LINGUIST List 14.1516

Mon May 26 2003

Qs: Cantonese/ESL; Transformation-Based Learning

Editor for this issue: Naomi Fox <foxlinguistlist.org>

Directory

1. Wai-ming Tsui, Cantonese ESL learners' perception/production
2. Torbj�rn Lager, Transformation-Based Learning Bibliography

Dear all,

 I am currently working on the acquisition of English plosives by
Cantonese ESL learners. I'll be grateful to anyone who is willing
enlighten me on the question: Why, as Cantonese has the distinction of
[+/-aspirate], can't Cantonese ESL learners be able to perceive and
produce the English word-medial plosives (preceding an unstressed
syllable) as unaspirated?

 English word-medial voiceless plosives preceding an unstressed
vowel are unaspirated. For example, for a word like pepper, the second
/p/ is not aspirated. In Cantonese, we have the contrast of aspirated
and unaspirated plosives, as in pin [phin] 'to edit' and bin [pin]
'edge' (transcribed in Cantonese Romanization scheme Jyutping and
IPA). I believe it is reasonable to assume that, since Cantonese has
the contrast of aspirated/unaspirated plosives, Cantonese ESL learners
are able to perceive, and subsequently produce, the English
word-medial voiceless plosives as unaspirated, rather than
aspirated. However, the fact is that most of them, including myself,
perceive and produce the plosives in question as aspirated.

 One possibility is that it is just the analogy from the variant of
the English plosives, as they are aspirated when preceding a stressed
vowel. This is unlike be the case because we produce the plosives
after /s/ as unaspirated, which is the same as the native speakers'
production.
 
 Further, if the word-medial plosive is [+voice], no aspiration is
present in Cantonese ESL learners' production. By the same token, they
do not perceive any aspiration there. In their production, it seems
that the duration of the preceding vowel is longer than the case that
if the plosive is [-voice].

Million thanks
Ming. 


- ------------------------------ TSUI, Wai-ming, 

Postgraduate student, 
Dept of Modern Languages and Intercultural Studies, 
Chinese University of Hong Kong. 

Hi All,

For a couple of years now, I've been compiling a bibliography of papers 
about the theory, use and implementation of Transformation-Based 
Learning (a machine learning method, invented by Eric Brill, which have 
been used to learn rules for many natural language processing tasks).

The relevant URL is:

 http://www.ling.gu.se/~lager/Mutbl/bibliography.html

I'd like to keep the bibliography up-to-date, so if you can suggest 
corrections and additions, please mail them directly to me, and I see to 
it that they get included.

Thanks,
Torbj�rn Lager