LINGUIST List 5.566

Thu 19 May 1994

Sum: Tone change

Editor for this issue: <>

Directory

1. Enid Wai-Ching Mok, Sum:tone change

This is a summary of all the responses I received from subscribers to the
Linguist List and Chinese Linguist List to my query on tone change. The
same summary is sent to both lists.

I'd like to thank the following individuals for their great help:

 * Halvor Eifring
 * Hua Lin
 * David Prager Branner
 * Wang Shi-Ping
 * Eugene Shing Chan
 * Wenchao Li
 * Lee Bickmore
 * John Goldsmith
 * Chilin Shih
 * Woody Mott
 * San.Duanmu
 * Bamba Moussa
 * Steven Blackwelder
 * Laura L. Koenig
 * Gregg Kinkley
 * Moira Yip
 * David Odden

*From: Halvor Eifring <halvor.eifringeasteur-orient.uio.no>
Norwegian has something that resembles tone neutralization, though very
different in character from the Chinese case.
Norwegian has a distinction between tonemes 1 and 2 in polysyllabic words
only, in monosyllabic words there is no toneme distinction, and for
phonetic and other reasons, one often reckons that all monosyllabic words
have toneme 1.
The vast majority of words, however, has only one accentuated syllable per
word, i.e. one syllable on which the toneme distinction may be realized.
This results in what I'd call tone neutralization in e.g. compounds.
The standard example of the toneme opposition in Norwegian is the
distinction between /1boner/ 'farmers' and /2boner/ 'beans' (the /o/ is
actually a roundeddistinction between /1boner/ 'farmers' and /2boner/
'beans'
(the /o/ is
front high-mid vowel, like German o with an umlaut). If these words occur
as the second element in compounds however, the toneme distinction is lost,
e.g./2he:dmarks,boner/ may mean either 'farmers from Hedmark' or 'beans from
Hedmark'. The first syllable /bon-/ still receives some stress (marked by
me with a comma), but apparently not enough to keep the tone distinction
intact.
Minnan dialect (Hokkien) also is supposed to have a "neutral tone". I don't
know much about it, but you can look it up in Robert L. Cheng and Susie S.
Cheng: Phonological Structure and Romanization of Taiwanese Hokkian [sic!],
Student Book Co., Taipei 1977 (written in Chinese) pp. 151ff. and, I'm
sure, many other places.

*From:Hua Lin <LINGHLUVVM.UVic.CA>
I am sending you a copy of my dissertation
which deals with some of the issures raised
in your questions about tone and tone reduction.

*From: David Prager Branner <charmiiu.washington.edu>
Tonal behavior in Chinese is beginning to be described with some
thoroughness, but all attempts to find universal rules to describe it
have been embarrassingly unsuccessful. I believe there is a short paper
by Anne Yue-Hashimoto in the Wang Li memorial volumes that attempts to
survey tone sandhi behavior and draw some general conclusions. But
absolute rules for explaining tone sandhi seem to me to be a long way
off, if not actually impossible.
Also, the relationship between tone and vowels has been explored to some
degree for isolated languages, such as Foochow (Fwujou, Fu2-zhou1, etc.),
which has been studied by Marjorie Chan (see her PhD from the University
of Washington, 1987?) and dialects spoken near Shining (Xi1-ning2, etc.),
studied by Liou Shiun'ning.
Tone neutralization is very common in Chinese dialects, but not well
described.

*From: Wang Shi-Ping <wspbdc.com.tw>
I once presented a paper about tonal neutralization of Taiwanese in NACCL 4,
Ann Arbor, 1992. Taiwanese tonal neutralization is a kind of tonal reduction
suggested by Dr. Larry Hyman (p.c., 1990 in UC-Berkeley). You might want to
ask him for the definition. I also did some experiments. Given the pitch
tract, sometimes it goes down to the bottom line 'flatly.' Sometimes the
contour disappears, which it becomes toneless. In my papers (1990,91,93), I
also proposed a high vowel /i/ is deleted when it bears neutral tone in the
case of 'I say ___ one time.' (gwa-kong____tsIt-pai) The high vowel /I/ is
sometimes deleted. It undergoes free variation. As for the interaction of
tone and stress, I also mention it occurs in the case of Taiwanese tonal
neutralization in my draft paper. I haven't got time to revise it.

*From: Eugene Shing Chan <eugenesits.iiu.MY>
I've been away from Chinese linguistics for many years while working
abroad, so I don't know the "latest" references.... however, I did do
some graduate work in your area a long time ago under Dr. William Leben
who specialized in tone phenomena. Check the linguistics bibliographies
for his work. He studied Mandarin, Thai, Yoruba and several other African
langauges as well as compared them to tone-accent systems such as
Serbo-Croatian and Japanese.

*From: Wenchao Li <wclivax.ox.ac.uk>
I can't claim to be an expert on tones, in fact, my research is on Mandarin
phonology exclusive of tonal phenomena. But in my search for structural
similarities between Beijing Mandarin and other Chinese dialects, in
which I've
looked at something like a hundred dialects, I think I can pretty much
confirm
that tonal neutralization does indeed occur in dialects other than Beijing
Mandarin. Most dialects in the Nothern Mandarin family have it. For
information regarding other dialects, try the Chinese Journal FANGYAN
(dialects).

*From: Lee Bickmore <LB527%ALBNYVMS.bitnetUACSC2.ALBANY.EDU>
There was actualy a BLS parallel session on tone which addressed many of
your questions. Try writing Larry Hyman to see if there is an official
proceedings of that event. I'm currently revising a ms. on "Tone and
Stress in
Lamba." Basically, you build trochaic feet across a certain stretch of the
verb. Then *if* there is a (floating) High tone in the verb, it will
associate
to the heads of the feet. When I finish the revised version (that I'm sending
to Phonology) I'd be happy to send you a copy if you like. I think it's
one of
the clearest cases of tone and stress interaction.

*From: John Goldsmith <gldsmthbloomfield.uchicago.edu>
The questions you ask about tone have been discussed
in hundreds of articles and scores of books: it's hard
to tell you where to start. My own work has primarily
concerned African tone languages, though I've also looked
at Mesoamerican and Asian tone languages.
Here's some references of mine:
Autosegmental and Metrical Phonology. 1990. Oxford: Basil Blackwell,
Ltd.
Autosegmental Studies in Bantu Tone, ed. by G.N. Clements and John
Goldsmith. Dordrecht: Foris Press, 1984.
Autosegmental Phonology. 1979. Garland Press. Published version of
MIT dissertation, 1976; also circulated by Indiana University
Linguistics Club, 1976-date.
Tone and Accent in Llogoori. In The Joy of Syntax, edited by D.
Brent*ari, G. Larson, and L. McLeod. John Benjamins.
Tone and Accent in Xhosa (with Karen Peterson and Joseph Drogo).
Current Approaches to African Linguistics (vol. 5), ed. Paul Newman
and Robert Botne. Dordrecht: Foris Publications.
Prosodic Trends in the Bantu Languages. In Autosegmental Studies in
Pitch Accent, edited by N. Smith and H. van der Hulst. Dordrecht:
Foris Publications.
The KiRundi Verb. With Firmard Sabimana. In Mod
edited by Francis Jouannet, pp. 19-62. Paris: Editions du CNRS.
The Rise of Rhythmic Structure in Bantu. Phonologica 1984, ed. W.
Dressler. Pp. 65-78 Cambridge: Cambridge University Press.
Tone and Accent and Getting the Two Together. Proceedings of the
Thirteenth Annual Meeting of the Berkeley Linguistics Society, edited
by Jon Aske, Natasha Beery, Laura Michaelis, and Hana Filip.

*From Chilin Shih <clsresearch.att.com>
Lots of works on intonation include experiments on
the interaction of vowel identity and pitch height,
referred to as intrinsic pitch.
The majority, if not all, intonation works mention
the interaction of stress and intonation, ranging from
simply mentioning the expansion of pitch range under
heavier stress, to prediction exactly how and how much
the pitch range expands.
Tone reduction is common is non-mandarin dialects.
Though in most of the cases I know of, the realization
of reduced tone is low. The kind of sensitivity to
the previous tone as in Mandarin is rare. However, I think
the Mandarin realization is fairly close to having a
single mid (referece line, or default value) target
near the end of the syllable. THe single target gives the
context sensitive tone shape.
Tonal reduction often changes a contour tone to a level tone.
But not all such changes much be considered reduction.
Tonal reduction is typically a by-product of duration reduction,
so you don't have time to realize to full tonal targets of
a full Mandarin tone. Tonal reduction could also change a
level tone into a "contour" tone: that's the property
of having a single tone target with differnt height specification
with it's environment.
It's possible for L->H change IN THE H CONTEXTS
to be a reduction phenomenon, I'll send you my paper on that.
A cautious note here. I said that tonal reduction
often takes the form of contour->level, but I stress that
not all contour->level are reduction. Likewise,
reduction COULD be level->contour. Whether it
is an reduction or not depends more on the duration,
amplitude (typical reduction stuff), than on the pitch
contour alone.

*From: Woody Mott <mottuhunix.uhcc.Hawaii.Edu>
I don't know if this qualifies, but Bao an
Hakka's fourth tone goes high level from high falling except before the low
second tone and in final position.

*From: San.Duanmuum.cc.umich.edu
In my paper 'Rime length, stress, and association domains', Journal of
East Asian Linguistics 2:1-2:1-44, 1993, I discussed some of the questions
you asked (but not all). In particular, I discuss the interaction between
stress and tone. In addition, tonal neutralization, in the ordinary
sense, occurs in Shanghai and other Wu dialects much more extensively than
in Mandarin. It also occurs in Lhasa Tibetan. Best wishes,

*From: BAMBA MOUSSA <k20520er.uqam.ca>
If you can read French, I will send you
a copy of my Ph.D. dissertation written in 1992 at the Universite du
Quebec a Montreal (UQAM). The thesis is entitled "De l'interaction entre
tons et accent". It proposes a metrical approach of tone languages. If
you are interested, let me know.

*From: Steven Blackwelder<sblackwelderfirstbyte.ccmail.compuserve.com>
 Most of what I know about tone neutralization came from my
 Mandarin-learning days, and I'll repeat it below at the risk of
 restating what you already know.
 The only serious descriptions of tone neutralization I've personally
seen
 are for standard Mandarin.
 The oldest ones I know are two of Y. R. Chao's most famous works, the
 first of which I've seen and the second I haven't (but read further):
 _A_Grammar_of_Spoken_Chinese_ (1968, Berkeley, CA: Univ. of California
 Press)
 _Mandarin_Primer_ (1964, Cambridge, MA: Harvard Univ. Press).
 _Chinese_Primer_, a mid-1980's publication from Princeton Univ.,
 claims to be the rewrite of _Mandarin_Primer_. It was the textbook
 for my first-year Chinese course at UCLA in 1986-87 (from fall '87 to
 the present UCLA has been using the Beijing Lang. Inst.'s textbook
 for intro.-Chinese courses).
 During my year-abroad at Beijing Univ. I took the introductory
 linguistics course in the Chinese Language and Literature department.
 During the phonetics/phonology portion of the course, the instructor
 mentioned that the historical development of tones may have been
 influenced by the re-interpretation of ancient voicing contrasts in
 syllable-initial stop consonants.
 For an engineering spin on Mandarin tones, you might profit from reading
 Lin-shan Lee _et_al_, "Improved Tone Concatenation Rules in a
 Formant-Based Chinese Text-to-Speech System,"
 _IEEE_Transactions_on_Speech_and_Audio_Processing_, Vol. 1, No. 3, July
 1993, pp. 287-294.
 With regard to tone neutralization outside Mandarin, I have a
 non-speaker's impression that something like that happens in Shanghai
 dialect.

*From: Laura L. Koenig
Yi Xu just did a dissertation (UConn and Haskins Labs) on tonal
coarticulation
(or, context effects) in Mandarin. Some of that was just published in the
latest issue of JASA. I realize You wanted stuff on lgs other than Mandarin,
but maybe it would still be of some use to you.
Xu, Yi (1993). Contextual tonal variation in Mandarin Chinese. Ph.D.
Dissertation, University of Connecticut. [Dept. of Linguistics]
Xu, Yi (1993). Production and perception of coarticulated tones. Journal
of the Acoustical Society of America 95(4):2240-2253.
The address of the lab is:
Haskins Laboratories
270 Crown Street
New Haven, CT 06510

*From: Gregg Kinkley <gkinkleyuhunix.uhcc.Hawaii.Edu>
I studied tonogenesis
across most of the language families of East and Southeast Asia, so I am
very interested in your questions. I must confess that I may not be
familiar with some of your terminolgy (probably because I have been out
of the literature for a while!) but at the risk of giving you a red
herring, at leat one of your questions (on vowels and tone) makes