Re: why can't the BBC

"Nathan Sanders" <nsanders@xxxxxxxxxxxx> wrote...
"Peter T. Daniels" <grammatim@xxxxxxxxxxx> wrote:
On May 23, 11:46 pm, Nathan Sanders <nsand...@xxxxxxxxxxxx> wrote:
> In article <BVKZj.4002$IK1.3...@xxxxxxxxxxxxxxxxxxxxxxxxxx>,
> "John Atkinson" <johna...@xxxxxxxxxxx> wrote:
> > "Nathan Sanders" <nsand...@xxxxxxxxxxxx> wrote...
> > > "Peter T. Daniels" <gramma...@xxxxxxxxxxx> wrote:
>
> > >> The contrasting reduced (which is different from unstressed) > > >> vowels
> > >> in
> > >> AmE are /+/ and /@/, Russias (@) ~ rushes (+).
>
> > > They aren't truly contrastive in a strict sense. Note that the
> > > morpheme boundaries in this pair are different: Russsia-s > > > versus
> > > rush-es. [@] can only show up stem-finally, while [+] appears
> > > elsewhere.
>
> > > Since this distribution is complementary and fully regular, > > > then they
> > > are better treated as allophones of the same phoneme.
>
> > I'm not equipped to express any opinion whether "@" and "+" in > > words
> > like this are different phonemes or not in Am E (so I won't), but > > why
> > does Peter write it as /+/? Is there any reason not to identify > > it with
> > /I/, the same way as he identifies "@" with /V/?
>
> According to Flemming and Johnson's measurements, the mean F2 of > this
> sound is around 1922 Hz, which means it is slightly fronter than > schwa
> (1797 Hz), but definitely not as front as [I] (2293 Hz).
>
> However, it does vary a lot (s.d. of 121 Hz), and the actual > frontness
> is largely dependent on the surrounding consonants.

If the two sounds are contrastive, it doesn't matter if they differ by
a single Hz.

If they only differ by a single Hz, they won't be contrastive!

I doubt if there's any human alive who can hear a 1 Hz difference in two sounds at around 2000 Hz. (Unless they're pure tones, and are sounded together for long enough to get a decent beat going.)

J.

.

Relevant Pages

  • Re: why cant the BBC
    ... They aren't truly contrastive in a strict sense. ... morpheme boundaries in this pair are different: Russsia-s versus ... are better treated as allophones of the same phoneme. ...
    (sci.lang)
  • Re: and in Argentian Spanisg (was: Chinese languages or language?)
    ... >>> The issue has also been discussed, numerous times and ad nauseum, in ... If two people differ as to whether the words are the same, ... > So do we see one phoneme here, or two, based on this experimental ... it doesn't matter, further information is required: ...
    (sci.lang)
  • Re: and in Argentian Spanisg (was: Chinese languages or language?)
    ... >> group:nl.taal, but in Dutch. ... >Doesn't matter. ... If two people differ as to whether the words are the same, ... So do we see one phoneme here, or two, based on this experimental ...
    (sci.lang)
  • Re: why cant the BBC
    ... morpheme boundaries in this pair are different: Russsia-s versus ... are better treated as allophones of the same phoneme. ... If they only differ by a single Hz, ...
    (sci.lang)