LINGUIST List 3.770

Sat 10 Oct 1992

Disc: Phoneticians II

Editor for this issue: <>

Directory

1. , Re: 3.766 Like, Phoneticians
2. John S. Coleman, 3.766 Like, Phoneticians
3. , 3.766 Phoneticians
4. "Larry G. Hutchinson", Re: 3.766 Like, Phoneticians

Message 1: Re: 3.766 Like, Phoneticians

Regarding Alexis Manaster-Ramer's recent posting on the issue of whether
phoneticians can't or don't hear properties of segments, and in particular,
his citation of the results purporting to show that neutralization is
phoentically incomplete, it is NOT the case that speakers produce
SYSTEMATIC if small phonetic differences between "neutralized" sounds.
In fact, speakers differ from one another and from token to token, which
has been to this phonetician one of the signs that the differences obtained
in these experiments are artifactual. One piece of additional support for
this view is an experiment by Iverson and Fourakis, in _Phonetica_ (1984,
I think -- I'll post a correction if this is mistaken) showing that when
subjects attention was drawn away from the underlying phonological contrast
(by having them do a morphological task that would produce the desired alter-
nations as a side effect), then they did not produce any phonetic differences
between an underlying [+voice] and undelrying [-voice] stop in the neutrali-
zation context. The other piece of evidence is discussed in a recent paper
by Port (and perhaps someone else) in _Journal of Phonetics_ suggesting that
the unsystematic phonetic differences that have been obtained in various
experiments are probably just another instance of the effect that Labov
and his coworkers have observed in excessively formal speech tasks like
reading from lists.

I will check both these references and give full citations to the list
in a subsequent posting for those who are interested.

I would also note that I don't disagree with Alexis's basic point, which I
take to be that expectations about what a sound should be, coupled with
knowledge of the message can distort even a well-trained phonetician's
perceptions. This point was made vis-a-vis intonation by Lieberman in
a paper in _Journal of the Acoustical Society of America_ in the early
sixties (again watch this space for the citation). This implies that
such errors are less likely if one is working a language other than one's
own. While they may be less frequent, I don't think they can be eliminated.
Positive evidence that this is so again comes from studies of intonation,
where Pierrehumbert and Beckman show in their book _Japanese Tone Structure_
(1988) that what native and non-native linguists had described as high
and low level contours in Japanese were actually rising and falling
contours respectively (and that what had been thought of evidence of
tone spreading across unspecified strings wasn't -- instead F0 is
interpolated between targets).
Best regards,
John Kingston

Alexis Manaster-Ramer asks:
> instrumental findings which show that
> words 'lucky' and 'bugger' have intervocalic fricatives seem
> to me quite consistent with the fact that pretty decent phoneticians
> often describe these as stops, namely, these fricatives are utterly
> different from a good ach-laut or gamma. Although here, too, it
> would be nice to know what exactly has been claimed about these
> phenomena. For example, have there been phoneticians who noticed
> them WITHOUT the aid of instruments?

Yes, numerous cases of this kind of lenition in English have
been reported by ear-trained phoneticians on the basis of
listening. Some examples and references may be found in
Simpson's latest paper in Linguistics 30, pp. 535-548.

The audible differences between devoiced final voiced stops
and voiceless stops in various languages has also been
extensively noted by ear-trained phoneticians long before
the recent flurry of papers on this subject in the instrumental
phonetics literature. Instrumental phonetics research by and
large serves to quantify phonetic phenomena, not reveal their
existence.

 --- John Coleman

 ---(Forwarded from: Alexis_Manaster_RamerMTS.cc.Wayne.edu, Dated: Fri, 9 Oct
 92 12:30:42 EDT)---

I think this will be my last response to John Coleman's constant
attacks on "cloth-eared" phoneticians. Just three basic points.

One, the whole
pointless discussion started with a simple query I posted about
whether anyone knew of published work that tends to show that
phoneticians, no matter how good, cannot hear certain distinctions
which languages nevertheless make. Whatever John thinks about this
issue, I am surely allowed to ask whether such literature exists.
Why try to shout down a simple request for information?

Two, John keeps making general statements as well as allusions to
nameless people. However, I do not see any references to studies
that show that phoneticians (or ANY human beings) can hear the
distinctions which have been claimed to show up in instrumental
studies of, for example, Polish, Russian, German, and Catalan
final underlyingly voiced vs. underlyingly voiceless obstruents.
On the other hand, as I have noted, there is extensive literature
that reports such instrumental results and strongly suggests, if
it does not say explicitly, that these distinctions are not
expected to be heard by either speakers or phoneticians.

Three, I should perhaps point out that the literature on these
instrumental findings has been roundly criticized, and there is
serious doubt as to whether the instrumental findings are even
correct (although I lean towards accepting them at least for
some of the languages mentioned). Now, it seems to me that if
instrumental phoneticians like Jassem and Fourakis cannot even
replicate the finding that these distinctions exist in the first
place,
then it is doubly doubtful that John's paragon non-cloth-eared
phoneticans can hear them.

Message 4: Re: 3.766 Like, Phoneticians

It's obvious that instruments can detect what our unaided senses cannot -
that's why we make them. But phonetic differences too small to be detected
by even trained phoneticians surely cannot be detected by the "normal"
language user either, and hence are not of LINGUISTIC interest. (Bloom-
field's point, I take it.) Such differences are below the level of
"systematic phonetics", to use a term used by Chomsky a long while back.

Chomsky himself got confused on this issue in 1964. He claimed that /k/ and
/p/ showed phonemic overlap, since INSTRUMENTS showed that the noise
perceived as [k] before some vowels was "identical" to the noise perceived
as [p] before other vowels. Even if this is so, he didn't lay a glove on
Bloch, whom he was criticizing on the issue of overlap and invariance.
It is the PERCEPTION of reality that phonolgy is about, and the same is true
for (linguistic) phonetics.