International Conference on Chinese Language Learning and Teaching in the Digital Age
Young Scholar Award Competition, Hong Kong, 25-27 November, 2011
Towards a Model of Second Language
Word Production and Recognition in Mandarin
Patrick Chun Kau Chu
PhD Candidate
School of Psychology
University of New South Wales, Sydney, Australia
Email correspondence: [email protected]
Website: http://www.patrickchu.net/
1
Aim of the study
• Understand how Cantonese speakers’ production and
recognition of second language (L2) Mandarin words is
influenced by the first language (L1) lexical system
2
Pronunciation relationships
between Cantonese and Mandarin words
• There are many homophones (同音字) in Cantonese and Mandarin
Character / meaning
聲 ‘sound’
Cantonese pronunciation
星 ‘star’
尾 ‘tail’
/sing/
Mandarin pronunciation
/sheng/
美 ‘beauty’
/mei/
/xing/
/wei/
/mei/
‘sound’ 聲音 /shengyin/  /xingyin/
‘tail’ 尾巴 /weiba/  /meiba/
‘tail’
=尾
Concept
L1 Cantonese
phonological
representation
Concept
route
‘beauty’
=美
mei
Concept
route
Lexical route
L2 Mandarin
phonological
representation
Mispronunciation occurrence
wei
mei
0%
25%
mi
3
Pronunciation relationships between
Cantonese and Mandarin (Zhang & Gao, 2000)
Cantonese onset (聲母) /m/
Mandarin
onset
m
w
b
Number
of words
160
33
2
Percentage
82%
17%
1%
Example
媽 ‘mother’
萬 ‘ten thousand’
剝 ‘to shell’
Cantonese
pronunciation
maa1
maan6
mok1
Mandarin
pronunciation
ma1
wan4
bo1, bao1
Cantonese rime (韻母) /ei/
Mandarin
rime
i
ei
in
ü
er
Number
of words
99
26
1
1
1
Percentage
77%
20%
1%
1%
1%
Example
皮 ‘skin’
悲 ‘sad’
您 ‘you’
履 ‘shoes’
餌 ‘bait’
Cantonese
pronunciation
pei4
bei1
neii5
lei5
lei6
Mandarin
pronunciation
pi2
bei1
nin2
lü3
er3
4
L2 Mandarin word production model
‘tail’
=尾
Concept
‘beauty’
=美
L1 Cantonese
phonological
representation
L1 Cantonese
sublexical
representation
mei
Sublexical
route
Concept
route
L2 Mandarin
sublexical
representation
L2 Mandarin
phonological
output buffer
wei
Sublexical correspondence
3%
Mispronunciation occurrence
0%
Lexical
route
m
ei
82%
17%
77%
m
w
i
17%
Lexical
route
Concept
route
ei
mi
wi
63%
13%
mei
14%
25%
Correlation between sublexical correspondence and mispronunciation occurrence = 0.35, p = .005
5
(Chu & Taft, LSHK ARF 2010)
Major tone relationships between
Cantonese and Mandarin (Zhang & Gao, 2000)
Cantonese
tone
1
2
3
4
5
6
Mandarin
tone
1
3
4
2
3
4
Percentage
93%
89%
91%
93%
76%
94%
Example
郊 ‘suburb’
找 ‘find’
怪 ‘strange’
牛 ‘cow’
偉 ‘great’
又 ‘again’
Regular-tone
Cantonese /
Mandarin
pronunciation
gaau1 / jiao1
zaau2 / zhao3
gwai3 / gwai4
ngau4 / niu2
wai5 / wei3
jau6 / you4
Exception
魔 ‘devil’
帽 ‘hat’
傘‘umbrella’
微 ‘little’
市 ‘city’
捕 ‘catch’
Cantonese /
Mandarin
pronunciation
mo1 / mo2
mou2 / mao4
saan3 / san3
mei4 / wei1
si5 / shi4
bou6 / bu3
Irregular-tone
6
L2 Mandarin word production model:
Incorporating the tone component
Regular-tone word: 找 ‘find’
Cantonese: zaau2 Mandarin: zhao3
‘find’
=找
Concept
L1 Cantonese
phonological
representation
‘hat’
=帽
zaau2
Concept
route
mou2
Sublexical route
Lexical
route
L1 Cantonese
sublexical
representation
Concept
route
z
53%
L2 Mandarin
sublexical
representation
L2 Mandarin
phonological
output buffer
Irregular-tone word: 帽 ‘hat’
Cantonese: mou2 Mandarin: mao4
zh
21%
aau
Sublexical route
Lexical
route
m
2
ou
2
18%
61%
30%
89%
5%
82%
17%
49%
47%
89%
5%
j
ao
iao
3
4
m
w
ao
u
3
4
z
zhao3
zao3
jiao3
29%
11%
5%
zhao4
2%
zao4
jiao4
mao4
mao3
mu3
wu3
mu4
wu4
0.6%
0.3%
2%
36%
34%
7%
2%
0.4%
7
Chinese character – Mandarin sound matching task
(Chu & Taft, ICPhs 2011)
• Materials
Word Type
Matched
Mismatched
zhao3
zhao4
>
<
mao4
mao3
Regular-tone words (e.g. 找 ‘find’ Cantonese: zaau2)
Irregular-tone words (e.g. 帽 ‘hat’ Cantonese: mou2)
• Dependent variables - Percentage of ‘yes’ responses
8
Further evidence for the use of the sublexical route
in L2 Mandarin word production
• Mandarin pinyin transcription task (Chu & Taft, EPC 2011)
Sublexical
unit
Cantonese
pronunciation
Onset
Regular word
Irregular word
Mandarin pronunciation
(% correspondence)
Example
Mandarin pronunciation
(% correspondence)
d
d (96%)
對 ‘correct’
t (2%)
突 ‘sudden’
Rime
aai
ai (59%)
帶 ‘bring’
a (2%)
拉 ‘pull’
Tone
Tone 4
Tone 2 (93%)
農 ‘farm’
Tone 4 (4%)
Mean error rate
Example
期 ‘period’
Error analysis of irregular words
9
L2 Mandarin word recognition model
Mispronunciation
魔 ‘devil’ mo2  mo1
Correct pronunciation
魔 ‘devil’ mo2
L2 Mandarin
phonological
input
mo1
mo2
L2 Mandarin
sublexical
representation
m
Concept
L1 Cantonese route
Lexical
route
98%
m
sublexical
representation
L1 Cantonese
phonological
representation
Concept
Sublexical route
Sublexical route
mo1
mo4
23%
‘devil’
=魔
‘grind’
=磨
o
31%
27%
o
ok
mok6
12%
17%
ut
ak
76%
4
mut6
3%
‘don’t’
=莫
m
2
13%
Lexical
route
98%
m
6
mo2
mak6
2%
‘sink’
=沒
Concept
route
‘wheat’
=麥
‘touch’
=摸
1
31%
27%
o
ok
12%
17%
ut
mo1
25%
2%
o
‘devil’
=魔
ak
83%
9%
1
3
mok1
mut3
22%
‘peel’
=剝
1%
‘plaster’
=抹
10
Empirical evidence for the
L2 Mandarin word recognition model
• Disyllabic word transcription task (Chu & Taft, ISB 2011)
魔鬼 ‘ghost’
– Correct pronunciation: mo2gui3
– Mispronunciation: mo1gui3 (nonword 摸鬼)
11
Conclusion
• Three-route L2 word production and recognition model:
– Concept, lexical and sublexical
• Beginning learners
– Use sublexical route in both L2 word production and recognition
• Advanced learners
– Shift from sublexical to lexical/concept route in L2 word production
– Use both sublexical and lexical/concept route in L2 word recognition
• Future research
– Lexical decision with cross-modal priming
– Eye-tracking paradigm with visual-world paradigm
– Event-related potential (ERP)
12