pronounce the difference by
the end of semester tho!

Phonetic nasalisation is, however, de-rigueur in English and I think
most languages with nasal consonants but not phonemically nasalised
vowels. Just about every English speaker will probably say something
along the lines of [m&~n] for "man", the exact
height/frontness/length/frog notwithstanding.

On 17/03/06, Mark J. Reed <[log in to unmask]> wrote:
> Hm.  [H] is IPA [ɥ], yes?  (Which is damn confusing.  I would have
> made [H] equivalent to IPA [ʜ]; there should be a universal rule of
> IPA->CXS that small caps become simple uppercase.)

I think the rule is actually that with vowels, CXS cap=IPA sc, whereas
with consonants, CXS cap=IPA turned, CXS cap+\=IPA sc. Note especially
[R], which isn't helped by French /r/ being [R] and not [R\] as one
might expect. Assuming there's a small cap and/or turned form. If not
the relationship's obviously arbitrary.

--
Tristan